Citation
The potential of Arzama densa (Lepidoptera: Noctuidae) for the control of waterhyacinth with special reference to the ecology of waterhyacinth (Eichhornia crassipes (Mart.) Solms)

Material Information

Title:
The potential of Arzama densa (Lepidoptera: Noctuidae) for the control of waterhyacinth with special reference to the ecology of waterhyacinth (Eichhornia crassipes (Mart.) Solms)
Creator:
Center, Ted Douglas, 1947-
Publication Date:
Language:
English

Subjects

Subjects / Keywords:
Crops ( jstor )
Eggs ( jstor )
Infestation ( jstor )
Insects ( jstor )
Larvae ( jstor )
Leaves ( jstor )
Parasites ( jstor )
Plants ( jstor )
Rhizomes ( jstor )
Species ( jstor )
City of Gainesville ( local )

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
This item is presumed in the public domain according to the terms of the Retrospective Dissertation Scanning (RDS) policy, which may be viewed at http://ufdc.ufl.edu/AA00007596/00001. The University of Florida George A. Smathers Libraries respect the intellectual property rights of others and do not claim any copyright interest in this item. Users of this work have responsibility for determining copyright status prior to reusing, publishing or reproducing this item for purposes other than what is allowed by fair use or other copyright exemptions. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder. The Smathers Libraries would like to learn more about this item and invite individuals or organizations to contact the RDS coordinator (ufdissertations@uflib.ufl.edu) with any additional information they can provide.
Resource Identifier:
22384888 ( ALEPH )
2342582 ( OCLC )

Downloads

This item has the following downloads:


Full Text






THE POTENTIAL OF ARZAMA DENSA (LEPIDOPTERA: NOCTUIDAE)

FOR THE CONTROL OF WATERHYACINTH WITH SPECIAL

REFERENCE TO THE ECOLOGY OF WATERHYACINTH

(EICHHORNIA CRASSIPES (MART.) SOLMS)














By

TED DOUGLAS CENTER










A DISSERTATION PRESENTED TO THE

GRADUATE COUNCIL OF THE UNIVERSITY OF FLORIDA

IN PARTIAL FULFILLMENT OF THE REQUIREMENTS

FOR THE DEGREE OF DOCTOR OF PHILOSOPHY


UNIVERSITY OF FLORIDA 1976










ACKNOWLEDGEMENTS


I wish to thank the numerous individuals who have assisted in these studies.

I would like to express my appreciation Dr. E. E. Grissell,

Dr. E. L. Todd, Dr. R. E. Woodruff, Dr. T.J. Walker, Dr. R. Carlson,

and Dr. C. W. Sabrosky for the identification of insect specimens; Dr. G. E. Allen, L. P. Kish, and Dr. E. I. Hazard for diagnosing insect diseases; C. Cagle, C. Siebenthaler, M. White, G. Presser, N. R. Spencer, and D. Butler for field and technical help; the University of Florida Soils Laboratory for analysing water samples; Dr. E. A. Farber for providing solar radiation data; the U.S. Department of Agriculture and the Florida Division of Plant Industries for providing space and facilities; Ann Owens and Susan Kynes for library and literature research assistance; Cath Siebenthaler for typing and editorial assistance in the original manuscript; N.R. Spencer and T. C. Carlysle for photographic and dark room assistance; and my graduate committee, Dr. D. H. Habeck, Dr. T. H. Walker, Dr. R. I. Sailer, Dr. G. E. Allen, and Dr. J. Reiskind

for critical reading of the manuscript.

I would especially like to thank Mr. Neal R. Spencer for providing space and facilities and the U.S. Army Corps of Engineers for providing funds.

I would also like to thank my wife, Debbie, whose patience and endurance saw me through to the conclusion of this work.

ii









TABLE OF CONTENTS


Page

ACKNOWLEDGEMENTS. .................. . . ii

LIST OF TABLES ................. . . .. ii

LIST OF ILLUSTRATIONS .... ... .. ...... .. . ix

ABSTRACT .. . . . . . . . .. .. . . . xiv

INTRODUCTION . . . . . . . . . . . . . 1

LITERATURE REVIEW . . . . . . . . . . . 3

Eichhornia crassipes. ................. . . . 3

Taxonomy . . .. ... . .. ... 3

Description and Account of Variation . . ..... 4 Economic Importance. ....... ..... . . 6

Distribution . . . . . . . . . . 15

Habitat . . . . . . . . . . . 23

Community Associations ............ 31

Growth and Development ............... 33

Morphology. ... ... . . . . 33

Perennation ........ . . . .... . . 38

Physiological Data. .......... . . 38

Phenology . . . . . . . . .. 40

Reproduction . . . . . . . . . 41

Floral Biology . ... ........ . 41

Seed Production and Dispersal .. ....... 45 Viability of Seeds and Germination ...... 46 Vegetative Reproduction. .. ... . . 50

Productivity and Standing Crop .... . .... 50



iii








Page

Control . . . . . . . . . . . 51

Arzama densa Wlk. ................. 53

Taxonomy ................... 53

Hosts Plants . . . . . . . . . 60

Biology and Life History of A. densa and Related

Species . . . . . . . . .. 63

Parasites, Predators, and Diseases ...... 68 CHAPTER. i. THE RELATIONSHIP BETWEEN THE PHENOLOGY AND BIOLOGY OF WATERHYACINTHS AND VARIOUS PHYSICAL AND BIOLOGICAL FACTORS 70

Introduction ................... .. 70

Methods and Materials ............... 73

Diurnal Waterhyacinth Productivity ...... 73 Annual Cycles and Insect Damage ........ 76

Site Description . . . . . . . . . 81

-Analyses . . ......... . .... 88

Results . . . ......... .... 91

Water Quality .. . ............ 91

Temperature and Solar Radiation ........ 104 Waterhyacinth Productivity .......... 114

Seasonal Variation in Photosynthetic Tissue .. 125 Seasonal Variation in Plant Density ...... 139 Seasonal Variation in Standing Crop ...... 150 Damage by Arzama densa . . .......... 157

Results of the Multivariate Analysis ..... 161

Discussion . . ......... ... .... 170





iv







Page

CHAPTER 2. THE CONSEQUENCES OF ATTACK BY ARZAMA DENSA WLK. ON SOME ECOLOGICAL CHARACTERISTICS AND MORPHOMETRIC FEATURES OF WATERHYACINTHS. .................. . . 185

Introduction . . . . . . . . . . . 185

Methods and Materials ................. 187

Analyses ................... . . 189

Results .................. . . .. 191

Plant Height .................. 191

Leaves. . . . . . . . . . . . 197

Plant Density . . . . . . . . . . 203

Biomass Estimates . . . . . . . . . 208

Productivity and Turnover Estimates . . . . 29 Plant Parts and Proportions . . . . . . 225

Discussion . . . . . . . . . . . . 241

CHAPTER 3. THE FEASIBILITY OF THE UTUILIZATION OF ARZAMA DENSA WLK. FOR THE BIOLOGICAL CONTROL OF WATERHYACINTH THE EFFECTS OF AN INTRODUCED POPULATION ON A SMALL POND COMMUNITY . . . . 244

Introduction . . . . . . . . . . . 244

Methods and Materials. ................. 247

Results . . . . . . . . . . . .. 250

Discussion . . . . . . . . ...... . . 265

CHAPTER 4. SOME NOTES OF THE BIONOMICS AND POPULATION DYNAMICS OF ARZAMA DENSA WLK .. . . . .................... . 271

Introduction . . . . . . . . .... . 271

Habits . . . ................. .... ...... 272

Fecundity .................. .. .... 275

Duration of Developmental Stages . . . . .. 278


v






Page

Population Cycles .................. .... 287

Mortality . . . . . . . .. . . .. 292

Discussion . . . . . . . . . . . . 304

Review of results and suggestions for further studies . 308 LITERATURE CITED . . . . . . . . . . . . 314

BIOGRAPHICAL SKETCH .. .. .. ... .. ..... .. .. 333









LIST OF TABLES

Table Page

1. Standing crop and productivity of waterhyacinths as estimated

by various authors. .... ... . ............. 52

2. Host plants of the Bellura-Arzama complex listed from various

literature sources . . ..................... 61

3. Hydrological budget for Lake Alice (March September 1973) .

. . . . . . . . . . . . . . . 85

4. A comparison of water quality measurements from Lake Alice,

Gainesville, Florida with previous reports. ....... 92

5. Metabolic and morphometric comparisons of the two morphological

types of waterhyacinth studied. ........ . . . 120

6. Average daily rates of change in biomass from initial and

final monthly values. ........ . . . . . . 156

7. Summary of multivariate regression analyses for annual variation plant characteristics. ...... . . . . . .162

8. Correlation coefficients (r) between independent variables..
. . . . . . . . . . . . . . . .164

9. Correlation coefficients (r) between dependent variables. .
. . . . . . . . . . . . . . . .167

10. Correlation coefficients (r) between dependent and independent variables and probabilities for a greater Ir ... . . 172 11. Ratios of the various plant parts and the percent change in the final values as compared to the initial values. ...... 236 12. Comparison of the samples from the release site with the control site based on various estimates of the plant and

insect populations. . . . . . . .. .. ... 251

vii









Table Page 13. Ratio of plant parts at the two sites on 12 December 1974. . . . . . . . . . . . . . . 263

14. Fecundity, egg viability, and egg stadia for 5 female Arzama densa collected as pupae in the field and mated in the

laboratory. .................. ...... 277

15. Summary of development data for A. densa. ......... 279

16. Annual summary of larval counts and mortality ........ 297 17. A summary of insects known to parasitize Arzama densa Wlk.

(from Vogel and Oliver 1969b in part) ........... 305



































viii









LIST OF ILLUSTRATIONS

Figure Page

1. An aerial view of Lake Alice on the University of Florida

campus . . . . . . . . . . 83

2. Water level taken at weekly intervals and precipitation at

Lake Alice from July 1974 through June 1975. ....... 87

3. Total carbonate and bicarbonate alkalinity and conductivity

of water samples taken from Lake Alice from June 1974

through June 1975. . . . . . . . . . . 95

4. Magnesium and total iron from Lake Alice water samples .. 97

5. Phosphorus concentrations present as phosphates and

nitrogen concentrations as total nitrate and nitrites from

water samples taken from Lake Alice. .......... 100

6. The negative logs of the hydrogen ion concentration (pH) of

water samples taken from Lake Alice. .......... 102

7. Potassium and sulfate ion concentrations of water samples

taken from Lake Alice. ................. 106

8. Maximum, minimum, and median weekly air and water temperatures at Lake Alice from late June 1974 through June

1975 . . . . . . . . . . . ....... 108

9. Solar radiation data from the University of Florida campus

from May 1974 through April 1975 ............ 112.

10. Diurnal curve for large waterhyacinth productivity

determined from CO2 gas exchange measured on Lake Alice with

an infrared CO2 gas analyser ............ . 116

11. Diurnal curve for small waterhyacinth productivity ... . 118




ix







Figure Page

12. A comparison of the standing crop and proportions of the

plant parts for the large and small waterhyacinth plants

used in the productivity studies .............. .123

13. Average daily solar radiation values per month for

Gainesville, Florida .................... .127

14. Annual phenological change in the average height of the

waterhyacinth plants on the marsh side of Lake Alice ... .129

15.. Annual variability in the number of leaves per waterhyacinth

plant from the study area ................. .131

16. Annual change in leaf density as determined from weekly

samples taken in the study area. ... ........... .133

17. The average area of the pseudolaminae of waterhyacinth

leaves . . . . . . . . . . . . .136

18. Leaf area index of the waterhyacinth population on Lake

Alice . ....... . . . . . . ... . . . . 138

19. Annual change in plant density as determined from weekly

samples taken in the study area. ............. .141

20. Average monthlycounts of the number of plants included in

each plant height class per square meter .......... .144

21. Statistics of skewness and kurtosis (peaking) derived from

each weekly frequency distribution of plant density by

height classes . . . . . . . . . . .146

22. The weekly waterhyacinth height class frequency distributions plotted three dimensionally on a time scale. ...... .149

23. Average weight per plant as a log function of the average

plant height . . . . . . . . . . . . 152



x








Figure Page 24. Standing crop values, both estimated and real, from Lake Alice. .................. ........ 154

25. Percentage of the leaves and rhizomes of the waterhyacinth population damaged through feeding activity of Arzama densa

at Lake Alice . .............. .. . .... 159

26. Plant density as a function of plant height. ........ 179

27. Average dry weight per waterhyacinth plant as a log function .of the average height. ................... 193

28. The average height per waterhyacinth plant (as measured from the longest leaf) as a function of the feeding activity of

Arzama densa larvae. .................. 196

29. The effects of varying levels of insect feeding activity on the average number of leaves per waterhyacinth plant expressed

as a percentage of predetermined means ........ .. 199

30. The effects of varying levels of insect feeding activity on the total number of waterhyacinth leaves per unit area

expressed as a percentage of predetermined means ..... 201 31. The effects of varying levels of insect feeding activity on the number of waterhyacinth plants per unit area expressed

as a percentage of predetermined means ........ . 205

32. The effects of varying insect concentrations on the total waterhyacinth biomass (expressed as both detritus and

living plant material) . . . . . . . . . 210

33. The effects of varying insect concentrations on the living waterhyacinth mass present per unit area ......... 212





xi







Figure Page 34. The effects of varying insect feeding activity on the amount of dead waterhyacinth plant material (detritus) per unit

area . . ......... . . . . . . . . . . 214

35. Detritus as a percentage of total waterhyacinth biomass as a function of insect feeding activity ......... ...218

36. Net waterhyacinth production as a function of insect feeding activity. . . . . . . . . . . ..... .221

37.. The ratio of conversion of living waterhyacinth plant material into detritus as a function of insect feeding activity. . .223 38. The effects of varying insect feeding activity on waterhyacinth green mass (pseudolaminae and petioles) .......... .227

39. The effects of varying insect feeding activity on waterhyacinth non-green mass (roots, rhizomes, and stolons) ........ 229 40. The effects of varying insect feeding activity on waterhyacinth root mass per unit area . .............. .... .231

41. The effects of varying insect feeding activity on the waterhyacinth rhizome mass present per unit area ........ .233 42. The effects of varying insect feeding activity on the waterhyacinth mass represented as stolons per unit area. ...... .235 43. A photographic comparison of the waterhyacinth stands at experimental and control sites at different times of the

year following the release of Arzama densa at the

former. . . . . . . . . . . . ... .253

44. A comparison of the standing crop of waterhyacinths at the control site and the release site ............. 259





xii







Figure Page 45. The mass represented by the various plant parts for an average waterhyacinth plant at both the control and

release sites. . ........... . . . . . .262

46. Probit analyses for developmental times of a greenhouse reared population of Arzama densa. ...... . . . .282

47. The head capsule diameter of Arzamna densa larvae at each

molt plotted against the larval age. .......... 285

48. The total number of Arzama densa larvae collected, either

living or dead, from the marsh side of Lake Alice. ...... 289

49. The age structure of the Arzama densa population during the

period of this study . . . . . . . . . . 291

50. The population of living Arzmna densa larvae (per square meter)

present on Lake Alice and the number of dead larvae expressed

as a percentage of the total . ............... .296

51. The total number of 4th and 7th instar Arzana densa larvae

per square meter as estimated from samples taken from the

marsh side of Lake Alice ................. 299

52. The number of parasites of 4th instar (CanpoZetis sp.) and

7th instar (Lydella radicis) Arzama densa larvae as

estimated from the number of pupae, or pupal exuviae found in

A. densa bores per square meter of waterhyacinth mat . .302












xiii








Abstract of Dissertation Presented to the
Graduate Council of the University of Florida in Partial
Fulfillment of the Requirements for the Degree of Doctor of Philosophy



THE POTENTIAL OF ARZAMA DENSA FOR THE CONTROL OF
WATERHYACINTH WITH SPECIAL REFERENCE TO THE
ECOLOGY OF WATERHYACINTH (EICHHORNIA CRASSIPES)



By


Ted Douglas Center


June 1976



Chairman: Dr. Dale H. Habeck
Major Department: Entomology

Waterhyacinth (Eichhornia crassipes (Mart.) Solms) is one of the world's worst aquatic weeds. Many countries have begun to consider biological control as a means to alleviate infestations. The United States has begun introductions of exotic insects for this purpose. To evaluate properly the success or failure of these introduced insects, background information on the ecology of waterhyacinths, the effects of insect attack on waterhyacinths, and the population biology of indigenous insects in the area of release is imperative. The purpose of this dissertation is to provide this information for Lake Alice, a primary study site on the University of Florida campus.

The ecology of waterhyacinth is approached in this dissertation through a comprehensive literature review and through field studies conducted to monitor various growth parameters. Diurnal productivity xiv




XV



curves comparing small and large plants revealed that the net efficiency of both plants was 1.6% (of incipient solar energy). The small plants grow faster than the large ones, however, by virtue of a larger P:R ration (2.05 vs. 1.46).

Photosynthetic efficiency is maintained by synchronization of the leaf area index with the annual solar energy flux. An annual increase in leaf area occurs first through an increase in leaf density and secondly through an increase in leaf (pseudolamina) size. The net result of these two growth phases is a peak in the leaf area index spanning the period of maximum solar radiation.

Intraspecific competition is strongly implicated in governing plant density and seems to account for observed changes in the population. Plant density is high in the winter reaching a maximum in April. This is followed by a decline in May and June as a result of the loss of plants in the smaller size classes. This loss is due to shading by the larger plants as they increase in size and leaf area.

Multivariate analyses indicate that solar radiation and minimum air temperatures were important in accounting for changes in standing crop, plant height, leaf area index, and the numbers of leaves per plant (all indices of biomass). The introduction of water quality parameters into the analyses resulted in confusion as causal relationships were difficult to establish.

Damage by Arzama densa Wlk. (Lepidoptera: Noctuidae) did not appear to affect the population of waterhyacinth studied. Greenhouse studies revealed that concentrations of 33 larvae per 100 plants could significantly reduce almost all characteristics examined and greatly accelerate




xvi




turnover. A seasonal aspect was implicated in the plant response as the insects appeared to be much more effective in the fall than the summer. This may be related to the energy budget of the plants under varying conditions of solar flux. Plant density increased in the summer in response to insect attack probably as a result of decreased intraspecific competition. A similar response would be expected to any factor which reduced competition provided adequate energy for growth was available.

To learn if A. densa could be used in biological control a greenhouse reared population was released on a small pond in August 1974. The waterhyacinth population was reduced and competing plants began to dominate the site. Ultimately cat-tail invaded and waterhyacinth failed to re-invade. A control site remained dominated by waterhyacinth.

Studies of natural A. densa populations on Lake Alice indicated that the failure of this insect to achieve sufficient levels to have an extensive effect on waterhyacinth was most likely due to the complex of parasites which attack them. Also, pickerelweed (Pontederia cordata) appears to be the preferred host of this insect which may partly explain the low populations observed on waterhyacinth. Further, seasonal changes in the plants' ability to withstand insect attack may obscure correlations between plant characteristics and insect damage.









INTRODUCTION


At the onset of this study various state and federal agencies were preparing for the release of exotic insects for the biological control of waterhyacinth (Eichhornia crassipes (Mart.) Solms) in Florida and the Southeastern U.S. In order to evaluate the effect of these insects it was apparent that prior information on the ecology of waterhyacinth at the release sites, particularly with regard to annual variability, and the effects of indigenous insects would be needed. The purpose of this dissertation was to provide part of this information for Lake Alice, the primary study site in the Gainesville area.

To achieve this end the annual sequence of events in the waterhyacinth population was studied as well as the actual and potential effects of natural and introduced populations of Arzama densa Wlk., a native insect which feeds on waterhyacinth. This dissertation is organized into five sections. The first section is a literature review organized into two parts. The first part reviews the biology of waterhyacinths and is organized in a manner similar to that suggest by Cavers and Mulligan (1972). The second part reviews the taxonomy and biology of A. densa Wlk. and related species.

The second section is a study of the waterhyacinth population on

Lake Alice. A fairly detailed description of the study site is provided. The phenology of various morphometric features of the waterhyacinth population is described with regard to the possible influence of various physical and biological factors. A short study on the productivity of waterhyacinth is also included in this section.


1







2

The third section investigates the potential effects of Arzama densa Wlk. on greenhouse cultures of waterhyacinth. These effects are evaluated in terms of various ecological and morphological traits of the plant.

The fourth section deals with the feasibility of augmenting natural populations of A. densa as a means of biological control. The effects of a small scale release are evaluated from a small pond near Paynes Prairie.

This fifth section constitutes notes on the biology and life

history of A. densa. Data on the natural population at Lake Alice is presented and the probable reasons for the failure of this insect to control waterhyacinth is discussed.

It is hoped that the information reported here will provide a

baseline from which future comparisons can be drawn after the release of exotic insects. It is also hoped that an increased understanding of the ecology of waterhyacinth in a situation relatively free of specific insect enemies has been gained.









LITERATURE REVIEW


Eichhornia crassipes (Mart.) Solms Taxonomy

Bock (1966) provides an excellent historical review of the literature dealing with the taxonomy of Eichhornia crassipes. The most current treatment of the genus appears to be that of Agostini (1974) which describes the species occurring in Venezuela. Five species are described (E. azurea, E. crassipes, E. diversifolia, E. heterosperma, and E. paradoxa) and a key provided. The synonomy provided for E. crassipes is as follows:

Eichhornia crassipes (Mart.) Solms in DC., Monogr. Phan. 4:527.
1883.
Pontederia crassipes Mart., Nov. Gen. 1:9. t. 4. 1824.
Piaropus crassipes (Mart.) Raf., Fl. Tell. 2: 81. 1837.
Eichhornia speciosa Kunth, Enum. Pl. 4: 131. 1843.
Eichhornia cordifolia Gandog., Bull. Soc. Bot. France 66:
294. 1920.

While the bionomial E. crassipes is in common usage today the

synonyms Piaropus crassipes and E. speciosa are common in the literature.

Because of the world-wide distribution of this plant it is known by a large variety of common names. Bock (1966) lists 48 common names for Eichhornia crassipes from 18 countries. The name waterhyacinth is used world-wide in scientific reports but the structure of the word has often been left up to the discretion of the user. It is often written as two words (water hyacinth), a hyphenated word (water-hyacinth), or as one word. Kelsey and Dayton (1942) in a list of standardized plant names use the single word, waterhyacinth. This usage seems appropriate since the plant isnot'related tothehyacinth as the two word name would


3





4



imply. For this reason I follow Kelsey and Dayton (1942) and use the spelling waterhyacinth throughout this dissertation.

Waterhyacinth is a member of the pickerelweed family which includes

9 genera and 36 species (Cook et al. 1974). The genera are Eichhornia (7 spp.), Eurystemon (1 sp.), Heteranthera (10 spp.), Hydrothrix (1 sp.), Monochoria (5 spp.), Pontederia (5 spp.), Reussia (4 spp.), Scholleropsis (1 sp.), and Zosterella (2 spp.). The majority of the members of this family. are confined to the Americas although two of the genera (Monochoria and ScholZZeropsis) appear to be Old World endemics. Keys and descriptions of the genera (world-wide) may be found in Cook et aZ.(1974). Lowden (1973) revised the genus Pontederia and united Reussia with it. Castellanos (1958) provides notes on the genus Pontederia in Brazil and keys and descriptions of the Brazilian species of Pontederiaceae (Castellanos 1959).

Description and Account of Variation

Following is a translation of the description of Eichhornia crassipes (Mart.) Solms given by Agostini (1974). Further descriptions can be found in Castellanos (1959), Sculthorpe (1967), Bock (1966), Penfound and Earle (1948), Buckman and Co. (1930), Webber (1897), and many others. The first definite description of this species (according to Bock 1966) was that of Kunth (1843).




5



Plants floating or sometimes fixed to the substrate,
the leaves in the form of a rosette with the stem
reduced and the plants connected by an elongated horizontal rhizome; numerous plumose roots issue
from each plant. The aerial leaves are variable
in shape; petioles of 2 to 30 cm long are more or less inflated; stipules 2-15 cm long with a small apical orbicular-reniform lamina with a lacerate
[serrate ?] margin; submerged leaves never evident.
Inflorescense variable, internodes between the
spathes nearly absent; inferior spathe with lamina
1-5 cm long, the sheath 3.5-7 cm long. Flowers
4-6 cm long; perianth light purple or rarely white,
tube 1.5-2.0 cm long, lobes 2.5-4.5 cm long, with
entire margins. Stamens all exserted, filaments villous-gladular. Capsule elliptical, trigonous, 12-15 mm long; seeds oblong-elliptical 1.2-1.5 x
0.5-0.6 mm with 10 longitudinal ridges. (Agostini 1974,P. 305)

The leaves of waterhyacinth are variable in the shape of the blade and the extent of development of the float. This variability is apparently dueto the plants' responseto environmental conditions. It was shown prior to 1930 that the size of the floats depends on such factors as light, temperature, and water quality (La Garde 1930). Rao (1920b) was convinced that increased water uptake (in a hypotonic medium) promoted the development of the float. This is discussed further by Bock (1966), Penfound and Earle (1948), and in an exceptionally good account by Misra (1969).

The waterhyacinth flowers are possibly trimorphic with regard to

the length of the style. Although medium and long styled forms are known, the existence of short styled forms is thought to be possible (Bock 1966). The midstylous form is normally predominant and Bock (1966) feels that only the mid and long styled forms exist. The flowers possess two whorls of anthers which Bock (1966) notes are long and short in the mid-styled form and short and mid-length in the long-styled form. This dimorphism is apparently regulated by a two gene system, one exerting an epistatic





6


effect on the other, and each gene with two alleles (Bock 1966; Ornduff 1966).

Very few cytological studies have been done on waterhyacinth. Banerjee (1974) found the chromosome number to be 2n = 32 in India, illustrated katryotypes, and described the chromosomes. She found the chromosome number to be very consistent but noted variants of 2n = 30 and 58. Bock (1966) also reported the diploid chromosome number to be 32 and noted that this had been reported by earlier authors as the probable number.

Economic Importance
Waterhyacinth is ranked world-wide as among the top 10 most important weeds and as the single most important aquatic weed (Holm 1969). Because of its floating habit and high productivity (Bock 1969) it competes with man for open water. Large build-ups interfere with hydroelectric operations in many areas (Holm et al. 1969; Rushing 1974). Its ability to interfere with navigation is well documented (Gay 1960; Evans 1963; Holm 1969; Webber 1897; Curtis 1900; Zeiger 1962). In the Panama Canal mats ofwaterhyacinth have become so thick as to interfere with the opening and closing of the locks (Pasco, pers. comm.). Gusio et al. (1965) cited a study in which the efficiency of canals in the Everglades were reduced 40-80% by large infestations of this plant. Irrigation operations are affected by the impediment of water flow and the clogging of pumps.
Waterhyacinths affect agriculture not only indirectly, as in irrigation, but also directly. Sugar and rice are cultivated in "flood-fallow"






7


situations where the land is flooded for several months. Aquatic weeds compete with the crops for the open surface, increase evaporation, and may provide reservoirs of crop pathogens (Nat. Sci. Res. Coun. of Guyana and N.A.S., 1973).

Water losses through evapotranspiration by waterhyacinth can be

considerable. Timmons (1960) showed that 17 states lost nearly 2 million acre-feet of irrigation water annually due to aquatic and ditch bank weeds. Holm et at. (1969) quoted an estimated value of this lost water as over $39 million. Rates of evaportranspiration by waterhyacinth are reported by Penfound and Earle (1948), Timmer and Weldon (1967), Misra (1969), Brezny et at, (1973), and Van der Weert and Kamerling (1974). These reports conflict, however, as experimental technique appears to be a large source of error in these studies. The ranges of the ratio of transpiration to open water evaporation are 1.66-6.6 (Penfound and Earle 1948), 3.7 (Timmer and Weldon 1967), 5.78-9.84 (Misra 1969), 1.02-1.36 (Brezny et aZ. 1973), and 1.20-1.58 (Van der Weert and Kamerling 1974).

Large mats of waterhyacinth covering the surface of water bodies block light to phytoplankton and submersed vegetation. This effectively prevents the liberation of oxygen through the photosynthetic processes of these organisms. Further, surface diffusion of oxygen is lowered as mixing is inhibited at the air-water interface. This results in a severe reduction of dissolved oxygen (Ultsch 1973). This renders the habitat unsuitable or lethal to many desirable species of fish. McVea and Boyd (1975) demonstrated that extensive pond coverage by waterhyacinth reduces phytoplankton growth and fish production. The composition of






8

the aquatic food chain may change from a plant-herbivore based community to a detritus-detritivore based community (e.g., Hansen et al. 1971) as a result of this loss of submersed primary productivity.

Waterhyacinths may also successfully compete with valuable wildlife forage thereby replacing it. This may destroy feeding areas for waterfowl (Gowanloch 1944). Local economies may be seriously damaged in areas which cater to recreational needs such as waterfowl hunting, fishing, boating, waterskiing, swimming, etc.

More seriously, riverine communities in the developing areas of the world which depend on fishing as a primary source of protein may be denied access to fishing grounds (Holm 1969). Holm (1969) further stated that impoundments for fish culturing may be destroyed by large masses of floating waterhyacinth. He stated that waterhyacinths constitute "...the most massive, most terrible and frightening weed problem" he had ever known.

Waterhyacinth possible pose a health threat by harboring vectors

and intermediate hosts of human diseases. The larvae and pupae of Mansonia uniformis (Theob.), a mosquito vector of filariasis in Asia, are known to attach to the roots of waterhyacinths (Burton 1960; McDonald 1970). Waterhyacinths may result in an increased production of mosquitoes by hindering insecticide application, interfering with predators, increasing the habitat available for certain species which attach to the plant, and by impeding runoff and water circulation thereby creating stagnant impoundments for breeding (Seabrook 1962). Mulrennan (1962) cautions that uncontrolled aquatic plant populations could lead to an increased incidence of mosquito-borne diseases such as malaria, encephalitis, and






9


other arboviruses. Mitchell (1974) indicated that aquatic plants, including waterhyacinths, may also harbor snails which are intermediate hosts of diseases such as fascioliasis and shistosomiasis. He mentioned that those species which do occur on waterhyacinth are assisted in their dispersal by the free-floating habit of this plant thereby spreading the associated diseases. Bock (1966) further reviews the literature dealing with health hazards caused by waterhyacinth.

It is difficult to arrive at sound figures regarding the monetary losses caused by waterhyacinth. Spencer (1973, 1974) quoted the following figures from a congressional report for losses prevented by waterhyacinth control in Louisiana in 1957.

Navigation $ 1,875,000
Flood Control unknown
Drainage 1,584,000
Agriculture 19,557,000
Fish & Wildlife 14,727,000
Public Health 250,000
$37,993,000

Penfound and Earle (1948) conservatively estimated that waterhyacinths were responsible for losses of $5 million annually as of 1948 in Louisiana. Spencer (1973, 1974) quoted figures from a Louisiana Fisheries and Wildlife report indicating losses of $65-75 million in 1947 in Louisiana due to aquatic weeds. In 1930 a report to the Jacksonville City Commission estimated that in the period from 1900-1930 the Federal Government spent $233,000 on waterhyacinth removal in the Jacksonville District to simply maintain "open channels for navigation" (Buckman and Co. 1930). By 1961 the total cost had risen to $1,861,788 in the same district (Tabita and Woods 1962). Wunderlich (1964) reported costs of clearing aquatic weeds ranging from $15-60 per acre. As of 1964 about







10

90,000 of Florida's 2,500,000 acres of fresh water and 70,000 to 100,000 acres of Louisiana's 2,000,000 acres of fresh water are covered with waterhyacinth (Ingersoll 1974). Hudson (pers. comm.) estimates that in 1975 the acreage of waterhyacinth in Florida has extended to more than 200,000 acres and the average cost of control per acre is about $25. He estimated that all agencies within the state in FY 1976 allocated $16 million for aquatic weed control, about 30% of which ($4.8 million) goes towards waterhyacinth control. This is an increase of almost $2 million over the previous year (FY 1975) for waterhyacinth control alone.

Thompson (pers comm.) indicated that between 1965 and 1974 the

U.S. Army Corps of Engineers spent $6.1 million in combined construction and operations funds for aquatic weed control in Louisiana alone. In the period between 1960 and 1964 the estimated cost was $1.7 million. Other weeds are of minor concern and for the most part 100% of this went toward waterhyacinth and alligator weed control. The State of Louisiana beginning its program in the mid 1940's spent $8.1 million as of 1973. Further costs included $1 million in 1974 and $1.1 million in 1975. Thompson further estimated that the average cost of treating an acre is between $32 and $35 in Louisiana. The most economical means being by helicopter ($13/acre) or fixed wing aircraft ($10/acre) when possible. The current estimate of acreage covered in Louisiana exceeds

1 million acres. This does not necessarily reflect an increase in acreage over Ingersoll's (1974) figure but is merely a more accurate estimate.

It is evident from these figures that the acreage covered with the plant is increasing while the cost of treating an acre is also









increasing. The result of this is a geometrically increasing trend in the overall cost of aquatic weed control by traditional means.

Because of the seriousness of waterhyacinth infestations the

beneficial aspects of this plant have been largely overlooked or ignored. Fringes of aquatic plants along rivers or lakes are often helpful in absorbing wind and wave action and preventing bank erosion (Tilghman 1963). Caldwell (1942) notes that the roots of waterhyacinth provide excellent cover for goldfish spawn. He promotes the growing of this plant for ornamental purposes stating that it is the "Biggest bargain in a pool plant .. .," and dismisses its detrimental attributes as an . attractive nuisance." Waterhyacinth was originally imported into this country for use as an ornamental (Raynes 1964) and the beautiful flower does give it a certain aesthetic appeal.

Tilghman (1962, 1963) spent many years fishing the St. Johns River and guiding fishing tours. He was vehement about the beneficial effects of waterhyacinth on fish propagation noting that the plant roots provide cover for spawn and support macro-invertebrates which are preyed upon by fish. Tilghman also noted that the plants helped clean the water thus improving the fish habitat.

Abu-Gideiri and Yousif (1974) studied the influence of waterhyacinth on planktonic development in the White Nile. They compared plankton populations and water chemistry parameters at a site south of Jebel Aulia Dam to similar studies done prior to 1958 before invasion of the area by waterhyacinth occurred. They found that overall planktonic densities had increased in the interim as a result of changes in the water quality (such as an increase in phosphates). They attributed







12
this change to the weed infestation but failed to consider cultural changes in the area. They concluded that the weed growth provides improved conditions for planktonic development and thus benefits fish production. The basis for these conclusions is obscure, however, and doesn't agree with the findings of other authors (eg. Wahlquist 1969b, McVea and Boyd 1975). While it seems probable that a fringe of the weed would be beneficial in some areas Bose (1945) commented ". . the various reports of fish mortalities in stagnant pools and ponds covered with waterhyacinth at once dispell the ideas and ruin the prospect that waterhyacinth should ever be fancied in tropical countries as 'one of the popular plants' for any kind of pisciculture."

Increasing attention has been directed towards using waterhyacinth. Pirie (1960) advocated utilizing waterhyacinth as a crop. Waterhyacinths are currently being considered for tertiary treatment of sewage effluent (see Dymond 1948; Sheffield 1967; Yount 1964; Yount and Crossman 1970; Boyd 1970b; Rogers 1971; Rogers and Davis 1971; Dunigan et al. 1975; Dunigan and Phelan 1975). Its possible use as feed for livestock (Chatterjee and Hye 1938; Baldwin et al. 1975; Bagnall et al. 1973, 1974; Baldwin 1973; Boyd 1968a, b; Combs 1970; Hentges 1970; and Salveson 1971). Apparently waterhyacinth is used as pig fodder in Singapore (Anonymous 1951). A complete cycle is developed when waterhyacinth is fed to pigs, wastes and fecal matter are washed from the piggeries into the pond, this fertilizes the pond to produce fish and more waterhyacinth which are both harvested.

Chatterjee and Hye (1938) found that waterhyacinth was high in potash with as much as 68% in the ash (5% on dry weight), comparable







13
with other fodders in nitrogen content (0.97 to 2.57% D.W.), rich in chlorine (3-4% D.W.), and richer than Napier and Guinea grass in lime (3.5% D.W.) and magnesia (0.96% D.W.). They also noted that i.tsphosphate content was low (0.36% D.W.) but that the digestible nutrients compared well with other fodders. Taylor and Robbins (1968) analyzed the composition of waterhyacinth and found the leaves to contain 15.8% dry matter which in turn was composed of 14.7% ash, 1.7% nitrogen, 10.7% crude protein and 17% crude fiber. The whole plants were 8.9% dry matter, 1.5% nitrogen and 9.6% crude protein. They also analyzed the plants for the amino acid composition. They concluded that the lysine content of waterhyacinth was sufficient to serve as an effective grain protein supplement.

Boyd (1968a) determined that waterhyacinth contained 12-18% (D.W.) crude protein. He subsequently fully analyzed the nutritive value of waterhyacinth and found the dry weight to be 5.9%, the crude protein to be 0.94% of the fresh weight (ca. 16% D.W.), cellulose ca. 28% (D.W.) total available carbohydrate 7.8%, ash 17%, and caloric content ca. 3.8 kcal/g. He further analyzed the inorganic nutrient content and the amino acid composition. Taylor et al. (1971) extracted protein from waterhyacinth and found that the percentage on a dry weight basis varied between 7.4 to 18.1%. They also analyzed the protein for the amino acid composition.

Knipling et al. (1970) compared the nutrient content of waterhyacinth from two different sites. They performed comparative analyses of various plant parts from the two sites for nitrogen, phosphorus, calcium, potassium, and magnesium as well as chlorophyll and water







14
content. They determined that the nutrient content in the plant tissues was not proportional to that of the water in which they were grown. Boyd (1974) has summarized the data on the composition of waterhyacinth and other aquatic plants.

Liang and Lovell (1971) evaluated waterhyacinth for use in channel catfish feed. They found that the addition of 5 to 10% waterhyacinth in vitamin free diets increased growth and reduced mortality in the fingerlings.

Bagnall et at. (1974) using waterhyacinth as feed supplements for cattle and sheep, for paper production, and for mulch determined that its processing as mulch was the most economically feasible use.

Azam (1941) proposed that underdeveloped countries encourage their people to utilize their spare time preparing various products made from waterhyacinth and thus supplement their income. Some of the products they suggested were paper, pressed board and tiles, detergents, cattle fodder, and manure. Nolan and Kirmse (1974) considered waterhyacinth unusable in the production of paper.

Iswaren and Sen (1973) found that an extract from waterhyacinth roots increased the yield of Brinjal (Solanum meZangena var. Pusa Kranti) from 507.2 g/plant to 1317.3 g/plant. Ganguly and Sircar (1964) found that a root extract from waterhyacinth increased the metabolic activity and the nitrogen and sugar content of Pisum sativwnum L. seedlings. Mukherjee et at. (1964) identified growth promoting substances in the roots of waterhyacinth which they believed to be bound auxins. Sheikh et al. (1964) noted that this extract was thermostable and found it to promote the growth of Phaseolus mungo var. roxburghii, the mycelium of Aspergillus niger, the growth of Rhizopus, and the multiplication of




15



yeast cells thereby promoting fermentation. S.M. Sircar and Chakravarty (1961) found that this root extract increased the yield of jute (Corchorus capsularis L.) and the production of fiber. P.K. Sircar et at. (1973) identified four gibberelin-like compounds from extracts of waterhyacinth shoots.

Other attempts to find marketable products of waterhyacinth include ink, upholstery stuffing, rope, bags, plastics, timber substitutes, and ice chests (Bock 1966). Morton (1962) gives a recipe for preparing waterhyacinths for food. For further reviews of the literature dealing with the utilization of waterhyacinth and other aquatic weeds see Bock (1966), Sculthorpe (1967), Little (.1968b ), Boyd (1972, 1974) and Mitchell (1974).

Distribution

It is generally accepted that South America is the area of origin of the waterhyacinth. Small (1936) indicates that it was originally discovered in the San Francisco River near Malhada, Brazil in 1824. Bock (1966) cites Hooker (1829) as listing several early collections from Brazil; Demerara River, Guiana; New Granada ( a former Spanish viceroyality including present day Venezuela, Ecuador, Colombia, and Panama); Guayaquil, Ecuador; and Buenos Aires, Argentina. She also cites early references indicating it was also native to the West Indies (eg. Schwartz 1928; Britton 1918). Castellanos (1959) lists its present South American distribution as Argentina, Paraguay, Brazil, Uruguay, Chile, Ecuador, Colombia, and Guyana. It has also been reported in Surinam (Little 1965, 1966; Holm et al. 1969). It has apparently been widespread in South America for many years as evidenced by the early records.





16


Misra (1969) cited a source which indicates that the center of origin for this species was probably the Pernambuco region of Brazil. A few authors have subscribed to other regions of origin outside of South America. Hildebrand (1946, p.477) states "The water hyacinth, Eichhornia crassipes, is a native of Japan and was carried about 70 years ago to South America, where it became widespread in fresh-water streams and lakes." He cites Gowanloch (1944) as the authority for this statement. Gowanloch apparently contradicts himself, however. In one paragraph he does indicate that waterhyacinth is native to Japan and was imported to South America. In the following paragraph he states, "When in 1884 an International Cotton Exposition was held in New Orleans, the Japanese Government representatives in their building on the Esposition grounds gave away as souveniers water hyacinths which they had imported from Venezuela." Small (1933) suggested that it may be native to Florida. A Ceylonese author claimed that Florida was its area of origin (Bock 1966).

Waterhyacinth is well known from the West Indies (Bock 1966).

Castellanos (1959) includes the Antilles within the range of distribution. Bancroft (1913) also indicated that the plant was present in the West Indies. Bock (1966) cites a paper which indicates that Puerto Rico is the center of dispersal for this species. She also found it naturalized in Jamaica. She suggests that it may have spread to the islands attached to boats or by floating from the mainland.

As might be expected waterhyacinth is also known from most of the Central American countries including Panama (Standley 1928; Hearne 1966), Costa Rica (Little 1965), Nicaragua (Little 1965, 1966; Holm et at. 1969), Honduras (Castellanos 1959), and El Salvador (Little 1965, 1966; Holm et at.






17

1969). I have not found any records of waterhyacinth occurring in Guatemala but its range does extend into Mexico (Castellanos 1959; Little 1965) where it is apparently well distributed.

As previously mentioned waterhyacinth was thought to have been introduced into the United States from Venezuela in 1884 at the International Cotton Exhibition in New Orleans by the Japanese delegation (Klorer 1909; Buckman and Co. 1930; Gowanloch 1944; Hildebrand 1946; Dymond 1948; Penfound and Earle 1948; Tabita and Woods 1962; Dutton 1964; Wunderlich 1964; Bock 1966). Some accounts indicate, however, that the plant may have been in the United States in the 1860's (Tabita and Woods 1962) or prior to the Civil War (Penfound and Earle 1948). The accepted theory maintains that the plant was given away as souvenirs at the New Orleans Cotton Exposition (Gowanloch 1944). The plants were taken for ornamental purposes (Klorer 1909; Dutton 1964; Wunderlich 1964) or for purposes of cultivating them for cattle fodder (Wunderlich 1964). In any case, it was felt that when the plants outgrew the limited amount of space given them they were cast out into natural bodies of water (Klorer 1909). By 1888 it was in the coastal fresh waters of Texas, Louisiana, Mississippi, and Alabama (Buckman and Co. 1930).
It was apparently introduced into Florida in 1890 (Webber 1897;

McLean 1922; Barber and Hayne 1925; La Garde 1930; Buckman and Co. 1930; Penfound and Earle 1948). Raynes (1964) reported it was first introduced in the St. Johns River at Edgewater about 4 miles above Palatka. Mr. J.E. Lucas was interviewed by a New York Sun reporter (Anonymous 1896) in 1896 and gave the following account:
"I know the man who brought the first plant to Florida," Mr. Lucas






18


said to a Sun reporter, "and he thought that he did the State a favor. I have it from his own lips, and I've known him since long before that time, for I used to carry him up the river in a launch year after year to his orange grove. He was Mr. Fuller, father of W.F. Fuller of Brooklyn, owner of Edgewater Grove, a property which he bought and improved, until now it is a beautiful place, seven miles above Palatka. Five years ago there wasn't a water hyacinth in the St. John's River, nor in the state, so far as I know. One season Mr. Fuller brought some there and put them in a pond on his premises. I understand that he brought them from Europe. They added very much to the beauty of the place, and they thrived so that he took some and threw them into the river. There they grew and blossomed abundantly, and they were greatly admired, and Mr. Fuller said to me one day: "The people of Florida ought to thank me for putting these plants here."

"But presently those in his pond had spread so that they covered it over. Then he cleared them all out. But it was too late to stop them from spreading all over the river. They worked their way and were blown up and down it for miles, and into the bayous, and finally up the Acklawaha [sic]. Two years ago they had become a serious menace to navigation, and protest after protest was sent to the Government. At last the War Department sent an agent to investigate, but he got to us just after the visitation of that heavy frost of two years ago, which killed all our orange trees. The hyacinths were killed too, apparently, and so the agent reported that nature had cleared the rivers and that there was nothing requiring the department's attention. But the plants were only







19

dead at the top. They grew again, and the startling conditions that you see in these pictures are a growth of only two years."

This account surprisingly indicates that the plants were introduced into Florida from Europe rather than from Louisiana as has generally been assumed (Buckman and Co. 1930; Tabita and Woods 1962; Wunderlich 1964).

Waterhyacinth was first discovered in Georgia in 1902 (Harper 1903) about one mile north of Valdosta. The first record in California is from 1904 near Clarksburg, Yolo Co. (Bock 1968). Johnson (1920) reported it in Fresno Co., California. Bock (1968) lists its present range in California from 10 mi NW Sacramento (ca. 38.50 N Lat.)to Ramona, San Diego Co. (ca. 330 N Lat.). She speculates that it was probably brought to California as an ornamental and released. The primary rivers infested are in Central California and include the Kings, Tuolumne, San Joaquim, and Sacramento River Systems.

The infestation of waterhyacinth in California is discontinuous with the North American range of this weed. Penfound and Earle (1948) stated that shortly after the turn of the century it had been reported from all the southeastern coastal states as far north as Virginia. A distribution map published by the U.S.D.A. (1970) indicates that the present range of this plant in the U.S. includes the Potomac River in Maryland-Virginia, west to southern Missouri, south to eastern Texas and southern Florida, and separately, central California.

Just when waterhyacinth spread to the Old World is not certain. Agarwal (1974) indicated it may have been introduced into India around 1896. McLean (1922) cited testimony indicating that it may have been present in Bengal as early as 1898 or 1899. It was apparently introduced






20


to the Buitenzorg Botanic Gardens in Java in 1894 (Bock 1966; Sculthorpe 1967). By 1905 it had appeared in Ceylon (Jepson 1933). It had become a serious problem in Cochin China (a part of S. Vietnam) by 1908, in Burma by 1913 and in Bengal by 1914 (McLean 1922). The Philippines acquired the plant around 1912 and by 1926 it was appearing in China as well as Borneo and Malaysia (Bock 1966). Records as to its entry in Japan are scarce but they were apparently introduced during this century as ornamentals (Shibata et at. 1965; Bock 1966). The first record of its occurance in Okinawa was in 1952 (Bock 1966) but it may have been there earlier.

From these beginnings it now occurs throughout India, Southeast Asia, and Indonesia (Bancroft 1913; Barber and Hayne 1925; Jepson 1933; Penfound and Earle 1948; Anonymous 1951; Sen 1961; Little 1965, 1968a; Holm et at. 1969; Chhibbar and Singh 1971; Haigh 1936; Hitchcock et al. 1949; McLean 1922; Agarwal 1974; Robertson and Thein 1932; Shibata et al. 1965; Gangstad et al. 1972; Ishaque 1952).

Waterhyacinth was first introduced into Australia in Queensland in 1895 (McLean 1922) and in New South Wales in 1896 (Maiden et al. 1906). It was apparently eradicated from New South Wales but a reinfestation occurred in the 1940's (Parsons 1963; Bill 1969), to South Australia by 1937 (Bill 1969) and to Victoria by 1939 (Parsons 1963). Bill (1969) notes that waterhyacinth is not a serious problem in Australia today except in some Queensland rivers.

Waterhyacinth also occurs in New Zealand (Walker 1954; Taylor 1955; Anonymous 1964b; Little 1965, 196Ba; Holm et at. 1969) although it is difficult to determine when it first appeared there. Taylor (1955)




21




seems to imply that it was discovered in 1949 at least in the Rotorua District. Walker (1954) reported it from the opposite side of North Island near Shannon. Matthews (1967) stated that there were 2 areas of infestation in 1948-50, 15 after 1950, and 70 by 1956. Another report (Anonymous 1964) indicated that there were at least 60 known infestations in New Zealand ranging from Opoua in the north to Shannon in the south. Manson and Manson (1958) noted its occurrence as far north as Kaitaica.

The spread of this plant has also taken in some of the Pacific Islands. It was reported from Hawaii in 1946 (Bock 1966) and Mune and Parham (1954) indicated that it was recognized as a pest in Fiji.
In Africa the plant is known from Kenya (Anonymous 1957), Zaire (Anonymous 1957; Lebrun 1958; Kirkpatrick 1958; Coste 1958; Berg 1959; Little 1965,1968a; Holm et al. 1969), Tanzania (Anonymous 1957; Little 1968); Uganda (Anonymous 1957), Angola (Lebrun 1958; Mendonca 1958), French Equatorial Africa (Lebrun 1958), Rhodesia (Lebrun 1958; Little 1968; Holm et al. 1969) Malawi (Lebrun 1958); Monzambique (Lebrun 1958; Mendonca 1958), South Africa (DuToit 1938; Penfound and Earle 1948; Lebrun 1958; Holm et al. 1969), Madagascar (Lebrun 1958), Sudan (Gay 1958, 1960; Davies 1959; Pettet 1964; Little 1965, 1966,1968a; Chadwick and Obeid 1966; Holm et al. 1969; Abu-Gideriri and Yousif 1974; Tag el Seed and Obeid 1975; Mohamed and Bebawi 1975), Senegal (Anonymous 1964; Little 1965; Holm et al. 1969) and Egypt (Little 1965; Holm et aZ.1969).
Waterhyacinth was first introduced into Africa either in South
Africa or Egypt. Sculthorpe (1967) cited a work on Egyptian flora which indicated that it made its appearance in Egypt in the period between






22


1879-1892. Bock (1966), however, indicated that it was not introduced into Egypt until 1912. It was introduced into South Africa around 1910 as an ornamental and by 1938 was reported from rivers in the Cape Peninsula, George, Knysna, Albany, Port Elizabeth, Uitenhage, Victoria East, and Natal (DuToit 1938). It had apparently reached South Rhodesia prior to 1937 as Europeans settling there reported its presence at that time (Holm et al. 1969; Bock 1966).

By 1942 waterhyacinth had spread into Mozambique in the Incomati estuary from Vila Luisa to Xanowano and apparently originating from the Transvaal of South Africa (Mendonca 1958). Kirkpatrick (1958) indicated that waterhyacinth was already present in Zaire (the Belgian Congo) in the Congo River in 1954. Coste (1958) felt that it was introduced in the period between 1950 to 1951. Other authors (Bock 1966; Holm et al. 1969) list 1952 for its introduction into the Congo. By 1955 it had spread over 1600 km of the river between Leopoldville and Stanleyville (Kirpatrick 1958). Gay (1958) first observed waterhyacinths occurring on the White Nile of the Sudan in 1958 along about 1000 km. It was apparently not abundant in the river prior to 1957 although it may have been present in 1956.

Senegal first reported the presence of waterhyacinth in 1964

(Anonymous 1964) from the Cape Vert peninsula and this may perhaps be the first record in the northwestern part of Africa. In spite of the warnings expressed by the Inter-African Phytosanitary Commission it was still available for purchase from street hawkers in Senegal in 1965.

Waterhyacinth is now distributed in all of the tropical and subtropical areas of the world. Its northernmost limits of distribution are probably near Sacramento, California (Lat. 38.50 N; Bock 1968),





23


the Potomac River near Washington, D.C. (ca. 30'N; Gowanloch and Bajkov 1948; U.S.D.A. 1970), Japan (30-350N; Holm uc al. 1969) and possibly Portugal (37-400N) as indicated by Holm et al. (1969) on their distribution map. The southern most limits of distribution appear to be Buenos Aires, Argentina (340S) and Concepcion, Chile (370S) in South America (Castellanos 1959), and Shannon, New Zealand (40-41OS; Anonymous 1964). The range in general seems to be bounded by the 400 North and South Latitude lines. Very little information is available on the altitudinal restrictions of this species although one paper (Anonymous 1957) states that it is limited in the tropics to an altitudinal zone of from sea level to 4500 feet (ca. 1400 m).

Habitat

Little is known of the ranges of environmental tolerances of

waterhyacinth. Webber (1897) noted the effects of freezing temperatures in Florida in the winter of 1894-95. He noted that the first freeze killed the top which caused the plant to float higher in the water. A second freeze killed this newly exposed portion. Most of the plants survived by resprouting from the unexposed portion of the rhizome.

Buckman and Co. (1930) stated that temperatures as low as 280F (-2.80C) may be withstood by the roots but will kill the tops. Temperatures lower than this will kill the roots as well. Hitchcock et al. (1950) observed waterhyacinths subjected to two days of freezing in New York. The plants were ice-covered when transferred to the greenhouse. Damage was apparently severe as the authors noted that all the foliage and all the roots were killed. Within 13 days the plants had recovered by resprouting from the rhizome tip. Misra (1969) placed





24



plants in a deepfreeze at -100C for 7-8 hours and they failed to revive. He found that at 150C the growth became restricted and the plant did not show any increase in growth up to a 90-day period.
Penfound and Earle (1948) exposed small plants in trays with 3

inches of water to various air temperatures for various durations. They found that at 330F and 270F (0.560C and -2.780C) all of the rhizomes resprouted when returned to room temperatures after being exposed for 12, 24, and 48 hours. At 230F (-50C) all resprouted after 12 and 24 hours but none survived after 48 hours. At 210F (-6.110C) some survived 12 hour exposures but none survived 24 or 48 hour exposures. At 190F (-7.220C) none resprouted after being exposed for 12, 24, or 48 hours. They concluded that the temperature effect depends upon the duration of exposure and that freezing of the rhizome tip results in the destruction of the plant.
Hitchcock et at. (1949) found that satisfactory growth

occurred in air temperatures of 21-270C. Silveira-Guido et at. (1965) stated that the plants grew well in water temperatures ranging from 17-350C. Bock (1966) measured air temperature ranges of 17-350C and water temperature ranges of 18.6-21.50C during the waterhyacinth growing season and winter mid-day temperatures of 5-10%0 for air and
5.80C for water in California. She also stated that the populations survived the winter of 1963-4 from which she monitored 28 da with air temperatures below freezing, 1964-65 with 25 day, and 1965-6 with 35 da although considerable mortality did occur.
Knipling et al. (1970) measured waterhyacinth growth along a gradient of water temperatures. They found the optimum to be 28-300C






25

although relatively high growth occurred over the range of 22-350C. Exposure to 100C nights reduced the amount of photosynthesis on following warm days.

Bock (1966, 1968) exposed plants to 26.70C-26.70C, 26.70C4.40C, and 4.40C-4.4oC day-night temperatures under both 16-18 and 8-16 L:D photoperiods. She found that growth was favored in the higher temperatures although it also appeared to be favored in the shortened photoperiod.

The maximum tolerable water temperatures appear to be around 33-340C. Penfound and Earle (1948) observed that the plants cannot tolerate water temperatures above 340C. Misra (1969) stated that in India the plants succumb at water temperatures above 330C. Knipling et al. (1970) found that growth began to be inhibited at about 330C and declined in a nearly linear manner at higher temperatures until, by 400C, negative growth was indicated. They noted that the plants were more tolerant of lower than optimum temperatures than of higher than optimum.

Light relations have been investigated by a few authors. Penfound iand Earle (1948) noted that in July the average light intensity was about 420 foot-candles above colonies of moderate-sized plants. Under the canopy of large plants the light intensity was about 170 ft-c representing a 60% decrease. They found that equitant elongate leaves are formed at intensities ranging from 130-500 ft-c and float leaves are formed at intensities over 500 ft-c. Under a walkway where the light intensity was 130 ft-c (31% of the July average) most of the plants were found to be dying. They also placed containers of waterhyacinth






26


under a table where the light intensity averaged 55 ft-c and all of the plants died in 2 mo. In connection with this they placed several plants in the dark and measured the starch depletion. By 7 da the starch content was reduced by 50% and by 12 da it was completely gone.

Hitchcock et al. (1949) grew plants in a greenhouse and supplied one group with supplemental heat, one group with supplemental light, and one group was left as a check. They found that the no. leaves per plant, the average leaf length, and the no. flowers produced were greatest in the high light condition.

In Africa (Anonymous 1957) it has been noted that light is seldom a limiting factor with respect to vegetation and frutification but it may have a more direct influence on germination.

As previously mentioned one study (Bock 1966, 1968) found that plants grown under the same temperature ranges grew better under the shorter photoperiod. This peculiarity was not explained.

Bock (1966) stated that waterhyacinth needed 60% full sunlight or better although she failed to define full sunlight. She placed plants under greenhouse benches when the light intensity at noon was 30-40% full sunlight. These were retained there from September to March and 67% mortality was observed.

Misra (1969) subjected plants to 40%, 70%, and 100% full sunlight (again undefined) and found that the no. leaves per plant and the percentage leaves with floats increased with increasing light intensity. Correspondingly a reduction in the average volume and diameter of the float occurred as the light intensity decreased.




27




Knipling et at. (1970) measured net productivity of attached
leaves under a range of light conditions. Photosynthesis increased from 7.8 mg CO 2/dm2 leaf surface/hr to 16.1 mg/dm2/hr as the light intensity increased from 1450 ft-c to 8000 ft-c. Dark respiration was found to range from 2.6 to 2.8 (average 2.7) mg/dm/hr.
Waterhyacinth is generally considered to tolerate a wide range of pH (Pieterse 1974). Haller and Sutton (1973) found they grew over a range of 4.0 to 10.0 although optimal growth occurred in acid to slightly alkaline conditions (4-8). Bock (1966) citing data from other authors concluded that waterhyacinth generally occurs in waters ranging in pH from 4 to 9. Chadwick and Obeid (1966) compared the growth of waterhyacinth and water lettuce (Pistia stratiodes L.) in cultures of varying pH. They found that waterhyacinth would grow at all levels (pH 3.0 to
8.2) but at 3.0 both dry-weight yield and offset production were minimal. They felt that pH values near 7.0 were optimal for waterhyacinth but values near 4.0 were optimal for water lettuce. Penfound and Earle (1948) reported pH values usually ranged between 6.2-6.8 in or near waterhyacinth mats in Louisiana but could survive extremes of 4-5 and 9-10. In the Guinean region of Africa pH is thought to be limiting at values of 4.2 or below (Berg 1959; Anonymous 1957).
Minschall and Scarth (1952) studied the effects of low ranges of pH (3.5-6.5) on the roots of waterhyacinth. They found that at values below 4.0 the roots exhibited decreased cell division and cell elongation. Cell division at pH 5.0 proceeded twice as fast as at 3.5. They further found that the plants could tolerate more acidity at cooler temperatures and the pH of the cell sap was always above that of the culture medium.






28


A few authors have suggested that stands of waterhyacinth may modify the pH of the water. Penfound and Earle (1948) noted that pond waters in the Mississippi River delta have an average pH of 7.2 whereas water in waterhyacinth mats are usually acid. Ultsch (1973) compared open water areas of a pond with areas covered with waterhyacinth and determined the yearly average pH to be 5.6 in the open areas and 5.4 in the areas with waterhyacinth. Haller and Sutton (1973) presented data which indicated that the plants cause a change in the direction of neutrality from both high and low initial pH values. Center and Balciunas (1975) compared water quality parameters from sites with and without waterhyacinth and found that those with the plants had lower pH (7.06 0.84) than those without (7.551.06) although the difference was not significant.

Moisture requirements of waterhyacinth and the effects of dessication upon its survival and growth have been only superficially examined. Webber (1897) noted that if the plants are to succeed a soil of loose texture thoroughly saturated with water is required. Parija (1934), however, found that they could survive 5.7% of water saturation in soil. Bock (1966) noted one instance when the plants survived 41 da in saturated soil. She speculated that the plants can withstand periods of dessication because excessive transpiration is prevented from the center of the rosette by the protective layer of dead outer leaves. Penfound and Earle (1948) found that waterhyacinth could survive drying periods up to 18 da depending upon climatic conditions and the surface they are exposed on. Sunny weather with the plants on galvanized metal killed the plants rapidly while rainy and cloudy weather or placing the plants in the





29


shade allowed them to survive longer. Misra (1969) found that when the rhizomes are air dried they progressively lose the ability to resprout as the moisture content decreases. They can tolerate a lower moisture content when dried in mud, however, than when dried in air. This may enable them to survive droughts in some instance.

As far as I have been able to determine Bock (1966) is the only one to have investigated the effects of humidity on the growth of waterhyacinth. She grew plants in a growth chamber both inside a plastic enclosure with high humidity and outside the enclosure. She concluded that high humidity favors growth.

With the recent interest in the utilization of waterhyacinth for

nutrient removal in sewage effluent increasing attention has been directed towards the nutrient requirements of this plant. Dymond (1948) found that the plants grew in both nutrient-rich and nutrient-poor water and that the nutrient content of the plant was higher in nutrient-rich water. Hitchcock et aZ. (1949) concluded that waterhyacinths have relatively low nutrient requirements as good growth occurred in solutions 0.01 to 0.001 times as strong as normal in water cultures. They also found that the growth response increased with added nutrients.

In Africa (Anonymous 1957) it has been noted that the lower limit of "mineralization" is very low but little is known of the upper limits. Chadwick and Obeid (1966) found that an increase in nitrogen levels caused a linear increase in the total yield and plant number but had little effect on the mean weight per plant. Knipling et al. (1970) studied two sites with notably different levels of orthophosphate and were surprised to find that the average standing crop yields were






30

similar at both sites. They further found that plants grown in varying phosphate solutions ranging from 0.075 ppm to 0.60 ppm did not signifi,cantly differ with respect to percentage weight gain over a 17 da period. Haller, et al. (1970) found that the critical phosphrus concentration for waterhyacinth growth was 0.01 ppm. Above this level phosphorus was absorbed in luxury amounts but a higher proportion of that available was absorbed at low concentrations. Haller and Sutton (1973) found that optimal growth occurred at 20 ppm phosphorus but levels higher than 40 ppm were toxic. They further found that the root weight was greatest at 0 ppm reflecting a tendency towards maximizing root absorptive surface in response to low nutrients.

Sutton and Blackburn (1971a, b) investigated the effects of varying copper solutions on growth and transpiration of waterhyacinth. They found that transpiration was reduced at 4.0 ppm with copper when grown in the solution for 1 week and at 2.0 ppm when grown for 2 weeks. Growth was inhibited by 3.5 ppm when subjected to the solution for 2 weeks. After one week the shoot dry weight was reduced at 8.0 ppm or above and the root dry weight by 16.0 ppm. The copper content of the shoot reflected the content of the water when the concentration was above 2.0 ppm but at levels below this the concentrations in the roots were independent of those in the water. The copper content of the roots increased linearly with the solution concentration.

Boyd and Scarsbrook (1975) found that the addition of 20:20:5 N:P2Os:K20 fertilizer to ponds increased the biomass yield of waterhyacinth. The fertilizer was added at 4 levels 0, 2.7 kg/ha, 10.8 kg/ha, and 21.6 kg/ha. It was interesting to note that the highest level of fertilization resulted in a yield less than the two intermediate levels.






31


Community Associations

Because of the worldwide distribution of waterhyacinth any comprehensive list of plants associated with it would be a tremendous task. A few authors have breached this subject on a local level, however. Harper (1903) noted the association of an orchid, Habenaria repens Nutt., with waterhyacinth in Georgia. Small (1936) listed a dozen plants which may be found growing on the floating mats of waterhyacinth and noted in New Orleans that it grows intimately with several other aquatic species. Penfound and Hathaway (1938) described plant communities in the marshlands of southeastern Louisiana. They found waterhyacinth associated with the cypress-gum swamps in strictly fresh water and presented an extensive list of other associated species. Penfound and Earle (1948) founda tremendous array of plants (63 species) occurring on mats of waterhyacinth. Eggler (1953) investigated the effects of 2, 4-D on other plant species associated with waterhyacinth and alligatorweed. Chadwick and Obeid (1966) investigated antagouismbetween Pistia stratiotes and Eichhornia crassipes. Bock (1966) listed several species associated with waterhyacinth in California and reviewed the work of several other authors. Abu-Gideiri and Yousif (1974) noted the composition of the plankton community in association with waterhyacinth stands in the Sudan.

Several workers have reported on the invertebrates associated with waterhyacinth but this has largely been the result of biological control investigations dealing primarily with insects. O'Hara (1967) quantitatively listed the invertebrates found in waterhyacinth mats. Hansen et al. (1971) listed some invertebrates present in the aquatic component of the waterhyacinth community and constructed a partial food web. They also studied the vertebrates present as did Goin (1943).





32


The entomofauna of waterhyacinth is quite large and diverse. Most of the information available regards those species which feed upon waterhyacinth and, thus, show potential as biological control agents. Sankran et at. (1966) investigated a grasshopper (Gesonula punctifrons Stal. :Acrididae) attacking waterhyacinth in India. Fred Bennett of CIBC in Trinidad has published many papers on the possibility of biological control of waterhyacinth and on the insects associated with it (Bennett 1967, 1968a, 1968b, 1970, 1972; Bennett and Zwolfer 1968). Other lists have been provided by Gordon and Coulson (1969), Coulson (1971), Perkins (1972, 1974) and Spencer (1973, 1974).
Sabrosky (1974) described a dipteran stemminer (Eugaurax setigena: Chloropidae) from South America. Barman (1974) investigated the growth and assimilation efficiences of an arctiid (Diacrisia virginica which is known to feed on waterhyacinth. Silveira-Guido and Perkins (1975) reported on the biology and host specificity of Cornops aquaticumn (Bruner), a grasshopper (Acrididae) from Argentina which attacks waterhyacinth. DeLoach (1975) provided indentification and biological notes on the genus Neochetina (Coleoptera: Curculionidae) that attack the Pontederiaceae in South America. DeLoach and Cordo (1976) provided information on the life cycle and biology of N. eichhorniae and N. bruchi, two species which have been released for the biological control of waterhyacinth in the United States. Warner (1970) described these two species.
Wallwork (1965) described a leaf-boring galumnoid mite (Orthogalumna terebrantis) from Uruguay which feeds on waterhyacinth which has subsequently been found in the United States (Bennett 1968a). Perkins (1973) studied the biology and host specificity of this species in Argentina.






33

Cordo and DeLoach (1975) investigated the ovipositional specificity and feeding habits of this mite also in Argentina. Del Fosse et al. (1975) determined the feeding mechanism of 0. terebrantis from the Florida strain.


Growth and Development
Morpholopgy.

Considerable confusion arises from the lack of uniformity in

naming.the vegetative structures of waterhyacinth. For this reason, I have adopted the terminology of Penfound and Earle (1948). The roots_4 .a numerous, fibrous, unbranched and adventitious (Couch 1971). They vary little in diameter (ca. 1 mm, Arnold 1940) but may range from 9 to 90
,~~--- ~~-------- ---- ~~~~-- -- --~~ ...~.~ ~~_~-.~-- -----cm in length (Penfound and Earle 1948). The length of the roots is strongly correlated with leaf length and may be correlated with water depth (Misra 1969). The roots are feathery in appearance due to the presence of numerous secondary lateral rootlets (du Toit 1938). Arnold (1940) found that the origin of these roots is unusual in that they make their first appearance in the immature pericycle a short distance from the promeristem and not, as in most plants, in mature tissue. The root is characterized by a distinct root cap which may extend up to

2.5 cm from the tip and is attached only at the tip (Olive 1894). The rootlets also possess root caps (Arnold 1940). The beginnings of the lacunar system are apparent in the roots approximately 1.5 cm from the tip (Olive 1894). The roots constitute 20-50% of the plants _biomu (Knipling et at. 1970 found it as high as 65%). The cortex is divided into three zones, which consist of a layer of parenchymatous tissue just under the epidermis, paenchymatous tissue surrounding the stele, and a






34


layer of lacunate tissue in between. The polyarch stele is surrounded by a weakly differentiated endodermis and pericycle (Couch 1971). Limited meristematic activity occurs at the apex and the roots possess a distinct epidermis (Sculthorpe 1967). The roots are known to become embedded into the mud although they are usually suspended freely in the water (Buckman and Co. 1930; Small 1936; du Toit 1938; Penfound and Earle 1948; Parsons 1963). The color of the roots is normally dark violet blue due to the presence of anthocyanin but when growing in the mud or in the dark the roots become white (Olive 1894; Penfound and Earle 1948). The roots arise from nodes on the rhizome (Penfound and Earle 1948).

The rhizome is the vegetative stem of the plant from which all other structures arise (Penfound and Earle 1948). It consists of an compact axis with short internodes and the leaves, roots, stolons and inflorescences are produced by the meristematic nodes which have generally small, compact cells. An area with a considerable number of intercellular air spaces exists around the periphery of the meristematic tissue (Couch 1971). The rhizome is approximately 95% water by weight and has a specific gravity of 0.805 (Penfound and Earle 1948). Olive (1894) indicated that there was no evidence of starch being stored within the rootstock (rhizome) but Penfound and Earle (1948) felt that the rhizome was the main organ of starch storage.

The rhizome may produce long horizontal internodes (stolons) which produce new shoots at the distal end which results in a sympodial branching pattern (Penfound and Earle 1948). These stolons arise from axillary stem buds (Bock 1966). Aerating spaces are abundant near the periphery and the collateral bundles are aggregated in the center (Olive 1894). The stolons are purple due to the presence of anthocyanin and range in diameter from 0.5-2.0 cm as;' : length up to 40 cb "ira 1969). The





35


specific gravity of the stolon is 0.818 and it consists of about 97% water by weight (Penfound and Earle 1948).

The most interesting morphological development of the waterhyacinth is its leaves. Arber (1918, 1920) found that the vascular bundles in the petiole are arranged with the xylem oriented towards the periphery. Those in the lamina may be arranged with the xylem up, down, or oblique. This is in contrast to plants with a true lamina which have the vascular bundles arranged with the xylem towards the upper leaf surface. She suggests that this indicates that the lamina is merely an extension of the apical end of the phyllode and not homologous with the laminae of a Dicotyledon. As such it should properly be called a pseudolamina and the basal portion a petiole. The two are connected by a narrow compact region called the isthmus and the narrow base below the float is referred to as the subfloat. A membranous ligule (= stipule of Agostini 1974) is present at the base of the subfloat (Penfound and Earle 1948) which possesses a small reniform lamina (Agostini 1974).

The petiole may be more or less inflatdtto form a bulb-like

structure commonly assumed to function in flojtiug the plant (Parsons 1963; McLean 1922; Chhibbar and Singh 1971; Olive 1894; Couch 1971). This has been contradicted by Rao (1920b) because the bladders are formed mostly above the water and the leaves float with or without them. Bock (1966) noted, however, that the bases of the inflated petioles just beneath the water formed a stable platform. Further, floating single plants with elongate petioles were unable to remain upright and if they remained on their side sent out new leaves with inflated petioles.






36


Several factors have been indicated as important in the development of the float. Rao (1920b) concluded that high osmotic pressure was the important factor although this may be altered by numerous factors. The lack of swelling may also be associated with high plant density (LaGarde 1930), anchorage in soil (LaGarde 1930; McLean 1922), shade and high temperatures (LaGarde 1930; Arber 1920). Conversely, bulbous petioles may be associated with the free-floating habit, full sunlight, or cooler temperatures (LaGarde 1930). Bock (1966) found she could not correlate petiole shape with shading.

Tce buoyancy of the plants is largely due to the presence of air spaces in the highly lacunated aerenchyma resulting in 70% air by volume (Couch 1971). The specific gravity of the float is 0.136 and of the pseudolamina is 0.741. The floats are 94% water by weight and the pseudolaminae are 89% (Penfound and Earle 1948). The petioles range in size from a few centimeters to as much as 1.5 meters in the equitant form (Buckman and Co. 1930). The angle between the leaves and the water surfaces ranges from 15 to 450 around the periphery of the rosette and from 75 to 900 in the center (Penfound and Earle 1948).

The leaves not only stabilize the plant and kee.p-itafloat-but-act as sails which catch the wind and move masses of them over the surface
- --------------------. *~....~.~------ -^~-.- 4~~~~~
of the water (McLean 1922; du Toit 1938). Further, the geometric arrangement of the leaves into a rosette with a large leaf area (as much as

8 m2/m2) and the erect habit of individual leaves is extremely efficient for light interception (Knipling et aZ. 1970).

The anatomy of the pseudolamina and petiole is further discussed by Olive (1894), Bock (1966), Sculthorpe (1967), Arber (1918, 1920),




37



and Penfound and Earle (1948). The lacunar system may enable the plant to utilize internal carbon dioxide (Billings and Godfrey 1967).

The inflorescence is displayed on a long peduncle (Penfound and Earle 1948) and is usually elevated a few centimeters above the leaves (du Toit 1938). Two unlike spathes subtend the inflorescence the lower being leaf-like and bearing a pseudolamina and the upper bract-like (Cook 1974). The inflorescence is a spike (du Toit 1938; Buckman and Co. 1930; Penfound and Earle 1948; Bock 1966) or may be considered spike-like or paniculate (Cook 1974; Bock 1966). The spike is 15-30 cm long (du Toit 1938; Mune and Parham 1954) and contains numerous flowers (6-20, du Toit 1938; 8, Parsons 1963; 10-12, McLean 1922; 6-12, Mune and Parham 1954; 4-29, Misra 1969) borne on a rachis with a flowerless subrachis below the inflorescence and above the spathes (Penfound and Earle 1948). The individual flowers consist of a hypanthium about 1.41.8 cm long (Misra 1969), 3 sepals, 3 petals, 6 stamens and a tricarpellate ovary (Penfound and Earle 1948). The petals and sepals are lavender in color (Bock 1966) and united at the base to form a 6-lobed tube (Cook 1974). The color of the flower is due to the anthocyanin, eichornin (Shibata et at. 1965). The tube is curved, glandular and pubescent near the base (Mune and Parham 1954). The perianth is slightly irregular with all 3 sepals and 2 petals similar in size and shape but the upper petal is somewhat wider and bears a distinctive yellow spot in the center bordered by a darker blue or violet area (Bock 1966; Buckman and Co. 1930). Buckman and Co. (1930) indicate that the function of this spot is obscure but others have indicated that it may function as a nectar guide to visiting bees (Sculthorpe 1967). The six stamens are arranged in two





38



whorls of 3 stamens each, of two different lengths and are adnate being fused to the corolla tube at the base of the filaments near the sinuses of the perianth lobe (Bock 1966). The filaments are white at the base, purple at the apex, and glandular (Misra 1969). The anthers are oblong and attached near the base (Bock 1966) and contain about 2000 pollen grains each (Penfound and Earle 1948). The ovary is superior, sessile, trilocular, contains numerous ovules in axile placentation (McLean 1922; Bock 1966) and conical in shape (Penfound and Earle 1948). The six stigmatic surfaces, by their close approximation, appear to be capitate but they are not (Penfound and Earle 1948; Bock 1966). The stigmate surface is covered with numerous glandular hairs (Bock 1966). The fruit is a loculicidal capsule containing seeds with an abundant mealy endosperm (McLean 1922). Approximately 50 seeds are produced per capsule (Penfound and Earle 1948).


Perennation.
Waterhyacinth is generally considered to be a perennial by virtue of its rhizome (Penfound and Earle 1948; Sculthrope 1969). Penfound and Earle (1948) felt that the rhizome may maintain a constant length over a pJofsula l years. Exact data on how long an individual rhizome may exist is not available but the plants are known to survive periods of freezing weather by resprouting from the rhizome (Bock 1966).

Physiological Data.
Most of the appropriate physiological data has already been discussed in scattered sections of this literature review but it bears repeating in a more organized discussion.






39

Many authors have studied transpiration rates and found variations

due to such factors as solar energy, wind speed, temperature, and technique. Misra (1969) found that waternjoss through a waterhyacinth mat was as high as 65.5 kgm /ida. This represented a water requirement of 6.74 kg of water per gram (dry wgt.) of biomass produced. The ratio of evapotranspiration to open water evaporation (ET:EO) ranged from 5.92 to 9.84.

Knipling et at. (1970) measured the moisture content of a stream of air before and after it had passed over a waterhyacinth leaf. They found during the day the transpiration rate increased from 1520 to 2450 mg/dm2 (leaf surface)/hr. in response to increasing light intensities. Dark transpiration values were also high, however, averaging 1430 mg/dm2/hr. In another experiment plants were grown in beakers in a variety of phosphorus concentrations and measured for daily water loss. There was no significant difference in evapotranspiration between the phosphorus concentrations. The ETEO ratio, however, was 305 g/da:100 g/da or 3:1. Dry matter production was 0.27 g/da indicating a water use efficiency ratio of 1129 gm H20/g plant dry wgt.

Other average values reported for the ET:EO ratio have been 3.2

(Penfound and Earle 1948), 7.8 (in India; Holm, et al. 1969), 3.7 (Timmer and Weldon 1967), 1.02-1.36 (Brezny et al. 1973), and 1.46 (Van der Weert and Kammerling 1974). The latter authors have found that 97% of evaporation from waterhyacinth covered situations is the result of the process of evapotranspiration.

Knipling et at. (1970) have also provided data on respiration and photosynthesis by measuring the CO2 concentration in an air stream passed over a leaf. Net photosynthesis increased from 7.8 to 16.1 mg C02/dm2/hr





40



with light intensities increasing from 1450 ft-c to 8000 ft-c. Respiration averaged 2.7 mg C02/dm2/hr. Ultsch and Anthony (1973) have found that waterhyacinth may have the capacity to utilize CO2 dissolved in the water under the mats as a source of carbon absorption throu ST his m9~ eaf for up to 1%-~ff the total carbon fixed. Billings and Godfrey (1967) have found that some hollow stemmed plants may use internal carbon dioxide generated from root and stem respiration in photosynthesis. This may be true in the case of waterhyacinth.

Ntrient uptake rates are not well worked out for this plant.

Waterhyacinth contains about 0.4% P and 2.6% N by weight (Boyd 1970b) or an N:P ratio of 6:1. If these represent constant proportions the rate of nutrient uptake is proportional to the growth rate of the plant. If the standing crop increases at a rate of 100 g/da the rate of uptake of N will be 2.6 g/da and of P 0.4 g/da. This agrees well with the results of Dunigan et al. (1975) who found the N:P ratio of uptake rates to be 5-6:1. The daily absorption rates from 6 liter containers were 2.4 ppm N and 0.4 ppm P in water concentrations of 50 and 100 ppm N and P. At concentrations of 250 ppm N and P the daily uptake rates were 3.5 and

0.7 ppm respectively. This implies a growth rate of 600 and 810 mg dry wgt/da. Mitsch (1975) indicates that this high ratio of N:P absorption indicates that nitrogen is generally more limiting to waterhyacinth than phosphorus.

Phenol.gy,

Data on the sequence and timing of events in the annual cycles of waterhyacinth population are scarce. Penfound and Earle (1948) measured the average length of the largest leaves over one growing season. The





41


maximum was reached in August and the period of maximum growth was between May and August. They also followed the flowering cycle over a period of years in Louisiana. In the years 1945-1947 anthesis began in April. They felt that a definite flowering rhythym occurs in a given colony of plants. Anthesis is maximum in June and declines through September although this may vary from colony to colony and a second period of flowering occurs in September and October and continues through November into December. Buckman and Co. (1930) reported that the plant is supposed to bloom every two to three months. In India flowering occurs throughout the year but is most abundant in the post-monsoon months (Sahai and Sinha 1970). A pre-monsoon flowering period (April and June) has been reported in India (Pieterse 1974). Sahai and Sinha (1970) further found that biomass accumulation was highest in January and February, and the area occupied (% cover) was greatest in February and March in India.

Reproduction
Floral B.ioQlogy

A single flowering spike contains a variable number of flowers. Bock (1966) found the average to range between 5 and 10 flowers per inflorescence although she indicated that other authors have observed up to 35 flowers per inflorescence. Small (1936) indicated that flowering occurred on a daily cycle appearing as buds up to 7:30 AM and opening by 8:00 AM. The mode of pollination may be allogamous or autogamous. Bock (1966) found that allogamous pollination may occur through the actions of several insect pollinators. She listed Apis mellifera, Halictus (1 sp.), and Lasioglossnum (2 sp.) as known pollinators and syrphid flies as possible






42


pollinators. Penfound and Earle (1948) observed honeybees, bumblebees, black unidentified bees, and sulfur butterflies visiting the flowers. They described three patterns of behavior of honeybees in visiting the flowers; visiting distal anthers only, alighting with the head among the proximal anthers and the abdomen on the stigma, and visiting the proximal anthers after alighting on the banner petal. They questioned the importance of insect pollinators in accounting for the production of seed in this species. Bock (1966) noted, however, that honeybees crawl down the floral tube to retrieve the nectar and in so doing receive pollen from both sets of anthers. She observed a great deal of cross-pollination.
Autogamous pollination occurs when the flower wilts and the stamens are twisted against the stigma (Penfound and Earle 1948; Bock 1966; Tag el Seed and Obeid 1975). Penfound and Earle (1948) found much more pollen on the stigma after the flowers had completely wilted than at any other time thus stressing the prevalance of autogamous pollination.

Since waterhyacinth flowers are at least dimorphic with regard to style length either legitimate (styles pollinated by anthers not of equivalent length) or illegitimate (styles pollinated by anthers not of equivalent length) crosses are possible (Ornduff 1966; Bock 1966; Frangois 1964-63). Both legitimate and illegitimate crosses result in seed production (Bock 1966). Franois (1964-63) reported that self-incompatibility was stronger in long styled forms than in short styled forms. Ornduff (1966) studied the breeding system of Pontederia cordata and compared it with E. crassipes. He concluded, as did Bock (1966), that both species exhibit relatively weak self-incompatability.
An interesting aspect of the floral biology of waterhyacinth is the





43




phenomenon of anthokinesis or the bending of the axis of the inflorescence following anthesis (Agharkar and Banerji 1930; LaGarde 1930; McLean 1922; Penfound and Earle 1948; Bock 1966; Misra 1969). LaGarde (1930) described this process as follows:

"As soon as the inflorescence starts wilting the upper portion of the stalk with the fertilized blossoms begins to bend downward. When this upper part has reached the surface of the water, usually after five days, the lower portion of the stalk commences to bend at the base, thus pushing the developing seed-pods under the surface of the water. This movement stops when the lower part of the stalk is level with the surface. The upper part carrying the seed pods is then submerged in the water at an angle of 450, the seed pods being covered and protected by the root system . The whole process of bending requires from six to seven days." (LaGarde 1930, p. 51).

Agharkar and Banerji (1930) quoted other workers who indicated

that anthokinesis was accompanied by a lengthening of the peduncle. They found, however, that the peduncle did not lengthen considerably and such lengthening was confined to an area of 1 to 2 cm below the terminal node. They further found that removal of the flowers or amputation of the peduncle above the node had no effect and curvature was normal. This was also true when they removed the peduncle and placed it in water.

Penfound and Earle (1948) studied that anthokinetic cycle and their results agree with other workers. They found that it requires about 14 days from the initiation of the floral bud until opening occurs. Floral opening begins about 8:00 AM, if all the flowers open the bending phase begins at about 5:00 PM of the same day. Bending occurs in three places: at the rhizome crown, jusc' below the two bract; of the inflorescence,





44


and in the rachis. Most of the flowers are inverted by 5:00 PM the following day. The complete cycle from flowering to complete geniculation takes 48 hours in the summer. This is contrary to LaGarde's (1930) finding that it takes 6 or 7 days. Bock (1966), in her studies, agreed with Penfound and Earle (1948).
Bock (1966) seemed to concur with the findings of Rao (1920a) in

that the bending was due to geotropism in that when the roots were packed with sponges and the plants held horizontally, no bending occurred. She disagreed with Agharkar and Banerji (1930) in that removal of the flowers would not permit bending to occur unless all of the flowers had wilted and bending had commenced first.
Misra (1969) found that curvature took place when the tips along with 2 terminal flowers were removed, when all of the flower buds were removed, and when all of the flowers were removed after they had opened. In all cases complete bending took as long as in the controls (35-40 hrs.). He found that this curvature was due to increased cell size along the outer edge of the curving portion. He felt that this process represented a free-running endogenous rhythm independent of auxins (geotropic in nature), photoperiod, temperature, and opening of the last flower as suggested by other authors.
Spermatogenesis has been described by Smith (1898) and Banerji and Gamgulee (1937) and oogenesis by Smith (1898). Pollen morphology and germination and development of the pollen tube have been investigated by Ganerji and Gangulee (1937), Bock (1966) and Tag el Seed and Obeid (1975). The embryology of the seed is discussed by Smith (1898), (Coker (1907), and Swamy (1966).





45


Seed Production and Dispersal

The degree of seed set seems to be extremely variable. Agharkar and Banerji (1930) found that 10 hours after anthesis through natural pollination (autogamous or allogamous not distinguished) 35% of the flowers were fertilized (30% with actively growing pollen tubes and another 15% with pollen grains present). Through artificial pollination, up to 71.3% of the fertilized flowers set fruit. McLean (1922) in Bengal found that only 1% of the flowers set any seed. Haigh (1936) found in Ceylon that 36 to 71% of the capsules produced may be empty. Backer (1951) found no seed set in Java and Misra (1969) found up to 48% of the fruits bear seed in India.

The conditions for seed set have been investigated but the results are confusing. Parija (1934) indicated that temperatures between 240C and 290C were necessary. Agharkar and Banerji (1930) felt that relative humidities above 90% were required. Bock (1966) found that seed was set when the relative humidity was never greater than 72%. Haigh (1936) found the number of seeds per inflorescence to be 86, 28, and 91 when the relative humidity was 90%, 70%, and 67% respectively. Tag el Seed and Obeid (1975) concluded that seed set was favored if pollination occurred immediately after the flowers opened. Thereafter, successful pollination was hindered by high temperature and low humidity which affected the stickiness and receptivity of the stigma.

Data on the quantity of seeds set per fruit or inflorescence also

indicate a great deal of variability. Haigh (1936) artificially pollinated flowers and found an average of 24 seeds/capsule with a maximum of 72. Bock (1966) reported the average in California to be 4.2 with a range of






46


1-16. Misra (1969) indicated that the average may be 15-41. Robertson and Thein (1932) found 50-150 seeds/capsule in Burma, Zeiger (1962) reported 3-250, and Frangois (1964-3) reported an average of 153.6 with a maximum of 244. Tag el seed (1972) reported an average of 98.95 with a range of 5 to 542.

The number of seeds per inflorescence depends upon the number of

flowers per inflorescence and the number of seeds per flower. Bock (1966) indicates that the average number of seeds per inflorescence is 3.44 in California. Tag el Seed and Obeid (1975) reported 1.5 capsules per inflorescence. Using the data from Tag el Seed (1972) for the average number of seeds per capsule (98.95) this expands to 148 seeds per inflorescence. Matthews (1967) indicated that a single spike may produce 5000 to 6000 seeds. Zeiger (1962) estimates that 45 million seeds may be produced per acre by medium sized plants. Penfound and Earle (1948) estimated a crop of 900,000 capsules/acre.

The mode of seed dispersal has not been studied to any extent. It seems apparent that since the seeds are deposited in the water the primary mode of dispersal would be through drifting. A few authors have indicated that birds and fur bearing animals may disperse seeds (Maiden et al. 1906; Holm et al. 1969; Gay 1960).


Viability of Seeds and Germination

Conditions for germination of waterhyacinth seeds have been

studied by many workers. Crocker (1907) refuted the idea of earlier workers that desiccation of the seeds is a necessary prerequisite to germination. He further found that green seeds kept at 230C germinated (20% within 1 week) while seeds kept at 60C did not, although they did




47



ripen. Seeds with mature coats failed to germinate when kept at either 230C or 290C. He then separated the embryo from ripe seeds or ruptured the seed coats and placed them in a bath at 290C. He noted that germination occurred very rapidly in both cases (96% after 1 day). He ruled out oxygen as a factor because they germinated equally well in boiled water covered with paraffin. He concluded that the hard seed coat and endosperm hinders water absorption and limits germination and that desiccation may, in fact, fracture the seed coat and promote germination.

Agharkar and Banerji (1930) indicated that a ripening period of

20 to 23 days was required for maturation of the fruit. After maturation they are severed from the axis by an abscission layer and float on the water surface for a day or two before sinking. Splits develop on the lateral walls through which seeds are discharged. They found that the seeds develop freely in tap or distilled water.

Parija (1930) suggests that germination takes place "in the

beginning of rains or whenever the humidity, soil moisture and temperature are suitable." (Parija 1930, p. 388). He felt that the function of the rain was to provide moisture, and expose the seeds in the mud providing access to oxygen.

Robertson and Thein (1932) noted that in every instance when they had found waterhyacinth seedlings it was in a depression which completely dries out during the dry season and floods again in the rainy season. They concluded that a period of drought alternating with a period of plentiful moisture was necessary for germination.

Haigh (1936) exposed seeds to varying treatments of always wet,

always dry, or alternately wet and dry. No germination occurred for three months unti eeds were placed in the sun. Within eleven days germination






48


had begun. They further found that drying, bubbling air in the water, and the addition of rotting waterhyacinth fragments would not promote germination when kept in the laboratory. To determine if heat or exposure to an intense light was the important factor they exposed seeds to a normal light bulb and to a blackened light bulb. Germination occurred only in the illuminated treatment. They concluded that bright sun is necessary for germination and that heat, in conjunction with high light intensity may also be required. Haigh (1940) later found that if disiccation is prolonged for a long enough period light is not necessary. He found that seeds collected in June 1935 would germinate in the laboratory as late as January 1937 (19 months).

Penfound and Earle (1948) concur with Haigh in his findings. They found that seed germination would occur on upturned plants indicating that either drying or increased light intensity was favorable for germination. They also concluded that scarification aided germination.

Hitchcock et aZ (1949) indicated that an after-ripening period of about 2 months was necessary for germination and under ideal conditions 100% germination was possible. Dry seeds, however, required twice as long (111-112 days) to germinate as seeds stored wet (64-67 days). They also found that relatively high water temperature (28-360C) favored germination but the seeds could survive very cold temperatures. When stored for 69 days at temperatures of -5, 0.5, 5, 10, and 220C and then placed in normal air temperatures germination occurred in every case except the -5oC treatment. When exposed for only 1 week even the -50C treatment gave 50% germination.

Hitchcock et al. (1950) investigated the effects of water depth on seed germination. At water depths of 2.5, 10.2, 20.3, 30.5, and 40.6 cm they obtained 40, 60, 72, 72, a 1i)% germination respc, tively. They






49



felt that the difference was due to longer heat retention in the deeper water at night. Under 15.2 cm of water in a brown glass bottle only 28% germination was observed.

Barton and Hotchkiss (1951) also studied the effects of temperature, light, and storage on seed germination. They concluded that a combination of high temperature and light is needed for germination of dormant seeds although temperatures as low as 50C did not impair germination when in direct sunlight (greenhouse) and alternating temperatures (5-300C, 5-35C, and 5-400C) allowed some germination even in the dark. They also found that a storage period of a month or longer hastened germination especially with less mature seeds.

Francois (1964-3) obtained good rates of germination (over 95%) by keeping his seeds in a 12:12 L:D photoperiod with a corresponding 400C:200C temperature regimen. Bock (1966) was convinced that seeds do not germinate in California and found that they do not remain viable there for longer than 2 months. Sculthorpe (1967) reflected the findings of other authors by indicating that the seeds are able to tolerate a long dry period and remain viable. Tag el Seed (1972) investigated seed germination under a wide range of chemical treatments and under low oxygen tension and low redox potential as well as many other environmental conditions. His extensive studies indicate that germination is stimulated by low redox potential and low oxygen tension expecially after wet storage, germination is most likely to occur in water warmed by intense light, the addition of organic matter to the substrate stimulates germination, the seeds will only germinate at the surface of the substrate, and aeration has no significant effect on germination.





50


The growth and development of the seedlings have been described

by Parija (1930), Robertson and Thein (1932), Haigh (1936), and Penfound and Earle (1948). Several authors have indicated that a water saturated medium is necessary for seedling survival (Hitchcock et at. 1949; Parija 1930; Haigh 1936) but forced immersion in water retards growth or kills the seedling (Parija 1930; Hitchcock et at. 1949). Penfound and Earle (1948) and Hitchcock et al. (1940) noted that seedlings would grow on waterhyacinth flotant and Pettet (1964) found them growing on the shore in "strand-lines" created by dead waterhyacinths. Hitchcock et at. (1950) noted that in nature factors which prevent young seedlings from surviving may be more important than factors which permit seed germination.

Vegetative Reproduction
Even though seed production by waterhyacinth may be massive, the

primary mode of reproduction is through vegetative propagation (Hitchcock et at. 1950). This occurs through the production of offsets, or suckers, produced on stolons (Penfound and Earle 1948). Hitchcock et at. (1950) found that offset production begins about 60 days after the plant germinates when the rosette attains a diameter of 7.6 to 10.2 cm. Penfound and Earle (1948) found that a mat extends its boundaries at a rate of 3 feet per month through vegetative reproduction under favorable conditions and the plants double their numbers every two weeks. Bock (1966, 1969) and Perkins (1972) have reviewed the literature dealing with the rates of offset production in different locations and situations.

Productivity and Standing Crop
Many authors have dealt with waterhyacinth productivity in one





51



form or another. Bock (1966, 1969) has done perhaps the most comprehensive study on productivity but she dealt with fresh weight and increment factors making comparisons with her data difficult. She also reviewed most of the literature on the subject and compared it to her data. Table 1 gives an updated compilation of various measures of standing crop and productivity of waterhyacinth from various sources. Control

The literature dealing with the various means of control is voluminous and I won't attempt to review it here. The Hyacinth Control Journal has been published annually since 1962 and is largely devoted to this subject. Furthermore, the various control methods have recently been reviewed. Robson (1974) has reviewed the methods for mechanical control of aquatic weeds and Blackburn (1974) has reviewed chemical control and the various compounds available in a recent UNESCO publication. In the same publication Bennett (1974) reviewed the biological control of aquatic weeds. Biological control has also been reviewed by Andres and Bennett (1975) and the use of plant pathogens in biological control efforts by Zettler and Freeman (1972), Freeman et at. (1974) and Charudattan (1975). Mitchell (1974) summarized techniques for the control of aquatic weeds through habitat management. Sculthorpe (1967) also discussed the various methods of aquatic weed control.





52










Table 1. Standing crop and productivity of waterhyacinths as estimated by various authors.


Source Standing Crop Productivity

Penfound and Earle (1948) 1.6-2.7 kg DW*/m2 Dymond (1949) 1.6 kg DW/m2 13-20 kg/m2/yr Penfound (1956) 0.4-1.3 kg DW/m2 12.7-14.6 g DW/m2/da
5.7-6.5 g C/m2/da

Westlake (1963) ---- 1.1-3.3 kg/m2/yr average 15 kg/m2/yr max. (19 g C/m2/da) Yount (1964) ---- 28 g C/m2/da Bock (1966) ---- 2.5% per day (Calif. average) Misra (1970) ---- 9.4-9.6 g OM*/m2/da (Aug. 1967)
3.48-8.98 g OM/m2/da (Apr.-Feb. 1968) Sahai and Sinha (1969) 0.46-0.72 kg DW/m2 3.8 g OM/m2/da max.
103.0 g OM/:2/yr
247.0 g OM/m2 net production to max. biomass

Knipling et uZ. (1970) 2.4-2.5 kg DW/m2 7.8-16.1 mg CO/dm2 leaf/hr net
2.6-2.8 mg C02/dm2/hr respiration Sinha and Sahai (1972) ---- 1.43 g OM/m2 leaf/hr net
0.56 g OM/m' leaf/hr respiration
1.99 g OM/m' leaf/hr gross

Ornes and Sutton (1975) 9.7 gm/m2 max. 1.05% per day (= 30 gm OM/r,;2/week) max.


*DW = dry weight; OM = organic matter.




53


Arzama densa Wlk.

Taxonomy

Walker (1864) described three genera and three species of moths in two families which are now known to be closely related. These were Edema obliqua (Notodontidae), Bellura gortynoides (Notodontidae), and Arzama densa (Gortynidae). His description of the latter genus and species follows:

Genus Arzama

Male. Body stout. Head with thick-set porrect hairs. Proboscis

short, slender. Palpi stout, porrect, pilose, not extending

beyond the hairs of the head; third joint extremely small,

not more than one-tenth the length of the second. Antennae moderately pectinated, rather short. Abdomen extending much beyond the hind wings, tapering towards the tip, which has a very small tuft. Legs stout, rather short; hind tibiae with

a short fringe; spurs long, stout. Wings rather short and narrow. Fore wings acute; exterior border almost straight, hardly oblique; second inferior vein almost as near to the

third as to the first; fourth not very remote from the third.

Arzama densa

Male. Reddish. Underside, abdomen and hind wings reddish cinereous.

Fore wings with an oblique very broad brownish band, which

contains the orbicular and reniform marks; the latter are red,

oblique, and narrow; a submarginal brown-bordered slightly dentate band, which is rather brighter than the ground hue.

Hind wings beneath with a round brown spot in the disk, and





54

with a slight exterior brownish band. Length of the body

9 lines; of the wings 16 lines.

Grote and Robinson (1868) described a second species of Arzama, A. obliquata. They compared this to Walker's type of A. densa in the British Museun and found they differed in the larger size of A. obliquata and different coloration. They apparently failed to compare it with Edema obliqua,however,

Herrich-Shaeffer (1868: Cited from Zoo. Record) provided generic and specific characters in full for A. densa from Cuba.

Grote (1873) described another species of Arzama, A. vulnifica,

which differed from A. densa Wlk. and A. obliquata G. & R. primarily by its dusky yellow color. He also noted that it was less robust than A. obliquata with the anterior wings more rounded posteriorly. Grote (1874) in his list of the Noctuidae of North America listed only these three species but in 1878 [1879] described a fourth species, A. diffusa from Maine. Gundlach (1881) redescribed Arzama densa Wlk. from specimens collected in Cuba. A fifth species, A. melanopyga, was subsequently described by Grote in 1881 (in Comstock 1881) from Florida. He pointed out characters which separate A. diffusa, A. vulnifica, 4. melanopyga, and Sphida obliquata (apparently a recombination for A. obliquata G. & R.). He noted that characters of the clypeus are of value in separating these two genera. In 1882 Grote synonymized Edema obliqua Wlk. with Sphida obliquata G. & R. These five species were listed together in the subfamily Aezaminae by Grote 1883 who noted that the species with the black anal tuft (melanopyga) is probably a variety of vulnifica. Riley (1885) stated that the genus Sphida Grt. had no existence in nature and Sphida obliquata G. & R. was synonymous with A. densa Wik.






55
In 1889 Grote placed these species in the tribe Arzamini which

included Arzama and Sphida, the former supposedly having a smooth front and the latter a tuberculate front. He considered Arzama as consisting of three species, apparently after having considered A. melanopyga to be a variety of A. vulnifica. This was also how the group was arranged in his checklist of 1890.

Smith (1893) decided that Walker's Bellura and Arzama were congeners and that Bellura had page priority. He considered A. densa Wlk., A. vulnifica Grt., and A. melanopyga Grt. synonyms of B. gortynoides Wlk. He recombined Arzama diffusa Grt. into B. diffusa (Grt.). Edema obliqua Wlk., Sphida obliquata (G. & R.) and Arzama obliquata G. & R. were considered synonyms of Bellura obliqua (Wlk.). Thus, the seven species were reduced to three, all in Bellura Wlk.

Beutenmiller (1902) redescribed B. obliqua (Wlk.) and B. gortynoides Wlk. but he also recognized B. melanopyga. All three were found in New York. Holland (1903) considered B. densa (Wlk.), B. vulnifica (Grt.), and B. meZanopyga to be synonyms of B. gortynoides Wlk. as did Smith (1893) but recognized the genus Sphida and considered S. obliquata (G. & R.) a synonym of S. obliqua (Wlk.). Hampson (1910) recognized the genera Sphida, by the single species S. obliqua, and Bellura. He considered A. densa Wlk. and A. vulnifica Grt. synonyms of B. gortynoides, and retained B. melanopyga and B. diffusa. He also presented a key for separating the three Bellura spp.

Dyar (1913) revised the genus Sphida, described three new species, and he provided a key. He retained S. obliqua (Wlk.) and considered E. obliqua Wlk. and A. obliquata G. & R. synonyms. The new species described were S. oecogenes from Washington, D. C., S. anoa from Miami,





56

and S. gargantua from California. He also included S. pleostigma Dyar and indicated that the description of this species was in a forthcoming paper.

Barnes and McDunnough (1914) considered Sphida Grt. synonymous with Arzama Wlk. thus making obliqua, densa, gargantua, and anoa all species of Arzama. They synonymized S. oecogenes Dyar with A. densa Wlk. but made no mention of S. pleostigma Dyar. They considered A. densa Wlk. distinct from B. gortynoides Wlk. by virtue of a frontal protuberance. Later they described another species of Arzama from New Jersey and named it A. brehmei in honor of its discoverer (Barnes and McDunnough 1916).

Grossbeck (1917) in a list of the insects of Florida recognized B. gortynoides Wlk., B. melanopyga Grt., S. obliqua Wlk., and S. anoa Dyar. Barnes and McDunnough (1917), apparently having identified S. obliqua for Grossbeck, noted that they made their determination before the publication of Dyar's S. anoa and indicated that the specimens they identified were probably S. anoa Dyar. This is confusing, however, because here they are recognizing S. obliqua Wlk. which they had earlier combined with Arzama.

Dyar (1922) re-evaluated the status of the genera Arzama and Bellura. He noted that Hampson (1910) placed A. densa Wlk. as a synonym of B. gortynoides on the assumption that both have a smooth clypeus. He also noted that Barnes and McDunnough (1914) found that the type specimen of A. densaWlk. did have a tubercle on the clypeus and resurrected the genus Arzama making Sphida a synonym of it but considered B. gortynoides Wlk. distinct. Dyar examined several specimens identified as B. gortynoides Wlk. and found that they all had tubercles on the clypeus and suspected that Walker's types would also. He felt this would probably synonymize





57

densa Wlk., gortynoides Wlk., and probably anoa Dyar. He therefore proposed the name Arzamopsis for those species with a smooth front and suggested that A. diffusa be the type species and A. melanopyga be included in the genus. He also described Arzama matanzanensis, a new species from Cuba.

Seitz (1923) again considered all of these species in the genus

BeZlura Wlk. The species listed were B. obliqua (Wlk.), B. densa (Wlk.) (= oecogenes Dyar), B. gargantua (Dyar), B. anoa (Dyar), B. matanzanensis (Dyar), B. pZeostigma (Dyar), B. gortynoides Wlk. (= vulnifica Grt.), B. melanopyga (Grt.), and B. diffusa (Grt.). He noted that B. pallida B. & Benj. and B. brehmei B. & McD. are probably races of B. obliqua (Wlk.) but may be distinct species.

Comstock (1936) discusses this group of insects in his introductory entomology text. He noted that the genus Bellura contained three North American species, B. melanopyga, B. diffusa, and B. gortynoides. He also recognized the genus Arzama and listed A. obliqua as "our most common species". He included these species in the subfamily Apatelinae.

Jones (1951) listed the macrolepidoptera of British Columbia and

included Arzama obliqua (Wlk.) and Bellura gortynoides Wlk. He noted, however, that the latter species is a doubtful record. He synonymized Dyar's Aarzamopsis [sic] with Bellura and considered B. vulnifica (Grt.) a synonym of B. gortynoides Wlk.

Tietz (1952) listed Arzama obliqua (Wlk.) and A. densa Wlk. from Pennsylvania. He considered pallida B. & Benj. a race of A. obliqua (Wlk.), obliquata Grt. a synonym of B. obliqua (Wlk.), and oecogenes Dyar a synonym of A. densa Wlk.




58





Forbes (1954) considered all species of Bellura, Arzama, and Sphida to be in the single genus Arzama. He divided the genus into two groups based on whether the front was flat or had a strong central bulge. In the first group he included gortynoides Wlk., diffusa Grt., and vulnifica Grt. and placed melanopyga Grt. as a synonym of vulnifica. In the second group he included obliqua Wlk., brehmei B. & McD., and densa Wlk.

Kimball (1965) listed the Lepidoptera of Florida and again recognized both Arzama and Bellure. He included A. obliqua (Wlk.), A. [brehmei B. & McD.], A. densa Wlk., A. anoa (Dyar), B. gortynoides Wlk., and B. melanopyga (Grt.). He noted that the two species of BeZZura were probably conspecific. He also referred to the specimen listed as B. gortynoides Wlk. by Grossbeck (1917) and noted that it was actually A. densa Wlk. making the former a synonym of the latter.

The only species listed by Tietz (1972) was Arzama gargantua (Dyar). Levine (1974) noted that B. vulnifica and B. gortynoides are separated largely by the color of their anal tuft, the former being brown and the latter white. He found that dark brown-tailed females (B. fulnifica Grt.) may produce white-tailed daughters (B. gortynoides Wlk.). This indicates that B. vulnifica is merely a form of B. gortynoides Wlk.

I received a personal correspondence from Dr. E. L. Todd from the Systematic Entomology Laboratory of the U. S. Department of Agriculture in April 1974. He explained the taxonomic situation with regard to these species as follows:

I consider that Bellura Walker 1864, type-species
B. gortynoides Walker by monotypy is the valid generic name.
Arzma Walker 1864, type-species A. densa Walker by monotypy
and Sphida Grote 1878 [1879], type-species Arzama obliquata




59




Grote & Robinson (=Edema obliqua Walker), I consider
to be junior synonyms. Bellura has page priority over
Arzama (Walker, 1864, List ..., pt. 32, p. 465 vs p. 645.).
In addition, so far as I can find, J. B. Smith, 1893, Bull.
U. S. Nat. Mus., No. 44, p. 181, was the first to treat
both names and he placed Arzama in the synonymy of Bellura.
Forbes, 1954, Cornell Exper. Stat. Mem. 329, p. 217-8, used
Arzama in error, but divided the genus into two sections (=Subgenera?). However, the character he uses to divide the two sections are invalid. Females of BelZZura do not
have simple antennae as he indicates, and the front may be developed in some forms. The extent of development of the
frons is a character that needs more study. It will also
be necessary to study the possibility that food plant
.varieties are involved. I have indicated to others that
I believe there are only two or three species in the genus,
gortynoides, obliqua, and possibly densa. Smith believed
that gortynoides and densa represented one species, and he
sank that latter as a synonym.

I think that it is apparent from the literature that the taxonomy

of this group is of an uncertain status. I agree with Todd that these

species probably represent one genus and the proper name of Arzama densa

Wlk. is BeZZllura densa (Wlk.). Because of the widespread current use of the

former name and the absence of a definitive study in literature I have

used the binomial Arzama densa Wlk. throughout this dissertation.







Host Plants 60

Table 2 lists host plant records for these species as indicated in various references. Because of the continual changes in the taxonomy of the group, however, these host records are not reliable. For example, Grossbeck (1917) listed BellZura gortynoides Wlk. from Mellonville, Florida, as inidcated by Hampson (1910) and implicated Typha as the host. Kimball (1965), however, indicated that the Mellonville record quoted by Grossbeck (1917) referred to Arzama densa. This creates uncertainty since the host record was not from Hampson (1910), who synonymized densa and gortynoides, but the geographic record was. Grossbeck (1917) apparently derived the host record from other sources. It is therefore impossible to determine which species Grossbeck's host record refers to. To partially alleviate this problem I have left the records in Table 2 with the binomial designated by the respective author intact regardless of synonyms. Where, in my opinion, there is sufficient agreement in the literature to indicate that a name is in synonymy with a more valid binomial, that species designation is included as a subcategory under the valid binomial.

Host plant synonymies also result in a great deal of confusion. For example, Nymphaea americana (Prov.) Miller & Standley listed as a host of Bellura melanopyga (Table 2, No. 2c) is listed by Muenscher (1967) as a synonym of Nuphar variegatum Engelm. Nymphaea advena (Table 2, No. 2) is also apparently a synonym of Nuphar variegatum (Fassett 1969). I do not believe any bona fide record exists of these species attacking any of the Nymphaea species.

I am not sure what the Nel iuitum sp. (Table 2, No. la) and the

Nelumbom sp. (Table 2, No. Ic) records refer to. They may mean Nelumbo but, if this is so, I doubt the veracity of the record. I also question the records for Sagittaria sp. (Table 2, Nos. la, Ic) and Sparganium sp. (Table 2, No.





61





Table 2. Host plants of the B uw-ir'::,a complex listed from various literature sources.


Species Host Plant Source

1. Bellura obLiqua (Wlk.) Typha /latifola L. Beutenmuller 1902; Seitz 1923.
"cattail" Rummel 1919.

a. Arar~ar obliqata G & R Sajittaria sp. Riley 1883a, b; Kellicott 1883b.
Nclwinblurn sp. Riley 1883a, b. Typhia latifulia L. Kellicott 1883b. "cattail reed" Brehlne 1888, 1889. Typha sp. Moffatt 1889. b. Sphida ,bliqua (Wlk.) Typha latijJliia L. Holland 1903; Hampson 1910.
Typha sp. Dyar 1913; Welch 1914; Grossbeck 1917.

c. Ara;arma vbZiqaa'(Wlk.) "corn" Mosher 1919.
'Tpha lit iJ;,lia L. Claassen 1921; Needham, et a. 1928; Guppy 1948; Jones 1951; Tietz 1952. "cattail" Comstock 1936; Forbes 1954; Kimball 1965.
LyUsihitorn ka muthate,nse Guppy 1948; Jones 1951. NWLubim sp. Tietz 1952. Ponntdev'ia orvdata L. Tietz 1952. ;Sagittaria sp. Tietz 1952. S:;parganiwum sp. Tietz 1952. Sylmp uacapue foet idus L. Tietz 1952 Typha sp. Crumb 1956.
d. Ai arena brehmai B. & McD. Typha anju;tijflia L. Forbes 1954.

2. Btllit gortynolid Wlk. Nyaq,haea adoena Robertson-Miller 1923.
Nupha, advena Ait. Levine 1974.
a. Ar arna oulZii i.a Grt. "yellow water-lily" Forbes 1954.
Nuph/a advena Ait. Levine 1974.
b. A. melanopyga Grt. Nuphar adv ena Ait. Comstock 1881.
"bonnett lily" Skinner 1903.

c. Bellura melanopyja (Grt.) "water lily" Hampson 1910; Seitz 1923.
N ruphaea amnr'i~ana (Prov.) Welch 1914; RobertsonMiller & Stand. Miller 1923; Needham, ct. al. 1928.
"pond lily" Grossbeck 1917; Comstock 1936.
Napha, variegatuwn Engelm. McGaha 1952. Nupha adv,'na Ait. McGaha 1952.
3. Bellura diffusa (Grt.) "pond lily" Comstock 1936.

a. Arzama di'ffusa Grt. "yellow water lily" Forbes 1954.

4. Belluix densa (Wlk.)

a. Aeiama d:,saL Wlk. Ponteder'ii 'ordata Forbes 1954; Vogel & Oliver 1969b.
Eiulhhornia ic'anzipes Vogel & Oliver 1969a, b. (Mart.) Solms
Calocasia c sculenta L. Habeck 1974. Typha latifoZla L. Tietz 1952. b. Bellura gotrtyzides Wlk. Typha sp. Grossbeck 1917 (Kimball 1954).
5. Betllur gargantua (Dyar)

a. Sphidca gygantua Dyar 'ypha latifolia L. Dyar 1913; Seitz 1923.

b. Arzanu gargantua (Dyar) Typha latifolia L. Comstock 1944.
'Iypha sp. Comstock & Dammers 1944; Tietz 1952.





62

In general there seems to be three families of plants attacked, the Typhaceae, the Nymphaeaceae, the Pontederiaceae. The Typhaceae are infested by Bellura obliqua (Wlk.) (in Typha latifolia L. and T. angustifolia L.) and Bellura gargantua (Dyar) (in T. Zatifolia L.). The Nymphaeaceae are infested by Beltura gortynoides Wlk. (Nuphar advena Ait. and N. variegatum Engelm.) and B. diffusa ("water lily"). The Pontederiaceae are infested by Arzama densa (Wlk.) (Pontederia cordata L. and Eichhornia crassipes (Mart.) Solms). This supports Todd's (pers. comm.) contention that possibly only three species are involved.

A fourth family, the Araceae, is strongly implicated within the host range of this family. Guppy's (1948) record (Table 2, No. ic) of Arzama obliqua (Walk.) from skunk cabbage (Lysichiton kamtschatcense= L. camtschatcense = L. anericanumwn Hult. & St. John, see Munz 1965) seems to be well founded. Tietz's (1952) citation of SympZocarpus foetidus L. (Table 2, No. Ic) probably refers to Guppy's paper. Habeck's (1974) record of Arzama densa Wlk. from dasheen (Calocasia esculenta L. = Colocasia esculenta (L.) Schott; Table 2, No. 4a) also seems substantiated. These represent two instances of the infestation of two different species of the Araceae from two widely separted regions (British Columbia, Guppy 1948 and Florida, Habeck 1974) by apparently two species of the Bellura complex. Takhtajan (1969) indicates that there is a close affinity between the Liliales (Ponteder.iaceae), Arales, and Typhales and they all are represented along a line of evolution in common with the Nymphaeales.

Mosher (1919) stated that Arzama obliqua has been reported from

corn. She did not cite any references to these reports, however, and I have not been able to substantiate this claim. Because crop plants such




63

as dasheen and corn have been implicated in the host range of this group of insects a great deal of study of host specificity is warranted and the taxonomic status of the group needs clarification. I do not doubt these records but I am dubious of the placement of species identified from these plants. On several occasions I have caged larvae of Arzama densa (Wlk.) on both dasheen (Colocasia esculenta Schott) and Xanthosoma sp. and found that they did not feed upon them. Further studies are severely needed to verify these host records.


Biology and Life History of Arzama densa Wlk. and Related Species.

The early literature on the biology of these species is sparse and occurs primarily as notes of correspondence in various journals. The first reference I have been able to find is that of Worthington (1878). He described the larva of Arzama obliquata (G. & R.) and noted that it was found "under the bark of a dead maple about three feet from the ground, where it had made for itself an oval cavity in the dust". He reared the adult and found that the pupal stadium was about 21 days (April 27 May 18).

Comstock (1881) described the larva and aquatic habits of Arzama meZanopyga Grt. He was the first to take note of the large dorsally situated pair of spiracles on the 9th abdominal segment which are characteristic of the larvae of this group.

Riley (1883a, b) described the eggs of Arzama obliquata G. & R.

(misspelled Arsame) as being laid in "curiously broadly convex or planoconvex masses enveloped in hair, and a cream colored mucuous secretion, when combined look much like spun silk on the inside, and on the outside like the glazed exudation of Orgyia leucostigma." He also noted the




64


large dorsally oriented last pair of spiracles. He stated that there were two annual broods, the second of which hibernated in stumps or moss near the water.

Kellicott (1883a, b), however, felt sure that in New York this

species was single brooded and pupated in May. He also noted that they overwinter in the soil or old wood.

Riley (1883a, b), in reply to Kellicott's comments, stated that there could be no doubt as to the digoneutic (=double brooded) nature of A. obliquata at Washinton (D. C.?).

Comstock (1888) referred to the habits of Azama (misspelled Argama) that infest the leaves of pond lilies. He distinguished these from truly aquatic larvae in that they "are obliged to come to the surface" for air.

Brehme (1888a) described the eggs, larva, and pupa of A. obliquata (G. & R.). He noted a developmental period of about 15 days for the eggs which were laid on cattail between the long leaves. He found the larval period to be 161 days and the pupal period to be 16 days making a total egg to adult span of 190 days.

Brehme (1888a) also stated that the larva returns to the top of the reed in its later larval stages and forms it pupa there. This sparked a series of correspondence in the Canadian Entomologist. Moffatt (1888a) stated that this was not its invariable habit in nature and he had found the pupa beneath the bark of a decaying stump some distance away from where the cattails grew. Brehme (1888b) replied that this may not be invariable but that the majority of them pupate in the reed. He cited a friend of his who had found the pupa in a stump but indicated that the larva had been feeding there and wondered if that wasn't true in Moffatt's




65


case. Moffatt (1888b) replied that there was no evidence that the larvae had fed in the stumps and that all of the larvae and pupae they collected were in similar situations, but admitted to not having looked in the Typha reeds for want of a boat.

Kellicott (1889) referred back to the communications between Riley and himself in 1883 and had decided that they were both right in that A. obliquata G. & R. produces two broods in Washington and one in New York. He also presented evidence confirming Moffatt's contention they they overwinter in stumps as larvae. Brehme (1889) later sent sections of Typha stalk to Moffatt with numerous burrows and two pupae. Moffatt (1889) subsequently reared a pair of the moths from this material. Brehme (1889) felt that Kellicott and Moffatt were mistaken in their assertion that the larvae overwinter in stumps because the specimens he sent to Moffatt were collected in the winter below the water in cattail reeds and some were even under ice. He also disagreed withRiley over the clustering mode of oviposition. He noted that he had always found eggs laid singly and felt that if it were otherwise it would be impossible for several larvae to live in one reed. Johnston (1889) agreed with Brehme as he had found abundant larvae and pupae in cattail in the winter in Ontario. He noted, however, that he had also found them on shore in old wood. He proposed that those on shore were merely wanderers. Beutenmller (1889) described the mature larvae of this species and indicated that he had found full brown specimens under decaying stumps. He later (Beutenmuller 1902) described the larva again under the name of BelZura obliqua (Walk.).

Hampson (1910) repreated Comstock's (1881) description of the larva of Arzama inelanopyga Grt. under the name of BellZZura melanopyga (Grt.).





66

Welch (1914) described the habits of Bellura melanopyga (Grote). He described two feeding periods, first being the leaf feeding period in which the young larvae mine the leaves of Nymphaea americana (=Nuphar variegatum). The second stage is the petiole period which occurs after the larva locates the midrib or the leaf-petiole junction and forms a large burrow in the petiole. He also experimented with other host plants. He found that in a starvation situation they would feed on white water lily, Castalia (= Nymphaea) odorata but when given a choice preferred the yellow water lily. Potamogeton natans and Sagittaria sp. were never attacked. He also discussed respiration, locomotion, and the natural enemies of this species.

Rummel (1919) noted that cattail plants bearing spikes were not infested with BeZZura obliqua.

Claassen (1921) pulbished a detailed account of the insects associated with Typha. Included within this was an excellent study on the biology of Arzama obliqua (Walk.). He found, in New York that there was only one generation per year and that the full grown larva overwinters in its burrow in the plant. He described the mode of oviposition to be in masses, similar to the description given by Riley (1883a, b). He found that each mass contained 35-60 eggs and each female produces ca 225 eggs making about 6 masses per female possible. Upon emergence the first instar larvae enter the leaf directly from the egg chorion. They feed within the longitudinal I-shaped partitions within the leaf mining downward. At the second molt they apparently become too large to feed within these partitions and emerge from the leaves and seek shelter behind the sheath of one of the outer leaves. They ultimately disperse, each to find its own plant, and become solitary burrowers in the stem and rhizome. These,then, exhibit two phases similar to those of Welch (1914)





67

although these should probably be called the leaf mining phase and the stem phase (rather than the petiole phase). He noted that the length of the pupal period averaged 17.6 days and described the egg, first instar larva, full grown larva, pupa, and adult (from Walker).

Robertson-Miller (1923) published many observations on the biology of Bellua gortynoides Wlk. and B. melanopyga Grt. Her information did not differ much from that of Welch (1914). She described the larvae of each and indicated that they did not appear to be much different. She described the egg masses and indicated that those of B. gortynoides were deposited in flat mats of about 20 eggs each. She noted that some of the eggs were covered with silvery white threads. The masses of B. melanopyga were similar to those of B. gortynoides. She found that B. gortynoides may pupate in the petiole, in soil, or in wood. When in the petiole the pupae of B. melanopyga was at the top of the burrow while those of B. gortynoides were lower down. She also found that B. gortynoides would feed of pickerelweed (Pontederia cordata) in captivity.

Needham et al. (1928) repeated the observations of Claassen (1921) on Arzama (= Bellura) obliqua (Wlk.) and of Welch (1914) on Bellura melanopyga Grt.

Comstock and Dammers (1944) described the full grown larva and pupa of Arzama gargantua Dyar. I see no distinction between this description and previous authors' descriptions of these stages of Bellura obliqua (Wlk.).

Guppy (1948) described the habits of Arzama (=Bellura) obliqua
attacking skunk cabbage (Lysichiton kamtschatcense) on Vancouver Island, B. C. He also indicated that they overwintered under loose bark on fallen logs.




68


Crumb (1956) provided a key to the larvae of the Amphipyrinae. The couplet separating Arzama used the large sub-dorsal spiracles on the 8th abdominal segment as a key character. He also described the larva of Arzama obliqua (Wlk.).

Vogel and Oliver published two papers (1969a, b) on Arzama densa Wlk. Their first paper was on the potential of this insect to control waterhyacinth. Their second paper was on the life history of A. densa. They provided cursory descriptions of the immature stages and determined the developmental times of the various stadia. Much of their data, however, is from larvae reared on artificial media which makes their results subject to question. These two papers will be further discussed later in this dissertation.

Levine (1974) found that in Indiana there were two complete generations per year of Bellura gortynoides Wlk. (= B. vulnifica Grt.). He found that the first generation (spring) pupateswithin the petiole of Nuphar advena. The second generation (fall) larvae swim to shore and overwinter as larvae under the bark of trees, in rotten wood, or in leaf litter. The eggs hatch in 6 days and there are 6 to 7 instars.



Parasites, Predators, and Diseases.

The first record of natural enemies which attack this group of

insects was that of Welch (1914) for Bellura melanopyga Grt. He noted that sunfish ate the larvae when they were swimming on the surface. He also observed water striders (Gerris sp.) attacking the larvae when they were on the surface of the leaves.

Claassen (1921) found Sturmia nigrita Town. (Diptera:Tachinidae)

parasitizing the larvae of Arzama obliqua (Wlk.). Robertson-Miller (1923)




69


found puparia of Masicira senilis associated with the burrows of Bellura gortynoides Wlk. Both of these names are probably synonyms of Lydella radicis (Town.) (Stone et al. 1965).

Comstock (1944) made note of the fact that he found no parasites associated with Arzama gargantua Dyar in California.

In the Thompson catalogue (1944) two parasites are listed from

Arzama obliqua (Wlk.). The first is Ceromasia senilis Mg. which may be a misidentification of Lydella radicis (Town.). The second is Pimpla roborator F. (=Exeristes) which is an ichneumonid. I question the veracity of this latter records, however, because the range is listed throughout Europe, Japan, and Guam. As far as I have been able to ascertain the Arzama -Bellura group is strictly New World.

Vogel and Oliver (1969b) listed several parasites and predators of

Arzama densa Wlk. They identified Lydella radicis (Town.) from the larvae, Ichneumon n. sp. and Eupteromalus viridescens (Walsh) (Hymenoptera: Pteromalidae) from the pupae, Telenomus arzamae Riley (Hymenoptera: Scelionidae) and Anastatus sp. (Hymenoptera: Eupelmidae) from the eggs. They also found Coleome!ZllZa maculata De Geer larvae (Coleoptera: Coccinellidae) preying on the eggs and young larvae, and Phyllopalpus pulchellus (Uhler) (Orthoptera: Gryllidae) and Chlaenius pusillus Say (Coleoptera: Carabidae) preying on the larvae.

Levine (1974) indicated that the eggs of first and second generations of Bellura gortynoides Walk. are also parasitized by Telenomus arzanae Riley and the second generation larval populations are parasitized by an ichneumonid and have a polyhedrosis virus.








CHAPTER I

THE RELATIONSHIP BETWEEN THE PHENOLOGY AND PRODUCTIVITY OF WATERHYACINTHS AND VARIOUS PHYSICAL AND BIOLOGICAL FACTORS.

Introduction

To evaulate the effects of insects for the biological control of weeds, a basic understanding of the ecology of the plant is essential. In realization of this the Canada Weed Committee has instituted a series on the biology of Canadian weeds (Cavers and Mulligan 1972). This is an attempt to pull together all the available knowledge on the biology of Canadian weeds that can be used in weed control efforts. Within this framework the phenology of the plant (annual variation), and the response of the plant to limiting factors and damage by indigenous insects is of special interest for the evaluation of biological control attempts. Omission of these considerations could result in the misinterpretation of pertinent data. For example, natural seasonal declines in the plant population could mistakenly be attributed to insect releases when the insects are also seasonal if patterns of seasonal variation of the plant are not known. Also, releases of insects may be more effective when correlated with critical periods in the annual cycle of the plant. Judgements for the timing of these releases can be made only on the basis of what is known about the plant.

Limiting factors can be defined as the necessary components of

the organism's environment which are least available and thereby control the life processes of the population. Liebig (1840) stated that a process is limited by the quantity of a single component present in minimal amounts relative to its optimal amounts. Sachs (1860) felt that biological processes required a certain minimal level of a limiting

70




71


factor to begin, attained an optimum at a certain level, and declined as levels of the limiting factor exceeded maximum tolerable levels. This is parallel to Shelford's (1913) "Law of Tolerance" where he essentially states that the failure of an organism may be due to an excess or deficiency of any one factor which may approach the maximum or minimum limits of tolerance of the organism for that factor. For an aquatic plant, such as waterhyacinth, these limiting factors include temperature, light, water, dissolved or available nutrients, space, etc.

Phytophagous insects probably cause a threshold type response in

the plants whereby the plant can sustain certain levels of damage without obvious deterioration until maximum tolerable limits are exceeded. As insect damage exceeds these threshold levels a rapid decline in the populations or standing crop may be evident. Levels of insect damage below this threshold may cause various plant response. When the population is at steady state (the stable maximum level restricted by the level of a limiting factor) insects may disrupt this stability causing the plant population or standing crop to fall below the carrying capacity of the system. This may have the effect of reducing intraspecific competition in the plant population. In this case the limiting factors would become increasingly more available and production may indirectly be stimulated. Hence yield may be increased under low insect concentrations where the insects prevent senescence of the population by increasing the rate of turnover.

This study was designed to measure the effects of various environmental factors as well as the effects of a natural buildup of an indigenous insect population (Arzama densa) on a stand of waterhyacinth. The parameters considered can be grouped into climatological conditions (temperature and solar radiation), limnological conditions (nutrients,




72



water quality and water level), intraspecific conditions (plant density, canopy effects, available space, etc.), and biotic stress (insect damage). These will be evaluated with possible interactions between them considered.

These concepts, possible interactions and all factors which control the plant must be considered and investigated. Attempts to evaluate the attack of an insect by studying only the insect or with only a superficial knowledge of the target plant are subject to erroneous conclusions and misinterpretation. Not only must the plant and the insect be studied but the insect-plant interrelationships must be established. This field has received increasing attention lately and may provide a basis for future biological control efforts.




73




Methods And Materials


Duirnal Waterhyacinth Productivity

Waterhyacinths of two distinct morphological types from the "open" side of the catwalk on Lake Alice were selected for in situ metabolism studies. Large plants, approximately 90 cm tall, with elongate petioles were measured for CO2 uptake on 11-12 August 1973. Small plants (<50 cm) with bulbous petioles were measured on 12-13 August 1973, A section of the mat approximately 0.5 m2 of each type was placed under a chamber constructed of a PVC pipe frame covered with clear poly-acetate. The base of the chamber was 71 cm x 71 cm (ca 0.5 m2).

Air was passed through the chamber with a blower and duct system. The duct entered the chamber at the base on one side. Air was supplied to the blower intake through a tube opening approximately 3 m above the water surface so the CO2 concentration would not be influenced by the plants surrounding the chamber. The rate of air flow was determined with a Hastings hot wire anemometer. The air was discharged from the chamber through a duct located on top.

Carbon dioxide concentrations were monitored at the chamber air intake duct and at the exhaust duct. The air at each location was collected through tubes which extended to a Beckman infra-red CO2 gas analyzer. Air flow was also measured at the intake and exhaust. This enabled the determination of the ppm C02/unit of air/time entering and leaving the chamber. The differential is the amount of C02 produced or consumed within the chamber.

The C02 analyzer readings had to be calibrated against a standard





74


to convert from a scale reading to ppm COZ. The scale reading is based on a comparison of two gases. Three pairs of gases were compared through the analyzer. Ambient air vs. ambient air (air entering the chamber) was compared to determine a zero point. The second comparison was chamber exhaust vs. ambient air. This difference represented the CO2 gradient through the chamber and was expressed as recorder scale division. The value of a scale division (sd) is determined according to the level of CO2 in the ambient air by the equation ppm/sd = aebx where x is the CO2 concentration of the ambient air. The ambient air C02 concentration was determined by a third comparison. In this case a standard was used of a known concentration. The standard was 300 ppm bottled gas and was compared against the ambient air. The CO02 concentration in the ambient air was determined by the equation ppm = ax2 + bx + c where x is the recorder reading. This involves lowering the amplification of the analyzer output by changing from "range 3" to "range 1". The range 1 equation is calculated by running the standard 300 ppm gas through the reference side of the analyzer and running other gases of known concentration through the sample side. The value of the sample gas is correlated with the recorder reading using the parabolic regression. Range 3 is calibrated using various known CO2 concentrations against a closed system aparatus with flow and pressure maintained constant. The closed system is injected with known quantities of pure CO02. An exponential regression is fitted for the ppm/sd against ppm CO2 of the various reference gases (ambient air in this case).

The volumetric CO2 concentration gradient (ppm C02) is converted
a gravimetric measurement (g C/m2) using the gas constant (0.14625 gm C.KO/m2




75



atm ppm C02). When multiplied by the flow rate this expression yields the rate of carbon metabolism (g C/hr) within the chamber. A more detailed explanation of this system is given by Carter et al. (1973).

Carbon metabolism for each type of plant was measured for 24 hrs.

Integration of the resultant production curves yielded both gross primary productivity and respiration. Respiration was assumed to be constant both day and night and was determined as the average nighttime value. Net production consisted of that portion of the curve above the compensation point (where Pg = R and Pn = 0). solar radiation was measured with a Weathermeasure Co. 24-hr. pyroheliograph in the 0.36-2.5pm range. Air temperature was recorded using a Yellow Springs Instrument thermistor apparatus.

Following the metabolism measurements the plants were harvested to obtain a biomass estimate. The total sample was divided into leaves, petioles, roots (= roots + rhizomes + stolons) and detritus and the various plant parts were weighed while fresh. A similarly divided subsample was taken and weighed before and after drying. From this subsample a wet to dry conversion factor was obtained so that the dry weight for each plant part and the total sample could be obtained. A subsample of the leaves (pseudolaminae) and petioles was pressed in a plant press and dried. The outlines of the dried leaves were traced on paper and the area measured with a planimeter. This determined a leaf area per gram of dried leaf conversion factor and the leaf area of the total sample was estimated from this. A similar procedure was employed with the petioles. From this the leaf area index (LAI) was determined which, in this case, is the total leaf area (leaves + petioles) per square meter as determined from only one side of the leaf.





76



Annual Cycles and Insect Damage

Estimates of various plant characteristics, of the Arzama densa

population, and of plant damage by A. densa were taken on a weekly basis from May 1974 to 30 April 1975. Sampling was done on a plot system using a rubber ring enclosing an inside area of 0.316 m2. The samples were taken each week in a pseudo-random manner. I have not been able to devise a satisfactory system of pinpointing a previously randomly selected point on a mat of waterhyacinths and then finding that point while trying to maneuver through the dense stand of plants. To eliminate the additional variables of seasonal plant species composition changes along the shoreline and different waterhyacinth growth characteristics only the central area of the lake was studied. The area in which samples were taken was defined by the catwalk on the west side and extended 25 m to the north and 25 m to the south of the central point on the catwalk. The eastern boundary was established by a small row of bushes 50-60 m from the catwalk that extended into the lake from the north shore. The study area, then, was 2500-3000 m2 in the central more or less homogenous region of the waterhyacinth mat on the marsh side of the catwalk. Sampling points were selected by throwing the ring in a high arc. After it fell into the mat it was reached using two Dow styrofoam billets (3 m X 0.5 m), one placed in front of the other sequentially. This allowed me to move (with some effort) over the mat on the water surface. Once the ring was reached the billets were used as platforms for counting and recording data.

The ring was manipulated down over the waterhyacinths until it was on the water surface. This involved making siu~1 tive decisions as to





77


which plants were inside and which plants were outside the sample. If the crown was in the ring the whole plant was considered in. This was still difficult to determine when the crown straddled the edge. The placement of these plants was left up to the discretion of the sampler. This border effect was probably the largest within sample source of error in the plot sampling.

Once the ring was placed each plant was withdrawn and measured. An offset was considered a separate plant only if the root system was developed. Measurements.. taken were the height of the plant, based on the distance from the tip of the longest leaf to the point where it attached to the rhizome, and the number of leaves per plant. Leaves were counted only if half or more on the pseudolamina was alive and unfurled. The number of plants in each plot was tallied to establish plant density. Each plant was carefully dissected and damage by Arzama densa noted. Damage was distinguished according to the degree of severity. Leaf damage was classified as to external feeding or petiole bores. Rhizome damage was classified as tip damage, rhizome bore, or rhizome fragmented. If a larva or parasite was found it was placed in a pill vial, given an identifying number and returned to the laboratory. The insect data will be discussed in a separate section of this dissertation. One sample required 3-4 man-hours. Three samples were taken each week.

Leaf area estimates were also taken weekly but on a different day. Ten plants were randomly selected along the catwalk for this measurement. The randomization procedure consisted of numbering the supporting pilings of the catwalk within the study area. Ten pilings were selected from generated random numbers. At each selected piling a single plant was





78



picked. This required a second randomization. One person involved in the process held a second generated random number between 1 and 10. A second person drew up to ten plants out of the water. When the nth plant (where n = the random number) was pulled out the first person notified the second and the plant was placed in a plastic bag and returned to the laboratory.

Before measuring the leaf area the petioles and leaf blades (pseudolamina) were separated. The petioles were rolled out with a rolling pin. This was necessary to compensate for the cylindrical shape of the petiole. The outline was then traced on a piece of drawing paper and measured with a planimeter. The leaf blades were pressed in a plant press and dried. They were then traced and measured the same way. I found that drying the leaves caused considerable shrinkage and a dry:fresh conversion factor had to be employed. The formula for this conversion was:
leaf area (fresh) = 1.437 X leaf area (dry)

The conversion factor was determined by measuring one sample (51 leaves) before and after drying. Each week figures for the average leaf (pseudolamina) area, average petiole area, and average total (pseudolamina + petiole) leaf area were derived. These figures were multiplied by the number of leaves per square meter from the plot samples to obtain the leaf or petiole area index. This figure represents the leaf or petiole surface area (considering only one side) per unit of substrate area (m2/m2).
Water samples were taken along the catwalk at the mid-point. The
water was collected a few centimeters below the surface at the level of the waterhyacinth roots. This level should best reflect the conditions the plants were being subjected to. The sampling station was on the downstream Fs of the study area so the tests would reflect minimum




79




nutrient levels. Two samples were collected each week. One sample was analyzed using a Hach DR-EL portable test kit for total alkalinity (carbonate + bicarbonate), total nitrates + nitrites, pH, total phosphates, and sulfates and a Hach micro-iron test kit (model IR-18-A) for iron. The second sample was taken to the University of Florida Soils Laboratory where it was analyzed for conductivity, magnesium, and potassium.

The methods applied to the water samples are as follows:

Alkalinity (total) Titration of Brom Cresol Green Methyl

Red indicator with 0.020 N. sulfuric acid 10 ml sample.

Conductivity Platinum electrode ohmmeter.

Iron 1, 10 Phenanthroline Method 25 ml sample.

Magnesium Atomic absorption spectrophotometer.

Nitrates and Nitrites (total) Cadmium reduction method

25 ml sample.

pH Colorimetric reading with a wide range indicator.

Phosphates (total) Colorimetric method 25 ml sample.

Potassium Flame emission spectrophotometer.

Sulfate Turbidimetric method.

The procedures employed in the Hach Test Kit are more for convenience and direct reading and are not as accurate as other techniques. For my purposes the loss in accuracy is outweighed by simplicity of the procedures. These procedures are probably accurate enough to indicate temporal differences but are probably not extremely definitive.

Maximum and minimum air and water temperatures were taken at the

same location as the water samples. Two Taylor (No. 5458) maximum-minimum self registering thermometers were mounted on a C-shaped styrofoam




80



block. The water temperature thermometer was placed vertically on the bottom of the block. The air temperature thermometer was placed vertically in the concave side. The block was mounted on a rider which slid over a piece of pipe which extended into the lake bottom. This allowed the thermometer to move up or down as the water level changed. The bulb of the underwater thermometer was about 4 cm below the surface and measured the conditions the submersed plant portions were subjected to. The air thermometer bulb was about 30 cm above the water surface and measured conditions under the leaf canopy. The concave side of the block was oriented towards the north so as to avoid direct exposure to the sun. The overhang on the block also helped prevent this.

Water level was measured at the northwest corner of the study area from a depth gauge established there previously by other investigators. Solar radiation data was obtained from Dr. E.A. Farber of the solar energy laboratory at the University of Florida.




81



Site Description

Lake Alice is located in the southwest corner of the University of Florida campus in Gainesville, Alachua Co., Florida (Topographic designation: Gainesville East quadrangle, T1OS, R19E, NE R20E, NW ). The lake was once a sinkhole fed by a small stream but damming off the west end in the late 1940's and later the addition of effluents from the campus sewage treatment plant and the heating plant resulted in its present configuration (see Figure 1). The lake area is approximately 33 ha and is divided into a marsh dominated by waterhyacinths and an area maintained by the University as open water. The marsh at the east end comprises approximately 65% (21 ha) of the lake surface and is separated from the open lake by a catwalk and fence constructed to retain the waterhyacinths. The depth of the marsh in generally less than 2 meters (Cason 1970). The "open" western end of the lake covers about 12 ha and is also generally less than 2 meters in depth with a few areas of about 5 meters, probably the original sinkholes (Mitsch 1975). The general flow of the lake is from the sewage plant and heating plant effluent at the eastern end through the marsh to the open lake at the western end where it discharges through two wells into the Florida aquifer.

The lake is situated on Ocala limestone which is dominated by a

karst topography. Solution sinkholes, fractures, and caverns are typical of this type of topography and are common in this area. Because of the silt that has accumulated on the bottom, however, the lake basin is maintained above the local water table (Cason 1970). The lake level is generally between 68 and 70 feet above mean sea level. Figure 2 indicates the lake level at the catwalk for the period of this study.























S.
o

u

S .- 4O





S4-, DU






LL.




4- U
-
LL4- 0 4-)




'- 4U
-) 3

U)~4



0r
4- C, u) 4'- )
c








( ) 4-) 4




S- ro c>





> V) (U





to 4-) r
O4
















LL
'r-'






83


















4Ly



























* A P:,

;4,r EL






84


As mentioned previously Lake Alice receives effluent from the campus sewage treatment facility and cooling water from the heating plant. The nutrient enriched water from the former and the warm water from the latter have probably contributed significantly to the eutrophication of this lake. Other sources include overflow from Hume Pond, also located on the university campus, runoff from the local watershed, and direct rainfall. Discharge of thesystemis through the wells mentioned earlier. Water loss also occurs through surface evaporation and evapotranspiration. Mitsch (1975) estimated the hydrological budget for the lake in terms of flows and storage (see Table 3). The water storage at a stage of 69 feet above mean sea level is estimated to be 254 x l03 m3. Water retention is low due to the high input-volume ratio. Brezonik et al.(1969) suggests that this may have a flushing effect causing low phytoplankton populations noted in the lake.

Discharge through the wells is regulated by valves and is frequently altered by campus personnel. The water level is often raised to facilitate mechanical removal of the waterhyacinth. During the hurricane season the water level is dropped to prevent flooding. Fluctuations are also caused when the discharge screens over the wells become clogged with debris. Water level appears to be correlated with seasonal precipitation patterns (Fig. 2) except for the months of December and January. During this time an oil spill occurred in the canal from the heating plant and sewage treatment facility. Flows from these two sources were minimized so the oil could be cleaned up. This resulted in a sharp drop in the lake level. Normal flow was restored the first part of February and a sharp increase in the lake level followed.




Full Text
Figure Pa9
45. The mass represented by the various plant parts for an
average waterhyacinth plant at both the control and
release sites 262
46. Probit analyses for developmental times of a greenhouse
reared population of Arzama densa 282
47. The head capsule diameter of Arzama densa larvae at each
molt plotted against the larval age 285
48.. The total number of Arzama densa larvae collected, either
living or dead, from the marsh side of Lake Alice 289
49. The age structure of the Arzama densa population during the
period of this study 29T
50. The population of living Arzama densa larvae (per square meter)
present on Lake Alice and the number of dead larvae expressed
as a percentage of the total 296
51. The total number of 4th and 7th instar Arzama densa larvae
per square meter as estimated from samples taken from the
marsh side of Lake Alice 299
52. The number of parasites of 4th instar (Campoletis sp.) and
7th instar (Lydella radiis) Arzama densa larvae as
estimated from the number of pupae, or pupal exuviae found in
A. densa bores per square meter of waterhyacinth mat . .302
xi i ¡


34
layer of lacunate tissue in between. The polyarch stele is surrounded by
a weakly differentiated endodermis and pericycle (Couch 1971). Limited
meristematic activity occurs at the apex and the roots possess a distinct
epidermis (Sculthorpe 1967). The roots are known to become embedded into
the mud although they are usually suspended freely in the water (Buckman
and Co. 1930; Small 1936; du Toit 1938; Penfound and Earle 1948; Parsons
1963). The color of the roots is normally dark violet blue due to the
presence of anthocyanin but when growing in the mud or in the dark the
roots become white (Olive 1894; Penfound and Earle 1948). The roots arise
from nodes on the rhizome (Penfound and Earle 1948).
The rhizome is the vegetative stem of the plant from which all other
structures arise (Penfound and Earle 1948). It consists of an compact
axis with short internodes and the leaves, roots, stolons and inflores
cences are produced by the meristematic nodes which have generally small,
compact cells. An area with a considerable number of intercellular air
spaces exists around the periphery of the meristematic tissue (Couch 1971).
The rhizome is approximately 95% water by weight and has a specific
gravity of 0.805 (Penfound and Earle 1948). Olive (1894) indicated that
there was no evidence of starch being stored within the rootstock (rhizome)
but Penfound and Earle (1948) felt that the rhizome was the main organ
of starch storage.
The rhizome may produce long horizontal internodes (stolons) which
/
produce new shoots at the distal end which results in a sympodial branch
ing pattern (Penfound and Earle 1948). These stolons arise from axillary
stem buds (Bock 1966). Aerating spaces are abundant near the periphery
and the collateral bundles are aggregated in the center (Olive 1894).
The stolons are purple due to the presence of anthocyanin and range in
diameter from 0.5-2.0 cm at i length up to 40 cm (M ra 1969). The


216
curve until it levels off at a point where the increase in detritus is
equal to the initial living material present. Figure 35 illustrates
this somewhat differently as an almost linear relationship between the
insect concentration and the amount of detritus present expressed as
a precentage of the total organic material present. This linear rela
tionship should hold until the ratio; of detritus to biomass approaches
unity. At this point an asymptote in the curve should become apparent
where higher concentrations of insects have a proportionately smaller
effect on this ratio (a ratio of greater than 1.0 is impossible). If
the net productivity is zero and the detritus: total biomass ratio is
1.00 then all of the plants were immediately killed upon release of the
insects. If this ratio is 1.00 and the net productivity is some value
greater than zero then all of the plants were killed at some time after
the initial release.


257
Various characteristics of the waterhyacinth populations and
estimates of insect damage are compared statistically in Table 12. All
of the plants counted at the control site were alive while 35% were
dead at the release site. Only 17% of the control plants had damage
to the rhizone while 73% of the plants had rhizome damage at the release
site. Further, only 6% of the control leaves were damaged as compared
to 39% at the release site. Plant density and leaf density were not
significantly different between the two sites but both height and leaves
per plant did decrease significantly (33% and 31% respectively). This
indicates that the plants were smaller as a result of insect attack but
not necessarily fewer in number.
Biomass estimates could not be compared statistically since only one
sample per site was taken. Differences in biomass between the sites
were obvious, however, and are illustrated in Figure 44. The changes in
biomass were much greater than any of the morphological characteristics
in Table 12. Standing crop (total living plant weight) at the release
site was only 25% of that at the control site. The change in stolon weight
was greatest with a demonstrated loss of 86%. This indicates a lack of
vegetative growth in the infested plants since stolon production is
necessary for offset production. Petioles (leaf bases) and rhizomes de
creased about 80% while roots decreased only 65%. The total photosynthetic
tissue (leaves and petioles) decreased more than the non-photosynthetic
tissue (roots, rhizomes and stolons). At the release site these were 21%
and 31% respectively of their values at the control site. This difference
appears to be due to the smaller change in the roots, the only part not
attacked by Avzama.


46
1-16. Misra (1969) indicated that the average may be 15-41. Robertson
and Thein (1932) found 50-150 seeds/capsule in Burma, Zeiger (1962)
reported 3-250, and Francois (1964-3) reported an average of 153.6 with
a maximum of 244. Tag el seed (1972) reported an average of 98.95 with
a range of 5 to 542.
The number of seeds per inflorescence depends upon the number of
flowers per inflorescence and the number of seeds per flower. Bock (1966)
indicates that the average number of seeds per inflorescence is 3.44 in
California. Tag el Seed and Obeid (1975) reported 1.5 capsules per
inflorescence. Using the data from Tag el Seed (1972) for the average
number of seeds per capsule (98.95) this expands to 148 seeds per inflor
escence. Matthews (1967) indicated that a single spike may produce 5000
to 6000 seeds. Zeiger (1962) estimates that 45 million seeds may be
produced per acre by medium sized plants. Penfound and Earle (1948)
estimated a crop of 900,000 capsules/acre.
The mode of seed dispersal has not been studied to any extent.
It seems apparent that since the seeds are deposited in the water the
primary mode of dispersal would be through drifting. A few authors have
indicated that birds and fur bearing animals may disperse seeds (Maiden
et al. 1906; Holm et al. 1969; Gay 1960).
Viability of Seeds and Germination
Conditions for germination of waterhyacinth seeds have been
studied by many workers. Crocker (1907) refuted the idea of earlier
workers that desiccation of the seeds is a necessary prerequisite to
germination. He further found that green seeds kept at 23C germinated
(20% within 1 week) while seeds kept at 6C did not, although they did


186
potential effect of Arzama densa on the ecology of the waterhyacinth
in terms of net productivity, standing crop, turnover rates, and other
characteristics of the plant community during two distinct seasons
(summer and fall).


16
Misra (1969) cited a source which indicates that the center of
origin for this species was probably the Pernambuco region of Brazil.
A few authors have subscribed to other regions of origin outside of South
America. Hildebrand (1946, p.477) states "The water hyacinth, Eichhornia
orassipes, is a native of Japan and was carried about 70 years ago to
South America, where it became widespread in fresh-water streams and
lakes." He cites Gowanloch (1944) as the authority for this statement.
Gowanloch apparently contradicts himself, however. In one paragraph
he does indicate that waterhyacinth is native to Japan and was imported
to South America. In the following paragraph he states, "When in 1884
an International Cotton Exposition was held in New Orleans, the Japanese
Government representatives in their building on the Esposition grounds
gave away as souveniers water hyacinths which they had imported from
Venezuela." Small (1933) suggested that it may be native to Florida.
A Ceylonese author claimed that Florida was its area of origin (Bock 1966).
Waterhyacinth is well known from the West Indies (Bock 1966).
Castellanos (1959) includes the Antilles within the range of distribution.
Bancroft (1913) also indicated that the plant was present in the West
Indies. Bock (1966) cites a paper which indicates that Puerto Rico is
the center of dispersal for this species. She also found it naturalized
in Jamaica. She suggests that it may have spread to the islands attached
to boats or by floating from the mainland.
As might be expected waterhyacinth is also known from most of the
Central American countries including Panama (Standley 1928; Hearne 1966),
Costa Rica (Little 1965), Nicaragua (Little 1965, 1966; Holm et al. 1969),
Honduras (Castellanos 1959), and El Salvador (Little 1965, 1966; Holm et al.


85
Table 3. Hydrological budget for Lake Alice (March September
1973). (From Mitsch 1975)
Source
Input*
Output*
Sewage Flow
10.5

Heating Plant Flow
43.5

Hume Stream
8.6

Direct Rainfall
0.8 2.0
--
Runoff
6.1 21.0

Transpiration

1.8 3.1
Evaporation

0.4 0.6
Discharge

58.0 76.8
TOTAL
69.5 85.6
60.2 80.5
*Flow x 103 m3/day


329
SAS. 1972. A user's guide to the statistical analysis system. North
Carolina State University, Raleigh. 260 pp.
Schwartz, 0. 1928. Die Pontederiaceen. Pflanzenreale 2: 13-14.
Sculthorpe, C. D. 1967. The biology of aquatic vascular plants. Edward
Arnold (Publ.) Ltd., London. 610 pp.
Seabrook, E. L. 1962. The correlation of mosquito breeding to hyacinth
plants. Hyacinth Control J. 1: 18-19.
Seitz, A. 1923. The Macrolepidoptera of the world. Alfred Kernen Publ.,
Stuttgart.
Sen, N. N. 1961. A note on the eradication of water hyacinth in the Ghana
bird sanctuary, Bharatpur Rajasthan. Indian Forest. 87(3): 168-169.
Sheffield, C. W. 1967. Water hyacinth for nutrient removal. Hyacinth
Control J. 6: 37-30.
Sheikh, N. M., S. A. Ahmed, and S. Hedayetullah. 1964. The effect of the
root extract of water hyacinth (Eiohhomia speoiosa Kunth), on the
growth of microorganisms and mash kalai (Phaseolus mungo var. Roxbuvghii) ,
and on alcoholic fermentation. Pakistan J. Sci. Ind. Res. 7: 96-102.
Shelford, V. E. 1913. Animal communities in temperate America.
University of Chicago Press, Chicago.
Shibata, M., K. Yamazaki, and N. Ishikura. 1965. Eichhornin, a new antho-
cynin isolated from the flower of the water hyacinth. Bot. Mag. Tokyo
78: 299-305.
Silveira-Guido, A., H. S. Montero, and J. C. Bruhn. 1965. U.S.D.A. final
project report S9-CR-1. Natural enemies of weed plants. Montevideo,
Uraguay.
Silveira-Guido, A., and B. D. Perkins. 1975. Biology and host specificity
of Cormops aquatioum, a potential biological control agent for water-
hyacinth. Envir. Entolmol. 4(3): 400-404.
Sinha, A. G., and R. Sahai. 1974. Contribution to the ecology of Indian
aquatics. IV. Rate of dry matter production of the leaves of several
common aquatic plants of Gorakhpur (India). Photosynthetica 8(2): 127-129.
Sircar, P. K., S. Banerjee, and S. M. Sircar. 1973. Gibberellin-like
activity in the shoot extract of water-hyacinth. (Eiohhomia cvassipes
Solms). Indian J. Agr. Sci. 43(1): 1-8.
Sircar, S. M., and R. Chakravarty. 1961. The effect of growth-regulating
substances of the root extract of water hyacinth (Eiohhomia speoiosa
Kunth) on jute (Corohorus capsularis Linn). Current Sci. 30(11: 428-430.


38
whorls of 3 stamens each, of two different lengths and are adnate being
fused to the corolla tube at the base of the filaments near the sinuses
of the perianth lobe (Bock 1966). The filaments are white at the base,
purple at the apex, and glandular (Misra 1969). The anthers are oblong
and attached near the base (Bock 1966) and contain about 2000 pollen
grains each (Penfound and Earle 1948). The ovary is superior, sessile,
trilocular, contains numerous ovules in axile placentation (McLean 1922;
Bock 1966) and conical in shape (Penfound and Earle 1948). The six
stigmatic surfaces, by their close approximation, appear to be capitate
but they are not (Penfound and Earle 1948; Bock 1966). The stigmate
surface is covered with numerous glandular hairs (Bock 1966). The fruit
is a loculicidal capsule containing seeds with an abundant mealy endo
sperm (McLean 1922). Approximately 50 seeds are produced per capsule
(Penfound and Earle 1948).
Perennation
Waterhyacinth is generally considered to be a perennial by virtue
of its rhizome (Penfound and Earle 1948; Sculthrope 1969). Penfound and
Earle (1948) felt that the rhizome may maintain a constant length over
aj3£rQt_ may exist is not available but the plants are known to _survive periods
of freezing weather by resprouting from the rhizome (Bock 1966).
Physiological Data
Most of the appropriate physiological data has already been discussed
in scattered sections of this literature review but it bears repeating in
a more organized discussion.


249
of each site was collected to estimate biomass. All living and dead
material from this sample was placed in a plastic bag and returned to
our laboratory. The plants were divided into petioles (leaf bases),
leaves (pseudolaminae), rhizome, roots, stolons, and detritus (dead
plant material). They were placed in ice cream cartons and dried to a
constant weight in an oven at 105 C for 2-3 da. After drying, the
containers were allowed to cool to room temperature and then weighed
on a Mettler to^loading balance.


Figure 6. The negative logs of the hydrogen ion concentration (pH)
of water samples taken from Lake Alice. The line is a
5-point moving average.


PROBIT
282


66
Welch (1914) described the habits of Bellura melanopyga (Grote).
He described two feeding periods, first being the leaf feeding period
in which the young larvae mine the leaves of Nymphaea americana (=Nuphar
variegatum). The second stage is the petiole period which occurs after
the larva locates the midrib or the leaf-petiole junction and forms a
large burrow in the petiole. He also experimented with other host plants.
He found that in a starvation situation they would feed on white water
lily, Castalia (= Nymphaea) odorata but when given a choice preferred the
yellow water lily. Potamogetn natans and Sagittaria sp. were never
attacked. He also discussed respiration, locomotion, and the natural
enemies of this species.
Rummel (1919) noted that cattail plants bearing spikes were not
infested with Bellura obliqua.
Claassen (1921) pulbished a detailed account of the insects associated
with Typha. Included within this was an excellent study on the biology
of Arsama obliqua (Walk.). He found, in New York that there was only
one generation per year and that the full grown larva overwinters in
its burrow in the plant. He described the mode of oviposition to be in
masses, similar to the description given by Riley (1883a, b). He found
that each mass contained 35-60 eggs and each female produces ca 225 eggs
making about 6 masses per female possible. Upon emergence the first
instar larvae enter the leaf directly from the egg chorion. They feed
within the longitudinal I-shaped partitions within the leaf mining
downward. At the second molt they apparently become too large to feed
within these partitions and emerge from the leaves and seek shelter behind
the sheath of one of the outer leaves. They ultimately disperse, each
to find its own plant, and become solitary burrowers in the stem and
rhizome. These,then, exhibit two phases similar to those of Welch (1914)


C^J


215
reaches the 1.00 level. This probably reflects the carrying capacity
of the plant community for this herbivore. The carrying capacity is
probably dependent upon the initial standing crop as well as the pro
ductivity of the plant, i.e., the rate at which the community can
replenish itself. The plants in the summer are more productive than
those in the fall and, as a result, can support a greater herbivore
population. The point at which the herbivores begin to have a negative
effect on yield is the response threshold. In the fall this threshold
occurs between the 0 and 0.33 levels. In the summer, however, the
threshold apparently is much higher at between the 0.67 and 1.00 levels
of infestation.
The change in the amount of detritus (dead organic material) present
is probably a good indicator of insect feeding activity provided that it
can be measured accurately and identified as waterhyacinth debris. The
net change is dependent upon the living material available during the
period of infestation, the level of insect feeding activity, and the rate
of degradation of the detritus by decomposers. Assuming that the decompo
sition rate per gram detritus is constant the amount of dead organic
material should directly indicate insect feeding activity up to the point
where it becomes limited by the amount of living material available for
conversion. A detrital response curve regressed on insect concentration
(Figure 34) would be expected to increase up to a point, tend to level
off, then show a rapid decline. The point of deflection for this curve
should occur at the point where plant productivity begins to be reduced
by the feeding activity of the insects. As insect concentrations become
larger (beyond the range of this experiment) this point would be reached
earlier in the growing period of the plant and cause a decline in the




Figure 14. Annual phenological change in the average height of the
waterhyacinth plants on the marsh side of Lake Alice. Height
was measured as the length of the longest leaf from the tip
of the pseudolamina to the base of the petiole at its attachment
to the rhizome. The mean is derived from all plants contained
within three 0.316 m2 samples. The dotted line represents pre
dicted values based on multivariate regression equations (see Table 7).


no
Even though frost did not severely damage the plants in the winter
the effect of the low temperatures may have been that of inhibiting
translocation of starch as described previously. Winter daytime temper
atures were generally quite warm but nighttime temperatures often fell to
the 10C range or below. If photosynthetic rates and carbohydrate trans
location are reduced in this temperature regimen a reduction in growth
would be expected during this time. Bock (1969) found that plants grown
at 4.4C nights and 4.4C days (16:8 L:D photoperiod) grew very little
(ca. 5%) during a 25 day test. Those grown at 26.7C days and 4.4C
nights did only slightly better (ca. 20%) while those grown at 26.7C
days and nights increased the most (ca. 105%). Apparently minimum
nighttime temperatures have a substantial effect on waterhyacinth growth
and even though daytime temperatures may be warm growth may be suppressed.
The inability of the plant to translocate starch in these low nighttime
temperature conditions along with the subsequent reduction in photo
synthetic efficiency seems to be the only explanation available for
this phenomenon and is probably a result of an interaction between temp
erature effects and physiological mechanisms (Knipling et al. 1970).
Figure 9 illustrates the annual curve for solar radiation during
the period of this study. The data is represented as daily averages for
the period between sampling dates and is in terms of calories/cm2(=Langleys).
A rapid increase in solar energy is indicated between the winter months
and late spring and appears to be at a maximum in early June. This is fol
lowed by a gradual decline which continues throughout the remainder of the
year and the minimum occurs at about the winter solstice. The summer maximum
does not seem to occur at the summer solstice and measurements in late June,
July, and August seen to be lower than would be expected based on day length


* I EAVES / M
1975
1974
LEAF DENSITY
OltllVIO DAU
MIDICTIC


17
1969). I have not found any records of waterhyacinth occurring in Guatemala
but its range does extend into Mexico (Castellanos 1959; Little 1965)
where it is apparently well distributed.
As previously mentioned waterhyacinth was thought to have been
introduced into the United States from Venezuela in 1884 at the
International Cotton Exhibition in New Orleans by the Japanese delegation
(Klorer 1909; Buckman and Co. 1930; Gowanloch 1944; Hildebrand 1946;
Dymond 1948; Penfound and Earle 1948; Tabita and Woods 1962; Dutton 1964;
Wunderlich 1964; Bock 1966). Some accounts indicate, however, that the
plant may have been in the United States in the 1860's (Tabita and Woods
1962) or prior to the Civil War (Penfound and Earle 1948). The accepted
theory maintains that the plant was given away as souvenirs at the
New Orleans Cotton Exposition (Gowanloch 1944). The plants were taken
for ornamental purposes (Klorer 1909; Dutton 1964; Wunderlich 1964) or
for purposes of cultivating them for cattle fodder (Wunderlich 1964). In
any case, it was felt that when the plants outgrew the limited amount of
space given them they were cast out into natural bodies of water (Klorer
1909). By 1888 it was in the coastal fresh waters of Texas, Louisiana,
Mississippi, and Alabama (Buckman and Co. 1930).
It was apparently introduced into Florida in 1890 (Webber 1897;
McLean 1922; Barber and Hayne 1925; La Garde 1930; Buckman and Co. 1930;
Penfound and Earle 1948). Raynes (1964) reported it was first introduced
in the St. Johns River at Edgewater about 4 miles above Palatka. Mr. J.E.
Lucas was interviewed by a New York Sun reporter (Anonymous 1896) in 1896
and gave the following account:
"I know the man who brought the first plant to Florida," Mr. Lucas


175
i
Other significant variables included in the four models which re
present biomass are water level, pH, iron, and sulfates. The possible
significance of water level and pH have already been discussed. Sulfates
are correlated with magnesium, potassium, alkalinity, and conductivity.
The inclusion of sulfates in the leaves per plant model may indirectly
indicate the importance of soluble salts. The inclusion of iron with a
negative coefficient in the same model indicates the increased uptake
of this essential nutrient as plant biomass increases. This agrees with
my observations on plants grown in greenhouse cultures.
Plant density and leaf density are highly correlated and form the
second category. These two parameters, unlike the biomass indices, do
not show a peak in the summer but rather in the spring. Potassium, iron,
phosphorus, and water level were included in both models (Table 7). The
signs for these coefficients are opposite those for the same variables
in the previous models. The negative coefficient for potassium indicates
that it is being absorbed as density increases. This supports the argument
given above that potassium is absorbed in the greatest quantities at the
time of fastest growth. Further, potassium in plants is known to accumulate
in those tissues that are growing rapidly (Robbins et al. 1964) and would
be expected to be absorbed in greater quantities when maximum growth occurs.
Iron, on the other hand, appears to be directly related to density.
Although it is difficult to determine why iron concentrations showed a
dramatic increase in the spring (Fig. 4) the strong correlation with plant
density cannot be ignored. The iron enriched water appears to have been
al least partially responsible for the increase in density evidenced in
the spring. Since iron is an improtant element for the synthesis of


Table 16. Annual summary of larval counts and mortality.
Instar
Total
Collected
# Dead
%
Mortality
1
9
0
0
2
44
1
2.27
3
27
2
7.41
4
55
29
55.53
5
28
2
7.14
6
25
2
8.00
7
83
30
36.14
Pupae
6
0
0
Pup. Ex.
2


Total
279
66
23.66


251
Table 12. Comparison of the samples from the release site with the control site nased on
various estimates of the plant and insect populations. Figures represent means
and standard deviations (in parentheses), the T-statistic for unpaired data,
and the release site values represented as a percent of the control.
A B
RELEASE
SITE
CONTROL
SITE
HP*)
A/B
100
HEIGHT1' '
37.
91
(f
16.387
56
58
(
20.197
7
14
7- o.cfoiT
67
00
PLANT DENSITY/M?
109
70
(
53.17)
97
05
(
22.45)
0
38
( .0.50)
113
03
LEAF DENSITY/M2
347
05
(
193.90)
504
22
('
86.91)
1
28
(0.30)
68
83
LEAVES/PLANT
3
16
(t
1.46)
5
19
('
2.42)
7
29
(0.001)
60
89
INSECT DENSITY/M2
51
69
(
16.24)
3
16
(
3.16)
5
08
(0.01)
1635
/6
INSECTS/PLANT
0
50
(
0.09)
0
03
(
0.03)
8
39
(0.01)
1666
67
l INSECT MORTALITY
0
(
0 )
44
44
(
50.92)
1
51
(0.30)

X LEAVES DAMAGED
39
00
(
12.12)
5
41
('
4.01)
4
56
(0.02)
720
89
X RHIZOME DAMAGE
72
65
(
13.57)
16
57
(
12.72)
5
22
(0.01)
438
44
DEAD PLANTS/M2
53
80
(<
24.71)
0
('
0 )
3
77
(0.05)

% PLANTS DEAD
34
49
(-
14.22)
0
('
0 )
4
20
(-0.05)
--
1'N = 104 for the experimental site and 92 for the control site since these measurements were
taken on each plant. Otherwise N = 3 since the parameters were only estimated once for each
plot.
*P = the probability of a larger T occuring under the hypothesis A / B.


263
Table 13. Ratios of plant parts at
Ratio Compared
Leaves:plant
Petioles:plant
Rhizome:plant
Root:pi ant
Stolonrplant
Root + Rhizome :shoot
Photosynthetic:pi ant
Non-photosynthetic:pi ant
Leaf¡petiole
Living:dead
the two sites on
12 December 1974
Release Site
Control Site
0.09
0.09
0.40
0.50
0.15
0.19
0.33
0.19
0.02
0.03
1.00
0.64
0.49
0.59
0.51
0.41
0.23
0.17
0.24
2.10




189
Analyses
After determining that there was no significant difference (student's
T-test for unpaired data) in any of the parameters measured between the
two post-infestation sets of data they were combined to obtain 10 obser
vations per treatment level for each parameter after one generation of
insect activity. The final results are expressed as the mean values
of 10 parameterized final values for each level of infestation and are
expressed as a percentage of the initial value.
The results were interpreted by means of the regression procedure
from the Statistical Analysis Systems on the University of Florida IBM
370 computer. The results were fitted according to the three following
regression models where Y = plant response and X = insect concentration:
1. Y = A + BX
2. Y = A + BX + CX2
3. Y = A + BX + CX2 + DX3
4. Y = A-Bx
The model which best fitted the data and presented a realistic
estimate of response trends was selected for plotting. The best fit
in the first three models was determined by examination of the sequential
sums of squares. The fourth model was compared with the others by means
of the regression coefficients (r). The model with the largest r was se
lected only if it was a great deal larger than the alternative model.
Otherwise the model most clearly representing the results in the simplest
form possible was chosen. In the majority of the cases the simple linear
model (equation 1) provided a satisfactory representation the the results.
It was originally expected that the data would conform to the log conver
sion model (eq. 4) but this regression consistently provided a lower r-value


78
picked. This required a second randomization. One person involved in
the process held a second generated random number between 1 and 10. A
J_ L
second person drew up to ten plants out of the water. When the nLr plant
(where n = the random number) was pulled out the first person notified
the second and the plant was placed in a plastic bag and returned to
the laboratory.
Before measuring the leaf area the petioles and leaf blades (pseudo
lamina) were separated. The petioles were rolled out with a rolling pin.
This was necessary to compensate for the cylindrical shape of the petiole.
The outline was then traced on a piece of drawing paper and measured with
a planimeter. The leaf blades were pressed in a plant press and dried.
They were then traced and measured the same way. I found that drying
the leaves caused considerable shrinkage and a dry:fresh conversion factor
had to be employed. The formula for this conversion was:
leaf area (fresh) = 1.437 X leaf area (dry)
The conversion factor was determined by measuring one sample (51 leaves)
before and after drying. Each week figures for the average leaf (pseudo
lamina) area, average petiole area, and average total (pseudolamina +
petiole) leaf area were derived. These figures were multiplied by the
number of leaves per square meter from the plot samples to obtain the
leaf or petiole area index. This figure represents the leaf or petiole
surface area (considering only one side) per unit of substrate area (m2/m2).
Water samples were taken along the catwalk at the mid-point. The
water was collected a few centimeters below the surface at the level of
the waterhyacinth roots. This level should best reflect the conditions
the plants were being subjected to. The sampling station was on the
downstream si of the study area so the tests would reflect minimum


Figure 13. Average daily solar radiation values per month for
Gainesville, Florida. The figure above each bar represents
the number of years the means are based upon. Adapted from
the Climatic Atlas of the United States, U.S. Dept, of Commerce.


Total Larvae & Pupae / M
68Z


54
with a slight exterior brownish band. Length of the body
9 lines; of the wings 16 lines.
Grote and Robinson (1868) described a second species of Arzama,
A. obliquata. They compared this to Walker's type of A. densa in the
British Museun and found they differed in the larger size of A. obliquata
and different coloration. They apparently failed to compare it with
Edema obliqua, however,
Herrich-Shaeffer (1868: Cited from Zoo. Record) provided generic
and specific characters in full for A. densa from Cuba.
Grote (1873) described another species of Arzama, A. vulnifioa,
which differed from A. densa Wlk. and A. obliquata G. & R. primarily by
its dusky yellow color. He also noted that it was less robust than A.
obliquata with the anterior wings more rounded posteriorly. Grote (1874)
in his list of the Noctuidae of North America listed only these three
species but in 1878 [1879] described a fourth species, A. diffusa from
Maine. Gundlach (1881) redescribed Arzama densa Wlk. from specimens
collected in Cuba. A fifth species, A. melanopyga, was subsequently
described by Grote in 1881 (in Comstock 1881) from Florida. He pointed
out characters which separate A. diffusa, A. vulnifioa, 4. melanopyga,
and Sphida obliquata (apparently a recombination for A. obliquata G. & R.).
He noted that characters of the clypeus are of value in separating these
two genera. In 1882 Grote synonymized Edema obliqua Wlk. with Sphida
obliquata G. & R. These five species were listed together in the sub
family Arzaminae by Grote 1883 who noted that the species with the black
anal tuft (melanopyga) is probably a variety of vulnifioa. Riley (1885)
stated that the genus Sphida Grt. had no existence in nature and Sphida
obliquata G. & R. was synonymous with A. densa Wlk.


174
decrease in phosphorus concentrations occurred across the marsh in the
summer. Hence plant growth probably does affect the phosphorus concen
tration but it is difficult to determine if these concentrations become
limiting to the plants. Phosphate-phosphorus levels never fell to the
limiting level of 0.10 mg/1 as determined by the experiments of Haller
et al. (1970) discussed earlier. The drop in phosphates may have been
due to the absorption of luxury amounts by the plants beyond their
immediate requirements.
Potassium was included in the plant height model after the effects
of climate and phosphorus were removed. The simple correlation coeffi
cient between height and potassium was negative and not significant
(Table 10). The coefficient in the regression model (Table 7), however,
was positive and significant. Potassium was also significantly correlated
with water level (Table 8). The indication from the model is that at
constant levels of sunlight, temperature, and phosphorus, plant height
increases as potassium concentrations increase or vice versa. Again,
one would expect the opposite, that is, as the plants become large ab
sorption would increase and the nutrient concentration correspondingly
decrease. The possibility exists that maximum nutrient absorption occurs
early in the growing season. Nutrients may be absorbed in large quantities
and stored within the plant. The period of maximum absorption may occur
simultaneously with the period of fastest growth. As the growth rate
slows nutrient absorption may also decrease. Hence, even though the plants
are larger during the summer the plants may be having a lesser effect
on nutrient levels. This may explain the apparent decline in potassium
levels in March and April (Figure 7). This same explanation is also
possible for nitrogen.


Figure 12. A comparison of the standing crop and proportions of the
plant parts for the large and small waterhyacinth plants
used in the productivity studies.


273
clustering mode of oviposition is somehow stimulated by the plant.
First instars were rarely encountered in the field. This appear
to be due to their short stadium, small size, and inconspicuous feeding
damage. Vogel and Oliver (1969b, p. 251) indicated that "young larvae
were found feeding on tender basal stems and young foliage" but made no
specific reference to the first instar. Welch (1914) reported that Bellura
melanopyga Grt. went though two feeding periods. The first was a leaf
mining period in which the early instars feed in the leaf blade of
Nymphaea [-Nuphar) americana. The second period was a petiole period
in which the older larvae burrowed into the leaf petiole. This
phenomenon was also observed by Claassen (1921) with Bellura obliqua
(Wlk.) on Typha latifolia L. I have found in the laboratory that when
neonate A. densa larvae are provided only with waterhyacinth leaves they
will form leaf mines and feed between the upper and lower epidermis.
They will also feed externally aggregating in folds in the leaf blade.
In the field I have never found the first instar in leaf mines. I have
found them in shallow burrows in the petiole just under the epidermis
in the region of the leaf isthmus. I have more frequently found them
at the base of the plant usually between a leaf petiole and the leaf
sheath or a wrapper leaf. This is consistent with Vogel and Oliver's
(1969b) observations. In a few instances where I have found egg masses
on Pontederia there was evidence of larval mines within the leaf blade.
Usually it appears that the larvae migrate from the egg mass, however.
Small exit holes through the leaf are frequently present under an egg
mass indicating that the larvae burrow through the leaf upon enclosion.
Those individuals which do mine the leaves would be easily overlooked
and may partially account for the relatively few first instar larvae


247
Methods and Materials
Eggs of Avzama densa Walker were collected from Pontederia cordata
at Putnam Hall, Putnam Co., Florida. These were surface sterilized in
a 0.5% hypochlorite solution for 20 minutes, placed in a 10% sodium thio
sulfate solution for 5 minutes (to neutralize the hypochlorite), rinsed
in distilled water and attached to filter paper with a 10% casein glue
solution. The glue was permitted to dry and the filter paper with the
eggs was placed in the lid of a baby food jar. Waterhyacinth leaves
from plants grown in a quarantine greenhouse were similarly washed
in the hypochlorite solution and placed in autoclaved baby food jars.
The lids were placed on the jars and sealed until the larvae emerged.
These sterilization procedures were necessary to retard the growth of
mold long enough for the larvae to eclose. Fresh leaves were added as
needed for food.
The larvae obtained were kept in jars for 2-3 days. They were then
placed on tables filled with waterhyacinths in a greenhouse on 25 June
1974. On 25 July half of the plants were harvested and the larvae and
4
pupae obtained. The remainder of the plants were harvested on 2 August.
Larvae obtained were placed individually in 50 dram snap top plastic
pill vials and provided with fresh waterhyacinth petioles. After
pupation, pupae were placed on Vermiculite in a pie pan with a wax-
paper lined cage over them. Adults were permitted to emerge in the cage
and a 5% sucrose solution was provided for food. They mated and the
female oviposited on the waxpaper lining. Eggs were collected from the
wax paper and treated in the same manner as the field collected eggs.
From 14 females I obtained 2872 eggs. Approximately 78% or 2253
eclosed. Due to the failure to release these immediately high mortality


150-
LEAF AREA
CO
cr>


3l9
Comstock, J. A., and C. M. Dammers. 1944. The larva and pupa of Arzama
gargantua Dyar. Bull. S. Calif. Acad. Sci. 43(2): 84-85.
Comstock, J. H. 1881. An aquatic noctuid larva. Arzama melanopyga Grote,
new species. Papilio 1(9): 147-149.
Comstock, J. H. 1888. Aquatic lepidopterous larvae. Amer. Natur. 22: 468-469.
Comstock, J. H. 1936. An introduction to entomology. Comstock Publ. Co.,
Inc., Ithaca, New York, 1044 pp.
Cook, C. D. K., B. J. Gut, E. M. Rix, J. Schneller, and M. Seitz. 1974.
Water plants of the world. A manual for the identification of the
genera of freshwater macrophytes. Dr. W. Junk b.v., Pub., The
Hague. 561 pp.
Cordo, H. A., and C. J. DeLoach. 1975. Ovipositional specificity and
feeding habits of the waterhyacinth mite, Orthogalwma terebrantis,
in Argentina. Envir. Entomol. 4(4): 561-565.
Coste, P. 1958. Un Fleau aquatique des pays tropicaux: la jacinthe d'eau.
Phytoma. 99: 24-28.
Couch, R. 1971. Morphology and anatomy of water hyacinths. Proc. 24th
S. Weed Sci. Soc.: 332.
Coulson, J. R. 1971. Prognosis for control of water hyacinth by arthro
pods. Hyacinth Control J. 9(1): 31-34.
Crocker, W. 1907. Germination of seeds of water plants. Bot. Gaz. 44:
375-380.
Crumb, S. E. 1956. The larvae of the Phalaenidae. U.S.D.A. Tech. Bull.
1135, 356 pp.
Curtiss, A. H. 1900. The water hyacinth in Florida. Plant World 3: 38-40.
Davies, H. R. J. 1959. Effect of the water hyacinth (Eichhomia arassipes)
in the Nile Valley. Nature 184(4692): 1085-1086.
Del Fosse, E. S., H. L. Cromroy, and D. H. Habeck. 1975. Determination
of the feeding mechanism of the waterhyacinth mite. Hyacinth Control
J. 13: 53-55.
DeLoach, C. J. 1975. Identification and biological notes on the species
of Neoahetina that attack Pontederiaceae in Argentina. Coleopt. Bull.
29(4): 257-265.
DeLoach, C. J., and H. A. Cordo. 1976. Life cycle and biology of Neoahetina
bruohi, a weevil attacking waterhyacinth in Argentina, with notes on
N. eiohhomiae. Ann. Entomol. Soc. Amer. (in press).
Dunigan, E. P., and R. A. Phelan. 1975. Nitrogen and phosphorus uptake by
water hyacinths in two farm ponds. Rep. Proj. Dep. Agron., LSU A&M
Coll., Agr. Exp. Sta.: 272-275.


Figure 51. The total number of 4th and 7th instar Arsama densa
larvae per square meter as estimated from samples taken
from the marsh site of Lake Alice. The solid lines
represent the total number of larvae encountered. The
dotted lines represent the number of that total which
showed no signs of parsitism or diseases. The vertical
bars depict the percentage of the total affected by the
various mortality factors.


Figure 3
Total carbonate and bicarbonate alkalinity and conductivity of
Juneri975P The^i611 frm Uke A1ice from July 1974 through
June 1975. The lines represent 5-point moving averages.




280
and adults (estimated from pupal exuvia) was measured at the time of
harvesting (35 and 44 da post egg collection).The larvae collected were
fed petiole sections until they pupated and the pupae were held until
they emerged. A cumulative tally was kept on the number of adults
emerging and the number of eggs deposited. These figures were later
converted to daily cumulative percentages based on the final totals.
The probit was derived from a conversion table for the cumulative
percentage of each of these and plotted against the log of the number
of days following egg collection. By fitting a line to these points
the day upon which 50% of the population transformed into pupae or
adults or the time at which 50% of the eggs were deposited was esti
mated. These probit analyses are illustrated in Figure 46. This
technique should be valid since the probability of occurrence of these
events is a sigmoid curve as a function of time. The vertical vectors
in Figure 46 indicate the points that the probit line crosses the 5.0
probit. This represents the 50% probability for occurrence of that
event and in a normally distributed population estimates the mean (see
Andrewartha, 1961, p. 65). The estimated dates of pupation, adult
emergence, and ovposition are 42, 50, anid 52 days respectively. It
may be noted that this predicts the emergence of the adults 2 days prior
to oviposition. This suggests a preoviposition period which is in
consistent with my previous findings. When the emergence of males and
females is plotted separately, however, the predicted average emergence
date for the males is 48 da and for the females 52 da. This is consistent
with the lack of a prolonged preoviposition period noted earlier.
Considering that these time periods are established from the
date the eggs are collected and not from the date of oviposition the


237
The quantity of carbohydrate available for storage is related to the
amount of photosynthetic material present to produce it and the efficiency
of the plant. The effect of reducing the amount of leaf tissue as a result
of insect feeding would be that of reducing the energy available for stor
age or maintainence. Except for the controls, leaf tissue changes (Fig. 38)
closely paralleled the changes in rhizome tissue at 123%, 74%, and 57% for
the 0.33, 0.67, and 1.00 infestation levels. The control final leaf weight
was 183% of the initial weight. Hence, insect feeding directly reduced
the photosynthetic tissue present and directly or indirectly reduced the
material stored as rhizome tissue.
The proportion of the living plant weight represented as green mass
increased in the controls from 0.735 to 0.776 or approximately 106%.
This figure was somewhat greater in the plots treated with insects (see
Table 11) but does not appear to be linearly related to insect concentra
tion. This increase in proportion was probably due to a decrease in the
non-photosynthetic tissue.
The non-photosynthetic tissue (Fig. 39) not only increased less in
the control treatment (ca. 146%) than photosynthetic tissue (ca. 183%)
but also decreased more in the insect treatments. This resulted in a
decrease in the root-rhizome:shoot ratio (Table 11) at all treatment
levels. This ratio decreased less in the control plots (82% of initial)
than in the treated plots (53-59%). Again, this does not appear to be
a linear relationship. This does indicate, however, that the effect of
insect activity decreased the stored material (rhizome) as well as the
ability of the plant to absorb nutrients relative to the photosynthetic
ability of the plant.


311
5. Insects, therefore, by reducing the canopy and stimulating
offset production may indirectly stimulate production.
6. This increased production may partly compensate for crop
reduction by herbivory.
7. Insects reduce standing crop by accelerating turnover.
8. Large amounts of nutrients are tied up as organic matter in
the plants.
9. Insect feeding may, therefore, result in a faster return of
these nutrients to the water which may also stimulate production.
10. Higher levels of insect infestation are likely to be needed in
the summer (probably the spring also) when solar radiation is
high or waxing than in the fall and winter (when solar energy is
low or waning) to achieve the same level of control.
11. A reduction in the size of the rhizome in proportion to the plant
by insect feeding is likely to hinder the ability of the plant to
survive the winter since spring regeneration occurs from the rhizome.
During the period of these studies, natural populations of A. densa
were consistently low. Heavy infestation by a complex of parasites appeared
to be the factor regulating population build-ups. This was very difficult
to analyze, however, because of the extreme degree of overlap in generations,
differential susceptibility of various instars to sampling, and low popu
lation levels from which to derive data.
A small scale field release of A. densa proved very effective in
controlling a small stand of waterhyacinth. Parasites failed to reduce the
larval population and severe damage to the plants resulted. The degree of
mortality as a result of these parasites was notably less than that of a


Figure 35. Detritus as a percentage of total waterhyacinth biomass as
a function of insect feeding activity. Curve fitted by eye.


NUMBER OF PLANTS PER SQUARE METER
144


32
The entomofauna of waterhyacinth is quite large and diverse. Most
of the information available regards those species which feed upon water-
hyacinth and, thus, show potential as biological control agents. Sankran
et al. (1966) investigated a grasshopper {Gesonula punctifrons Stal.
:Acrididae) attacking waterhyacinth in India. Fred Bennett of CIBC in
Trinidad has published many papers on the possibility of biological con
trol of waterhyacinth and on the insects associated with it (Bennett 1967,
1968a, 1968b, 1970, 1972; Bennett and Zwolfer 1968). Other lists have
been provided by Gordon and Coulson (1969), Coulson (1971), Perkins (1972,
1974) and Spencer (1973, 1974).
Sabrosky (1974) described a dipteran stem miner (Eugaurax setigena:
Chloropidae) from South America. Barman (1974) investigated the growth
and assimilation efficiences of an arctiid (Diaovisia virginioa which is
known to feed on waterhyacinth. Silveira-Guido and Perkins (1975) reported
on the biology and host specificity of Comops aquatioum (Bruner), a
grasshopper (Acrididae) from Argentina which attacks waterhyacinth.
DeLoach (1975) provided indentification and biological notes on the genus
Neoohetina (Coleptera: Curculionidae) that attack the Pontederiaceae in
South America. DeLoach and Cordo (1976) provided information on the life
cycle and biology of N. eichhomiae and N. bruohi3 two species which have
been released for the biological control of waterhyacinth in the United
States. Warner (1970) described these two species.
Wallwork (1965) described a leaf-boring galumnoid mite (Orthogalwma
terebrantis) from Uruguay which feeds on waterhyacinth which has subse
quently been found in the United States (Bennett 1968a). Perkins (1973)
studied the biology and host specificity of this species in Argentina.


Figure Pa9e
24. Standing crop values, both estimated and real, from Lake
Alice 154
25. Percentage of the leaves and rhizomes of the waterhyacinth
population damaged through feeding activity of Arzama densa
at Lake Alice 159
26. Plant density as a function of plant height 179
27. Average dry weight per waterhyacinth plant as a log function
. of the average height 193
28. The average height per waterhyacinth plant (as measured from
the longest leaf) as a function of the feeding activity of
Arzama densa larvae 196
29. The effects of varying levels of insect feeding activity on
the average number of leaves per waterhyacinth plant expressed
as a percentage of predetermined means 199
30. The effects of varying levels of insect feeding activity on
the total number of waterhyacinth leaves per unit area
expressed as a percentage of predetermined means 201
31. The effects of varying levels of insect feeding activity on
the number of waterhyacinth plants per unit area expressed
as a percentage of predetermined means 205
32. The effects of varying insect concentrations on the total
waterhyacinth biomass (expressed as both detritus and
living plant material). 210
33. The effects of varying insect concentrations on the living
waterhyacinth mass present per unit area 212
XI


268
open stand than were those in the greenhouse experiments.
The small difference in total biomass present tends to substantiate
the feeling that net productivity was not reduced. A number of
explanations are possible for this. Productivity at this time of year may
be low, hence what was measured was merely the amount initially present.
This is possible since I have noted in other studies that the standing
crop begins to decline in late summer. A second explanation is that the
insect infestation decreases intraspecific competition thereby increasing
the productivity of the remaining plants. This increased productivity
may make up part of the difference caused by the insects.
Jameson (1963) pointed out that carbohydrate storage is directly
correlated with winter hardiness. If the primary organ of carbohydrate
storage in waterhyacinths is the rhizome (Penfound and Earl, 1948) then
the feeding activity of Avzama severely reduced the carbohydrate reserves .
This is illustrated in Figure 45 where the rhizome weight per plant at the
release site is only 15% of that at the control site. Harris (1973)
stated, however, that an injury that lowers carbohydrate levels either
directly or indirectly by stimulating auxin production and growth is
likely to be partly compensated for by an increase in photosynthetic
efficiency.
Penfound and Earle (1948) found that the rhizome length remains
fairly constant throughout the growing season. They attributed this to
an equilibrium between rates of decay at the older portion and rates of
increase at the crown. If this is true we may assume that carbohydrate
reserves also remain fairly constant and at the end of the growing season
are sufficient to maintain the plant through the winter and provide


CO
00


88
Analyses
Multivariate analyses were performed on the annual data in an attempt
to account for observed variation in the plant characteristics in terms
of the various environmental parameters. A stepwise regression procedure
(SAS STEPWISE) was first employed to determine which linear combination
of independent variables would provide the best fit for the actual data.
The dependent variables analyzed were standing crop, plant height, plant
density, leaf density, average leaves per plant, and leaf area index.
Each was regressed against solar radiation, minimum air temperature,
maximum air temperature, minimum water temperature, maximum water temp
erature, % rhizome damage by Arzama densa, % leaf damage by A. densa, all
nine water quality parameters, and water level. Three procedures were em
ployed to determine the best regression equation. These were the forward
selection procedure, the backward elimination procedure, and the stepwise
procedure. All form linear models in different ways (see SAS manual, 1972
for further explanation). Each provides an analysis of variance, regression
coefficients and statistics of fit for the model. In all cases the step
wise procedure provided the most significant fits to the data.
Two additional variables were entered which were derived from other
variables. Since it was assumed that the amount of light intercepted by
the plants would be inversely proportional to the degree of self shading
the solar radiation variable was divided by the leaf area index to form
a new variable. This was entered into the analyses in place of incident
solar radiation but it failed to increase the significance of any of the
linear combinations (i.e., incident solar radiation was just as good or
better). It was also speculated that available space may contribute to


278
Duration of Developmental Stages
First, second, and third instar larvae were obtained from field
collections and from laboratory reared material. Larvae were placed
individually in 1 oz. diet cups and provided with petiole sections from
either Eichhomia crassipes or Pontederia cordata. The cups were kept
in an enviromental chamber at 25 C and 16:8 L:D photoperiod. The
plant material was checked daily and replaced as needed. The larval
instar was also checked daily and recorded for each cup. Head capsules
were saved and later measured. Table 15 summarizes the developmental data
from this study. Only three larvae pupated and this occurred following
the seventh instar. I feel that this is the typical number of instars.
Other larvae went into eighth and ninth instars before death occurred.
The presence of extra instars is typical in laboratory reared Lepidoptera
when under stress (Leppla, pers. comm.). Several factors may have been
responsible in this study. The cups used to contain the larvae were
samll. They were translucent and not transparent. The larvae may need
rhizome material at some stage in their life and they were only fed
petioles. Humidity was high in the cups as condensation was frequently
noted. Hence, light, humidity, space, food quantity and food quality
could have become stress factors ultimately resulting in these extra
molts. Developmental data was included for these individuals only
through the seventh instar.
Total developmental time was estimated indirectly from the
greenhouse experiment testing the effects of A. densa damage on water-
hyacinth (Section 2). The approximate age of the larvae was known at the
time of the release. The proportion represented as larvae, pupae, and
282


'i-/.
35
specific gravity of the stolon is 0.818 and it consists of about 97%
water by weight (Penfound and Earle 1948).
The most interesting morphological development of the waterhyacinth
is its leaves. Arber (1918, 1920) found that the vascular bundles in the
petiole are arranged with the xylem oriented towards the periphery. Those
in the lamina may be arranged with the xylem up, down, or oblique. This
is in contrast to plants with a true lamina which have the vascular
bundles arranged with the xylem towards the upper leaf surface. She
suggests that this indicates that the lamina is merely an extension of
the apical end of the phyllode and not homologous with the laminae of a
Dicotyledon. As such it should properly be called a pseudolamina and
the basal portion a petiole. The two are connected by a narrow compact
region called the isthmus and the narrow base below the float is referred
to as the subfloat. A membranous ligule (= stipule of Agostini 1974)
is present at the base of the subfloat (Penfound and Earle 1948) which
possesses a small reniform lamina (Agostini 1974).
The petiole may be more or less inflated to form a bulb-like
structure commonly assumed to function in floating the plant (Parsons
1963; McLean 1922; Chhibbar and Singh 1971; Olive 1894; Couch 1971).
This has been contradicted by Rao (1920b) because the bladders are
formed mostly above the water and the leaves float with or without
them. Bock (1966) noted, however, that the bases of the inflated petioles
just beneath the water formed a stable platform. Further, floating
single plants with elongate petioles were unable to remain upright
and if they remained on their side sent out new leaves with inflated
petioles.


181
favors those plants that can produce the most offsets earliest. Late in
the season selection favors those plants that can compete for light most
effectively. This results in the pattern of density observed in Figure
19. This explanation applies only to waterhyacinth populations which
are limited by space and is the result of genetic flexibility within
the population rather than different genetic strains. If open water
continues to be available in the summer the reproductive phase may con
tinue until all available space is utilized as long as some other fact
or does not become limiting. This phase of reproduction is likely to
occur only at the fringe of the mat nearest the available space. The
plants further back within the mat are more likely to be limited by
competition for light and will generally be increasing in height. Con
sidering that waterhyacinths are colonizer species this pattern of growth
is probably very adaptive. Increasing density in open areas enables
the plant to colonize rapidly. It further increases the number of propagules
available for reaching new downstream areas to produce daughter colonies.
Once the plants become established in an area their ability to compete
must increase in order to maintain the colony. In its native habitat the
ability to increase in height enables it to compete both inter- and intra-
specifically since several similar species occur with it (e,g. Eiohhomia
azurea, Reussia sp.). In the United States, however, there are relatively
few large floating aquatic macrophytes and most interspecific competition
is with plants in the littoral zone. In areas of deep water where emer
gent vegetation does not exist competition for light is intraspecific.
Since relatively few herbivores act to reduce this intraspecific compe
tition the population is limited only by climate and nutrients. Hence,


19
dead at the top. They grew again, and the startling conditions that you
see in these pictures are a growth of only two years."
This account surprisingly indicates that the plants were introduced
into Florida from Europe rather than from Louisiana as has generally been
assumed (Buckman and Co. 1930; Tabita and Woods 1962; Wunderlich 1964).
Waterhyacinth was first discovered in Georgia in 1902 (Harper 1903)
about one mile north of Valdosta. The first record in California is
from 1904 near Clarksburg, Yolo Co. (Bock 1968). Johnson (1920) reported
it in Fresno Co., California. Bock (1968) lists its present range in
California from 10 mi NW Sacramento (ca. 38.5 N Lat.)to Ramona, San
Diego Co. (ca. 33 N Lat.). She speculates that it was probably brought
to California as an ornamental and released. The primary rivers infested
are in Central California and include the Kings, Tuolumne, San Joaquim,
and Sacramento River Systems.
The infestation of waterhyacinth in California is discontinuous
with the North American range of this weed. Penfound and Earle (1948)
stated that shortly after the turn of the century it had been reported
from all the southeastern coastal states as far north as Virginia. A
distribution map published by the U.S.D.A. (1970) indicates that the
present range of this plant in the U.S. includes the Potomac River in
Maryland-Virginia, west to southern Missouri, south to eastern Texas
and southern Florida, and separately, central California.
Just when waterhyacinth spread to the Old World is not certain.
Agarwal (1974) indicated it may have been introduced into India around
1896. McLean (1922) cited testimony indicating that it may have been
present in Bengal as early as 1898 or 1899. It was apparently introduced


in less need for supportive tissue to display the leaves. The pro
portion of the plant represented by roots increased and the root-
rhizome:shoot ratio increased. This is typical of small plants
growing in open stands. The photosynthetic tissue ratio decreased
relative to the non-photosynthetic tissue. This is in contrast to
similar experiments in a greenhouse (see previous section) where
Avzcoma feeding appeared to result in a decrease in the root rhizome:
shoot ratio. Since the surviving plants had a well developed root
system at the release site they would have probably recovered if the
insects had been removed and if the season was favorable.


curves comparing small and large plants revealed that the net efficiency
of both plants was 1.6% (of incipient solar energy). The small plants
grow faster than the large ones, however, by virtue of a larger P:R
ration (2.05 vs. 1.46).
Photosynthetic efficiency is maintained by synchronization of the
leaf area index with the annual solar energy flux. An annual increase
in leaf area occurs first through an increase in leaf density and
secondly through an increase in leaf (pseudolamina) size. The net result
of these two growth phases is a peak in the leaf area index spanning
the period of maximum solar radiation.
Intraspecific competition is strongly implicated in governing plant
density and seems to account for observed changes in the population.
Plant density is high in the winter reaching a maximum in April. This
is followed by a decline in May and June as a result of the loss of
plants in the smaller size classes. This loss is due to shading by the
larger plants as they increase in size and leaf area.
Multivariate analyses indicate that solar radiation and minimum
air temperatures were important in accounting for changes in standing
crop, plant height, leaf area index, and the numbers of leaves per plant
(all indices of biomass). The introduction of water quality parameters
into the analyses resulted in confusion as causal relationships were
difficult to establish.
Damage by Arzama densa Wlk. (Lepidoptera: Noctuidae) did not appear
to affect the population of waterhyacinth studied. Greenhouse studies
revealed that concentrations of 33 larvae per 100 plants could signifi
cantly reduce almost all characteristics examined and greatly accelerate


218


292
A summer decline in the population is evident from the data. From
early May through late July A. densa is rare on Lake Alice. I have
frequently required eggs during the summer for various reasons. Since
they were unavailable on waterhyacinth, stands of Pontederia sp. at
Putnam Hall, Putnam Co., Fla. and at Paynes Prairie, Alachua Co., Fla.
were checked. Eggs and larvae were found to be present throughout the
summer. No population studies were conducted on this host plant but the
need for these studies is evident. Pontederia seems to be the primary
host for A. densa and an understanding of the population cycles on this
host would be an invaluable in interpreting seasonal population differences
on waterhyacinth.
Mortality
Because of the lack of data with regard to egg counts I have very
little information on egg mortality from populations on waterhyacinth.
From 9 egg masses collected between September and December 1974 the range
of egg parasitism by the scelinoid Telenomus arzamae Riley was between
0 and 68% (average = 26%). Because mortality was not always complete
these figures are affected by the age of the eggs. Further, many egg
masses were collected after the larvae and the adult parasites had
emerged and it was difficult to determine whether the chorion of a
particular egg had been vacated by a larva or a parasite.
To further estimate egg mortality 80 eggs were collected from a
caged female. The eggs were laid on gauze and were deposited in single
layers rather than in the typical convex masses. The gauze was cut up
so as to partition the eggs into 4 groups of 20 eggs each. These were
placed on waterhyacinth leaves on a small pond near Paynes Prairie on
4 September 1973. They were left on the plants for 4 days, recollected


Figure 48. The total number of Arzcona densa larvae collected, either living or dead, from the
marsh side of Lake Alice. Each point represents a mean per square meter derived from
three 0.316 mz samples collected at weekly intervals from May 1974 through April 1975.


322
Guscio, F. J., T. R. Bartley, and A. N. Beck. 1965. Water resources
problems generated by obnoxious plants. J. Waterways Harbor
Div. Amer. Soc. Civ. Eng. 10: 47-60.
Habeck, D. H. 1974. Arzama densa as a pest of dasheen. Fla. Entomol.
57(4): 409-410.
Haigh, J. C. 1936. Notes on the water hyacinth (Eiohhomia crassipes
Solms) in Ceylon. Ceylon J. Sci. Sec. 12(2A): 97-108.
Haigh, J. C. 1940. The propagation of water hyacinth (Eiohhomia crassipes
Solms) by seed. Trop. Agr., Ceylon 94(5): 296-297.
Haller, W. T., E. B. Knipling, and S. H. West. 1970. Phosphorous absorption
by and distribution in water hyacinths. Proc. Fla. Soil and Crop Sci.
Soc. 30: 64-68.
Haller, W. T., and D. L. Sutton. 1973. Effect of pH and high phosphorus
concentrations on growth of waterhyacinth. Hyacinth Control J.
11: 59-61.
Hampson, G. F. 1910. Catalogue of the Lepidoptera Phalaenae in the
British Museum. Vol. 9. Noctuidae, Acronyctinae, Part 3. Brit. Mus.
Natur. Hist., London, 552 pp.
Hansen, K. L., E. G. Ruby, and R. L. Thompson. 1971. Trophic relation
ships in the water hyacinth community. Quart. J. Fla Acad. Sci.
34(2): 107-113.
Harper, R. M. 1903. The water-hyacinth in Georgia. Plant World 6:
164-165.
Harris, P. 1972. Insects in the population dynamics of plants. Pages
201-209 in H. F. van Emden, ed. Insect/plant relationships.
Symp. Roy. Entomol. Soc. London, No. 6. Blackwell Scientific Publ.
Oxford.
Harris, P. 1973. Weed vulnerability to damage by biological control
agents. C.I.B.C. Mise. Pub. 6: 29-39.
Hearne, J. S. 1966. The Panama Canal's aquatic weed problem. Hyacinth
Control J. 5: 1-5.
Hentges, J. F., Jr. 1970. Processed aquatic weeds for animal nutrition.
Proc. Aquatic Plant Res. Conf., University of Florida, Gainesville:
62-67.
Herrich-Schaffer, G. A. W. 1868. Korrespondenzblatt, fr Sammler von
Insecten, insbesondere von Schmetterlingen. Regensburg, Manz: 156.
Hildebrand, E. M. 1946. Herbicidal action of 2,4-D acid on the water
hyacinths, Eiohhomia crassipes. Sci. 103: 477-479.


Figure 43(c)
Figure 43(d)


59
Grote & Robinson (=Edema obliqua Walker), I consider
to be junior synonyms. Belluva has page priority over
Avzama (Walker, 1864, List ..., pt. 32, p. 465 vs p. 645.).
In addition, so far as I can find, J. B. Smith, 1893, Bull.
U. S. Nat. Mus., No. 44, p. 181, was the first to treat
both names and he placed Avzama in the synonymy of Bellura.
Forbes, 1954, Cornell Exper. Stat. Mem. 329, p. 217-8, used
Avzama in error, but divided the genus into two sections
(=Subgenera?). However, the character he uses to divide
the two sections are invalid. Females of Belluva do not
have simple antennae as he indicates, and the front may be
developed in some forms. The extent of development of the
frons is a character that needs more study. It will also
be necessary to study the possibility that food plant
. varieties are involved. I have indicated to others that
I believe there are only two or three species in the genus,
govtynoides, obliqua3 and possibly densa. Smith believed
that govtynoides and densa represented one species, and he
sank that latter as a synonym.
I think that it is apparent from the literature that the taxonomy
of this group is of an uncertain status. I agree with Todd that these
species probably represent one genus and the proper name of Avzama densa
Wlk. is Belluva densa (Wlk.). Because of the widespread current use of the
former name and the absence of a definitive study in literature I have
used the binomial Avzama densa Wlk. throughout this dissertation.


176
chlorophyll it may be an important limiting factor in the waterhyacinth
community.
Phosphorus was also directly related to leaf density and plant density.
Maximum densities, therefore, occur when phosphorus concentrations are
high. This is in contrast to plant height which achieves its maximum when
phosphorus quantities are low. This may be interpreted as indicating
that high phosphate concentrations early in the growing season are important
in initiating the spurt of rapid growth in the spring. As the plants
become larger and biomass increases more phosphorus is absorbed and the
concentration decreases.
Magnesium in the leaf density model had a negative coefficient.
This may have reflected the uptake of magnesium as leaf tissue increased.
This would be expected since magnesium is an important constituent of
the chlorophyll molecule. This change is not obvious in Figure 4, however.
Climatological variables did not appear to have the importance in
the density models that they had in the biomass models. Solar radiation
was included in the leaf density model but the coefficient was not signi
ficant. Maximum water temperature was considered the most important
variable in the plant density model and was negatively related to it.
This probably reflects inter-relationships of different characteristics
within the plant population, however. That is, competition for light is
reduced in the winter thereby allowing an increase in density. Since
temperature is so important in the biomass model and biomass is inversely
related to density.
The multivariate analysis as presented does not take into consider
ation auto-regulatory features within the waterhyacinth population. The
productivity studies show that small plants grow more rapidly than large


75
atm ppm C02). When multiplied by the flow rate this expression yields the
rate of carbon metabolism (g C/hr) within the chamber. A more detailed
explanation of this system is given by Carter et al. (1973).
Carbon metabolism for each type of plant was measured for 24 hrs.
Integration of the resultant production curves yielded both gross primary
productivity and respiration. Respiration was assumed to be constant both
day and night and was determined as the average nighttime value. Net
production consisted of that portion of the curve above the compensation
point (where Pg = R and = 0). solar radiation was measured with a
Weathermeasure Co. 24-hr. pyroheliograph in the 0.36-2.5um range. Air
temperature was recorded using a Yellow Springs Instrument thermistor
apparatus.
Following the metabolism measurements the plants were harvested to
obtain a biomass estimate. The total sample was divided into leaves,
petioles, roots (= roots + rhizomes + stolons) and detritus and the
various plant parts were weighed while fresh. A similarly divided sub
sample was taken and weighed before and after drying. From this subsample
a wet to dry conversion factor was obtained so that the dry weight for
each plant part and the total sample could be obtained. A subsample of
the leaves (pseudolaminae) and petioles was pressed in a plant press and
dried. The outlines of the dried leaves were traced on paper and the
area measured with a planimeter. This determined a leaf area per gram
of dried leaf conversion factor and the leaf area of the total sample
was estimated from this. A similar procedure was employed with the
petioles. From this the leaf area index (LAI) was determined which, in
this case, is the total leaf area (leaves + petioles) per square meter
as determined from only one side of the leaf.


236
Table 11. Ratios of the various plant parts and the percent
change in the final values as compared to the initial
values. The simple correlation coefficient (r) is
derived from a linear regression analysis.
Insect Concentration
Ratio
0
0.33
0.67
1.00
r
Leaf Wgt.
Initial
0.731
0.704
0.735
0.694
-0.172+
0.214+
Plant Wgt.
Final
0.775
0.788
0.827
0.793
%
106.02
111.93
112.52
114.27
0.399*
Rhizome Wgt.
Initial
0.087
0.082
0.092
0.079
-0.192+
Plant Wgt.
Final
0.116
0.089
0.078
0.079
-0.559**
%
133.33
108.54
84.78
100.00
-0.475**
Root Wgt,
Plant Wgt.
Initial
Final
0.149
0.090
0.185
0.082
0.135
0.056
0.172
0.091
0.040+
0.023+
%
60.40
44.32
41.48
52.91
-0.175+
Stolon Wgt.
Initial
0.028
0.027
0.035
0.051
0.389+
FlanT Wgt.
Final
0.020
0.040
0.023
0.039
0.142+
%
71.43
148.15
65.71
76.47
-0.072+
Root-Rhiz.
Initial
0.325
0.386
0.311
0.373
0.084+
Wgt.
Final
0.267
0.224
0.164
0.218
-0.380*
Leaf Wgt.
%
82.16
57.90
52.60
58.55
-0.550**
+ Probability (p) of a greater |r| >0.05; *p<0.05; **p<0.01.


274
observed in the field.
Second and third instars are found in a variety of places. They
are most often located at the base of the plant frequently between
two tightly appressed petioles or under a wrapper leaf usually feeding on
new leaf growth. Occasionally they will form burrows within a petiole
or shallow grooves on the outside of a petiole. By the fourth instar they
become almost exclusively internal feeders, boring the petioles and feed
ing on the apical tip of the rhizome. By the sixth instar they create
large burrows doing considerable damage to the petioles and may bore
three or four centimeters into the crown. The most extensive damage is
created by the seventh instar. The tunnels may extend into three or four
adjacent petioles, the full length of the rhizome, even through the stolon
into an adjacent plant. The damage to the rhizome may be so extensive
as to cause severe rotting and fragmentation.
Pupation occurs within a petiole usually in the basal portion. A
pupal chamber is hollowed out and a window is opened 2-3 cm. above the
pupa to permit egress of the adult. No cocoon is formed although a
silken suspensory apparatus may be constructed below the abdomen to
cradle the pupa within the burrow. The pupa is oriented parallel to the
long axis of the petiole with the head toward the distal end. Occasionally
pupation occurs in the rhizome.
The adults seem to rest during the day withing the foliage of the
waterhyacinth mat or in the vegetation along the shoreline. They are
quite active at night and are frequently collected at light traps (Frost
1975). I have found in the laboratory that females may mate and oviposit
within a few hours after emergence when caged with males in the dark.
This indicates a very brief pre-ovipositional period. To confirm this I


[ 221


50
The growth and development of the seedlings have been described
by Parija (1930), Robertson and Thein (1932), Haigh (1936), and Penfound
and Earle (1948). Several authors have indicated that a water saturated
medium is necessary for seedling survival (Hitchcock et al. 1949; Parija
1930; Haigh 1936) but forced immersion in water retards growth or kills
the seedling (Parija 1930; Hitchcock et al. 1949). Penfound and Earle
(1948) and Hitchcock et al. (1940) noted that seedlings would grow on
waterhyacinth flotant and Pettet (1964) found them growing on the shore
in "strand-lines" created by dead waterhyacinths. Hitchcock et al. (1950)
noted that in nature factors which prevent young seedlings from surviving
may be more important than factors which permit seed germination.
Vegetative Reproduction
Even though seed production by waterhyacinth may be massive, the
primary mode of reproduction is through vegetative propagation (Hitchcock
et al. 1950). This occurs through the production of offsets, or suckers,
produced on stolons (Penfound and Earle 1948). Hitchcock et al. (1950)
found that offset production begins about 60 days after the plant germinates
when the rosette attains a diameter of 7.6 to 10.2 cm. Penfound and Earle
(1948) found that a mat extends its boundaries at a rate of 3 feet per
month through vegetative reproduction under favorable conditions and the
plants double their numbers every two weeks. Bock (1966, 1969) and
Perkins (1972) have reviewed the literature dealing with the rates of
offset production in different locations and situations.
Productivity and Standing Crop
Many authors have dealt with waterhyacinth productivity in one


170
Overall, with the exceptions of standing crop and possibly leaf
area index, the models produced reflect observed trends in the plant
characteristics fairly accurately. More sophisticated modeling techniques
could probably improve these fits if non-linear responses could be con
sidered .
Discussion
The object of this study was to determine the degree of effect of
damage by Avzama densa on various characteristics of the waterhyacinth
population. Neither leaf damage nor rhizome damage proved to be a
significant factor in any of the models derived from the multivariate
regression analyses (Table 7). The obvious conclusion based only on these
analyses is that a. densa damage did not affect the plants or, more
precisely, did not account for a significant amount of the variation
observed in the plant characteristics measured. Upon examination of the
correlation matrix for the independent variables (Table 8), however, it
is found that both leaf and rhizome damage are highly correlated with the
climatological variables and with water level. Further, these correlation
coefficients are all negative indicating that insect damage is high when
sunlight, temperature, and water level are low. Because of thes rela
tionships it is not reasonable to exclude insect damage as an important
factor since the correlated variables are important. If all of the variables
were independent a term for insect damage may have been included in the
models. In lieu of this independence the stepwise analysis first selects
the parameter which best reduces the variability. Further parameters are
used to account for the variability remaining after variability due to
the first parameter is removed. When the first and second parameters are
highly correlated the exclusion of variability due to the first may also


Figure 44. A comparison of the standing crop of waterhyacinths at the
control site and the release site. Total biomass includes
both living and dead plant material. The photosynthetic
mass (green mass) includes pseudolaminae (leaves) and
petioles. The non-photosynthetic mass includes roots,
rhizomes, and stolons collectively.


Figure 18. Leaf area index of the waterhyacinth population on Lake Alice.
Each point represents the product of the leaf density per
square meter and the average area per leaf as determined from
the smoothed curves (Figures 16 and 17). The dotted line
represents predicted values based on multivariate regression
equations (see Table 7).


30
similar at both sites. They further found that plants grown in varying
phosphate solutions ranging from 0.075 ppm to 0.60 ppm did not signifi
cantly differ with respect to percentage weight gain over a 17 da period.
Haller, et al. (1970) found that the critical phosphrus concentration
for waterhyacinth growth was 0.01 ppm. Above this level phosphorus was
absorbed in luxury amounts but a higher proportion of that available
was absorbed at low concentrations. Haller and Sutton (1973) found
that optimal growth occurred at 20 ppm phosphorus but levels higher than
40 ppm were toxic. They further found that the root weight was greatest
at 0 ppm reflecting a tendency towards maximizing root absorptive sur
face in response to low nutrients.
Sutton and Blackburn (1971a, b) investigated the effects of vary
ing copper solutions on growth and transpiration of waterhyacinth. They
found that transpiration was reduced at 4.0 ppm with copper when grown
in the solution for 1 week and at 2.0 ppm when grown for 2 weeks. Growth
was inhibited by 3.5 ppm when subjected to the solution for 2 weeks.
After one week the shoot dry weight was reduced at 8.0 ppm or above and
the root dry weight by 16.0 ppm. The copper content of the shoot reflected
the content of the water when the concentration was above 2.0 ppm but at
levels below this the concentrations in the roots were independent of
those in the water. The copper content of the roots increased linearly
with the solution concentration.
Boyd and Scarsbrook (1975) found that the addition of 20:20:5
N:P205:K20 fertilizer to ponds increased the biomass yield of water
hyacinth. The fertilizer was added at 4 levels 0, 2.7 kg/ha, 10.8 kg/ha,
and 21.6 kg/ha. It was interesting to note that the highest level of
fertilization resulted in a yield less than the two intermediate levels.


320
4
Dunigan, E. P., Z. H. Shamsuddin, and R. A. Phelan. 1975. Water hyacinths
tested for cleaning polluted water. La. Agr. 18(2): 12-13.
Du Toit, R. 1938. Water hyacinth. Farming in S. Afr. 13(132): 16-17.
Dutton, J. 6. 1964. Louisiana's hyacinth control program. La. Conserv.
16(1-2): 6-8.
Dyar, H. 6. 1913. The species of Sphida Grote. Insec. Inscit. Menst.
1: 18-19.
Dyar, H. G. 1922. A note on Belluva gortynoides Wlk. Insec. Inscit.
Menst. 10: 50.
Dymond, G. C. 1948. The water-hyacinth: a Cinderella of the plant world.
Pages 221-227 in J. P. J. Van Vurens, ed. Soil fertility and sewage.
Faber and Faber, Ltd. London.
Eggler, W. A. 1953. The use of 2,4-D in the control of water hyacinth
and alligatorweed in the Mississippi Delta, with certain ecological
implications. Ecology 34(2): 409-414.
Evans, A. C. 1963. The grip of the water hyacinth. New Sci. 19(358):
666-668.
Forbes, W. T. M. 1954. The Lepidoptera of New York and neighboring states
Part 3. Noctuidae. Cornell University Agr. Exp. Sta. Mem. 329, 433 pp
Francois, J. 1964-3. Observations sur 1 'htrostylie chez Eiohhomia
1ovassipes (Mart.) Solms. Bull, des Seances 1964-3: 501-519.
Freeman, T. E., F. W. Zettler, and R. Charudattan. 1974. Phytopathogens
as biocontrols for aquatic weeds. PANS 20(2): 181-184.
Frick, K. E. 1974. Augmenting the weed control effectiveness of
phytopagous insects. Biol. Control Weeds Symp., Entomol. Soc. Amer.
Nat. Meeting, Minneapolis, Minnesota, Dec. 2-5, 1974, 19 pp.
Frost, S. W. 1975. Third supplement to insects taken in light traps at
the Archbold Biological Station, Highland County, Florida. Fla.
Entomol. 58(1): 35-42.
Gangstad, E. 0., D. E. Seaman, and M. L. Nelson. 1972. Potential growth
of aquatic plants of the lower Mekong River Basin Laos-Thialand.
Hyacinth Control J. 10: 4-9.
Ganguly, S., and S. M. Sircar. 1964. Cell growth and metabolism of pea
(Pisum sativum L.) internodes as affected by the growth substances
from the root of water hyacinth (Eichhomia oras sipes). Bull. Bot.
Soc. Bengal 18(1): 83-86.
Gay, P. A. 1958. Eichhomia orassipes in the Nile of the Sudan. Nature
182(4634): 538-539.


ib 1 e 6. Average daily rates of change in biomass from initial and final monthly
values.
Month
Standing Crop
Interval
(Days)
Daily Increment*
1st Estimate
Last Estimate
Jan
787
611
27
0.9907
Feb
798
757
20
0.9974
Mar
835
988
21
1.0080
Apr
929
1483
28
1.0168
May
1176
1545
28
1.0098
Jun
1750
2190
23
1.0098
Jul
1802
1682
26
0.9974
Aug
1451
1436
21
0.9995
Sep
1254
1500
21
1.0086
Oct
1501
1092
28
0.9887
Nov
1283
968
21
0.9867
Dec
1067
1005
20
0.9970
TIT"
T where T = interval, NT = final estimate, Nn = initial estimate.
To 0


Figure 8. Maximum, minimum, and median weekly air and water temperatures
at Lake Alice from late June 1974 through June 1975. The
winter air temperatures were unseasonably warm during this
study period.


(JM^UJ/O) NOIlDnaOdd NOIIV lOSNI
lie
I I ME
AMBIENT TEMPERATURE 1*0


142
These observations were somewhat surprising as a decrease in density
was expected as solar radiation levels fell and temperatures became colder.
Plant density was definitely higher in the winter, however, than in the
summer. This appears to be related to the leaf area index and plant height.
The average weekly frequency distribution for the various plant height
classes for each month is shown in Figure 20. In January the distri-
bition was narrow (range 60 cm) and skewed towards the small plants. The
dominant size class was 31-40 cm which contained approximately 32 plants
(26%). Taken together with the two smaller size classes, 55% of the plants
were found to be less than 40 cm during this month.
In February the distribution was narrower yet (range 50 cm) but the
predominant size class was larger (41-50 cm) and contained 31% of the
plants. Still the greater proportion of the plants were smaller than the
predominant class (42%) and taken together with the predominant class
represent 73% of the population.
In March the range of sizes seemed to broaden (60 cm) and there was
a lack of a single predominant size class. Four size classes (21-30, 31-40,
41-50, 51-60) accounted for 22%, 18%, 21%, and 24% of the population
respectively or 85% of the population collectively. By April when the
greatest increase in density occurred the size classes appeared to be
nearly normally distributed. The predominant class was 51-60 cm in
height and contained 38 plants (24%) and together with the 41-50 cm
class accounted for 42% of the population. Smaller size classes contained
26% of the population and larger plants accounted for 16%.
In May the distribution seemed to bifurcate. Two modes were apparent,
the first in the 41-50 range and the second in the 71-80 and 81-90 ranges.


LIST OF ILLUSTRATIONS
Figure Page
1. An aerial view of Lake Alice on the University of Florida
campus
2. Water level taken at weekly intervals and precipitation at
Lake Alice from July 1974 through June 1975 87
3. Total carbonate and bicarbonate alkalinity and conductivity
of water samples taken from Lake Alice from June 1974
through June 1975 95
4. Magnesium and total iron from Lake Alice water samples ... 97
5. Phosphorus concentrations present as phosphates and
nitrogen concentrations as total nitrate and nitrites from
water samples taken from Lake Alice 100
6. The negative logs of the hydrogen ion concentration (pH) of
water samples taken from Lake Alice 102
7. Potassium and sulfate ion concentrations of water samples
taken from Lake Alice 106
8. Maximum, minimum, and median weekly air and water tempera
tures at Lake Alice from late June 1974 through June
1975 108
9. Solar radiation data from the University of Florida campus
from May 1974 through April 1975 112
10. Diurnal curve for large waterhyacinth productivity
determined from CO^ gas exchange measured on Lake Alice with
an infrared CO2 gas analyser 116
11. Diurnal curve for small waterhyacinth productivity 118


this change to the weed infestation but failed to consider cultural
changes in the area. They concluded that the weed growth provides
improved conditions for planktonic development and thus benefits fish
production. The basis for these conclusions is obscure, however, and
doesn't agree with the findings of other authors (e^g. Wahlquist 1969b,
McVea and Boyd 1975). While it seems probable that a fringe of the
weed would be beneficial in some areas Bose (1945) commented ". . the
various reports of fish mortalities in stagnant pools and ponds covered
with waterhyacinth at once dispell the ideas and ruin the prospect that
waterhyacinth should ever be fancied in tropical countries as 'one of
the popular plants' for any kind of pisciculture."
Increasing attention has been directed towards using waterhyacinth.
Pirie (1960) advocated utilizing waterhyacinth as a crop. Waterhyacinths
are currently being considered for tertiary treatment of sewage effluent
(see Dymond 1948; Sheffield 1967; Yount 1964; Yount and Crossman 1970;
Boyd 1970b; Rogers 1971; Rogers and Davis 1971; Dunigan et al. 1975;
Dunigan and Phelan 1975). Its possible use as feed for livestock
(Chatterjee and Hye 1938; Baldwin et al. 1975; Bagnall et al. 1973, 1974;
Baldwin 1973; Boyd 1968a, b; Combs 1970; Hentges 1970; and Salveson 1971)
Apparently waterhyacinth is used as pig fodder in Singapore (Anonymous
1951). A complete cycle is developed when waterhyacinth is fed to pigs,
wastes and fecal matter are washed from the piggeries into the pond,
this fertilizes the pond to produce fish and more waterhyacinth which
are both harvested.
Chatterjee and Hye (1938) found that waterhyacinth was high in
potash with as much as 68% in the ash (5% on dry weight), comparable


Figure 7. Potassium and sulfate ion concentrations of water samples
taken from Lake Alice. The lines represent 5-point moving
averages.


334
Mr. A. A. Kirk of the Australian division of C.S.I.R.O. to help collect,
culture, and ship wood wasp parasites from Arizona for their biological
control program. His Master's Degree dealt with the biology and coevo
lution of seed beetles and their host plants.
He is a member of the Ecological Society of America and the Entomological
Society of America.
He is married to the former Deborah Jean Learned.


125
Seasonal Variation in Photosynthetic Tissue
Virtually all parameters associated with photosynthetic tissue showed
strong seasonal tendencies. Plant height, leaves per plant, leaves per
unit area, area per leaf, leaf area index, area per petiole, petiole area
index, and total leaf area index show seasonal maximum ranges but these
peaks vary in time depending upon the parameter measured.
Plant height (Fig 14) seems to follow solar radiation curves but lags
a month or two behind. Figure 13 gives average daily solar radiation for
Gainesville on a monthly basis as 8 to 12 year averages. The data for
the period of this study generally agrees with this (Figure 4c). Maximum
radiation occurs in late May or early June but maximum plant height
(mean) is not achieved until late June and July. Minimum solar radiation
ranges occur in December but plant height does not reach its lowest level
until late January.
The number of leaves per plant (Fig. 15) appeared to be extremely
variable. At the beginning of the study (May) the range was between
6 and 7 leaves per plant. The following May, however, it failed to return
to this level (ranges 4-5). A decline was observed for this parameter
throughout the summer followed by an increase in September. The gradual
decline in the fall and winter seemed to parallel the decline in solar
radiation but a sharp spring increase did not occur. This is explainable
in terms of plant density and leaf density (Figs. 16 and 19). Even though
the number of leaves per plant was low in the spring the number of leaves
per square meter was at a maximum because plant density was high during
this time. It may be concluded then that a plant has the greatest number
of leaves in the summer when it is at its maximum height but the maximum


265
Discussion
I assumed at the beginning of this experiment that parasite
populations would increase and ultimately reduce the introduced
Arsama densa population at the release site. By the end of the
second generation, however, this had not become apparent. The popu
lation of F-| seventh instar larvae was 72% greater than the population
of first instar larvae initially released. Furthermore, none of the
larvae collected at the release site were dead as a result of para
sitism while 44% of those at the control site were. Not only did the
initial population survive and reproduce contrary to my expectations
but it increased in the subsequent generation which was apparently
also surviving well.
Since this study was not designed to evaluate the population
dynamics of A. densa I can only speculate on the reasons for the success
of this population. I feel that by synchronizing the generations the
consequences of parasitism were effectively reduced. The parasites at
this site may have been "programmed" to low host populations and over
lapping generations. By inundating the natural population which had
individuals at various stages of development with the introduced
population with indidivudals all at the same stage of development the
synchronization of the age-specific parasites with the host may have been
imbalanced. For example, if parasites of the seventh instar larvae were
issuing at the time of release of the first instar larvae the ability
of the parasite population to increase would not change. Only those
parasites that are present at the time that the introduced host popu
lation is at an appropriate age would have an increased chance of


74
to convert from a scale reading to ppm CO^. The scale reading is based
on a comparison of two gases. Three pairs of gases were compared through
the analyzer. Ambient air vs. ambient air (air entering the chamber) was
compared to determine a zero point. The second comparison was chamber
exhaust vs. ambient air. This difference represented the CO2 gradient
through the chamber and was expressed as recorder scale division. The
value of a scale division (sd) is determined according to the level of
C02 in the ambient air by the equation ppm/sd = ae^x where x is the C02
concentration of the ambient air. The ambient air CO2 concentration
was determined by a third comparison. In this case a standard was used
of a known concentration. The standard was 300 ppm bottled gas and was
compared against the ambient air. The CO2 concentration in the ambient
air was determined by the equation ppm = ax2 + bx + c where x is the
recorder reading. This involves lowering the amplification of the analyzer
output by changing from "range 3" to "range 1". The range 1 equation is
calculated by running the standard 300 ppm gas through the reference
side of the analyzer and running other gases of known concentration through
the sample side. The value of the sample gas is correlated with the
recorder reading using the parabolic regression. Range 3 is calibrated
using various known C02 concentrations against a closed system aparatus
with flow and pressure maintained constant. The closed system is injected
with known quantities of pure CO2. An exponential regression is fitted
for the ppm/sd against ppm C02 of the various reference gases (ambient
air in this case).
The volumetric C02 concentration gradient (ppm C02) is converted
a gravimetric measurement (g C/m2) using the gas constant (0.14625 gm C-K/m2


Figure 19. Annual change in plant density as determined from weekly
samples taken in the study area. Each point represents an
average derived from counts taken from three 0.316 m2
samples. Only offsets with some root development were
counted as distinct plants. The dotted line represents
predicted values based on multivariate regression equations
(see Table 7).


63
as dasheen and corn have been implicated in the host range of this
group of insects a great deal of study of host specificity is warranted
and the taxonomic status of the group needs clarification. I do not
doubt these records but I am dubious of the placement of species identified
from these plants. On several occasions I have caged larvae of Arzama
densa (Wlk.) on both dasheen (Colooasia esoulenta Schott) and Xanthosoma
sp. and found that they did not feed upon them. Further studies are
severely needed to verify these host records.
Biology and Life History of Arzama densa Wlk. and Related Species.
The early literature on the biology of these species is sparse and
occurs primarily as notes of correspondence in various journals. The
first reference I have been able to find is that of Worthington (1878).
He described the larva of Arzama obliquata (G. & R.) and noted that it
was found "under the bark of a dead maple about three feet from the
ground, where it had made for itself an oval cavity in the dust". He
reared the adult and found that the pupal stadium was about 21 days
(April 27 May 18).
Comstock (1881) described the larva and aquatic habits of Arzama
melanopyga Grt. He was the first to take note of the large dorsally
situated pair of spiracles on the 9th abdominal segment which are
characteristic of the larvae of this group.
Riley (1883a, b) described the eggs of Arzama obliquata G. & R.
(misspelled Arsame) as being laid in "curiously broadly convex or plano
convex masses enveloped in hair, and a cream colored mucuous secretion,
when combined look much like spun silk on the inside, and on the outside
like the glazed exudation of Orgyia leuaostigma." He also noted the


242
is doubtful. Also, the root mass and the root-rhizome:shoot ratio was
small indicating a loss of efficiency. The increase in plants in the
summer is more likely a vegatative response of the surviving plants
before they are severely damaged to the increased space available. In
this manner the community maintains a larger standing crop and supports
a larger insect concentration before yield is affected.
Survival over periods of stress, such as winter freezing, would
probably be reduced by insect infestations. Penfound and Earle (1948)
stated that as the leaves are killed by frost the plant tends to
float higher increasing the susceptability of the rhizome to frost.
Insects have the same effect of removing mass and causing the rhizome
to become more exposed to temperature extremes. Further, the ability
of the plant to regenerate after periods of stress is probably decreased
as a result of insect attack because of a depletion of carbohydrate
reserves in the rhizome.
Suprisingly, productivity did not decrease significantly as a result
of insect attack. This is probably the result of a time factor. That is,
the insects failed to produce a noticeable effect on plant growth until
they reached a stage late in their development. Only the later
instars do severe damage to the rhizome. Once this point was reached
productivity probably was reduced but it occurred so late in the experiment
that it failed to show up in the results. This same phenomenon was
apparent in the total biomass estimates. The only factors that did not
show a significant response to insect attack were net production, total
biomass, and plant density.
All factors associated with the standing crop showed a highly sig-
/
nificant decrease as insect concentrations increased. The proportions
of the various parts of the plants changed but these changes, for the


238
While rhizome weight increased the most in the controls, roots
increased the least (Figure 40) or more acurately did not increase (102%).
This agrees with field observations where I have found that small water-
hyacinth plants have nearly the same root mass as th large plants when
growing in similar conditions. When the plants are small crowding is
minimized and nutrients would be more limiting than light. The plants,
therefore, probably maximize root growth early in their growing period.
As the plants grow and light becomes more limiting due to intraspecific
competition the available energy is probably shunted more towards shoot
development and less towards root development. This would explain the
decrease in the proportions of the plants represented as roots and the
decrease in the root-rhizome:shoot ratio observed in the controls in
Table 11.
Roots increased the least in the control and decreased the most in
the treated plots (28% of initial values with 1.00 insect per plant).
Values for final proportions ranged between 41 and 53% of the initial
proportions after insect feeding. This is not significantly different
than the control (60%). Because shading by adjacent plants is reduced
and increased mount of light is available for growth. For optimum
regeneration of the plant, nutrients need to be absorbed rapidly for
the full photosynthetic potential to be realized. Ideally, then, the
proportion of the plants represented as roots should become larger as
is normally the case in small plants. The root proportion does not
change as a result of insect attack, therefore regeneration of damaged
plants is probably slower than would be expected. Because light energy
and nutrient energy react multiplicatively in plant production the
availability of one affects the utilization of the other (i.e., growth is


79
nutrient levels. Two samples were collected each week. One sample
was analyzed using a Hach DR-EL portable test kit for total alkalinity
(carbonate + bicarbonate), total nitrates + nitrites, pH, total phosphates,
and sulfates and a Hach micro-iron test kit (model IR-18-A) for iron.
The second sample was taken to the University of Florida Soils Laboratory
where it was analyzed for conductivity, magnesium, and potassium.
The methods applied to the water samples are as follows:
Alkalinity (total) Titration of Brom Cresol Green Methyl
Red indicator with 0.020 N. sulfuric acid 10 ml sample.
Conductivity Platinum electrode ohmmeter.
Iron 1, 10 Phenanthroline Method 25 ml sample.
Magnesium Atomic absorption spectrophotometer.
Nitrates and Nitrites (total) Cadmium reduction method -
25 ml sample.
pH Colorimetric reading with a wide range indicator.
Phosphates (total) Colorimetric method 25 ml sample.
Potassium Flame emission spectrophotometer.
Sulfate Turbidimetric method.
The procedures employed in the Hach Test Kit are more for convenience
and direct reading and are not as accurate as other techniques. For my
purposes the loss in accuracy is outweighed by simplicity of the proce
dures. These procedures are probably accurate enough to indicate temporal
differences but are probably not extremely definitive.
Maximum and minimum air and water temperatures were taken at the
same location as the water samples. Two Taylor (No. 5458) maximum-minimum
self registering thermometers were mounted on a C-shaped styrofoam


28
A few authors have suggested that stands of waterhyacinth may
modify the pH of the water. Penfound and Earle (1948) noted that pond
waters in the Mississippi River delta have an average pH of 7.2 whereas
water in waterhyacinth mats are usually acid. Ultsch (1973) compared
open water areas of a pond with areas covered with waterhyacinth and
determined the yearly average pH to be 5.6 in the open areas and 5.4
in the areas with waterhyacinth. Haller and Sutton (1973) presented
data which indicated that the plants cause a change in the direction of
neutrality from both high and low initial pH values. Center and Balciunas
(1975) compared water quality parameters from sites with and without
waterhyacinth and found that those with the plants had lower pH (7.06
0.84) than those without (7.551.06) although the difference was not
significant.
Moisture requirements of waterhyacinth and the effects of dessication
upon its survival and growth have been only superficially examined.
Webber (1897) noted that if the plants are to succeed a soil of loose
texture thoroughly saturated with water is required. Parija (1934), however,
found that they could survive 5.7% of water saturation in soil. Bock (1966)
noted one instance when the plants survived 41 da in saturated soil.
She speculated that the plants can withstand periods of dessication
because excessive transpiration is prevented from the center of the
rosette by the protective layer of dead outer leaves. Penfound and Earle
(1948) found that waterhyacinth could survive drying periods up to 18 da
depending upon climatic conditions and the surface they are exposed on.
Sunny weather with the plants on galvanized metal killed the plants
rapidly while rainy and cloudy weather or placing the plants in the


276
(205 and 225 eggs/ ) are lower than that of Vogel and Oliver 1969b..
They reported an average of 328 eggs per female with 8.25% infertility
based on 10 mated females.


121
leaf tissue of the large plants is only 59% of that of the small plants
even where differences in solar radiation are taken into account. The
contribution of the photosynthetic layer in the petioles is not well
understood but it is thought to be of minor importance in primary pro
duction (Knipling et at. 1970). Further Knipling et al. (1970) found
that the light was rapidly extinguished beneath the upper canopy (ca.
25% at mid-height). Hence the primary function of petiole is probably
the supportive function of leaf display and the leaf tissue is probably
photosynthetically more important than the total photosynthetic tissue.
Table 5 lists the various metabolic comparisons standardized with regard
to the leaf area index, photosynthetic tissue, leaf tissue, and standing
crop. In all cases gross primary productivity is greater in the small
plants than in the large ones even though the reverse is true strictly
on a per unit area basis. The fact that the amount of leaf tissue pre
sent is 9 times greater in the large plants while the GPP/gm leaf tissue
is 40% less indicates that the increased leaf area interferes with light
reception and probably results in a greater deviation from potential
productivity.
The ratio of plant parts (Fig. 12) may be important in terms of
supporting photosynthetic processes and may partly account for the
difference in growth rates. As mentioned previously the petioles function
in displaying the leaves but probably do not contribute greatly to photo
synthesis. Hence, even though they are necessary, they represent a sub
stantial metabolic cost to the plant. Petioles account for 46% of the
weight of a large plant and only 19.1% of a small plant. A considerably
greater portion of the small plant is non-photosynthetic in nature than


NITROGEN
PHOSPHORUS
(Wdd) SrUdOHdSOHd-aiVHdSOHd
n


turnover. A seasonal aspect was implicated in the plant response as the
insects appeared to be much more effective in the fall than the summer.
This may be related to the energy budget of the plants under varying
conditions of solar flux. Plant density increased in the summer in
response to insect attack probably as a result of decreased intraspecifi
competition. A similar response would be expected to any factor which
reduced competition provided adequate energy for growth was available.
To learn if A. densa could be used in biological control a green
house reared population was released on a small pond in August 1974.
The waterhyacinth population was reduced and competing plants began to
dominate the site. Ultimately cat-tail invaded and waterhyacinth failed
to re-invade. A control site remained dominated by waterhyacinth.
Studies of natural A. densa populations on Lake Alice indicated
that the failure of this insect to achieve sufficient levels to have
an extensive effect on waterhyacinth was most likely due to the complex
of parasites which attack them. Also, pickerelweed (Pontederia cordata)
appears to be the preferred host of this insect which may partly explain
the low populations observed on waterhyacinth. Further, seasonal changes
in the plants' ability to withstand insect attack may obscure correla
tions between plant characteristics and insect damage.


177
plants presumably because the large plants are closer to steady state
with a smaller P:R ratio. This is in spite of the fact that net efficiencies
are approximately equal for both size classes. One would expect, then,
that the plants would grow faster in the spring when the plants were
small than in the summer when they were large even if external conditions
were equal. This cannot be attributed entirely to an increased metabolic
load in large plants because the gross primary productivity per gram of
leaf tissue decreases as the plants become larger (see Table 5). This
infers that a unit of large plant leaf tissue is less efficient than an
equal unit of small plant leaf tissue. Respiration per gram biomass is
almost double in the small plants so this cannot account for the difference.
One possible reason for this difference is senescence of the older leaves.
A second is intraspecific competition for light. Both of these explanations
are probably partly true. As the leaves become older they probably do
naturally become less efficient. Also they are more likely to have been
attacked by diseases, mites, insects, and other factors which may reduce
their effectiveness. Also, as the plants become larger they are more prone
to self shading. Hence, even though the amount of light received is the
same as the small plants the amount received per unit of photosynthetic
area would be inversely proportional to the degree of competition.
This pattern of self regulation by intraspecific competition seems
to account for changes in plant density within the waterhyacinth community
better than any of the physical parameters included in the plant
density model. As shown earlier, small plant size classes are lost and
plant density decreases as the plants become larger. Figure 26 further
illustrates this. In this figure density is plotted as log function of
plant height. This slightly improves the correlation coefficient derived


51
form or another. Bock (1966, 1969) has done perhaps the most comprehen
sive study on productivity but she dealt with fresh weight and increment
factors making comparisons with her data difficult. She also reviewed
most of the literature on the subject and compared it to her data. Table 1
gives an updated compilation of various measures of standing crop and
productivity of waterhyacinth from various sources.
Control
The literature dealing with the various means of control is vol
uminous and I won't attempt to review it here. The Hyacinth Control
Journal has been published annually since 1962 and is largely devoted
to this subject. Furthermore, the various control methods have recently
been reviewed. Robson (1974) has reviewed the methods for mechanical
control of aquatic weeds and Blackburn (1974) has reviewed chemical
control and the various compounds available in a recent UNESCO publication.
In the same publication Bennett (1974) reviewed the biological control
of aquatic weeds. Biological control has also been reviewed by Andres
and Bennett (1975) and the use of plant pathogens in biological con
trol efforts by Zettler and Freeman (1972), Freeman et at. (1974) and
Charudattan (1975). Mitchell (1974) summarized techniques for the con
trol of aquatic weeds through habitat management. Sculthorpe (1967)
also discussed the various methods of aquatic weed control.


314
LITERATURE CITED
Abu-Gideiri, Y.B., and A. M. Yousif. 1974. The influence of Eichhomia
crassipes Solm.on plantonic development in the White Nile.
Arch. Hydrobiol. 74(4): 463-467
Agarwal, S. C. 1974. Water hyacinth in West Bengal and its control.
Plant Prot. Bull. (Delhi) 23(3): 21-24.
Agharkar, S. P., and I. Banerji. 1930. Studies in the pollination and
seed formation of water-hyacinth (Eichhomia speciosa Kunth)
Agr. J. India 25(4): 286-296.
Agostini, G. 1974. El genero Eichhomia (Pontederiaceae) en Venezuela.
Acta Bot. Venezuelica 9(1-4): 303-310.
American Public Health Association. 1965. Standard methods for the
examination of water and wastewater, including bottom sediments and
sludges. Amer. Public Health Assoc., Inc. N.Y. 769 pp.
Andres, L. A., and F. D. Bennett. 1975. Biological control of aquatic
weeds. Annu. Rev. Entomol. 20: 31-46.
Andrewartha, H. G. 1961. Introduction to the study of animal populations
University of Chicago Press, Chicago. 281 pp.
Anonymous. 1896. Clogged by hyacinths Navigation on the St. Johns,
Florida, seriously obstructed. The Sun [newspaper], New York,
Sept. 20, 1896.
Anonymous. 1951. The water hyacinth problem and pig farming. Sci. Cult.
17(6): 231-232.
Anonymous. 1957a. Commission for Techincal Co-operation in Africa
South of the Sahara. Report of the symposium of Eichhomia crassipes
Leopoldville, 1957. C.C.T.A./C.S.A. pubis. 27: 1-31.
Anonymous. 1957b. The water hyacinth now a "prohibited imigrant."
E. Afr. Agr. J. 22(3): 129.
Anonymous. 1964a. Occurence of water hyacinth. F.A.0. Plant Prot. Bull
12(4): 93.
Anonymous. 1964b. Water hyacinth is one oi the world's worst weeds.
New Zeal. FI. Agr. 108: 232.
Arber, A. 1918. The phyllode theory of the monocotyledonous leaf, with
special reference to anatomical evidence. Ann. Bot. 32(128): 465-501
Arber, A. 1920. Water plants, a study of aquatic angiosperms. University
Press, Cambridge. 436 pp.
Arnold, C. A. 1940. A note on the origin of the lateral rootlets of
Eichhomia crassipes (Mart.) Solrns. Amer. J. Bot. 27: 728-730.


27
Knipling et al. (1970) measured net productivity of attached
leaves under a range of light conditions. Photosynthesis increased from
7.8 mg C02/dm2 leaf surface/hr to 16.1 mg/dm2/hr as the light intensity
increased from 1450 ft-c to 8000 ft-c. Dark respiration was found to
range from 2.6 to 2.8 (average 2.7) mg/dm/hr.
Waterhyacinth is generally considered to tolerate a wide range of
pH (Pieterse 1974). Haller and Sutton (1973) found they grew over a
range of 4.0 to 10.0 although optimal growth occurred in acid to slightly
alkaline conditions (4-8). Bock (1966) citing data from other authors
concluded that waterhyacinth generally occurs in waters ranging in pH
from 4 to 9. Chadwick and Obeid (1966) compared the growth of water
hyacinth and water lettuce (Pistia stratiodes L.) in cultures of varying
pH. They found that waterhyacinth would grow at all levels (pH 3.0 to
8.2) but at 3.0 both dry-weight yield and offset production were minimal.
They felt that pH values near 7.0 were optimal for waterhyacinth but
values near 4.0 were optimal for water lettuce. Penfound and Earle (1948)
reported pH values usually ranged between 6.2-6.8 in or near waterhyacinth
mats in Louisiana but could survive extremes of 4-5 and 9-10. In the
Guinean region of Africa pH is thought to be limiting at values of 4.2
or below (Berg 1959; Anonymous 1957).
Minschall and Scarth (1952) studied the effects of low ranges of
pH (3.5-6.5) on the roots of waterhyacinth. They found that at values
below 4.0 the roots exhibited decreased cell division and cell elongation.
Cell division at pH 5.0 proceeded twice as fast as at 3.5. They further
found that the plants could tolerate more acidity at cooler temperatures
and the pH of the cell sap was always above that of the culture medium.


22
1879-1892. Bock (1966), however, indicated that it was not introduced
into Egypt until 1912. It was introduced into South Africa around 1910
as an ornamental and by 1938 was reported from rivers in the Cape Pen
insula, George, Knysna, Albany, Port Elizabeth, Uitenhage, Victoria
East, and Natal (DuToit 1938). It had apparently reached South Rhodesia
prior to 1937 as Europeans settling there reported its presence at that
time (Holm et al. 1969; Bock 1966).
By 1942 waterhyacinth had spread into Mozambique in the Incomati
estuary from Vila Luisa to Xanowano and apparently originating from the
Transvaal of South Africa (Mendonca 1958). Kirkpatrick (1958) indicated
that waterhyacinth was already present in Zaire (the Belgian Congo) in
the Congo River in 1954. Coste (1958) felt that it was introduced in
the period between 1950 to 1951. Other authors (Bock 1966; Holm et al.
1969) list 1952 for its introduction into the Congo. By 1955 it had
spread over 1600 km of the river between Leopoldville and Stanleyville
(Kirpatrick 1958). Gay (1958) first observed waterhyacinths occurring
on the White Nile of the Sudan in 1958 along about 1000 km. It was
apparently not abundant in the river prior to 1957 although it may have
been present in 1956.
Senegal first reported the presence of waterhyacinth in 1964
(Anonymous 1964) from the Cape Vert peninsula and this may perhaps be
the first record in the northwestern part of Africa. In spite of the
warnings expressed by the Inter-African Phytosanitary Commission it
was still available for purchase from street hawkers in Senegal in 1965.
Waterhyacinth is now distributed in all of the tropical and sub
tropical areas of the world. Its northernmost limits of distribution
are probably near Sacramento, California (Lat. 38.5 N; Bock 1968),


8
the aquatic food chain may change from a plant-herbivore based community
to a detritus-detritivore based community (e.g., Hansen et al. 1971) as
a result of this loss of submersed primary productivity.
Waterhyacinths may also successfully compete with valuable wildlife
forage thereby replacing it. This may destroy feeding areas for waterfowl
(Gowanloch 1944). Local economies may be seriously damaged in areas which
cater to recreational needs such as waterfowl hunting, fishing, boating,
waterskiing, swimming, etc.
More seriously, riverine communities in the developing areas of
the world which depend on fishing as a primary source of protein may
be denied access to fishing grounds (Holm 1969). Holm (1969) further
stated that impoundments for fish culturing may be destroyed by large
masses of floating waterhyacinth. He stated that waterhyacinths con
stitute "...the most massive, most terrible and frightening weed
problem" he had ever known.
Waterhyacinth possible pose a health threat by harboring vectors
and intermediate hosts of human diseases. The larvae and pupae of Mansonia
uniformis (Theob.), a mosquito vector of filariasis in Asia, are known
to attach to the roots of waterhyacinths (Burton 1960; McDonald 1970).
Waterhyacinths may result in an increased production of mosquitoes by
hindering insecticide application, interfering with predators, increasing
the habitat available for certain species which attach to the plant, and
by impeding runoff and water circulation thereby creating stagnant
impoundments for breeding (Seabrook 1962). Mulrennan (1962) cautions
that uncontrolled aquatic plant populations could lead to an increased
incidence of mosquito-borne diseases such as malaria, encephalitis, and


239
limited by the necessary resource which is least available). When light
energy is abundant, as in a sparsely populated stand, nutrient levels
probably limit growth. One would expect the strategy of the plants in
this situation to be that of maximizing root development thereby increasing
nutrient absorption and minimizing the limiting effect of nutrients. In
a dense stand where intraspecific competition for light is intense light
availability would be expected to be limiting. In this situation selection
would favor those plants which maximized photosynthetic tissue and the
tissues necessary for photosynthetic display. Hence, a greater proportion
of the available energy would be expected to be utilized in producing
photosynthetic tissue. Small plants, therefore, would be expected
to have a larger root-rhizome:shoot ratio than large plants. The effect
of the insects in this experiment resulted in smaller plants but the
root-rhizome:shoot ratio did not become larger than that of the large
plants in the control (it in fact was generally smaller). As a result
the available light increased but the plants could probably not efficiently
utilize this increase because of the limiting effect of the absorptive
ability of the small root mass. Small slowly growing plants were pro
duced as a result instead of small rapidly growing plants or large slowly
growing plants. The insects caused not only a decrease in the standing
crop but probably also inhibited the natural regenerative ability of the
plants remaining.
The stolon weight decreased directly as a result of insect activity
(Figure 42). Stolon weight as a proportion of plant weight should reflect
offset production. If the effect of the insects was that of increasing
the offsets produced then the stolon weight and the stolon weight:plant
weight (since plant weight is reduced)should increase. This did not


169
serve as indices of biomass (see Table 9) are regulated by climate as
solar radiation and air temperature are most important. Leaf density and
plant density appear to be regulated more by hydroponic conditions as
various water quality parameters are emphasized.
Predicted values were generated for each dependent variable based
on the known independent variable values. Each has been plotted as an
annual curve with the actual observed curve (see Figs. 14, 15, 16, 18,
19, and 24). Biomass estimates (Fig. 24) seemed to fit well in the win
ter, spring and fall but values were underestimated in the summer. This
is probably because of the assumption of linear effects inherent in the
model. As mentioned previously this assumption is probably not justified.
Plant height (Fig. 14) was approximated fairly accurately. The observed
drop in late August was not apparent, however, and an increase in late
January was predicted which did not occur. This is apparently the result
of a brief period of warm weather which, according to the model, should
have resulted in a brief increase in height. Otherwise the included
variables satisfactorily account for variations in height.
Values predicted for leaf area index produce a curve of approxi
mately the same shape as the observed data (Fig. 18). The peak was not
predicted until early June, however, where it actually occurred in mid-
May. The predicted peak corresponds to the peak in solar radiation.
The predicted curves for leaves per plant (Fig. 15) and leaf !
density (Fig. 16) conform extremely well to the actual data. Plant
density (Fig. 19) is also well represented but the spring peak is not
as dramatic as was observed. This is probably because the change at
this time was exponential and the model treats it in a linear fashion.


71
factor to begin, attained an optimum at a certain level, and declined
as levels of the limiting factor exceeded maximum tolerable levels.
This is parallel to Shelford's (1913) "Law of Tolerance" where he
essentially states that the failure of an organism may be due to an
excess or deficiency of any one factor which may approach the maximum
or minimum limits of tolerance of the organism for that factor. For
an aquatic plant, such as waterhyacinth, these limiting factors include
temperature, light, water, dissolved or available nutrients, space, etc
Phytophagous insects probably cause a threshold type response in
the plants whereby the plant can sustain certain levels of damage without
obvious deterioration until maximum tolerable limits are exceeded. As
insect damage exceeds these threshold levels a rapid decline in the
(
populations or standing crop may be evident. Levels of insect damage
)
below this threshold may cause various plant response. When the popu-
J
lation is at steady state (the stable maximum level restricted by the
level of a limiting factor) insects may disrupt this stability causing
the plant population or standing crop to fall below the carrying capacity
of the system. This may have the effect of reducing intraspecific com
petition in the plant population. In this case the limiting factors
would become increasingly more available and production may indirectly
be stimulated. Hence yield may be increased under low insect concen
trations where the insects prevent senescence of the population by
increasing the rate of turnover.
This study was designed to measure the effects of various environ
mental factors as well as the effects of a natural buildup of an indi
genous insect population (Arsama densa) on a stand of waterhyacinth.
The parameters considered can be grouped into climatological conditions
(temperature and solar radiation), limnological conditions (nutrients,


Table 5.
Metabolic and morphometric comparisons of the two morphological types of
waterhyacinth studied.
Parameter
Large Plants
(11-12 Aug. 1973)
Small Plants
(12-13 Aug. 1973)
Standing Crop (gm/m^)
2131.75. __
588.99
Leaf Area Index (m2/m2)
7 78
3.30
Photosynthetic 1 issue (gm/nv1)
1250.98
446.87
Leaf (Pseudolamina) Tissue (gm/m)
267.53
30.38
Incident Solar Radiaion
3750.
4900.
Gross Primary Productivity (gm C/m -da)
19.30 (25.23* )
15.60
GPP/LAI
2.48 ( 3.24* )
4.73
GPP/gm Photo. Tissue
0.015 ( 0.020*)
0.034
GPP/gm Leaf Tissue
0.072 ( 0.094*)
0.513
GPP/Standing Crop
0.009 ( 0.012*)
0.026
Net Primary Productivity (gm C/m -da)
6.1 ( 8.0* )
8.0
NPP/LAI
0.784 ( 1.03* )
2.424
NPP/gm Photo. Tissue
0.007 ( 0.009*)
0.018
NPP/gm Leaf Tissue
0.005 ( 0.006*)
0.263
NPP/Standing Crop
0.003 ( 0.004*)
0.014
Plant Respiration (gm C/m -da)
13.2
7.6
Respiration/Standing Crop
0.006
0.013
Total Efficiency (kcal GPP/Sol. Rad. x 100)
5.15/,
3.18%
Net Efficiency (kcal NPP/Sol. Rad. x 100)
1.63/
1.63%
Gross Primary Productivity:Respiration
1 46
2.05
Daily Rate of Increase (1 + Stan3PPCrop)
1.006 ( 1.008*)
1.030
Non-Photo. Tissue:Photo. Tissue
0.71
3.13
Assumes incident solar radiation equal to that of the small plants. This adjustment
is necessary for comparisons since considerably greater solar radiation values were
recorded during the day of the small plant study.


287
Population Cycles
Vogel and Oliver (1969b) felt that Avzama densa produced at least
2 and possibly 3 generations per year in Louisiana. Their estimate was
based on their determination of the length of the developmental period
and a 120 day winter diapause. Larvae were present at all times of the
year, however, and appeared to be most abundant in September and October
(Vogel and Oliver 1969a).
Larvae were sampled from 1 May 1974 to 30 April 1975 at weekly
intervals on Lake Alice, Gainesville, Florida from the sampling plots
described in Section 1. The larvae collected were placed in snap top
pill vials and returned to the laboratory. They were maintained by
feeding them waterhyacinth petiole sections. Each larva was classified
as to instar at the time of collection. If they died I attempted to
establish the cause of death. When parasites emerged they were identified .
If no parasites issued shortly after the death of the larvae they were
checked for diseases by Dr. George Allen. Dead larvae collected in the
field were handled the same way. Parasite pupae found in Avzama
burrows were collected and reared. Pupal exuviae were noted as well as
the probable instar of the dead host larva.
Figure 48 illustrates the total population density of A. densa
larvae and pupae on Lake Alice for the 1974-75 sampling period. This
includes all larvae both living and dead. The same data is presented
in Figure 49 by individual instar. While eggs were seldom encountered
by sampling they were noted in the field from September through March.
This information indicates the existence of continuously brooded, over
lapping generations. Eggs were most abundant in January and February
which leads me to believe that no winter reproductive diapause occurs.


190
than the linear model. Because of this and because the log model produced
predicted values based on geometric means which underestimated the arith
metic means this model was never selected. If a greater number of higher
levels of insect infestation were used in the experiment I believe the
exponential nature of the response curves may have become more evident.
Even though the regressions in this paper are linear the true relationship
between the plant response and the insect concentration may not be
linear.


31
Community Associations
Because of the worldwide distribution of waterhyacinth any com
prehensive list of plants associated with it would be a tremendous task.
A few authors have breached this subject on a local level, however.
Harper (1903) noted the association of an orchid, Habenaria repens Nutt.,
with waterhyacinth in Georgia. Small (1936) listed a dozen plants which
may be found growing on the floating mats of waterhyacinth and noted in
New Orleans that it grows intimately with several other aquatic species.
Penfound and Hathaway (1938) described plant communities in the marshlands
of southeastern Louisiana. They found waterhyacinth associated with the
cypress-gum swamps in strictly fresh water and presented an extensive list
of other associated species. Penfound and Earle (1948) found a tremendous
array of plants (63 species) occurring on mats of waterhyacinth. Eggler
(1953) investigated the effects of 2, 4-D on other plant species associated
with waterhyacinth and al 1igatorweed. Chadwick and Obeid (1966) investi
gated antagonism between Pistia stratiotes and Eichhomia crassipes. Bock
(1966) listed several species associated with waterhyacinth in California
and reviewed the work of several other authors. Abu-Gideiri and Yousif
(1974) noted the composition of the plankton community in association
with waterhyacinth stands in the Sudan.
Several workers have reported on the invertebrates associated
with waterhyacinth but this has largely been the result of biological
control investigations dealing primarily with insects. O'Hara (1967)
quantitatively listed the invertebrates found in waterhyacinth mats.
Hansen et al. (1971) listed some invertebrates present in the aquatic
component of the waterhyacinth community and constructed a partial food
web. They also studied the vertebrates present as did Goin (1943).


92
Table 4.
A comparison of water quality measurements from Lake Alice, Gainesville,
Florida with previous reports.1
Source of Data
Time Period
Point Sampled
Brezonik et al. (1969)a
Dec 1968
Composite of 3 Stations
Mitsch (1975)b
Jan Sept 1973
Marsh Discharge
This Study0
July 1974-June 1975
Marsh Discharge
Sp. Conductance
533
379 484
464.53 (61.37)
Alkalinity
178
133 182
175.96 (24.27)
pH
7.4
7.6 8.3
7.33 (0.25)
nh3 (n)
0.01
0.053 0.184
...
NO3 (N)
0
0.008 0.938

no2 (N)
...
0.01 0.03
...
N02 + NO3

...
0.99 (0.62)
Ortho-P
0.07
0.73 1.00
...
Total-P
0.59
0.86 2.10
1.12 (0.45)
Chloride
16.7

...
Sodium
16.0
15 19
...
Calcium
71
14-65
...
Iron
0.01
...
0.097 (0.073)
Potassium

0.5 6.5
3.97 (2.02)
Sul fates
...
103.9 (38.1)
Magnesium
...
5.5 16.0
11.35 (3.38)
* Single determination.
b Range based on monthly averages.
c Mean (standard deviation) of all samples for the time period.
1 Units in mg/1 except for conductance (umho/cm), pH (pH units), and alkalinity
(mg/1 as HC03 in Brezonik and Mitsch and as HCOj + COf in this study).


55
In 1889 Grote placed these species in the tribe Arzamini which
included Arsama and Sphida, the former supposedly having a smooth front
and the latter a tuberculate front. He considered Arzama as consisting
of three species, apparently after having considered A. melanopyga to be
a variety of A. vulnifica. This was also how the group was arranged in
his checklist of 1890.
Smith (1893) decided that Walker's Bellura and Arzama were congeners
and that Bellura had page priority. He considered A. densa Wlk., A.
vulnifica Grt., and A. melanopyga Grt. synonyms of B. gortynoides Wlk.
He recombined Arzama diffusa Grt. into B. diffusa (Grt.). Edema obliqua
Wlk., Sphida obliquata (G. & R.) and Arzama obliquata G. & R. were
considered synonyms of Bellura obliqua (Wlk.). Thus, the seven species
were reduced to three, all in Bellura Wlk.
BeutenmT1er (1902) redescribed B. obliqua (Wlk.) and B. gortynoides
Wlk. but he also recognized B. melanopyga. All three were found in New
York. Holland (1903) considered B. densa (Wlk.), B. vulnifica (Grt.),
and B. melanopyga to be synonyms of B. gortynoides Wlk. as did Smith
(1893) but recognized the genus Sphida and considered S. obliquata (G. & R.)
a synonym of S. obliqua (Wlk.). Hampson (1910) recognized the genera
Sphida, by the single species S. obliqua, and Bellura. He considered
A. densa Wlk. and A. vulnifica Grt. synonyms of B. gortynoides, and
retained B. melanopyga and B. diffusa. He also presented a key for
separating the three Bellura spp.
Dyar (1913) revised the genus Sphida, described three new species,
and he provided a key. He retained s. obliqua (Wlk.) and considered
E. obliqua Wlk. and A. obliquata G. & R. synonyms. The new species
described were S. oecogenes from Washington, D. C., S. anoa from Miami,


160
population. An increase in damage was apparent beginning in July and
continuing through November when a peak of approximately 25% occurred.
A decline followed through December until late January when the damage
ranged between 5 and 10%. It remained at this level through February and
March with a brief increase in April. By mid-May the level of damage
was somewhat greater than the previous year.
Rhizome damage closely paralleled leaf damage but was generally
larger peaking in November where 43% of the plants had rhizome
damage (Fig. 25). Rhizome damage was very low from May through September
(Figure 20). An increase was apparent in October but most of the damage
was minor. Most of the severe damage occurred in November, December, and
January. By spring the relatively low larval population coupled with the
high plant density resulted in relatively low levels of damage. The in
crease noted in April was primarily minor damage.
The frequency of attack of a given plant size class roughly corresponds
to the frequency of that size class (Figure 20). This is generally true
throughout the year except in December and January when the smaller plants
are more abundant but most of the damage occurs to the larger plants. This
may be partly responsible for the loss of plants in the larger classes
apparent between December and January. Otherwise it appear that there is
no selection by the insect for the size of plant attacked. Larger plants
may be attacked more frequently at certain times but this is probably
due to the fact that they are older and have been exposed to attack for
a longer period of time. If the selection of a plant in a given size
class by a larva is a random process then the frequency of attack within
that class should correspond to the relative abundance of that class within
the frequency distribution. Without further analyses this appears to be the case.


40
with light intensities increasing from 1450 ft-c to 8000 ft-c. Respiration
averaged 2.7 mg C02/dm2/hr. Ultsch and Anthony (1973) have found that
waterhyacinth may have the capacity to utilize C02 dissolved in the
wc^ter^under the mats as a source of carbon by absorption throuc
;oots. This majc-aeetTT f or up^Kr'->99Laf the total carbon fixed. Billings
and Godfrey (1967) have found that some hollow stemmed plants may use
internal carbon dioxide generated from root and stem respiration in
photosynthesis. This may be true in the case of waterhyacinth.
Nutrient uptake rates are not well worked out for this plant.
Waterhyacinth contains about 0.4% P and 2.6% N by weight (Boyd 1970b)
or an N:P ratio of 6:1. If these represent constant proportions the rate
of nutrient uptake is proportional to the growth rate of the plant.
If the standing crop increases at a rate of 100 g/da the rate of uptake
of N will be 2.6 g/da and of P 0.4 g/da. This agrees well with the
results of Dunigan et al. (1975) who found the N:P ratio of uptake rates
to be 5-6:1. The daily absorption rates from 6 liter containers were 2.4
ppm N and 0.4 ppm P in water concentrations of 50 and 100 ppm N and P.
At concentrations of 250 ppm N and P the daily uptake rates were 3.5 and
0.7 ppm respectively. This implies a growth rate of 600 and 810 mg dry
wgt/da. Mitsch (1975) indicates that this high ratio of N:P absorption
indicates that nitrogen is generally more limiting to waterhyacinth
than phosphorus.
Phenology
Data on the sequence and timing of events in the annual cycles of
waterhyacinth population are scarce. Penfound and Earle (1948) measured
the average length of the largest leaves over one growing season. The


9
other arboviruses. Mitchell (1974) indicated that aquatic plants,
including waterhyacinths, may also harbor snails which are intermediate
hosts of diseases such as fascioliasis and shistosomiasis. He mentioned
that those species which do occur on waterhyacinth are assisted in
their dispersal by the free-floating habit of this plant thereby spreading
the associated diseases. Bock (1966) further reviews the literature
dealing with health hazards caused by waterhyacinth.
It is difficult to arrive at sound figures regarding the monetary
losses caused by waterhyacinth. Spencer (1973, 1974) quoted the following
figures from a congressional report for losses prevented by waterhyacinth
control in Louisiana in 1957.
Navigation
Flood Control
Drainage
Agriculture
Fish & Wildlife
Public Health
14,727,000
250,000
$37,993,000
Penfound and Earle (1948) conservatively estimated that water-
hyacinths were responsible for losses of $5 million annually as of 1948
in Louisiana. Spencer (1973, 1974) quoted figures from a Louisiana
Fisheries and Wildlife report indicating losses of $65-75 million in 1947
in Louisiana due to aquatic weeds. In 1930 a report to the Jacksonville
City Commission estimated that in the period from 1900-1930 the Federal
Government spent $233,000 on waterhyacinth removal in the Jacksonville
District to simply maintain "open channels for navigation" (Buckman and
Co. 1930). By 1961 the total cost had risen to $1,861,788 in the same
district (Tabita and Woods 1962). Wunderlich (1964) reported costs of
clearing aquatic weeds ranging from $15-60 per acre. As of 1964 about


47
ripen. Seeds with mature coats failed to germinate when kept at either
23C or 29C. He then separated the embryo from ripe seeds or ruptured
the seed coats and placed them in a bath at 29C. He noted that germi
nation occurred very rapidly in both cases (96% after 1 day). He ruled
out oxygen as a factor because they germinated equally well in boiled
water covered with paraffin. He concluded that the hard seed coat and
endosperm hinders water absorption and limits germination and that
desiccation may, in fact, fracture the seed coat and promote germination.
Agharkar and Banerji (1930) indicated that a ripening period of
20 to 23 days was required for maturation of the fruit. After maturation
they are severed from the axis by an abscission layer and float on the
water surface for a day or two before sinking. Splits develop on the
lateral walls through which seeds are discharged. They found that the
seeds develop freely in tap or distilled water.
Parija (1930) suggests that germination takes place "in the
beginning of rains or whenever the humidity, soil moisture and temperature
are suitable." (Parija 1930, p. 388). He felt that the function of the rain
was to provide moisture, and expose the seeds in the mud providing access to
oxygen.
Robertson and Thein (1932) noted that in every instance when they had
found waterhyacinth seedlings it was in a depression which completely dries
out during the dry season and floods again in the rainy season. They concluded
that a period of drought alternating with a period of plentiful moisture was
necessary for germination.
Haigh (1936) exposed seeds to varying treatments of always wet,
always dry, or alternately wet and dry. No germination occurred for three
months unti eeds were placed in the sun. Within eleven days germination


182
in Florida's nutrient rich waters and moderate climate large stands of
tall plants are common.
The pattern of growth of waterhyacinth seems to involve several
phases on Lake Alice. A period of "no-growth" occurs in January and
February where the standing crop begins to increase an increase in leaf
density occurs. This begins as early as February and seems to be the
first phase of growth. The leaf density peak occurs in late March. The
second phase of growth is an increase in plant density. This begins in
early March and the peak occurs in late April. The number of leaves per
plant increases as leaf density begins its increase but levels off in
March and April only to begin a new increase in May. The spring increase
in leaf density is due both to an increase in the number of leaves per
plant and an increase in plant density.
The third phase of growth is an increase in height which begins in
late March and peaks in July. The increase begins when both leaf density
and plant density are high and may be a response to this. Standing crop
begins to increase at the same time as height and peaks at about the same
time.
The fourth growth phase is an increase in leaf size and does not
begin until May but reaches a peak in early July along with height and
standing crop. Leaf production appears to be manifest first in an in
crease in leaf density, second in an increase in the number of leaves per
plant, and finally in an increase in leaf size. The adaptive strategy
seems to be that of maximizing leaf area. Under conditions of low com
petition this can best be achieved by producing more offshoots. Under
conditions of intense competition each plant produces more leaves and the
leaves increase in size. This may be interpreted as a diversion of


183
available energy in the path of an energy gradient. When the solar energy
gradient is stronger peripherally lateral growth occurs. As peripheral
light penetration diminishes due to increased competition and the light
gradient becomes relatively stronger vertically, vertical growth begins
to occur. Either way the increase in the leaf area index from early
February through late May is almost continuous.
Growth appears to slow in the summer and a dramatic decrease in
both leaf and plant density occurs. A summer decline in the number of
leaves per plant, plant height and standing crop occurs in late July,
August, and early September. The reasons for this are not apparent but
it may be the result of a change in the carrying capacity of the system.
These characteristics level off for a short while until a general decline
begins in mid-September. This decline continues until the winter lows
are reached in January. As plant height, leaf size, leaves per plant, and
standing crop decline leaf density and plant density begin to increase.
This increase continues until January when a slight decline occurs.
These annual cycles illustrate the plasticity of the waterhyacinth
population in adapting to different situations. The population is regu
lated primarily by climatological factors, by water quality, and by the
intensity of intraspecific competition. Rapid adjustment in the morpho
metry of the mat occurs as these factors change. It is not unreasonable
to expect that this capacity to adjust may apply to attack by insects. By
reducing intraspecific competition insects may indirectly cause an increase
in density. There probably is, of course, a damage threshold beyond which
further damage by insects could severely affect the ability of the water-
hyacinth population to adjust. I would expect this threshold to be high,


ROOTS
125-
* i lili
0.2 0.4 0.6 0.8 1.0
INITIAL INSECT CONCENTRATION
0


20
to the Buitenzorg Botanic Gardens in Java in 1894 (Bock 1966; Sculthorpe
1967). By 1905 it had appeared in Ceylon (Jepson 1933). It had become a
serious problem in Cochin China (a part of S. Vietnam) by 1908, in Burma
by 1913 and in Bengal by 1914 (McLean 1922). The Philippines acquired
the plant around 1912 and by 1926 it was appearing in China as well as
Borneo and Malaysia (Bock 1966). Records as to its entry in Japan are
scarce but they were apparently introduced during this century as
ornamentals (Shibata et al. 1965; Bock 1966). The first record of its
occurance in Okinawa was in 1952 (Bock 1966) but it may have been there
earlier.
From these beginnings it now occurs throughout India, Southeast
Asia, and Indonesia (Bancroft 1913; Barber and Hayne 1925; Jepson 1933;
Penfound and Earle 1948; Anonymous 1951; Sen 1961; Little 1965, 1968a;
Holm et al. 1969; Chhibbar and Singh 1971; Haigh 1936; Hitchcock et al.
1949; McLean 1922; Agarwal 1974; Robertson and Thein 1932; Shibata et al.
1965; Gangstad et al. 1972; Ishaque 1952).
Waterhyacinth was first introduced into Australia in Queensland
in 1895 (McLean 1922) and in New South Wales in 1896 (Maiden et al.
1906). It was apparently eradicated from New South Wales but a reinfest
ation occurred in the 1940's (Parsons 1963; Bill 1969), to South
Australia by 1937 (Bill 1969) and to Victoria by 1939 (Parsons 1963).
Bill (1969) notes that waterhyacinth is not a serious problem in
Australia today except in some Queensland rivers.
Waterhyacinth also occurs in New Zealand (Walker 1954; Taylor 1955;
Anonymous 1964b; Little 1965,1968a; Holm et al. 1969) although it is
difficult to determine when it first appeared there. Taylor (1955)


The responses of the plant population to insect feeding tend to
verify similar experiments using greenhouse cultures of waterhyacinth.
Plant density increased a small amount but this was probably an
opportunistic response to available space as the larger plants initially
present died and light penetration increased. The degree of this
response appears to be related to the quantity of light available and
is probably seasonal in nature. The plants produced were much smaller
thus floating higher in the water and were probably more susceptible
to temperature extremes.
A reduction in the leaf canopy was apparent as height, leaf density,
number of leaves per plant, and the biomass of photosynthetic organs
decreased. This does not necessarily indicate a reduction in net pro
ductivity since smaller plants tend to be more efficient than larger
ones (Browne et al. 1974). The proportion of the plant represented as
leaf blades did not change while the proportion represented as roots
increased. The major change appeared to be in the petiole proportion
which is primarily a structure for supporting and displaying the leaves.
In this open stand this supportive structure would not be as important
as in a dense stand. As a result the relative photosynthetic ability
of the plant per gram of biomass was probably not affected but the
relative ability to absorb nutrients probably increased. This is
reflected in the increased root-rhizome to shoot ratio. In the green
house studies a decrease in the root-rhizome:shoot ratio was evident.
This is in contrast to the increase noted here. Perhaps, in the former
case, an insufficient amount of time was available for root regrowth
prior to harvesting. The plants produced at the release site in the
field experiment were much more typical of small plants growing in


Figure 23. Average weight per plant as a log function of the average
plant height. Data from biomass samples taken from Lake
Alice.




Figure 22. The weekly waterhyacinth height class frequency distributions
plotted three dimensionally on a time scale. The horizontal
axis represents size classes increasing from left to right.
The vertical axis represents density, the number of plants in
each height class per square meter. The third axis, perpen
dicular to the plane of the paper, represents time (weeks).


203
Plant Density
When discussing the effects of stress on the density of waterhyacinth
(number per unit area) one must be careful to define the seasonal stage
of the plant and the type of stand it is growing in. Eiohhomia orassipes
requires abundant solar energy, waters rich in nutrients, and open space
for maximum productivity (see the discussion on the seasonal ecology of
this plant presented earlier). Most factors which are likely to stress
waterhyacinth (such as herbicides, frost, or insects) tend to open up
the canopy. The plants become smaller and as a result the P/R ratio and
net efficiency probably increases (Brown et at. 1964). In this situation
they become r-strategists and reproduce rapidly competing successfully
with other species for the available space. This is observed in the spring
when there is an apparent "overshoot" in the plant population. The data
from experiments with laser beams for the control of this plant illustrate
the same principle (Long and Smith 1975). After the laser treatment a
decline in the rate of change was observed but it was immediately followed
by a sharp increase until the experimental plots were almost identical to
the control plots.
In this r-selection situation species with high rates of reproduction
and growth are more likely to survive in an uncrowded situation (E. P. Odum,
1969). When the waterhyacinth stand matures and occupies the total avail
able space an equilibrium density is reached. In this situation the
plants appear to be -strategists where selection favors species with
lower growth potential but which are more competitive under equilibrium
densities (E. P. Odum, 1969). The effect of the insect (or other stress
factors) appears to be that of causing the plant to "switch" strategies
/
by disturbing the equilibrium density. Discontinuation of the stress


202
Figure 30 illustrates the response curves in terms of leaves/unit
area. The changes in this measurement were quite different between the
two seasons. In the fall the percentage of the initial number of leaves
present showed a significant decline with increasing insect
During the summer, however, no change was apparent. Since
the number of leaves per plant was evident this stability i
per unit area was apparently due to the increase in plants.
concentrations,
a decrease in
n the leaves


¡07
described by Lloyd and Dybas (1966) for the periodical cicada and by Janzen
(1969) for seeds. In both cases it was suggested that this strategy employs
a sudden increase in the population of susceptible individuals allowing
them to escape before their resceptive predators could respond numerically.
The synchronization of age involving the escape of insects from parasites
may be considered part of this strategy. This, in effect, minimizes the
time period available for successful oviposition by the age-specific
parasite and again requires a rapid numerical response. Since the interval
between instars suitable for a particular parasite would be maximized
the probability of continual parasitism would be minimized.
Large scale testing of this theory of augmentation of the Arzama
densa population is precluded by the inability to mass rear suitable
numbers of larvae for release. Further, a great deal of basic research
on the biology and population dynamics of both Arzama densa and its para
sites should be completed. The potential for control of waterhyacinth by
this insect species exists and warrants further study. It may be used in
other countries simply by importing it free of parasites and predators
or in this country through more complex means similar to those suggested
above. A thorough understanding of the host specificity of this insect,
its taxonomic status, and its populations on Pontederia should take high
priority in future studies.


330
Skinner, H. 1903. Arzama melanopyga. in Doings of Societies. Entomol.
News: 209-210.
Small, J. K. 1933. Manual of the southeastern flora. University of
North Carolina Press, Chapel Hill. 1554 pp.
Small, J. K. 1936. The water-hyacinth a time clock. J. New York Bot.
Gard. 37: 35-41.
Smith, J. B. 1893. Catalogue of Noctuidae. U.S.N.M. Bull. 44:1-424.
Smith, W. R. 1898. A contribution to the life history of the Ponte-
deriaceae. Bot. Gaz. 25: 324-337.
Spencer, N. R. 1973. Insect enemies of aquatic weeds. Proc. 3rd Int.
Symp. Biol. Control of Weeds: 39-47.
Spencer, N. R. 1974. Insect enemies of aquatic weeds. PANS 20(4): 444-450.
Standley, P. C. 1928. Flora of the Panama Canal Zone. Contrib. U.S.
Nat. Herb. 27: 111-112, Washington; U.S. Gov't. Printing Office.
Stone, A., C. W. Sabrosky, W. W. Wirth, R. H. Foote, and J. R. Coulson.
1965. A catalog of the Diptera of America north of Mexico. U.S.
Gov't. Printing Office, Washington, D.C.. 1696 pp.
Sutton, D. L., and R. D. Blackburn. 1971a. Uptake of copper by water
hyacinth. Hyacinth Control J. 9(1): 18-20.
Sutton, D. L., and R. D. Blackburn. 1971b. Uptake of copper by parrot-
feather and water hyacinth. Proc. 24th Annu. S. Weed Sci. Soc.: 331.
Swamy, B. G. L. 1966. The origin and organization of the embryonic
shoot apex in Eiohhomia orassipes. Bull. Torr. Bot. Club 93(1 ): 20-34.
Tabita, A., and J. W. Woods. 1962. History of hyacinth control in
Florida. Hyacinth Control J. 1: 19-23.
Tag el Seed, M. 1972. Some aspects of the biology and control of Eiohhomia
orassipes (Mart.) Solms. PhD thesis, University of Khartoum. 274 pp.
Tag el Seed, M., and M. Obeid. 1975. Sexual reproduction of Eiohhomia
orassipes (Mart.) Solms in the Nile. Weeds Res. 15(1): 7-12.
Takhtajan, A. 1969. Flowering plants, origin and dispersal (translation
by C. Jeffrey). Smithsonian Institution Press, City of Washington.
310 pp.
Taylor, C. R. 1955. Control of water hyacinth in Rotorua district.
New Zeal. J. Agr. 90: 188-189, 191, 193, 195.


139
followed by a sharp increase. On the whole the leaf area index seems to
follow solar radiation cycles. Cold weather causes a depression in the
leaf area index but the effect was short due to the relatively mild winter.
The strategy of the plant seems to be to maximize photosynthetic display.
This is done first in the spring by increasing leaf density and offset
production followed by an increase in leaves per plant and leaf size in
the summer as intraspecific competition becomes more intese.
As was discussed previously, the importance of the petiole in photo
synthesis is not known. It is assumed, however, to contribute little.
For this reason the petiole area index has not been graphed. The area
per petiole is usually very close to the area per leaf. Hence, the total
leaf area (pseudolamina and petiole) is approximately twice the leaf
area and the total leaf area index is approximately twice the leaf area
index. All three (leaf area, petiole area, and total leaf area) strongly
parallel the curve for mean maximum height (Fig. 14).
Seasonal Variation in Plant Density
Plant density (Fig 19) did not seem to follow the same trends as
the various estimates of photosynthetic tissue. A major peak occurred
in late April when the density reached 186 plants per square meter.
This was followed by an equally abrupt decline in May. By June the density
was between 70 and 90 plants per square meter and it remained in this
range until September. At this time the density began to increase and a
secondary peak of 130 to 140 plants per square meter was achieved in early
January. This level dropped slightly in February but the spring increase
began in early March.


Figure 30. The effects of varying levels of insect feeding activity
on the total number of waterhyacinth leaves per unit area
expressed as a percentage of predetermined means. Legend
as in Figure 28.
Summer: Y = 135.11 + 1.72X, r = 0.012
Y = 13-.29 85.54X, r -0,572
Fall:


113
alone. This is probably
afternoon thunderstorms
(see rainfall in Figure
the result of cloud cover associated with
common in this part of Florida during the
2).
frequent
summer


240
appear to be true in this experiment. The stolon:plant weight ratio
(Table 11) was not significantly different between the infestation levels
tested. This agrees with the plant density changes discussed earlier.


4
imply. For this reason I follow Kelsey and Dayton (1942) and use the
spelling waterhyacinth throughout this dissertation.
Waterhyacinth is a member of the pickerelweed family which includes
9 genera and 36 species (Cook et al. 1974). The genera are Eiohhomia
(7 spp.), Eurystemon (1 sp.), Heteranthera (10 spp.), Hydrothrix (1 sp.),
Monochoria (5 spp.), Pontederia (5 spp.), Reussia (4 spp.), Soholleropsis
(1 sp.), and Zosterella (2 spp.). The majority of the members of this
family.are confined to the Americas although two of the genera (Mono
choria and Soholleropsis) appear to be Old World endemics. Keys and
descriptions of the genera (world-wide) may be found in Cook et al.(1974).
Lowden (1973) revised the genus Pontederia and united Reussia with it.
Castellanos (1958) provides notes on the genus Pontederia in Brazil and
keys and descriptions of the Brazilian species of Pontederiaceae
(Castellanos 1959).
Description and Account of Variation
Following is a translation of the description of Eiohhomia orassipes
(Mart.) Solms given by Agostini (1974). Further descriptions can be
found in Castellanos (1959), Sculthorpe (1967), Bock (1966), Penfound
and Earle (1948), Buckman and Co. (1930), Webber (1897), and many others.
The first definite description of this species (according to Bock 1966)
was that of Kunth (1843).


PERCENTAGE OF INITIAL VALUE
196


253
Figure 43(a)
Figure 43(b)


243
most part, were not out of line with those of the controls. This was
surprising because the smaller plants produced were expected to be
similar to the plants growing in open stands (i.e., a high root-rhizome:
shoot ratio). This did not appear to be true.
The turnover ratio was the most revealing characteristic measured.
Living material was killed and transformed to detritus at rates almost
directly proportional to the insect concentrations. The total amount
of living material initially present survived approximately 25% as long
with 1 larva per plant as did the plants without insects. This acceler
ated turnover could severely affect the competitive ability of the plants.
In summary, Avzama densa severely affected almost all aspects of
waterhyacinth growth. It does appear to be a good agent for biological
control and mass releases should be attempted. The chances for success
would be greater in the fall than in the summer and I believe concern
over a resultant increase in plant density and dispersal is unwarranted.
%
The major problem with this insect is that parasite buildups tend to
causes pulses in the effects of A. densa and as noted previously
waterhyacinths recuperate rapidly after cessation of stress factors.
Exotic insects which are not limited by parasites and exert continuous
pressure on the plant community will probably have a greater long term
possibility of providing a permanent control. Releases of A. densa
may be beneficial in conjunction with exotics to bring the plant popula
tion down to levels more easily controlled by the exotics.


2
The third section investigates the potential effects of Arzama densa
Wlk. on greenhouse cultures of waterhyacinth. These effects are evaluated
in terms of various ecological and morphological traits of the plant.
The fourth section deals with the feasibility of augmenting natural
populations of A. densa as a means of biological control. The effects
of a small scale release are evaluated from a small pond near Paynes
Prairie.
This fifth section constitutes notes on the biology and life
history of A. densa. Data on the natural population at Lake Alice is
presented and the probable reasons for the failure of this insect to
control waterhyacinth is discussed.
It is hoped that the information reported here will provide a
baseline from which future comparisons can be drawn after the release
of exotic insects. It is also hoped that an increased understanding of
the ecology of waterhyacinth in a situation relatively free of specific
insect enemies has been gained.


Figure 15. Annual variability in the number of leaves per waterhyacinth
plant from the study area. The means are derived from weekly
samples and represent averages from all plants contained in
three 0.316 m2 samples. Only leaves which had unfurled were
counted. The dotted line represents predicted values based
on multivariate regression equations (see Table 7),


15
yeast cells thereby promoting fermentation. S.M. Sircar and Chakravarty
(1961) found that this root extract increased the yield of jute (Corehorus
eapsularis L.) and the production of fiber. P.K. Sircar et al. (1973)
identified four gibberelin-1 ike compounds from extracts of waterhyacinth
shoots.
Other attempts to find marketable products of waterhyacinth include
ink, upholstery stuffing, rope, bags, plastics, timber substitutes, and
ice chests (Bock 1966). Morton (1962) gives a recipe for preparing water-
hyacinths for food. For further reviews of the literature dealing with
the utilization of waterhyacinth and other aquatic weeds see Bock (1966),
Sculthorpe (1967), Little ( 1968b ), Boyd (1972, 1974) and Mitchell
(1974).
Distribution
It is generally accepted that South America is the area of origin
of the waterhyacinth. Small (1936) indicates that it was originally
discovered in the San Francisco River near Malhada, Brazil in 1824.
Bock (1966) cites Hooker (1829) as listing several early collections
from Brazil; Demerara River, Guiana; New Granada ( a former Spanish
viceroyality including present day Venezuela, Ecuador, Colombia, and
Panama); Guayaquil, Ecuador; and Buenos Aires, Argentina. She also
cites early references indicating it was also native to the West Indies
(eg. Schwartz 1928; Britton 1918). Castellanos (1959) lists its present
South American distribution as Argentina, Paraguay, Brazil, Uruguay,
Chile, Ecuador, Colombia, and Guyana. It has also been reported in
Surinam (Little 1965, 1966; Holm et al. 1969). It has apparently been
widespread in South America for many years as evidenced by the early
records.


Figure 20. Average monthly counts of the number of plants included in
each plant height class per square meter. The monthly
values were derived from the averages of all the weekly
samples taken during a given month. The dark vertical
bars represent the frequency of damage to the rhizome by
Arzama densa in each height class.


323
Hitchcock, A. E., P. W. Zimmerman, H. Kirkpatrick, Jr., and T. T. Earle.
1949. Water hyacinth: its growth, reproduction and practical control
by 2,4-D. Contrib. Boyce Thompson Inst. 15(7): 363-401.
Hitchcock, A. E., P. W. Zimmerman, H. Kirkpatrick, Jr., and T. T. Earle.
1950. Growth and reproduction of water hyacinth and alligator weed
and their control by means of 2,4-D. Contrib. Boyce Thompson Inst.
16(3): 91-130.
Holland, W. J. 1903. The moth book. Doubleday, Page, and Co., New York.
Holm, L. G. 1969. Weed problems in developing countries. Weed Sci.
17(1): 113-118.
Holm, L. G., L. W. Weldon, and R. D. Blackburn. 1969. Aquatic weeds.
Sci. 166: 699-709.
Hooker, W. J. 1829. Pontederia azurea. Large-flowered Pontederia.
Curtis' Bot. Mag., Entry 2932.
Ingersoll, J. M. 1974. The water hyacinth. Aquatic Plant Control Prog.
Tech. Rep. 6: A3-A33, U.S. Army Eng., WES, Vicksburg, Mississippi.
Ishaque, M. 1952. Water hyacinth: a problem of East Pakistan and its
control. Proc. Pakistan Sci. Conf. 4(3): 10-11.
Iswaran, V., and A. Sen. 1973. Influence of extract of water hyacinth
(Eiahhomia arassipes) on the yield of brinjal (Solanwn melongena)
var. Pusa Kranti. Sci. Cult. 39(9): 394.
Jameson^ D. A. 1963. Responses of individual plants to harvesting.
Bot. Rev. 29(4): 532-594.
Janzen, D. H. 1969. Seed eaters versus seed size, number toxicity and
dispersal. Evolution 23(1): 1-27.
Jepson, F. P. 1933. The water hyacinth problem in Ceylon. Trop. Agr.
Mag. Ceylon Agr. Soc. 81: 339-355.
Johnson, E. 1920. Fresno County will fight water hyacinth. Calif.
Dept. Agr. Monthly Bull. 9(5-6): 202-203.
Johnston, J. 1889. Arzama obliquata. in Correspondence Can. Entomol.
21: 79.
Jones, J. R. J. L. 1951. An annotated checklist of the Macrolepidoptera
of British Columbia. Entomol. Soc. B.C. Occ. Paper 1: 85.
Kellicott, D. S. 1883a. in Correspondence Can. Entomol. 15: 171.
Kellicott, D. S. 1883b. in Correspondence Can. Entomol. 15: 174-175.
Kellicott, D. S. 1889. Arzama obliquata. Can. Entomol. 21: 39.


THE POTENTIAL OF ARZAMA DENSA (LEPIDOPTERA: NOCTUIDAE)
FOR THE CONTROL OF WATERHYACINTH WITH SPECIAL
REFERENCE TO THE ECOLOGY OF WATERHYACINTH
(EICHHORNIA CRASSIPES (MART.) SOLMS)
By
TED DOUGLAS CENTER
A DISSERTATION PRESENTED TO THE
GRADUATE COUNCIL OF THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS
FOR THE DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
1976

ACKNOWLEDGEMENTS
I wish to thank the numerous individuals who have assisted in
these studies.
I would like to express my appreciation Dr. E. E. Grissell,
Dr. E. L. Todd, Dr. R. E. Woodruff, Dr. T.J. Walker, Dr. R. Carlson,
and Dr. C. W. Sabrosky for the identification of insect specimens;
Dr. G. E. Allen, L. P. Kish, and Dr. E. I. Hazard for diagnosing insect
diseases; C. Cagle, C. Siebenthaler, M. White, G. Presser, N. R. Spencer,
and D. Butler for field and technical help; the University of Florida
Soils Laboratory for analysing water samples; Dr. E. A. Farber for
providing solar radiation data; the U.S. Department of Agriculture and
the Florida Division of Plant Industries for providing space and facil
ities; Ann Owens and Susan Kynes for library and literature research
assistance; Cath Siebenthaler for typing and editorial assistance in the
original manuscript; N.R. Spencer and T. C. Carlysle for photographic
and dark room assistance; and my graduate committee, Dr. D. H. Habeck,
Dr. T. H. Walker, Dr. R. I. Sailer, Dr. G. E. Allen, and Dr. J. Reiskind
for critical reading of the manuscript.
I would especially like to thank Mr. Neal R. Spencer for providing
space and facilities and the U.S. Army Corps of Engineers for providing
funds.
I would also like to thank my wife, Debbie, whose patience and
endurance saw me through to the conclusion of this work.

TABLE OF CONTENTS
Page
ACKNOWLEDGEMENTS
LIST OF TABLES VU
LIST OF ILLUSTRATIONS IX
ABSTRACT xiv
INTRODUCTION 1
LITERATURE REVIEW 3
Eiohhomia crassipes 3
Taxonomy 3
Description and Account of Variation 4
Economic Importance 6
Distribution 15
Habitat 23
Community Associations 31
Growth and Development 33
Morphology 33
Perennation 38
Physiological Data 38
Phenology 40
Reproduction 41
Floral Biology 41
Seed Production and Dispersal 45
Viability of Seeds and Germination 46
Vegetative Reproduction 50
Productivity and Standing Crop 50
iii

Page
Control 51
Arzama densa Wlk 53
Taxonomy 53
Hosts Plants 60
Biology and Life History of A. densa and Related
Species 63
Parasites, Predators, and Diseases 68
CHAPTER. 1. THE RELATIONSHIP BETWEEN THE PHENOLOGY AND BIOLOGY
OF WATERHYACINTHS AND VARIOUS PHYSICAL AND BIOLOGICAL FACTORS 70
Introduction 70
Methods and Materials 73
Diurnal Waterhyacinth Productivity 73
Annual Cycles and Insect Damage 76
Site Description 81
- Analyses 88
Results 91
Water Quality 9,1
Temperature and Solar Radiation 104
Waterhyacinth Productivity 114
Seasonal Variation in Photosynthetic Tissue . 125
Seasonal Variation in Plant Density 139
Seasonal Variation in Standing Crop 150
Damage by Arzama densa 157
Results of the Multivariate Analysis 161
Discussion 170
TV

Page
CHAPTER 2. THE CONSEQUENCES OF ATTACK BY ARZAMA DENSA WLK. ON
SOME ECOLOGICAL CHARACTERISTICS AND MORPHOMETRIC FEATURES OF
WATERHYACINTHS 185
Introduction 185
Methods and Materials 187
Analyses 189
Results 191
Plant Height 191
Leaves 197
Plant Density 203
Biomass Estimates 208
Productivity and Turnover Estimates 219
Plant Parts and Proportions 225
Discussion 241
CHAPTER 3. THE FEASIBILITY OF THE UTILIZATION OF ARZAMA DENSA WLK.
FOR THE BIOLOGICAL CONTROL OF WATERHYACINTH THE EFFECTS OF AN
INTRODUCED POPULATION ON A SMALL POND COMMUNITY 244
Introduction 244
Methods and Materials 247
Results 250
Discussion 265
CHAPTER 4. SOME NOTES OF THE BIONOMICS AND POPULATION DYNAMICS
OF ARZAMA DENSA WLK 271
Introduction 271
Habits 272
Fecundity 275
Duration of Developmental Stages 278
v

Page
Population Cycles 287
Mortality 292
Discussion 304
Review of results and suggestions for further studies . 308
LITERATURE CITED 314
BIOGRAPHICAL SKETCH 333

LIST OF TABLES
Table
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
Page
Standing crop and productivity of waterhyacinths as estimated
by various authors 52
Host plants of the Bellura-Arzama complex listed from various
literature sources.
Hydrological budget for Lake Alice (March September 1973)
61
85
A comparison of water quality measurements from Lake Alice,
Gainesville, Florida with previous reports 92
Metabolic and morphometric comparisons of the two morphological
types of waterhyacinth studied 120
Average daily rates of change in biomass from initial and
final monthly values 156
Summary of multivariate regression analyses for annual varia
tion plant characteristics 152
Correlation coefficients (r) between independent variables. .
164
Correlation coefficients (r) between dependent variables. .
.167
Correlation coefficients (r) between dependent and independent
variables and probabilities for a greater ¡r | 172
Ratios of the various plant parts and the percent change in
the final values as compared to the initial values 236
Comparison of the samples from the release site with the
control site based on various estimates of the plant and
insect populations
vii
251

Table
Page
13. Ratio of plant parts at the two sites on 12 December
1974 263
14. Fecundity, egg viability, and egg stadia for 5 female Arsama
densa collected as pupae in the field and mated in the
laboratory 277
15. Summary of development data for A. densa 279
16. Annual summary of larval counts and mortality 297
17. A summary of insects known to parasitize Arsama densa Wlk.
(from Vogel and Oliver 1969b in part) 305
v i i i

LIST OF ILLUSTRATIONS
Figure Page
1. An aerial view of Lake Alice on the University of Florida
campus
2. Water level taken at weekly intervals and precipitation at
Lake Alice from July 1974 through June 1975 87
3. Total carbonate and bicarbonate alkalinity and conductivity
of water samples taken from Lake Alice from June 1974
through June 1975 95
4. Magnesium and total iron from Lake Alice water samples ... 97
5. Phosphorus concentrations present as phosphates and
nitrogen concentrations as total nitrate and nitrites from
water samples taken from Lake Alice 100
6. The negative logs of the hydrogen ion concentration (pH) of
water samples taken from Lake Alice 102
7. Potassium and sulfate ion concentrations of water samples
taken from Lake Alice 106
8. Maximum, minimum, and median weekly air and water tempera
tures at Lake Alice from late June 1974 through June
1975 108
9. Solar radiation data from the University of Florida campus
from May 1974 through April 1975 112
10. Diurnal curve for large waterhyacinth productivity
determined from CO^ gas exchange measured on Lake Alice with
an infrared CO2 gas analyser 116
11. Diurnal curve for small waterhyacinth productivity 118

Page
Figure
12. A comparison of the standing crop and proportions of the
plant parts for the large and small waterhyacinth plants
used in the productivity studies 123
13. Average daily solar radiation values per month for
Gainesville, Florida 127
14. Annual phenological change in the average height of the
waterhyacinth plants on the marsh side of Lake Alice . .129
15. Annual variability in the number of leaves per waterhyacinth
plant from the study area 131
16. Annual change in leaf density as determined from weekly
samples taken in the study area 133
17. The average area of the pseudolaminae of waterhyacinth
leaves 136
18. Leaf area index of the waterhyacinth population on Lake
Alice 138
19. Annual change in plant density as determined from weekly
samples taken in the study area 141
20. Average monthly counts of the number of plants included in
each plant height class per square meter 144
21. Statistics of skewness and kurtosis (peaking) derived from
each weekly frequency distribution of plant density by
height classes 146
22. The weekly waterhyacinth height class frequency distribu
tions plotted three dimensionally on a time scale 149
23. Average weight per plant as a log function of the average
plant height 152
x

Figure Pa9e
24. Standing crop values, both estimated and real, from Lake
Alice 154
25. Percentage of the leaves and rhizomes of the waterhyacinth
population damaged through feeding activity of Arzama densa
at Lake Alice 159
26. Plant density as a function of plant height 179
27. Average dry weight per waterhyacinth plant as a log function
. of the average height 193
28. The average height per waterhyacinth plant (as measured from
the longest leaf) as a function of the feeding activity of
Arzama densa larvae 196
29. The effects of varying levels of insect feeding activity on
the average number of leaves per waterhyacinth plant expressed
as a percentage of predetermined means 199
30. The effects of varying levels of insect feeding activity on
the total number of waterhyacinth leaves per unit area
expressed as a percentage of predetermined means 201
31. The effects of varying levels of insect feeding activity on
the number of waterhyacinth plants per unit area expressed
as a percentage of predetermined means 205
32. The effects of varying insect concentrations on the total
waterhyacinth biomass (expressed as both detritus and
living plant material). 210
33. The effects of varying insect concentrations on the living
waterhyacinth mass present per unit area 212
XI

Page
Figure
34.The effects of varying insect feeding activity on the amount
of dead waterhyacinth plant material (detritus) per unit
area 214
35. Detritus as a percentage of total waterhyacinth biomass as
a function of insect feeding activity 218
36. Net waterhyacinth production as a function of insect feeding
activity 221
37.. The ratio of conversion of living waterhyacinth plant material
into detritus as a function of insect feeding activity. . .223
38. The effects of varying insect feeding activity on waterhyacinth
green mass (pseudolaminae and petioles) 227
39. The effects of varying insect feeding activity on waterhyacinth
non-green mass (roots, rhizomes, and stolons) 229
40. The effects of varying insect feeding activity on waterhyacinth
root mass per unit area 231
41. The effects of varying insect feeding activity on the water
hyacinth rhizome mass present per unit area 233
42. The effects of varying insect feeding activity on the water
hyacinth mass represented as stolons per unit area 235
43. A photographic comparison of the waterhyacinth stands at
experimental and control sites at different times of the
year following the release of Avzama densa at the
former 253
44. A comparison of the standing crop of waterhyacinths at the
control site and the release site 259
xii

Figure Pa9
45. The mass represented by the various plant parts for an
average waterhyacinth plant at both the control and
release sites 262
46. Probit analyses for developmental times of a greenhouse
reared population of Arzama densa 282
47. The head capsule diameter of Arzama densa larvae at each
molt plotted against the larval age 285
48.. The total number of Arzama densa larvae collected, either
living or dead, from the marsh side of Lake Alice 289
49. The age structure of the Arzama densa population during the
period of this study 29T
50. The population of living Arzama densa larvae (per square meter)
present on Lake Alice and the number of dead larvae expressed
as a percentage of the total 296
51. The total number of 4th and 7th instar Arzama densa larvae
per square meter as estimated from samples taken from the
marsh side of Lake Alice 299
52. The number of parasites of 4th instar (Campoletis sp.) and
7th instar (Lydella radiis) Arzama densa larvae as
estimated from the number of pupae, or pupal exuviae found in
A. densa bores per square meter of waterhyacinth mat . .302
xi i ¡

Abstract of Dissertation Presented to the
Graduate Council of the University of Florida in Partial
Fulfillment of the Requirements for the Degree of Doctor of Philosophy
THE POTENTIAL OF ARZAMA DENSA FOR THE CONTROL OF
WATERHYACINTH WITH SPECIAL REFERENCE TO THE
ECOLOGY OF WATERHYACINTH (EICHHORNIA CRASSIPES)
By
Ted Douglas Center
June 1976
Chairman: Dr. Dale H. Habeck
Major Department: Entomology
Waterhyacinth (Eichhovnia erassipes (Mart.) Solms) is one of the
world's worst aquatic weeds. Many countries have begun to consider
biological control as a means to alleviate infestations. The United
States has begun introductions of exotic insects for this purpose. To
evaluate properly the success or failure of these introduced insects,
background information on the ecology of waterhyacinths, the effects
of insect attack on waterhyacinths, and the population biology of indi
genous insects in the area of release is imperative. The purpose of this
dissertation is to provide this information for Lake Alice, a primary
study site on the University of Florida campus.
The ecology of waterhyacinth is approached in this dissertation
through a comprehensive literature review and through field studies
conducted to monitor various growth parameters. Diurnal productivity
xiv

curves comparing small and large plants revealed that the net efficiency
of both plants was 1.6% (of incipient solar energy). The small plants
grow faster than the large ones, however, by virtue of a larger P:R
ration (2.05 vs. 1.46).
Photosynthetic efficiency is maintained by synchronization of the
leaf area index with the annual solar energy flux. An annual increase
in leaf area occurs first through an increase in leaf density and
secondly through an increase in leaf (pseudolamina) size. The net result
of these two growth phases is a peak in the leaf area index spanning
the period of maximum solar radiation.
Intraspecific competition is strongly implicated in governing plant
density and seems to account for observed changes in the population.
Plant density is high in the winter reaching a maximum in April. This
is followed by a decline in May and June as a result of the loss of
plants in the smaller size classes. This loss is due to shading by the
larger plants as they increase in size and leaf area.
Multivariate analyses indicate that solar radiation and minimum
air temperatures were important in accounting for changes in standing
crop, plant height, leaf area index, and the numbers of leaves per plant
(all indices of biomass). The introduction of water quality parameters
into the analyses resulted in confusion as causal relationships were
difficult to establish.
Damage by Arzama densa Wlk. (Lepidoptera: Noctuidae) did not appear
to affect the population of waterhyacinth studied. Greenhouse studies
revealed that concentrations of 33 larvae per 100 plants could signifi
cantly reduce almost all characteristics examined and greatly accelerate

turnover. A seasonal aspect was implicated in the plant response as the
insects appeared to be much more effective in the fall than the summer.
This may be related to the energy budget of the plants under varying
conditions of solar flux. Plant density increased in the summer in
response to insect attack probably as a result of decreased intraspecifi
competition. A similar response would be expected to any factor which
reduced competition provided adequate energy for growth was available.
To learn if A. densa could be used in biological control a green
house reared population was released on a small pond in August 1974.
The waterhyacinth population was reduced and competing plants began to
dominate the site. Ultimately cat-tail invaded and waterhyacinth failed
to re-invade. A control site remained dominated by waterhyacinth.
Studies of natural A. densa populations on Lake Alice indicated
that the failure of this insect to achieve sufficient levels to have
an extensive effect on waterhyacinth was most likely due to the complex
of parasites which attack them. Also, pickerelweed (Pontederia cordata)
appears to be the preferred host of this insect which may partly explain
the low populations observed on waterhyacinth. Further, seasonal changes
in the plants' ability to withstand insect attack may obscure correla
tions between plant characteristics and insect damage.

INTRODUCTION
At the onset of this study various state and federal agencies were
preparing for the release of exotic insects for the biological control
of waterhyacinth (Eichhomia cvassipes (Mart.) Solms) in Florida and
the Southeastern U.S. In order to evaluate the effect of these insects
it was apparent that prior information on the ecology of waterhyacinth
at the release sites, particularly with regard to annual variability,
and the effects of indigenous insects would be needed. The purpose of
this dissertation was to provide part of this information for Lake
Alice, the primary study site in the Gainesville area.
To achieve this end the annual sequence of events in the water
hyacinth population was studied as well as the actual and potential
effects of natural and introduced populations of Arzama densa Wlk., a
native insect which feeds on waterhyacinth. This dissertation is or
ganized into five sections. The first section is a literature review
organized into two parts. The first part reviews the biology of water-
hyacinths and is organized in a manner similar to that suggest by
Cavers and Mulligan (1972). The second part reviews the taxonomy and
biology of A. densa Wlk. and related species.
The second section is a study of the waterhyacinth population on
Lake Alice. A fairly detailed description of the study site is provided.
The phenology of various morphometric features of the waterhyacinth
population is described with regard to the possible influence of various
physical and biological factors. A short study on the productivity of
waterhyacinth is also included in this section.
1

2
The third section investigates the potential effects of Arzama densa
Wlk. on greenhouse cultures of waterhyacinth. These effects are evaluated
in terms of various ecological and morphological traits of the plant.
The fourth section deals with the feasibility of augmenting natural
populations of A. densa as a means of biological control. The effects
of a small scale release are evaluated from a small pond near Paynes
Prairie.
This fifth section constitutes notes on the biology and life
history of A. densa. Data on the natural population at Lake Alice is
presented and the probable reasons for the failure of this insect to
control waterhyacinth is discussed.
It is hoped that the information reported here will provide a
baseline from which future comparisons can be drawn after the release
of exotic insects. It is also hoped that an increased understanding of
the ecology of waterhyacinth in a situation relatively free of specific
insect enemies has been gained.

LITERATURE REVIEW
Eichhomia crassipes (Mart.) Solms
Taxonomy
Bock (1966) provides an excellent historical review of the liter
ature dealing with the taxonomy of Eichhomia crassipes. The most
current treatment of the genus appears to be that of Agostini (1974)
which describes the species occurring in Venezuela. Five species are
described (E. azurea, E. crassipes3 E. diversifolia, E. heterosperma,
and E. paradoxa) and a key provided. The synonomy provided for E.
crassipes is as follows:
Eiohhomia crassipes (Mart.) Solms in DC., Monogr. Phan. 4:527.
1883.
Pontederia crassipes Mart., Nov. Gen. 1:9. t. 4. 1824.
Piaropus crassipes (Mart.) Raf., FI. Tell. 2: 81. 1837.
Eiohhomia speciosa Kunth, Enum. PI. 4: 131. 1843.
Eiohhomia cordifolia Gandog., Bull. Soc. Bot. France 66:
294. 1920.
While the bionomial E. crassipes is in common usage today the
synonyms Piaropus crassipes and E. speciosa are common in the literature.
Because of the world-wide distribution of this plant it is known
by a large variety of common names. Bock (1966) lists 48 common names
for Eiohhomia crassipes from 18 countries. The name waterhyacinth is
used world-wide in scientific reports but the structure of the word
has often been left up to the discretion of the user. It is often
written as two words (water hyacinth), a hyphenated word (water-hyacinth),
or as one word. Kelsey and Dayton (1942) in a list of standardized plant
names use the single word, waterhyacinth. This usage seems appropriate
since the plant is not related to thehyacinth as the two word name would
3

4
imply. For this reason I follow Kelsey and Dayton (1942) and use the
spelling waterhyacinth throughout this dissertation.
Waterhyacinth is a member of the pickerelweed family which includes
9 genera and 36 species (Cook et al. 1974). The genera are Eiohhomia
(7 spp.), Eurystemon (1 sp.), Heteranthera (10 spp.), Hydrothrix (1 sp.),
Monochoria (5 spp.), Pontederia (5 spp.), Reussia (4 spp.), Soholleropsis
(1 sp.), and Zosterella (2 spp.). The majority of the members of this
family.are confined to the Americas although two of the genera (Mono
choria and Soholleropsis) appear to be Old World endemics. Keys and
descriptions of the genera (world-wide) may be found in Cook et al.(1974).
Lowden (1973) revised the genus Pontederia and united Reussia with it.
Castellanos (1958) provides notes on the genus Pontederia in Brazil and
keys and descriptions of the Brazilian species of Pontederiaceae
(Castellanos 1959).
Description and Account of Variation
Following is a translation of the description of Eiohhomia orassipes
(Mart.) Solms given by Agostini (1974). Further descriptions can be
found in Castellanos (1959), Sculthorpe (1967), Bock (1966), Penfound
and Earle (1948), Buckman and Co. (1930), Webber (1897), and many others.
The first definite description of this species (according to Bock 1966)
was that of Kunth (1843).

5
Plants floating or sometimes fixed to the substrate,
the leaves in the form of a rosette with the stem
reduced and the plants connected by an elongated
horizontal rhizome; numerous plumose roots issue
from each plant. The aerial leaves are variable
in shape; petioles of 2 to 30 cm long are more or
less inflated; stipules 2-15 cm long with a small
apical orbicular-reniform lamina with a lacerate
[serrate ?] margin; submerged leaves never evident.
Inflorescense variable, internodes between the
spathes nearly absent; inferior spathe with lamina
1-5 cm long, the sheath 3.5-7 cm long. Flowers
4-6 cm long; perianth light purple or rarely white,
tube 1.5-2.0 cm long, lobes 2.5-4.5 cm long, with
entire margins. Stamens all exserted, filaments
villous-gladular. Capsule elliptical, trigonous,
12-15 mm long; seeds oblong-elliptical 1.2-1.5 x
0.5-0.6 mm with 10 longitudinal ridges. (Agostini 1974,P. 305)
The leaves of waterhyacinth are variable in the shape of the blade
and the extent of development of the float. This variability is appar
ently dueto the plants' responseto environmental conditions. It was
shown prior to 1930 that the size of the floats depends on such factors
as light, temperature, and water quality (La Garde 1930). Rao (1920b)
was convinced that increased water uptake (in a hypotonic medium)
promoted the development of the float. This is discussed further by
Bock (1966), Penfound and Earle (1948), and in an exceptionally good
account by Misra (1969).
The waterhyacinth flowers are possibly trimorphic with regard to
the length of the style. Although medium and long styled forms are known,
the existence of short styled forms is thought to be possible (Bock 1966).
The midstylous form is normally predominant and Bock (1966) feels that
only the mid and long styled forms exist. The flowers possess two whorls
of anthers which Bock (1966) notes are long and short in the mid-styled
form and short and mid-length in the long-styled form. This dimorphism
is apparently regulated by a two gene system, one exerting an epistatic

6
effect on the other, and each gene with two alleles (Bock 1966; Ornduff
1966).
Very few cytological studies have been done on waterhyacinth.
Banerjee (1974) found the chromosome number to be 2n = 32 in India,
illustrated kavyotypes, and described the chromosomes. She found the
chromosome number to be very consistent but noted variants of 2n = 30
and 58. Bock (1966) also reported the diploid chromosome number to
be 32 and noted that this had been reported by earlier authors as
the probable number.
Economic Importance
Waterhyacinth is ranked world-wide as among the top 10 most im
portant weeds and as the single most important aquatic weed (Holm 1969).
Because of its floating habit and high productivity (Bock 1969) it com
petes with man for open water. Large build-ups interfere with hydro
electric operations in many areas (Holm et al. 1969; Rushing 1974). Its
ability to interfere with navigation is well documented (Gay 1960; Evans
1963; Holm 1969; Webber 1897; Curtis 1900; Zeiger 1962). In the Panama
Canal mats of waterhyacinth have become so thick as to interfere with
the opening and closing of the locks (Pasco, pers. comm.). Gusio et al.
(1965) cited a study in which the efficiency of canals in the Everglades
were reduced 40-80% by large infestations of this plant. Irrigation
operations are affected by the impediment of water flow and the clogging
of pumps.
Waterhyacinths affect agriculture not only indirectly, as in irri
gation, but also directly. Sugar and rice are cultivated in "flood-fallow"

7
situations where the land is flooded for several months. Aquatic weeds
compete with the crops for the open surface, increase evaporation, and
may provide reservoirs of crop pathogens (Nat. Sci. Res. Coun. of Guyana
and N.A.S., 1973).
Water losses through evapotranspiration by waterhyacinth can be
considerable. Timmons (1960) showed that 17 states lost nearly 2 million
acre-feet of irrigation water annually due to aquatic and ditch bank
weeds. Holm et al. (1969) quoted an estimated value of this lost water
as over $39 million. Rates of evaportranspiration by waterhyacinth are
reported by Penfound and Earle (1948), Timmer and Weldon (1967), Misra
(1969), Brezny et al, (1973), and Van der Weert and Kamerling (1974).
These reports conflict, however, as experimental technique appears to
be a large source of error in these studies. The ranges of the ratio
of transpiration to open water evaporation are 1.66-6.6 (Penfound and
Earle 1948), 3.7 (Timmer and Weldon 1967), 5.78-9.84 (Misra 1969),
1.02-1.36 (Brezny et al. 1973), and 1.20-1.58 (Van der Weert and
Kamerling 1974).
Large mats of waterhyacinth covering the surface of water bodies
block light to phytoplankton and submersed vegetation. This effectively
prevents the liberation of oxygen through the photosynthetic processes
of these organisms. Further, surface diffusion of oxygen is lowered as
mixing is inhibited at the air-water interface. This results in a severe
reduction of dissolved oxygen (Ultsch 1973). This renders the habitat
unsuitable or lethal to many desirable species of fish. McVea and Boyd
(1975) demonstrated that extensive pond coverage by waterhyacinth
reduces phytoplankton growth and fish production. The composition of

8
the aquatic food chain may change from a plant-herbivore based community
to a detritus-detritivore based community (e.g., Hansen et al. 1971) as
a result of this loss of submersed primary productivity.
Waterhyacinths may also successfully compete with valuable wildlife
forage thereby replacing it. This may destroy feeding areas for waterfowl
(Gowanloch 1944). Local economies may be seriously damaged in areas which
cater to recreational needs such as waterfowl hunting, fishing, boating,
waterskiing, swimming, etc.
More seriously, riverine communities in the developing areas of
the world which depend on fishing as a primary source of protein may
be denied access to fishing grounds (Holm 1969). Holm (1969) further
stated that impoundments for fish culturing may be destroyed by large
masses of floating waterhyacinth. He stated that waterhyacinths con
stitute "...the most massive, most terrible and frightening weed
problem" he had ever known.
Waterhyacinth possible pose a health threat by harboring vectors
and intermediate hosts of human diseases. The larvae and pupae of Mansonia
uniformis (Theob.), a mosquito vector of filariasis in Asia, are known
to attach to the roots of waterhyacinths (Burton 1960; McDonald 1970).
Waterhyacinths may result in an increased production of mosquitoes by
hindering insecticide application, interfering with predators, increasing
the habitat available for certain species which attach to the plant, and
by impeding runoff and water circulation thereby creating stagnant
impoundments for breeding (Seabrook 1962). Mulrennan (1962) cautions
that uncontrolled aquatic plant populations could lead to an increased
incidence of mosquito-borne diseases such as malaria, encephalitis, and

9
other arboviruses. Mitchell (1974) indicated that aquatic plants,
including waterhyacinths, may also harbor snails which are intermediate
hosts of diseases such as fascioliasis and shistosomiasis. He mentioned
that those species which do occur on waterhyacinth are assisted in
their dispersal by the free-floating habit of this plant thereby spreading
the associated diseases. Bock (1966) further reviews the literature
dealing with health hazards caused by waterhyacinth.
It is difficult to arrive at sound figures regarding the monetary
losses caused by waterhyacinth. Spencer (1973, 1974) quoted the following
figures from a congressional report for losses prevented by waterhyacinth
control in Louisiana in 1957.
Navigation
Flood Control
Drainage
Agriculture
Fish & Wildlife
Public Health
14,727,000
250,000
$37,993,000
Penfound and Earle (1948) conservatively estimated that water-
hyacinths were responsible for losses of $5 million annually as of 1948
in Louisiana. Spencer (1973, 1974) quoted figures from a Louisiana
Fisheries and Wildlife report indicating losses of $65-75 million in 1947
in Louisiana due to aquatic weeds. In 1930 a report to the Jacksonville
City Commission estimated that in the period from 1900-1930 the Federal
Government spent $233,000 on waterhyacinth removal in the Jacksonville
District to simply maintain "open channels for navigation" (Buckman and
Co. 1930). By 1961 the total cost had risen to $1,861,788 in the same
district (Tabita and Woods 1962). Wunderlich (1964) reported costs of
clearing aquatic weeds ranging from $15-60 per acre. As of 1964 about

90,000 of Florida's 2,500,000 acres of fresh water and 70,000 to 100,000
acres of Louisiana's 2,000,000 acres of fresh water are covered with
waterhyacinth (Ingersoll 1974). Hudson (pers. comm.) estimates that in
1975 the acreage of waterhyacinth in Florida has extended to more than
200,000 acres and the average cost of control per acre is about $25. He
estimated that all agencies within the state in FY 1976 allocated $16
million for aquatic weed control, about 30% of which ($4.8 million)
goes towards waterhyacinth control. This is an increase of almost $2
million over the previous year (FY 1975) for waterhyacinth control
alone.
Thompson (pers comm.) indicated that between 1965 and 1974 the
U.S. Army Corps of Engineers spent $6.1 million in combined construction
and operations funds for aquatic weed control in Louisiana alone. In
the period between 1960 and 1964 the estimated cost was $1.7 million.
Other weeds are of minor concern and for the most part 100% of this
went toward waterhyacinth and alligator weed control. The State of
Louisiana beginning its program in the mid 1940's spent $8.1 million
as of 1973. Further costs included $1 million in 1974 and $1.1 million
in 1975. Thompson further estimated that the average cost of treating
an acre is between $32 and $35 in Louisiana. The most economical means
being by helicopter ($13/acre) or fixed wing aircraft ($10/acre) when
possible. The current estimate of acreage covered in Louisiana exceeds
1 million acres. This does not necessarily reflect an increase in acreage
over Ingersoll's (1974) figure but is merely a more accurate estimate.
It is evident from these figures that the acreage covered with
the plant is increasing while the cost of treating an acre is also

11
increasing. The result of this is a geometrically increasing trend in
the overall cost of aquatic weed control by traditional means.
Because of the seriousness of waterhyacinth infestations the
beneficial aspects of this plant have been largely overlooked or ignored.
Fringes of aquatic plants along rivers or lakes are often helpful in
absorbing wind and wave action and preventing bank erosion (Tilghman
1963). Caldwell (1942) notes that the roots of waterhyacinth provide
excellent cover for goldfish spawn. He promotes the growing of this
plant for ornamental purposes stating that it is the "Biggest bargain
in a pool plant . and dismisses its detrimental attributes as an
" . attractive nuisance." Waterhyacinth was originally imported into
this country for use as an ornamental (Raynes 1964) and the beautiful
flower does give it a certain aesthetic appeal.
Tilghman (1962, 1963) spent many years fishing the St. Johns River
and guiding fishing tours. He was vehement about the beneficial effects
of waterhyacinth on fish propagation noting that the plant roots provide
cover for spawn and support macro-invertebrates which are preyed upon
by fish. Tilghman also noted that the plants helped clean the water
thus improving the fish habitat.
Abu-Gideiri and Yousif (1974) studied the influence of water
hyacinth on planktonic development in the White Nile. They compared
plankton populations and water chemistry parameters at a site south
of Jebel Aulia Dam to similar studies done prior to 1958 before in
vasion of the area by waterhyacinth occurred. They found that overall
planktonic densities had increased in the interim as a result of changes
in the water quality (such as an increase in phosphates). They attributed

this change to the weed infestation but failed to consider cultural
changes in the area. They concluded that the weed growth provides
improved conditions for planktonic development and thus benefits fish
production. The basis for these conclusions is obscure, however, and
doesn't agree with the findings of other authors (e^g. Wahlquist 1969b,
McVea and Boyd 1975). While it seems probable that a fringe of the
weed would be beneficial in some areas Bose (1945) commented ". . the
various reports of fish mortalities in stagnant pools and ponds covered
with waterhyacinth at once dispell the ideas and ruin the prospect that
waterhyacinth should ever be fancied in tropical countries as 'one of
the popular plants' for any kind of pisciculture."
Increasing attention has been directed towards using waterhyacinth.
Pirie (1960) advocated utilizing waterhyacinth as a crop. Waterhyacinths
are currently being considered for tertiary treatment of sewage effluent
(see Dymond 1948; Sheffield 1967; Yount 1964; Yount and Crossman 1970;
Boyd 1970b; Rogers 1971; Rogers and Davis 1971; Dunigan et al. 1975;
Dunigan and Phelan 1975). Its possible use as feed for livestock
(Chatterjee and Hye 1938; Baldwin et al. 1975; Bagnall et al. 1973, 1974;
Baldwin 1973; Boyd 1968a, b; Combs 1970; Hentges 1970; and Salveson 1971)
Apparently waterhyacinth is used as pig fodder in Singapore (Anonymous
1951). A complete cycle is developed when waterhyacinth is fed to pigs,
wastes and fecal matter are washed from the piggeries into the pond,
this fertilizes the pond to produce fish and more waterhyacinth which
are both harvested.
Chatterjee and Hye (1938) found that waterhyacinth was high in
potash with as much as 68% in the ash (5% on dry weight), comparable

13
with other fodders in nitrogen content (0.97 to 2.57% D.W.), rich in
chlorine (3-4% D.W.), and richer than Napier and Guinea grass in lime
(3.5% D.W.) and magnesia (0.96% D.W.). They also noted that i.ts phosphate
content was low (0.36% D.W.) but that the digestible nutrients compared
well with other fodders. Taylor and Robbins (1968) analyzed the com
position of waterhyacinth and found the leaves to contain 15.8% dry
matter which in turn was composed of 14.7% ash, 1.7% nitrogen, 10.7%
crude protein and 17% crude fiber. The whole plants were 8.9% dry
matter, 1.5% nitrogen and 9.6% crude protein. They also analyzed the
plants for the amino acid composition. They concluded that the lysine
content of waterhyacinth was sufficient to serve as an effective grain
protein supplement.
Boyd (1968a) determined that waterhyacinth contained 12-18%
(D.W.) crude protein. He subsequently fully analyzed the nutritive
value of waterhyacinth and found the dry weight to be 5.9%, the crude
protein to be 0.94% of the fresh weight (ca. 16% D.W.), cellulose ca.
28% (D.W.) total available carbohydrate 7.8%, ash 17%, and caloric
t
content ca. 3.8 kcal/g. He further analyzed the inorganic nutrient
content and the amino acid composition. Taylor et al. (1971) extracted
protein from waterhyacinth and found that the percentage on a dry
weight basis varied between 7.4 to 18.1%. They also analyzed the
protein for the amino acid composition.
Knipling et al. (1970) compared the nutrient content of water
hyacinth from two different sites. They performed comparative analyses
of various plant parts from the two sites for nitrogen, phosphorus,
calcium, potassium, and magnesium as well as chlorophyll and water

14
content. They determined that the nutrient content in the plant tissues
was not proportional to that of the water in which they were grown. Boyd
(1974) has summarized the data on the composition of waterhyacinth and
other aquatic plants.
Liang and Lovell (1971) evaluated waterhyacinth for use in channel
catfish feed. They found that the addition of 5 to 10% waterhyacinth in
vitamin free diets increased growth and reduced mortality in the fingerlings.
Bagnall et al. (1974) using waterhyacinth as feed supplements for
cattle and sheep, for paper production, and for mulch determined that
its processing as mulch was the most economically feasible use.
Azam (1941) proposed that underdeveloped countries encourage their
people to utilize their spare time preparing various products made from
waterhyacinth and thus supplement their income. Some of the products
they suggested were paper, pressed board and tiles, detergents, cattle
fodder, and manure. Nolan and Kirmse (1974) considered waterhyacinth
jsable in the production of paper.
Iswaren and Sen (1973) found that an extract from waterhyacinth
roots increased the yield of Brinjal (Solomon melangena var. Pusa
Kranti) from 507.2 g/plant to 1317.3 g/plant. Ganguly and Sircar (1964)
found that a root extract from waterhyacinth increased the metabolic
activity and the nitrogen and sugar content of Vision sativum L. seedlings.
Mukherjee et al. (1964) identified growth promoting substances in the
roots of waterhyacinth which they believed to be bound auxins. Sheikh
et al. (1964) noted that this extract was thermostable and found it to
promote the growth of Phaseolus mungo var. roxburghii, the mycelium of
Aspergillus niger, the growth of Rhizopus, and the multiplication of

15
yeast cells thereby promoting fermentation. S.M. Sircar and Chakravarty
(1961) found that this root extract increased the yield of jute (Corehorus
eapsularis L.) and the production of fiber. P.K. Sircar et al. (1973)
identified four gibberelin-1 ike compounds from extracts of waterhyacinth
shoots.
Other attempts to find marketable products of waterhyacinth include
ink, upholstery stuffing, rope, bags, plastics, timber substitutes, and
ice chests (Bock 1966). Morton (1962) gives a recipe for preparing water-
hyacinths for food. For further reviews of the literature dealing with
the utilization of waterhyacinth and other aquatic weeds see Bock (1966),
Sculthorpe (1967), Little ( 1968b ), Boyd (1972, 1974) and Mitchell
(1974).
Distribution
It is generally accepted that South America is the area of origin
of the waterhyacinth. Small (1936) indicates that it was originally
discovered in the San Francisco River near Malhada, Brazil in 1824.
Bock (1966) cites Hooker (1829) as listing several early collections
from Brazil; Demerara River, Guiana; New Granada ( a former Spanish
viceroyality including present day Venezuela, Ecuador, Colombia, and
Panama); Guayaquil, Ecuador; and Buenos Aires, Argentina. She also
cites early references indicating it was also native to the West Indies
(eg. Schwartz 1928; Britton 1918). Castellanos (1959) lists its present
South American distribution as Argentina, Paraguay, Brazil, Uruguay,
Chile, Ecuador, Colombia, and Guyana. It has also been reported in
Surinam (Little 1965, 1966; Holm et al. 1969). It has apparently been
widespread in South America for many years as evidenced by the early
records.

16
Misra (1969) cited a source which indicates that the center of
origin for this species was probably the Pernambuco region of Brazil.
A few authors have subscribed to other regions of origin outside of South
America. Hildebrand (1946, p.477) states "The water hyacinth, Eichhornia
orassipes, is a native of Japan and was carried about 70 years ago to
South America, where it became widespread in fresh-water streams and
lakes." He cites Gowanloch (1944) as the authority for this statement.
Gowanloch apparently contradicts himself, however. In one paragraph
he does indicate that waterhyacinth is native to Japan and was imported
to South America. In the following paragraph he states, "When in 1884
an International Cotton Exposition was held in New Orleans, the Japanese
Government representatives in their building on the Esposition grounds
gave away as souveniers water hyacinths which they had imported from
Venezuela." Small (1933) suggested that it may be native to Florida.
A Ceylonese author claimed that Florida was its area of origin (Bock 1966).
Waterhyacinth is well known from the West Indies (Bock 1966).
Castellanos (1959) includes the Antilles within the range of distribution.
Bancroft (1913) also indicated that the plant was present in the West
Indies. Bock (1966) cites a paper which indicates that Puerto Rico is
the center of dispersal for this species. She also found it naturalized
in Jamaica. She suggests that it may have spread to the islands attached
to boats or by floating from the mainland.
As might be expected waterhyacinth is also known from most of the
Central American countries including Panama (Standley 1928; Hearne 1966),
Costa Rica (Little 1965), Nicaragua (Little 1965, 1966; Holm et al. 1969),
Honduras (Castellanos 1959), and El Salvador (Little 1965, 1966; Holm et al.

17
1969). I have not found any records of waterhyacinth occurring in Guatemala
but its range does extend into Mexico (Castellanos 1959; Little 1965)
where it is apparently well distributed.
As previously mentioned waterhyacinth was thought to have been
introduced into the United States from Venezuela in 1884 at the
International Cotton Exhibition in New Orleans by the Japanese delegation
(Klorer 1909; Buckman and Co. 1930; Gowanloch 1944; Hildebrand 1946;
Dymond 1948; Penfound and Earle 1948; Tabita and Woods 1962; Dutton 1964;
Wunderlich 1964; Bock 1966). Some accounts indicate, however, that the
plant may have been in the United States in the 1860's (Tabita and Woods
1962) or prior to the Civil War (Penfound and Earle 1948). The accepted
theory maintains that the plant was given away as souvenirs at the
New Orleans Cotton Exposition (Gowanloch 1944). The plants were taken
for ornamental purposes (Klorer 1909; Dutton 1964; Wunderlich 1964) or
for purposes of cultivating them for cattle fodder (Wunderlich 1964). In
any case, it was felt that when the plants outgrew the limited amount of
space given them they were cast out into natural bodies of water (Klorer
1909). By 1888 it was in the coastal fresh waters of Texas, Louisiana,
Mississippi, and Alabama (Buckman and Co. 1930).
It was apparently introduced into Florida in 1890 (Webber 1897;
McLean 1922; Barber and Hayne 1925; La Garde 1930; Buckman and Co. 1930;
Penfound and Earle 1948). Raynes (1964) reported it was first introduced
in the St. Johns River at Edgewater about 4 miles above Palatka. Mr. J.E.
Lucas was interviewed by a New York Sun reporter (Anonymous 1896) in 1896
and gave the following account:
"I know the man who brought the first plant to Florida," Mr. Lucas

18
said to a Sun reporter, "and he thought that he did the State a favor.
I have it from his own lips, and I've known him since long before that
time, for I used to carry him up the river in a launch year after year
to his orange grove. He was Mr. Fuller, father of W.F. Fuller of
Brooklyn, owner of Edgewater Grove, a property which he bought and
improved, until now it is a beautiful place, seven miles above Palatka.
Five years ago there wasn't a water hyacinth in the St. John's River,
nor in the state, so far as I know. One season Mr. Fuller brought some
there and put them in a pond on his premises. I understand that he brought
them from Europe. They added very much to the beauty of the place, and
they thrived so that he took some and threw them into the river. There
they grew and blossomed abundantly, and they were greatly admired, and
Mr. Fuller said to me one day: "The people of Florida ought to thank me
for putting these plants here."
"But presently those in his pond had spread so that they covered
it over. Then he cleared them all out. But it was too late to stop them
from spreading all over the river. They worked their way and were blown
up and down it for miles, and into the bayous, and finally up the
Acklawaha [sic]. Two years ago they had become a serious menace to navi
gation, and protest after protest was sent to the Government. At last
the War Department sent an agent to investigate, but he got to us just
after the visitation of that heavy frost of two years ago, which killed
all our orange trees. The hyacinths were killed too, apparently, and so
the agent reported that nature had cleared the rivers and that there was
nothing requiring the department's attention. But the plants were only

19
dead at the top. They grew again, and the startling conditions that you
see in these pictures are a growth of only two years."
This account surprisingly indicates that the plants were introduced
into Florida from Europe rather than from Louisiana as has generally been
assumed (Buckman and Co. 1930; Tabita and Woods 1962; Wunderlich 1964).
Waterhyacinth was first discovered in Georgia in 1902 (Harper 1903)
about one mile north of Valdosta. The first record in California is
from 1904 near Clarksburg, Yolo Co. (Bock 1968). Johnson (1920) reported
it in Fresno Co., California. Bock (1968) lists its present range in
California from 10 mi NW Sacramento (ca. 38.5 N Lat.)to Ramona, San
Diego Co. (ca. 33 N Lat.). She speculates that it was probably brought
to California as an ornamental and released. The primary rivers infested
are in Central California and include the Kings, Tuolumne, San Joaquim,
and Sacramento River Systems.
The infestation of waterhyacinth in California is discontinuous
with the North American range of this weed. Penfound and Earle (1948)
stated that shortly after the turn of the century it had been reported
from all the southeastern coastal states as far north as Virginia. A
distribution map published by the U.S.D.A. (1970) indicates that the
present range of this plant in the U.S. includes the Potomac River in
Maryland-Virginia, west to southern Missouri, south to eastern Texas
and southern Florida, and separately, central California.
Just when waterhyacinth spread to the Old World is not certain.
Agarwal (1974) indicated it may have been introduced into India around
1896. McLean (1922) cited testimony indicating that it may have been
present in Bengal as early as 1898 or 1899. It was apparently introduced

20
to the Buitenzorg Botanic Gardens in Java in 1894 (Bock 1966; Sculthorpe
1967). By 1905 it had appeared in Ceylon (Jepson 1933). It had become a
serious problem in Cochin China (a part of S. Vietnam) by 1908, in Burma
by 1913 and in Bengal by 1914 (McLean 1922). The Philippines acquired
the plant around 1912 and by 1926 it was appearing in China as well as
Borneo and Malaysia (Bock 1966). Records as to its entry in Japan are
scarce but they were apparently introduced during this century as
ornamentals (Shibata et al. 1965; Bock 1966). The first record of its
occurance in Okinawa was in 1952 (Bock 1966) but it may have been there
earlier.
From these beginnings it now occurs throughout India, Southeast
Asia, and Indonesia (Bancroft 1913; Barber and Hayne 1925; Jepson 1933;
Penfound and Earle 1948; Anonymous 1951; Sen 1961; Little 1965, 1968a;
Holm et al. 1969; Chhibbar and Singh 1971; Haigh 1936; Hitchcock et al.
1949; McLean 1922; Agarwal 1974; Robertson and Thein 1932; Shibata et al.
1965; Gangstad et al. 1972; Ishaque 1952).
Waterhyacinth was first introduced into Australia in Queensland
in 1895 (McLean 1922) and in New South Wales in 1896 (Maiden et al.
1906). It was apparently eradicated from New South Wales but a reinfest
ation occurred in the 1940's (Parsons 1963; Bill 1969), to South
Australia by 1937 (Bill 1969) and to Victoria by 1939 (Parsons 1963).
Bill (1969) notes that waterhyacinth is not a serious problem in
Australia today except in some Queensland rivers.
Waterhyacinth also occurs in New Zealand (Walker 1954; Taylor 1955;
Anonymous 1964b; Little 1965,1968a; Holm et al. 1969) although it is
difficult to determine when it first appeared there. Taylor (1955)

21
seems to imply that it was discovered in 1949 at least in the Rotorua
District. Walker (1954) reported it from the opposite side of North
Island near Shannon. Matthews (1967) stated that there were 2 areas of
infestation in 1948-50, 15 after 1950, and 70 by 1956. Another report
(Anonymous 1964) indicated that there were at least 60 known infest
ations in New Zealand ranging from Opoua in the north to Shannon in
the south. Manson and Manson (1958) noted its occurrence as far north
as Kaitaica.
The spread of this plant has also taken in some of the Pacific
Islands. It was reported from Hawaii in 1946 (Bock 1966) and Mune and
Parham (1954) indicated that it was recognized as a pest in Fiji.
In Africa the plant is known from Kenya (Anonymous 1957), Zaire
(Anonymous 1957; Lebrun 1958; Kirkpatrick 1958; Coste 1958; Berg 1959;
Little 1965,1968a; Holm et al. 1969), Tanzania (Anonymous 1957; Little
1968); Uganda (Anonymous 1957), Angola (Lebrun 1958; Mendonca 1958),
French Equatorial Africa (Lebrun 1958), Rhodesia (Lebrun 1958; Little
1968; Holm et al. 1969) Malawi (Lebrun 1958); Monzambique (Lebrun 1958;
Mendonca 1958), South Africa (DuToit 1938; Penfound and Earle 1948;
Lebrun 1958; Holm et al. 1969), Madagascar (Lebrun 1958), Sudan (Gay
1958, 1960; Davies 1959; Pettet 1964; Little 1965, 1966,1968a; Chadwick
and Obeid 1966; Holm et al. 1969; Abu-Gideriri and Yousif 1974; Tag el
Seed and Obeid 1975; Mohamed and Bebawi 1975), Senegal (Anonymous 1964;
Little 1965; Holm et al. 1969) and Egypt (Little 1965; Holm et al.1969).
Waterhyacinth was first introduced into Africa either in South
Africa or Egypt. Sculthorpe (1967) cited a work on Egyptian flora which
indicated that it made its appearance in Egypt in the period between

22
1879-1892. Bock (1966), however, indicated that it was not introduced
into Egypt until 1912. It was introduced into South Africa around 1910
as an ornamental and by 1938 was reported from rivers in the Cape Pen
insula, George, Knysna, Albany, Port Elizabeth, Uitenhage, Victoria
East, and Natal (DuToit 1938). It had apparently reached South Rhodesia
prior to 1937 as Europeans settling there reported its presence at that
time (Holm et al. 1969; Bock 1966).
By 1942 waterhyacinth had spread into Mozambique in the Incomati
estuary from Vila Luisa to Xanowano and apparently originating from the
Transvaal of South Africa (Mendonca 1958). Kirkpatrick (1958) indicated
that waterhyacinth was already present in Zaire (the Belgian Congo) in
the Congo River in 1954. Coste (1958) felt that it was introduced in
the period between 1950 to 1951. Other authors (Bock 1966; Holm et al.
1969) list 1952 for its introduction into the Congo. By 1955 it had
spread over 1600 km of the river between Leopoldville and Stanleyville
(Kirpatrick 1958). Gay (1958) first observed waterhyacinths occurring
on the White Nile of the Sudan in 1958 along about 1000 km. It was
apparently not abundant in the river prior to 1957 although it may have
been present in 1956.
Senegal first reported the presence of waterhyacinth in 1964
(Anonymous 1964) from the Cape Vert peninsula and this may perhaps be
the first record in the northwestern part of Africa. In spite of the
warnings expressed by the Inter-African Phytosanitary Commission it
was still available for purchase from street hawkers in Senegal in 1965.
Waterhyacinth is now distributed in all of the tropical and sub
tropical areas of the world. Its northernmost limits of distribution
are probably near Sacramento, California (Lat. 38.5 N; Bock 1968),

23
the Potomac River near Washington, D.C. (ca. 30N; Gowanloch and Bajkov
1948; U.S.D.A. 1970), Japan (30-35N; Holm aval. 1969) and possibly
Portugal (37-40N) as indicated by Holm et al. (1969) on their distri
bution map. The southern most limits of distribution appear to be Buenos
Aires, Argentina (34S) and Concepcion, Chile (37S) in South America
(Castellanos 1959), and Shannon, New Zealand (40-41S; Anonymous 1964).
The range in general seems to be bounded by the 40 North and South
Latitude lines. Very little information is available on the altitudinal
restrictions of this species although one paper (Anonymous 1957) states
that it is limited in the tropics to an altitudinal zone of from sea
level to 4500 feet (ca. 1400 m).
Habitat
Little is known of the ranges of environmental tolerances of
waterhyacinth. Webber (1897) noted the effects of freezing temperatures
in Florida in the winter of 1894-95. He noted that the first freeze
killed the top which caused the plant to float higher in the water.
A second freeze killed this newly exposed portion. Most of the plants
survived by resprouting from the unexposed portion of the rhizome.
Buckman and Co. (1930) stated that temperatures as low as 28F
(-2.8C) may be withstood by the roots but will kill the tops. Temper
atures lower than this will kill the roots as well. Hitchcock et al.
(1950) observed waterhyacinths subjected to two days of freezing in
New York. The plants were ice-covered when transferred to the green
house. Damage was apparently severe as the authors noted that all the
foliage and all the roots were killed. Within 13 days the plants had
recovered by resprouting from the rhizome tip. Misra (1969) placed

24
plants in a deepfreeze at -10C for 7-8 hours and they failed to revive.
He found that at 15C the growth became restricted and the plant did
not show any increase in growth up to a 90-day period.
Penfound and Earle (1948) exposed small plants in trays with 3
inches of water to various air temperatures for various durations. They
found that at 33F and 27F (0.56C and -2.78C) all of the rhizomes
resprouted when returned to room temperatures after being exposed for
12, 24, and 48 hours. At 23F (-5C) all resprouted after 12 and 24
hours but none survived after 48 hours. At 21F (-6.11C) some survived
12 hour exposures but none survived 24 or 48 hour exposures. At 19F
(-7.22C) none resprouted after being exposed for 12, 24, or 48 hours.
They concluded that the temperature effect depends upon the duration
of exposure and that freezing of the rhizome tip results in the
destruction of the plant.
Hitchcock et al. (1949) found that satisfactory growth
occurred in air temperatures of 21-27C. Silveira-Gui.do et al. (1965)
stated that the plants grew well in water temperatures ranging from
17-35C. Bock (1966) measured air temperature ranges of 17-35C and
water temperature ranges of 18.6-21.5C during the waterhyacinth
growing season and winter mid-day temperatures of 5-10C for air and
5.8C for water in California. She also stated that the populations
survived the winter of 1963-4 from which she monitored 28 da with
air temperatures below freezing, 1964-65 with 25 day, and 1965-6
with 35 da although considerable mortality did occur.
Knipling et al. (1970) measured waterhyacinth growth along a
gradient of water temperatures. They found the optimum to be 28-30C

25
although relatively high growth occurred over the range of 22-35C.
Exposure to 10C nights reduced the amount of photosynthesis on
following warm days.
Bock (1966, 1968) exposed plants to 26.7C-26.7C, 26.7C-
4.4C, and 4.4C-4.4C day-night temperatures under both 16-18 and
8-16 L:D photoperiods. She found that growth was favored in the higher
temperatures although it also appeared to be favored in the shortened
photoperiod.
The maximum tolerable water temperatures appear to be around
33-34C. Penfound and Earle (1948) observed that the plants cannot
tolerate water temperatures above 34C. Misra (1969) stated that in
India the plants succumb at water temperatures above 33C. Knipling
et al. (1970) found that growth began to be inhibited at about 33C
and declined in a nearly linear manner at higher temperatures until,
by 40C, negative growth was indicated. They noted that the plants
were more tolerant of lower than optimum temperatures than of higher
than optimum.
Light relations have been investigated by a few authors. Penfound
and Earle (1948) noted that in July the average light intensity was
about 420 foot-candles above colonies of moderate-sized plants. Under
the canopy of large plants the light intensity was about 170 ft-c repre
senting a 60% decrease. They found that equitant elongate leaves are
formed at intensities ranging from 130-500 ft-c and float leaves are
formed at intensities over 500 ft-c. Under a walkway where the light
intensity was 130 ft-c (31% of the July average) most of the plants
were found to be dying. They also placed containers of waterhyacinth

26
under a table where the light intensity averaged 55 ft-c and all of
the plants died in 2 mo. In connection with this they placed several
plants in the dark and measured the starch depletion. By 7 da the
starch content was reduced by 50% and by 12 da it was completely gone.
Hitchcock et al. (1949) grew plants in a greenhouse and supplied
one group with supplemental heat, one group with supplemental light,
and one group was left as a check. They found that the no. leaves per
plant, the average leaf length, and the no. flowers produced were greatest
in the high light condition.
In Africa (Anonymous 1957) it has been noted that light is seldom
a limiting factor with respect to vegetation and frutification but it
may have a more direct influence on germination.
As previously mentioned one study (Bock 1966, 1968) found that
plants grown under the same temperature ranges grew better under the
shorter photoperiod. This peculiarity was not explained.
Bock (1966) stated that waterhyacinth needed 60% full sunlight
or better although she failed to define full sunlight. She placed
plants under greenhouse benches when the light intensity at noon was
30-40% full sunlight. These were retained there from September to March
and 67% mortality was observed.
Misra (1969) subjected plants to 40%, 70%, and 100% full sunlight
(again undefined) and found that the no. leaves per plant and the per
centage leaves with floats increased with increasing light intensity.
Correspondingly a reduction in the average volume and diameter of the
float occurred as the light intensity decreased.

27
Knipling et al. (1970) measured net productivity of attached
leaves under a range of light conditions. Photosynthesis increased from
7.8 mg C02/dm2 leaf surface/hr to 16.1 mg/dm2/hr as the light intensity
increased from 1450 ft-c to 8000 ft-c. Dark respiration was found to
range from 2.6 to 2.8 (average 2.7) mg/dm/hr.
Waterhyacinth is generally considered to tolerate a wide range of
pH (Pieterse 1974). Haller and Sutton (1973) found they grew over a
range of 4.0 to 10.0 although optimal growth occurred in acid to slightly
alkaline conditions (4-8). Bock (1966) citing data from other authors
concluded that waterhyacinth generally occurs in waters ranging in pH
from 4 to 9. Chadwick and Obeid (1966) compared the growth of water
hyacinth and water lettuce (Pistia stratiodes L.) in cultures of varying
pH. They found that waterhyacinth would grow at all levels (pH 3.0 to
8.2) but at 3.0 both dry-weight yield and offset production were minimal.
They felt that pH values near 7.0 were optimal for waterhyacinth but
values near 4.0 were optimal for water lettuce. Penfound and Earle (1948)
reported pH values usually ranged between 6.2-6.8 in or near waterhyacinth
mats in Louisiana but could survive extremes of 4-5 and 9-10. In the
Guinean region of Africa pH is thought to be limiting at values of 4.2
or below (Berg 1959; Anonymous 1957).
Minschall and Scarth (1952) studied the effects of low ranges of
pH (3.5-6.5) on the roots of waterhyacinth. They found that at values
below 4.0 the roots exhibited decreased cell division and cell elongation.
Cell division at pH 5.0 proceeded twice as fast as at 3.5. They further
found that the plants could tolerate more acidity at cooler temperatures
and the pH of the cell sap was always above that of the culture medium.

28
A few authors have suggested that stands of waterhyacinth may
modify the pH of the water. Penfound and Earle (1948) noted that pond
waters in the Mississippi River delta have an average pH of 7.2 whereas
water in waterhyacinth mats are usually acid. Ultsch (1973) compared
open water areas of a pond with areas covered with waterhyacinth and
determined the yearly average pH to be 5.6 in the open areas and 5.4
in the areas with waterhyacinth. Haller and Sutton (1973) presented
data which indicated that the plants cause a change in the direction of
neutrality from both high and low initial pH values. Center and Balciunas
(1975) compared water quality parameters from sites with and without
waterhyacinth and found that those with the plants had lower pH (7.06
0.84) than those without (7.551.06) although the difference was not
significant.
Moisture requirements of waterhyacinth and the effects of dessication
upon its survival and growth have been only superficially examined.
Webber (1897) noted that if the plants are to succeed a soil of loose
texture thoroughly saturated with water is required. Parija (1934), however,
found that they could survive 5.7% of water saturation in soil. Bock (1966)
noted one instance when the plants survived 41 da in saturated soil.
She speculated that the plants can withstand periods of dessication
because excessive transpiration is prevented from the center of the
rosette by the protective layer of dead outer leaves. Penfound and Earle
(1948) found that waterhyacinth could survive drying periods up to 18 da
depending upon climatic conditions and the surface they are exposed on.
Sunny weather with the plants on galvanized metal killed the plants
rapidly while rainy and cloudy weather or placing the plants in the

29
shade allowed them to survive longer. Misra (1969) found that when the
rhizomes are air dried they progressively lose the ability to resprout
as the moisture content decreases. They can tolerate a lower moisture
content when dried in mud, however, than when dried in air. This may
enable them to survive droughts in some instance.
As far as I have been able to determine Bock (1966) is the only
one to have investigated the effects of humidity on the growth of
waterhyacinth. She grew plants in a growth chamber both inside a plastic
enclosure with high humidity and outside the enclosure. She concluded
that high humidity favors growth.
With the recent interest in the utilization of waterhyacinth for
nutrient removal in sewage effluent increasing attention has been directed
towards the nutrient requirements of this plant. Dymond (1948) found that
the plants grew in both nutrient-rich and nutrient-poor water and that
the nutrient content of the plant was higher in nutrient-rich water.
Hitchcock et al. (1949) concluded that waterhyacinths have relatively
low nutrient requirements as good growth occurred in solutions 0.01 to
0.001 times as strong as normal in water cultures. They also found that
the growth response increased with added nutrients.
In Africa (Anonymous 1957) it has been noted that the lower limit
of "mineralization" is very low but little is known of the upper limits.
Chadwick and Obeid (1966) found that an increase in nitrogen levels
caused a linear increase in the total yield and plant number but had
little effect on the mean weight per plant. Knipling et al. (1970)
studied two sites with notably different levels of orthophosphate and
were surprised to find that the average standing crop yields were

30
similar at both sites. They further found that plants grown in varying
phosphate solutions ranging from 0.075 ppm to 0.60 ppm did not signifi
cantly differ with respect to percentage weight gain over a 17 da period.
Haller, et al. (1970) found that the critical phosphrus concentration
for waterhyacinth growth was 0.01 ppm. Above this level phosphorus was
absorbed in luxury amounts but a higher proportion of that available
was absorbed at low concentrations. Haller and Sutton (1973) found
that optimal growth occurred at 20 ppm phosphorus but levels higher than
40 ppm were toxic. They further found that the root weight was greatest
at 0 ppm reflecting a tendency towards maximizing root absorptive sur
face in response to low nutrients.
Sutton and Blackburn (1971a, b) investigated the effects of vary
ing copper solutions on growth and transpiration of waterhyacinth. They
found that transpiration was reduced at 4.0 ppm with copper when grown
in the solution for 1 week and at 2.0 ppm when grown for 2 weeks. Growth
was inhibited by 3.5 ppm when subjected to the solution for 2 weeks.
After one week the shoot dry weight was reduced at 8.0 ppm or above and
the root dry weight by 16.0 ppm. The copper content of the shoot reflected
the content of the water when the concentration was above 2.0 ppm but at
levels below this the concentrations in the roots were independent of
those in the water. The copper content of the roots increased linearly
with the solution concentration.
Boyd and Scarsbrook (1975) found that the addition of 20:20:5
N:P205:K20 fertilizer to ponds increased the biomass yield of water
hyacinth. The fertilizer was added at 4 levels 0, 2.7 kg/ha, 10.8 kg/ha,
and 21.6 kg/ha. It was interesting to note that the highest level of
fertilization resulted in a yield less than the two intermediate levels.

31
Community Associations
Because of the worldwide distribution of waterhyacinth any com
prehensive list of plants associated with it would be a tremendous task.
A few authors have breached this subject on a local level, however.
Harper (1903) noted the association of an orchid, Habenaria repens Nutt.,
with waterhyacinth in Georgia. Small (1936) listed a dozen plants which
may be found growing on the floating mats of waterhyacinth and noted in
New Orleans that it grows intimately with several other aquatic species.
Penfound and Hathaway (1938) described plant communities in the marshlands
of southeastern Louisiana. They found waterhyacinth associated with the
cypress-gum swamps in strictly fresh water and presented an extensive list
of other associated species. Penfound and Earle (1948) found a tremendous
array of plants (63 species) occurring on mats of waterhyacinth. Eggler
(1953) investigated the effects of 2, 4-D on other plant species associated
with waterhyacinth and al 1igatorweed. Chadwick and Obeid (1966) investi
gated antagonism between Pistia stratiotes and Eichhomia crassipes. Bock
(1966) listed several species associated with waterhyacinth in California
and reviewed the work of several other authors. Abu-Gideiri and Yousif
(1974) noted the composition of the plankton community in association
with waterhyacinth stands in the Sudan.
Several workers have reported on the invertebrates associated
with waterhyacinth but this has largely been the result of biological
control investigations dealing primarily with insects. O'Hara (1967)
quantitatively listed the invertebrates found in waterhyacinth mats.
Hansen et al. (1971) listed some invertebrates present in the aquatic
component of the waterhyacinth community and constructed a partial food
web. They also studied the vertebrates present as did Goin (1943).

32
The entomofauna of waterhyacinth is quite large and diverse. Most
of the information available regards those species which feed upon water-
hyacinth and, thus, show potential as biological control agents. Sankran
et al. (1966) investigated a grasshopper {Gesonula punctifrons Stal.
:Acrididae) attacking waterhyacinth in India. Fred Bennett of CIBC in
Trinidad has published many papers on the possibility of biological con
trol of waterhyacinth and on the insects associated with it (Bennett 1967,
1968a, 1968b, 1970, 1972; Bennett and Zwolfer 1968). Other lists have
been provided by Gordon and Coulson (1969), Coulson (1971), Perkins (1972,
1974) and Spencer (1973, 1974).
Sabrosky (1974) described a dipteran stem miner (Eugaurax setigena:
Chloropidae) from South America. Barman (1974) investigated the growth
and assimilation efficiences of an arctiid (Diaovisia virginioa which is
known to feed on waterhyacinth. Silveira-Guido and Perkins (1975) reported
on the biology and host specificity of Comops aquatioum (Bruner), a
grasshopper (Acrididae) from Argentina which attacks waterhyacinth.
DeLoach (1975) provided indentification and biological notes on the genus
Neoohetina (Coleptera: Curculionidae) that attack the Pontederiaceae in
South America. DeLoach and Cordo (1976) provided information on the life
cycle and biology of N. eichhomiae and N. bruohi3 two species which have
been released for the biological control of waterhyacinth in the United
States. Warner (1970) described these two species.
Wallwork (1965) described a leaf-boring galumnoid mite (Orthogalwma
terebrantis) from Uruguay which feeds on waterhyacinth which has subse
quently been found in the United States (Bennett 1968a). Perkins (1973)
studied the biology and host specificity of this species in Argentina.

33
Cordo and DeLoach (1975) investigated the ovipositional specificity
and feeding habits of this mite also in Argentina. Del Fosse et al.
(1975) determined the feeding mechanism of 0. terebrantis from the
Florida strain.
Growth and Development
Morphology
Considerable confusion arises from the lack of uniformity in
naming, the vegetative structures of waterhyacinth. For this reason, I
have adopted the terminology of Penfound and Earle (1948). The roots ace.
numerous, fibrous, unbranched and adventitious (Couch 1971). They vary
little in diameter (ca. 1 mm, Arnold 1940) but may range from 9 to 90
cm in length (Penfound and Earle 1948). The length of the roots is
strongly correlated with leaf length and may be correlated with water
depth (Misra 1969). The roots are feathery in appearance due to the
presence of numerous secondary lateral rootlets (du Toit 1938). Arnold
(1940) found that the origin of these roots is unusual in that they
make their first appearance in the immature pericycle a short distance
from the promeristem and not, as in most plants, in mature tissue. The
root is characterized by a distinct root cap which may extend up to
2.5 cm from the tip and is attached only at the tip (Olive 1894). The
rootlets also possess root caps (Arnold 1940). The beginnings of the
lacunar system are apparent in the roots approximately 1.5 cm from the
tip (Olive 1894). The roots constitute 20-50% of the plants biomass
(Knipling et al. 1970 found it as high as 65%). The cortex is divided
into three zones, which consist of a layer of parenchymatous tissue just
under the epidermis, parenchymatous tissue surrounding the stele, and a

34
layer of lacunate tissue in between. The polyarch stele is surrounded by
a weakly differentiated endodermis and pericycle (Couch 1971). Limited
meristematic activity occurs at the apex and the roots possess a distinct
epidermis (Sculthorpe 1967). The roots are known to become embedded into
the mud although they are usually suspended freely in the water (Buckman
and Co. 1930; Small 1936; du Toit 1938; Penfound and Earle 1948; Parsons
1963). The color of the roots is normally dark violet blue due to the
presence of anthocyanin but when growing in the mud or in the dark the
roots become white (Olive 1894; Penfound and Earle 1948). The roots arise
from nodes on the rhizome (Penfound and Earle 1948).
The rhizome is the vegetative stem of the plant from which all other
structures arise (Penfound and Earle 1948). It consists of an compact
axis with short internodes and the leaves, roots, stolons and inflores
cences are produced by the meristematic nodes which have generally small,
compact cells. An area with a considerable number of intercellular air
spaces exists around the periphery of the meristematic tissue (Couch 1971).
The rhizome is approximately 95% water by weight and has a specific
gravity of 0.805 (Penfound and Earle 1948). Olive (1894) indicated that
there was no evidence of starch being stored within the rootstock (rhizome)
but Penfound and Earle (1948) felt that the rhizome was the main organ
of starch storage.
The rhizome may produce long horizontal internodes (stolons) which
/
produce new shoots at the distal end which results in a sympodial branch
ing pattern (Penfound and Earle 1948). These stolons arise from axillary
stem buds (Bock 1966). Aerating spaces are abundant near the periphery
and the collateral bundles are aggregated in the center (Olive 1894).
The stolons are purple due to the presence of anthocyanin and range in
diameter from 0.5-2.0 cm at i length up to 40 cm (M ra 1969). The

'i-/.
35
specific gravity of the stolon is 0.818 and it consists of about 97%
water by weight (Penfound and Earle 1948).
The most interesting morphological development of the waterhyacinth
is its leaves. Arber (1918, 1920) found that the vascular bundles in the
petiole are arranged with the xylem oriented towards the periphery. Those
in the lamina may be arranged with the xylem up, down, or oblique. This
is in contrast to plants with a true lamina which have the vascular
bundles arranged with the xylem towards the upper leaf surface. She
suggests that this indicates that the lamina is merely an extension of
the apical end of the phyllode and not homologous with the laminae of a
Dicotyledon. As such it should properly be called a pseudolamina and
the basal portion a petiole. The two are connected by a narrow compact
region called the isthmus and the narrow base below the float is referred
to as the subfloat. A membranous ligule (= stipule of Agostini 1974)
is present at the base of the subfloat (Penfound and Earle 1948) which
possesses a small reniform lamina (Agostini 1974).
The petiole may be more or less inflated to form a bulb-like
structure commonly assumed to function in floating the plant (Parsons
1963; McLean 1922; Chhibbar and Singh 1971; Olive 1894; Couch 1971).
This has been contradicted by Rao (1920b) because the bladders are
formed mostly above the water and the leaves float with or without
them. Bock (1966) noted, however, that the bases of the inflated petioles
just beneath the water formed a stable platform. Further, floating
single plants with elongate petioles were unable to remain upright
and if they remained on their side sent out new leaves with inflated
petioles.

36
Several factors have been indicated as important in the development
of the float. Rao (1920b) concluded that high osmotic pressure was the
important factor although this may be altered by numerous factors. The
lack of swelling may also be associated with high plant density (LaGarde
1930), anchorage in soil (LaGarde 1930; McLean 1922), shade and high
temperatures (LaGarde 1930; Arber 1920). Conversely, bulbous petioles
may be associated with the free-floating habit, full sunlight, or cooler
temperatures (LaGarde 1930). Bock (1966) found she could not correlate
petiole shape with shading.
spaces in the highly lacunated aerenchyma resulting in 70% air by volume
(Couch 1971). The specific gravity of the float is 0.136 and of the
pseudolamina is 0.741. The floats are 94% water by weight and the pseudo
laminae are 89% (Penfound and Earle 1948). The petioles range in size
from a few centimeters to as much as 1.5 meters in the equitant form
(Buckman and Co. 1930). The angle between the leaves and the water
surfaces ranges from 15 to 45 around the periphery of the rosette and
from 75 to 90 in the center (Penfound and Earle 1948).
The leaves not only stabilize the plant and keep ttafToat Twrt-act
as sails which catch the wind and move masses of them over the surface
of the water (McLean 1922; du Toit 1938). Further, the geometric arrange
ment of the leaves into a rosette with a large leaf area (as much as
8 m2/m2) and the erect habit of individual leaves is extremely efficient
for light interception (Knipling et al. 1970).
The anatomy of the pseudolamina and petiole is further discussed
by Olive (1894), Bock (1966), Sculthorpe (1967), Arber (1918, 1920),

37
and Penfound and Earle (1948). The lacunar system may enable the plant
to utilize internal carbon dioxide (Billings and Godfrey 1967).
The inflorescence is displayed on a long peduncle (Penfound and
Earle 1948) and is usually elevated a few centimeters above the leaves
(du Toit 1938). Two unlike spathes subtend the inflorescence the lower
being leaf-like and bearing a pseudolamina and the upper bract-like
(Cook 1974). The inflorescence is a spike (du Toit 1938; Buckman and
Co. 1930; Penfound and Earle 1948; Bock 1966) or may be considered
spike-like or paniculate (Cook 1974; Bock 1966). The spike is 15-30 cm
long (du Toit 1938; Mune and Parham 1954) and contains numerous flowers
(6-20, du Toit 1938; 8, Parsons 1963; 10-12, McLean 1922; 6-12, Mune and
Parham 1954; 4-29, Misra 1969) borne on a rachis with a flowerless sub-
rachis below the inflorescence and above the spathes (Penfound and
Earle 1948). The individual flowers consist of a hypanthium about 1.4-
1.8 cm long (Misra 1969), 3 sepals, 3 petals, 6 stamens and a tricarpel late
ovary (Penfound and Earle 1948). The petals and sepals are lavender in
color (Bock 1966) and united at the base to form a 6-lobed tube (Cook
1974). The color of the flower is due to the anthocyanin, eichornin
(Shibata et at. 1965). The tube is curved, glandular and pubescent near
the base (Mune and Parham 1954). The perianth is slightly irregular with
all 3 sepals and 2 petals similar in size and shape but the upper petal
is somewhat wider and bears a distinctive yellow spot in the center
bordered by a darker blue or violet area (Bock 1966; Buckman and Co. 1930).
Buckman and Co. (1930) indicate that the function of this spot is obscure
but others have indicated that it may function as a nectar guide to
visiting bees (Sculthorpe 1967). The six stamens are arranged in two

38
whorls of 3 stamens each, of two different lengths and are adnate being
fused to the corolla tube at the base of the filaments near the sinuses
of the perianth lobe (Bock 1966). The filaments are white at the base,
purple at the apex, and glandular (Misra 1969). The anthers are oblong
and attached near the base (Bock 1966) and contain about 2000 pollen
grains each (Penfound and Earle 1948). The ovary is superior, sessile,
trilocular, contains numerous ovules in axile placentation (McLean 1922;
Bock 1966) and conical in shape (Penfound and Earle 1948). The six
stigmatic surfaces, by their close approximation, appear to be capitate
but they are not (Penfound and Earle 1948; Bock 1966). The stigmate
surface is covered with numerous glandular hairs (Bock 1966). The fruit
is a loculicidal capsule containing seeds with an abundant mealy endo
sperm (McLean 1922). Approximately 50 seeds are produced per capsule
(Penfound and Earle 1948).
Perennation
Waterhyacinth is generally considered to be a perennial by virtue
of its rhizome (Penfound and Earle 1948; Sculthrope 1969). Penfound and
Earle (1948) felt that the rhizome may maintain a constant length over
aj3£rQt_ may exist is not available but the plants are known to _survive periods
of freezing weather by resprouting from the rhizome (Bock 1966).
Physiological Data
Most of the appropriate physiological data has already been discussed
in scattered sections of this literature review but it bears repeating in
a more organized discussion.

39
Many authors have studied transpiration rates and found variations
due to such factors as solar energy, wind speed, temperature, and technique
through a waterhyacinth mat was as high
Misra (1969) found that water
as 65.5 kg/m /da. This represented a water requirement of 6.74 kg of
water-per gram (dry wgt.) of biomass produced. The ratio of evapotrans-
piration to open water evaporation (Ej:Eq) ranged from 5.92 to 9.84.
Knipling et al. (1970) measured the moisture content of a stream
of air before and after it had passed over a waterhyacinth leaf. They
found during the day the transpiration rate increased from 1520 to 2450
mg/dm2 (leaf surface)/hr. in response to increasing light intensities.
Dark transpiration values were also high, however, averaging 1430 mg/dm2/hr
In another experiment plants were grown in beakers in a variety of phos
phorus concentrations and measured for daily water loss. There was no
significant difference in evapotranspiration between the phosphorus con
centrations. The Ej'Eq ratio, however, was 305 g/da:100 g/da or 3:1.
Dry matter production was 0.27 g/da indicating a water use efficiency
ratio of 1129 gm H20/g plant dry wgt.
Other average values reported for the E :E ratio have been 3.2
(Penfound and Earle 1948), 7.8 (in India; Holm, et al. 1969), 3.7 (Timmer
and Weldon 1967), 1.02-1.36 (Brezny et al. 1973), and 1.46 (Van der Weert
and Kammerling 1974). The latter authors have found that 97% of evaporation
from waterhyacinth covered situations is the result of the process of
evapotranspiration.
Knipling et al. (1970) have also provided data on respiration and
photosynthesis by measuring the C02 concentration in an air stream passed
over a leaf. Net photosynthesis increased from 7.8 to 16.1 mg C02/dm2/hr

40
with light intensities increasing from 1450 ft-c to 8000 ft-c. Respiration
averaged 2.7 mg C02/dm2/hr. Ultsch and Anthony (1973) have found that
waterhyacinth may have the capacity to utilize C02 dissolved in the
wc^ter^under the mats as a source of carbon by absorption throuc
;oots. This majc-aeetTT f or up^Kr'->99Laf the total carbon fixed. Billings
and Godfrey (1967) have found that some hollow stemmed plants may use
internal carbon dioxide generated from root and stem respiration in
photosynthesis. This may be true in the case of waterhyacinth.
Nutrient uptake rates are not well worked out for this plant.
Waterhyacinth contains about 0.4% P and 2.6% N by weight (Boyd 1970b)
or an N:P ratio of 6:1. If these represent constant proportions the rate
of nutrient uptake is proportional to the growth rate of the plant.
If the standing crop increases at a rate of 100 g/da the rate of uptake
of N will be 2.6 g/da and of P 0.4 g/da. This agrees well with the
results of Dunigan et al. (1975) who found the N:P ratio of uptake rates
to be 5-6:1. The daily absorption rates from 6 liter containers were 2.4
ppm N and 0.4 ppm P in water concentrations of 50 and 100 ppm N and P.
At concentrations of 250 ppm N and P the daily uptake rates were 3.5 and
0.7 ppm respectively. This implies a growth rate of 600 and 810 mg dry
wgt/da. Mitsch (1975) indicates that this high ratio of N:P absorption
indicates that nitrogen is generally more limiting to waterhyacinth
than phosphorus.
Phenology
Data on the sequence and timing of events in the annual cycles of
waterhyacinth population are scarce. Penfound and Earle (1948) measured
the average length of the largest leaves over one growing season. The

41
maximum was reached in August and the period of maximum growth was
between May and August. They also followed the flowering cycle over a
period of years in Louisiana. In the years 1945-1947 anthesis began in
April. They felt that a definite flowering rhythym occurs in a given
colony of plants. Anthesis is maximum in June and declines through
September although this may vary from colony to colony and a second
period of flowering occurs in September and October and continues through
November into December. Buckman and Co. (1930) reported that the plant
is supposed to bloom every two to three months. In India flowering occurs
throughout the year but is most abundant in the post-monsoon months
(Sahai and Sinha 1970). A pre-monsoon flowering period (April and June)
has been reported in India (Pieterse 1974). Sahai and Sinha (1970)
further found that biomass accumulation was highest in January and
February, and the area occupied [% cover) was greatest in February and
March in India.
Reproduction
Floral Biology
A single flowering spike contains a variable number of flowers.
Bock (1966) found the average to range between 5 and 10 flowers per
inflorescence although she indicated that other authors have observed up
to 35 flowers per inflorescence. Small (1936) indicated that flowering
occurred on a daily cycle appearing as buds up to 7:30 AM and opening by
8:00 AM. The mode of pollination may be allogamous or autogamous. Bock
(1966) found that allogamous pollination may occur through the actions
of several insect pollinators. She listed Apis mellifera, Haliotus (1 sp.),
and Lasioglossum (2 sp.) as known pollinators and syrphid flies as possible

42
pollinators. Penfound and Earle (1948) observed honeybees, bumblebees,
black unidentified bees, and sulfur butterflies visiting the flowers.
They described three patterns of behavior of honeybees in visiting the
flowers; visiting distal anthers only, alighting with the head among the
proximal anthers and the abdomen on the stigma, and visiting the proximal
anthers after alighting on the banner petal. They questioned the import
ance of insect pollinators in accounting for the production of seed in
this species. Bock (1966) noted, however, that honeybees crawl down the
floral tube to retrieve the nectar and in so doing receive pollen from
both sets of anthers. She observed a great deal of cross-pollination.
Autogamous pollination occurs when the flower wilts and the stamens
are twisted against the stigma (Penfound and Earle 1948; Bock 1966; Tag
el Seed and Obeid 1975). Penfound and Earle (1948) found much more pollen
on the stigma after the flowers had completely wilted than at any other
time thus stressing the prevalance of autogamous pollination.
Since waterhyacinth flowers are at least dimorphic with regard to
style length either legitimate (styles pollinated by anthers not of equi
valent length) or illegitimate (styles pollinated by anthers not of equi
valent length) crosses are possible (Ornduff 1966; Bock 1966; Frangois
1964-63). Both legitimate and illegitimate crosses result in seed production
(Bock 1966). Frangois (1964-63) reported that self-incompatibility was
stronger in long styled forms than in short styled forms. Ornduff (1966)
studied the breeding system of Pontederia cordata and compared it with
E. crassipes. He concluded, as did Bock (1966), that both species exhibit
relatively weak self-incompatability.
An interesting aspect of the floral biology of waterhyacinth is the

43
phenomenon of anthokinesis or the bending of the axis of the inflorescence
following anthesis (Agharkar and Banerji 1930; LaGarde 1930; McLean 1922;
Penfound and Earle 1948; Bock 1966; Misra 1969). LaGarde (1930) described
this process as follows:
"As soon as the inflorescence starts wilting the upper portion of
the stalk with the fertilized blossoms begins to bend downward. When
this upper part has reached the surface of the water, usually after five
days, the lower portion of the stalk commences to bend at the base, thus
pushing the developing seed-pods under the surface of the water. This
movement stops when the lower part of the stalk is level with the surface.
The upper part carrying the seed pods is then submerged in the water at
an angle of 45, the seed pods being covered and protected by the root
system . The whole process of bending requires from six to seven days."
(LaGarde 1930, p. 51).
Agharkar and Banerji (1930) quoted other workers who indicated
that anthokinesis was accompanied by a lengthening of the peduncle. They
found, however, that the peduncle did not lengthen considerably and such
lengthening was confined to an area of 1 to 2 cm below the terminal node.
They further found that removal of the flowers or amputation of the peduncle
above the node had no effect and curvature was normal. This was also true
when they removed the peduncle and placed it in water.
Penfound and Earle (1948) studied that anthokinetic cycle and their
results agree with other workers. They found that it requires about 14
days from the initiation of the floral bud until opening occurs. Floral
opening begins about 8:00 AM, if all the flowers open the bending phase
begins at about 5:00 PM of the same day. Bending occurs in three places:
at the rhizome crown, j r' elow the two bracts of the inflorescence,

44
and in the rachis. Most of the flowers are inverted by 5:00 PM the
following day. The complete cycle from flowering to complete geniculation
takes 48 hours in the summer. This is contrary to LaGarde's (1930)
finding that it takes 6 or 7 days. Bock (1966), in her studies, agreed
with Penfound and Earle (1948).
Bock (1966) seemed to concur with the findings of Rao (1920a) in
that the: bending was due to geotropism in that when the roots were packed
with sponges and the plants held horizontally, no bending occurred. She
disagreed with Agharkar and Banerji (1930) in that removal of the flowers
would not permit bending to occur unless all of the flowers had wilted
and bending had commenced first.
Misra (1969) found that curvature took place when the tips along
with 2 terminal flowers were removed, when all of the flower buds were
removed, and when all of the flowers were removed after they had opened.
In all cases complete bending took as long as in the controls (35-40 hrs.).
He found that this curvature was due to increased cell size along the
outer edge of the curving portion. He felt that this process represented
a free-running endogenous rhythm independent of auxins (geotropic in
nature), photoperiod, temperature, and opening of the last flower as
suggested by other authors.
Spermatogenesis has been described by Smith (1898) and Banerji and
Gamgulee (1937) and oogenesis by Smith (1898). Pollen morphology and
germination and development of the pollen tube have been investigated
by Ganerji and Gangulee (1937), Bock (1966) and Tag el Seed and Obeid
(1975). The embryology of the seed is discussed by Smith (1898), (Coker
(1907), and Swamy (1966).

45
Seed Production and Dispersal
The degree of seed set seems to be extremely variable. Agharkar
and Banerji (1930) found that 10 hours after anthesis through natural
pollination (autogamous or allogamous not distinguished) 35% of the
flowers were fertilized (30% with actively growing pollen tubes and
another 15% with pollen grains present). Through artificial pollination,
up to 71.3% of the fertilized flowers set fruit. McLean (1922) in Bengal
found that only 1% of the flowers set any seed. Haigh (1936) found in
Ceylon that 36 to 71% of the capsules produced may be empty. Backer (1951)
found no seed set in Java and Misra (1969) found up to 48% of the fruits
bear seed in India.
The conditions for seed set have been investigated but the results
are confusing. Parija (1934) indicated that temperatures between 24C
and 29C were necessary. Agharkar and Banerji (1930) felt that relative
humidities above 90% were required. Bock (1966) found that seed was set
when the relative humidity was never greater than 72%. Haigh (1936)
found the number of seeds per inflorescence to be 86, 28, and 91 when
the relative humidity was 90%, 70%, and 67% respectively. Tag el Seed
and Obeid (1975) concluded that seed set was favored if pollination
occurred immediately after the flowers opened. Thereafter, successful
pollination was hindered by high temperature and low humidity which
affected the stickiness and receptivity of the stigma.
Data on the quantity of seeds set per fruit or inflorescence also
indicate a great deal of variability. Haigh (1936) artificially pollinated
flowers and found an average of 24 seeds/capsule with a maximum of 72.
Bock (1966) reported the average in California to be 4.2 with a range of

46
1-16. Misra (1969) indicated that the average may be 15-41. Robertson
and Thein (1932) found 50-150 seeds/capsule in Burma, Zeiger (1962)
reported 3-250, and Francois (1964-3) reported an average of 153.6 with
a maximum of 244. Tag el seed (1972) reported an average of 98.95 with
a range of 5 to 542.
The number of seeds per inflorescence depends upon the number of
flowers per inflorescence and the number of seeds per flower. Bock (1966)
indicates that the average number of seeds per inflorescence is 3.44 in
California. Tag el Seed and Obeid (1975) reported 1.5 capsules per
inflorescence. Using the data from Tag el Seed (1972) for the average
number of seeds per capsule (98.95) this expands to 148 seeds per inflor
escence. Matthews (1967) indicated that a single spike may produce 5000
to 6000 seeds. Zeiger (1962) estimates that 45 million seeds may be
produced per acre by medium sized plants. Penfound and Earle (1948)
estimated a crop of 900,000 capsules/acre.
The mode of seed dispersal has not been studied to any extent.
It seems apparent that since the seeds are deposited in the water the
primary mode of dispersal would be through drifting. A few authors have
indicated that birds and fur bearing animals may disperse seeds (Maiden
et al. 1906; Holm et al. 1969; Gay 1960).
Viability of Seeds and Germination
Conditions for germination of waterhyacinth seeds have been
studied by many workers. Crocker (1907) refuted the idea of earlier
workers that desiccation of the seeds is a necessary prerequisite to
germination. He further found that green seeds kept at 23C germinated
(20% within 1 week) while seeds kept at 6C did not, although they did

47
ripen. Seeds with mature coats failed to germinate when kept at either
23C or 29C. He then separated the embryo from ripe seeds or ruptured
the seed coats and placed them in a bath at 29C. He noted that germi
nation occurred very rapidly in both cases (96% after 1 day). He ruled
out oxygen as a factor because they germinated equally well in boiled
water covered with paraffin. He concluded that the hard seed coat and
endosperm hinders water absorption and limits germination and that
desiccation may, in fact, fracture the seed coat and promote germination.
Agharkar and Banerji (1930) indicated that a ripening period of
20 to 23 days was required for maturation of the fruit. After maturation
they are severed from the axis by an abscission layer and float on the
water surface for a day or two before sinking. Splits develop on the
lateral walls through which seeds are discharged. They found that the
seeds develop freely in tap or distilled water.
Parija (1930) suggests that germination takes place "in the
beginning of rains or whenever the humidity, soil moisture and temperature
are suitable." (Parija 1930, p. 388). He felt that the function of the rain
was to provide moisture, and expose the seeds in the mud providing access to
oxygen.
Robertson and Thein (1932) noted that in every instance when they had
found waterhyacinth seedlings it was in a depression which completely dries
out during the dry season and floods again in the rainy season. They concluded
that a period of drought alternating with a period of plentiful moisture was
necessary for germination.
Haigh (1936) exposed seeds to varying treatments of always wet,
always dry, or alternately wet and dry. No germination occurred for three
months unti eeds were placed in the sun. Within eleven days germination

48
had begun. They further found that drying, bubbling air in the water,
and the addition of rotting waterhyacinth fragments would not promote
germination when kept in the laboratory. To determine if heat or exposure
to an intense light was the important factor they exposed seeds to a
normal light bulb and to a blackened light bulb. Germination occurred
only in the illuminated treatment. They concluded that bright sun is
necessary for germination and that heat, in conjunction with high light
intensity may also be required. Haigh (1940) later found that if dislo
cation is prolonged for a long enough period light is not necessary.
He found that seeds collected in June 1935 would germinate in the labor
atory as late as January 1937 (19 months).
Penfound and Earle (1948) concur with Haigh in his findings. They
found that seed germination would occur on upturned plants indicating
that either drying or increased light intensity was favorable for germin
ation. They also concluded that scarification aided germination.
Hitchcock et at (1949) indicated that an after-ripening period of
about 2 months was necessary for germination and under ideal conditions
100% germination was possible. Dry seeds, however, required twice as
long (111-112 days) to germinate as seeds stored wet (64-67 days). They
also found that relatively high water temperature (28-36C) favored ger
mination but the seeds could survive very cold temperatures. When stored
for 69 days at temperatures of -5, 0.5, 5, 10, and 22C and then placed in
normal air temperatures germination occurred in every case except the -5C
treatment. When exposed for only 1 week even the -5C treatment gave 50%
germination.
Hitchcock et al. (1950) investigated the effects of water depth on
seed germination. At water depths of 2.5, 10.2, 20.3, 30.5, and 40.6 cm
they obtained 40, 60, 72, 72, a '>00% germination respi. tively. They

49
felt that the difference was due to longer heat retention in the deeper
water at night. Under 15.2 cm of water in a brown glass bottle only 28%
germination was observed.
Barton and Hotchkiss (1951) also studied the effects of temper
ature, light, and storage on seed germination. They concluded that a com
bination of high temperature and light is needed for germination of dor
mant seeds although temperatures as low as 5C did not impair germination
when in direct sunlight (greenhouse) and alternating temperatures (5-30C,
5-35C, and 5-40C) allowed some germination even in the dark. They also
found that a storage period of a month or longer hastened germination
especially with less mature seeds.
Francois (1964-3) obtained good rates of germination (over 95%)
by keeping his seeds in a 12:12 L:D photoperiod with a corresponding
40C:20C temperature regimen. Bock (1966) was convinced that seeds do
not germinate in California and found that they do not remain viable
there for longer than 2 months. Sculthorpe (1967) reflected the findings
of other authors by indicating that the seeds are able to tolerate a long
dry period and remain viable. Tag el Seed (1972) investigated seed germ
ination under a wide range of chemical treatments and under low oxygen
tension and low redox potential as well as many other environmental
conditions. His extensive studies indicate that germination is stimu
lated by low redox potential and low oxygen tension expecially after wet
storage, germination is most likely to occur in water warmed by intense
light, the addition of organic matter to the substrate stimulates germ
ination, the seeds will only germinate at the surface of the substrate,
and aeration has no significant effect on germination.

50
The growth and development of the seedlings have been described
by Parija (1930), Robertson and Thein (1932), Haigh (1936), and Penfound
and Earle (1948). Several authors have indicated that a water saturated
medium is necessary for seedling survival (Hitchcock et al. 1949; Parija
1930; Haigh 1936) but forced immersion in water retards growth or kills
the seedling (Parija 1930; Hitchcock et al. 1949). Penfound and Earle
(1948) and Hitchcock et al. (1940) noted that seedlings would grow on
waterhyacinth flotant and Pettet (1964) found them growing on the shore
in "strand-lines" created by dead waterhyacinths. Hitchcock et al. (1950)
noted that in nature factors which prevent young seedlings from surviving
may be more important than factors which permit seed germination.
Vegetative Reproduction
Even though seed production by waterhyacinth may be massive, the
primary mode of reproduction is through vegetative propagation (Hitchcock
et al. 1950). This occurs through the production of offsets, or suckers,
produced on stolons (Penfound and Earle 1948). Hitchcock et al. (1950)
found that offset production begins about 60 days after the plant germinates
when the rosette attains a diameter of 7.6 to 10.2 cm. Penfound and Earle
(1948) found that a mat extends its boundaries at a rate of 3 feet per
month through vegetative reproduction under favorable conditions and the
plants double their numbers every two weeks. Bock (1966, 1969) and
Perkins (1972) have reviewed the literature dealing with the rates of
offset production in different locations and situations.
Productivity and Standing Crop
Many authors have dealt with waterhyacinth productivity in one

51
form or another. Bock (1966, 1969) has done perhaps the most comprehen
sive study on productivity but she dealt with fresh weight and increment
factors making comparisons with her data difficult. She also reviewed
most of the literature on the subject and compared it to her data. Table 1
gives an updated compilation of various measures of standing crop and
productivity of waterhyacinth from various sources.
Control
The literature dealing with the various means of control is vol
uminous and I won't attempt to review it here. The Hyacinth Control
Journal has been published annually since 1962 and is largely devoted
to this subject. Furthermore, the various control methods have recently
been reviewed. Robson (1974) has reviewed the methods for mechanical
control of aquatic weeds and Blackburn (1974) has reviewed chemical
control and the various compounds available in a recent UNESCO publication.
In the same publication Bennett (1974) reviewed the biological control
of aquatic weeds. Biological control has also been reviewed by Andres
and Bennett (1975) and the use of plant pathogens in biological con
trol efforts by Zettler and Freeman (1972), Freeman et at. (1974) and
Charudattan (1975). Mitchell (1974) summarized techniques for the con
trol of aquatic weeds through habitat management. Sculthorpe (1967)
also discussed the various methods of aquatic weed control.

52
Table 1. Standing crop and productivity of waterhyacinths as estimated by various authors.
Source
Standing Crop
1.6-2.7 kg DW*/m2
Productivity
Penfound and Earle (1948)
Dymond (1949)
1.6 kg DW/m2
13-20 kg/m2/yr
Penfound (1956)
0.4-1.3 kg DW/m2
12.7-14.6 g DW/m2/da
5.7-6.5 g C/m2/da
Westlake (1963)
----
1.1-3.3 kg/m2/yr average
15 kg/m2/yr max. (19 g C/m2/da)
Yount (1964)

28 g C/m2/da
Bock (1966)
----
2.5/ per day (Calif, average)
Misra (1970)

9.4-9.6 g 0M*/m2/da (Aug. 1967)
3.48-8.98 g 0M/m2/da (Apr.-Feb. 1968)
Sahai and Sinha (1969)
0.46-0.72 kg DW/m2
3.8 g 0M/m2/da max.
103.0 g 0M/r:2/yr
247.0 g OM/m2 net production to max.
biomass
Knipling et al. (1970)
2.4-2.5 kg DW/m2
7.8-16.1 mg CO?/dm2 leaf/hr net
2.6-2.8 mg C02/dm2/hr respiration
Sinha and Sahai (1972)

1.43 g 0M/m2 leaf/hr net
0.56 g OM/m2 leaf/hr respiration
1.99 g OM/m2 leaf/hr gross
Ornes and Sutton (1975) 9.7 gm/m2 max.
*DW = dry weight; 0M = organic matter.
1.05X per day (= 30 gm 0M/r,;2/week) max.

53
Arzama densa Wlk.
Taxonomy
Walker (1864) described three genera and three species of moths in
two families which are now known to be closely related. These were
Edema obliqua (Notodontidae), Bellura gortynoides (Notodontidae), and
Arzama densa (Gortynidae). His description of the latter genus and
species follows:
Genus Arzama
Male. Body stout. Head with thick-set porrect hairs. Proboscis
short, slender. Palpi stout, porrect, pilose, not extending
beyond the hairs of the head; third joint extremely small,
not more than one-tenth the length of the second. Antennae
moderately pectinated, rather short. Abdomen extending much
beyond the hind wings, tapering towards the tip, which has a
very small tuft. Legs stout, rather short; hind tibiae with
a short fringe; spurs long, stout. Wings rather short and
narrow. Fore wings acute; exterior border almost straight,
hardly oblique; second inferior vein almost as near to the
third as to the first; fourth not very remote from the third.
Arzama densa
Male. Reddish. Underside, abdomen and hind wings reddish cinereous.
Fore wings with an oblique very broad brownish band, which
contains the orbicular and reniform marks; the latter are red,
oblique, and narrow; a submarginal brown-bordered slightly
dentate band, which is rather brighter than the ground hue.
Hind wings beneath with a round brown spot in the disk, and

54
with a slight exterior brownish band. Length of the body
9 lines; of the wings 16 lines.
Grote and Robinson (1868) described a second species of Arzama,
A. obliquata. They compared this to Walker's type of A. densa in the
British Museun and found they differed in the larger size of A. obliquata
and different coloration. They apparently failed to compare it with
Edema obliqua, however,
Herrich-Shaeffer (1868: Cited from Zoo. Record) provided generic
and specific characters in full for A. densa from Cuba.
Grote (1873) described another species of Arzama, A. vulnifioa,
which differed from A. densa Wlk. and A. obliquata G. & R. primarily by
its dusky yellow color. He also noted that it was less robust than A.
obliquata with the anterior wings more rounded posteriorly. Grote (1874)
in his list of the Noctuidae of North America listed only these three
species but in 1878 [1879] described a fourth species, A. diffusa from
Maine. Gundlach (1881) redescribed Arzama densa Wlk. from specimens
collected in Cuba. A fifth species, A. melanopyga, was subsequently
described by Grote in 1881 (in Comstock 1881) from Florida. He pointed
out characters which separate A. diffusa, A. vulnifioa, 4. melanopyga,
and Sphida obliquata (apparently a recombination for A. obliquata G. & R.).
He noted that characters of the clypeus are of value in separating these
two genera. In 1882 Grote synonymized Edema obliqua Wlk. with Sphida
obliquata G. & R. These five species were listed together in the sub
family Arzaminae by Grote 1883 who noted that the species with the black
anal tuft (melanopyga) is probably a variety of vulnifioa. Riley (1885)
stated that the genus Sphida Grt. had no existence in nature and Sphida
obliquata G. & R. was synonymous with A. densa Wlk.

55
In 1889 Grote placed these species in the tribe Arzamini which
included Arsama and Sphida, the former supposedly having a smooth front
and the latter a tuberculate front. He considered Arzama as consisting
of three species, apparently after having considered A. melanopyga to be
a variety of A. vulnifica. This was also how the group was arranged in
his checklist of 1890.
Smith (1893) decided that Walker's Bellura and Arzama were congeners
and that Bellura had page priority. He considered A. densa Wlk., A.
vulnifica Grt., and A. melanopyga Grt. synonyms of B. gortynoides Wlk.
He recombined Arzama diffusa Grt. into B. diffusa (Grt.). Edema obliqua
Wlk., Sphida obliquata (G. & R.) and Arzama obliquata G. & R. were
considered synonyms of Bellura obliqua (Wlk.). Thus, the seven species
were reduced to three, all in Bellura Wlk.
BeutenmT1er (1902) redescribed B. obliqua (Wlk.) and B. gortynoides
Wlk. but he also recognized B. melanopyga. All three were found in New
York. Holland (1903) considered B. densa (Wlk.), B. vulnifica (Grt.),
and B. melanopyga to be synonyms of B. gortynoides Wlk. as did Smith
(1893) but recognized the genus Sphida and considered S. obliquata (G. & R.)
a synonym of S. obliqua (Wlk.). Hampson (1910) recognized the genera
Sphida, by the single species S. obliqua, and Bellura. He considered
A. densa Wlk. and A. vulnifica Grt. synonyms of B. gortynoides, and
retained B. melanopyga and B. diffusa. He also presented a key for
separating the three Bellura spp.
Dyar (1913) revised the genus Sphida, described three new species,
and he provided a key. He retained s. obliqua (Wlk.) and considered
E. obliqua Wlk. and A. obliquata G. & R. synonyms. The new species
described were S. oecogenes from Washington, D. C., S. anoa from Miami,

and S. gargantua from California. He also included S. pleostigma Dyar
and indicated that the description of this species was in a forthcoming
paper.
Barnes and McDunnough (1914) considered Sphida Grt. synonymous with
Arzama Wlk. thus making obliqua, densa, gargantua, and anoa all species
of Arzama. They synonymized S. oecogenes Dyar with A. densa Wlk. but
made no mention of S. pleostigma Dyar. They considered A. densa Wlk.
distinct from B. gortynoides Wlk. by virtue of a frontal protuberance.
Later they described another species of Arzama from New Jersey and named
it A. brehmei in honor of its discoverer (Barnes and McDunnough 1916).
Grossbeck (1917) in a list of the insects of Florida recognized
B. gortynoides Wlk., B. melanopyga Grt., S. obliqua Wlk., and S. anoa
Dyar. Barnes and McDunnough (1917), apparently having identified S. obliqua
for Grossbeck, noted that they made their determination before the pub
lication of Dyar's S. anoa and indicated that the specimens they identified
were probably S. anoa Dyar. This is confusing, however, because here
they are recognizing S. obliqua Wlk. which they had earlier combined with
Arzama.
Dyar (1922) re-evaluated the status of the genera Arzama and Bellura.
He noted that Hampson (1910) placed A. densa Wlk. as a synonym of B.
gortynoides on the assumption that both have a smooth clypeus. He also
noted that Barnes and McDunnough (1914) found that the type specimen of
A. densa Wlk. did have a tubercle on the clypeus and resurrected the genus
Arzama making Sphida a synonym of it but considered B. gortynoides Wlk.
distinct. Dyar examined several specimens identified as B. gortynoides
Wlk. and found that they all had tubercles on the clypeus and suspected
that Walker's types would also. He felt this would probably synonymize

57
densa Wlk., gortynoides Wlk., and probably anoa Dyar. He therefore
proposed the name Arzamopsis for those species with a smooth front and
suggested that A. diffusa be the type species and .4. melanopyga be
included in the genus. He also described Arzama matanzanensis, a new
species from Cuba.
Seitz (1923) again considered all of these species in the genus
Bellura Wlk. The species listed were B. obliqua (Wlk.), B. densa (Wlk.)
(= oeoogenes Dyar), B. gargantua (Dyar), B. anoa (Dyar), B. matanzanensis
(Dyar), B. pleostigma (Dyar), B. gortynoides Wlk. (= vulnifiea Grt.),
B. melanopyga (Grt.), and B. diffusa (Grt.). He noted that B. pallida
B. & Benj. and B. brehmei B. & McD. are probably races of B. obliqua
(Wlk.) but may be distinct species.
Comstock (1936) discusses this group of insects in his introductory
entomology text. He noted that the genus Bellura contained three North
American species, B. melanopyga, B. diffusa, and B. gortynoides. He
also recognized the genus Arzama and listed A. obliqua as "our most
common species". He included these species in the subfamily Apatelinae.
Jones (1951) listed the macrolepidoptera of British Columbia and
included Arzama obliqua (Wlk.) and Bellura gortynoides Wlk. He noted,
however, that the latter species is a doubtful record. He synonymized
Dyar's Aarzamopsis [sic] with Bellura and considered B. vulnifiea (Grt.)
a synonym of B. gortynoides Wlk.
Tietz (1952) listed Arzama obliqua (Wlk.) and A. densa Wlk. from
Pennsylvania. He considered pallida B. & Benj. a race of A. obliqua
(Wlk.), obliquata Grt. a synonym of B. obliqua (Wlk.), and oeoogenes
Dyar a synonym of A. densa Wlk.

58
Forbes (1954) considered all species of Belluva, Avzama, and Sphida
to be in the single genus Avzama. He divided the genus into two groups
based on whether the front was flat or had a strong central bulge. In
the first group he included govtynoides Wlk., diffusa Grt., and vulnifica
Grt. and placed melanopyga Grt. as a synonym of vulnifica. In the second
group he included obliqua Wlk., bvehmei B. & McD., and densa Wlk.
Kimball (1965) listed the Lepidoptera of Florida and again recognized
both Avzama and Belluve. He included A. obliqua (Wlk.), A. [bvehmei B. &
McD.], A. densa Wlk., A. anoa (Dyar), B. govtynoides Wlk., and B.
melanopyga (Grt.). He noted that the two species of Belluva were
probably conspecific. He also referred to the specimen listed as B.
govtynoides Wlk. by Grossbeck (1917) and noted that it was actually
A. densa Wlk. making the former a synonym of the latter.
The only species listed by Tietz (1972) was Avzama gavgantua (Dyar).
Levine (1974) noted that B. vulnifica and B. govtynoides are separated
largely by the color of their anal tuft, the former being brown and the
latter white. He found that dark brown-tailed females (B. fulnifica Grt.)
may produce white-tailed daughters (b. govtynoides Wlk.). This indicates
that B. vulnifica is merely a form of B. govtynoides Wlk.
I received a personal correspondence from Dr. E. L. Todd from the
Systematic Entomology Laboratory of the U. S. Department of Agriculture
in April 1974. He explained the taxonomic situation with regard to these
species as follows:
I consider that Belluva Walker 1864, type-species
B. govtynoides Walker by monotypy is the valid generic name.
Avzama Walker 1864, type-species A. densa Walker by monotypy
and Sphida Grote 1878 [1879], type-species Avzama obliquata

59
Grote & Robinson (=Edema obliqua Walker), I consider
to be junior synonyms. Belluva has page priority over
Avzama (Walker, 1864, List ..., pt. 32, p. 465 vs p. 645.).
In addition, so far as I can find, J. B. Smith, 1893, Bull.
U. S. Nat. Mus., No. 44, p. 181, was the first to treat
both names and he placed Avzama in the synonymy of Bellura.
Forbes, 1954, Cornell Exper. Stat. Mem. 329, p. 217-8, used
Avzama in error, but divided the genus into two sections
(=Subgenera?). However, the character he uses to divide
the two sections are invalid. Females of Belluva do not
have simple antennae as he indicates, and the front may be
developed in some forms. The extent of development of the
frons is a character that needs more study. It will also
be necessary to study the possibility that food plant
. varieties are involved. I have indicated to others that
I believe there are only two or three species in the genus,
govtynoides, obliqua3 and possibly densa. Smith believed
that govtynoides and densa represented one species, and he
sank that latter as a synonym.
I think that it is apparent from the literature that the taxonomy
of this group is of an uncertain status. I agree with Todd that these
species probably represent one genus and the proper name of Avzama densa
Wlk. is Belluva densa (Wlk.). Because of the widespread current use of the
former name and the absence of a definitive study in literature I have
used the binomial Avzama densa Wlk. throughout this dissertation.

Host Plants
60
Table 2 lists host plant records for these species as indicated in
various references. Because of the continual changes in the taxonomy of
the group, however, these host records are not reliable. For example,
Grossbeck (1917) listed Bellura gortynoid.es Wlk. from Mellonville, Florida,
as inidcated by Hampson (1910) and implicated Typha as the host. Kimball
(1965), however, indicated that the Mellonville record quoted by Grossbeck
(1917) referred to Arsama densa. This creates uncertainty since the host
record was not from Hampson (1910), who synonymized densa and gortynoides,
but the geographic record was. Grossbeck (1917) apparently derived the
host record from other sources. It is therefore impossible to determine
which species Grossbeck's host record refers to. To partially alleviate
this problem I have left the records in Table 2 with the binomial designated
by the respective author intact regardless of synonyms. Where, in my
opinion, there is sufficient agreement in the literature to indicate that
a name is in synonymy with a more valid binomial, that species designation
is included as a subcategory under the valid binomial.
Host plant synonymies also result in a great deal of confusion.
For example, Nymphaea americana (Prov.) Miller & Standiey listed as
a host of Bellura melanopyga (Table 2, No. 2c) is listed by Muenscher
(1967) as a synonym of Nuphar variegatum Engelm. Nymphaea advena (Table
2, No. 2) is also apparently a synonym of Nuphar variegatum (Fassett
1969). I do not believe any bona fide record exists of these species
attacking any of the Nymphaea species.
I am not sure what the Nelumbium sp. (Table 2, No. la) and the
Nelumbom sp. (Table 2, No. lc) records refer to. They may mean Nelumbo
but, if this is so, I doubt the veracity of the record. I also question
the records for Sagittaria sp. (Table 2, Nos. la, lc) and Sparganium
sp. (Table 2, No. 1

Table 2. Host plants of the Bellum-Arzama complex listed from various literature sources.
Species
Host Plant
Source
1. Bellura obliqua (Wlk.)
Typha latfolia L.
"cattail"
Beutenmuller 1902;
Seitz 1923.
Rummel 1919.
a. Arzama obliquata G & R
Sagittaria sp.
Nelumbium sp.
Typha latifolia L.
"cattail reed"
'Typha sp.
Riley 1883a, b;
Kellicott 1883b.
Riley 1883a, b.
Kellicott 1883b.
Brehme 1888, 1889.
Moffatt 1889.
b. Sphida obliqua (Wlk.)
Typha lati folia L.
Typha sp.
Holland 1903; Hampson 1910.
Dyar 1913; Welch 1914;
Grossbeck 1917.
C. Arzama obZi'qua~(Wlk.)
"corn"
Typha lat folia L.
"cattail
Lysiohiton kamtsohatense
Nelumbom sp.
Pontederia cordata L.
Sagittaria sp.
Sparganium sp.
Symplocarpus foetidus L.
Typha sp.
Mosher 1919.
Claassen 1921; Needham,
et al. 1928; Guppy 1948;
Jones 1951; Tietz 1952.
Comstock 1936; Forbes 1954;
Kimball 1965.
Guppy 1948; Jones 1951.
Tietz 1952.
Tietz 1952.
Tietz 1952.
Tietz 1952.
Tietz 1952
Crumb 1956.
d. Arzama brehmei B. & McD.
Typha angust folia L.
Forbes 1954.
2. Be llura gortynoides Wlk.
Nywphaea advena
Nuphar advctia Ait.
Robertson-Miller 1923.
Levine 1974.
a. Arzama vulnifica Grt.
"yellow water-lily"
Nuphar advena Ait.
Forbes 1954.
Levine 1974.
b. A. melcmopyga Grt.
Nuphar advena Ait.
"bonnett lily"
Comstock 1881.
Skinner 1903.
c. Hallara melanopyga (Grt.)
"water lily"
Nymphaea americana (Prov.)
Miller & Stand.
"pond lily"
Nuphar variegatwn Engelm.
Nuphar advena Ait.
Hampson 1910; Seitz 1923.
Welch 1914; Robertson-
Miller 1923; Needham,
et al. 1928.
Grossbeck 1917; Comstock
1936.
McGaha 1952.
McGaha 1952.
3. Bellura diffusa (Grt.)
"pond lily"
Comstock 1936.
a. Arzarria diffusa Grt.
"yellow water lily"
Forbes 1954.
4. Bellum densa (Wlk.)
a. Arzama densa Wlk.
Pontederia cordata
Eichhomia crassipes
(Mart.) Solms
Calocasia esculenta L.
Typha latifolla L.
Forbes 1954; Vogel &
Oliver 1969b.
Vogel & Oliver 1969a, b.
Habeck 1974.
Tietz 1952.
b. Bellura gortynoidea Wlk.
Typha sp.
Grossbeck 1917
(Kimball 1954).
5. Bellura garyantua (Dyar)
a. Sphida garyantua Dyar
Typha latifolia L.
Dyar 1913; Seitz 1923.
b. Arzama garyantua (Dyar)
Typha latifolia L.
Typha sp.
Comstock 1944.
Comstock & Dammers 1944;
Tietz 1952.

62
In general there seems to be three families of plants attacked,
the Typhaceae, the Nymphaeaceae, the Pontederiaceae. The Typhaceae
are infested by Bellura obliqua (Wlk.) (in Typha latifolia L. and T.
angustifolia L.) and Bellura gargantua (Dyar) (in T. latifolia L.).
The Nymphaeaceae are infested by Bellura gortynoides Wlk. (Nuphar
advena Ait. and N. variegatum Engelm.) and B. diffusa ("water lily").
The Pontederiaceae are infested by Arsama densa (Wlk.) (Pontederia
cordata L. and Eichhomia crassipes (Mart.) Solms). This supports Todd's
(pers. comm.) contention that possibly only three species are involved.
A fourth family, the Araceae, is strongly implicated within the
host range of this family. Guppy's (1948) record (Table 2, No. lc)
of Arzama obliqua (Walk.) from skunk cabbage (Lysichiton kamtschatcense=
L. camtschatcense = L. americanum Hult. & St. John, see Munz 1965) seems
to be well founded. Tietz's (1952) citation of Symplocarpus foetidus
L. (Table 2, No. lc) probably refers to Guppy's paper. Habeck's (1974)
record of Arzama densa Wlk. from dasheen (Calocasia esculenta L. =
Colocasia esculenta (L.) Schott; Table 2, No. 4a) also seems substan
tiated. These represent two instances of the infestation of two
different species of the Araceae from two widely separted regions
(British Columbia, Guppy 1948 and Florida, Habeck 1974) by apparently
two species of the Bellura complex. Takhtajan (1969) indicates
that there is a close affinity between the Liliales (Pontederiaceae),
Arales, and Typhales and they all are represented along a line of
evolution in common with the Nymphaeales.
Mosher (1919) stated that Arzama obliqua has been reported from
corn. She did not cite any references to these reports, however, and I
have not been able to substantiate this claim. Because crop plants such

63
as dasheen and corn have been implicated in the host range of this
group of insects a great deal of study of host specificity is warranted
and the taxonomic status of the group needs clarification. I do not
doubt these records but I am dubious of the placement of species identified
from these plants. On several occasions I have caged larvae of Arzama
densa (Wlk.) on both dasheen (Colooasia esoulenta Schott) and Xanthosoma
sp. and found that they did not feed upon them. Further studies are
severely needed to verify these host records.
Biology and Life History of Arzama densa Wlk. and Related Species.
The early literature on the biology of these species is sparse and
occurs primarily as notes of correspondence in various journals. The
first reference I have been able to find is that of Worthington (1878).
He described the larva of Arzama obliquata (G. & R.) and noted that it
was found "under the bark of a dead maple about three feet from the
ground, where it had made for itself an oval cavity in the dust". He
reared the adult and found that the pupal stadium was about 21 days
(April 27 May 18).
Comstock (1881) described the larva and aquatic habits of Arzama
melanopyga Grt. He was the first to take note of the large dorsally
situated pair of spiracles on the 9th abdominal segment which are
characteristic of the larvae of this group.
Riley (1883a, b) described the eggs of Arzama obliquata G. & R.
(misspelled Arsame) as being laid in "curiously broadly convex or plano
convex masses enveloped in hair, and a cream colored mucuous secretion,
when combined look much like spun silk on the inside, and on the outside
like the glazed exudation of Orgyia leuaostigma." He also noted the

64
large dorsally oriented last pair of spiracles. He stated that there
were two annual broods, the second of which hibernated in stumps or
moss near the water.
Kellicott (1883a, b), however, felt sure that in New York this
species was single brooded and pupated in May. He also noted that they
overwinter in the soil or old wood.
Riley (1883a, b), in reply to Kellicott's comments, stated that
there could be no doubt as to the digoneutic (=double brooded) nature
of A. obliquata at Washinton (D. C.?).
Comstock (1888) referred to the habits of Arzama (misspelled Argoma)
that infest the leaves of pond lilies. He distinguished these from
truly aquatic larvae in that they "are obliged to come to the surface"
for air.
Brehme (1888a) described the eggs, larva, and pupa of A. obliquata
(G. & R.). He noted a developmental period of about 15 days for the
eggs which were laid on cattail between the long leaves. He found the
larval period to be 161 days and the pupal period to be 16 days making
a total egg to adult span of 190 days.
Brehme (1888a) also stated that the larva returns to the top of the
reed in its later larval stages and forms it pupa there. This sparked
a series of correspondence in the Canadian Entomologist. Moffatt (1888a)
stated that this was not its invariable habit in nature and he had found
the pupa beneath the bark of a decaying stump some distance away from
where the cattails grew. Brehme (1888b) replied that this may not be
invariable but that the majority of them pupate in the reed. He cited
a friend of his who had found the pupa in a stump but indicated that the
larva had been feeding there and wondered if that wasn't true in Moffatt's

65
case. Moffatt (1888b) replied that there was no evidence that the
larvae had fed in the stumps and that all of the larvae and pupae they
collected were in similar situations, but admitted to not having looked
in the Typha reeds for want of a boat.
Kellicott (1889) referred back to the communications between Riley
and himself in 1883 and had decided that they were both right in that
A. obliquata G. & R. produces two broods in Washington and one in New York.
He also presented evidence confirming Moffatt1s contention they they
overwinter in stumps as larvae. Brehme (1889) later sent sections of
Typha stalk to Moffatt with numerous burrows and two pupae. Moffatt (1889)
subsequently reared a pair of the moths from this material. Brehme (1889)
felt that Kellicott and Moffatt were mistaken in their assertion that the
larvae overwinter in stumps because the specimens he sent to Moffatt were
collected in the winter below the water in cattail reeds and some were
even under ice. He also disagreed with Riley over the clustering mode of
oviposition. He noted that he had always found eggs laid singly and felt
that if it were otherwise it would be impossible for several larvae to
live in one reed. Johnston (1889) agreed with Brehme as he had found
abundant larvae and pupae in cattail in the winter in Ontario. He noted,
however, that he had also found them on shore in old wood. He proposed
that those on shore were merely wanderers. Beutenmilller (1889) described
the mature larvae of this species and indicated that he had found full
brown specimens under decaying stumps. He later (Beutenmller 1902)
described the larva again under the name of Bellura obliqua (Walk.).
Hampson (1910) repreated Comstock's (1881) description of the larva
of Avzama melanopyga Grt. under the name of Belluva melanopyga (Grt.).

66
Welch (1914) described the habits of Bellura melanopyga (Grote).
He described two feeding periods, first being the leaf feeding period
in which the young larvae mine the leaves of Nymphaea americana (=Nuphar
variegatum). The second stage is the petiole period which occurs after
the larva locates the midrib or the leaf-petiole junction and forms a
large burrow in the petiole. He also experimented with other host plants.
He found that in a starvation situation they would feed on white water
lily, Castalia (= Nymphaea) odorata but when given a choice preferred the
yellow water lily. Potamogetn natans and Sagittaria sp. were never
attacked. He also discussed respiration, locomotion, and the natural
enemies of this species.
Rummel (1919) noted that cattail plants bearing spikes were not
infested with Bellura obliqua.
Claassen (1921) pulbished a detailed account of the insects associated
with Typha. Included within this was an excellent study on the biology
of Arsama obliqua (Walk.). He found, in New York that there was only
one generation per year and that the full grown larva overwinters in
its burrow in the plant. He described the mode of oviposition to be in
masses, similar to the description given by Riley (1883a, b). He found
that each mass contained 35-60 eggs and each female produces ca 225 eggs
making about 6 masses per female possible. Upon emergence the first
instar larvae enter the leaf directly from the egg chorion. They feed
within the longitudinal I-shaped partitions within the leaf mining
downward. At the second molt they apparently become too large to feed
within these partitions and emerge from the leaves and seek shelter behind
the sheath of one of the outer leaves. They ultimately disperse, each
to find its own plant, and become solitary burrowers in the stem and
rhizome. These,then, exhibit two phases similar to those of Welch (1914)

67
although these should probably be called the leaf mining phase and the
stem phase (rather than the petiole phase). He noted that the length
of the pupal period averaged 17.6 days and described the egg, first
instar larva, full grown larva, pupa, and adult (from Walker).
Robertson-Miller (1923) published many observations on the biology
of Bellura gortynoides Wlk. and B. melanopyga Grt. Her information did
not differ much from that of Welch (1914). She described the larvae of
each and indicated that they did not appear to be much different. She
described the egg masses and indicated that those of B. gortynoides
were deposited in flat mats of about 20 eggs each. She noted that
some of the eggs were covered with silvery white threads. The masses
of B. melanopyga were similar to those of B. gortynoides. She found that
B. gortynoides may pupate in the petiole, in soil, or in wood. When
in the petiole the pupae of B. melanopyga was at the top of the burrow
while those of B. gortynoides were lower down. She also found that
B. gortynoides would feed of pickerel weed (Pontederia cordata) in
captivity.
Needham et al. (1928) repeated the observations of Claassen (1921)
on Arzama (= Bellura) obliqua (Wlk.) and of Welch (1914) on Bellura
melanopyga Grt.
Comstock and Dammers (1944) described the full grown larva and pupa
of Arzama gargantua Dyar. I see no distinction between this description
and previous authors' descriptions of these stages of Bellura obliqua (Wlk.).
Guppy (1948) described the habits of Arzama (=Bellura) obliqua
attacking skunk cabbage (Lysichiton kamtschatcense) on Vancouver Island,
B. C. He also indicated that they overwintered under loose bark on
fallen logs.

68
Crumb (1956) provided a key to the larvae of the Amphipyrinae. The
couplet separating Arzama used the large sub-dorsal spiracles on the 8th
abdominal segment as a key character. He also described the larva of
Arzama obliqua (Wlk.).
Vogel and Oliver published two papers (1969a, b) on Arzama densa
Wlk. Their first paper was on the potential of this insect to control
waterhyacinth. Their second paper was on the life history of A. densa.
They provided cursory descriptions of the immature stages and determined
the developmental times of the various stadia. Much of their data,
however, is from larvae reared on artificial media which makes their
results subject to question. These two papers will be further discussed
later in this dissertation.
Levine (1974) found that in Indiana there were two complete genera
tions per year of Bellura gortynoides Wlk. (= B. vulnifiea Grt.). He
found that the first generation (spring) pupates within the petiole of
Nuphar advena. The second generation (fall) larvae swim to shore and
overwinter as larvae under the bark of trees, in rotten wood, or in
leaf litter. The eggs hatch in 6 days and there are 6 to 7 instars.
Parasites, Predators, and Diseases.
The first record of natural enemies which attack this group of
insects was that of Welch (1914) for Bellura melanopyga Grt. He noted
that sunfish ate the larvae when they were swimming on the surface. He
also observed water striders (Gerris sp.) attacking the larvae when they
were on the surface of the leaves.
Claassen (1921) found Sturmia nigrita Town. (Diptera:Tachinidae)
parasitizing the larvae of Arzama obliqua (Wlk.). Robertson-Miller (1923)

69
found puparia of Masieira senilis associated with the burrows of Bellura
gortynoides Wlk. Both of these names are probably synonyms of Lydella
radiis (Town.) (Stone et at. 1965).
Comstock (1944) made note of the fact that he found no parasites
associated with Arzama gargantua Dyar in California.
In the Thompson catalogue (1944) two parasites are listed from
Arzama obliqua (Wlk.). The first is Ceromasia senilis Mg. which may be
a misidentification of Lydella radiis (Town.). The second is Pimpla
roborator F. (=Exeristes) which is an ichneumonid. I question the
veracity of this latter records, however, because the range is listed
throughout Europe, Japan, and Guam. As far as I have been able to as
certain the Arzama Bellura group is strictly New World.
Vogel and Oliver (1969b) listed several parasites and predators of
Arzama densa Wlk. They identified Lydella radiis (Town.) from the larvae,
Ichneumon n. sp. and Eupteromalus virideseens (Walsh) (Hymenoptera:
Pteromalidae) from the pupae, Telenomus arzamae Riley (Hymenoptera:
Scelionidae) and Anastatus sp. (Hymenoptera: Eupelmidae) from the eggs.
They also found Coleomegilla maeulata De Geer larvae (Coleptera:
Coccinel 1 idae) preying on the eggs and young larvae, and Phyllopalpus
pulehellus (Uhler) (Orthoptera: Gryllidae) and Chlaenius pusillus Say
(Coleptera: Carabidae) preying on the larvae.
Levine (1974) indicated that the eggs of first and second generations
of Bellura gortynoides Walk, are also parasitized by Telenomus arzamae
Riley and the second generation larval populations are parasitized by
an ichneumonid and have a polyhedrosis virus.

CHAPTER I
THE RELATIONSHIP BETWEEN THE PHENOLOGY AND PRODUCTIVITY OF
WATERHYACINTHS AND VARIOUS PHYSICAL AND BIOLOGICAL FACTORS.
Introduction
To evaulate the effects of insects for the biological control of
weeds, a basic understanding of the ecology of the plant is essential.
In realization of this the Canada Weed Committee has instituted a series
on the biology of Canadian weeds (Cavers and Mulligan 1972). This is
an attempt to pull together all the available knowledge on the biology
of Canadian weeds that can be used in weed control efforts. Within this
framework the phenology of the plant (annual variation), and the response
of the plant to limiting factors and damage by indigenous insects is of
special interest for the evaluation of biological control attempts.
Omission of these considerations could result in the misinterpretation
of pertinent data. For example, natural seasonal declines in the plant
population could mistakenly be attributed to insect releases when the
insects are also seasonal if patterns of seasonal variation of the plant
are not known. Also, releases of insects may be more effective when
correlated with critical periods in the annual cycle of the plant.
Judgements for the timing of these releases can be made only on the
basis of what is known about the plant.
Limiting factors can be defined as the necessary components of
the organism's environment which are least available and thereby con
trol the life processes of the population. Liebig (1840) stated that a
process is limited by the quantity of a single component present in
minimal amounts relative to its optimal amounts. Sachs (1860) felt that
biological processes required a certain minimal level of a limiting
70

71
factor to begin, attained an optimum at a certain level, and declined
as levels of the limiting factor exceeded maximum tolerable levels.
This is parallel to Shelford's (1913) "Law of Tolerance" where he
essentially states that the failure of an organism may be due to an
excess or deficiency of any one factor which may approach the maximum
or minimum limits of tolerance of the organism for that factor. For
an aquatic plant, such as waterhyacinth, these limiting factors include
temperature, light, water, dissolved or available nutrients, space, etc
Phytophagous insects probably cause a threshold type response in
the plants whereby the plant can sustain certain levels of damage without
obvious deterioration until maximum tolerable limits are exceeded. As
insect damage exceeds these threshold levels a rapid decline in the
(
populations or standing crop may be evident. Levels of insect damage
)
below this threshold may cause various plant response. When the popu-
J
lation is at steady state (the stable maximum level restricted by the
level of a limiting factor) insects may disrupt this stability causing
the plant population or standing crop to fall below the carrying capacity
of the system. This may have the effect of reducing intraspecific com
petition in the plant population. In this case the limiting factors
would become increasingly more available and production may indirectly
be stimulated. Hence yield may be increased under low insect concen
trations where the insects prevent senescence of the population by
increasing the rate of turnover.
This study was designed to measure the effects of various environ
mental factors as well as the effects of a natural buildup of an indi
genous insect population (Arsama densa) on a stand of waterhyacinth.
The parameters considered can be grouped into climatological conditions
(temperature and solar radiation), limnological conditions (nutrients,

72
water quality and water level), intraspecific conditions (plant density,
canopy effects, available space, etc.)* and biotic stress (insect damage).
These will be evaluated with possible interactions between them considered.
These concepts, possible interactions and all factors which control
the plant must be considered and investigated. Attempts to evaluate the
attack of an insect by studying only the insect or with only a super-
ficial knowledge of the target plant are subject to erroneous conclusions
and misinterpretation. Not only must the plant and the insect be studied
but the insect-plant interrelationships must be established. This field
has received increasing attention lately and may provide a basis for
future biological control efforts.

73
Methods And Materials
Duirnal Waterhyacinth Productivity
Waterhyacinths of two distinct morphological types from the "open"
side of the catwalk on Lake Alice were selected for in situ metabolism
studies. Large plants, approximately 90 cm tall, with elongate petioles
were measured for C02 uptake on 11-12 August 1973. Small plants (<50 cm)
with bulbous petioles were measured on 12-13 August 1973. A section of
the mat approximately 0.5 m2 of each type was placed under a chamber
constructed of a PVC pipe frame covered with clear poly-acetate. The
base of the chamber was 71 cm x 71 cm (ca 0.5 m2).
Air was passed through the chamber with a blower and duct system.
The duct entered the chamber at the base on one side. Air was supplied
to the blower intake through a tube opening approximately 3 m above the
water surface so the C02 concentration would not be influenced by the
plants surrounding the chamber. The rate of air flow was determined
with a Hastings hot wire anemometer. The air was discharged from the
chamber through a duct located on top.
Carbon dioxide concentrations were monitored at the chamber air
intake duct and at the exhaust duct. The air at each location was col
lected through tubes which extended to a Beckman infra-red C02 gas analyzer.
Air flow was also measured at the intake and exhaust. This enabled the
determination of the ppm C02/unit of air/time entering and leaving the
chamber. The differential is the amount of CO2 produced or consumed
within the chamber.
The CO2 analyzer readings had to be calibrated against a standard

74
to convert from a scale reading to ppm CO^. The scale reading is based
on a comparison of two gases. Three pairs of gases were compared through
the analyzer. Ambient air vs. ambient air (air entering the chamber) was
compared to determine a zero point. The second comparison was chamber
exhaust vs. ambient air. This difference represented the CO2 gradient
through the chamber and was expressed as recorder scale division. The
value of a scale division (sd) is determined according to the level of
C02 in the ambient air by the equation ppm/sd = ae^x where x is the C02
concentration of the ambient air. The ambient air CO2 concentration
was determined by a third comparison. In this case a standard was used
of a known concentration. The standard was 300 ppm bottled gas and was
compared against the ambient air. The CO2 concentration in the ambient
air was determined by the equation ppm = ax2 + bx + c where x is the
recorder reading. This involves lowering the amplification of the analyzer
output by changing from "range 3" to "range 1". The range 1 equation is
calculated by running the standard 300 ppm gas through the reference
side of the analyzer and running other gases of known concentration through
the sample side. The value of the sample gas is correlated with the
recorder reading using the parabolic regression. Range 3 is calibrated
using various known C02 concentrations against a closed system aparatus
with flow and pressure maintained constant. The closed system is injected
with known quantities of pure CO2. An exponential regression is fitted
for the ppm/sd against ppm C02 of the various reference gases (ambient
air in this case).
The volumetric C02 concentration gradient (ppm C02) is converted
a gravimetric measurement (g C/m2) using the gas constant (0.14625 gm C-K/m2

75
atm ppm C02). When multiplied by the flow rate this expression yields the
rate of carbon metabolism (g C/hr) within the chamber. A more detailed
explanation of this system is given by Carter et al. (1973).
Carbon metabolism for each type of plant was measured for 24 hrs.
Integration of the resultant production curves yielded both gross primary
productivity and respiration. Respiration was assumed to be constant both
day and night and was determined as the average nighttime value. Net
production consisted of that portion of the curve above the compensation
point (where Pg = R and = 0). solar radiation was measured with a
Weathermeasure Co. 24-hr. pyroheliograph in the 0.36-2.5um range. Air
temperature was recorded using a Yellow Springs Instrument thermistor
apparatus.
Following the metabolism measurements the plants were harvested to
obtain a biomass estimate. The total sample was divided into leaves,
petioles, roots (= roots + rhizomes + stolons) and detritus and the
various plant parts were weighed while fresh. A similarly divided sub
sample was taken and weighed before and after drying. From this subsample
a wet to dry conversion factor was obtained so that the dry weight for
each plant part and the total sample could be obtained. A subsample of
the leaves (pseudolaminae) and petioles was pressed in a plant press and
dried. The outlines of the dried leaves were traced on paper and the
area measured with a planimeter. This determined a leaf area per gram
of dried leaf conversion factor and the leaf area of the total sample
was estimated from this. A similar procedure was employed with the
petioles. From this the leaf area index (LAI) was determined which, in
this case, is the total leaf area (leaves + petioles) per square meter
as determined from only one side of the leaf.

76
Annual Cycles and Insect Damage
Estimates of various plant characteristics, of the Arzama densa
population, and of plant damage by A. densa were taken on a weekly basis
from May 1974 to 30 April 1975. Sampling was done on a plot system
using a rubber ring enclosing an inside area of 0.316 m2. The samples
were taken each week in a pseudo-random manner. I have not been able
to devise a satisfactory system of pinpointing a previously randomly
selected point on a mat of waterhyacinths and then finding that point
while trying to maneuver through the dense stand of plants. To elim
inate the additional variables of seasonal plant species composition
changes along the shoreline and different waterhyacinth growth charac
teristics only the central area of the lake was studied. The area in
which samples were taken was defined by the catwalk on the west side
and extended 25 m to the north and 25 m to the south of the central
point on the catwalk. The eastern boundary was established by a small
row of bushed 50-60 m from the catwalk that extended into the lake from
the north shore. The study area, then, was 2500-3000 m2 in the central
more or less homogenous region of the waterhyacinth mat on the marsh
side of the catwalk. Sampling points were selected by throwing the ring
in a high arc. After it fell into the mat it was reached using two Dow
styrofoam billets (3 m X 0.5 m), one placed in front of the other sequen
tially. This allowed me to move (with some effort) over the mat on the
water surface. Once the ring was reached the billets were used as plat
forms for counting and recording data.
The ring was manipulated down over the waterhyacinths until it was
on the water surface. This involved making subjective decisions as to

77
which plants were inside and which plants were outside the sample. If
the crown was in the ring the whole plant was considered in. This was
still difficult to determine when the crown straddled the edge. The
placement of these plants was left up to the discretion of the sampler.
This border effect was probably the largest within sample source of
error in the plot sampling.
Once the ring was placed each plant was withdrawn and measured.
An offset was considered a separate plant only if the root system was
developed. Measurements, taken were the height of the plant, based
on the distance from the tip of the longest leaf to the point where it
attached to the rhizome, and the number of leaves per plant. Leaves
were counted only if half or more on the pseudolamina was alive and
unfurled. The number of plants in each plot was tallied to establish
plant density. Each plant was carefully dissected and damage by Arsama
densa noted. Damage was distinguished according to the degree of sever
ity. Leaf damage was classified as to external feeding or petiole bores.
Rhizome damage was classified as tip damage, rhizome bore, or rhizome
fragmented. If a larva or parasite was found it was placed in a pill
vial, given an identifying number and returned to the laboratory. The
insect data will be discussed in a separate section of this dissertation.
One sample required 3-4 man-hours. Three samples were taken each week.
Leaf area estimates were also taken weekly but on a different day.
Ten plants were randomly selected along the catwalk for this measurement.
The randomization procedure consisted of numbering the supporting pilings
of the catwalk within the study area. Ten pilings were selected from
generated random numbers. At each selected piling a single plant was

78
picked. This required a second randomization. One person involved in
the process held a second generated random number between 1 and 10. A
J_ L
second person drew up to ten plants out of the water. When the nLr plant
(where n = the random number) was pulled out the first person notified
the second and the plant was placed in a plastic bag and returned to
the laboratory.
Before measuring the leaf area the petioles and leaf blades (pseudo
lamina) were separated. The petioles were rolled out with a rolling pin.
This was necessary to compensate for the cylindrical shape of the petiole.
The outline was then traced on a piece of drawing paper and measured with
a planimeter. The leaf blades were pressed in a plant press and dried.
They were then traced and measured the same way. I found that drying
the leaves caused considerable shrinkage and a dry:fresh conversion factor
had to be employed. The formula for this conversion was:
leaf area (fresh) = 1.437 X leaf area (dry)
The conversion factor was determined by measuring one sample (51 leaves)
before and after drying. Each week figures for the average leaf (pseudo
lamina) area, average petiole area, and average total (pseudolamina +
petiole) leaf area were derived. These figures were multiplied by the
number of leaves per square meter from the plot samples to obtain the
leaf or petiole area index. This figure represents the leaf or petiole
surface area (considering only one side) per unit of substrate area (m2/m2).
Water samples were taken along the catwalk at the mid-point. The
water was collected a few centimeters below the surface at the level of
the waterhyacinth roots. This level should best reflect the conditions
the plants were being subjected to. The sampling station was on the
downstream si of the study area so the tests would reflect minimum

79
nutrient levels. Two samples were collected each week. One sample
was analyzed using a Hach DR-EL portable test kit for total alkalinity
(carbonate + bicarbonate), total nitrates + nitrites, pH, total phosphates,
and sulfates and a Hach micro-iron test kit (model IR-18-A) for iron.
The second sample was taken to the University of Florida Soils Laboratory
where it was analyzed for conductivity, magnesium, and potassium.
The methods applied to the water samples are as follows:
Alkalinity (total) Titration of Brom Cresol Green Methyl
Red indicator with 0.020 N. sulfuric acid 10 ml sample.
Conductivity Platinum electrode ohmmeter.
Iron 1, 10 Phenanthroline Method 25 ml sample.
Magnesium Atomic absorption spectrophotometer.
Nitrates and Nitrites (total) Cadmium reduction method -
25 ml sample.
pH Colorimetric reading with a wide range indicator.
Phosphates (total) Colorimetric method 25 ml sample.
Potassium Flame emission spectrophotometer.
Sulfate Turbidimetric method.
The procedures employed in the Hach Test Kit are more for convenience
and direct reading and are not as accurate as other techniques. For my
purposes the loss in accuracy is outweighed by simplicity of the proce
dures. These procedures are probably accurate enough to indicate temporal
differences but are probably not extremely definitive.
Maximum and minimum air and water temperatures were taken at the
same location as the water samples. Two Taylor (No. 5458) maximum-minimum
self registering thermometers were mounted on a C-shaped styrofoam

80
block. The water temperature thermometer was placed vertically on the
bottom of the block. The air temperature thermometer was placed vertically
in the concave side. The block was mounted on a rider which slid over a
piece of pipe which extended into the lake bottom. This allowed the
thermometer to move up or down as the water level changed. The bulb of
the underwater thermometer was about 4 cm below the surface and measured
the conditions the submersed plant portions were subjected to. The air
thermometer bulb was about 30 cm above the water surface and measured
conditions under the leaf canopy. The concave side of the block was
oriented towards the north so as to avoid direct exposure to the sun.
The overhang on the block also helped prevent this.
Water level was measured at the northwest corner of the study area
from a depth gauge established there previously by other investigators.
Solar radiation data was obtained from Dr. E.A. Farber of the solar
energy laboratory at the University of Florida.

81
Site Description
Lake Alice is located in the southwest corner of the University of
Florida campus in Gainesville, Alachua Co., Florida (Topographic desig
nation: Gainesville East quadrangle, TIOS, R19E, N£?s, R20E, NW^). The
lake was once a sinkhole fed by a small stream but damming off the west
end in the late 1940's and later the addition of effluents from the
campus sewage treatment plant and the heating plant resulted in its pre
sent configuration (see Figure 1). The lake area is approximately 33 ha
and is divided into a marsh dominated by waterhyacinths and an area main
tained by the University as open water. The marsh at the east end com
prises approximately 65% (21 ha) of the lake surface and is separated
from the open lake by a catwalk and fence constructed to retain the
waterhyacinths. The depth of the marsh in generally less than 2 meters
(Cason 1970). The "open" western end of the lake covers about 12 ha
and is also generally less than 2 meters in depth with a few areas of
about 5 meters, probably the original sinkholes (Mitsch 1975). The
general flow of the lake is from the sewage plant and heating plant
effluent at the eastern end through the marsh to the open lake at the
western end where it discharges through two wells into the Florida
aquifer.
The lake is situated on Ocala limestone which is dominated by a
karst topography. Solution sinkholes, fractures, and caverns are typical
of this type of topography and are common in this area. Because of the
silt that has accumulated on the bottom, however, the lake basin is
maintained above the local water table (Cason 1970). The lake level is
generally between 68 and 70 feet above mean sea level. Figure 2 indi
cates the lake level at the catwalk for the period of this study.

Figure 1. An aerial view of Lake Alice on the University of Florida campus.
The white square designates the study area. Notice the darker color
of the waterhyacinths in the main flow of the sewage effluent.

83

84
As mentioned previously Lake Alice receives effluent from the campus
sewage treatment facility and cooling water from the heating plant. The
nutrient enriched water from the former and the warm water from the lat
ter have probably contributed significantly to the eutrophication of
this lake. Other sources include overflow from Hume Pond, also located
on the university campus, runoff from the local watershed, and direct
rainfall. Discharge of the system is through the wells mentioned earlier.
Water loss also occurs through surface evaporation and evapotranspir-
ation. Mitsch (1975) estimated the hydrological budget for the lake in
terms of flows and storage (see Table 3). The water storage at a stage
of 69 feet above mean sea level is estimated to be 254 x 103 m3. Water
retention is low due to the high input-volume ratio. Brezonik et aZ.(1969)
suggests that this may have a flushing effect causing low phytoplankton
populations noted in the lake.
Discharge through the wells is regulated by valves and is frequently
altered by campus personnel. The water level is often raised to facili
tate mechanical removal of the waterhyacinth. During the hurricane season
the water level is dropped to prevent flooding. Fluctuations are also
caused when the discharge screens over the wells become clogged with
debris. Water level appears to be correlated with seasonal precipitation
patterns (Fig. 2) except for the months of December and January. During
this time an oil spill occurred in the canal from the heating plant and
sewage treatment facility. Flows from these two sources were minimized
so the oil could be cleaned up. This resulted in a sharp drop in the
lake level. Normal flow was restored the first part of February and a
sharp increase in the lake level followed.

85
Table 3. Hydrological budget for Lake Alice (March September
1973). (From Mitsch 1975)
Source
Input*
Output*
Sewage Flow
10.5

Heating Plant Flow
43.5

Hume Stream
8.6

Direct Rainfall
0.8 2.0
--
Runoff
6.1 21.0

Transpiration

1.8 3.1
Evaporation

0.4 0.6
Discharge

58.0 76.8
TOTAL
69.5 85.6
60.2 80.5
*Flow x 103 m3/day

Figure 2.
Water level taken at weekly intervals and precipitation at
Lake Alice from July 1974 through June 1975.

LAKE LEVEL (cm)
o S
/

(mo) mvdNivt

88
Analyses
Multivariate analyses were performed on the annual data in an attempt
to account for observed variation in the plant characteristics in terms
of the various environmental parameters. A stepwise regression procedure
(SAS STEPWISE) was first employed to determine which linear combination
of independent variables would provide the best fit for the actual data.
The dependent variables analyzed were standing crop, plant height, plant
density, leaf density, average leaves per plant, and leaf area index.
Each was regressed against solar radiation, minimum air temperature,
maximum air temperature, minimum water temperature, maximum water temp
erature, % rhizome damage by Arzama densa, % leaf damage by A. densa, all
nine water quality parameters, and water level. Three procedures were em
ployed to determine the best regression equation. These were the forward
selection procedure, the backward elimination procedure, and the stepwise
procedure. All form linear models in different ways (see SAS manual, 1972
for further explanation). Each provides an analysis of variance, regression
coefficients and statistics of fit for the model. In all cases the step
wise procedure provided the most significant fits to the data.
Two additional variables were entered which were derived from other
variables. Since it was assumed that the amount of light intercepted by
the plants would be inversely proportional to the degree of self shading
the solar radiation variable was divided by the leaf area index to form
a new variable. This was entered into the analyses in place of incident
solar radiation but it failed to increase the significance of any of the
linear combinations (i.e., incident solar radiation was just as good or
better). It was also speculated that available space may contribute to

89
growth. A variable for space was formed which was merely the inverse of
the leaf area index and added to the data. In many cases this did increase
the significance of the regression equations. Upon further consideration
it became apparent that the inclusion of this variable was not valid since
it was derived from a dependent variable. Available space would obviously
be greater when the standing crop is at a minimum and would certainly be
inversely correlated with it. An increase in space may very well increase
the rate of growth but rates were not being analyzed. The dependent varia
bles represent the state of each parameter which is regressed against the
corresponding states of the independent variables. For these reasons
space and shading were excluded in the final analysis.
Once the model for the best fit was derived it was entered into a
regression program (SAS REGR Procedure). This program provided the same
analysis as the stepwise procedure but in addition it predicted values
for the dependent variables based on the multivariate model with each set
of independent variables.
The linear model assumes orthogonality between independent variables.
Since many of the parameters were inter-related (i.e., sun and air temp
erature, rhizome damage and leaf damage, etc.) the assumption of orthogonality
is violated. In evaluating a model the procedure may select the parameter
which best reduces variability and important correlated variables may be
lost. For this reason a correlation procedure (SAS CORR Procedure) was em
ployed to determine which variables were significantly correlated. As a
result, if a factor is found to significantly contribute to variation in
the dependent variable it can be cross checked in the correlation matrix
to determine what other variables may be working with it. The dependent
variables were also checked against one another for significant correlations
using the same procedure.

90
A major weakness in the use of a multivariate linear model in this
type of study is that it assumes independence between independent variables.
In actuality probably very few of the variables are completely independent.
For example sunlight and nutrient levels are both assumed to be linearly
related to the state of the variable for standing crop. It is further
assumed that each contribute independently and in an additive fashion.
This is not true, however, as variables such as nutrient loads and solar
radiation interact multiplicatively and the effects of one are limited by
the state of the other (see H. T. Odum 1974). This interaction may be
linear, exponential, or logistic depending upon whether or not either is
present in limiting quantities. The linear model does not account for these
complex relationships and should not be considered as a basis for making
generalizations about interrelationships in the system. I feel, however,
that this type of analysis can provide an indication of which variables
are important in accounting for variation but it cannot be interpreted as
a mathmematical expression of the functions of these variables (i.e., the
coefficients in the model have no real meaning).

91
Results
Water Quality
Lake Alice has been variously classified as eutrophic, highly eutrophic,
and senescent (Brezonik 1969, Brezonik et al. 1969). Brezonik et al. (1969)
reported that lake nutrient levels do not reflect the nutrient enrichened
sewage source that feeds the lake. They postulated that this was due to
ability of the waterhyacinths to absorb these nutrients. This was substan
tiated by Mitsch (1975) who found that various nutrient concentrations
decline in the direction of flow across the marsh. Brezonik et al. (1969)
also found that phytoplankton counts were extremely low and suggested
that this may be the result of light blockage by the waterhyacinths as
well as a flushing effect of the large volume of cooling water from the
heating plant. Table 4 lists various water quality ranges for the lake
from Mitsch (1975), Brezonik et al. (1969) and from this study. The former
two investigators followed standard procedures as established by the
American Public Health Association (1965). I did not have the facilities
to follow these procedures and used a water chemistry test kit for pH,
nitrogen, phosphates, sulfates, iron, and alkalinity. The samples were
further analyzed for conductivity, potassium, and magnesium by the
University of Florida Soils Laboratory. In spite of this divergence in
technique and accuracy the values from my study compare favorably with
those from the other studies. However, my values are not definitive and
are only general indices of the parameters measured.
Figures 3-7 illustrate seasonal changes in the water quality
parameters measured. Because of the many factors affecting nutrient loads
it is difficult to explain the variations observed. The amount of waste

92
Table 4.
A comparison of water quality measurements from Lake Alice, Gainesville,
Florida with previous reports.1
Source of Data
Time Period
Point Sampled
Brezonik et al. (1969)a
Dec 1968
Composite of 3 Stations
Mitsch (1975)b
Jan Sept 1973
Marsh Discharge
This Study0
July 1974-June 1975
Marsh Discharge
Sp. Conductance
533
379 484
464.53 (61.37)
Alkalinity
178
133 182
175.96 (24.27)
pH
7.4
7.6 8.3
7.33 (0.25)
nh3 (n)
0.01
0.053 0.184
...
NO3 (N)
0
0.008 0.938

no2 (N)
...
0.01 0.03
...
N02 + NO3

...
0.99 (0.62)
Ortho-P
0.07
0.73 1.00
...
Total-P
0.59
0.86 2.10
1.12 (0.45)
Chloride
16.7

...
Sodium
16.0
15 19
...
Calcium
71
14-65
...
Iron
0.01
...
0.097 (0.073)
Potassium

0.5 6.5
3.97 (2.02)
Sul fates
...
103.9 (38.1)
Magnesium
...
5.5 16.0
11.35 (3.38)
* Single determination.
b Range based on monthly averages.
c Mean (standard deviation) of all samples for the time period.
1 Units in mg/1 except for conductance (umho/cm), pH (pH units), and alkalinity
(mg/1 as HC03 in Brezonik and Mitsch and as HCOj + COf in this study).

93
treated varies with the number of students present at the university at
different times of the year, output from the sewage treatment plant varies,
the amount of water (hence, the diluting effect) from the heating plant
varies, rainfall varies as does runoff from near fertilized farmland.
All of these factors affect nutrient concentrations. Since the samples
were taken from the study area, however, they should reflect the relative
conditions the plants were growing in at the time. The reasons for differences
in the nutrient loads are not important but the differences themselves are.
Alkalinity (Fig. 3) generally ranged from 140 to 210 mg/1. Concen
trations were fairly constant. The greatest fluctuations occurred in
December when a sharp drop was apparent and in January when an equally abrupt
increase occurred. The concentration at the end of the study (11 June 1975)
was somewhat higher than that at the beginning (20 June 1974). Alkalinity
is a measure of the buffering capacity of the lake and in indication
of eutrophication.
Conductivity (Fig. 3) ranged from 340-580 ymhos/cm and is a measure
of the electrical conductance of the water resulting from soluble salt
concentrations. The conductivity values fluctuated greatly throughout
the year. In general periods of low conductivity seemed to correspond
to periods of frequent rainfall. This is presumably due to a dilution
of the soluble salts present relative to the water storage of the lake.
Iron (Fig 4) was measured because I have observed in greenhouse
cultures that waterhyacinths grow much better in nutrient solutions
high in iron concentrations. Further, it appeared that the iron in the
solutions was rapidly absorbed by the plants. Iron is a constituent of
cytochromes and as such is an essential micronutrient for plant growth.
Concentrations in Lake Alice were less than 0.10 mg/1 most of the year

Figure 3
Total carbonate and bicarbonate alkalinity and conductivity of
Juneri975P The^i611 frm Uke A1ice from July 1974 through
June 1975. The lines represent 5-point moving averages.

ALKALINITY (ppm CO, HC03)
cn
CONDUCTIVITY (fnmhos/cm)

Figure 4. Magnesium and total iron from Lake Alice water samples.
Note the sharp increase in iron concentrations in April.
The lines represent 5-point moving averages.

MAGNESIUM
97
MAGNESIUM (ppm)
O u"> O u-j
(wdd) NOdl

98
but a curious sharp rise occurred in April where the maximum concentration
of 0.40 mg/1 occurred. Concentrations dropped in May and returned to
initial levels in June.
Magnesium is an important structural component of the chlorophyll
molecule and is also used by plants in the metabolism of carbohydrates.
It is therefore also an essential element for plant growth. Magnesium
levels remained fairly constant throughout the year usually ranging
between 9 and 17 mg/1 (Fig. 4).
Nitrogen is a major nutrient for plants and becomes available in
the form of nitrates. Figure 5 illustrates the values for the sum of
nitrates for the study period. Since nitrites are usually fairly low
the curve probably gives a fair representation of relative nitrate values.
The range of total nitrates and nitrites was between 0.3 and 3.5. The
lower values occurred in July through October and increased through the
winter. A decline began in January and continued through March. Concen
trations began to increase gradually in the spring and by June the values
were higher than those from the previous year. Nitrate concentrations
may be inversely related to water level as similar but opposite trends
are noted in Figure 2.
Figure 6 shows the annual variation in pH measured over the study
period. Values did not vary much usually ranging between 7.0 and 7.7. A
decrease was noted in late December and again in April and May. At these
times the pH value became as low as 6.9. Maximum values occurred in
November and January when they reached 7.8. Except for the spring data
pH seems to parallel total alkalinity.

Figure 5
Phosphorus concentrations present as phosphates and nitroqen
concentrations as total nitrate and nitrites from water
samples taken from Lake Alice. The lines represent 5-point
moving averages. w

NITROGEN
PHOSPHORUS
(Wdd) SrUdOHdSOHd-aiVHdSOHd
n

Figure 6. The negative logs of the hydrogen ion concentration (pH)
of water samples taken from Lake Alice. The line is a
5-point moving average.

pH UNITS
ZOL

103
Unfortunately orthophosphate-phosphorus was not determined in this
study. Figure 5 represents the phosphorus concentration present as total
phosphates. While this does not define orthophosphate availability it
probably does give an index of it. Orthophosphate is the only form of
phosphate derived from natural sources and is present in organic wastes
and fertilizers (Vernon 1969). Hence, it is probably the predominant form
in Lake Alice. Phosphate-phosphorus concentrations were lowest in July
and August 1974 (ca. 0.5 mg/1) and increased through December. Concen
trations remained high into May and June 1975 and failed to return to
the 1974 levels. This may have been an artifact of the lake level as it
also failed to return to its previous July level. Also an increase in
phosphate concentrations in April and May appear to correspond to a drop
in the water level at the same time.
Phosphorous is a primary factor limiting production in aquatic eco
systems. Haller et al. (1970) found that P-concentrations below 0.1 mg/1
were limiting to waterhyacinth growth. Above this concentration the plants
absorbed P in luxury amounts. P-concentrations remained above this cri
tical concentration throughout the year in Lake Alice. I assume, therefore,
that P does not become limiting to waterhyacinth at this site. The fact
that phosphorous is lowest when plant biomass is highest probably reflects
the increasing absorption of these luxury amounts by an increased plant
standing crop. Mitsch (1975), however, found in a model simulation that
phosphorous concentrations in Lake Alice appeared to be unaffected by
waterhyacinth uptake or any other annual cycle. He further found that
there was very little decrease in phosphorous in the direction of flow
across the waterhyacinth marsh. He hypothesized rather that nitrates may

104
be more limiting in this system than phosphates since stronger winter
peaks and summer minima were apparent in annual nitrogen cycles and signi
ficant drops in the nitrate concentration occurred across the marsh.
Dunigan et al. (1975) found that phosphate-phosphorous concentrations
above 50 ppm were not significantly affected by the growth of waterhyacinths
but both nitrate-nitrogen and ammonia-nitrogen were. They further found
that the N:P ratio of uptake rates was 5-6:1.
Potassium concentrations (Fig. 7) remained fairly constant and did
not show any strong annual variation. Values between June and November
generally ranged between 2 and 4 mg/1. Concentrations increased somewhat
from December through February and an abrupt decline occurred in March
and April By June potassium concentrations were about the same as they
had been the previous year.
Sulfates (Fig. 7) were extremely variable but a bimodal tendency
was observed. Concentrations appeared to be maximum in the fall and
spring and minimum in the winter and summer. Since plants take up rela
tively little sulfate compared to the amount available (Ruttner 1972)
this pattern is probably due to factors other than waterhyacinth growth.
Temperature and Solar Radiation
Figure 8 indicates the weekly maximum, minimum, and median air and
water temperatures at Lake Alice during this study. Knipling et al. (1970)
found that maximum growth of waterhyacinths was favored at water temperatures
of 28-30C. Water temperatures greater than 40C were lethal and growth
decreased linearly as temperatures were reduced to 15C. In comparisons
of plants exposed to 30C days:30C nights with plants exposed to 30C
days:10C nights they found that the plants exposed to the lower nighttime

Figure 7. Potassium and sulfate ion concentrations of water samples
taken from Lake Alice. The lines represent 5-point moving
averages.

POTASSIUM
SULFATES
106
*
(lUdd) wniSSV-LOd

Figure 8. Maximum, minimum, and median weekly air and water temperatures
at Lake Alice from late June 1974 through June 1975. The
winter air temperatures were unseasonably warm during this
study period.

o
00

109
air temperatures had photosynthetic rates 15-30% lower during the day.
They also found that starch accumulation in the cholorplasts of 10C plants
was 2.5 times greater than the 30C plants. They attributed this to the
failure of the plant to translocate the previous day's starch accumulation
from the chloroplasts.
During the period of this study water temperature ranged between 10C
and 32C except for two weeks in January where the weekly minima were
recorded as about 4C. The median water temperature was never less than
12C nor more than 31C. The highest maxima occurred in late August and
early September and the lowest minima occurred in December and January.
The optimum median temperatures (28-30C as determined by Knipling et al.)
occurred in June through October but median temperatures above 20C occurred
from mid-March through late November. It may be concluded that water temp
eratures were generally favorable for waterhyacinth growth most of the
year although winter lows in December and January probably hindered growth.
The lethal limit (maximum) of 39-40C was never approached so it is not
reasonable to expect a summer decline resulting from high water temperatures.
Air temperature ranged from a minimum of 28 in December and January
to a maximum 41C in August. The median weekly temperature was never less
than 10C nor greater than 30C. Minimum temperatures, however, were con
sistently less than 10C from early November to late April. Summer temp
eratures were relatively constant and averaged about 27C. The winter was
comparatively mild with only 8 days having minimum temperatures of 0C or
less. Six of these freezes occurred in December and 2 occurred in January.
None were serious enough to severely damage the waterhyacinths. The only
detrimental effect noticed was a browning of the leaf tips on some of the
larger plants.

no
Even though frost did not severely damage the plants in the winter
the effect of the low temperatures may have been that of inhibiting
translocation of starch as described previously. Winter daytime temper
atures were generally quite warm but nighttime temperatures often fell to
the 10C range or below. If photosynthetic rates and carbohydrate trans
location are reduced in this temperature regimen a reduction in growth
would be expected during this time. Bock (1969) found that plants grown
at 4.4C nights and 4.4C days (16:8 L:D photoperiod) grew very little
(ca. 5%) during a 25 day test. Those grown at 26.7C days and 4.4C
nights did only slightly better (ca. 20%) while those grown at 26.7C
days and nights increased the most (ca. 105%). Apparently minimum
nighttime temperatures have a substantial effect on waterhyacinth growth
and even though daytime temperatures may be warm growth may be suppressed.
The inability of the plant to translocate starch in these low nighttime
temperature conditions along with the subsequent reduction in photo
synthetic efficiency seems to be the only explanation available for
this phenomenon and is probably a result of an interaction between temp
erature effects and physiological mechanisms (Knipling et al. 1970).
Figure 9 illustrates the annual curve for solar radiation during
the period of this study. The data is represented as daily averages for
the period between sampling dates and is in terms of calories/cm2(=Langleys).
A rapid increase in solar energy is indicated between the winter months
and late spring and appears to be at a maximum in early June. This is fol
lowed by a gradual decline which continues throughout the remainder of the
year and the minimum occurs at about the winter solstice. The summer maximum
does not seem to occur at the summer solstice and measurements in late June,
July, and August seen to be lower than would be expected based on day length

Figure 9. Solar radiation data from the University of Florida campus
from May 1974 through April 1975. Each point represents a
daily average spanning a seven-day period. Solar radiation
begins to decline before the summer solstice as a result of
afternoon cloud cover associated with summer rains. The
unit of energy (Langleys) represents calories per square
centimeter per day. (Data from Dr. E. A. Farber.)

X DAILY SOLAR RAOIATION (Langlcya)
2LL

113
alone. This is probably
afternoon thunderstorms
(see rainfall in Figure
the result of cloud cover associated with
common in this part of Florida during the
2).
frequent
summer

114
Waterhyacinth Productivity
Figures 10 and 11 illustrate diurnal curves for the productivity of
small and large waterhyacinths. Incident solar radiation and ambient
temperature curves for the two days in which this study was done are also
given. This data is the result of infrared C02 gas analysis described
in the methods section. This study was carried out cooperatively with
Sandra Brown, Ken Dugger, and Bill Mitsch and it was agreed that each
investigator would use the results freely as his research dictated. Even
though this agreement was made I do wish to point out that this is not
entirely my own material.
Gross primary production of the large plants (Fig. 10) was determined
to be 19.3 g Carbon/m2/da. Respiration was estimated at 13.2 g C/m2/da.
With the assumption that 1 g carbon = 10 kcal these figures are trans
formed into 193 kcal/m2/da and 132 kcal/m2/da. This indicates a value
of 61 kcal/m2/da for net primary production. The ratio' of this value
and the incident solar energy indicates a net efficiency of 1.6%. This
translates into a net gain of 13.55 gm organic matter (assuming 1 gm 0M =
4.5 kcal). Since the standing crop was 2140 gm/sq. m. a net gain of 0.63%
for the 24 hour period is estimated (0.83% standardizing to the 4900 kcal
solar radiation measurement of the small plants).
The gross primary production of the small plants (Fig. 11) was 15.6
g C/m2/da (156 kcal/m2/da). Respiration was estimated at 7.6 gm C/m2/da
(76 kcal/m2/da) and net primary productivity at 8.0 gm C/m2/da (80 kcal/m2/da)
The net efficiency for the small plants then is also 1.6%. Since the
standing crop is smaller this represents a relatively larger organic
matter gain. The net productivity of 80 gm C equals an organic matter gain

Figure 10. Diurnal curve for large waterhyacinth productivity determined
from CO2 gas exchange measured on Lake Alice with an infrared
C02 gas analyser. Respiration rates were determined from the
average night-time values. Gross production is defined as the
area under the curve above the respiration line. Net produc
tion is the area under the curve above the compensation point
(0 gC/m2/hr), The lower curves represent solar energy and
temperature.

PRODUCTION
116
TIME
AMBIENT TEMPERATURE <*C>

Figure 11.
Diurnal curve for small waterhyacinth productivity. See
Figure 10 for explanation.

(JM^UJ/O) NOIlDnaOdd NOIIV lOSNI
lie
I I ME
AMBIENT TEMPERATURE 1*0

119
of 17.78 gm or an increase of approximately 3.02%. Hence, one would
expect the small plants to increase relatively more rapidly than the
large plants.
The reasons for this difference are many. The increased metabolic
load in the larger plants due to the larger standing crop results in a
lower gross primary productivity:respiration ratio (1.46 vs. 2.05).
This ratio at steady state is 1.00 which indicates that the larger plants
are closer to steady state than the smaller ones. This infers that a
greater portion of the gross production is spent in maintaining existing
plant structure than in producing new material. This may be important in
biological control considerations for a herbivore which merely removes
leaf tissue without doing damage to the growing portion of the plant
may indirectly stimulate growth.
Even though the total metabolic load was greater in the large plants
the respiration per gram plant biomass was more than double in the smaller
plants (see Table 5). This may be an indication of a more active metabolic
rate associated with a faster growth rate.
Intraspecific competition for light may be another factor affecting
the observed difference in growth rates. The leaf area index of the
large plants was more than twice that of the small plants. The total
amount of photosynthetic tissue was 3 times greater in the large plants
and the leaf (pseudolamina) tissue was 9 times greater. In spite of these
large differences net efficiencies were equal and gross efficiency was
only 62% greater in the large plants. Hence, this greater amount of
photosynthetic tissue is probably not as effective per unit as is the
smaller amount of the small plants. In fact the gross productivity/gm

Table 5.
Metabolic and morphometric comparisons of the two morphological types of
waterhyacinth studied.
Parameter
Large Plants
(11-12 Aug. 1973)
Small Plants
(12-13 Aug. 1973)
Standing Crop (gm/m^)
2131.75. __
588.99
Leaf Area Index (m2/m2)
7 78
3.30
Photosynthetic 1 issue (gm/nv1)
1250.98
446.87
Leaf (Pseudolamina) Tissue (gm/m)
267.53
30.38
Incident Solar Radiaion
3750.
4900.
Gross Primary Productivity (gm C/m -da)
19.30 (25.23* )
15.60
GPP/LAI
2.48 ( 3.24* )
4.73
GPP/gm Photo. Tissue
0.015 ( 0.020*)
0.034
GPP/gm Leaf Tissue
0.072 ( 0.094*)
0.513
GPP/Standing Crop
0.009 ( 0.012*)
0.026
Net Primary Productivity (gm C/m -da)
6.1 ( 8.0* )
8.0
NPP/LAI
0.784 ( 1.03* )
2.424
NPP/gm Photo. Tissue
0.007 ( 0.009*)
0.018
NPP/gm Leaf Tissue
0.005 ( 0.006*)
0.263
NPP/Standing Crop
0.003 ( 0.004*)
0.014
Plant Respiration (gm C/m -da)
13.2
7.6
Respiration/Standing Crop
0.006
0.013
Total Efficiency (kcal GPP/Sol. Rad. x 100)
5.15/,
3.18%
Net Efficiency (kcal NPP/Sol. Rad. x 100)
1.63/
1.63%
Gross Primary Productivity:Respiration
1 46
2.05
Daily Rate of Increase (1 + Stan3PPCrop)
1.006 ( 1.008*)
1.030
Non-Photo. Tissue:Photo. Tissue
0.71
3.13
Assumes incident solar radiation equal to that of the small plants. This adjustment
is necessary for comparisons since considerably greater solar radiation values were
recorded during the day of the small plant study.

121
leaf tissue of the large plants is only 59% of that of the small plants
even where differences in solar radiation are taken into account. The
contribution of the photosynthetic layer in the petioles is not well
understood but it is thought to be of minor importance in primary pro
duction (Knipling et at. 1970). Further Knipling et al. (1970) found
that the light was rapidly extinguished beneath the upper canopy (ca.
25% at mid-height). Hence the primary function of petiole is probably
the supportive function of leaf display and the leaf tissue is probably
photosynthetically more important than the total photosynthetic tissue.
Table 5 lists the various metabolic comparisons standardized with regard
to the leaf area index, photosynthetic tissue, leaf tissue, and standing
crop. In all cases gross primary productivity is greater in the small
plants than in the large ones even though the reverse is true strictly
on a per unit area basis. The fact that the amount of leaf tissue pre
sent is 9 times greater in the large plants while the GPP/gm leaf tissue
is 40% less indicates that the increased leaf area interferes with light
reception and probably results in a greater deviation from potential
productivity.
The ratio of plant parts (Fig. 12) may be important in terms of
supporting photosynthetic processes and may partly account for the
difference in growth rates. As mentioned previously the petioles function
in displaying the leaves but probably do not contribute greatly to photo
synthesis. Hence, even though they are necessary, they represent a sub
stantial metabolic cost to the plant. Petioles account for 46% of the
weight of a large plant and only 19.1% of a small plant. A considerably
greater portion of the small plant is non-photosynthetic in nature than

Figure 12. A comparison of the standing crop and proportions of the
plant parts for the large and small waterhyacinth plants
used in the productivity studies.

123
LEAVES PETIOLES TOT. PHOT. NON-PHOT.

124
of the large plant (76% vs. 42%). The ratio of non-photosynthetic tissue:
photosynthetic tissue (Table 5) is over 4 times greater in the small plants
than that of the large plants. This non-photosynthetic tissue is primarily
roots. Hence, the small plants are probably more efficient at absorbing
nutrients to supply the photosynthetic and metabolic processes involved
in growth and production. All of these factors are probably responsible
for the faster growth rate of the small plants.
Because of the difficulty of obtaining the data discussed here
repetitions were not possible at this time. As a result, statistical
comparisons of the differences observed between the two types of plants
could not be made. In spite of this, such direct methods of measuring
productivity are far superior to traditionally used indirect measures
such as the estimation of changes in standing crop by periodical har
vesting. The latter method is often easily repeated and, as a result,
may be statistically more appealing. The technique employed here, however,
has the advantage of stoichemetric interpretation without the necessity
of transferring the plants to an artificial laboratory situation as is
normally required in metabolism studies.

125
Seasonal Variation in Photosynthetic Tissue
Virtually all parameters associated with photosynthetic tissue showed
strong seasonal tendencies. Plant height, leaves per plant, leaves per
unit area, area per leaf, leaf area index, area per petiole, petiole area
index, and total leaf area index show seasonal maximum ranges but these
peaks vary in time depending upon the parameter measured.
Plant height (Fig 14) seems to follow solar radiation curves but lags
a month or two behind. Figure 13 gives average daily solar radiation for
Gainesville on a monthly basis as 8 to 12 year averages. The data for
the period of this study generally agrees with this (Figure 4c). Maximum
radiation occurs in late May or early June but maximum plant height
(mean) is not achieved until late June and July. Minimum solar radiation
ranges occur in December but plant height does not reach its lowest level
until late January.
The number of leaves per plant (Fig. 15) appeared to be extremely
variable. At the beginning of the study (May) the range was between
6 and 7 leaves per plant. The following May, however, it failed to return
to this level (ranges 4-5). A decline was observed for this parameter
throughout the summer followed by an increase in September. The gradual
decline in the fall and winter seemed to parallel the decline in solar
radiation but a sharp spring increase did not occur. This is explainable
in terms of plant density and leaf density (Figs. 16 and 19). Even though
the number of leaves per plant was low in the spring the number of leaves
per square meter was at a maximum because plant density was high during
this time. It may be concluded then that a plant has the greatest number
of leaves in the summer when it is at its maximum height but the maximum

Figure 13. Average daily solar radiation values per month for
Gainesville, Florida. The figure above each bar represents
the number of years the means are based upon. Adapted from
the Climatic Atlas of the United States, U.S. Dept, of Commerce.

100
Mean Daily Solar Radiation (Langleys)
M
§
u>
o
o
o
o
500

Figure 14. Annual phenological change in the average height of the
waterhyacinth plants on the marsh side of Lake Alice. Height
was measured as the length of the longest leaf from the tip
of the pseudolamina to the base of the petiole at its attachment
to the rhizome. The mean is derived from all plants contained
within three 0.316 m2 samples. The dotted line represents pre
dicted values based on multivariate regression equations (see Table 7).

AV HEIGHT (CM )
PLANT HEIGHT
j
F
M
A
M
no


Figure 15. Annual variability in the number of leaves per waterhyacinth
plant from the study area. The means are derived from weekly
samples and represent averages from all plants contained in
three 0.316 m2 samples. Only leaves which had unfurled were
counted. The dotted line represents predicted values based
on multivariate regression equations (see Table 7),

CO

Figure 16
Annual change in leaf density as determined from weekly
samples taken in the study area. Each point represents a
mean based on the number of leaves in each of three 0.316
m2 samples. The dotted line represents predicted values
based on multivariate regression equations (see Table 7).

* I EAVES / M
1975
1974
LEAF DENSITY
OltllVIO DAU
MIDICTIC

134
leaf density occurs in the spring when the plants are at their maximum
density. Leaf density appeared to return very near its May 1974 level
in May 1975.
Plant height, plant density, leaf density, and leaves per plant
appear to be interrelated. It is confusing to consider any one of these
as an indicator of photosynthetic production. A more valuable index is
the leaf area index which takes into account the average area per leaf and
the leaf density. Figure 17 represents the curve for the average area per
leaf. This appears to be similar to the plant height curve but does not
show the brief period of decline in the late summer. A single distinct
peak occurs in July. The leaf area index is the product of the average
area per leaf and the number of leaves per square meter and is represented
in Figure 18. The leaf area index shows a strong increase in the spring
actually beginning as early as February. A peak occurs in May as a result
of both increasing area per leaf and a high leaf density. A secondary
peak occurs in July and seems to be due only to the increase in the area
per leaf. A rapid decline occurs from mid-July through early October.
An increasing trend is observable in October but drops off very sharply
within a one week period in November. The first cold weather (<2C)
occurred at this time and was apparently responsible for this. It is not
readily apparent what caused the short term rise in the leaf area index
prior to this decline but a number of factors may be responsible. Leaf
density was showing a steady increase while the area per leaf increased
slightly. Maximum air temperatures appeared to increase somewhat at the
same time and phosphate concentrations were increasing. The leaf area
index reached its minimum in January following the last two frosts of the
year (Jan 14 & 15) but this depression lasted only a few weeks and was

Figure 17. The average area of the pseudolaminae of waterhyacinth
leaves. Each point represents the mean leaf size derived
from measurements of all the leaves of 10 waterhyacinth
plants selected randomly at weekly intervals. Only one
side of the pseudolamina was measured.

150-
LEAF AREA
CO
cr>

Figure 18. Leaf area index of the waterhyacinth population on Lake Alice.
Each point represents the product of the leaf density per
square meter and the average area per leaf as determined from
the smoothed curves (Figures 16 and 17). The dotted line
represents predicted values based on multivariate regression
equations (see Table 7).

CO
00

139
followed by a sharp increase. On the whole the leaf area index seems to
follow solar radiation cycles. Cold weather causes a depression in the
leaf area index but the effect was short due to the relatively mild winter.
The strategy of the plant seems to be to maximize photosynthetic display.
This is done first in the spring by increasing leaf density and offset
production followed by an increase in leaves per plant and leaf size in
the summer as intraspecific competition becomes more intese.
As was discussed previously, the importance of the petiole in photo
synthesis is not known. It is assumed, however, to contribute little.
For this reason the petiole area index has not been graphed. The area
per petiole is usually very close to the area per leaf. Hence, the total
leaf area (pseudolamina and petiole) is approximately twice the leaf
area and the total leaf area index is approximately twice the leaf area
index. All three (leaf area, petiole area, and total leaf area) strongly
parallel the curve for mean maximum height (Fig. 14).
Seasonal Variation in Plant Density
Plant density (Fig 19) did not seem to follow the same trends as
the various estimates of photosynthetic tissue. A major peak occurred
in late April when the density reached 186 plants per square meter.
This was followed by an equally abrupt decline in May. By June the density
was between 70 and 90 plants per square meter and it remained in this
range until September. At this time the density began to increase and a
secondary peak of 130 to 140 plants per square meter was achieved in early
January. This level dropped slightly in February but the spring increase
began in early March.

Figure 19. Annual change in plant density as determined from weekly
samples taken in the study area. Each point represents an
average derived from counts taken from three 0.316 m2
samples. Only offsets with some root development were
counted as distinct plants. The dotted line represents
predicted values based on multivariate regression equations
(see Table 7).

ao-
PLANT
DENSITY
OllltVIO 0*TA
raioicTie
N D

142
These observations were somewhat surprising as a decrease in density
was expected as solar radiation levels fell and temperatures became colder.
Plant density was definitely higher in the winter, however, than in the
summer. This appears to be related to the leaf area index and plant height.
The average weekly frequency distribution for the various plant height
classes for each month is shown in Figure 20. In January the distri-
bition was narrow (range 60 cm) and skewed towards the small plants. The
dominant size class was 31-40 cm which contained approximately 32 plants
(26%). Taken together with the two smaller size classes, 55% of the plants
were found to be less than 40 cm during this month.
In February the distribution was narrower yet (range 50 cm) but the
predominant size class was larger (41-50 cm) and contained 31% of the
plants. Still the greater proportion of the plants were smaller than the
predominant class (42%) and taken together with the predominant class
represent 73% of the population.
In March the range of sizes seemed to broaden (60 cm) and there was
a lack of a single predominant size class. Four size classes (21-30, 31-40,
41-50, 51-60) accounted for 22%, 18%, 21%, and 24% of the population
respectively or 85% of the population collectively. By April when the
greatest increase in density occurred the size classes appeared to be
nearly normally distributed. The predominant class was 51-60 cm in
height and contained 38 plants (24%) and together with the 41-50 cm
class accounted for 42% of the population. Smaller size classes contained
26% of the population and larger plants accounted for 16%.
In May the distribution seemed to bifurcate. Two modes were apparent,
the first in the 41-50 range and the second in the 71-80 and 81-90 ranges.

Figure 20. Average monthly counts of the number of plants included in
each plant height class per square meter. The monthly
values were derived from the averages of all the weekly
samples taken during a given month. The dark vertical
bars represent the frequency of damage to the rhizome by
Arzama densa in each height class.

NUMBER OF PLANTS PER SQUARE METER
144

145
The latter two classes comprised 45% of the population and 40% were
smaller. Approximately 16% of the population was larger which was similar
to the previous month. By June this subpopulation of small plants began
to disappear and the distribution was skewed strongly toward the larger
size classes. The 91-100 and 100-110 cm classes were co-dominant with
54% of the population. The smaller classes comprises 34% of the total
and the larger only 10%. This continued in July and the same two size
classes represented 65% of the population. The contribution of the small
plants was minor with only a 23% representation. The predominant size
class was 101-110 cm in July and was the maximum height achieved by a
dominant class. The distribution was similar in August and September
but the predominant class was 91-100 cm in both months.
In October the predominance of the larger size classes was beginning
to decrease and the smaller plants were becoming more important. A net
increase in density occurred which was apparently responsible for the
increase in the smaller size classes. The two predominant classes (91-100,
101-110) comprised only 35% of the population. This trend continued in
November but the two predominant classes were smaller (71-80, 81-90) and
53% of the plants were smaller. By December the predominant class was 31-40
cm and the frequency distribution was broad. Six classes (21-80 cm) were
co-dominant.
Statistics for skewness (assymetry) and kurtosis (peakedness) were
determined for each weekly frequency distribution according to the methods
described by Sokal and Rohlf (1969). Kurtosis was plotted as a dependent
function of skewness and a hyperbolic regression fitted to the data (Fig.
21). Positive values for skewness indicate that the distribution is
skewed towards the ight (high values in the distribution). Negative

Figure 21. Statistics of skewness and kurtosis (peaking) derived from
each weekly frequency distribution of plant density by
height classes. This figure indicates that as the frequency
of plants becomes skewed towards the larger height classes
the degree of peaking in the distribution increases sharply
(i.e. the diversity of height classes represented decreases

Pages
are
misnumbered
following
this
insert

4^
CT>

147
values indicate that the distribution is skewed left (towards low values).
Positive values for kurtosis indicate a high degree of peaking where a few
classes contain most of the individuals in the distribution. Negative
values indicate a broad distribution with less distinct peaks. Both
skewness and kurtosis are approximately 0 in a normal distribution. The
regression indicates that when the population is not strongly skewed to
wards the larger size classes the distribution tends to be normal or de
pressed with several classes well represented. As the degree of skewness
towards the large size classes increases, however, the population shows
much stronger peaks indicating the increased predominance of a few size
classes. This supports the contention that the increased dominance by
the large plants results in a loss of the smaller size classes and a
decrease in density.
In general, then, as the predominant size class becomes larger there
appears to be a loss of plants in the smaller size classes and plant den
sity decreases. This is further illustrated in Figure 22 where each
weekly frequency distribution is plotted from January through December
in a three dimensional manner. This is particularly true in the summer
when the largest plants are also the predominant class. As the leaves
from the larger plants die the small plants become better represented
and the density increases. Density, then, appears to be auto-regulatory
and responds to the changes in the canopy. Even though the photophase
is decreasing the amount of light may increase in the lower canopy as
the leaves of the larger plants die off. This may stimulate offset pro
duction as is evidenced by the close inverse association between plant
height and plant density. There appears to be an optimum, however, and
this may have occur ed in April. At this time the degree of intraspecific

Figure 22. The weekly waterhyacinth height class frequency distributions
plotted three dimensionally on a time scale. The horizontal
axis represents size classes increasing from left to right.
The vertical axis represents density, the number of plants in
each height class per square meter. The third axis, perpen
dicular to the plane of the paper, represents time (weeks).


150
shading was not so intense as to select for only the larger height
classes. Solar radiation was waxing and a dramatic increase in density
resulted. Following this, as the plants became taller and the average
area per leaf increased, intraspecific competition invariably selected
for the plants which could maximize their energy budgets, namely the
larger height classes. As these large plants became larger the light
available to the smaller plants became less. This resulted in the loss
of the smaller size classes and the skewed distributions observed in
the summer. As a word of caution then, a decrease in density does not
necessarily indicate a reduction in the condition of a waterhyacinth
stand. This could easily be misunderstood by persons evaluating the
effect of insects on waterhyacinths.
Seasonal Variation in the Standing Crop
Standing crop was not measured directly on a weekly basis. However,
biomass samples were taken in December 1974 during the period of this
study and in April, May, June, July and August 1975 after weekly data
collections were terminated. From these biomass samples the average
weight per plant was estimated and is plotted against the average height
(measured at the same time) in Figure 23. An exponential regression was
fitted to this data and the regression formula derived. The correlation
coefficient for this regression was 0.92 with 10 degrees of freedom and
is highly significant (<0.01). The estimates for average height from the
original weekly samples were entered into this regression equation to
determine the estimated average weight per plant. This value was multiplied
by the plant density for the same sampling period to obtain an estimate
of the standing crop per square meter in terms of grams dry weight. These

Figure 23. Average weight per plant as a log function of the average
plant height. Data from biomass samples taken from Lake
Alice.

36

Figure 24. Standing crop values, both estimated and real, from Lake Alice
The points represent estimates derived from the average values
for height from May 1974 through April 1975 using the equation
in Figure 23 to determine an estimated plant weight. These
were multiplied by plant density to obtain an estimate for
standing crop. The vertical bars represent actual measurements
taken from nine 0.25 m2 samples taken each month from April 1975
through February 1976. The dotted line represents predicted values
from multivariate regression analyses (see Table 7).

2.5
2.0-
STANDING CROP
t
ISTIMATtO DATA
rtiOICTIO
>T oiv OISftVtO DATA
r ni i Art rj-fib.'7*
1.5
cn
-P

estimates are plotted in Figure 24. Subsequent actual measurements taken
from April 1975 February 1976 are also plotted (vertical bars) and
conform very well with the estimated curve. As a result I feel that the
estimated curve is a fairly accurate representation of acutal conditions
Standing crop was lowest in January and February when it ranged
between 600 and 1000 gm DW/m2. A period of exponential growth began in
March and continued through April. The discontinuity of the curve at
the end of April is due to differences between the initial measurements
in May 1974 and the final measurements in 1975. It is therefore not
apparent whether this exponential trend continued. It does not seem to,
however, because there was a drop in biomass between April and May of
1975 in the actual biomass samples. Following this decline a nearly
linear rise occurred through late June. During this period of active
growth the growth rate was fastest in March and April. The daily incre
ment factor (see Bock 1969) for this period (5 March 23 April) is
estimated at 1.015 (1.5% per day). The same value for the period be
tween 1 May and 10 July is 1.010 (1.0% per day). The average monthly
values for daily increments are listed in Table 6.
Following the peak in July a rapid decline occurred. This was
followed by a stable period in mid-September. A general decline began
the first part of October and continued through January. Biomass re
mained stable in February before beginning a spring resurgence.
Surprisingly, the greatest monthly decline did not occur during
the coldest months but rather in October and November. This may have
been due to heavy damage by Arzama densa which occurred at that time.
The period of maximum growth (April) corresponds to the greatest monthly

ib 1 e 6. Average daily rates of change in biomass from initial and final monthly
values.
Month
Standing Crop
Interval
(Days)
Daily Increment*
1st Estimate
Last Estimate
Jan
787
611
27
0.9907
Feb
798
757
20
0.9974
Mar
835
988
21
1.0080
Apr
929
1483
28
1.0168
May
1176
1545
28
1.0098
Jun
1750
2190
23
1.0098
Jul
1802
1682
26
0.9974
Aug
1451
1436
21
0.9995
Sep
1254
1500
21
1.0086
Oct
1501
1092
28
0.9887
Nov
1283
968
21
0.9867
Dec
1067
1005
20
0.9970
TIT"
T where T = interval, NT = final estimate, Nn = initial estimate.
To 0

157
increase in solar radiation (see Fig. 9). the drop in May was probably
due to the loss of small plants as a result of increased intraspecific
competition. There appears to be two growth phases, the early spring
phase due to both increasing plant density and individual plant growth
and the early summer phase due to individual plant growth only. This
seems to parallel the theory of cyclic r and K-selection and will be
discussed later in this section.
Damage by Arsama densa
Larvae of Arsama densa cause considerable damage to the waterhyacinth
plants by boring into the leaves and rhizomes. Eggs are laid in masses
usually on the pseudolamina and upon emergence the neonates mine the
leaves or move to the base of the plant and feed on the tender wrapper
leaves. As they become larger (ca. 4th instar) they bore into the petiole.
Prior to this time they primarily feed externally on the photosynthetic
tissue. By the 5th instar they are capable of killing the leaves and
may bore into the rhizome. At first, damage to the rhizome occurs at the
apical tip but later the 6th and 7th instars tunnel into it creating
large burrows and possibly fragmenting the plant. Damage by the larger
larvae often results in the death of the plant.
Figure 25 illustrates the leaves damaged by A. densa as a percentage
of the total leaf density. This figure represents total leaf damage and
does not distinguish between external feeding and petiole bores. Figure
25 also illustrates the percentage of the plants with rhizome damage.
Again, the degrees of damage are not distinguished. Figure 20 illustrates
monthly averages for rhizome damage by plant size classes.
Leaf damage (Fig. 25) was very low in the summer as was the larval

Figure 25. Percentage of the leaves and rhizomes of the waterhyacinth
population damaged through feeding activity of Arzama densa
at Lake A1ice.

159

160
population. An increase in damage was apparent beginning in July and
continuing through November when a peak of approximately 25% occurred.
A decline followed through December until late January when the damage
ranged between 5 and 10%. It remained at this level through February and
March with a brief increase in April. By mid-May the level of damage
was somewhat greater than the previous year.
Rhizome damage closely paralleled leaf damage but was generally
larger peaking in November where 43% of the plants had rhizome
damage (Fig. 25). Rhizome damage was very low from May through September
(Figure 20). An increase was apparent in October but most of the damage
was minor. Most of the severe damage occurred in November, December, and
January. By spring the relatively low larval population coupled with the
high plant density resulted in relatively low levels of damage. The in
crease noted in April was primarily minor damage.
The frequency of attack of a given plant size class roughly corresponds
to the frequency of that size class (Figure 20). This is generally true
throughout the year except in December and January when the smaller plants
are more abundant but most of the damage occurs to the larger plants. This
may be partly responsible for the loss of plants in the larger classes
apparent between December and January. Otherwise it appear that there is
no selection by the insect for the size of plant attacked. Larger plants
may be attacked more frequently at certain times but this is probably
due to the fact that they are older and have been exposed to attack for
a longer period of time. If the selection of a plant in a given size
class by a larva is a random process then the frequency of attack within
that class should correspond to the relative abundance of that class within
the frequency distribution. Without further analyses this appears to be the case.

161
Results of the Multivariate Analysis
The results of the multivariate regression analysis are summarized
in Table 7. The expression for each dependent variable may be derived
from this table by reading across the row according to the form of the
general model. All regressions were highly significant with a probability
of a greater F occurring by chance at 0.0001. Some of the regression
coefficients were not significant at the 0.05 level but were included if
the probability for the null hypothesis (Ho:B=0) was less than 0.10.
Solar radiation and minimum air temperature were the most important
variables in the models for standing crop, plant height, leaf area index,
and leaves per plant. Since all of the climatological variables were
highly correlated (see Table 8) the inclusion of these two reflects the
importance of climate to the maintenance of the waterhyacinth stand.
Nutrient effects were variable depending upon the model. Nitrogen
was important only in the expression for standing crop but the coefficient
was marginally significant at the 0.05 level. This infers that high levels
of nitrogen were associated with larger standing crop values and vice-versa.
This positive relationship was somewhat unexpected since the nitrogen
concentration should drop as biomass increased and greater quantities of
nitrogen are absorbed. Phophate-phosphorus was included in the models
for plant height, leaf density, and plant density. The coefficient for
plant height was negative and probably reflects the uptake of phosphorus
as height increases. The positive coefficients for leaf density and plant
density indicate that these values were high at the same time that phosphorus
concentrations were high. Potassium was a significant factor in the same
three models but the relationship was reversed. High potassium concentra
tion corresponded to low leaf and plant density v ues and to high values
for plant height.

Table 7. Summary of multivariate regression analyses for annual variation of various plant characteristics. The general model is
Y Bj + B¡X] + ... + BnXn where Y = the dependent variable, Bo = the intercept, B-_-Bri = coefficients for the independent
variables and Xi-Xr = trie dependent variables.
DEPENDENT
VARIABLE
INTERCEPT
Mi
B.X.

B.X.
B Xu
Ms
BfX£ B7X7
F
R2
STANDING
CROP
185.34
(0.2817)*
27.08 Min.-Air
(0.0001)
1.36-Sol. Rad.
(0.0004)
116.14-N
(0.0575)

--

29.91
(0.0001)**
0.647
PLANT
HEIGHT
27.70
(0.0792)
0.05-Sol. Rad.
(0.0005)
0.81-Min. Air
(0.0030)
1.91 Max.-Air
(0.0059)
-13.25-P
(0.0003)
1.26-K
(0.0281)
-0.09-Alk. -8.65-Level
(0.0824) (0.0007)
44.10
(0.0001)
0.873
LEAF AREA
INDEX
8.44
(0.0008)
0.004 Sol.-Rad.
(0.0001)
0.024-Min. Air
(0.0462)
-0.88-pH
(0.0061)
--
--
32.77
(0.0001 )
0.667
LEAVES PER
PLANT
(0.0001)
0.002-Sol. Rad.
(0.0079)
0.053-Min. Air
(0.0001)
-2.72-FE
(0.0036)
0.006-S0-
(0.0029)
--
--
35.77
(0.0001)
0.749
LEAF
DENSITY
177.20
(0.0350)
219.14 P
(0.0001)
57.37-Level
(0.0253)
-19.66K
(0.0015)
440.54-Fe
(0.0097)
-7.45-Mg.
(0.0307)
0.22-Sol. Rad.
(0.0712)
17.30
(0.0001)
0.693
PLANT
DENSITY
137.93
(0.0001)
-3.58-Max. H:0
(0.0001)
28.85-P
(0.0001)
144.96-Fe
(0.0001)
-2.21 K
(0.0783)
10.46-Level
(0.0255)
--
27.46
(0.0001)
0.745
* The statistic in parenthesis under each coefficient represents the results of a T-test performed on each coefficient and is the probability
of a greater value of T under the null hypothesis that B = 0.
** The statistic in parenthesis under each F represents the results of a multivariate analysis of variance and is the probability of a greater
F occurring by chance.

163
Iron appears to be an important nutrient in at least three models.
The coefficient for iron was negative in the expression for leaves per
plant. The number of leaves per plant is correlated with standing crop
(see Table 9), hence, the negative coefficient for iron can be taken to
indicate uptake of iron as the plant biomass increases. The positive
coefficients in the models for leaf and plant density indicate that
iron concentrations are high when these two variables are high. Since
maximum plant and leaf densities occur early in the growing season iron
levels may be high because the plants have not yet affected it. The
peak for plant density occurs immediately after the peak for iron con
centration (see Figs. 4 and 19). This indicates that a causal relation
ship may exist between the two.
Potassium was included in the expression for plant height, leaf
density, and plant density. A decrease in the potassium concentration
between late February and late April (Fig. 3) corresponds to the peaks
for leaf and plant density and accounts for the negative potassium co
efficient for these two variables. The positive relationship between
potassium and plant height (Fig. 14) is not obvious by mere inspection
of the data. The indication is that as potassium concentrations increase,
plant height also increases. The effects of more important variables
probably obscure this relationship between potassium and plant height.
Magnesium was significant in the model for leaf density. This was
somewhat surprising since magnesium concentrations were relatively con
stant through the year (see Fig. 4). The coefficient for leaf density
was negative but this inverse association is not obvious.
Hydrogen ion concentration (pH) was included in the model for leaf
area index with a negative coefficient. Values of pH had a narrow range

Table 8. Correlation coefficients (r) between independent variables.
Statistics inparentheses represent the probability of a
greater |r| under the null hypothesis.
Solar Min. Air Max. Air Min. H20 Max. H^O
Radiation Temp. Temp. Temp. Temp.
Solar Radiation
1.000
(0.0000)
0.615
(0.0001)
0.558
(0.0001 )
0.608
(0.0001)
0.612
(0.0001)
Min. Air
1.000 0.815 0.848 0.828
(0.0000) (0.0001) (0.0001) (0.0001)
Max. Air
1.000 0.796 0.830
(0.0000) (0.0001) (0.0001)
Min. H20
1.000 0.813
(0.0000) (0.0001)
Max. H20
1.000
(0.0000)
Rhizome Damage
Leaf Damage
Phosphorus
Nitrogen
Iron
Conductivity
Potassium
Magnesium
Alkalinity
pH
Sulfates
Lake Level

Table 8.
(Continued)
Nitrate +
Rhizome Leaf Phosphate Nitrite -
Damaqe
Damage
Phosphorus
Nitroqen
Iron
Conductivity
Solar Radiation
-0.423
-0.388
-0.292
-0.328
-0.100
0.228
(0.0001)
(0.0001)
(0.0024)
(0.0006)
(0.3064)
(0.0186)
Min. Air
-0.512
-0.492
-0.546
-0.406
-0.204
0.112
(0.0001)
(0.0001)
(0.0001)
(0.0001)
(0.0364)
(0.2549)
Max. Air
-0.361
-0.288
-0.639
-0.489
-0.284
0.045
(0.0001)
(0.0001)
(0.0001)
(0.0001)
(0.0032)
(0.6499)
Min. H20
-0.334
-0.270
-0.646
-0.525
-0.149
0.182
(0.0005)
(0.0051)
(0.0001)
(0.0001)
(0.1279)
(0.0621)
Max. H,0
-0.301
-0.277
-0.648
-0.505
-0.210
0.247
(0.0017)
(0.0040)
(0.0001)
(0.0001)
(0.0304)
(0.0106)
Rhizome Damage
1.000
0.961
0.139
0.267
0.156
0.091
(0.0000)
(0.0001)
(0.1548)
(0.0057)
(0.1106)
(0.3526)
Leaf Damage
1.000
0.096
0.243
0.111
0.063

(0.0000)
(0.3274)
(0.0121)
(0.2558)
(0.5214)
Phosphorus
1.000
0.389
0.134
0.057
--
--
(0.0000)
(0.0001)
(0.1723)
(0.5647)
Nitrogen
1.000
-0.061
-0.186
--
--
--
(0.0000)
(0.5356)
(0.0565)
1 ron
_ _
1.000
0.180

--
--
(0.0000)
(0.0643)
Conductivity
_ _
_ _
1.000
--
--
--
--
--
(0.0000)
Potassium

--
--
--

--
Magnesium

--
--
--


Alkalinity
--
--




ph
Sul fates
--
--

--
Lake Level

166
Table 8. (Continued)
Potassium
Maqnesium
Alkalinity
pH
Sul fates
Lake
Level
Solar Radiation
-0.249
0.175
0.091
-0.279
0.380
0.591
(0.0101)
(0.0731)
(0.3549)
(0.0037)
(0.0001)
(0.0001)
Min. Air
-0.188
0.044
0.083
-0.160
0.186
0.404
(0.0537)
(0.6510)
(0.3979)
(0.1016)
(0.0564)
(0.0001)
Max. Air
-0.224
-0.043
0.127
-0.041
0.189
0.493
(0.0207)
(0.6594)
(0.1935)
(0.6730)
(0.0523)
(0.0001)
Min. H^O
-0.266
0.022
-0.042
-0.158
0.254
0.522
(0.0058)
(0.8197)
(0.6667)
(0.1060)
(0.0087)
(0.0001)
Max. H,0
-0.279
0.152
0.308
0.080
0.308
0.473
(0.0038)
(0.1198)
(0.0013)
(0.4157)
(0.0013)
(0.0001)
Rhizome Damage
-0.037
0.165
0.003
0.381
0.058
-0.376
(0.7060)
(0.0912)
(0.9747)
(0.0001)
(0.5538)
(0.0001)
Leaf Damage
-0.029
0.135
-0.045
0.363
0.063
-0.319
(0.7675)
(0.1675)
(0.6501)
(0.0001)
(0.5227)
(0.0009)
Phosphorus
0.221
0.019
-0.004
-0.291
-0.130
-0.382
(0.0227)
(0.8444)
(0.9674)
(0.0025)
(0.1841)
(0.0001)
Nitrogen
0.416)
-0.066
-0.133
0.133
-0.237
-0.613
(0.0001)
(0.5003)
(0.2506
(0.2502)
(0.0145)
(0.0001)
Iron
-0.247
0.122
0.094
-0.195
0.165
0.092
(0.0108)
(0.2536)
(0.3380)
(0.0454)
(0.0911)
(0.3504)
Conductivity
-0.415
0.394
0.602
-0.087
0.762
0.237
(0.0001)
(0.0001)
(0.0001)
(0.3758)
(0.0001)
(0.0144)
Potassium
1.000
-0.146
-0.157
-0.052
-0.422
-0.267
(0.0000)
(0.1343)
(0.1077)
(0.5939)
(0.0001)
(0.0057)
Magnesium
1.000
0.233
-0.077
0.522
-0.024

(0.000)
(0.0161)
(0.4333)
(0.0001)
(0.8066)
Alkalinity
1.000
0.214
0.466
0.089
--
--
(0.0000)
(0.0275)
(0.0001)
(0.3632)
pH
1.000
-0.094
-0.141
--
--

(0.0000)
(0.3360)
(0.1485)
Sulfates
1.000
0.189
--
--

(0.0000)
(0.0522)
Lake Level
_ _
_ _
_ _
_ _
_ _
1.000
--

--
--

(0.0000)

Table 9. Correlation coefficients (r) between dependent variables. Values
in parentheses represent the probability of a greater |r| under
the null hypothesis.
Standing
Crop
Plant
Height
Leaf
Area
Index
Leaves
Per
Plant
Leaf
Density
Plant
Density
Standing Crop
1.000
_ _
_ _
(0.0000)
--
--
--
--
--
Plant Height
0.846
1.000

--
--
--
(0.0001)
(0.0000)
--
--
--
--
Leaf Area Index
0.820
0.664
1.000
--

(0.0001)
(0.0001)
(0.0000)

--
Leaves Per Plant
0.599
0.790
0.630
1.000
(0.0001)
(0.0001)
(0.0001)
(0.0000)
--
--
Leaf Density
-0.027
-0.436
0.181
-0.143
1.000
(0.8447)
(0.0015)
(0.1922)
(0.3082)
(0.0000)
--
Plant Density
-0.325
-0.760
-0.187
-0.651
0.806
1.000
(0.0166)
(0.0001)
(0.1756)
(0.0001)
(0.0001)
(0.0000)

168
over the year but were generally lower in April and May (see Fig. 6).
This may be the result of increased plant respiration during this phase
of active growth. Respiration increases the CO^ concentration which re
acts with water to form carbonic acid. Carbonic acid dissociates ancL^
increases the hydrogen ion concentration thus lowering pH. These changes
are usually buffered by reacting with limestone to form carbonates and
bicarbonates. This accounts for the correlation between pH and alkalinity
(total carbonates and bicarbonates) in Table 8. The drop in pH, then,
is probably a result of plant growth and not a cause of it.
Alkalinity was included in the plant height model but the coefficient
was not significant. Sulfates were significant in the model for leaves
per plant and had a positive coefficient. Significant correlations with
sulfates include conductivity, magnesium, potassium and alkalinity.
Overall, then, sulfates are probably an index of soluble salts.
Water level was included in the plant height model with a negative
coefficient and in the plant density model with a positive coefficient.
Considering the water sources for Lake Alice low water levels probably
result in a concentration of nutrients and high levels in a dilution.
The negative coefficient for plant height indicates, then, that the plants
are taller when nutrients are low. Conversely the plants are most dense
when nutrients are high. Significant negative correlations exist between
water level and nitrogen and phosphorus.
Maximum water temperature was the most important variable in the
plant density model and was negatively related to it. Hence, low maximum
water temperatures indicate high densities.
In general, the models seem to indicate that the variables that

169
serve as indices of biomass (see Table 9) are regulated by climate as
solar radiation and air temperature are most important. Leaf density and
plant density appear to be regulated more by hydroponic conditions as
various water quality parameters are emphasized.
Predicted values were generated for each dependent variable based
on the known independent variable values. Each has been plotted as an
annual curve with the actual observed curve (see Figs. 14, 15, 16, 18,
19, and 24). Biomass estimates (Fig. 24) seemed to fit well in the win
ter, spring and fall but values were underestimated in the summer. This
is probably because of the assumption of linear effects inherent in the
model. As mentioned previously this assumption is probably not justified.
Plant height (Fig. 14) was approximated fairly accurately. The observed
drop in late August was not apparent, however, and an increase in late
January was predicted which did not occur. This is apparently the result
of a brief period of warm weather which, according to the model, should
have resulted in a brief increase in height. Otherwise the included
variables satisfactorily account for variations in height.
Values predicted for leaf area index produce a curve of approxi
mately the same shape as the observed data (Fig. 18). The peak was not
predicted until early June, however, where it actually occurred in mid-
May. The predicted peak corresponds to the peak in solar radiation.
The predicted curves for leaves per plant (Fig. 15) and leaf !
density (Fig. 16) conform extremely well to the actual data. Plant
density (Fig. 19) is also well represented but the spring peak is not
as dramatic as was observed. This is probably because the change at
this time was exponential and the model treats it in a linear fashion.

170
Overall, with the exceptions of standing crop and possibly leaf
area index, the models produced reflect observed trends in the plant
characteristics fairly accurately. More sophisticated modeling techniques
could probably improve these fits if non-linear responses could be con
sidered .
Discussion
The object of this study was to determine the degree of effect of
damage by Avzama densa on various characteristics of the waterhyacinth
population. Neither leaf damage nor rhizome damage proved to be a
significant factor in any of the models derived from the multivariate
regression analyses (Table 7). The obvious conclusion based only on these
analyses is that a. densa damage did not affect the plants or, more
precisely, did not account for a significant amount of the variation
observed in the plant characteristics measured. Upon examination of the
correlation matrix for the independent variables (Table 8), however, it
is found that both leaf and rhizome damage are highly correlated with the
climatological variables and with water level. Further, these correlation
coefficients are all negative indicating that insect damage is high when
sunlight, temperature, and water level are low. Because of thes rela
tionships it is not reasonable to exclude insect damage as an important
factor since the correlated variables are important. If all of the variables
were independent a term for insect damage may have been included in the
models. In lieu of this independence the stepwise analysis first selects
the parameter which best reduces the variability. Further parameters are
used to account for the variability remaining after variability due to
the first parameter is removed. When the first and second parameters are
highly correlated the exclusion of variability due to the first may also

171
remove the effects of the second. In this case the second variable is
considered as not important and is excluded from the model. This may
very well have been the case with insect damage. The effects of insect
damage may have been obscured in this manner by the greater effects of
sunlight, temperature, and water level. To see if this possibility existed
a simple correlation analysis was performed comparing each independent
variable with each dependent variable on a one to one basis (Table 10).
Both estimates of Arzama damage were highly negatively correlated with
plant height, leaf area index and leaves per plant. While this does not
necessarily infer that a causal relationship exists between insect
damage and the plant characteristics the possibility is present. Further
more sophisticiated analyses may be able to isolate this effect but the
results of this study are inconclusive with regard to damage by Arzama
densa.
The models produced seemed to fall into two broad categories, bio
mass and density. The first includes standing crop, plant height, leaf
area index, and leaves per plant. These four characteristics are inter
related as evidenced by the significant positive correlations (Table 9)
between them. Because of this interrelationship all are probably indices
of biomass and all are low in the winter, increase in the spring, reach
their peaks in early summer, and decline in the fall. All of these are
primarily climatologically limited as evidenced from the inclusion of
solar radiation and minimum temperature in each multiple regression
model (Table 7).
The major nutrients (N, P, K) are all included in at least one of
the four above mentioned models. The coefficients for these are somewhat
difficult to interpret. Nitrogen is included in the standing crop model.

Table 10. Correlation coefficients (r) between dependent and independent
variables and probabilities for a greater |r|.
Standing Plant
Crop Height
Solar Radiation
0.200
(0.0403)
0.620
(0.0001)
Min. Air Temp.
0.211
(0.0299)
0.840
(0.0001)
Max. Air Temp.
0.174
(0.0746)
0.757
(0.0001)
Min. H ,0 Temp.
0.193
(0.0473)
0.806
(0.0001)
Max. H;,0 Temp.
0.185
(0.0577
0.823
(0.0001)
Rhizome Damage
-0.133
(0.1752)
-0.386
(0.0043)
Leaf Damage
-0.128
(0.1921)
-0.344
(0.0118)
Phosphorus
-0.126
(0.1964)
-0.675
(0.0001)
Nitrogen
-0.051
(0.6041)
-0.275
(0.0465)
I ron
-0.040
(0.6824)
-0.347
(0.0109)
Conductivity
0.022
(0.8261)
0.058
(0.6795)
Potassium
-0.020
(0.8352)
-0.090
(0.5200)
Magnesium
0.017
(0.8661)
0.098
(0.4865)
Alkalinity
-0.011
(0.9090)
0.015
(0.9165)
pH
-0.079
(0.4233)
-0.039
(0.7790)
Sul fates
/
0.034
(0.7264)
0.196
(0.1587)
Lake Level
0.096
(0.3294)
l) 2>J9
(O.ui'S/)
Leaf
Area
Index
Leaves
Per
Plant
Leaf
Density
Plant
Density
0.764
0.698
0.150
-0.258
(0.0001)
(0.0001)
(0.2844)
(0.0619)
0.600
0.762
-0.259
-0.600
(0.0001)
(0.0001)
(0.0610)
(0.0001)
0.512
0.665
-0.271
-0.595
(0.0001)
(0.0001)
(0.0493)
(0.0001)
0.585
0.701
-0.253
-0.584
(0.0001)
(0.0001)
(0.0680)
(0.0001)
0.487
0.730
-0.308
-0.657
(0.0002)
(0.0001)
(0.0249)
(0.0001)
-0.467
-0.473
-0.073
0.203
(0.0004)
(0.0003)
(0.6020)
(0.1444)
-0.441
-0.407
-0.095
0.149
(0.0009)
(0.0025)
(0.4974)
(0.2860)
-0.214
-0.409
0.555
0.650
(0.1243)
(0.0023)
(0.0001)
(0.0000)
-0.257
-0.400
-0.030
0.236
(0.0636)
(0.0030)
(0.8331)
(0.0891)
-0.091
-0.319
0.398
0.546
(0.5186)
(0.0197)
(0.0032)
(0.0001)
0.117
0.309
0.313
0.050
(0.4045)
(0.0242)
(0.0224)
(0.7214)
-0.232
-0.276
-0.288
-0.066
(0.0947)
(0.0452)
(0.0367)
(0.6379)
0.071
0.227
-0.076
-0.088
(0.6119)
(0.1028)
(0.5867)
(0.5303)
-0.078
0.197
0.102
-0.060
(0.5811)
(0.1579)
(0.4681)
(0.6708)
-0.438
-0.124
-0.360
-0.271
(0.0010)
(0.3770)
(0.0080)
(0.0498)
0.176
0.410
0.124
-0.128
(0.2096)
(0.0023)
(0.3751)
(0.3603)
0.462
0.409
0.195
-0.103
(0.0005)
(0.0024)
(0.1626)
(0.4647)

173
The simple correlation coefficient between standing crop and nitrogen
(Table 10) is negative and not significant. After the effects of sun
light and solar radiation are removed this relationship is positive as
is evidenced by the positive coefficient in the model. Hence this states
that at constant levels of light and temperature the standing crop in
creases as nitrogen levels increase. However, since absorption of nutrients
increases as standing crop increases one would expect a negative rela
tionship between the two if nitrogen input from the various sources re
mains constant. This inverse relationship seems to be apparent in Figures
5 and 24. The drop in nitrogen concentration in the summer, however, is
more likely due to an increase in the water level resulting in a dilution
of the nutrient load. This is evidenced by the significant correlation
between nitrogen and lake level (Table 8). Mitsch (1975) demonstrated
that a decrease in nitrates across the marsh does occur and is greatest
in the summer when the waterhyacinth standing crop is high. Nitrogen,
then, may very well be limiting to waterhyacinths and the biomass supported
by the available nutrients may increase as relative nitrogen concentrations
increase.
*-
Phosphorus is generally considered one of the primary limiting factors
is aquatic systems (H. T. Odum 1953). It is included in the model for
plant height and the coefficient is negative. The simple correlation co
efficient between phosphorus and height is also negative and significant.
The model infers that with climatological effects removed, plant height
increases as phosphorus decreases. This suggests an increasing absorption
of phosphorus as the biomass increases. Phosphorus concentration, however,
is also affected by the water level. High phosphorus concentrations are
correlated with low water levels (Table 8). Mitsch (1975) showed a small

174
decrease in phosphorus concentrations occurred across the marsh in the
summer. Hence plant growth probably does affect the phosphorus concen
tration but it is difficult to determine if these concentrations become
limiting to the plants. Phosphate-phosphorus levels never fell to the
limiting level of 0.10 mg/1 as determined by the experiments of Haller
et al. (1970) discussed earlier. The drop in phosphates may have been
due to the absorption of luxury amounts by the plants beyond their
immediate requirements.
Potassium was included in the plant height model after the effects
of climate and phosphorus were removed. The simple correlation coeffi
cient between height and potassium was negative and not significant
(Table 10). The coefficient in the regression model (Table 7), however,
was positive and significant. Potassium was also significantly correlated
with water level (Table 8). The indication from the model is that at
constant levels of sunlight, temperature, and phosphorus, plant height
increases as potassium concentrations increase or vice versa. Again,
one would expect the opposite, that is, as the plants become large ab
sorption would increase and the nutrient concentration correspondingly
decrease. The possibility exists that maximum nutrient absorption occurs
early in the growing season. Nutrients may be absorbed in large quantities
and stored within the plant. The period of maximum absorption may occur
simultaneously with the period of fastest growth. As the growth rate
slows nutrient absorption may also decrease. Hence, even though the plants
are larger during the summer the plants may be having a lesser effect
on nutrient levels. This may explain the apparent decline in potassium
levels in March and April (Figure 7). This same explanation is also
possible for nitrogen.

175
i
Other significant variables included in the four models which re
present biomass are water level, pH, iron, and sulfates. The possible
significance of water level and pH have already been discussed. Sulfates
are correlated with magnesium, potassium, alkalinity, and conductivity.
The inclusion of sulfates in the leaves per plant model may indirectly
indicate the importance of soluble salts. The inclusion of iron with a
negative coefficient in the same model indicates the increased uptake
of this essential nutrient as plant biomass increases. This agrees with
my observations on plants grown in greenhouse cultures.
Plant density and leaf density are highly correlated and form the
second category. These two parameters, unlike the biomass indices, do
not show a peak in the summer but rather in the spring. Potassium, iron,
phosphorus, and water level were included in both models (Table 7). The
signs for these coefficients are opposite those for the same variables
in the previous models. The negative coefficient for potassium indicates
that it is being absorbed as density increases. This supports the argument
given above that potassium is absorbed in the greatest quantities at the
time of fastest growth. Further, potassium in plants is known to accumulate
in those tissues that are growing rapidly (Robbins et al. 1964) and would
be expected to be absorbed in greater quantities when maximum growth occurs.
Iron, on the other hand, appears to be directly related to density.
Although it is difficult to determine why iron concentrations showed a
dramatic increase in the spring (Fig. 4) the strong correlation with plant
density cannot be ignored. The iron enriched water appears to have been
al least partially responsible for the increase in density evidenced in
the spring. Since iron is an improtant element for the synthesis of

176
chlorophyll it may be an important limiting factor in the waterhyacinth
community.
Phosphorus was also directly related to leaf density and plant density.
Maximum densities, therefore, occur when phosphorus concentrations are
high. This is in contrast to plant height which achieves its maximum when
phosphorus quantities are low. This may be interpreted as indicating
that high phosphate concentrations early in the growing season are important
in initiating the spurt of rapid growth in the spring. As the plants
become larger and biomass increases more phosphorus is absorbed and the
concentration decreases.
Magnesium in the leaf density model had a negative coefficient.
This may have reflected the uptake of magnesium as leaf tissue increased.
This would be expected since magnesium is an important constituent of
the chlorophyll molecule. This change is not obvious in Figure 4, however.
Climatological variables did not appear to have the importance in
the density models that they had in the biomass models. Solar radiation
was included in the leaf density model but the coefficient was not signi
ficant. Maximum water temperature was considered the most important
variable in the plant density model and was negatively related to it.
This probably reflects inter-relationships of different characteristics
within the plant population, however. That is, competition for light is
reduced in the winter thereby allowing an increase in density. Since
temperature is so important in the biomass model and biomass is inversely
related to density.
The multivariate analysis as presented does not take into consider
ation auto-regulatory features within the waterhyacinth population. The
productivity studies show that small plants grow more rapidly than large

177
plants presumably because the large plants are closer to steady state
with a smaller P:R ratio. This is in spite of the fact that net efficiencies
are approximately equal for both size classes. One would expect, then,
that the plants would grow faster in the spring when the plants were
small than in the summer when they were large even if external conditions
were equal. This cannot be attributed entirely to an increased metabolic
load in large plants because the gross primary productivity per gram of
leaf tissue decreases as the plants become larger (see Table 5). This
infers that a unit of large plant leaf tissue is less efficient than an
equal unit of small plant leaf tissue. Respiration per gram biomass is
almost double in the small plants so this cannot account for the difference.
One possible reason for this difference is senescence of the older leaves.
A second is intraspecific competition for light. Both of these explanations
are probably partly true. As the leaves become older they probably do
naturally become less efficient. Also they are more likely to have been
attacked by diseases, mites, insects, and other factors which may reduce
their effectiveness. Also, as the plants become larger they are more prone
to self shading. Hence, even though the amount of light received is the
same as the small plants the amount received per unit of photosynthetic
area would be inversely proportional to the degree of competition.
This pattern of self regulation by intraspecific competition seems
to account for changes in plant density within the waterhyacinth community
better than any of the physical parameters included in the plant
density model. As shown earlier, small plant size classes are lost and
plant density decreases as the plants become larger. Figure 26 further
illustrates this. In this figure density is plotted as log function of
plant height. This slightly improves the correlation coefficient derived

Figure 26. Plant density as a function of plant height. The greatest deviation from
the regression line occurred in April (dotted line) when density reached
its highest level. By May, however, the points again followed the regression.

C^J

180
from the simple correlation in Table 9 (-0.806 vs. -0.760). Most of the
data points fit the regression line extremely well. The greatest deviation
from this occurs in April when a much higher density occurs than would
have been expected based on plant height alone. Because density rapidly
fell back to normal ranges I believe that this represents an "overshoot"
response by the plants to favorable conditions. Early in the season nu
trients are high and competition for light is low. As a result the plants
respond by producing a great number of offshoots. This numerical response
greatly intensifies competition and as the canopy rapidly increases light
is effectively cut off from the smaller plants resulting in their loss
and a decrease in density. The surviving plants would be those that
maximized their ability to receive incident solar radiation. This would
involve an increase in petiole length so that leaf display would be above
the neighboring plants. Hence, in the summer a low density population
of tall plants was present.
This phenonmenon strikes me as very similar to a theory of insect
cycles discussed by Wellington (1960) based on tent caterpillar popu
lations. He found that active genetic strains are present at low den
sities and tend to reproduce rapidly. Sluggish strains are present in
high densities and are adapted to crowded conditions. Conditions of
low competition are present in a closed waterhyacinth community in the
spring. In this situation the plant appears to be adapted to vegetatively
reproduce rapidly. During the summer under conditions of high competition
the plants seem to be at steady-state and appear to be limited by envir-
omental conditions. Like the tent caterpillars they seem to be adapted
to conditions of intense competition. Hence early in the season selection

181
favors those plants that can produce the most offsets earliest. Late in
the season selection favors those plants that can compete for light most
effectively. This results in the pattern of density observed in Figure
19. This explanation applies only to waterhyacinth populations which
are limited by space and is the result of genetic flexibility within
the population rather than different genetic strains. If open water
continues to be available in the summer the reproductive phase may con
tinue until all available space is utilized as long as some other fact
or does not become limiting. This phase of reproduction is likely to
occur only at the fringe of the mat nearest the available space. The
plants further back within the mat are more likely to be limited by
competition for light and will generally be increasing in height. Con
sidering that waterhyacinths are colonizer species this pattern of growth
is probably very adaptive. Increasing density in open areas enables
the plant to colonize rapidly. It further increases the number of propagules
available for reaching new downstream areas to produce daughter colonies.
Once the plants become established in an area their ability to compete
must increase in order to maintain the colony. In its native habitat the
ability to increase in height enables it to compete both inter- and intra-
specifically since several similar species occur with it (e,g. Eiohhomia
azurea, Reussia sp.). In the United States, however, there are relatively
few large floating aquatic macrophytes and most interspecific competition
is with plants in the littoral zone. In areas of deep water where emer
gent vegetation does not exist competition for light is intraspecific.
Since relatively few herbivores act to reduce this intraspecific compe
tition the population is limited only by climate and nutrients. Hence,

182
in Florida's nutrient rich waters and moderate climate large stands of
tall plants are common.
The pattern of growth of waterhyacinth seems to involve several
phases on Lake Alice. A period of "no-growth" occurs in January and
February where the standing crop begins to increase an increase in leaf
density occurs. This begins as early as February and seems to be the
first phase of growth. The leaf density peak occurs in late March. The
second phase of growth is an increase in plant density. This begins in
early March and the peak occurs in late April. The number of leaves per
plant increases as leaf density begins its increase but levels off in
March and April only to begin a new increase in May. The spring increase
in leaf density is due both to an increase in the number of leaves per
plant and an increase in plant density.
The third phase of growth is an increase in height which begins in
late March and peaks in July. The increase begins when both leaf density
and plant density are high and may be a response to this. Standing crop
begins to increase at the same time as height and peaks at about the same
time.
The fourth growth phase is an increase in leaf size and does not
begin until May but reaches a peak in early July along with height and
standing crop. Leaf production appears to be manifest first in an in
crease in leaf density, second in an increase in the number of leaves per
plant, and finally in an increase in leaf size. The adaptive strategy
seems to be that of maximizing leaf area. Under conditions of low com
petition this can best be achieved by producing more offshoots. Under
conditions of intense competition each plant produces more leaves and the
leaves increase in size. This may be interpreted as a diversion of

183
available energy in the path of an energy gradient. When the solar energy
gradient is stronger peripherally lateral growth occurs. As peripheral
light penetration diminishes due to increased competition and the light
gradient becomes relatively stronger vertically, vertical growth begins
to occur. Either way the increase in the leaf area index from early
February through late May is almost continuous.
Growth appears to slow in the summer and a dramatic decrease in
both leaf and plant density occurs. A summer decline in the number of
leaves per plant, plant height and standing crop occurs in late July,
August, and early September. The reasons for this are not apparent but
it may be the result of a change in the carrying capacity of the system.
These characteristics level off for a short while until a general decline
begins in mid-September. This decline continues until the winter lows
are reached in January. As plant height, leaf size, leaves per plant, and
standing crop decline leaf density and plant density begin to increase.
This increase continues until January when a slight decline occurs.
These annual cycles illustrate the plasticity of the waterhyacinth
population in adapting to different situations. The population is regu
lated primarily by climatological factors, by water quality, and by the
intensity of intraspecific competition. Rapid adjustment in the morpho
metry of the mat occurs as these factors change. It is not unreasonable
to expect that this capacity to adjust may apply to attack by insects. By
reducing intraspecific competition insects may indirectly cause an increase
in density. There probably is, of course, a damage threshold beyond which
further damage by insects could severely affect the ability of the water-
hyacinth population to adjust. I would expect this threshold to be high,

184
however, and to vary seasonally. I would expect the plants to be most
vulnerable in the fall when solar energy is declining and the ability to
store carbohydrates for winter survival is critical. By decreasing this
ability the integrated effects of insect damage and freeze damage may
retard the spring growth phase. Although Arsama damage was highest in
the fall it apparently was not severe enough to have any long lasting
effect on the waterhyacinth community.

CHAPTER II
THE CONSEQUENCES OF ATTACK BY ARZAMA DENSA WLK. ON SOME ECOLOGICAL
CHARACTERISTICS AND MORPHORMETRIC FEATURES OF WATERHYACINTH.
Introduction
The effects of an insect attack on a host plant depends not only
on the biology of the insect but also on the ecological response of the
plant. Harris (1962) noted that insect attacks may decrease plant abun
dance, have no effect on plant abundance, or actually stimulate plant
growth. He further stated that insects which feed on aquatic plants
may cause sectioning of the stems from which propagation occurs and
increase the spread of the plant. Bennett (in Harris 1972, apparently
referring to Vogel and Oliver (1969a)) stated that it has been demon
strated that a large noctuid (probably Arsama densa) which attacks
waterhyacinth (Eichhomia erassipes) may create more plants and spread
the weed. Vogel and Oliver (1969a) attributed this increase in the
number of plants with increasing insect concentrations to a reduction
in dominance of the apical bud. Their hypothesis was that by feeding
on the apical meristem the insect caused the expression of the lateral
buds thereby increasing the number of offsets produced.
If herbivory can cause the spread of a weed and thus increase the
probelm, it is imperative to determine the mechanism by which this
occurs. A number of explanations other than a reduction in apical
dominance are possible for this increase in offsets. A reduction in
intraspecific competition or an increase in nutrient due to an accelerated
turnover rate may contribute to this. Seasonal effects must also be
considered. The purpose of this study is to examine in detail the
185

186
potential effect of Arzama densa on the ecology of the waterhyacinth
in terms of net productivity, standing crop, turnover rates, and other
characteristics of the plant community during two distinct seasons
(summer and fall).

187
Methods and Materials
The sides of four greenhouse tables 12 inches deep were constructed
from 1 inch by 12 inch redwood. The tables were lined with 6 mil
polyethylene sheeting and filled with water. The water on all tables
was fertilized equally, as needed, with 20:20:20: water soluble commercial
plant fertilizer with minor elements. Sequestered iron was also added
to obtain concentrations of 2.5 ppm. Square foot grids were constructed
on each table by stretching nylon twine across the top of the tables and
tying it off on nails spaced one foot apart. Two tables provided 18 one
sq. ft. quadrats each and two of the tables which were somewhat larger
provided 27 quadrats each. One small waterhyacinth plant in the inflated
petiole stage taken from Lake Alice was placed within each quadrat on all
tables and the plants were allowed to grow for several weeks until the
tables were completely covered. Five quadrats were then randomly selected
from each table and the plants were harvested to obtain base measurements
for the parameters to be evaluated. In the first experiment average
plant height, no. leaves/quadrat, no. of leaves/plant, and no. plants/
quadrat were counted and to obtain initial baseline estimates for the
table. The same variables were estimated in the second experiment in
addition to total dry weight subdivided into living and dead organic
material. The living material was further subdivided into leaves, rhizomes,
roots, and stolons. Where applicable these measurements were evaluated
in terms of both unit area and individual plant.
Arsama densa eggs were collected for the first experiment from
pickerel weed (Pontederia cordata) at a lake near Putnam Hall, Florida,
on 25 June 1974. These were allowed to hatch and neonates were placed
on the tables in sufficient numbers to achieve 0, 0.33, 0.67, and 1.00

188
larva/plant levels of infestation. Five quadrats were again selected
on 25 July and again on 2 August 1974 and the same parameters measured.
Also the no. of leaves damaged, no. rhizomes damaged, severity of the
damage, no. larvae, no. pupae, no. pupal cases, no. dead larvae,
apparent instar of each larva, etc. was determined. Final values were
taken on two separate dates (when the insects were approximately 30 and
38 days old) so as to bracket the damage estimate. Underestimates of
damage would be obtained if the plants were harvested before the insects
reached full size or after they ceased feeding and the plants had begun
to recover. For this reason, it was desireable to harvest the plants
at the time when the insects had begun to pupate. The final values are
therefore expressed as the means of ten quadrats taken five at a time on
two separate occasions. To further standardize the results the final
values are expressed as a percent of the initial value (i.e., ynit^aT
X 100).
The eggs for the second experiment were the generation of the
larvae from the first experiment and were collected from waterhyacinths.
Initial samples were taken 1 October 1974 and the neonates were intro
duced 3 October. The levels of infestation were the same as in the
first experiment. This experiment was allowed to proceed somewhat longer
until 14 and 25 November (42 to 53 days). Other than the increased
duration and the greater number of parameters measured this experiment
was like the first one.

189
Analyses
After determining that there was no significant difference (student's
T-test for unpaired data) in any of the parameters measured between the
two post-infestation sets of data they were combined to obtain 10 obser
vations per treatment level for each parameter after one generation of
insect activity. The final results are expressed as the mean values
of 10 parameterized final values for each level of infestation and are
expressed as a percentage of the initial value.
The results were interpreted by means of the regression procedure
from the Statistical Analysis Systems on the University of Florida IBM
370 computer. The results were fitted according to the three following
regression models where Y = plant response and X = insect concentration:
1. Y = A + BX
2. Y = A + BX + CX2
3. Y = A + BX + CX2 + DX3
4. Y = A-Bx
The model which best fitted the data and presented a realistic
estimate of response trends was selected for plotting. The best fit
in the first three models was determined by examination of the sequential
sums of squares. The fourth model was compared with the others by means
of the regression coefficients (r). The model with the largest r was se
lected only if it was a great deal larger than the alternative model.
Otherwise the model most clearly representing the results in the simplest
form possible was chosen. In the majority of the cases the simple linear
model (equation 1) provided a satisfactory representation the the results.
It was originally expected that the data would conform to the log conver
sion model (eq. 4) but this regression consistently provided a lower r-value

190
than the linear model. Because of this and because the log model produced
predicted values based on geometric means which underestimated the arith
metic means this model was never selected. If a greater number of higher
levels of insect infestation were used in the experiment I believe the
exponential nature of the response curves may have become more evident.
Even though the regressions in this paper are linear the true relationship
between the plant response and the insect concentration may not be
linear.

191
Results
Plant Height
As waterhyacinth stands become denser they become contained
laterally and instead of producing more shoots they increase in height.
As the plants become taller the biomass increases exponentially as is
illustrated in Figure 27. As the living tissue increases the respira
tory demand must also increase. Due to crowding and intraspecific com
petition for light the amount of photosynthetic tissue needed to main
tain the plant must also increase. The ultimate result is that the Pro-
ductivity:Respiration (P/R) ratio decreases until it approximates 1
(Brown et at. 1974). This steady-state is achieved where the energy
produced is just sufficient to meet the respiratory demand of the plant
and net production approximates zero. The standing crop at steady-
state depends upon the carrying capacity of the habitat. A nutrient-
rich body of water supports a larger steady-state standing crop than
a nutrient poor body of water. As long as nutrients are constantly
replenished and solar energy remains high a steady-state equilibrium
will be maintained. Density and biomass become balanced with respect
to their sources of energy unless other factors interact to disrupt
this stability. Because plant height is related to stand density as
well as the plant weight it also reflects external changes. Since the
taller leaves (from which I measured height) are older, they are more
prone to insect attack due to an increased duration of exposure. Height,
then should be a very sensitive index of feeding activity by herbivores
providing other factors remain stable.

Figure 27. Average dry weight per waterhyacinth plant as a log function
of the average height.

MEAN PLANT WEIGHT G.DW
VO
OJ

194
Figure 28 represents the response curve for plant height to one
generation of feeding by Arsama densa with varying initial insect
concentrations in the summer and in the fall. At the time the plants
were infested the stands had approached a steady-state as reflected by
the small increase in the controls (11% and 5% respectively). As feeding
intensity increased a nearly linear decline in both experiments was
observed. The 0.33 level of infestation appeared to be unduly low in
the summer experiment but this was the consequence of a severe attack
by spider mites (Tetranyahus gloveri Banks) which occured on that table.
A somewhat greater response to herbivore activity was observed in the
fall than in the summer. This is as expected and appears to be due to
an interaction effect between decreased solar radiation in the fall and
herbivore activity.

Figure 28. The average height per waterhyacinth plant (as measured from
the longest leaf) as a function of the feeding activity of
Arzama densa larvae. The means represent data taken after a
generation of feeding attack expressed as a percentage of the
mean values for the same variables taken before the insects
were introduced. The circles represent the mean values taken
from ten 0.093 sq. meter samples. The brackets enclose the
range of the standard error of the mean. The open circles
(dotted line) are values from the summer. The solid circles
(solid line) are values from the fall. The summer data for
the 0.33 insect concentration was excluded from the regression
analysis because of the effects of a heavy spider mite infestation.
Summer: Y = 111.63 36.23X, r = -0.8045
Fall: Y = 103.92 49.97X, r = -0.8168

PERCENTAGE OF INITIAL VALUE
196

197
Leaves
The terminology of waterhyacinth leaves is an area of controversy
among botanists. In general there are two distinct portions, the more
or less swollen base and the blade. According to Arber (cited in
Penfound and Earle, 1948) the blade is not a true lamina but an extension
of the petiole hence it is often referred to as a pseudo-lamina. For the
prupose of this discussion I will refer to the swollen base as the
petiole and to the pseudo-lamina as the leaf blade.
In open stands the petioles are bulbous containing numerous air
spaces and function as floats and, as mentioned previously, tend to be
short. In dense stands, however, the petioles are elongate and tend to
lack the bulbous float depending instead on the support of the dense
anchorlike roots and the intertwining of adjacent plants to hold them
erect.
Since A. densa feeds on the leaves as well as the rhizome the
effects of feeding activity should be manifest in a leaf number response
(Figs. 29 & 30). By killing the plant a decrease in leaves/unit area
should be observable when plant density also decreases. Since plant
density responses vary (Figure 31) the number of leaves per unit area
will not always reflect insect activity. A better index appears to be
leaves per plant. As the insect feeds on the leaves the number per
plant should decline. Likewise, if new offsets are produced they would
tend to be smaller and have fewer leaves than their parent plants.
Figure 29 illustrates the response curves of leaves per plant to varying
insect concentrations in the summer and the fall. With the exception of
the controls the means were almost identical in both experiments. The
difference in the controls is probably due to a seasonal effect as
explained earlier.

Figure 29. The effects of varying levels of insect feeding activity
on the average number of leaves per waterhyacinth plant
expressed as a percentage of predetermined means. Legend
as in Figure 28.
Summer: Y = 102.16 40.83X, r = -0.76
Fall: Y = 112.22 53.60X, r = -0.69


Figure 30. The effects of varying levels of insect feeding activity
on the total number of waterhyacinth leaves per unit area
expressed as a percentage of predetermined means. Legend
as in Figure 28.
Summer: Y = 135.11 + 1.72X, r = 0.012
Y = 13-.29 85.54X, r -0,572
Fall:


202
Figure 30 illustrates the response curves in terms of leaves/unit
area. The changes in this measurement were quite different between the
two seasons. In the fall the percentage of the initial number of leaves
present showed a significant decline with increasing insect
During the summer, however, no change was apparent. Since
the number of leaves per plant was evident this stability i
per unit area was apparently due to the increase in plants.
concentrations,
a decrease in
n the leaves

203
Plant Density
When discussing the effects of stress on the density of waterhyacinth
(number per unit area) one must be careful to define the seasonal stage
of the plant and the type of stand it is growing in. Eiohhomia orassipes
requires abundant solar energy, waters rich in nutrients, and open space
for maximum productivity (see the discussion on the seasonal ecology of
this plant presented earlier). Most factors which are likely to stress
waterhyacinth (such as herbicides, frost, or insects) tend to open up
the canopy. The plants become smaller and as a result the P/R ratio and
net efficiency probably increases (Brown et at. 1964). In this situation
they become r-strategists and reproduce rapidly competing successfully
with other species for the available space. This is observed in the spring
when there is an apparent "overshoot" in the plant population. The data
from experiments with laser beams for the control of this plant illustrate
the same principle (Long and Smith 1975). After the laser treatment a
decline in the rate of change was observed but it was immediately followed
by a sharp increase until the experimental plots were almost identical to
the control plots.
In this r-selection situation species with high rates of reproduction
and growth are more likely to survive in an uncrowded situation (E. P. Odum,
1969). When the waterhyacinth stand matures and occupies the total avail
able space an equilibrium density is reached. In this situation the
plants appear to be -strategists where selection favors species with
lower growth potential but which are more competitive under equilibrium
densities (E. P. Odum, 1969). The effect of the insect (or other stress
factors) appears to be that of causing the plant to "switch" strategies
/
by disturbing the equilibrium density. Discontinuation of the stress

Figure 31. The effects of varying levels of insect feeding activity on
the number of waterhyacinth plants per unit area expressed
as a percentage of predetermined means. Legend as in
Figure 28.
Summer: Y = 122.77 + 81.07X, r = 0.4209
Fall: Y = 123.95 35.53X, r = -0.2758


206
would allow the equilibrium density to again be achieved and the available
space occupied according to the carrying capacity of the habitat.
If these assumptions are true, then one would expect the seasonal
consequences of insect attack to vary. The carrying capacity of the
habitat depends to a large extent on solar flux. In the summer when
solar energy is nearly maximum one would expect insect infestation to
ultimately result in an increase in the absolute density of the plant
since energy is high and space is available. In the late fall, when
solar energy is waning, these consequences may be much different.
Because of the reduced energy available the carrying capacity is reduced.
An increase in available space at this time may ultimately result in
little change or a decrease in plant density as other factors (i.e. solar
energy) become limiting. The data from this experiment (Figure 31) tends
to support this hypothesis. If insect damage to the apical bud was solely
responsible for these increases in density without the influence of other
factors, then it would be reasonable to expect the plants to react simi
larly in both experiments. Instead I observed an increase in density in
the summer experiment and a slight decrease in the fall experiment which
favors the seasonal interaction explanation for density changes.
I do not intend to dispute the fact that damage to the terminal bud
may cause a response in the direction of lateral growth. One could cite
many examples of similar phenomena in many different plants. From field
observations, however, it is worth noting that the offshoots produced
from plants with severe damage from A. densa do not appear to be as
vigorous as normal offshoots. These offshoots are often deformed and the
leaves often appear to be rather thick and leathery. I merely intend to
demonstrate that the evolutionary strategy of the plant is to produce

207
offsets when sufficient space and solar energy is available. This occurs
with or without an interaction of insect attack. From a biological con
trol standpoint the fact remains, whatever the cause, that an increased
number of plants may be produced and the possibility of augmenting the
spread of this plant does exist. I contend that the activity of the
insect does not influence spread. Spread occurs primarily by the small
floating plants which grow peripherally from a stand of larger plants
toward the open water. The limiting factor is space, if none is available
insects can do no harm. If space is available the plants will grow into
it anyhow, providing conditions are suitable. If the total area available
is completely occupied by waterhyacinths the density of the plant popu
lation is inconsequential. The only undesirable effect that may be the
result of insect activity is that of fragmenting the stand causing
daughter colonies to float off more frequently and become established
in new areas. As will be discussed subsequently, other parameters illus
trate more fully the true impact of A. densa feeding activity on
waterhyacinth. Because of differences in the growth pattern of the
plant, density is not a good index of herbivore effectiveness.

208
Biomass Estimates
The effect of varying insect concentrations on the total biomass
per unit area is shown in Figure 32. Total biomass, as defined here,
includes a living component (standing crop) and a non-living component
(detritus), Figures 33 and 34, i.e., the total organic material present.
Biomass estimates were only taken in the fall experiment, therefore
seasonal comparisons cannot be considered. The summer standing crop
(Fig. 33) could be estimated, however, using predicted values from the
regression equation in Figure 27 for average weights per plant multiplied
by the number of plants per unit area. The value for standing crop at
the 0.33 level of infestation was corrected for mite damage by interpo
lating the expected values for height and plant density from figures
la and Id assuming approximate linearity between the 0 and 0.67 levels
of infestation.
Total biomass (Figure 32) revealed a rather unexpected response to
insect concentration. At all four levels of infestation an increase was
evident. The control (0 infestation level) increased 188%. The plots
treated with insects, however, responded very similarly at all three
levels of infestation resulting in approximately a 160% increase over
the initial biomass present. The effects of insects on total biomass
was not significant.
The standing crop (grams living material per unit area) declined
significantly with increased initial insect concentrations in the fall
experiment. If it is assumed that the estimated regression for the
standing crop values in the summer experiment is reasonably correct a
quite different response is apparent. Instead of a rapid decline even
with low infestation levels no response is apparent until the infestation

Figure 32. The effects of varying insect concentrations on the total
waterhyacinth biomass (expressed as both detritus and living
plant material). Data was taken only from the fall experi
ment. Legend as in Figure 28.
Regression: Y = 180.75 26.58X, r = -0.1828

260
TOTAL BIOMASS
220-
140-
100
f
0 0.2 0.4 0.6 0.8
INITIAL INSECT CONCENTRATION
1.0

Figure 33. The effects of varying insect concentrations on the living
waterhyacinth mass present per unit area. Data was only
taken from the fall experiment, however, estimates were made
by calculating the average weight per plant from Figure 27
based on the average plant height and multiplying by the plant
density (open circles and dotted line). The summer curve was
fitted by eye. Otherwise, legend as in Figure 28.
Regression (Fall):
Y = 161.99 122.71X, r = -0,6584

PERCENTAGE OF INITIAL VALUE

Figure 34. The effects of varying insect feeding activity on the
amount of dead waterhyacinth plant material (detritus)
per unit area. Data was taken only from the fall experi
ment. Legend as in Figure 28.
Regression: Y = 269.15 3.76X + 2980.51X2 -
2456.33X3, r = -0.7984

PERCENTAGE OF INITIAL VALUE

215
reaches the 1.00 level. This probably reflects the carrying capacity
of the plant community for this herbivore. The carrying capacity is
probably dependent upon the initial standing crop as well as the pro
ductivity of the plant, i.e., the rate at which the community can
replenish itself. The plants in the summer are more productive than
those in the fall and, as a result, can support a greater herbivore
population. The point at which the herbivores begin to have a negative
effect on yield is the response threshold. In the fall this threshold
occurs between the 0 and 0.33 levels. In the summer, however, the
threshold apparently is much higher at between the 0.67 and 1.00 levels
of infestation.
The change in the amount of detritus (dead organic material) present
is probably a good indicator of insect feeding activity provided that it
can be measured accurately and identified as waterhyacinth debris. The
net change is dependent upon the living material available during the
period of infestation, the level of insect feeding activity, and the rate
of degradation of the detritus by decomposers. Assuming that the decompo
sition rate per gram detritus is constant the amount of dead organic
material should directly indicate insect feeding activity up to the point
where it becomes limited by the amount of living material available for
conversion. A detrital response curve regressed on insect concentration
(Figure 34) would be expected to increase up to a point, tend to level
off, then show a rapid decline. The point of deflection for this curve
should occur at the point where plant productivity begins to be reduced
by the feeding activity of the insects. As insect concentrations become
larger (beyond the range of this experiment) this point would be reached
earlier in the growing period of the plant and cause a decline in the

216
curve until it levels off at a point where the increase in detritus is
equal to the initial living material present. Figure 35 illustrates
this somewhat differently as an almost linear relationship between the
insect concentration and the amount of detritus present expressed as
a precentage of the total organic material present. This linear rela
tionship should hold until the ratio; of detritus to biomass approaches
unity. At this point an asymptote in the curve should become apparent
where higher concentrations of insects have a proportionately smaller
effect on this ratio (a ratio of greater than 1.0 is impossible). If
the net productivity is zero and the detritus: total biomass ratio is
1.00 then all of the plants were immediately killed upon release of the
insects. If this ratio is 1.00 and the net productivity is some value
greater than zero then all of the plants were killed at some time after
the initial release.

Figure 35. Detritus as a percentage of total waterhyacinth biomass as
a function of insect feeding activity. Curve fitted by eye.

218

219
Productivity and Turnover Estimates
The relationship between biomass and detritus can be more graphically
illustrated if net productivity and turnover rates are considered (Figs.
36 and 37). Net productivity, as used here, is defined as the final
quantity of organic material present (total biomass present at the end of
the experiment detritus present at the beginning) per quantity of living
material initially present and is expressed as a percentage. A value of
100 for net productivity would indicate no change in biomass and is the
minimum value possible. Statistically Figure 36 indicates that there was
no significant change in net productivity as a result of insect feeding
activity. Intuitively there does appear to be some effect, however, as
productivity was approximately 170% at all levels of infestation whereas
it was approximately 205% for the control.
Assuming that productivity was not affected by the insect population
and with the knowledge that standing crop was significantly reduced, it
can be deduced that the important function of the insects was that of
accelerating turnover. The effects of insect feeding activity on the
relative turnover ratio is illustrated in Figure 37. The turnover ratio
as used here refers to the change in detritus per unit area per gram of
living plant material initially present. It is inversely related to
turnover time or the time required for the initial living material to be
converted to detritus. An approximately linear relationship is apparent
between insect activity and the turnover ratios. These ratios translate
into a range of turnover times of approximately 153 days for the control
to 40 days for the 1.00 infestation level (where T.0. time = j-qX
47.5 da (mean duration)). It is obvious, then, that the turnover time
decreases exponentially with increasing insect concentrations. These

Figure 36. Net waterhyacinth production as a function of insect feeding
activity. Net production, as used here, refers to the final
quantity of organic matter present,excluding the initial
amount of detritus, as a percentage of the initial living
plant material. Data only from the fall experiment. Legend
as in Figure 28.
Regression: Y = 194.42 32.54X, r = -0.1897

[ 221

Figure 37. The ratio of conversion of living waterhyacinth plant material
into detritus as a function of insect feeding activity. The
figures are based on the change in detritus over one generation
of Arsama densa population divided by the amount of convertible
living material present at the time the insects were introduced.
Data only from the fall experiment. Legend as in Figure 28.
Regression: Y = 0.3254 + 0.9007X, r = -0.7521

A DETRITUS ST. CROP
223

224
relationships should remain fairly constant regardless of season since
they consider only initial living material present and detritus produced.

225
Plant Parts and Proportions
The response curves for the weights of various plant parts to
varying levels of insect infestations show very similar trends and
tend towards exponential declines (Figures 38-42). Regression anal
ysis, however, showed that the log response curves did not improve
the correlation coefficients when compared to a linear response curve.
The results, therefore, are plotted as straight line relationships.
If further data were available beyond the levels of infestation tested
in this experiment a curvilinear response might be more evident.
Of the four plant parts rhizomes showed the greatest increase in
the control treatment at approximately 228% of the initial value. The
0.33, 0.67, and 1.00 levels showed responses of 128%, 65%, and 48%
respectively. This indicates that in a situation without an insect
infestation and greatest proportion of the carbon fixed is stored in
the rhizome. The proportion of rhizome weight to total living plant
weight tends to support this. The rhizome represented 8.7% of the ini
tial plant weight and increased to 11.6%, this represents an increase
of 133% (see Table 11). Penfound and Earle (1948) stated that the
rhizome was the main organ of storage. It is apparent in this study
that a great deal of the energy assimilated by the plant is stored in
the rhizome as carbohydrates. Insects, by causing a decrease in the
ability of the plant to create this storage, cause a depletion in the
carbohydrate reserves. This directly affects the ability of the plants
to survive periods of further stress and to resprout from the rhizome
if the leaves are killed. This may reduce the ability of the plant to
survive periods of cold, herbicide treatments, pathogens, and further
insect attacks.

Figure 38. The effects of varying insect feeding activity on waterhyacinth
green mass (pseudo-laminae and petioles). Data only from the
fall experiment. Legend as in Figure 28.
Regression: Y = 173.27 128.00X, r = -0.6501


Figure 39. The effects of varying insect feeding activity on waterhyacinth
non-green mass (roots, rhizomes, and stolons). Data only from
the fall experiment. Legend as in Figure 28.
Regression: Y = 131.54 109.68X, r = -0.6749


Figure 40. The effects of varying insect feeding activity on waterhyacinth
root mass per unit area. Data only from fall experiment.
Legend as in Figure 28.
Regression: Y = 90.78 73.92X, r = -0.6452

ROOTS
125-
* i lili
0.2 0.4 0.6 0.8 1.0
INITIAL INSECT CONCENTRATION
0

Figure 41. The effects of varying insect feeding activity on the
waterhyacinth rhizome mass present per unit area. Data
only from the fall experiment. Legend as in Figure 28.
Regression: Y = 207.43 180.22X, r
-0.6946

RHIZOMES
*\*
\
0.2 0.4 0.6 0.8
INITIAL INSECT CONCENTRATION
0
1.0

Figure 42. The effects of varying insect feeding activity on the
waterhyacinth mass represented as stolons per unit area.
Data only from the fall experiment. Legend as in Figure
28.
Regression: Y = 136.84 97.01X, r = -0.5169


236
Table 11. Ratios of the various plant parts and the percent
change in the final values as compared to the initial
values. The simple correlation coefficient (r) is
derived from a linear regression analysis.
Insect Concentration
Ratio
0
0.33
0.67
1.00
r
Leaf Wgt.
Initial
0.731
0.704
0.735
0.694
-0.172+
0.214+
Plant Wgt.
Final
0.775
0.788
0.827
0.793
%
106.02
111.93
112.52
114.27
0.399*
Rhizome Wgt.
Initial
0.087
0.082
0.092
0.079
-0.192+
Plant Wgt.
Final
0.116
0.089
0.078
0.079
-0.559**
%
133.33
108.54
84.78
100.00
-0.475**
Root Wgt,
Plant Wgt.
Initial
Final
0.149
0.090
0.185
0.082
0.135
0.056
0.172
0.091
0.040+
0.023+
%
60.40
44.32
41.48
52.91
-0.175+
Stolon Wgt.
Initial
0.028
0.027
0.035
0.051
0.389+
FlanT Wgt.
Final
0.020
0.040
0.023
0.039
0.142+
%
71.43
148.15
65.71
76.47
-0.072+
Root-Rhiz.
Initial
0.325
0.386
0.311
0.373
0.084+
Wgt.
Final
0.267
0.224
0.164
0.218
-0.380*
Leaf Wgt.
%
82.16
57.90
52.60
58.55
-0.550**
+ Probability (p) of a greater |r| >0.05; *p<0.05; **p<0.01.

237
The quantity of carbohydrate available for storage is related to the
amount of photosynthetic material present to produce it and the efficiency
of the plant. The effect of reducing the amount of leaf tissue as a result
of insect feeding would be that of reducing the energy available for stor
age or maintainence. Except for the controls, leaf tissue changes (Fig. 38)
closely paralleled the changes in rhizome tissue at 123%, 74%, and 57% for
the 0.33, 0.67, and 1.00 infestation levels. The control final leaf weight
was 183% of the initial weight. Hence, insect feeding directly reduced
the photosynthetic tissue present and directly or indirectly reduced the
material stored as rhizome tissue.
The proportion of the living plant weight represented as green mass
increased in the controls from 0.735 to 0.776 or approximately 106%.
This figure was somewhat greater in the plots treated with insects (see
Table 11) but does not appear to be linearly related to insect concentra
tion. This increase in proportion was probably due to a decrease in the
non-photosynthetic tissue.
The non-photosynthetic tissue (Fig. 39) not only increased less in
the control treatment (ca. 146%) than photosynthetic tissue (ca. 183%)
but also decreased more in the insect treatments. This resulted in a
decrease in the root-rhizome:shoot ratio (Table 11) at all treatment
levels. This ratio decreased less in the control plots (82% of initial)
than in the treated plots (53-59%). Again, this does not appear to be
a linear relationship. This does indicate, however, that the effect of
insect activity decreased the stored material (rhizome) as well as the
ability of the plant to absorb nutrients relative to the photosynthetic
ability of the plant.

238
While rhizome weight increased the most in the controls, roots
increased the least (Figure 40) or more acurately did not increase (102%).
This agrees with field observations where I have found that small water-
hyacinth plants have nearly the same root mass as th large plants when
growing in similar conditions. When the plants are small crowding is
minimized and nutrients would be more limiting than light. The plants,
therefore, probably maximize root growth early in their growing period.
As the plants grow and light becomes more limiting due to intraspecific
competition the available energy is probably shunted more towards shoot
development and less towards root development. This would explain the
decrease in the proportions of the plants represented as roots and the
decrease in the root-rhizome:shoot ratio observed in the controls in
Table 11.
Roots increased the least in the control and decreased the most in
the treated plots (28% of initial values with 1.00 insect per plant).
Values for final proportions ranged between 41 and 53% of the initial
proportions after insect feeding. This is not significantly different
than the control (60%). Because shading by adjacent plants is reduced
and increased mount of light is available for growth. For optimum
regeneration of the plant, nutrients need to be absorbed rapidly for
the full photosynthetic potential to be realized. Ideally, then, the
proportion of the plants represented as roots should become larger as
is normally the case in small plants. The root proportion does not
change as a result of insect attack, therefore regeneration of damaged
plants is probably slower than would be expected. Because light energy
and nutrient energy react multiplicatively in plant production the
availability of one affects the utilization of the other (i.e., growth is

239
limited by the necessary resource which is least available). When light
energy is abundant, as in a sparsely populated stand, nutrient levels
probably limit growth. One would expect the strategy of the plants in
this situation to be that of maximizing root development thereby increasing
nutrient absorption and minimizing the limiting effect of nutrients. In
a dense stand where intraspecific competition for light is intense light
availability would be expected to be limiting. In this situation selection
would favor those plants which maximized photosynthetic tissue and the
tissues necessary for photosynthetic display. Hence, a greater proportion
of the available energy would be expected to be utilized in producing
photosynthetic tissue. Small plants, therefore, would be expected
to have a larger root-rhizome:shoot ratio than large plants. The effect
of the insects in this experiment resulted in smaller plants but the
root-rhizome:shoot ratio did not become larger than that of the large
plants in the control (it in fact was generally smaller). As a result
the available light increased but the plants could probably not efficiently
utilize this increase because of the limiting effect of the absorptive
ability of the small root mass. Small slowly growing plants were pro
duced as a result instead of small rapidly growing plants or large slowly
growing plants. The insects caused not only a decrease in the standing
crop but probably also inhibited the natural regenerative ability of the
plants remaining.
The stolon weight decreased directly as a result of insect activity
(Figure 42). Stolon weight as a proportion of plant weight should reflect
offset production. If the effect of the insects was that of increasing
the offsets produced then the stolon weight and the stolon weight:plant
weight (since plant weight is reduced)should increase. This did not

240
appear to be true in this experiment. The stolon:plant weight ratio
(Table 11) was not significantly different between the infestation levels
tested. This agrees with the plant density changes discussed earlier.

241
Discussion
It is evident from these two experiments that a seasonal effect
interacts with insect activity and produces varying plant responses
depending upon the time the infestation occurs. The assumption that
insects directly cause an increase in plant density appears to be over
simplified. The canopy is reduced as is evidenced by reductions in
both height and photosynthetic material. This allows an increased pene
tration of available light and may accelerate the growth of the remaining
plants. This may be due to a cessation of rather than a direct result
of stress. Any stress factor that would reduce the effect of intraspecific
shading by reducing the canopy would probably result in a short term
increase in the number of plants present until a new equilibrium density
is achieved. Competitors may become increasingly important and suppress
this density increase by interspecific competition in the field. In
these greenhouse experiments competing species were not present so this
response may be exagerated.
The effect of insect damage to the terminal bud could possibly
contribute to an increase in the number of plants. I feel that the
contribution by insects to this process is minimal. Penfound and
Earl (1948) found that decapitated rhizomes failed to produce new
sprouts only when 4 cm of the rhizome tip was removed. The rhizomes
of the majority of the plants were thoroughly fragmented as a result
of attack by the larger larvae and the plants were dead. Fewer than 10%
of the surviving plants had rhizome damage. Those with rhizome damage
produced rather sickly offshoots. The offsets remaining after insect
attack were not the type that usually occur in open stands. The petioles
were not bulbous so as to function as floats so their ability to disperse

242
is doubtful. Also, the root mass and the root-rhizome:shoot ratio was
small indicating a loss of efficiency. The increase in plants in the
summer is more likely a vegatative response of the surviving plants
before they are severely damaged to the increased space available. In
this manner the community maintains a larger standing crop and supports
a larger insect concentration before yield is affected.
Survival over periods of stress, such as winter freezing, would
probably be reduced by insect infestations. Penfound and Earle (1948)
stated that as the leaves are killed by frost the plant tends to
float higher increasing the susceptability of the rhizome to frost.
Insects have the same effect of removing mass and causing the rhizome
to become more exposed to temperature extremes. Further, the ability
of the plant to regenerate after periods of stress is probably decreased
as a result of insect attack because of a depletion of carbohydrate
reserves in the rhizome.
Suprisingly, productivity did not decrease significantly as a result
of insect attack. This is probably the result of a time factor. That is,
the insects failed to produce a noticeable effect on plant growth until
they reached a stage late in their development. Only the later
instars do severe damage to the rhizome. Once this point was reached
productivity probably was reduced but it occurred so late in the experiment
that it failed to show up in the results. This same phenomenon was
apparent in the total biomass estimates. The only factors that did not
show a significant response to insect attack were net production, total
biomass, and plant density.
All factors associated with the standing crop showed a highly sig-
/
nificant decrease as insect concentrations increased. The proportions
of the various parts of the plants changed but these changes, for the

243
most part, were not out of line with those of the controls. This was
surprising because the smaller plants produced were expected to be
similar to the plants growing in open stands (i.e., a high root-rhizome:
shoot ratio). This did not appear to be true.
The turnover ratio was the most revealing characteristic measured.
Living material was killed and transformed to detritus at rates almost
directly proportional to the insect concentrations. The total amount
of living material initially present survived approximately 25% as long
with 1 larva per plant as did the plants without insects. This acceler
ated turnover could severely affect the competitive ability of the plants.
In summary, Avzama densa severely affected almost all aspects of
waterhyacinth growth. It does appear to be a good agent for biological
control and mass releases should be attempted. The chances for success
would be greater in the fall than in the summer and I believe concern
over a resultant increase in plant density and dispersal is unwarranted.
%
The major problem with this insect is that parasite buildups tend to
causes pulses in the effects of A. densa and as noted previously
waterhyacinths recuperate rapidly after cessation of stress factors.
Exotic insects which are not limited by parasites and exert continuous
pressure on the plant community will probably have a greater long term
possibility of providing a permanent control. Releases of A. densa
may be beneficial in conjunction with exotics to bring the plant popula
tion down to levels more easily controlled by the exotics.

CHAPTER III
THE FEASIBILITY OF THE UTILIZATION OF ARZAMA DENSA WLK. FOR
THE BIOLOGICAL CONTROL OF WATERHYACINTH THE EFFECTS
OF AN INTRODUCED POPULATION ON A SMALL POND COMMUNITY.
Introduction
It has been suggested that Arsama densa Walker could be used for
the biological control of waterhyacinth Eichhomia erassipes (Mart.)
Solms) by supplementing natural populations if a staisfactory method
of mass-rearing was developed (Vogel and Oliver 1969a). Frick (1974)
also suggested the possibility of augmenting populations of native
insects to increase their effectiveness in weed control. While this
tactic has been discussed by various investigators in the field of bio
logical control there are few examples of studies where this has been
attempted in an effort to control weeds.
Sufficient numbers of A. densa were reared on living waterhyacinth
plants in a greenhouse to release for a small scale field test. Three
variables were important in the location and timing of this release.
First, the site tested had to be small to achieve an adequate insect:
plant concentration. Second, the release had to be synchronized at a
time when parasite populations were low and the naturals, densa popu
lation was increasing. Third, the release had to be strategically
made so as to damage the plants at a time when they were most vulnerable
to attack. With this criteria in mind a small site near Paynes Prairie
in Alachua Co. was selected where the previous summer the largest buildup
of the natural A. densa population occurred in the late summer and fall.
Also, an attack late in the growing season of the plants should increase
244

245
their vulnerability to winter cold and decrease their chances of sur
viving until spring. Hence, the release was made in mid-August and ade
quate time for two A. densa generations before evaluating the effects
of the release.
I had two strategies in mind to achieve the desired results with
the initial release. The first was that of predator satiation (see
Lloyd and Dybas 1966: Janzen 1969) where a sufficient number of larvae
had to be released to satiate the predators and parasites present thus
permitting an adequate number to escape. The larvae also had to be of
a uniform age class so as to carry this phenomenon through to affect
all age-specific parasites. Egg parasites were avoided by allowing the
eggs to hatch prior to release. An ichneumonied parasite (Campoletis sp.)
is present in low populations in the fall and does not become abundant
until late winter. Hence, a large release in the fall should not be
seriously affected by this parasite. A second larval parasite (Lydella
radiis), a tachinid, attacks the seventh instar and occurs in low
concentrations throughout the year. Apparently, because in natural popu
lations of Arsama the generations overlap extensively seventh instar
larvae are almost always present and create a constant reservoir for this
parasite. This parasite is always present but since the seventh instar
populations are never large the parasite populations cannot build up.
By synchronizing the release eventually resulting in an abnormally large
population of seventh instar larvae, also synchronized, it was hoped that
the parasites would fail to make a numerical response in time to signi
ficantly affect the population. An ichneumonid pupal parasite (Chasmias sp.)
is occasionally present but was not considered a threat. I expected this

246
strategy to break down in the second generation and abnormally high
parasite populations to ultimately cause a decline in the A. densa
population.
The second strategy employed was that of an inundative release.
Assuming that we could obtain a reasonable survival rate in the first
generation a sufficient number of larvae had to be released to severely
damage the plant population before parasite buildups caused a decline
in the,A. densa population. The minimum insect concentration to achieve
this was determined to be 0.3 larvae per plant based on field observation
of natural populations of A. densa as well as greenhouse studies using
various insect concentrations.

247
Methods and Materials
Eggs of Avzama densa Walker were collected from Pontederia cordata
at Putnam Hall, Putnam Co., Florida. These were surface sterilized in
a 0.5% hypochlorite solution for 20 minutes, placed in a 10% sodium thio
sulfate solution for 5 minutes (to neutralize the hypochlorite), rinsed
in distilled water and attached to filter paper with a 10% casein glue
solution. The glue was permitted to dry and the filter paper with the
eggs was placed in the lid of a baby food jar. Waterhyacinth leaves
from plants grown in a quarantine greenhouse were similarly washed
in the hypochlorite solution and placed in autoclaved baby food jars.
The lids were placed on the jars and sealed until the larvae emerged.
These sterilization procedures were necessary to retard the growth of
mold long enough for the larvae to eclose. Fresh leaves were added as
needed for food.
The larvae obtained were kept in jars for 2-3 days. They were then
placed on tables filled with waterhyacinths in a greenhouse on 25 June
1974. On 25 July half of the plants were harvested and the larvae and
4
pupae obtained. The remainder of the plants were harvested on 2 August.
Larvae obtained were placed individually in 50 dram snap top plastic
pill vials and provided with fresh waterhyacinth petioles. After
pupation, pupae were placed on Vermiculite in a pie pan with a wax-
paper lined cage over them. Adults were permitted to emerge in the cage
and a 5% sucrose solution was provided for food. They mated and the
female oviposited on the waxpaper lining. Eggs were collected from the
wax paper and treated in the same manner as the field collected eggs.
From 14 females I obtained 2872 eggs. Approximately 78% or 2253
eclosed. Due to the failure to release these immediately high mortality

248
occurred in the jars and approximately only 1500 larvae survived.
These were released 16 August on a small pond approximately 3.2 km
south of Paynes Prairie on 1-75, Alachua Co,, Florida. The pond had a
surface area of approximately 50 sq. m. and was covered with water-
hyacinths which were approximately a meter tall. From previous data
I estimated the density of the plants at this site and time of year to
be about 85 per sq. m. Our infestation level then was approximately
0.35 larvae per plant or 30 larvae per square meter.
A second pond 0.8 km South of the first pond was selected as a
control. This site was somewhat smaller than the first but the water-
hyacinths were very similar in both density and height. Both ponds were
formed at culverts under the interstate highway and both were formed
from the same watershed.
Since I hypothesized that the effects of the insect feeding activity
would be most evident after the first frost no sampling was done until
12 December. Because of the destructive nature of the sampling and
the small size of the ponds and the amount of time required to process
each sample the number of samples taken were necessarily small. Only
three samples (0.32 sq. m. ) were selected at each site and were along
an east-west transect the first being near the west bank, the second
in the center, and the third near the east bank.
The larvae present at this time represented the F-j generation of
those released. The height of each plant was measured as well as the
plant density, leaf density, and the number of leaves per plant. Insect
damage on each plant was measured in terms of both leaf and rhizome
damage. The larvae were counted, the instar noted and any parasites
present were recorded. One sample (0.32 sq. m;.) near the center of

249
of each site was collected to estimate biomass. All living and dead
material from this sample was placed in a plastic bag and returned to
our laboratory. The plants were divided into petioles (leaf bases),
leaves (pseudolaminae), rhizome, roots, stolons, and detritus (dead
plant material). They were placed in ice cream cartons and dried to a
constant weight in an oven at 105 C for 2-3 da. After drying, the
containers were allowed to cool to room temperature and then weighed
on a Mettler to^loading balance.

250
Results
Extensive damage by A. densa was evident within two weeks at the
release site. While the initial releases were made in one small area,
the larvae rapidly spread over the entire pond. Within 50 days abundant
egg masses were noted indicating the beginning of a second generation.
At the time of sampling (December 12) the Avzama population had
increased to 52 larvae per sq. m. (Table 12) as compared to 3 per sq. m.
(living larvae at the control site. Mose of these were 6th or 7th
instar and represented the final stages of the second generation. This
population was equivalent to 0.50 larvae/plant at the experimental site
and only 0.03 larvae/plant at the control site.
None of the insects in the unusually high population at the
release site were parasitized. Nearly 45% of the larvae in the relatively
low population at the control site were dead as a result of parasitism.
After 17 weeks of insect feeding the plants in the release site were
severely damaged. As expected, the occurrence of a light frost the
first week in December accentuated this damage. Many of the plants,
although severely damaged by the insects, appeared green and healthy prior
to this time. The freezing temperatures killed a large percentage of
these damaged leaves. At the control site only the tips of the leaves
suffered damage from this initial frost. Figures 43(a)-43(h) show the
release and the control sites in a sequence up to one year after the
initial release. Figure 43(c) shows the release site after the first
frost.

251
Table 12. Comparison of the samples from the release site with the control site nased on
various estimates of the plant and insect populations. Figures represent means
and standard deviations (in parentheses), the T-statistic for unpaired data,
and the release site values represented as a percent of the control.
A B
RELEASE
SITE
CONTROL
SITE
HP*)
A/B
100
HEIGHT1' '
37.
91
(f
16.387
56
58
(
20.197
7
14
7- o.cfoiT
67
00
PLANT DENSITY/M?
109
70
(
53.17)
97
05
(
22.45)
0
38
( .0.50)
113
03
LEAF DENSITY/M2
347
05
(
193.90)
504
22
('
86.91)
1
28
(0.30)
68
83
LEAVES/PLANT
3
16
(t
1.46)
5
19
('
2.42)
7
29
(0.001)
60
89
INSECT DENSITY/M2
51
69
(
16.24)
3
16
(
3.16)
5
08
(0.01)
1635
/6
INSECTS/PLANT
0
50
(
0.09)
0
03
(
0.03)
8
39
(0.01)
1666
67
l INSECT MORTALITY
0
(
0 )
44
44
(
50.92)
1
51
(0.30)

X LEAVES DAMAGED
39
00
(
12.12)
5
41
('
4.01)
4
56
(0.02)
720
89
X RHIZOME DAMAGE
72
65
(
13.57)
16
57
(
12.72)
5
22
(0.01)
438
44
DEAD PLANTS/M2
53
80
(<
24.71)
0
('
0 )
3
77
(0.05)

% PLANTS DEAD
34
49
(-
14.22)
0
('
0 )
4
20
(-0.05)
--
1'N = 104 for the experimental site and 92 for the control site since these measurements were
taken on each plant. Otherwise N = 3 since the parameters were only estimated once for each
plot.
*P = the probability of a larger T occuring under the hypothesis A / B.

Figure 43. A photographic comparison of the waterhyacinth stands at
experimental and control sites at different times of the year following
the release of Arzama densa at the former.
a) Experimental site shortly after the release of Arzama densa
(August 1974).
b) Release site in October.
c) Release site after two generations of insect damage and after
the first winter freezes (January). The predominant plant is
Hydrocotyle. A patch of dead waterhyacinth is noticeable to
the right.
d) The release site in the spring (March 1975). Most of the water
is covered with Hydrocotyle.
e) The release site in July 1975 as the Hydrocotyle stand begins
to open up.
f) The release site one year after the initial release (August
1975). The site is dominated by cattail (Typha sp.) Notice
the small stand of waterhyacinth in the background.
g) The control site in the spring (March 1975). Compare this
with Figure 43d.
h) The control site one year after the initiation of this study.
Compare this with Figure 43f.

253
Figure 43(a)
Figure 43(b)

Figure 43(c)
Figure 43(d)

Figure 43(f)

256
Figure 43(g)
Figure 43(h)

257
Various characteristics of the waterhyacinth populations and
estimates of insect damage are compared statistically in Table 12. All
of the plants counted at the control site were alive while 35% were
dead at the release site. Only 17% of the control plants had damage
to the rhizone while 73% of the plants had rhizome damage at the release
site. Further, only 6% of the control leaves were damaged as compared
to 39% at the release site. Plant density and leaf density were not
significantly different between the two sites but both height and leaves
per plant did decrease significantly (33% and 31% respectively). This
indicates that the plants were smaller as a result of insect attack but
not necessarily fewer in number.
Biomass estimates could not be compared statistically since only one
sample per site was taken. Differences in biomass between the sites
were obvious, however, and are illustrated in Figure 44. The changes in
biomass were much greater than any of the morphological characteristics
in Table 12. Standing crop (total living plant weight) at the release
site was only 25% of that at the control site. The change in stolon weight
was greatest with a demonstrated loss of 86%. This indicates a lack of
vegetative growth in the infested plants since stolon production is
necessary for offset production. Petioles (leaf bases) and rhizomes de
creased about 80% while roots decreased only 65%. The total photosynthetic
tissue (leaves and petioles) decreased more than the non-photosynthetic
tissue (roots, rhizomes and stolons). At the release site these were 21%
and 31% respectively of their values at the control site. This difference
appears to be due to the smaller change in the roots, the only part not
attacked by Avzama.

Figure 44. A comparison of the standing crop of waterhyacinths at the
control site and the release site. Total biomass includes
both living and dead plant material. The photosynthetic
mass (green mass) includes pseudolaminae (leaves) and
petioles. The non-photosynthetic mass includes roots,
rhizomes, and stolons collectively.

o
BIOMASS [kgDW/W*]
O
PETIOLES
LEAVES
ROOTS
RHIZOMES
mn
x
m
r*
m
>
m
O
5
# *
- z¡ *
m m
STOLONS
DETRITUS
TOTAL
LIVING
TOTAL PHOTO
SYNTHETIC
TOTAL NON
PHOTOSYNTHETIC
TOTAL
BIOMASS
rv>
c_n
U3

260
Total biomass (living and dead material) was only 11% less at the
release site than at the control site. This appears to indicate that net
primary production was not dramatically reduced as a result of insect
activity. The amount of this represented as detritus (dead material)
more than doubled at the release site. Hence, living plant material
decreased, dead plant material increased, but the total of the two showed
little change.
Insect infestation did not significantly reduce the plant density,
in fact, a small increase was evident. The yield in biomass per sq. m.,
however, decreased. This indicates that the weight per plant decreased
more than is apparent from the total standing crop. Figure 45 illustrates
the biomass per plant for the various plant parts. In all cases a greater
change is noted when parameterized in this manner. The change in standing
crop, then, is not the result of the insects killing a portion of the
plants and leaving a portion intact. This would result in a smaller
standing crop but the weight per plant change would be less than or equal
to the standing crop change. Rather, insect attack resulted in a popu
lation of smaller plants. These were probably offsets produced in
response to an increase in available space as leaves from neighboring
plants died thus reducing the amount of shading in the mat.
Since the degree of change as a result of insect attack varied with
the plant parts the plant proportions must have changed. Table 13 lists
the ratios of the various plant parts at both sites. In general, the
plant proportions at the release site were typical of small plants. The
ratio of leaf weight to plant weight was similar at both sites. The
ratio of petiole weight to plant weight was less at the release site
probably due to the reduction in intraspecific competition resulting

Figure 45. The mass represented by the various plant parts for an
average waterhyacinth plant at both the control and release
sites.

AVERAGE WEIGHT PER PLANT [9m]
12 DEC. I974
CONTROL SITE
RELEASE SITE
in
no
CTi
no

263
Table 13. Ratios of plant parts at
Ratio Compared
Leaves:plant
Petioles:plant
Rhizome:plant
Root:pi ant
Stolonrplant
Root + Rhizome :shoot
Photosynthetic:pi ant
Non-photosynthetic:pi ant
Leaf¡petiole
Living:dead
the two sites on
12 December 1974
Release Site
Control Site
0.09
0.09
0.40
0.50
0.15
0.19
0.33
0.19
0.02
0.03
1.00
0.64
0.49
0.59
0.51
0.41
0.23
0.17
0.24
2.10

in less need for supportive tissue to display the leaves. The pro
portion of the plant represented by roots increased and the root-
rhizome:shoot ratio increased. This is typical of small plants
growing in open stands. The photosynthetic tissue ratio decreased
relative to the non-photosynthetic tissue. This is in contrast to
similar experiments in a greenhouse (see previous section) where
Avzcoma feeding appeared to result in a decrease in the root rhizome:
shoot ratio. Since the surviving plants had a well developed root
system at the release site they would have probably recovered if the
insects had been removed and if the season was favorable.

265
Discussion
I assumed at the beginning of this experiment that parasite
populations would increase and ultimately reduce the introduced
Arsama densa population at the release site. By the end of the
second generation, however, this had not become apparent. The popu
lation of F-| seventh instar larvae was 72% greater than the population
of first instar larvae initially released. Furthermore, none of the
larvae collected at the release site were dead as a result of para
sitism while 44% of those at the control site were. Not only did the
initial population survive and reproduce contrary to my expectations
but it increased in the subsequent generation which was apparently
also surviving well.
Since this study was not designed to evaluate the population
dynamics of A. densa I can only speculate on the reasons for the success
of this population. I feel that by synchronizing the generations the
consequences of parasitism were effectively reduced. The parasites at
this site may have been "programmed" to low host populations and over
lapping generations. By inundating the natural population which had
individuals at various stages of development with the introduced
population with indidivudals all at the same stage of development the
synchronization of the age-specific parasites with the host may have been
imbalanced. For example, if parasites of the seventh instar larvae were
issuing at the time of release of the first instar larvae the ability
of the parasite population to increase would not change. Only those
parasites that are present at the time that the introduced host popu
lation is at an appropriate age would have an increased chance of

266
ovipositional success. Hence, the parasite build-up would be slower
than expected as long as the host population remained synchronized. This
assumes negligible recruitment to the parasite population from outside
the release area. This is probably not a valid assumption but the
parasites did fail to control A. densa before sufficient damage was done
to the waterhyacinth population. I plant to conduct a similar study
including a life table analysis of a field released population of A. densa
in the future.

The responses of the plant population to insect feeding tend to
verify similar experiments using greenhouse cultures of waterhyacinth.
Plant density increased a small amount but this was probably an
opportunistic response to available space as the larger plants initially
present died and light penetration increased. The degree of this
response appears to be related to the quantity of light available and
is probably seasonal in nature. The plants produced were much smaller
thus floating higher in the water and were probably more susceptible
to temperature extremes.
A reduction in the leaf canopy was apparent as height, leaf density,
number of leaves per plant, and the biomass of photosynthetic organs
decreased. This does not necessarily indicate a reduction in net pro
ductivity since smaller plants tend to be more efficient than larger
ones (Browne et al. 1974). The proportion of the plant represented as
leaf blades did not change while the proportion represented as roots
increased. The major change appeared to be in the petiole proportion
which is primarily a structure for supporting and displaying the leaves.
In this open stand this supportive structure would not be as important
as in a dense stand. As a result the relative photosynthetic ability
of the plant per gram of biomass was probably not affected but the
relative ability to absorb nutrients probably increased. This is
reflected in the increased root-rhizome to shoot ratio. In the green
house studies a decrease in the root-rhizome:shoot ratio was evident.
This is in contrast to the increase noted here. Perhaps, in the former
case, an insufficient amount of time was available for root regrowth
prior to harvesting. The plants produced at the release site in the
field experiment were much more typical of small plants growing in

268
open stand than were those in the greenhouse experiments.
The small difference in total biomass present tends to substantiate
the feeling that net productivity was not reduced. A number of
explanations are possible for this. Productivity at this time of year may
be low, hence what was measured was merely the amount initially present.
This is possible since I have noted in other studies that the standing
crop begins to decline in late summer. A second explanation is that the
insect infestation decreases intraspecific competition thereby increasing
the productivity of the remaining plants. This increased productivity
may make up part of the difference caused by the insects.
Jameson (1963) pointed out that carbohydrate storage is directly
correlated with winter hardiness. If the primary organ of carbohydrate
storage in waterhyacinths is the rhizome (Penfound and Earl, 1948) then
the feeding activity of Avzama severely reduced the carbohydrate reserves .
This is illustrated in Figure 45 where the rhizome weight per plant at the
release site is only 15% of that at the control site. Harris (1973)
stated, however, that an injury that lowers carbohydrate levels either
directly or indirectly by stimulating auxin production and growth is
likely to be partly compensated for by an increase in photosynthetic
efficiency.
Penfound and Earle (1948) found that the rhizome length remains
fairly constant throughout the growing season. They attributed this to
an equilibrium between rates of decay at the older portion and rates of
increase at the crown. If this is true we may assume that carbohydrate
reserves also remain fairly constant and at the end of the growing season
are sufficient to maintain the plant through the winter and provide

269
sufficient energy for regrowth in the spring. By reducing these reserves
the effect of the insect infestation was that of decreasing the winter
hardiness of the plants and preventing regrowth in the spring.
While total biomass showed little change between the two sites the
turnover rate obviously was greater at the release site and accounted for
the decrease in standing crop. This is manifest in the increased detrital
production. The ratio of dead:living plant material was almost nine
times greater at the release site than at the control site.
Following the decline in the waterhyacinth population several changes
in the species composition were noted at the release site. In August,
at the time of release, the pond was covered with a pure stand of water-
hyacinth (Figure 43(a)). By October a large proportion of the leaves were
beginning to wilt and become brown and Hydrocotyle had begun to appear
amongst the waterhyacinth plants (Figure 43(b)). In January only a few
patches of dead waterhyacinth were evident (Figure 43(c)) and hydrocotyle
dominated the surface although there were some areas of open water.
Hydrocotyle increased and by March (Figure 43(d)) it covered the entire
surface of the pond. In July the surface had again begun to open and
hydrocotyle was less dominant. A mixture of Hydrocotyle, Polygonum,
Bidens, and Ludwigia was present and the small stand of Typha had begun
to expand (Figure 43(e)). Waterhyacinth was again present but only in a
small patch on the southeast side. By mid-August Hydrocotyle was present
only in small patches and Typha was dominant. Waterhyacinth was present
in a pure stand the whole year at the control site with the exception
of a small fringe of Hydrocotyle which appeared in the spring.
I had expected waterhyacinth to reoccupy and dominate the pond at

270
the release site by early summer but, this did not occur. Apparently
the other species had sufficient time to become established and prevent
the spread of the waterhyacinths into the center of the pond. All of the
plants mentioned earlier are rooted and may form a physical barrier to
the waterhyacinths. The long term success in controlling waterhyacinth
experienced at this site would probably not occur where the water was
too deep for rooted emergents to gain a firm foothold and occupy the
space available. In this situation the waterhyacinths would readily
float in from other areas and again become dominant. Nevertheless, I
feel that this study has shown not only that waterhyacinth is vulnerable
to biological control and that this control can be achieved, at least
temporarily, by the manipulation of populations of native insects. It has
also proven the overall effectiveness of Arzama densa as a control agent.
Harris (1973) has suggested that in some cases we may need an infestation
of one insect to reduce the infestation of a weed and a second one to
keep it low. Perhaps indigenous insect populations, such as Arzama densa,
can be used to initially reduce the weed infestation with subsequent
releases of exotic insects, such as Neoahetina spp., to exert a more
constant stress and maintain a low weed infestation.

CHAPTER IV
NOTES ON THE BIONOMICS AND POPULATION DYNAMICS OF ARZAMA DENSA WLK.
Introduction
Arzama densa Walker (Noctuidae: Amphipyrinae) is a large moth whose
larvae are semi-aquatic in habit. The taxonomy of its species group is
poorly understood and deserves further attention (see literature review
section). It is obviously closely allied to the species generally included
in the genus Bellura as both adult morphology and larval habits show
striking similarities. The separation of these species into two genera is
based largely on the armature of the frons and characteristics of the
antennae. Todd (pers. com.) feels that the validity of these characters is
questionable and two are probably congeners. He therefore proposes the com
bination Bellura densa (Wlk.) as proper for this species and further sug
gests that it may be conspecific with B. gortynoides Wlk. as proposed by
Smith (1893). Because the status of this group is questionable I have used
the name Arzama densa Wlk. throughout this dissertation. It must be pointed
out that this may not be a valid name and future taxonomic studies are
needed. While I agree that Bellura and Arzama are probably congeneric, I do
not feel that A. densa and B. gortynoides are conspecific. I have had the
opportunity to observe both species in the immature stages and from the
viewpoint of a nontaxonomist they certainly appear to be distinct. In order
to resolve the taxonomic relationships within this group larval characters,
oviposition behavior, host plant preferences, and other aspects of the bio
logy and immature stages should be considered in a bio-systematic approach.
This study was not designed to investigate the life history of Arzama
densa as it was originally assumed that this had been adequately investi
gated by Vogel and Oliver (1969). In the course of my experiments
271

272
and field studies, however, I have accumulated many observations on the
bionomics of this species which I feel are significant. I will present-
these observations at this point, but since the methods employed are
diverse, I will omit a methods section and instead explain them as the
results are presented.
Habits
The eggs of A. densa are laid in a mass on the upper surface of the
leaves of Eiohhormia arassipes (Mart.) Sol ms or Pontederia sp. The egg
mass is very similar to that described by Riley (1883a, p. 174) for Arzama
obliquata (= Bellura obliqua) as "broadly convex or plano-convex masses
enveloped in hair, and a cream colored mucous secretion". Vogel and
Oliver (1969) described the masses as being covered with light yellow
body hairs. The egg masses I have observed have been tan or creme colored
rather than yellow. Vogel and Olvier (1969) further indicated that each
egg mass contained 30-40 eggs. From 11 masses collected in 1973 from
waterhyacinth I have found the range to be 19-66 with an average of 42
(+ 14.57 s.d.). Further, I have frequently found eggs laid singly or
in groups of two or three often on the petiole of a leaf near the plant
base. These are usually not provided with a covering and may be left
by a resting female as an artifact of a previous egg extrusion. I have
frequently noted a few eggs clinging to the abdomens of caged females
after oviposition. These remnants may account for the single ovipositions
noted in the field.
From these same caged females I have observed that the eggs are not
deposited in clusters when provided with an artificial substrate (i.e.
wax paper, guaze). When leaf bouquets are available the eggs deposited
/
on the leaves are usually in the typical masses. This suggests that this

273
clustering mode of oviposition is somehow stimulated by the plant.
First instars were rarely encountered in the field. This appear
to be due to their short stadium, small size, and inconspicuous feeding
damage. Vogel and Oliver (1969b, p. 251) indicated that "young larvae
were found feeding on tender basal stems and young foliage" but made no
specific reference to the first instar. Welch (1914) reported that Bellura
melanopyga Grt. went though two feeding periods. The first was a leaf
mining period in which the early instars feed in the leaf blade of
Nymphaea [-Nuphar) americana. The second period was a petiole period
in which the older larvae burrowed into the leaf petiole. This
phenomenon was also observed by Claassen (1921) with Bellura obliqua
(Wlk.) on Typha latifolia L. I have found in the laboratory that when
neonate A. densa larvae are provided only with waterhyacinth leaves they
will form leaf mines and feed between the upper and lower epidermis.
They will also feed externally aggregating in folds in the leaf blade.
In the field I have never found the first instar in leaf mines. I have
found them in shallow burrows in the petiole just under the epidermis
in the region of the leaf isthmus. I have more frequently found them
at the base of the plant usually between a leaf petiole and the leaf
sheath or a wrapper leaf. This is consistent with Vogel and Oliver's
(1969b) observations. In a few instances where I have found egg masses
on Pontederia there was evidence of larval mines within the leaf blade.
Usually it appears that the larvae migrate from the egg mass, however.
Small exit holes through the leaf are frequently present under an egg
mass indicating that the larvae burrow through the leaf upon enclosion.
Those individuals which do mine the leaves would be easily overlooked
and may partially account for the relatively few first instar larvae

274
observed in the field.
Second and third instars are found in a variety of places. They
are most often located at the base of the plant frequently between
two tightly appressed petioles or under a wrapper leaf usually feeding on
new leaf growth. Occasionally they will form burrows within a petiole
or shallow grooves on the outside of a petiole. By the fourth instar they
become almost exclusively internal feeders, boring the petioles and feed
ing on the apical tip of the rhizome. By the sixth instar they create
large burrows doing considerable damage to the petioles and may bore
three or four centimeters into the crown. The most extensive damage is
created by the seventh instar. The tunnels may extend into three or four
adjacent petioles, the full length of the rhizome, even through the stolon
into an adjacent plant. The damage to the rhizome may be so extensive
as to cause severe rotting and fragmentation.
Pupation occurs within a petiole usually in the basal portion. A
pupal chamber is hollowed out and a window is opened 2-3 cm. above the
pupa to permit egress of the adult. No cocoon is formed although a
silken suspensory apparatus may be constructed below the abdomen to
cradle the pupa within the burrow. The pupa is oriented parallel to the
long axis of the petiole with the head toward the distal end. Occasionally
pupation occurs in the rhizome.
The adults seem to rest during the day withing the foliage of the
waterhyacinth mat or in the vegetation along the shoreline. They are
quite active at night and are frequently collected at light traps (Frost
1975). I have found in the laboratory that females may mate and oviposit
within a few hours after emergence when caged with males in the dark.
This indicates a very brief pre-ovipositional period. To confirm this I

275
dissected a 10 day old female pupa and found fully formed eggs. It
appears then that oviposition can occur almost immediately after
emergence and mating.
Fecundity
Several pupae were collected from waterhyacinths in July 1973.
These were individually placed in baby food jars which were held in sealed
aquaria lined with damp paper towelling. Upon emergence a single male
was paired with a single female in a one gallon ice cream carton. The
adults and pupae were held in an environmental chamber at 25 C, 16:8 L:D
photoperiod. In all, five pairs of adults were obtained from the field
collected pupae. The cartons were checked daily for eggs. The eggs were
removed and counted, held in baby food jars in the manner described in
Section 2, and allowed to eclose. The neonates were removed each day and
the egg developmental time noted. These results are summarized in
Table 14. The average fecundity was 225 eggs/female, 72.6% were laid on
the first day although oviposition generally continued for three days.
Viability was 80.7% and the average developmental time was 5.6 da.
Fecundity was checked again when larvae were reared for a field
release (Section 3). The pupa were placed in vermiculiteand held
in a cage constructed of hardware cloth and lined with wax paper. In
stead of isolating pairs of adults all were kept in the cage and fed
a sucrose solution. A total of 18 males and 17 females were reared
(C?:sex ratio = 1.06:1). Three females emerged after the death of
the last male and did not contribute fertile eggs. From the 14 females
that did mate 2872 eggs were obtained. This represents an average of
205 eggs per female. Only 4.49% of the eggs were sterile, 17.06% were
fertile but failed to eclose, and 78.45% eclosed. Both of these estimates

276
(205 and 225 eggs/ ) are lower than that of Vogel and Oliver 1969b..
They reported an average of 328 eggs per female with 8.25% infertility
based on 10 mated females.

277
Table 14. Fecundity, egg viability, and egg stadia for 5 female
Arsama densa collected as pupae in the field and mated
in the laboratory.
No.
#Eggs/
%Laid on
Average egg
# Eggs
%
first day
stadium (da.)
Hatched
Hatch
1
194
87
5.0
167
86.1
2
228
74
5.3
186
81.6
3
257
60
6.0
252
98.0
4
262
60
6.5
176
67.2
5
184
82
5.0
130
70.7
X
225
72.6
5.6
182
80.7
S.D.
35.5
12.4
0.7
44.4
12.4

278
Duration of Developmental Stages
First, second, and third instar larvae were obtained from field
collections and from laboratory reared material. Larvae were placed
individually in 1 oz. diet cups and provided with petiole sections from
either Eichhomia crassipes or Pontederia cordata. The cups were kept
in an enviromental chamber at 25 C and 16:8 L:D photoperiod. The
plant material was checked daily and replaced as needed. The larval
instar was also checked daily and recorded for each cup. Head capsules
were saved and later measured. Table 15 summarizes the developmental data
from this study. Only three larvae pupated and this occurred following
the seventh instar. I feel that this is the typical number of instars.
Other larvae went into eighth and ninth instars before death occurred.
The presence of extra instars is typical in laboratory reared Lepidoptera
when under stress (Leppla, pers. comm.). Several factors may have been
responsible in this study. The cups used to contain the larvae were
samll. They were translucent and not transparent. The larvae may need
rhizome material at some stage in their life and they were only fed
petioles. Humidity was high in the cups as condensation was frequently
noted. Hence, light, humidity, space, food quantity and food quality
could have become stress factors ultimately resulting in these extra
molts. Developmental data was included for these individuals only
through the seventh instar.
Total developmental time was estimated indirectly from the
greenhouse experiment testing the effects of A. densa damage on water-
hyacinth (Section 2). The approximate age of the larvae was known at the
time of the release. The proportion represented as larvae, pupae, and
282

Table 15. Summary of developmental data for A. ier\si.
'
1
2
Average
for all
Sugarcane borer diet
Diet
Eiakkomia
crassipes
Fontedria cordata
larvae from 1 & 2.
(from Vogel S Oliver 1969)
Temperature (C)
25
25
25
21
Photoperiod (L:D)
14:
10
14:
10
14:
10
14:
10
Instar
# Observed
X (iS.D.) da
? Observed
X (:S.D.) da
# Observed
X (:S.D.) da
# Observed
X (S.0.) da
Egg
5 (1125)
5.6:0.7**
5 (1125)
5.6:0.7**
5 (1125)
5.6:0.7**
14
5.4:0.5
1
36
3.3*0.5
25
3.2:0.4
61
3.2-0.4
14
3.8:0.7
2
31
3.2*0.8
22
3.5:1.1
53
3.3:0.9
14
5.2:0.8
3
26
3.8*1.1
21
3.3:0.6
47
3.6:1.0
14
5.6:1.6
4
34
4.7*1.5
15
4.5:1.2
49
4.7:1.4
14
5.4:2.0
5
30
5.5*1.5
13
7.7:1.4
43
6.1:1.8
14
7.1:2.2
6
17
6.9*1.7
11
10.9:2.5
28
8.5:2.8
8
12.5:2.6
7
8
10.8:3.0
4
12.5:1.3
12
11.3:2.6
2
9.5:0.7
Pupa
0
N.O.*
3
10.0: 0
3
10.0: 0
14
14.6:1.8
Total
--
53.8*

61.2

56.3

69.1
* Assumes 10 da pupal stadium although none were observed in these experiements. Total egg to egg duration estimated from
the sum of the means for all stadia.
** Egg developmental data is the same for both host plants since the larvae used were from the same egg masses. The mean is
derived from the average of 5 estimates representing a total of 1125 eggs (see Table I).
279

280
and adults (estimated from pupal exuvia) was measured at the time of
harvesting (35 and 44 da post egg collection).The larvae collected were
fed petiole sections until they pupated and the pupae were held until
they emerged. A cumulative tally was kept on the number of adults
emerging and the number of eggs deposited. These figures were later
converted to daily cumulative percentages based on the final totals.
The probit was derived from a conversion table for the cumulative
percentage of each of these and plotted against the log of the number
of days following egg collection. By fitting a line to these points
the day upon which 50% of the population transformed into pupae or
adults or the time at which 50% of the eggs were deposited was esti
mated. These probit analyses are illustrated in Figure 46. This
technique should be valid since the probability of occurrence of these
events is a sigmoid curve as a function of time. The vertical vectors
in Figure 46 indicate the points that the probit line crosses the 5.0
probit. This represents the 50% probability for occurrence of that
event and in a normally distributed population estimates the mean (see
Andrewartha, 1961, p. 65). The estimated dates of pupation, adult
emergence, and ovposition are 42, 50, anid 52 days respectively. It
may be noted that this predicts the emergence of the adults 2 days prior
to oviposition. This suggests a preoviposition period which is in
consistent with my previous findings. When the emergence of males and
females is plotted separately, however, the predicted average emergence
date for the males is 48 da and for the females 52 da. This is consistent
with the lack of a prolonged preoviposition period noted earlier.
Considering that these time periods are established from the
date the eggs are collected and not from the date of oviposition the

Figure 46. Probit analyses for developmental times of a greenhouse
reared population of Arzama densa. The vertical vectors
represent the dates at which the 5.0 probit (50% probability)
for pupation, adult emergence, and oviposition respectively
were reached.

PROBIT
282

283
values may be underestimated. The addition of 6 da (egg developmental
period) may be added to these figures to bracket the estimate. Hence
pupation occurs 42-48 da post-poviposition, adult emergence 50-56 da
(males 48-54; females 52-58), and oviposition 52-58 da. This conforms
to the hypothesized developmental schedule given in Table 15 developed
from direct laboratory observations. By summing the developmental times
for each instar in Table 15 pupation is expected at 46 da post-oviposition
and adult emergence at 56 da. This is considerably less than Vogel
and Oliver's (1969) estimates but they reared their larvae on an arti
ficial medium and at a lower temperature. I suspect that these two
factors account for the difference.
In attempting to rear large numbers of A. densa larvae for other
experiments I have used this data to estimate the time at which I should
harvest the plants to obtain primarily pupae. Eggs were collected in the
field and neonates reared in the laboratory. The first instar larvae
were placed on waterhyacinth plants in ten large troughs in an airhouse.
The plants were harvested from each trough between 43 and 48 da after
the collection of the eggs. In all cases 50% or more of the insects
recovered were pupae. Hence, I feel that the developmental data pre
sented here is reasonably accurate.
The data for head capsule measurements and the larval age at
each molt is summarized in Figure 47. The cross bars represent the
standard deviation for each parameter and the point of intersection
represents the means. The figure to the right and the figure below
represent the number of observations in the mean for the head capsule
measurement and the age respectively. The figure above represents the
instar these figures are derived from. I have found the head capsule

Figure 47. The head capsule diameter of Avzama densa larvae at each
molt plotted against the larval age. The data was derived
from larvae reared in cups containing pickerelweed or water
hyacinth petiole sections in an environmental chamber at
25C and 14:10 L:D photophase. The stars represent the
head capsule size of 6th and 7th instar field collected
larvae and extend the trend to that expected had the larvae
developed "normally."

Head Capsule Diameter iirnn)
285
Larval Age At Molt (da post eclosin)

286
measurements for the first 5 instars very useful for identifying the
stages of field collected larvae. The values for the sixth and
seventh instars, however, seem to be much smaller than those from the
field. These data are suspect and the smaller size may have resulted
from the rearing conditions.

287
Population Cycles
Vogel and Oliver (1969b) felt that Avzama densa produced at least
2 and possibly 3 generations per year in Louisiana. Their estimate was
based on their determination of the length of the developmental period
and a 120 day winter diapause. Larvae were present at all times of the
year, however, and appeared to be most abundant in September and October
(Vogel and Oliver 1969a).
Larvae were sampled from 1 May 1974 to 30 April 1975 at weekly
intervals on Lake Alice, Gainesville, Florida from the sampling plots
described in Section 1. The larvae collected were placed in snap top
pill vials and returned to the laboratory. They were maintained by
feeding them waterhyacinth petiole sections. Each larva was classified
as to instar at the time of collection. If they died I attempted to
establish the cause of death. When parasites emerged they were identified .
If no parasites issued shortly after the death of the larvae they were
checked for diseases by Dr. George Allen. Dead larvae collected in the
field were handled the same way. Parasite pupae found in Avzama
burrows were collected and reared. Pupal exuviae were noted as well as
the probable instar of the dead host larva.
Figure 48 illustrates the total population density of A. densa
larvae and pupae on Lake Alice for the 1974-75 sampling period. This
includes all larvae both living and dead. The same data is presented
in Figure 49 by individual instar. While eggs were seldom encountered
by sampling they were noted in the field from September through March.
This information indicates the existence of continuously brooded, over
lapping generations. Eggs were most abundant in January and February
which leads me to believe that no winter reproductive diapause occurs.

Figure 48. The total number of Arzcona densa larvae collected, either living or dead, from the
marsh side of Lake Alice. Each point represents a mean per square meter derived from
three 0.316 mz samples collected at weekly intervals from May 1974 through April 1975.

Total Larvae & Pupae / M
68Z

Figure 49. The age structure of the Arzama densa population during the period of this study.

POPULATION ESTIMATES (LAMV.'M')
!
I\D
VO

292
A summer decline in the population is evident from the data. From
early May through late July A. densa is rare on Lake Alice. I have
frequently required eggs during the summer for various reasons. Since
they were unavailable on waterhyacinth, stands of Pontederia sp. at
Putnam Hall, Putnam Co., Fla. and at Paynes Prairie, Alachua Co., Fla.
were checked. Eggs and larvae were found to be present throughout the
summer. No population studies were conducted on this host plant but the
need for these studies is evident. Pontederia seems to be the primary
host for A. densa and an understanding of the population cycles on this
host would be an invaluable in interpreting seasonal population differences
on waterhyacinth.
Mortality
Because of the lack of data with regard to egg counts I have very
little information on egg mortality from populations on waterhyacinth.
From 9 egg masses collected between September and December 1974 the range
of egg parasitism by the scelinoid Telenomus arzamae Riley was between
0 and 68% (average = 26%). Because mortality was not always complete
these figures are affected by the age of the eggs. Further, many egg
masses were collected after the larvae and the adult parasites had
emerged and it was difficult to determine whether the chorion of a
particular egg had been vacated by a larva or a parasite.
To further estimate egg mortality 80 eggs were collected from a
caged female. The eggs were laid on gauze and were deposited in single
layers rather than in the typical convex masses. The gauze was cut up
so as to partition the eggs into 4 groups of 20 eggs each. These were
placed on waterhyacinth leaves on a small pond near Paynes Prairie on
4 September 1973. They were left on the plants for 4 days, recollected

and placed in petri dishes until all of the egg parasites emerged. By
17 September (13 da. post-oviposition) 80 adult T. arzamae had issued
resulting in 100% egg parasitizm. This leads me to believe that the
layered conformation of the typical egg mass protects the lower layers
of eggs from this parasite. Subsequent preliminary examinations of
parasitized egg clusters indicate that only the outside layer of eggs
are parasitized. The thick coating over the eggs probably prevents the
parasites from working their way down in between layers. Because of the
short ovipositor of this parasite only the outermost layer of eggs is
vulnerable to attack. Since only about one third of the eggs are so
protected I would expect the maximum egg mortality due to this parasite
to be about 67%. I have collected eggs for rearing purposes from various
locations at all times of the year and have found T. arzamae continually
present.
Vogel and Oliver (1969b) indicated that a second egg parasite,
Anastatus sp., was present in Louisiana. I have not found this in
Florida. I have found the ladybird larva, Coleomegilla imaculata DeGeer,
commonly feeding on the eggs on Pontederia sp. Vogel and Oliver
(1969b) also list this species. I have further found an unidentified
cecidomyiid larva commonly attacking the eggs on Pontederia.
Figure 49 illustrates the structure of the A. densa population based
on larval and pupal instars. The susceptibility of a particular instar
to sampling is dependent upon the length of the stadium and the prominence
of the damage. The duration of the seventh larval stadium is approximately
three times as long as the first larval stadium. Damage by the seventh
instar larva is very conspicuous and easily detected while damage by
the yaungerlarvae is less so. I believe that this accounts for the

294
apparent predominance of the latter stages in the life cycle. This
together with the large degree of overlap between generations makes
analyses of age-specific mortality factors extremely difficult.
Figure 50 illustrates the annual curves for the density of larvae
and pupae collectively and population mortality. The percent mortality
is derived from the number of individuals found dead in the field
relative to the total living and dead. The mortality curve shows a
configuration very similar to the population curve but lags slightly
behind it. This would be expected when most of the mortality is due to
parasitoids. Higher parasitism occurs when the population is high but
mortality as a result of this parasitism does not occur until somewhat
later.
The data for the total annual population counts and the proportion
of each instar found dead is summarized in Table 16. It is apparent
that most of the mortality observed in the field occurred during the
fourth and seventh larval stadia. This was primarily the result of
parasitism by Campoletis sp. oxylus grp. (Ichneumonidae) to the fourth
instar and by Lydella sp. (Tachinidae) to the seventh instar.
Figure 51 summarizes the population data for the 4th and 7th instar
larvae. Illustrated for each instar is the percentage of the total found
dead in the field and the percentage parasitized but still living. The
dashed lines indicating the number escaping is an estimate of the number
of living larvae free of parasites and pathogens. Parasitism by Lydella
and Campoletis is high throughout the year. A few 7th instar larvae
die as a result of infection by the microsporidian Nosema neoatrix
and other causes but the majority of the total observed mortality is a
/
result of parasitoids.

Figure 50. The population of living Arzama densa larvae (per square meter)
present on Lake Alice and the number of dead larvae expressed as
a percentage of the total.

296

Table 16. Annual summary of larval counts and mortality.
Instar
Total
Collected
# Dead
%
Mortality
1
9
0
0
2
44
1
2.27
3
27
2
7.41
4
55
29
55.53
5
28
2
7.14
6
25
2
8.00
7
83
30
36.14
Pupae
6
0
0
Pup. Ex.
2


Total
279
66
23.66

Figure 51. The total number of 4th and 7th instar Arsama densa
larvae per square meter as estimated from samples taken
from the marsh site of Lake Alice. The solid lines
represent the total number of larvae encountered. The
dotted lines represent the number of that total which
showed no signs of parsitism or diseases. The vertical
bars depict the percentage of the total affected by the
various mortality factors.


300
Because of the extreme degree of overlap in generations and be
cause the influence of outside populations from a second host plant is
not know analyses of the population dynamics and mortality factors is
extremely difficult. In general, I hypothesize that mortality of the
eggs, 4th instar larvae, and 7th instar larvae are the important
factors regulating the population. Following is a hypothetical explanation
of the seasonal trends based on the available data.
During the summer, populations of host larvae are very low, hence,
the parasitoid populations would be expected to be low. As a result
a larval buildup becomes possible in the fall and early winter as the
parasite populations fail to respond quickly enough to eliminate the
host. The parasite buildup is probably slow as the host population never
becomes exceedingly high. As a result enough larvae escape to permit
the observed fall buildup. By early winter a general decline becomes
apparent in the A. densa population presumably as a result of the increased
parasite pressure. The population appears to begin to oscillate at this
time. By late winter 100% mortality occurs consistently to both 4th and
7th instar larvae. The final population buildup occurs in early spring.
Apparently the parasitoid population is high and allows few host larvae
to escape. The few 7th instar larvae that survive to pupate emerge as
adults, and oviposit in the spring producing progeny subject to severe
parasite pressure. The majority of the larvae escaping the egg parasites
are parasitized by Campoletis sp. and fail to survive through the fourth
instar. Those that do survive are subjected to further parasitism by
Lydella sp. By late spring the population is at a very low level and
remains so throughout the summer.

Figure 52. The number of parasites of 4th instar (Campoletis sp.) and 7th instar
[Lydella radiis) Arsama densa larvae as estimated from the number of
pupae, or pupal exuviae found in A. densa bores per square meter of
waterhyacinth mat.

o
PARASITE POPUI ATION ESTIMATES
zoe

Figure 52 illustrates estimates of Lydella radiis and Campoletis sp.
populations during the study period. The parasite numbers are based
on the number of pupae or puparia and the pupal exuvia found associated
with dead A. densa larvae. A single Campoletis pupa is usually found
in association with a 4th instar larva although on at least one occasion
two pupae were found in association with a single larva. Lydella, on
the other hand produces 1-5 puparia per 7th instar larva. The normal
range is more on the order of 2-3. This accounts for the more prominent
peaks observed in the parasite populations for Lydella. Although
parasite populations of both species appear to be more intense in the fall
and early winter, the effects of this intensity may be tempered by the
relatively high asynchronous host population present at this time. While
a large number of larvae are being removed by parasites they are replaced
by younger larvae. Initially the parasites cannot respond numerically
fast enough to take advantage of the newly recruited larvae.
Since the parasite populations are dependent upon thelevel of the
host populations they remain at a fairly low level. As the parasite
populations gradually increase a subsequent decline in the larval popu
lation begins to take place. As this continues the number of parasites
present relative to the number of host larvae present increases. As a
result the parasite populations are able to more fully exploit the avail
able host populations in the spring. Ultimately the recuitment of new
individuals is insufficient to maintain the host population and a
dramatic decline occurs. The low host populations in the summer results
in a decline in the parasite populations. This permits the subsequent
buildup of the larval population in the fall. It is not apparent whether
the source of this fall population is the low population present earlier

304
or if it is the result of immigration from other populations and
possibly from the other host plant [Pontederia).
A parasite of A. densa pupa has been encountered occasionally
although never from the Lake Alice population. It is an ichneumonid
(Chasmias seelestus Cr.) and in two years of collecting larvae and
pupae has only been found twice. I doubt that this species has a
serious impact on the A. densa population. Vogel and Oliver (1969b)
listed two other species, Ichneumon n. sp. (Ichneumonidae) and
Eupteromalus viridescens (Walsh), as pupal parasites of this host in
Louisiana. Table 17 is a list of the various parasites attacking A.
densa. At least seven parasites have been associated with A. densa
representing 5 different families of insects. At least four of these
occur in Florida.
Discussion
Arsama densa Walker (1865) occurs naturally in Florida feeding on
pickerelweed (Pontederia sp.) and has extended its host range to include
the introduced waterhyacinth (Eichomia crassipes (Mart.) Solms). The
larger larvae are capable of causing severe damage to waterhyacinth when
populations reach high levels. This has been observed in the field but
these outbreaks are generally very localized and of short duration.
Severe pressure by a diverse parasite complex appears to restrict such
outbreaks and maintain A. densa populations at low levels on waterhyacinth.
The results of the studies from Lake Alice indicate that many of
the larvae escape parasitism in the fall but since host stages are con
tinuously available the parasite populations eventually build up and
suppress the A. densa population. This suggests the possibility of the
manipulation of A. densa populations for the control of waterhyacinth.

Table 17. A summary of insects known to parsitize Arsama densa Wlk. (from Vogel
and Oliver 1969b in part).
Fami1y
Species
Host stage attacked
Locality
Seelionidae
Telenomus arsama Riley
eggs
Fla., La
Eupelmidae
Anastatus sp.
eggs
La.
Tachinidae
Lydella radiois (Townsend)
7th instar larvae
Fla., La
Icheneumonidae
Ca/npoletis sp. oxylus group
4th instar larvae
Fla.
Ichneumon n. sp.
pupae
La.
Chasmias scelestus Cr.
pupae
Fla.
Pteromalidae
Eupteromalus viridesccns (Walsh)
pupae
La.

306
First, since the parasites are always present but in low numbers
due to the relative rarity of host material perhaps large quantities of
larvae may be released to supplement the natural population. If this
release is made in the late summer or early fall when enough larvae
are escaping parasitism to permit a natural population increase perhaps
sufficient numbers of larvae may survive to cause extensive damage to
the waterhyacinth crop before excessive mortality occurs. This is based
on the assumption that the parasites will be unable to make a numerical
response in time to exploit the host population.
Second, if the parasites are "programmed" to the asynchronous
overlapping generations of the host possibly the host population can be
forced into synchrony. The age-specific parasite population is permitted
to increase as the host population increases because of the continuous
availability of the proper stage host. If inundative releases are made
of larvae similar in age the probability of successful oviposition by
the parasite would only increase during the time period that the host
population is at a suitable age. Because the parasites issuing from
previously parasitized hosts would be representative of the natural
population the levels would be too low to exploit the introduced popu
lation. Further, those parasites issuing from the introduced population
would not be able to further parasitize the introduced population unless
its life cycle is of equal duration to that of the host. Assuming this
is not the case, the majority of the parasite population would not
immediately be able to locate suitable host material and the progeny
of the released population would suffer low mortality.
This strategy is similar to the strategy of predator satiation

¡07
described by Lloyd and Dybas (1966) for the periodical cicada and by Janzen
(1969) for seeds. In both cases it was suggested that this strategy employs
a sudden increase in the population of susceptible individuals allowing
them to escape before their resceptive predators could respond numerically.
The synchronization of age involving the escape of insects from parasites
may be considered part of this strategy. This, in effect, minimizes the
time period available for successful oviposition by the age-specific
parasite and again requires a rapid numerical response. Since the interval
between instars suitable for a particular parasite would be maximized
the probability of continual parasitism would be minimized.
Large scale testing of this theory of augmentation of the Arzama
densa population is precluded by the inability to mass rear suitable
numbers of larvae for release. Further, a great deal of basic research
on the biology and population dynamics of both Arzama densa and its para
sites should be completed. The potential for control of waterhyacinth by
this insect species exists and warrants further study. It may be used in
other countries simply by importing it free of parasites and predators
or in this country through more complex means similar to those suggested
above. A thorough understanding of the host specificity of this insect,
its taxonomic status, and its populations on Pontederia should take high
priority in future studies.

308
Review of Results and Suggestions for Further Studies
The productivity study discussed in the second section of this disser
tation showed that the net solar efficiency of waterhyacinth was similar
in both small and large plants (1.6%). Because of a high P:R ratio the
small plants grow faster (in terms of weight gain relative to standing
crop) than the large plants.
Three phases are apparent in the annual growth of waterhyacinth on
Lake Alice. A spring growth period is characterized by an initial increase
in plant and leaf density followed by an increase in plant height accomp
anied by a decline in plant density. This may be explained in part by
energy allocation under differing conditions of density. Early in the
season when the canopy is open and the plants are small, more energy is
allocated towards producing offsets than towards increasing individual
plant size. As space becomes more limiting more energy is put into
increasing the size of the individual plant, making it more able to com
pete for available light, and less into offset production. In a dense
stand the small offsets would probably have a small chance of surviving
in the low light conditions under the canopy. It would be maladaptive,
then, to produce them in this situation. As the plants increase further
in height the small plants die which accounts for the sharp drop in
absolute density.
A late summer and fall phase is defined by plant senescence and a
gradual decline in plant size. This is accompanied by an equally gradual
increase in plant density. An increase in damage by Avzama densa Wlk.
also occurred at this time but, because of multiple effects, the degree
to which A. densa contributed to this decline is not ascertainable.

Multivariate analyses failed to implicate A. densa as a factor in
accounting for seasonal variability in the plant characteristics. Climate
was considered the most important factor regulating variables estimating
standing crop. Water quality seemed to be more important in the variables
associated with density. Because changes in water quality (nutrient loads)
are as likely to be a result of changes in the plants as well as a cause
of those changes I am not satisfied that these models (plant and leaf
density) reflect dependent relationships even though statistically "good"
fits were obtained.
Intraspecific competition for light and space seems to be strongly
implicated in changes in plant density. A significant negative correlation
exists between plant density and plant height. Also, as the plant height
distribution becomes more strongly skewed towards large plants the number
of height classes important in the total distribution drops sharply.
Further, there appears to be an almost total loss of small plants during
the summer when plant height is maximum.
A third phase is the winter "no growth" or dormancy phase. During
the 2-3 mo. period little change occurred in most of the characteristics
observed.
Greenhouse experiments in which the levels of infestation by A. densa
were controlled more effectively brought out the relationship between the
various plant characteristics and feeding by this insect. Repetitions of
the experiment in the summer and the fall produced quite different results.
In general, the plants were much more sensitive to attack by insects in
the fall, and less so in the summer. Height declined in both experiments
but the slope of the decline was greater in the fall. The changes in the

310
number of leaves per plant were remarkably similar in both experiments.
Leaf density declined in the fall but remained constant in the summer.
Plant density increased with increasing insect concentration in the
summer but decreased in the fall. Standing crop was only estimated in
the summer but it appeared to be affected only at the highest level of
infestation at that time whereas in the fall in inverse linear relationship
at all levels was observed.
Net community productivity did not show a dramatic decline in the
fall experiment but turnover rates accelerated in direct proportion to
insect concentrations. As a result, the standing crop declined even
though productivity was little affected. This was not evaluated in the
summer.
Also in the fall experiment a slight tendency for the proportion
of the plant represented as leaves to increase under insect attack was
observed as well as a tendency for the proportion represented as rhi
zome to decrease. This resulted in an overall tendency for the root +
rhizome:shoot ratio to decrease.
Some interesting conclusions and inferences can be drawn from the
field and laboratory studies combined. These are enumerated as follows:
1. Insects are apt to cause a decrease in plant size.
2. Plant density is apt to decrease with increasing plant size as
a result of intraspecific competition for light and space.
3. Insects may, therefore, indirectly cause an increase in plant
density by reducing intraspecific competition.
While both small and large plants are equally efficient, small
plants are apt to grow relatively faster by virtue of a larger
P:R ratio.
4.

311
5. Insects, therefore, by reducing the canopy and stimulating
offset production may indirectly stimulate production.
6. This increased production may partly compensate for crop
reduction by herbivory.
7. Insects reduce standing crop by accelerating turnover.
8. Large amounts of nutrients are tied up as organic matter in
the plants.
9. Insect feeding may, therefore, result in a faster return of
these nutrients to the water which may also stimulate production.
10. Higher levels of insect infestation are likely to be needed in
the summer (probably the spring also) when solar radiation is
high or waxing than in the fall and winter (when solar energy is
low or waning) to achieve the same level of control.
11. A reduction in the size of the rhizome in proportion to the plant
by insect feeding is likely to hinder the ability of the plant to
survive the winter since spring regeneration occurs from the rhizome.
During the period of these studies, natural populations of A. densa
were consistently low. Heavy infestation by a complex of parasites appeared
to be the factor regulating population build-ups. This was very difficult
to analyze, however, because of the extreme degree of overlap in generations,
differential susceptibility of various instars to sampling, and low popu
lation levels from which to derive data.
A small scale field release of A. densa proved very effective in
controlling a small stand of waterhyacinth. Parasites failed to reduce the
larval population and severe damage to the plants resulted. The degree of
mortality as a result of these parasites was notably less than that of a

312
nearby natural population. This suggests that in some situations local
populations can escape parasitism. The mechanism for this is not clear
but it may have been due to synchronization of the age distribution in
the released populations.
About one half way through this research it became apparent that
A. densa was almost always more abundant on pickerelweed (Pontederia
cordata L.) than on waterhyacinth. These observations were casual,
however, and no data has been obtained from these populations. Some im
portant questions that should be answered could be derived from con
current studies of populations on both of these host plants. It would be
interesting to determine if the two populations are in phase or out of
phase. Is the population on waterhyacinth the result of dispersing in
dividuals from pickerelweed after the population from the latter host
builds up to a high level? Does waterhyacinth represent a secondary host
allowing the population to continue while pickerelweed is scarce? Do
A. densa populations control P. cordata to any extent? Is there any dif
ference in rates of parasitism of A. densa on its two host plants? To
what extend do populations from both hosts interact? Are these two popu
lations temporally separated? These and many other questions would greatly
enhance our understanding of the observations on the A. densa populations
reported here.
Life table studies of Arzama densa would be helpful in evaluating
methods of population augmentation. Such studies of natural populations
would be difficult if not impossible, however, because of the problems
associated with obtaining adequate samples discussed above. A possible
approach to this might be mass releases with subsequent life table studies
of the released population.

A prerequisite to any further research on A. densa, however, should
be a detailed bio-systematic study of the species group to which this
insect belongs. Neither host plant data nor life history data can be
accepted until the systematic relationships are established. Various
countries have expressed an interest in A. densa for control of water-
hyacinth. As long as economic crops are implicated within the host range
of this species its use in foreign countries should be discouraged. Host
specificity studies with larvae from differing natural host plants should
be carried out to either verify or disprove the existing host records.

314
LITERATURE CITED
Abu-Gideiri, Y.B., and A. M. Yousif. 1974. The influence of Eichhomia
crassipes Solm.on plantonic development in the White Nile.
Arch. Hydrobiol. 74(4): 463-467
Agarwal, S. C. 1974. Water hyacinth in West Bengal and its control.
Plant Prot. Bull. (Delhi) 23(3): 21-24.
Agharkar, S. P., and I. Banerji. 1930. Studies in the pollination and
seed formation of water-hyacinth (Eichhomia speciosa Kunth)
Agr. J. India 25(4): 286-296.
Agostini, G. 1974. El genero Eichhomia (Pontederiaceae) en Venezuela.
Acta Bot. Venezuelica 9(1-4): 303-310.
American Public Health Association. 1965. Standard methods for the
examination of water and wastewater, including bottom sediments and
sludges. Amer. Public Health Assoc., Inc. N.Y. 769 pp.
Andres, L. A., and F. D. Bennett. 1975. Biological control of aquatic
weeds. Annu. Rev. Entomol. 20: 31-46.
Andrewartha, H. G. 1961. Introduction to the study of animal populations
University of Chicago Press, Chicago. 281 pp.
Anonymous. 1896. Clogged by hyacinths Navigation on the St. Johns,
Florida, seriously obstructed. The Sun [newspaper], New York,
Sept. 20, 1896.
Anonymous. 1951. The water hyacinth problem and pig farming. Sci. Cult.
17(6): 231-232.
Anonymous. 1957a. Commission for Techincal Co-operation in Africa
South of the Sahara. Report of the symposium of Eichhomia crassipes
Leopoldville, 1957. C.C.T.A./C.S.A. pubis. 27: 1-31.
Anonymous. 1957b. The water hyacinth now a "prohibited imigrant."
E. Afr. Agr. J. 22(3): 129.
Anonymous. 1964a. Occurence of water hyacinth. F.A.0. Plant Prot. Bull
12(4): 93.
Anonymous. 1964b. Water hyacinth is one oi the world's worst weeds.
New Zeal. FI. Agr. 108: 232.
Arber, A. 1918. The phyllode theory of the monocotyledonous leaf, with
special reference to anatomical evidence. Ann. Bot. 32(128): 465-501
Arber, A. 1920. Water plants, a study of aquatic angiosperms. University
Press, Cambridge. 436 pp.
Arnold, C. A. 1940. A note on the origin of the lateral rootlets of
Eichhomia crassipes (Mart.) Solrns. Amer. J. Bot. 27: 728-730.

315
Azam, M. A. 1941. Utilisation of water hyacinth in the manufacture of
paper and pressed boards. Sci. Cult. 6(11): 656-661.
Backer, C. A. 1951. Pontedariaceae. Pages 255-261 in C. 6. G. J. Van
Steenia, ed. Flora Malesiana, vol. 4.
Bagnall, L. 0., J.A. Baldwin, and J. F. Hentges, Jr. 1974. Processing
and storage of waterhyacinth silage. Hyacinth Control J. 12: 73-79.
Bagnall, L. 0., T. de S. Furman, J. F. Hentges, Jr., W. J. Nolan, and
R. L. Shirley. 1974. Feed and fiber from effluent-grown water hyacinth.
Waste Water Use in the Production of Food and Fiber Proc. EPA-660/2-74-041
Bagnall, L. 0., J. F. Hentges, Jr., and R. L. Shirley. 1973. Processing,
chemical composition and nutritive value of aquatic weeds as affecting
the economics of their removal and utilization. Inst. Food Agr.
Sci., University of Florida, Gainesville.
Baldwin, J. A. 1973. Utilization of ensiled water hyacinths in ruminant
diets. M.S.A. Thesis. University of Florida, Gainesville.
Baldwin, J. A., J. F. Hentges, Jr., L. 0. Bagnall, and R. L. Shirley.
1975. Comparison of pangolagrass and water hyacinth silages as
diets for sheep. J. Anim. Sci. 40(5): 968-971.
Bancroft, K. 1913. The water-hyacinth. Agr. Bull. F. M. S. 1: 228.
Banerjee, M. 1974. Cytological studies on some members of the family
Ponterderiaceae. Cytol. 39: 483-491.
Banerji, I., and H. C. Gangulee. 1937. Spermatogenesis in Eiahhomia
orassipes Solms. Indian Bot. Soc. J. 16: 289-295.
Barber, M. A., and T. B. Hayne. 1925. Water hyacinth and the breeding
of Anopheles. Public Health Rep. 40(47): 2557-2562.
Barman, E. H., Jr. 1974. Gross growth, net growth, and assimilation
efficiencies of Diacrisia virginioa (Artiidae: Lepidoptera) fed
leaves of water-hyacinth. Assoc. SE. Biol. Bull. 21(2): 37.
Barnes, W., and F. H. Benjamin. 1924. Notes and new species. Contrib.
Natur. Hist. N. Amer. Lepidoptera. 5(3): 169.
Barnes, W., and J. H. McDunnough. 1914. Synonymic notes on North American
Lepidoptera. Contrib. Natur. Hist. N. Amer. Lepidoptera. 2(5): 200-201.
Barnes, W., and J. H. McDunnough. 1916. Synonymic notes on North American
Heterocera. Contrib. Natur. Hist. N. Amer. Lepidoptera. 3(3): 166-167.
Barnes, W., and J. H. McDunnough. 1917. Remarks on Grossbeck's list of
Florida Lepidoptera. Contrib. Natur. Hist. N. Amer. Lepidoptera.
3(4): 217.
Barton, L. V., and J. E. Hotchkiss. 1951. Germination of seeds of Eiahhornia
orassipes Solms. Contrib. Boyce Thompson Inst. 16: 215-220.

316
Bennett, F. D. 1967. Notes on the possibility of biological control of
the water hyacinth Eiehhornia crassipes. PANS 13(4C): 304-309.
Bennett, F. D. 1968a. Investigations on insects attacking water hyacinth
in Florida, British Honduras, and Jamaica. C.I.B.C. Rep. (unpublished).
9 pp.
Bennett, F. D. 1968b. Insects and mites as potential controlling agents
of water hyacinth (Eiehhornia crassipes (Mart.) Solms). Proc. 9th
Brit. Weed Control Conf.: 832-835.
Bennett, F. D. 1970. Insects attacking water hyacinth in the West Indies,
British Honduras and the U.S.A. Hyacinth Control J. 8(2): 10-14.
Bennett, F. D. 1972. Survey and assessment of the natural enemies of
water hyacinth, Eiehhornia crassipes. PANS 18(3): 310-311.
Bennett, F. D. 1974. Biological control. Pages 99-106 in 0. S. Mitchell,
ed. Aquatic vegetation and its use and control. UNESCO, Paris.
Bennett, F. D., and H. Zwolfer. 1968. Exploration for natural enemies of
the water hyacinth in northern South America and Trinidad. Hyacinth
Control J. 7: 44-52.
Berg, A. 1959. Analyse des conditions impropres au developpement de la
jacinthe d'eau Eiehhornia crassipes (Mart.) Solms dans certaines
rivieres de la cuvette congolaise. Bull. Agr. Congo Beige.
50(2): 365-393.
Beutenmller, W. 1889. On early stages of some Lepidoptera. Can. Entomol.
21: 160.
Beutenmller, W. 1902. Descriptive catalogue of the Noctuidae found within
fifty miles of New York City. Part II. Bull. Amer. Mus. Natur. Hist.
16: 413-458.
Bill, S. M. 1969. The water weed problem in Australia. Hyacinth Control
J. 8(1): 1-6.
Billings, W. D., and P. J. Godfrey. 1967. Photosynthetic utilization of
internal carbon dioxide by hoi low-stemmed plants. Sci. 158: 121-123.
Blackburn, R. D. 1974. Chemical control. Pages 85-98 in D. S. Mitchell,
ed. Aquatic vegetation and its use and control. UNESCO, Paris.
Bock, J. H. 1966. An ecological study of Eiehhornia crassipes with
special emphasis on its reproductive biology. PhD Thesis in Botany,
University of California. 175 pp.
Bock, J. H. 1968. The water hyacinth in California. Madrono 19(7): 281-283.
Bock, J. H. 1969. Productivity of the water hyacinth Eiehhornia crassipes
(Mart.) Solms. Ecology 50(3): 460-464.
Bose, P. K. 1945. The problem of water-hyacinth in Bengal. Sci. Cult.
11(4): 167-171.

317
Boyd, C. E. 1968a. Evaluation of some common aquatic weeds as possible
feedstuffs. Hyacinth Control J. 7: 26-27.
Boyd, C. E. 1968b. Fresh-water plants: a potential source of protein.
Econ. Bot. 22(4): 359-368.
Boyd, C. E. 1970a. Amino acid, protein and caloric content of vascular
aquatic macrophytes. Ecology 51(5): 902-906.
Boyd, C. E. 1970b. Vascular aquatic plants for mineral nutrient removal
from polluted waters. Econ. Bot. 24(1): 95-103.
Boyd, C. E. 1972. A bibliography of interest in the utilization of
vascular aquatic plants. Econ. Bot. 26: 74-84
Boyd, C. E. 1974. Utilization of aquatic plants. Pages 107-115 in
D. S. Mitchell, ed. Aquatic vegetation and its use and control.
UNESCO, Paris.
Boyd, C. E., and E. Scarsbrook. 1975. Influence of nutrient additions and
initial density of plants on production of waterhyacinth, Eiohhormia
arassipes. Aquatic Bot. 1(3): 253-261.
Brehme, H. H. 1888a. Early stages of Arzama obliquata 6. and R. Can.
Entomol. 20: 119.
Brehme, H. H. 1888b. Arzama obliquata G. and R. in Correspondence.
Can. Entomol. 20: 180.
Brehme, H. H. 1889. Arzama obliquata. in Correspondence. Can. Entomol.
21: 78.
Brezny, 0., I. Mehta, and R. K. Sharma. 1973. Studies on evapotrans-
piration of some aquatic weeds. Weed Sci. 21(3): 197-204.
Brezonik, P. L. 1969. Eutrophication in Florida lakes. Proc. Fla.
Envir. Eng. Conf. Water Poll. Control. Gainesville, Fla. Eng. & Ind.
Exp. Sta. 24(3), Bull. Ser. 135: 124-129.
Brezonik, P. L., W. H. Morgan, E. E. Shannon, and H. D. Putnam. 1969.
Eutrophication factors in North Central Florida lakes. Eng. Prog.
Univ. Fla. 23(8), Bull. Ser. 134, Water Resources Center
Pub. 5, 101 pp.
Britton, N. L. 1918. Flora of Bermuda. Charles Scribner and Sons, New York.
Brown, S., T. D. Center, K. R. Dugger, and W. J. Mitsch. 1974. Management
model and associated field measurements of Lake Alice. Unpub. rep.
Buckman and Company. 1930. Report to Jacksonville City Commission on
projected water hyacinth control projects. 41 pp.

318
Burton G. J. I960. Studies on the bionomics of mosquito vectors which
transmit filariasis in India. II. The role of water hyacinth (Eichhomia
speciosa Kunth) as an important host plant in the life cycle of Mansonia
uniformis (Theobald) with notes on the differentiation of the late
embryonic and newly hatched stages of Mansonia uniformis (Theobald)
and Mansonia annulifera (Theobald). Indian J. Malario!. 14(2): 81-106.
Caldwell, S. Y. 1942. Orchid of the lily pool. Flower Grower 29: 269.
Carter, M. R., L. A. Burns, T. R. Cavinder, K. R. Dugger, P. L. Fore,
D. B. Hicks, H. L. Revells, and T. W. Schmidt. 1973. Ecosystems
analysis of the big cypress swamp and estuaries. U.S. E.P.A. Region
IV, Atlanta.
Cason, J. H. 1970. Lake Alice -- A study of potential pollution of the
Floridan Aquifer. Compass 47(4): 206-210.
Castellanos, A. 1958. Ntula sobre el gneroPontederia en Brasil.
Arquivos do Jardim Botnico do Rio de Janeiro 15: 59-67.
Castellanos, A. 1959. Las Pontederiaceae de Brasil. Arquivos do
Jardim Botnico do Rio de Janeiro 16: 149-216.
Cavers, P. B., and G. A. Mulligan. 1972. A new series -- The biology of
Canadian weeds. Can. J. Plant Sci. 52: 651-654.
Center, T. D. and J. Balciunas. 1975. The effects of water quality on
the distribution of alligatorweed and waterhyacinth. Aquatic Plant
Control Prog., Tech. Rep. 10: B3-B13, U.S. Army Eng., WES, Vicksburg,
Mississippi.
Chadwick, M. J. and M. Obeid. 1966. A comparative study of the growth of
Eiehhomia orassipes and Pistia stratiotes in water culture. J. Ecol.
54: 563-575.
Charudattan, R. 1975. Use of plant pathogens for control of aquatic weeds.
Ecol. Res. Ser., U.S. E.P.A. 660-3-75-001: 127-153.
Chatterjee, I., and M. A. Hye. 1938. Can water hyacinth be used as a
cattle feed? Agr. and Live-stock in India 8(5): 547-553.
Chhibbar, S. S., and G. D. Singh. 1971. Paddy straw and water hyacinth
silage. Indian Farming 20: 24-26, 32.
Claassen, P. W. 1921. Typha insects: their ecological relationships.
Cornell University Agr. Exp. Sta. Mem. 47: 458-531.
Coker, W. C. 1907. The development of the seed in the Pontederiaceae.
Bot. Gaz. 44: 293-301.
Combs, G. E. 1970. Aquatic plants for swine feeding. Proc. Aquatic
Plant Res. Conf., University of Florida, Gainesville.
Comstock, J. A. 1944. Four California moths associated with cat-tails.
Bull. S. Calif. Acad. Sci. 43(2): 81-83.

3l9
Comstock, J. A., and C. M. Dammers. 1944. The larva and pupa of Arzama
gargantua Dyar. Bull. S. Calif. Acad. Sci. 43(2): 84-85.
Comstock, J. H. 1881. An aquatic noctuid larva. Arzama melanopyga Grote,
new species. Papilio 1(9): 147-149.
Comstock, J. H. 1888. Aquatic lepidopterous larvae. Amer. Natur. 22: 468-469.
Comstock, J. H. 1936. An introduction to entomology. Comstock Publ. Co.,
Inc., Ithaca, New York, 1044 pp.
Cook, C. D. K., B. J. Gut, E. M. Rix, J. Schneller, and M. Seitz. 1974.
Water plants of the world. A manual for the identification of the
genera of freshwater macrophytes. Dr. W. Junk b.v., Pub., The
Hague. 561 pp.
Cordo, H. A., and C. J. DeLoach. 1975. Ovipositional specificity and
feeding habits of the waterhyacinth mite, Orthogalwma terebrantis,
in Argentina. Envir. Entomol. 4(4): 561-565.
Coste, P. 1958. Un Fleau aquatique des pays tropicaux: la jacinthe d'eau.
Phytoma. 99: 24-28.
Couch, R. 1971. Morphology and anatomy of water hyacinths. Proc. 24th
S. Weed Sci. Soc.: 332.
Coulson, J. R. 1971. Prognosis for control of water hyacinth by arthro
pods. Hyacinth Control J. 9(1): 31-34.
Crocker, W. 1907. Germination of seeds of water plants. Bot. Gaz. 44:
375-380.
Crumb, S. E. 1956. The larvae of the Phalaenidae. U.S.D.A. Tech. Bull.
1135, 356 pp.
Curtiss, A. H. 1900. The water hyacinth in Florida. Plant World 3: 38-40.
Davies, H. R. J. 1959. Effect of the water hyacinth (Eichhomia arassipes)
in the Nile Valley. Nature 184(4692): 1085-1086.
Del Fosse, E. S., H. L. Cromroy, and D. H. Habeck. 1975. Determination
of the feeding mechanism of the waterhyacinth mite. Hyacinth Control
J. 13: 53-55.
DeLoach, C. J. 1975. Identification and biological notes on the species
of Neoahetina that attack Pontederiaceae in Argentina. Coleopt. Bull.
29(4): 257-265.
DeLoach, C. J., and H. A. Cordo. 1976. Life cycle and biology of Neoahetina
bruohi, a weevil attacking waterhyacinth in Argentina, with notes on
N. eiohhomiae. Ann. Entomol. Soc. Amer. (in press).
Dunigan, E. P., and R. A. Phelan. 1975. Nitrogen and phosphorus uptake by
water hyacinths in two farm ponds. Rep. Proj. Dep. Agron., LSU A&M
Coll., Agr. Exp. Sta.: 272-275.

320
4
Dunigan, E. P., Z. H. Shamsuddin, and R. A. Phelan. 1975. Water hyacinths
tested for cleaning polluted water. La. Agr. 18(2): 12-13.
Du Toit, R. 1938. Water hyacinth. Farming in S. Afr. 13(132): 16-17.
Dutton, J. 6. 1964. Louisiana's hyacinth control program. La. Conserv.
16(1-2): 6-8.
Dyar, H. 6. 1913. The species of Sphida Grote. Insec. Inscit. Menst.
1: 18-19.
Dyar, H. G. 1922. A note on Belluva gortynoides Wlk. Insec. Inscit.
Menst. 10: 50.
Dymond, G. C. 1948. The water-hyacinth: a Cinderella of the plant world.
Pages 221-227 in J. P. J. Van Vurens, ed. Soil fertility and sewage.
Faber and Faber, Ltd. London.
Eggler, W. A. 1953. The use of 2,4-D in the control of water hyacinth
and alligatorweed in the Mississippi Delta, with certain ecological
implications. Ecology 34(2): 409-414.
Evans, A. C. 1963. The grip of the water hyacinth. New Sci. 19(358):
666-668.
Forbes, W. T. M. 1954. The Lepidoptera of New York and neighboring states
Part 3. Noctuidae. Cornell University Agr. Exp. Sta. Mem. 329, 433 pp
Francois, J. 1964-3. Observations sur 1 'htrostylie chez Eiohhomia
1ovassipes (Mart.) Solms. Bull, des Seances 1964-3: 501-519.
Freeman, T. E., F. W. Zettler, and R. Charudattan. 1974. Phytopathogens
as biocontrols for aquatic weeds. PANS 20(2): 181-184.
Frick, K. E. 1974. Augmenting the weed control effectiveness of
phytopagous insects. Biol. Control Weeds Symp., Entomol. Soc. Amer.
Nat. Meeting, Minneapolis, Minnesota, Dec. 2-5, 1974, 19 pp.
Frost, S. W. 1975. Third supplement to insects taken in light traps at
the Archbold Biological Station, Highland County, Florida. Fla.
Entomol. 58(1): 35-42.
Gangstad, E. 0., D. E. Seaman, and M. L. Nelson. 1972. Potential growth
of aquatic plants of the lower Mekong River Basin Laos-Thialand.
Hyacinth Control J. 10: 4-9.
Ganguly, S., and S. M. Sircar. 1964. Cell growth and metabolism of pea
(Pisum sativum L.) internodes as affected by the growth substances
from the root of water hyacinth (Eichhomia oras sipes). Bull. Bot.
Soc. Bengal 18(1): 83-86.
Gay, P. A. 1958. Eichhomia orassipes in the Nile of the Sudan. Nature
182(4634): 538-539.

321
Gay, P. A. 1960. Ecological studies of Eichhomia arassipes Solms in
the Sudan. I. Analysis of spread in the Nile. J. Ecol. 48: 183-191.
Goin, C. J. 1943. The lower vertebrate fauna of the water hyacinth
community in northern Florida. Proc. Fla. Acad. Sci. 6(3-4): 143-153.
Gordon, R. D., and J. R. Coulson. 1969. Report on field observations
of arthropods on water hyacinth in Florida, Louisiana, and Texas,
July 1969. Aquatic Plant Control Prog., Tech. Rep. 6: B3-B37,
U.S. Army Eng., WES, Vicksburg, Mississippi.
Gowanloch, J. N. 1944. The economic status of the water hyacinth in
Louisiana. La. Conserv. 2(9): 3,6,8.
Gowanloch, J. N., and A. D. Bajkov. 1948. Water hyacinth program.
La. Dept. Wildlife and Fish., Biennial Rep. (1946/1947) 2: 66-124.
Grossbeck, J. A. 1917. Insects of Florida. IV. Lepidoptera. (F. E.
Watson, ed.). Bull. Amer. Mus. Natur. Hist. 37: 1-147.
Grote, A. R. 1873. Descriptions of North American Noctuidae No. 3.
Trans. Amer. Entorno!. Soc. 4: 293-310.
Grote, A. R. 1874. List of the Noctuidae of North America. Bull.
Buffalo Soc. Natur. Sci. 2: 1-77.
Grote, A. R. 1878 [1879]. Descriptions of Noctuidae chiefly from
California. Bull. U.S. Geol. and Geogr. Surv. 4(1): 169-188,
Washington, 1878. Cited from: New Publications, Bull. Brooklyn
Entomol. Soc. 2: 36, 1879.
Grote, A. R. 1882. An illustrated essay on the Noctuidae of North
America with a colony of butterflies. E. W. Classey, Ltd. Middlesex,
England.
Grote, A. R. 1883. Introduction to a study of the North American
Noctuidae. Proc. Amer. Phil. Soc. 21: 157.
Grote, A. R. 1889. The Noctuidae of North America and Europe compared.
Can. Entomol. 21: 226-230.
Grote, A. R. 1890. North American Lepidoptera. Revised check list of
North American Noctuidae. Part I. Thyatirinae Noctuinae.
Bremen, Germany. 52 pp.
Grote, A. R., and C. T. Robinson. 1867-8. Descriptions of American
Lepidoptera No. 3. Trans. Amer. Entomol. Soc. 1: 323-360.
Gundlach, J. 1881. Contribucin la Entomologia Cubana. Havana, Cuba.
Guppy, R. 1948. Some notes on the habits of Avzoana obliqua on Vancouver
Island (Lepidoptera: Phalaenidae). Proc. Entomol. Soc. B.C. 44: 17-18.

322
Guscio, F. J., T. R. Bartley, and A. N. Beck. 1965. Water resources
problems generated by obnoxious plants. J. Waterways Harbor
Div. Amer. Soc. Civ. Eng. 10: 47-60.
Habeck, D. H. 1974. Arzama densa as a pest of dasheen. Fla. Entomol.
57(4): 409-410.
Haigh, J. C. 1936. Notes on the water hyacinth (Eiohhomia crassipes
Solms) in Ceylon. Ceylon J. Sci. Sec. 12(2A): 97-108.
Haigh, J. C. 1940. The propagation of water hyacinth (Eiohhomia crassipes
Solms) by seed. Trop. Agr., Ceylon 94(5): 296-297.
Haller, W. T., E. B. Knipling, and S. H. West. 1970. Phosphorous absorption
by and distribution in water hyacinths. Proc. Fla. Soil and Crop Sci.
Soc. 30: 64-68.
Haller, W. T., and D. L. Sutton. 1973. Effect of pH and high phosphorus
concentrations on growth of waterhyacinth. Hyacinth Control J.
11: 59-61.
Hampson, G. F. 1910. Catalogue of the Lepidoptera Phalaenae in the
British Museum. Vol. 9. Noctuidae, Acronyctinae, Part 3. Brit. Mus.
Natur. Hist., London, 552 pp.
Hansen, K. L., E. G. Ruby, and R. L. Thompson. 1971. Trophic relation
ships in the water hyacinth community. Quart. J. Fla Acad. Sci.
34(2): 107-113.
Harper, R. M. 1903. The water-hyacinth in Georgia. Plant World 6:
164-165.
Harris, P. 1972. Insects in the population dynamics of plants. Pages
201-209 in H. F. van Emden, ed. Insect/plant relationships.
Symp. Roy. Entomol. Soc. London, No. 6. Blackwell Scientific Publ.
Oxford.
Harris, P. 1973. Weed vulnerability to damage by biological control
agents. C.I.B.C. Mise. Pub. 6: 29-39.
Hearne, J. S. 1966. The Panama Canal's aquatic weed problem. Hyacinth
Control J. 5: 1-5.
Hentges, J. F., Jr. 1970. Processed aquatic weeds for animal nutrition.
Proc. Aquatic Plant Res. Conf., University of Florida, Gainesville:
62-67.
Herrich-Schaffer, G. A. W. 1868. Korrespondenzblatt, fr Sammler von
Insecten, insbesondere von Schmetterlingen. Regensburg, Manz: 156.
Hildebrand, E. M. 1946. Herbicidal action of 2,4-D acid on the water
hyacinths, Eiohhomia crassipes. Sci. 103: 477-479.

323
Hitchcock, A. E., P. W. Zimmerman, H. Kirkpatrick, Jr., and T. T. Earle.
1949. Water hyacinth: its growth, reproduction and practical control
by 2,4-D. Contrib. Boyce Thompson Inst. 15(7): 363-401.
Hitchcock, A. E., P. W. Zimmerman, H. Kirkpatrick, Jr., and T. T. Earle.
1950. Growth and reproduction of water hyacinth and alligator weed
and their control by means of 2,4-D. Contrib. Boyce Thompson Inst.
16(3): 91-130.
Holland, W. J. 1903. The moth book. Doubleday, Page, and Co., New York.
Holm, L. G. 1969. Weed problems in developing countries. Weed Sci.
17(1): 113-118.
Holm, L. G., L. W. Weldon, and R. D. Blackburn. 1969. Aquatic weeds.
Sci. 166: 699-709.
Hooker, W. J. 1829. Pontederia azurea. Large-flowered Pontederia.
Curtis' Bot. Mag., Entry 2932.
Ingersoll, J. M. 1974. The water hyacinth. Aquatic Plant Control Prog.
Tech. Rep. 6: A3-A33, U.S. Army Eng., WES, Vicksburg, Mississippi.
Ishaque, M. 1952. Water hyacinth: a problem of East Pakistan and its
control. Proc. Pakistan Sci. Conf. 4(3): 10-11.
Iswaran, V., and A. Sen. 1973. Influence of extract of water hyacinth
(Eiahhomia arassipes) on the yield of brinjal (Solanwn melongena)
var. Pusa Kranti. Sci. Cult. 39(9): 394.
Jameson^ D. A. 1963. Responses of individual plants to harvesting.
Bot. Rev. 29(4): 532-594.
Janzen, D. H. 1969. Seed eaters versus seed size, number toxicity and
dispersal. Evolution 23(1): 1-27.
Jepson, F. P. 1933. The water hyacinth problem in Ceylon. Trop. Agr.
Mag. Ceylon Agr. Soc. 81: 339-355.
Johnson, E. 1920. Fresno County will fight water hyacinth. Calif.
Dept. Agr. Monthly Bull. 9(5-6): 202-203.
Johnston, J. 1889. Arzama obliquata. in Correspondence Can. Entomol.
21: 79.
Jones, J. R. J. L. 1951. An annotated checklist of the Macrolepidoptera
of British Columbia. Entomol. Soc. B.C. Occ. Paper 1: 85.
Kellicott, D. S. 1883a. in Correspondence Can. Entomol. 15: 171.
Kellicott, D. S. 1883b. in Correspondence Can. Entomol. 15: 174-175.
Kellicott, D. S. 1889. Arzama obliquata. Can. Entomol. 21: 39.

324
Kelsey, H.P., and W. A. Dayton. 1942. Standardized plant names.
J. Horace McFarland Co., Harrisburg, Pennsylvania. 675 pp.
Kimball, C. P. 1965. The Lepidoptera of Florida. Arthropods of Florida
and neighboring land areas. Vol. 1. Fla. Dept. Agr., Div. Plant
Ind., Gainesville, Florida 363 pp.
Kirkpatrick, H. 1958. Eradication and control of water hyacinth.
Results of a compaign in the Belgian Congo in 1955-57. World
Crops 10(8): 286-288.
Klorer, J. 1909. The water hyacinth problem. J. Assoc. Eng. Soc.
42: 33-48.
Knipling, E. B., S. H. West, and W. T. Haller. 1970. Growth character
istics, yield potential, and nutritive content of water hyacinths.
Proc. Fla. Soil and Crop Soc. 30: 51-63.
Kunth, C. S. 1843. Enumeratio plantarum, Vol. 4. Stuttgart: J. G. Cottae
La Garde, R. V. 1930. A plant that stopped navigation. Missouri Bot.
Gard. Bull. 18: 48-51.
Lebrun, J. 1958. The present situation in respect of the problems of
Eiohhomia crassipes in Africa south of the Sahara. PANS
4(C): 87-99.
Levine, E. 1974. Biology of Belluva gortynoides Walker (=vulnifica
(Grote)), the yellow water lily borer (Lepidoptera: Noctuidae).
Proc. Indiana Acad. Sci. 83: 214-215.
Liang, J. K., and R. T. Lovell. 1971. Nutritional value of water hyacinth
in channel catfish feeds. Hyacinth Control J. 9(1): 40-44.
Liebig, J. 1840. Chemistry and its application to agriculture and
physiology. Taylor and Walton, London. 352 pp.
Little, E. C. S. 1965. The world wide distribution of the water hyacinth.
Hyacinth Control J. 4: 30-32.
Little, E. C. S.1966. The invasion of man-made lakes by plants.
Pages 75-86 in R. H. Lowe-McConnel1, ed. Man-Made lakes. Proc. 15th
Symp. Inst. Biol., London, Academic Press, London.
Little, E. C. S. 1968a. The control of water weeds. Weed Res. 8(2): 79-105
Little E. C.S.(ed.). 1968b. Handbook of utilization of aquatic plants.
F.A.O., Rome, Italy. 121 pp.
Lloyd, M., and H. S. Dybas. 1966. The periodical cicada problem. II.
Evolution. Evolution 20(4): 466-505.

325
Long, K. S., and P. A. Smith. 1975. Effects of C02 laser on water hyacinth
growth. Aquatic Plant Control Prog., Tech. Rep. 11, 124 pp. U.S.
Army Eng., WES, Vicksburg, Mississippi.
Lowden, R. M. 1973. Revision of the genus Pontederia L. Rhodora 75: 426-487.
McDonald, J. L. 1970. Prelimanary results on experimental detection of
Mansonia uniformis (Theob.) mosquito immatures. Mosq. News 30(4):
614-619.
McGaha, Y. J. 1952. The limnological relations of insects to certain
aquatic flowering plants. Trans. Amer. Microscop. Soc. 71: 355-381.
McLean, K. 1922. Water hyacinth. A serious pest in Bengal. Agr. J. India
17(1): 23-40.
McVea, C., and C. E. Boyd. 1975. Effects of waterhyacinth cover on water
chemistry, phytoplankton, and fish in ponds. J. Environ. Qual. 4(3):
375-378.
Maiden, J. H., F. B. Guthrie, and T. E. Burrows. 1906. Water hyacinth
in New South Wales. Agr. Gaz. N.S.W. 1906: 1212-1221.
Manson, J. G., and B. E. Manson. 1958. Water hyacinth reproduces by seed
in New Zealand. New Zeal. J. Agr. 96: 191.
Matthews, L. J. 1967. Seedling establishment of water hyacinth. PANS
13(1C): 7-8.
Mendonca, A. 1958. Present state of the problem of Eichhomia orassipes
in Mozambique and Angola. PANS 4(C): 236-239.
Minshall, W. H., and G. W. Scarth. 1952. Effect of growth in acid media
on the permeability of water hyacinth and frogbit root cells.
Can. J. Bot. 32: 188-208.
Misra, R. 1969. Ecological studies of noxious weeds, common to India and
America, which are becoming an increasing problem in the Upper Gangetic
Plains. Final PL-480 Tech. Rep., Part II: 409-441, Dept. Bot. Baaras
Hindu University, Varanasi-5 (India).
Mitchell, D. S. 1974. Aquatic vegetation and its use and control. UNESCO,
Paris. 135 pp.
Mitsch, W. J. 1975. Systems analysis of nutrient disposal in cypress
wetlands and lake ecosystems in Florida. PhD Thesis, University of
Florida. 421 pp.
Moffat, J. A. 1888a. Arzama obliquata. in Correspondence Can. Entomol.
20: 139.
Moffat, J. A. 1888b. Arzama. obliquata. in Correspondence Can. Entomol.
20: 238-239

326
Moffat, J. A. 1889. Arzama obliquata. In Correspondence Can. Entomol.
21: 99.
Mohamed, B. F., and F. F. Bebawi. 1975. The distribution of Eiohhomia
orassipes (Mart.) Solms in the White Nile, Sudan. MS. 13 pp.
Morton, J. F. 1962. Eiohhomia orassipes Solms. Hyacinth Control J.
1: 31. Reprinted from Wild plants for survival in South Florida.
Hurricane House, Miami, 1962. Page 18.
Mosher, E. 1919. Notes on lepidopterous borers found in plants with special
reference to the European corn borer. J. Econ. Entomol. 12: 258-268.
Muenscher, W. C. 1967. Aquatic plants of the United States. Comstock
Publ. Co., Inc., Ithaca, New York. 374 pp.
Mukherjee, R. K., A. Bhanja, P. R. Burman, and S. M. Sircar. 1964
Presence of bound auxin in the roots of water hyacinth (Eichhomia
orassipes). Bull. Bot. Soc. Bengal 18(1): 87-90.
Mulrennan, J. A. 1962. The relationship of mosquito breeding to aquatic
plant production Hyacinth Control J. 1: 6-7.
Mue, T. L., and J. W. Parham. 1954. Water hyacinth Bekabekairaga
(Eiohhomia orassipes, Solms). Fiji Agr. J. 25: 82-83.
Munz, P. A. 1965. A California flora. University of California Press,
Berkeley. 1681 pp.
National Science Research Council of Guyana and National Academy of Sciences,
U.S.A. 1973. Some prospects for aquatic weed management in Guyana.
Workshop on Aquatic Weed Management and Utilization, Georgetown,
Guyana. 39 pp.
Needham, J. G., S. W. Frost, and B. H. Tothill. 1928. Leaf-mining
insects. The Williams and Wilkins Co.. 351 pp.
Nolan, W. J., and D. W. Kirmse. 1974. The papermaking properties of
waterhyacinth. Hyacinth Control. 12: 90-97.
Odum, E. P. 1969. The strategy of ecosystem development. Science 164:
262-270.
Odum, H. T. 1953. Dissolved phosphorus in Florida waters. Fla. Geol.
Surv., Rep. of Invest. No. 9. 40 pp.
Odum, H. T. 1974. Energy basis for man and nature, or energy: crisis
to steady state. McGraw-Hill, Co., publ. pending.
O'Hara, J. 1967(1968). Invertebrates found in water hyacinth mats.
Quart. J. Fla. Acad. Sci. 30(1): 73-80.
Olive, E. W. 1894. Contribution to the histology of the Pontederiaceae.
Bot. Gaz. 19: 178-184.
Ornduff, R. 1966. The breeding system of Pontederia cordata L. Bull.
Torr. Bot. Club 93(6): 407-416.

327
Ornes, W. H., and D. L. Sutton. 1975. Removal of phosphorus from static
sewage effluent by waterhyacinth. Hyacinth Control J. 13: 56-58
Parija, P. 1930. A preliminary note on the physiology of the seedlings
of the water-hyacinth (Eiohhomia speoiosa). Agr. J. India 25(5):
386-391.
Parija, P. 1934. Physiological investigations on waterhyacinth (Eiohhomia
cvassipes) in Orissa with notes on some other aquatic weeds. Indian
J. Agr. Sci. 4: 399-429.
Parsons, W. T. 1963. Water hyacinth, a pest of world waterways. J. Agr.
Victoria 61: 23-27.
Penfound, W. T. 1956. Primary production of vascular aquatic plants.
Limnol. Oceanogr. 1: 92-101.
Penfound, W. T., and T. T. Earle. 1948. The biology of the water hyacinth.
Ecol. Monogr. 18(4): 448-472.
Penfound, W. T., and E. S. Hathaway. 1938. Plant communities in the
marshlands of southeastern Louisiana. Ecol. Monogr. 8(1): 1-56.
Perkins, B. D. 1972. Potential for waterhyacinth management with
biological agents. Proc. Annu. Tall Timbers Conf. on Ecol. Anim.
Control by Habitat Manage., Feb. 24-25, 1972: 53-64.
Perkins, B. D. 1973. Preliminary studies of a strain of the waterhyacinth
mite from Argentina. C.I.B.C. Mise. Publ. 6: 179-184.
Perkins, B. D. 1974. Arthropods that stress waterhyacinth. PANS 20(3):
304-314.
Pettet, A. 1964. Seedlings of Eiohhomia orassipes: a possible complication
to control measures in the Sudan. Nature 201(4918): 516-517.
Pieterse, A. H. 1974. The water hyacinth. Trop. Abstr. 29(2): X263-X483.
Pirie, N. W. 1960. Water hyacinth: a curse or a crop? Nature 185(4706): 116.
Rao, P. S. J. 1920a. Note on the geotropic curvature of the inflorescene
in Eiohhomia speoiosa Kunth (water hyacinth). Indian Bot. Soc. J.
1: 217-218.
Rao, P. S. J. 1920b. The formation of leaf-bladders in Eiohhomia
speoiosa Kunth (water hyacinth). Indian Bot. Soc. J. 1: 219-225.
Raynes, J. J. 1964. Aquatic plant control. Hyacinth Control J. 3: 2-4.
Riley, C. V. 1883a. in Meeting of the Entomological Club of the American
Association for the Advancement of Science. Can. Entomol. 15(9): 174, 176.

328
Riley, C. V. 1883b. in Meeting of the Entomological Club of the American
Association for the Advancement of Science. Entomol. Soc. Ont. Rep.
1883: 18, 21.
Riley, C. V. 1885. Notes: June 4, 1885. Proc. Entomol. Soc. Wash. 1: 30.
Robbins, W. W., T. E. Weier, and C. R. Stocking. 1964. Botany. An intro
duction of plant science. John Wiley & Sons, Inc., New York. 614pp.
Robertson, H. F., and B. A. Thein. 1932. The occurrence of water hyacinth
Eiohhomia arassipes Solms) seedlings under natural conditions in
Burma. Agr. and Live-stock in India 2(4): 383-390.
Robertson-Mi1ler, E. 1923. Observations on the Be llura. Ann. Entomol. Soc.
Amer. 16: 374-386.
Robson, T. 0. 1974. Mechanical control. Pages 72-84 in D. S. Mitchell,
ed. Aquatic vegetation and its use and control. UNESCO, Paris.
Rogers, H. H., Jr. 1971. Nutrient removal by waterhyacinth. M. S. Thesis,
Auburn University. 52 pp.
Rogers, H. H., Jr., and D. E. Davis. 1971. Nutrient absorption from sewage
effluent by aquatic weeds. Proc. 24th S. Weed Sci. Soc.: 352.
Rummel, C. 1919. Trapping for larvae of Apeen tesis. Bull. Brooklyn Entomol.
Soc. 14: 62-63.
Rushing, W. N. 1974. Waterhyacinth research in Puerto Rico. Hyacinth
Control J. 12: 48-52.
Ruttner, F. 1957. Fundamentals of limnology. University of Toronto
Press, Toronto. 242 pp.
Sabrosky, C. W. 1974. Eugaurax setigena, (Diptera: Chloropidae) a new
stem miner in water hyacinth. Fla. Entomol. 57(4): 347-348.
Sachs, J. 1860. Physiologische Untersuchungen ber die Abhngigkeit
der Keimung von der Temperatur. Jahrbiicher Wiss. Bot. 2: 338-377.
Sahai, R., and A. B. Sinha. 1970. Contribution to the ecology of Indian
aquatics. I. Seasonal changes in biomass of water hyacinth (Eiohhomia
arassipes (Mart.) Solms). Hydrobiologia 35(34): 376-382.
Salveson, R. E. 1971. Utilization of aquatic plants in steer diets:
voluntary intake and digestibility. M.S.A. Thesis, University of
Florida, Gainesville.
Sankaran, T., D. Srinath, and K. Krishna. 1966. Studies on Gesonula
punatifrons Stal (Orthoptera: Arcrididae: Crytacanthacridinae)
attacking water-hyacinth in India. Entomophaga 11(5): 433-440.

329
SAS. 1972. A user's guide to the statistical analysis system. North
Carolina State University, Raleigh. 260 pp.
Schwartz, 0. 1928. Die Pontederiaceen. Pflanzenreale 2: 13-14.
Sculthorpe, C. D. 1967. The biology of aquatic vascular plants. Edward
Arnold (Publ.) Ltd., London. 610 pp.
Seabrook, E. L. 1962. The correlation of mosquito breeding to hyacinth
plants. Hyacinth Control J. 1: 18-19.
Seitz, A. 1923. The Macrolepidoptera of the world. Alfred Kernen Publ.,
Stuttgart.
Sen, N. N. 1961. A note on the eradication of water hyacinth in the Ghana
bird sanctuary, Bharatpur Rajasthan. Indian Forest. 87(3): 168-169.
Sheffield, C. W. 1967. Water hyacinth for nutrient removal. Hyacinth
Control J. 6: 37-30.
Sheikh, N. M., S. A. Ahmed, and S. Hedayetullah. 1964. The effect of the
root extract of water hyacinth (Eiohhomia speoiosa Kunth), on the
growth of microorganisms and mash kalai (Phaseolus mungo var. Roxbuvghii) ,
and on alcoholic fermentation. Pakistan J. Sci. Ind. Res. 7: 96-102.
Shelford, V. E. 1913. Animal communities in temperate America.
University of Chicago Press, Chicago.
Shibata, M., K. Yamazaki, and N. Ishikura. 1965. Eichhornin, a new antho-
cynin isolated from the flower of the water hyacinth. Bot. Mag. Tokyo
78: 299-305.
Silveira-Guido, A., H. S. Montero, and J. C. Bruhn. 1965. U.S.D.A. final
project report S9-CR-1. Natural enemies of weed plants. Montevideo,
Uraguay.
Silveira-Guido, A., and B. D. Perkins. 1975. Biology and host specificity
of Cormops aquatioum, a potential biological control agent for water-
hyacinth. Envir. Entolmol. 4(3): 400-404.
Sinha, A. G., and R. Sahai. 1974. Contribution to the ecology of Indian
aquatics. IV. Rate of dry matter production of the leaves of several
common aquatic plants of Gorakhpur (India). Photosynthetica 8(2): 127-129.
Sircar, P. K., S. Banerjee, and S. M. Sircar. 1973. Gibberellin-like
activity in the shoot extract of water-hyacinth. (Eiohhomia cvassipes
Solms). Indian J. Agr. Sci. 43(1): 1-8.
Sircar, S. M., and R. Chakravarty. 1961. The effect of growth-regulating
substances of the root extract of water hyacinth (Eiohhomia speoiosa
Kunth) on jute (Corohorus capsularis Linn). Current Sci. 30(11: 428-430.

330
Skinner, H. 1903. Arzama melanopyga. in Doings of Societies. Entomol.
News: 209-210.
Small, J. K. 1933. Manual of the southeastern flora. University of
North Carolina Press, Chapel Hill. 1554 pp.
Small, J. K. 1936. The water-hyacinth a time clock. J. New York Bot.
Gard. 37: 35-41.
Smith, J. B. 1893. Catalogue of Noctuidae. U.S.N.M. Bull. 44:1-424.
Smith, W. R. 1898. A contribution to the life history of the Ponte-
deriaceae. Bot. Gaz. 25: 324-337.
Spencer, N. R. 1973. Insect enemies of aquatic weeds. Proc. 3rd Int.
Symp. Biol. Control of Weeds: 39-47.
Spencer, N. R. 1974. Insect enemies of aquatic weeds. PANS 20(4): 444-450.
Standley, P. C. 1928. Flora of the Panama Canal Zone. Contrib. U.S.
Nat. Herb. 27: 111-112, Washington; U.S. Gov't. Printing Office.
Stone, A., C. W. Sabrosky, W. W. Wirth, R. H. Foote, and J. R. Coulson.
1965. A catalog of the Diptera of America north of Mexico. U.S.
Gov't. Printing Office, Washington, D.C.. 1696 pp.
Sutton, D. L., and R. D. Blackburn. 1971a. Uptake of copper by water
hyacinth. Hyacinth Control J. 9(1): 18-20.
Sutton, D. L., and R. D. Blackburn. 1971b. Uptake of copper by parrot-
feather and water hyacinth. Proc. 24th Annu. S. Weed Sci. Soc.: 331.
Swamy, B. G. L. 1966. The origin and organization of the embryonic
shoot apex in Eiohhomia orassipes. Bull. Torr. Bot. Club 93(1 ): 20-34.
Tabita, A., and J. W. Woods. 1962. History of hyacinth control in
Florida. Hyacinth Control J. 1: 19-23.
Tag el Seed, M. 1972. Some aspects of the biology and control of Eiohhomia
orassipes (Mart.) Solms. PhD thesis, University of Khartoum. 274 pp.
Tag el Seed, M., and M. Obeid. 1975. Sexual reproduction of Eiohhomia
orassipes (Mart.) Solms in the Nile. Weeds Res. 15(1): 7-12.
Takhtajan, A. 1969. Flowering plants, origin and dispersal (translation
by C. Jeffrey). Smithsonian Institution Press, City of Washington.
310 pp.
Taylor, C. R. 1955. Control of water hyacinth in Rotorua district.
New Zeal. J. Agr. 90: 188-189, 191, 193, 195.

331
Taylor, K. G., R. P. Bates, and R. C. Robbins. 1971. Extraction of
protein from water hyacinth. Hyacinth Control J. 9(1): 20-22.
Taylor, K. G., and R. C. Robbins. 1968. The amino acid composition of
water hyacinth (Eiohhomia orassipes) and its value as a protein
supplement. Hyacinth Control J. 7: 24-25.
Thompson, W. R. 1944. A catalogue of the parasites and predators of
insect pests. Sect. 1. Parasite host catalogue. Pt. 5. Parasites of
the Lepidoptera. The Imperial Parasite Service, Belleville,
Ontario, Canada. 130pp.
Tietz, H. M. 1952. The Lepidoptera of Pennsylvania. Pensylvania State
College, PA. 194 pp.
Tietz, H. M. 1972. An index to the described life history, early stages
and hosts of the Macrolepidoptera of the Continental United States
and Canada. A.C. Allyn, publ., 2 Vols.. 1041 pp.
Tilghman, N. J. 1962. The value of water hyacinth in the propogation
of fish. Hyacinth Control J. 1:8.
Tilghman, N. J. 1963. The St. Johns River hyacinth story. Hyacinth
Control J. 2: 13-14.
Timmer, C. E., and L. W. Weldon. 1967. Evapotranspiration and pollution
of water by water hyacinth. Hyacinth Control J. 6: 34-37.
Timmons, F. L. 1960. Weed control in western irrigation and drainage
systems. U.S. Dept. Agr., Agr. Res. Serv., ARS 34-14. 22 pp.
UTtsch, G. R. 1973. The effects of water hyacinths (Eiohhomia orassipes)
on the microenvironment of aquatic communities. Arch. Hydrobiol.
72(4): 460-473.
Ultsch, G. R., and D. S. Anthony. 1973. The role of the aquatic exchange
of carbon dioxide in the ecology of the water hyacinth (Eiohhomia
orassipes). Florida Sci. 36(1): 16-22.
U.S.D.A. 1970. Selected weeds of the United States. Agr. Handbook 366,
Agr. Res. Serv., U.S. Gov't Printing Office, Washington, D. C., 463 pp.
Van der Weert, R., and G. E. Kamerling. 1974. Evapotranspiration of water
hyacinth (Eiohhomia orassipes). J. Hydrol. 22(34): 201-212.
Vernon, R. 0. 1969. Review of Florida's water pollution problems. Ground
water. Proc. Fla. Envir. Eng. Conf. Water Poll. Control. Gainesville,
Fla. Eng. & Ind. Exp. Sta. 24(3), Bull. Ser. 135: 28-43.
Vogel, E., and A. D. Oliver, Jr. 1969a. Evaluation of Arzama densa as
an aid in the control of water hyacinth in Louisiana. J. Econ.
Entolmol. 62(1): 142-145.

332
Vogel, E., and A. D. Oliver, Jr. 1969b. Life history and some factors
affecting the population of Arzama densa in Louisiana. Ann. Entolmol.
Soc. Amer. 62(4): 749-752.
Wahlquist, H. 1969a. Effects of waterhyacinth on fish production.
F.A.O. Fish Cult. Bull 2(1): 5.
Wahlquist, H. 1969b. Effect of water hyacinths and fertilization on
fish-food organisms and production of bluegill and redear sunfish
in experimental ponds. Proc. 23rd Annu. Conf. SE. Assoc. Game and
Fish Comm.: 373-384.
Walker, C. 1954. Water hyacinth is a dangerous weed. New Zeal. J. Agr.
89: 605-606.
Walker, F. 1864. List of the specimens of lepidopterousinsects in the
collection of the British Museum. Part 32. Suppl. part 2. Catalogue
of the Lepidoptera Heterocera. Seventh Ser. 35 Parts., London.
Wallwork, J. A. 1965. A leaf-boring galumnoid mite (Acari: Cryptostigmata)
from Uruguay. Acarologia 7(4): 758-764.
Warner, R. E. 1970. Neochetina eiohhormiae, a new species of weevil from
waterhyacinth, and biological notes on it and N. bruohi. Proc. Entomol.
Soc. Washington 72(4): 487-496.
Webber, H. J. 1897. The water hyacinth and its relation to navigation in
Florida. U.S.D.A. Bull. 18: 1-20.
Welch, P. S. 1914. Habits of the larva of Bellura melanopyga Grote
(Lepidoptera). Biol. Bull. 27: 97-114.
Wellington, W. G. 1960. Qualitative changes in natural populations
during changes in abundance. Can. J. Zool. 38: 289-314.
Westlake, D. F. 1963. Comparisons of plant productivity. Biol. Rev.
38: 385-425.
Worthington, C. E. 1878. Arzama obliquata. in Miscellaneous Memoranda.
Can. Entomol. 10: 15-16.
Wunderlich, W. E. 1964. Water hyacinth control in Louisiana. Hyacinth
Control J. 3: 4-7.
Yount, J. L. 1964. Aquatic nutrient reduction-potential and possible
methods. Rep. 35th Annu. Meeting, Florida Anti-Mosquite Assoc.: 83-85.
Yount, J. L., and R. A. Crossman, Jr. 1970. Eutrophication control by
plant harvesting. J. Water Poll. Control Fed. 42(5) Part 2: R173-183.
Zeiger, C. F. 1962. Hyacinth-obstruction to navigation. Hyacinth Control
J. 1: 16-17.
Zettler, F. W., and T. E. Freeman. 1972. Plant pathogens as biocontrols
of aquatic weeds. Annu. Rev. Phytopathol. 10: 455-470.

333
BIOGRAPHICAL SKETCH
Ted Douglas Center was born 15 August 1947 in Dayton, Montgomery
Co., Ohio. He attended Belmont Elementary School and Belmont High School
where he graduated in 1965. Following high school he attended Ohio
University in Athens, Ohio for one year, after which he transferred to
Foothill College in Los Altos Hills, California. After a year in
California he transferred to Northern Arizona University in 1967 from which
he received his Bachelor of Science degree in Zoology in 1970 and his
Master of Science degree in Biology in 1972. He transferred his studies
to the University of Florida in September 1971 where he is currently
completing the requirements for a Ph.D in Entomology working on the
biological control of aquatic weeds.
Ted Center's work experience began at the age of 14 when he became
employed at the Dayton Museum of Natural History. He remained on the
staff working part-time during high school and full-time during the
summers from 1961 through 1969. His duties included regular television
appearances on a local children's show, instructing museum nature classes,
presenting lectures to various civic groups, care of the museum's live
animal collection, curating and preparing specimens for the museum's
collections, and participation in various research projects. While in
California he performed similar duties at the Palo Alto Children's
Museum.
While in Arizona he was employed part-time in the Biology Department
of N.A.U. curating the insect collection and preparing bird and mammal
specimens. He also worked part-time for the Geology Department of the
Museum of Northern Arizona. In the summer of 1971 he was employed by

334
Mr. A. A. Kirk of the Australian division of C.S.I.R.O. to help collect,
culture, and ship wood wasp parasites from Arizona for their biological
control program. His Master's Degree dealt with the biology and coevo
lution of seed beetles and their host plants.
He is a member of the Ecological Society of America and the Entomological
Society of America.
He is married to the former Deborah Jean Learned.

I certify that I
conforms to acceptable
adequate, in scope and
Doctor of Philosophy.
I certify that I
conforms to acceptable
adequate, in scope and
Doctor of Philosophy.
I certify that I
conforms to acceptable
adequate, in scope and
Doctor of Philosophy.
I certify that I
conforms to acceptable
adequate, in scope and
Doctor of Philosophy.
I certify that I
conforms to acceptable
adequate, in scope and
Doctor of Philosophy.
have read this study and that in my opinion it
standards of scholarly presentation and is fully
quality, as a dissertation for the degree of
Dale H. Habeck, Chairman
Professor of Entomology
have read this study and that in my opinion it
standards of scholarly presentation and is fully
quality, as a dissertation for the degree of
er
Professor of Entomology
have read this study and that in my opinion it
standards of scholarly presentation and is fully
quality, as a dissertation for the degree of
George E/ Allen
Professor of Entomology
have read this study and that in my opinion it
standards of scholarly presentation and is fully
quality, as a dissertation for the degree of
Reece I. Sai
have read this study and that in my opinion it
standards of scholarly presentation and is fully
quality, as a dissertation for the degree of
John Reiskind
Associate Professor Zoology

This dissertation was submitted to the Graduate Faculty of the College
of Agriculture and to the Graduate Council, and was accepted as partial
fulfillment of the requirements for the degree of Doctor of Philosophy.
June 1976
2 oJk .
Deal/,/College of Agriculture
Dean, Graduate School



23
the Potomac River near Washington, D.C. (ca. 30N; Gowanloch and Bajkov
1948; U.S.D.A. 1970), Japan (30-35N; Holm aval. 1969) and possibly
Portugal (37-40N) as indicated by Holm et al. (1969) on their distri
bution map. The southern most limits of distribution appear to be Buenos
Aires, Argentina (34S) and Concepcion, Chile (37S) in South America
(Castellanos 1959), and Shannon, New Zealand (40-41S; Anonymous 1964).
The range in general seems to be bounded by the 40 North and South
Latitude lines. Very little information is available on the altitudinal
restrictions of this species although one paper (Anonymous 1957) states
that it is limited in the tropics to an altitudinal zone of from sea
level to 4500 feet (ca. 1400 m).
Habitat
Little is known of the ranges of environmental tolerances of
waterhyacinth. Webber (1897) noted the effects of freezing temperatures
in Florida in the winter of 1894-95. He noted that the first freeze
killed the top which caused the plant to float higher in the water.
A second freeze killed this newly exposed portion. Most of the plants
survived by resprouting from the unexposed portion of the rhizome.
Buckman and Co. (1930) stated that temperatures as low as 28F
(-2.8C) may be withstood by the roots but will kill the tops. Temper
atures lower than this will kill the roots as well. Hitchcock et al.
(1950) observed waterhyacinths subjected to two days of freezing in
New York. The plants were ice-covered when transferred to the green
house. Damage was apparently severe as the authors noted that all the
foliage and all the roots were killed. Within 13 days the plants had
recovered by resprouting from the rhizome tip. Misra (1969) placed


14
content. They determined that the nutrient content in the plant tissues
was not proportional to that of the water in which they were grown. Boyd
(1974) has summarized the data on the composition of waterhyacinth and
other aquatic plants.
Liang and Lovell (1971) evaluated waterhyacinth for use in channel
catfish feed. They found that the addition of 5 to 10% waterhyacinth in
vitamin free diets increased growth and reduced mortality in the fingerlings.
Bagnall et al. (1974) using waterhyacinth as feed supplements for
cattle and sheep, for paper production, and for mulch determined that
its processing as mulch was the most economically feasible use.
Azam (1941) proposed that underdeveloped countries encourage their
people to utilize their spare time preparing various products made from
waterhyacinth and thus supplement their income. Some of the products
they suggested were paper, pressed board and tiles, detergents, cattle
fodder, and manure. Nolan and Kirmse (1974) considered waterhyacinth
jsable in the production of paper.
Iswaren and Sen (1973) found that an extract from waterhyacinth
roots increased the yield of Brinjal (Solomon melangena var. Pusa
Kranti) from 507.2 g/plant to 1317.3 g/plant. Ganguly and Sircar (1964)
found that a root extract from waterhyacinth increased the metabolic
activity and the nitrogen and sugar content of Vision sativum L. seedlings.
Mukherjee et al. (1964) identified growth promoting substances in the
roots of waterhyacinth which they believed to be bound auxins. Sheikh
et al. (1964) noted that this extract was thermostable and found it to
promote the growth of Phaseolus mungo var. roxburghii, the mycelium of
Aspergillus niger, the growth of Rhizopus, and the multiplication of


25
although relatively high growth occurred over the range of 22-35C.
Exposure to 10C nights reduced the amount of photosynthesis on
following warm days.
Bock (1966, 1968) exposed plants to 26.7C-26.7C, 26.7C-
4.4C, and 4.4C-4.4C day-night temperatures under both 16-18 and
8-16 L:D photoperiods. She found that growth was favored in the higher
temperatures although it also appeared to be favored in the shortened
photoperiod.
The maximum tolerable water temperatures appear to be around
33-34C. Penfound and Earle (1948) observed that the plants cannot
tolerate water temperatures above 34C. Misra (1969) stated that in
India the plants succumb at water temperatures above 33C. Knipling
et al. (1970) found that growth began to be inhibited at about 33C
and declined in a nearly linear manner at higher temperatures until,
by 40C, negative growth was indicated. They noted that the plants
were more tolerant of lower than optimum temperatures than of higher
than optimum.
Light relations have been investigated by a few authors. Penfound
and Earle (1948) noted that in July the average light intensity was
about 420 foot-candles above colonies of moderate-sized plants. Under
the canopy of large plants the light intensity was about 170 ft-c repre
senting a 60% decrease. They found that equitant elongate leaves are
formed at intensities ranging from 130-500 ft-c and float leaves are
formed at intensities over 500 ft-c. Under a walkway where the light
intensity was 130 ft-c (31% of the July average) most of the plants
were found to be dying. They also placed containers of waterhyacinth


Figure 50. The population of living Arzama densa larvae (per square meter)
present on Lake Alice and the number of dead larvae expressed as
a percentage of the total.


171
remove the effects of the second. In this case the second variable is
considered as not important and is excluded from the model. This may
very well have been the case with insect damage. The effects of insect
damage may have been obscured in this manner by the greater effects of
sunlight, temperature, and water level. To see if this possibility existed
a simple correlation analysis was performed comparing each independent
variable with each dependent variable on a one to one basis (Table 10).
Both estimates of Arzama damage were highly negatively correlated with
plant height, leaf area index and leaves per plant. While this does not
necessarily infer that a causal relationship exists between insect
damage and the plant characteristics the possibility is present. Further
more sophisticiated analyses may be able to isolate this effect but the
results of this study are inconclusive with regard to damage by Arzama
densa.
The models produced seemed to fall into two broad categories, bio
mass and density. The first includes standing crop, plant height, leaf
area index, and leaves per plant. These four characteristics are inter
related as evidenced by the significant positive correlations (Table 9)
between them. Because of this interrelationship all are probably indices
of biomass and all are low in the winter, increase in the spring, reach
their peaks in early summer, and decline in the fall. All of these are
primarily climatologically limited as evidenced from the inclusion of
solar radiation and minimum temperature in each multiple regression
model (Table 7).
The major nutrients (N, P, K) are all included in at least one of
the four above mentioned models. The coefficients for these are somewhat
difficult to interpret. Nitrogen is included in the standing crop model.


24
plants in a deepfreeze at -10C for 7-8 hours and they failed to revive.
He found that at 15C the growth became restricted and the plant did
not show any increase in growth up to a 90-day period.
Penfound and Earle (1948) exposed small plants in trays with 3
inches of water to various air temperatures for various durations. They
found that at 33F and 27F (0.56C and -2.78C) all of the rhizomes
resprouted when returned to room temperatures after being exposed for
12, 24, and 48 hours. At 23F (-5C) all resprouted after 12 and 24
hours but none survived after 48 hours. At 21F (-6.11C) some survived
12 hour exposures but none survived 24 or 48 hour exposures. At 19F
(-7.22C) none resprouted after being exposed for 12, 24, or 48 hours.
They concluded that the temperature effect depends upon the duration
of exposure and that freezing of the rhizome tip results in the
destruction of the plant.
Hitchcock et al. (1949) found that satisfactory growth
occurred in air temperatures of 21-27C. Silveira-Gui.do et al. (1965)
stated that the plants grew well in water temperatures ranging from
17-35C. Bock (1966) measured air temperature ranges of 17-35C and
water temperature ranges of 18.6-21.5C during the waterhyacinth
growing season and winter mid-day temperatures of 5-10C for air and
5.8C for water in California. She also stated that the populations
survived the winter of 1963-4 from which she monitored 28 da with
air temperatures below freezing, 1964-65 with 25 day, and 1965-6
with 35 da although considerable mortality did occur.
Knipling et al. (1970) measured waterhyacinth growth along a
gradient of water temperatures. They found the optimum to be 28-30C


estimates are plotted in Figure 24. Subsequent actual measurements taken
from April 1975 February 1976 are also plotted (vertical bars) and
conform very well with the estimated curve. As a result I feel that the
estimated curve is a fairly accurate representation of acutal conditions
Standing crop was lowest in January and February when it ranged
between 600 and 1000 gm DW/m2. A period of exponential growth began in
March and continued through April. The discontinuity of the curve at
the end of April is due to differences between the initial measurements
in May 1974 and the final measurements in 1975. It is therefore not
apparent whether this exponential trend continued. It does not seem to,
however, because there was a drop in biomass between April and May of
1975 in the actual biomass samples. Following this decline a nearly
linear rise occurred through late June. During this period of active
growth the growth rate was fastest in March and April. The daily incre
ment factor (see Bock 1969) for this period (5 March 23 April) is
estimated at 1.015 (1.5% per day). The same value for the period be
tween 1 May and 10 July is 1.010 (1.0% per day). The average monthly
values for daily increments are listed in Table 6.
Following the peak in July a rapid decline occurred. This was
followed by a stable period in mid-September. A general decline began
the first part of October and continued through January. Biomass re
mained stable in February before beginning a spring resurgence.
Surprisingly, the greatest monthly decline did not occur during
the coldest months but rather in October and November. This may have
been due to heavy damage by Arzama densa which occurred at that time.
The period of maximum growth (April) corresponds to the greatest monthly


104
be more limiting in this system than phosphates since stronger winter
peaks and summer minima were apparent in annual nitrogen cycles and signi
ficant drops in the nitrate concentration occurred across the marsh.
Dunigan et al. (1975) found that phosphate-phosphorous concentrations
above 50 ppm were not significantly affected by the growth of waterhyacinths
but both nitrate-nitrogen and ammonia-nitrogen were. They further found
that the N:P ratio of uptake rates was 5-6:1.
Potassium concentrations (Fig. 7) remained fairly constant and did
not show any strong annual variation. Values between June and November
generally ranged between 2 and 4 mg/1. Concentrations increased somewhat
from December through February and an abrupt decline occurred in March
and April By June potassium concentrations were about the same as they
had been the previous year.
Sulfates (Fig. 7) were extremely variable but a bimodal tendency
was observed. Concentrations appeared to be maximum in the fall and
spring and minimum in the winter and summer. Since plants take up rela
tively little sulfate compared to the amount available (Ruttner 1972)
this pattern is probably due to factors other than waterhyacinth growth.
Temperature and Solar Radiation
Figure 8 indicates the weekly maximum, minimum, and median air and
water temperatures at Lake Alice during this study. Knipling et al. (1970)
found that maximum growth of waterhyacinths was favored at water temperatures
of 28-30C. Water temperatures greater than 40C were lethal and growth
decreased linearly as temperatures were reduced to 15C. In comparisons
of plants exposed to 30C days:30C nights with plants exposed to 30C
days:10C nights they found that the plants exposed to the lower nighttime


Host Plants
60
Table 2 lists host plant records for these species as indicated in
various references. Because of the continual changes in the taxonomy of
the group, however, these host records are not reliable. For example,
Grossbeck (1917) listed Bellura gortynoid.es Wlk. from Mellonville, Florida,
as inidcated by Hampson (1910) and implicated Typha as the host. Kimball
(1965), however, indicated that the Mellonville record quoted by Grossbeck
(1917) referred to Arsama densa. This creates uncertainty since the host
record was not from Hampson (1910), who synonymized densa and gortynoides,
but the geographic record was. Grossbeck (1917) apparently derived the
host record from other sources. It is therefore impossible to determine
which species Grossbeck's host record refers to. To partially alleviate
this problem I have left the records in Table 2 with the binomial designated
by the respective author intact regardless of synonyms. Where, in my
opinion, there is sufficient agreement in the literature to indicate that
a name is in synonymy with a more valid binomial, that species designation
is included as a subcategory under the valid binomial.
Host plant synonymies also result in a great deal of confusion.
For example, Nymphaea americana (Prov.) Miller & Standiey listed as
a host of Bellura melanopyga (Table 2, No. 2c) is listed by Muenscher
(1967) as a synonym of Nuphar variegatum Engelm. Nymphaea advena (Table
2, No. 2) is also apparently a synonym of Nuphar variegatum (Fassett
1969). I do not believe any bona fide record exists of these species
attacking any of the Nymphaea species.
I am not sure what the Nelumbium sp. (Table 2, No. la) and the
Nelumbom sp. (Table 2, No. lc) records refer to. They may mean Nelumbo
but, if this is so, I doubt the veracity of the record. I also question
the records for Sagittaria sp. (Table 2, Nos. la, lc) and Sparganium
sp. (Table 2, No. 1


43
phenomenon of anthokinesis or the bending of the axis of the inflorescence
following anthesis (Agharkar and Banerji 1930; LaGarde 1930; McLean 1922;
Penfound and Earle 1948; Bock 1966; Misra 1969). LaGarde (1930) described
this process as follows:
"As soon as the inflorescence starts wilting the upper portion of
the stalk with the fertilized blossoms begins to bend downward. When
this upper part has reached the surface of the water, usually after five
days, the lower portion of the stalk commences to bend at the base, thus
pushing the developing seed-pods under the surface of the water. This
movement stops when the lower part of the stalk is level with the surface.
The upper part carrying the seed pods is then submerged in the water at
an angle of 45, the seed pods being covered and protected by the root
system . The whole process of bending requires from six to seven days."
(LaGarde 1930, p. 51).
Agharkar and Banerji (1930) quoted other workers who indicated
that anthokinesis was accompanied by a lengthening of the peduncle. They
found, however, that the peduncle did not lengthen considerably and such
lengthening was confined to an area of 1 to 2 cm below the terminal node.
They further found that removal of the flowers or amputation of the peduncle
above the node had no effect and curvature was normal. This was also true
when they removed the peduncle and placed it in water.
Penfound and Earle (1948) studied that anthokinetic cycle and their
results agree with other workers. They found that it requires about 14
days from the initiation of the floral bud until opening occurs. Floral
opening begins about 8:00 AM, if all the flowers open the bending phase
begins at about 5:00 PM of the same day. Bending occurs in three places:
at the rhizome crown, j r' elow the two bracts of the inflorescence,


Figure 36. Net waterhyacinth production as a function of insect feeding
activity. Net production, as used here, refers to the final
quantity of organic matter present,excluding the initial
amount of detritus, as a percentage of the initial living
plant material. Data only from the fall experiment. Legend
as in Figure 28.
Regression: Y = 194.42 32.54X, r = -0.1897


103
Unfortunately orthophosphate-phosphorus was not determined in this
study. Figure 5 represents the phosphorus concentration present as total
phosphates. While this does not define orthophosphate availability it
probably does give an index of it. Orthophosphate is the only form of
phosphate derived from natural sources and is present in organic wastes
and fertilizers (Vernon 1969). Hence, it is probably the predominant form
in Lake Alice. Phosphate-phosphorus concentrations were lowest in July
and August 1974 (ca. 0.5 mg/1) and increased through December. Concen
trations remained high into May and June 1975 and failed to return to
the 1974 levels. This may have been an artifact of the lake level as it
also failed to return to its previous July level. Also an increase in
phosphate concentrations in April and May appear to correspond to a drop
in the water level at the same time.
Phosphorous is a primary factor limiting production in aquatic eco
systems. Haller et al. (1970) found that P-concentrations below 0.1 mg/1
were limiting to waterhyacinth growth. Above this concentration the plants
absorbed P in luxury amounts. P-concentrations remained above this cri
tical concentration throughout the year in Lake Alice. I assume, therefore,
that P does not become limiting to waterhyacinth at this site. The fact
that phosphorous is lowest when plant biomass is highest probably reflects
the increasing absorption of these luxury amounts by an increased plant
standing crop. Mitsch (1975), however, found in a model simulation that
phosphorous concentrations in Lake Alice appeared to be unaffected by
waterhyacinth uptake or any other annual cycle. He further found that
there was very little decrease in phosphorous in the direction of flow
across the waterhyacinth marsh. He hypothesized rather that nitrates may


93
treated varies with the number of students present at the university at
different times of the year, output from the sewage treatment plant varies,
the amount of water (hence, the diluting effect) from the heating plant
varies, rainfall varies as does runoff from near fertilized farmland.
All of these factors affect nutrient concentrations. Since the samples
were taken from the study area, however, they should reflect the relative
conditions the plants were growing in at the time. The reasons for differences
in the nutrient loads are not important but the differences themselves are.
Alkalinity (Fig. 3) generally ranged from 140 to 210 mg/1. Concen
trations were fairly constant. The greatest fluctuations occurred in
December when a sharp drop was apparent and in January when an equally abrupt
increase occurred. The concentration at the end of the study (11 June 1975)
was somewhat higher than that at the beginning (20 June 1974). Alkalinity
is a measure of the buffering capacity of the lake and in indication
of eutrophication.
Conductivity (Fig. 3) ranged from 340-580 ymhos/cm and is a measure
of the electrical conductance of the water resulting from soluble salt
concentrations. The conductivity values fluctuated greatly throughout
the year. In general periods of low conductivity seemed to correspond
to periods of frequent rainfall. This is presumably due to a dilution
of the soluble salts present relative to the water storage of the lake.
Iron (Fig 4) was measured because I have observed in greenhouse
cultures that waterhyacinths grow much better in nutrient solutions
high in iron concentrations. Further, it appeared that the iron in the
solutions was rapidly absorbed by the plants. Iron is a constituent of
cytochromes and as such is an essential micronutrient for plant growth.
Concentrations in Lake Alice were less than 0.10 mg/1 most of the year


272
and field studies, however, I have accumulated many observations on the
bionomics of this species which I feel are significant. I will present-
these observations at this point, but since the methods employed are
diverse, I will omit a methods section and instead explain them as the
results are presented.
Habits
The eggs of A. densa are laid in a mass on the upper surface of the
leaves of Eiohhormia arassipes (Mart.) Sol ms or Pontederia sp. The egg
mass is very similar to that described by Riley (1883a, p. 174) for Arzama
obliquata (= Bellura obliqua) as "broadly convex or plano-convex masses
enveloped in hair, and a cream colored mucous secretion". Vogel and
Oliver (1969) described the masses as being covered with light yellow
body hairs. The egg masses I have observed have been tan or creme colored
rather than yellow. Vogel and Olvier (1969) further indicated that each
egg mass contained 30-40 eggs. From 11 masses collected in 1973 from
waterhyacinth I have found the range to be 19-66 with an average of 42
(+ 14.57 s.d.). Further, I have frequently found eggs laid singly or
in groups of two or three often on the petiole of a leaf near the plant
base. These are usually not provided with a covering and may be left
by a resting female as an artifact of a previous egg extrusion. I have
frequently noted a few eggs clinging to the abdomens of caged females
after oviposition. These remnants may account for the single ovipositions
noted in the field.
From these same caged females I have observed that the eggs are not
deposited in clusters when provided with an artificial substrate (i.e.
wax paper, guaze). When leaf bouquets are available the eggs deposited
/
on the leaves are usually in the typical masses. This suggests that this


Figure 37. The ratio of conversion of living waterhyacinth plant material
into detritus as a function of insect feeding activity. The
figures are based on the change in detritus over one generation
of Arsama densa population divided by the amount of convertible
living material present at the time the insects were introduced.
Data only from the fall experiment. Legend as in Figure 28.
Regression: Y = 0.3254 + 0.9007X, r = -0.7521




81
Site Description
Lake Alice is located in the southwest corner of the University of
Florida campus in Gainesville, Alachua Co., Florida (Topographic desig
nation: Gainesville East quadrangle, TIOS, R19E, N£?s, R20E, NW^). The
lake was once a sinkhole fed by a small stream but damming off the west
end in the late 1940's and later the addition of effluents from the
campus sewage treatment plant and the heating plant resulted in its pre
sent configuration (see Figure 1). The lake area is approximately 33 ha
and is divided into a marsh dominated by waterhyacinths and an area main
tained by the University as open water. The marsh at the east end com
prises approximately 65% (21 ha) of the lake surface and is separated
from the open lake by a catwalk and fence constructed to retain the
waterhyacinths. The depth of the marsh in generally less than 2 meters
(Cason 1970). The "open" western end of the lake covers about 12 ha
and is also generally less than 2 meters in depth with a few areas of
about 5 meters, probably the original sinkholes (Mitsch 1975). The
general flow of the lake is from the sewage plant and heating plant
effluent at the eastern end through the marsh to the open lake at the
western end where it discharges through two wells into the Florida
aquifer.
The lake is situated on Ocala limestone which is dominated by a
karst topography. Solution sinkholes, fractures, and caverns are typical
of this type of topography and are common in this area. Because of the
silt that has accumulated on the bottom, however, the lake basin is
maintained above the local water table (Cason 1970). The lake level is
generally between 68 and 70 feet above mean sea level. Figure 2 indi
cates the lake level at the catwalk for the period of this study.


100
Mean Daily Solar Radiation (Langleys)
M
§
u>
o
o
o
o
500


LITERATURE REVIEW
Eichhomia crassipes (Mart.) Solms
Taxonomy
Bock (1966) provides an excellent historical review of the liter
ature dealing with the taxonomy of Eichhomia crassipes. The most
current treatment of the genus appears to be that of Agostini (1974)
which describes the species occurring in Venezuela. Five species are
described (E. azurea, E. crassipes3 E. diversifolia, E. heterosperma,
and E. paradoxa) and a key provided. The synonomy provided for E.
crassipes is as follows:
Eiohhomia crassipes (Mart.) Solms in DC., Monogr. Phan. 4:527.
1883.
Pontederia crassipes Mart., Nov. Gen. 1:9. t. 4. 1824.
Piaropus crassipes (Mart.) Raf., FI. Tell. 2: 81. 1837.
Eiohhomia speciosa Kunth, Enum. PI. 4: 131. 1843.
Eiohhomia cordifolia Gandog., Bull. Soc. Bot. France 66:
294. 1920.
While the bionomial E. crassipes is in common usage today the
synonyms Piaropus crassipes and E. speciosa are common in the literature.
Because of the world-wide distribution of this plant it is known
by a large variety of common names. Bock (1966) lists 48 common names
for Eiohhomia crassipes from 18 countries. The name waterhyacinth is
used world-wide in scientific reports but the structure of the word
has often been left up to the discretion of the user. It is often
written as two words (water hyacinth), a hyphenated word (water-hyacinth),
or as one word. Kelsey and Dayton (1942) in a list of standardized plant
names use the single word, waterhyacinth. This usage seems appropriate
since the plant is not related to thehyacinth as the two word name would
3


173
The simple correlation coefficient between standing crop and nitrogen
(Table 10) is negative and not significant. After the effects of sun
light and solar radiation are removed this relationship is positive as
is evidenced by the positive coefficient in the model. Hence this states
that at constant levels of light and temperature the standing crop in
creases as nitrogen levels increase. However, since absorption of nutrients
increases as standing crop increases one would expect a negative rela
tionship between the two if nitrogen input from the various sources re
mains constant. This inverse relationship seems to be apparent in Figures
5 and 24. The drop in nitrogen concentration in the summer, however, is
more likely due to an increase in the water level resulting in a dilution
of the nutrient load. This is evidenced by the significant correlation
between nitrogen and lake level (Table 8). Mitsch (1975) demonstrated
that a decrease in nitrates across the marsh does occur and is greatest
in the summer when the waterhyacinth standing crop is high. Nitrogen,
then, may very well be limiting to waterhyacinths and the biomass supported
by the available nutrients may increase as relative nitrogen concentrations
increase.
*-
Phosphorus is generally considered one of the primary limiting factors
is aquatic systems (H. T. Odum 1953). It is included in the model for
plant height and the coefficient is negative. The simple correlation co
efficient between phosphorus and height is also negative and significant.
The model infers that with climatological effects removed, plant height
increases as phosphorus decreases. This suggests an increasing absorption
of phosphorus as the biomass increases. Phosphorus concentration, however,
is also affected by the water level. High phosphorus concentrations are
correlated with low water levels (Table 8). Mitsch (1975) showed a small


68
Crumb (1956) provided a key to the larvae of the Amphipyrinae. The
couplet separating Arzama used the large sub-dorsal spiracles on the 8th
abdominal segment as a key character. He also described the larva of
Arzama obliqua (Wlk.).
Vogel and Oliver published two papers (1969a, b) on Arzama densa
Wlk. Their first paper was on the potential of this insect to control
waterhyacinth. Their second paper was on the life history of A. densa.
They provided cursory descriptions of the immature stages and determined
the developmental times of the various stadia. Much of their data,
however, is from larvae reared on artificial media which makes their
results subject to question. These two papers will be further discussed
later in this dissertation.
Levine (1974) found that in Indiana there were two complete genera
tions per year of Bellura gortynoides Wlk. (= B. vulnifiea Grt.). He
found that the first generation (spring) pupates within the petiole of
Nuphar advena. The second generation (fall) larvae swim to shore and
overwinter as larvae under the bark of trees, in rotten wood, or in
leaf litter. The eggs hatch in 6 days and there are 6 to 7 instars.
Parasites, Predators, and Diseases.
The first record of natural enemies which attack this group of
insects was that of Welch (1914) for Bellura melanopyga Grt. He noted
that sunfish ate the larvae when they were swimming on the surface. He
also observed water striders (Gerris sp.) attacking the larvae when they
were on the surface of the leaves.
Claassen (1921) found Sturmia nigrita Town. (Diptera:Tachinidae)
parasitizing the larvae of Arzama obliqua (Wlk.). Robertson-Miller (1923)


90,000 of Florida's 2,500,000 acres of fresh water and 70,000 to 100,000
acres of Louisiana's 2,000,000 acres of fresh water are covered with
waterhyacinth (Ingersoll 1974). Hudson (pers. comm.) estimates that in
1975 the acreage of waterhyacinth in Florida has extended to more than
200,000 acres and the average cost of control per acre is about $25. He
estimated that all agencies within the state in FY 1976 allocated $16
million for aquatic weed control, about 30% of which ($4.8 million)
goes towards waterhyacinth control. This is an increase of almost $2
million over the previous year (FY 1975) for waterhyacinth control
alone.
Thompson (pers comm.) indicated that between 1965 and 1974 the
U.S. Army Corps of Engineers spent $6.1 million in combined construction
and operations funds for aquatic weed control in Louisiana alone. In
the period between 1960 and 1964 the estimated cost was $1.7 million.
Other weeds are of minor concern and for the most part 100% of this
went toward waterhyacinth and alligator weed control. The State of
Louisiana beginning its program in the mid 1940's spent $8.1 million
as of 1973. Further costs included $1 million in 1974 and $1.1 million
in 1975. Thompson further estimated that the average cost of treating
an acre is between $32 and $35 in Louisiana. The most economical means
being by helicopter ($13/acre) or fixed wing aircraft ($10/acre) when
possible. The current estimate of acreage covered in Louisiana exceeds
1 million acres. This does not necessarily reflect an increase in acreage
over Ingersoll's (1974) figure but is merely a more accurate estimate.
It is evident from these figures that the acreage covered with
the plant is increasing while the cost of treating an acre is also


CHAPTER III
THE FEASIBILITY OF THE UTILIZATION OF ARZAMA DENSA WLK. FOR
THE BIOLOGICAL CONTROL OF WATERHYACINTH THE EFFECTS
OF AN INTRODUCED POPULATION ON A SMALL POND COMMUNITY.
Introduction
It has been suggested that Arsama densa Walker could be used for
the biological control of waterhyacinth Eichhomia erassipes (Mart.)
Solms) by supplementing natural populations if a staisfactory method
of mass-rearing was developed (Vogel and Oliver 1969a). Frick (1974)
also suggested the possibility of augmenting populations of native
insects to increase their effectiveness in weed control. While this
tactic has been discussed by various investigators in the field of bio
logical control there are few examples of studies where this has been
attempted in an effort to control weeds.
Sufficient numbers of A. densa were reared on living waterhyacinth
plants in a greenhouse to release for a small scale field test. Three
variables were important in the location and timing of this release.
First, the site tested had to be small to achieve an adequate insect:
plant concentration. Second, the release had to be synchronized at a
time when parasite populations were low and the naturals, densa popu
lation was increasing. Third, the release had to be strategically
made so as to damage the plants at a time when they were most vulnerable
to attack. With this criteria in mind a small site near Paynes Prairie
in Alachua Co. was selected where the previous summer the largest buildup
of the natural A. densa population occurred in the late summer and fall.
Also, an attack late in the growing season of the plants should increase
244


AV HEIGHT (CM )
PLANT HEIGHT
j
F
M
A
M
no



44
and in the rachis. Most of the flowers are inverted by 5:00 PM the
following day. The complete cycle from flowering to complete geniculation
takes 48 hours in the summer. This is contrary to LaGarde's (1930)
finding that it takes 6 or 7 days. Bock (1966), in her studies, agreed
with Penfound and Earle (1948).
Bock (1966) seemed to concur with the findings of Rao (1920a) in
that the: bending was due to geotropism in that when the roots were packed
with sponges and the plants held horizontally, no bending occurred. She
disagreed with Agharkar and Banerji (1930) in that removal of the flowers
would not permit bending to occur unless all of the flowers had wilted
and bending had commenced first.
Misra (1969) found that curvature took place when the tips along
with 2 terminal flowers were removed, when all of the flower buds were
removed, and when all of the flowers were removed after they had opened.
In all cases complete bending took as long as in the controls (35-40 hrs.).
He found that this curvature was due to increased cell size along the
outer edge of the curving portion. He felt that this process represented
a free-running endogenous rhythm independent of auxins (geotropic in
nature), photoperiod, temperature, and opening of the last flower as
suggested by other authors.
Spermatogenesis has been described by Smith (1898) and Banerji and
Gamgulee (1937) and oogenesis by Smith (1898). Pollen morphology and
germination and development of the pollen tube have been investigated
by Ganerji and Gangulee (1937), Bock (1966) and Tag el Seed and Obeid
(1975). The embryology of the seed is discussed by Smith (1898), (Coker
(1907), and Swamy (1966).


277
Table 14. Fecundity, egg viability, and egg stadia for 5 female
Arsama densa collected as pupae in the field and mated
in the laboratory.
No.
#Eggs/
%Laid on
Average egg
# Eggs
%
first day
stadium (da.)
Hatched
Hatch
1
194
87
5.0
167
86.1
2
228
74
5.3
186
81.6
3
257
60
6.0
252
98.0
4
262
60
6.5
176
67.2
5
184
82
5.0
130
70.7
X
225
72.6
5.6
182
80.7
S.D.
35.5
12.4
0.7
44.4
12.4


MAGNESIUM
97
MAGNESIUM (ppm)
O u"> O u-j
(wdd) NOdl


Figure 1. An aerial view of Lake Alice on the University of Florida campus.
The white square designates the study area. Notice the darker color
of the waterhyacinths in the main flow of the sewage effluent.


Figure 47. The head capsule diameter of Avzama densa larvae at each
molt plotted against the larval age. The data was derived
from larvae reared in cups containing pickerelweed or water
hyacinth petiole sections in an environmental chamber at
25C and 14:10 L:D photophase. The stars represent the
head capsule size of 6th and 7th instar field collected
larvae and extend the trend to that expected had the larvae
developed "normally."


26
under a table where the light intensity averaged 55 ft-c and all of
the plants died in 2 mo. In connection with this they placed several
plants in the dark and measured the starch depletion. By 7 da the
starch content was reduced by 50% and by 12 da it was completely gone.
Hitchcock et al. (1949) grew plants in a greenhouse and supplied
one group with supplemental heat, one group with supplemental light,
and one group was left as a check. They found that the no. leaves per
plant, the average leaf length, and the no. flowers produced were greatest
in the high light condition.
In Africa (Anonymous 1957) it has been noted that light is seldom
a limiting factor with respect to vegetation and frutification but it
may have a more direct influence on germination.
As previously mentioned one study (Bock 1966, 1968) found that
plants grown under the same temperature ranges grew better under the
shorter photoperiod. This peculiarity was not explained.
Bock (1966) stated that waterhyacinth needed 60% full sunlight
or better although she failed to define full sunlight. She placed
plants under greenhouse benches when the light intensity at noon was
30-40% full sunlight. These were retained there from September to March
and 67% mortality was observed.
Misra (1969) subjected plants to 40%, 70%, and 100% full sunlight
(again undefined) and found that the no. leaves per plant and the per
centage leaves with floats increased with increasing light intensity.
Correspondingly a reduction in the average volume and diameter of the
float occurred as the light intensity decreased.


286
measurements for the first 5 instars very useful for identifying the
stages of field collected larvae. The values for the sixth and
seventh instars, however, seem to be much smaller than those from the
field. These data are suspect and the smaller size may have resulted
from the rearing conditions.


119
of 17.78 gm or an increase of approximately 3.02%. Hence, one would
expect the small plants to increase relatively more rapidly than the
large plants.
The reasons for this difference are many. The increased metabolic
load in the larger plants due to the larger standing crop results in a
lower gross primary productivity:respiration ratio (1.46 vs. 2.05).
This ratio at steady state is 1.00 which indicates that the larger plants
are closer to steady state than the smaller ones. This infers that a
greater portion of the gross production is spent in maintaining existing
plant structure than in producing new material. This may be important in
biological control considerations for a herbivore which merely removes
leaf tissue without doing damage to the growing portion of the plant
may indirectly stimulate growth.
Even though the total metabolic load was greater in the large plants
the respiration per gram plant biomass was more than double in the smaller
plants (see Table 5). This may be an indication of a more active metabolic
rate associated with a faster growth rate.
Intraspecific competition for light may be another factor affecting
the observed difference in growth rates. The leaf area index of the
large plants was more than twice that of the small plants. The total
amount of photosynthetic tissue was 3 times greater in the large plants
and the leaf (pseudolamina) tissue was 9 times greater. In spite of these
large differences net efficiencies were equal and gross efficiency was
only 62% greater in the large plants. Hence, this greater amount of
photosynthetic tissue is probably not as effective per unit as is the
smaller amount of the small plants. In fact the gross productivity/gm


PRODUCTION
116
TIME
AMBIENT TEMPERATURE <*C>


90
A major weakness in the use of a multivariate linear model in this
type of study is that it assumes independence between independent variables.
In actuality probably very few of the variables are completely independent.
For example sunlight and nutrient levels are both assumed to be linearly
related to the state of the variable for standing crop. It is further
assumed that each contribute independently and in an additive fashion.
This is not true, however, as variables such as nutrient loads and solar
radiation interact multiplicatively and the effects of one are limited by
the state of the other (see H. T. Odum 1974). This interaction may be
linear, exponential, or logistic depending upon whether or not either is
present in limiting quantities. The linear model does not account for these
complex relationships and should not be considered as a basis for making
generalizations about interrelationships in the system. I feel, however,
that this type of analysis can provide an indication of which variables
are important in accounting for variation but it cannot be interpreted as
a mathmematical expression of the functions of these variables (i.e., the
coefficients in the model have no real meaning).


187
Methods and Materials
The sides of four greenhouse tables 12 inches deep were constructed
from 1 inch by 12 inch redwood. The tables were lined with 6 mil
polyethylene sheeting and filled with water. The water on all tables
was fertilized equally, as needed, with 20:20:20: water soluble commercial
plant fertilizer with minor elements. Sequestered iron was also added
to obtain concentrations of 2.5 ppm. Square foot grids were constructed
on each table by stretching nylon twine across the top of the tables and
tying it off on nails spaced one foot apart. Two tables provided 18 one
sq. ft. quadrats each and two of the tables which were somewhat larger
provided 27 quadrats each. One small waterhyacinth plant in the inflated
petiole stage taken from Lake Alice was placed within each quadrat on all
tables and the plants were allowed to grow for several weeks until the
tables were completely covered. Five quadrats were then randomly selected
from each table and the plants were harvested to obtain base measurements
for the parameters to be evaluated. In the first experiment average
plant height, no. leaves/quadrat, no. of leaves/plant, and no. plants/
quadrat were counted and to obtain initial baseline estimates for the
table. The same variables were estimated in the second experiment in
addition to total dry weight subdivided into living and dead organic
material. The living material was further subdivided into leaves, rhizomes,
roots, and stolons. Where applicable these measurements were evaluated
in terms of both unit area and individual plant.
Arsama densa eggs were collected for the first experiment from
pickerel weed (Pontederia cordata) at a lake near Putnam Hall, Florida,
on 25 June 1974. These were allowed to hatch and neonates were placed
on the tables in sufficient numbers to achieve 0, 0.33, 0.67, and 1.00


248
occurred in the jars and approximately only 1500 larvae survived.
These were released 16 August on a small pond approximately 3.2 km
south of Paynes Prairie on 1-75, Alachua Co,, Florida. The pond had a
surface area of approximately 50 sq. m. and was covered with water-
hyacinths which were approximately a meter tall. From previous data
I estimated the density of the plants at this site and time of year to
be about 85 per sq. m. Our infestation level then was approximately
0.35 larvae per plant or 30 larvae per square meter.
A second pond 0.8 km South of the first pond was selected as a
control. This site was somewhat smaller than the first but the water-
hyacinths were very similar in both density and height. Both ponds were
formed at culverts under the interstate highway and both were formed
from the same watershed.
Since I hypothesized that the effects of the insect feeding activity
would be most evident after the first frost no sampling was done until
12 December. Because of the destructive nature of the sampling and
the small size of the ponds and the amount of time required to process
each sample the number of samples taken were necessarily small. Only
three samples (0.32 sq. m. ) were selected at each site and were along
an east-west transect the first being near the west bank, the second
in the center, and the third near the east bank.
The larvae present at this time represented the F-j generation of
those released. The height of each plant was measured as well as the
plant density, leaf density, and the number of leaves per plant. Insect
damage on each plant was measured in terms of both leaf and rhizome
damage. The larvae were counted, the instar noted and any parasites
present were recorded. One sample (0.32 sq. m;.) near the center of


ACKNOWLEDGEMENTS
I wish to thank the numerous individuals who have assisted in
these studies.
I would like to express my appreciation Dr. E. E. Grissell,
Dr. E. L. Todd, Dr. R. E. Woodruff, Dr. T.J. Walker, Dr. R. Carlson,
and Dr. C. W. Sabrosky for the identification of insect specimens;
Dr. G. E. Allen, L. P. Kish, and Dr. E. I. Hazard for diagnosing insect
diseases; C. Cagle, C. Siebenthaler, M. White, G. Presser, N. R. Spencer,
and D. Butler for field and technical help; the University of Florida
Soils Laboratory for analysing water samples; Dr. E. A. Farber for
providing solar radiation data; the U.S. Department of Agriculture and
the Florida Division of Plant Industries for providing space and facil
ities; Ann Owens and Susan Kynes for library and literature research
assistance; Cath Siebenthaler for typing and editorial assistance in the
original manuscript; N.R. Spencer and T. C. Carlysle for photographic
and dark room assistance; and my graduate committee, Dr. D. H. Habeck,
Dr. T. H. Walker, Dr. R. I. Sailer, Dr. G. E. Allen, and Dr. J. Reiskind
for critical reading of the manuscript.
I would especially like to thank Mr. Neal R. Spencer for providing
space and facilities and the U.S. Army Corps of Engineers for providing
funds.
I would also like to thank my wife, Debbie, whose patience and
endurance saw me through to the conclusion of this work.


Figure 24. Standing crop values, both estimated and real, from Lake Alice
The points represent estimates derived from the average values
for height from May 1974 through April 1975 using the equation
in Figure 23 to determine an estimated plant weight. These
were multiplied by plant density to obtain an estimate for
standing crop. The vertical bars represent actual measurements
taken from nine 0.25 m2 samples taken each month from April 1975
through February 1976. The dotted line represents predicted values
from multivariate regression analyses (see Table 7).


208
Biomass Estimates
The effect of varying insect concentrations on the total biomass
per unit area is shown in Figure 32. Total biomass, as defined here,
includes a living component (standing crop) and a non-living component
(detritus), Figures 33 and 34, i.e., the total organic material present.
Biomass estimates were only taken in the fall experiment, therefore
seasonal comparisons cannot be considered. The summer standing crop
(Fig. 33) could be estimated, however, using predicted values from the
regression equation in Figure 27 for average weights per plant multiplied
by the number of plants per unit area. The value for standing crop at
the 0.33 level of infestation was corrected for mite damage by interpo
lating the expected values for height and plant density from figures
la and Id assuming approximate linearity between the 0 and 0.67 levels
of infestation.
Total biomass (Figure 32) revealed a rather unexpected response to
insect concentration. At all four levels of infestation an increase was
evident. The control (0 infestation level) increased 188%. The plots
treated with insects, however, responded very similarly at all three
levels of infestation resulting in approximately a 160% increase over
the initial biomass present. The effects of insects on total biomass
was not significant.
The standing crop (grams living material per unit area) declined
significantly with increased initial insect concentrations in the fall
experiment. If it is assumed that the estimated regression for the
standing crop values in the summer experiment is reasonably correct a
quite different response is apparent. Instead of a rapid decline even
with low infestation levels no response is apparent until the infestation


Figure 41. The effects of varying insect feeding activity on the
waterhyacinth rhizome mass present per unit area. Data
only from the fall experiment. Legend as in Figure 28.
Regression: Y = 207.43 180.22X, r
-0.6946


69
found puparia of Masieira senilis associated with the burrows of Bellura
gortynoides Wlk. Both of these names are probably synonyms of Lydella
radiis (Town.) (Stone et at. 1965).
Comstock (1944) made note of the fact that he found no parasites
associated with Arzama gargantua Dyar in California.
In the Thompson catalogue (1944) two parasites are listed from
Arzama obliqua (Wlk.). The first is Ceromasia senilis Mg. which may be
a misidentification of Lydella radiis (Town.). The second is Pimpla
roborator F. (=Exeristes) which is an ichneumonid. I question the
veracity of this latter records, however, because the range is listed
throughout Europe, Japan, and Guam. As far as I have been able to as
certain the Arzama Bellura group is strictly New World.
Vogel and Oliver (1969b) listed several parasites and predators of
Arzama densa Wlk. They identified Lydella radiis (Town.) from the larvae,
Ichneumon n. sp. and Eupteromalus virideseens (Walsh) (Hymenoptera:
Pteromalidae) from the pupae, Telenomus arzamae Riley (Hymenoptera:
Scelionidae) and Anastatus sp. (Hymenoptera: Eupelmidae) from the eggs.
They also found Coleomegilla maeulata De Geer larvae (Coleptera:
Coccinel 1 idae) preying on the eggs and young larvae, and Phyllopalpus
pulehellus (Uhler) (Orthoptera: Gryllidae) and Chlaenius pusillus Say
(Coleptera: Carabidae) preying on the larvae.
Levine (1974) indicated that the eggs of first and second generations
of Bellura gortynoides Walk, are also parasitized by Telenomus arzamae
Riley and the second generation larval populations are parasitized by
an ichneumonid and have a polyhedrosis virus.


49
felt that the difference was due to longer heat retention in the deeper
water at night. Under 15.2 cm of water in a brown glass bottle only 28%
germination was observed.
Barton and Hotchkiss (1951) also studied the effects of temper
ature, light, and storage on seed germination. They concluded that a com
bination of high temperature and light is needed for germination of dor
mant seeds although temperatures as low as 5C did not impair germination
when in direct sunlight (greenhouse) and alternating temperatures (5-30C,
5-35C, and 5-40C) allowed some germination even in the dark. They also
found that a storage period of a month or longer hastened germination
especially with less mature seeds.
Francois (1964-3) obtained good rates of germination (over 95%)
by keeping his seeds in a 12:12 L:D photoperiod with a corresponding
40C:20C temperature regimen. Bock (1966) was convinced that seeds do
not germinate in California and found that they do not remain viable
there for longer than 2 months. Sculthorpe (1967) reflected the findings
of other authors by indicating that the seeds are able to tolerate a long
dry period and remain viable. Tag el Seed (1972) investigated seed germ
ination under a wide range of chemical treatments and under low oxygen
tension and low redox potential as well as many other environmental
conditions. His extensive studies indicate that germination is stimu
lated by low redox potential and low oxygen tension expecially after wet
storage, germination is most likely to occur in water warmed by intense
light, the addition of organic matter to the substrate stimulates germ
ination, the seeds will only germinate at the surface of the substrate,
and aeration has no significant effect on germination.


308
Review of Results and Suggestions for Further Studies
The productivity study discussed in the second section of this disser
tation showed that the net solar efficiency of waterhyacinth was similar
in both small and large plants (1.6%). Because of a high P:R ratio the
small plants grow faster (in terms of weight gain relative to standing
crop) than the large plants.
Three phases are apparent in the annual growth of waterhyacinth on
Lake Alice. A spring growth period is characterized by an initial increase
in plant and leaf density followed by an increase in plant height accomp
anied by a decline in plant density. This may be explained in part by
energy allocation under differing conditions of density. Early in the
season when the canopy is open and the plants are small, more energy is
allocated towards producing offsets than towards increasing individual
plant size. As space becomes more limiting more energy is put into
increasing the size of the individual plant, making it more able to com
pete for available light, and less into offset production. In a dense
stand the small offsets would probably have a small chance of surviving
in the low light conditions under the canopy. It would be maladaptive,
then, to produce them in this situation. As the plants increase further
in height the small plants die which accounts for the sharp drop in
absolute density.
A late summer and fall phase is defined by plant senescence and a
gradual decline in plant size. This is accompanied by an equally gradual
increase in plant density. An increase in damage by Avzama densa Wlk.
also occurred at this time but, because of multiple effects, the degree
to which A. densa contributed to this decline is not ascertainable.


11
increasing. The result of this is a geometrically increasing trend in
the overall cost of aquatic weed control by traditional means.
Because of the seriousness of waterhyacinth infestations the
beneficial aspects of this plant have been largely overlooked or ignored.
Fringes of aquatic plants along rivers or lakes are often helpful in
absorbing wind and wave action and preventing bank erosion (Tilghman
1963). Caldwell (1942) notes that the roots of waterhyacinth provide
excellent cover for goldfish spawn. He promotes the growing of this
plant for ornamental purposes stating that it is the "Biggest bargain
in a pool plant . and dismisses its detrimental attributes as an
" . attractive nuisance." Waterhyacinth was originally imported into
this country for use as an ornamental (Raynes 1964) and the beautiful
flower does give it a certain aesthetic appeal.
Tilghman (1962, 1963) spent many years fishing the St. Johns River
and guiding fishing tours. He was vehement about the beneficial effects
of waterhyacinth on fish propagation noting that the plant roots provide
cover for spawn and support macro-invertebrates which are preyed upon
by fish. Tilghman also noted that the plants helped clean the water
thus improving the fish habitat.
Abu-Gideiri and Yousif (1974) studied the influence of water
hyacinth on planktonic development in the White Nile. They compared
plankton populations and water chemistry parameters at a site south
of Jebel Aulia Dam to similar studies done prior to 1958 before in
vasion of the area by waterhyacinth occurred. They found that overall
planktonic densities had increased in the interim as a result of changes
in the water quality (such as an increase in phosphates). They attributed


AVERAGE WEIGHT PER PLANT [9m]
12 DEC. I974
CONTROL SITE
RELEASE SITE
in
no
CTi
no


Figure 11.
Diurnal curve for small waterhyacinth productivity. See
Figure 10 for explanation.


6
effect on the other, and each gene with two alleles (Bock 1966; Ornduff
1966).
Very few cytological studies have been done on waterhyacinth.
Banerjee (1974) found the chromosome number to be 2n = 32 in India,
illustrated kavyotypes, and described the chromosomes. She found the
chromosome number to be very consistent but noted variants of 2n = 30
and 58. Bock (1966) also reported the diploid chromosome number to
be 32 and noted that this had been reported by earlier authors as
the probable number.
Economic Importance
Waterhyacinth is ranked world-wide as among the top 10 most im
portant weeds and as the single most important aquatic weed (Holm 1969).
Because of its floating habit and high productivity (Bock 1969) it com
petes with man for open water. Large build-ups interfere with hydro
electric operations in many areas (Holm et al. 1969; Rushing 1974). Its
ability to interfere with navigation is well documented (Gay 1960; Evans
1963; Holm 1969; Webber 1897; Curtis 1900; Zeiger 1962). In the Panama
Canal mats of waterhyacinth have become so thick as to interfere with
the opening and closing of the locks (Pasco, pers. comm.). Gusio et al.
(1965) cited a study in which the efficiency of canals in the Everglades
were reduced 40-80% by large infestations of this plant. Irrigation
operations are affected by the impediment of water flow and the clogging
of pumps.
Waterhyacinths affect agriculture not only indirectly, as in irri
gation, but also directly. Sugar and rice are cultivated in "flood-fallow"


TABLE OF CONTENTS
Page
ACKNOWLEDGEMENTS
LIST OF TABLES VU
LIST OF ILLUSTRATIONS IX
ABSTRACT xiv
INTRODUCTION 1
LITERATURE REVIEW 3
Eiohhomia crassipes 3
Taxonomy 3
Description and Account of Variation 4
Economic Importance 6
Distribution 15
Habitat 23
Community Associations 31
Growth and Development 33
Morphology 33
Perennation 38
Physiological Data 38
Phenology 40
Reproduction 41
Floral Biology 41
Seed Production and Dispersal 45
Viability of Seeds and Germination 46
Vegetative Reproduction 50
Productivity and Standing Crop 50
iii


Page
Control 51
Arzama densa Wlk 53
Taxonomy 53
Hosts Plants 60
Biology and Life History of A. densa and Related
Species 63
Parasites, Predators, and Diseases 68
CHAPTER. 1. THE RELATIONSHIP BETWEEN THE PHENOLOGY AND BIOLOGY
OF WATERHYACINTHS AND VARIOUS PHYSICAL AND BIOLOGICAL FACTORS 70
Introduction 70
Methods and Materials 73
Diurnal Waterhyacinth Productivity 73
Annual Cycles and Insect Damage 76
Site Description 81
- Analyses 88
Results 91
Water Quality 9,1
Temperature and Solar Radiation 104
Waterhyacinth Productivity 114
Seasonal Variation in Photosynthetic Tissue . 125
Seasonal Variation in Plant Density 139
Seasonal Variation in Standing Crop 150
Damage by Arzama densa 157
Results of the Multivariate Analysis 161
Discussion 170
TV


2.5
2.0-
STANDING CROP
t
ISTIMATtO DATA
rtiOICTIO
>T oiv OISftVtO DATA
r ni i Art rj-fib.'7*
1.5
cn
-P


Table 15. Summary of developmental data for A. ier\si.
'
1
2
Average
for all
Sugarcane borer diet
Diet
Eiakkomia
crassipes
Fontedria cordata
larvae from 1 & 2.
(from Vogel S Oliver 1969)
Temperature (C)
25
25
25
21
Photoperiod (L:D)
14:
10
14:
10
14:
10
14:
10
Instar
# Observed
X (iS.D.) da
? Observed
X (:S.D.) da
# Observed
X (:S.D.) da
# Observed
X (S.0.) da
Egg
5 (1125)
5.6:0.7**
5 (1125)
5.6:0.7**
5 (1125)
5.6:0.7**
14
5.4:0.5
1
36
3.3*0.5
25
3.2:0.4
61
3.2-0.4
14
3.8:0.7
2
31
3.2*0.8
22
3.5:1.1
53
3.3:0.9
14
5.2:0.8
3
26
3.8*1.1
21
3.3:0.6
47
3.6:1.0
14
5.6:1.6
4
34
4.7*1.5
15
4.5:1.2
49
4.7:1.4
14
5.4:2.0
5
30
5.5*1.5
13
7.7:1.4
43
6.1:1.8
14
7.1:2.2
6
17
6.9*1.7
11
10.9:2.5
28
8.5:2.8
8
12.5:2.6
7
8
10.8:3.0
4
12.5:1.3
12
11.3:2.6
2
9.5:0.7
Pupa
0
N.O.*
3
10.0: 0
3
10.0: 0
14
14.6:1.8
Total
--
53.8*

61.2

56.3

69.1
* Assumes 10 da pupal stadium although none were observed in these experiements. Total egg to egg duration estimated from
the sum of the means for all stadia.
** Egg developmental data is the same for both host plants since the larvae used were from the same egg masses. The mean is
derived from the average of 5 estimates representing a total of 1125 eggs (see Table I).
279


Figure 49. The age structure of the Arzama densa population during the period of this study.


Figure 16
Annual change in leaf density as determined from weekly
samples taken in the study area. Each point represents a
mean based on the number of leaves in each of three 0.316
m2 samples. The dotted line represents predicted values
based on multivariate regression equations (see Table 7).


MEAN PLANT WEIGHT G.DW
VO
OJ


Figure 10. Diurnal curve for large waterhyacinth productivity determined
from CO2 gas exchange measured on Lake Alice with an infrared
C02 gas analyser. Respiration rates were determined from the
average night-time values. Gross production is defined as the
area under the curve above the respiration line. Net produc
tion is the area under the curve above the compensation point
(0 gC/m2/hr), The lower curves represent solar energy and
temperature.


57
densa Wlk., gortynoides Wlk., and probably anoa Dyar. He therefore
proposed the name Arzamopsis for those species with a smooth front and
suggested that A. diffusa be the type species and .4. melanopyga be
included in the genus. He also described Arzama matanzanensis, a new
species from Cuba.
Seitz (1923) again considered all of these species in the genus
Bellura Wlk. The species listed were B. obliqua (Wlk.), B. densa (Wlk.)
(= oeoogenes Dyar), B. gargantua (Dyar), B. anoa (Dyar), B. matanzanensis
(Dyar), B. pleostigma (Dyar), B. gortynoides Wlk. (= vulnifiea Grt.),
B. melanopyga (Grt.), and B. diffusa (Grt.). He noted that B. pallida
B. & Benj. and B. brehmei B. & McD. are probably races of B. obliqua
(Wlk.) but may be distinct species.
Comstock (1936) discusses this group of insects in his introductory
entomology text. He noted that the genus Bellura contained three North
American species, B. melanopyga, B. diffusa, and B. gortynoides. He
also recognized the genus Arzama and listed A. obliqua as "our most
common species". He included these species in the subfamily Apatelinae.
Jones (1951) listed the macrolepidoptera of British Columbia and
included Arzama obliqua (Wlk.) and Bellura gortynoides Wlk. He noted,
however, that the latter species is a doubtful record. He synonymized
Dyar's Aarzamopsis [sic] with Bellura and considered B. vulnifiea (Grt.)
a synonym of B. gortynoides Wlk.
Tietz (1952) listed Arzama obliqua (Wlk.) and A. densa Wlk. from
Pennsylvania. He considered pallida B. & Benj. a race of A. obliqua
(Wlk.), obliquata Grt. a synonym of B. obliqua (Wlk.), and oeoogenes
Dyar a synonym of A. densa Wlk.


21
seems to imply that it was discovered in 1949 at least in the Rotorua
District. Walker (1954) reported it from the opposite side of North
Island near Shannon. Matthews (1967) stated that there were 2 areas of
infestation in 1948-50, 15 after 1950, and 70 by 1956. Another report
(Anonymous 1964) indicated that there were at least 60 known infest
ations in New Zealand ranging from Opoua in the north to Shannon in
the south. Manson and Manson (1958) noted its occurrence as far north
as Kaitaica.
The spread of this plant has also taken in some of the Pacific
Islands. It was reported from Hawaii in 1946 (Bock 1966) and Mune and
Parham (1954) indicated that it was recognized as a pest in Fiji.
In Africa the plant is known from Kenya (Anonymous 1957), Zaire
(Anonymous 1957; Lebrun 1958; Kirkpatrick 1958; Coste 1958; Berg 1959;
Little 1965,1968a; Holm et al. 1969), Tanzania (Anonymous 1957; Little
1968); Uganda (Anonymous 1957), Angola (Lebrun 1958; Mendonca 1958),
French Equatorial Africa (Lebrun 1958), Rhodesia (Lebrun 1958; Little
1968; Holm et al. 1969) Malawi (Lebrun 1958); Monzambique (Lebrun 1958;
Mendonca 1958), South Africa (DuToit 1938; Penfound and Earle 1948;
Lebrun 1958; Holm et al. 1969), Madagascar (Lebrun 1958), Sudan (Gay
1958, 1960; Davies 1959; Pettet 1964; Little 1965, 1966,1968a; Chadwick
and Obeid 1966; Holm et al. 1969; Abu-Gideriri and Yousif 1974; Tag el
Seed and Obeid 1975; Mohamed and Bebawi 1975), Senegal (Anonymous 1964;
Little 1965; Holm et al. 1969) and Egypt (Little 1965; Holm et al.1969).
Waterhyacinth was first introduced into Africa either in South
Africa or Egypt. Sculthorpe (1967) cited a work on Egyptian flora which
indicated that it made its appearance in Egypt in the period between


Figure 27. Average dry weight per waterhyacinth plant as a log function
of the average height.


250
Results
Extensive damage by A. densa was evident within two weeks at the
release site. While the initial releases were made in one small area,
the larvae rapidly spread over the entire pond. Within 50 days abundant
egg masses were noted indicating the beginning of a second generation.
At the time of sampling (December 12) the Avzama population had
increased to 52 larvae per sq. m. (Table 12) as compared to 3 per sq. m.
(living larvae at the control site. Mose of these were 6th or 7th
instar and represented the final stages of the second generation. This
population was equivalent to 0.50 larvae/plant at the experimental site
and only 0.03 larvae/plant at the control site.
None of the insects in the unusually high population at the
release site were parasitized. Nearly 45% of the larvae in the relatively
low population at the control site were dead as a result of parasitism.
After 17 weeks of insect feeding the plants in the release site were
severely damaged. As expected, the occurrence of a light frost the
first week in December accentuated this damage. Many of the plants,
although severely damaged by the insects, appeared green and healthy prior
to this time. The freezing temperatures killed a large percentage of
these damaged leaves. At the control site only the tips of the leaves
suffered damage from this initial frost. Figures 43(a)-43(h) show the
release and the control sites in a sequence up to one year after the
initial release. Figure 43(c) shows the release site after the first
frost.


270
the release site by early summer but, this did not occur. Apparently
the other species had sufficient time to become established and prevent
the spread of the waterhyacinths into the center of the pond. All of the
plants mentioned earlier are rooted and may form a physical barrier to
the waterhyacinths. The long term success in controlling waterhyacinth
experienced at this site would probably not occur where the water was
too deep for rooted emergents to gain a firm foothold and occupy the
space available. In this situation the waterhyacinths would readily
float in from other areas and again become dominant. Nevertheless, I
feel that this study has shown not only that waterhyacinth is vulnerable
to biological control and that this control can be achieved, at least
temporarily, by the manipulation of populations of native insects. It has
also proven the overall effectiveness of Arzama densa as a control agent.
Harris (1973) has suggested that in some cases we may need an infestation
of one insect to reduce the infestation of a weed and a second one to
keep it low. Perhaps indigenous insect populations, such as Arzama densa,
can be used to initially reduce the weed infestation with subsequent
releases of exotic insects, such as Neoahetina spp., to exert a more
constant stress and maintain a low weed infestation.


Table 17. A summary of insects known to parsitize Arsama densa Wlk. (from Vogel
and Oliver 1969b in part).
Fami1y
Species
Host stage attacked
Locality
Seelionidae
Telenomus arsama Riley
eggs
Fla., La
Eupelmidae
Anastatus sp.
eggs
La.
Tachinidae
Lydella radiois (Townsend)
7th instar larvae
Fla., La
Icheneumonidae
Ca/npoletis sp. oxylus group
4th instar larvae
Fla.
Ichneumon n. sp.
pupae
La.
Chasmias scelestus Cr.
pupae
Fla.
Pteromalidae
Eupteromalus viridesccns (Walsh)
pupae
La.


333
BIOGRAPHICAL SKETCH
Ted Douglas Center was born 15 August 1947 in Dayton, Montgomery
Co., Ohio. He attended Belmont Elementary School and Belmont High School
where he graduated in 1965. Following high school he attended Ohio
University in Athens, Ohio for one year, after which he transferred to
Foothill College in Los Altos Hills, California. After a year in
California he transferred to Northern Arizona University in 1967 from which
he received his Bachelor of Science degree in Zoology in 1970 and his
Master of Science degree in Biology in 1972. He transferred his studies
to the University of Florida in September 1971 where he is currently
completing the requirements for a Ph.D in Entomology working on the
biological control of aquatic weeds.
Ted Center's work experience began at the age of 14 when he became
employed at the Dayton Museum of Natural History. He remained on the
staff working part-time during high school and full-time during the
summers from 1961 through 1969. His duties included regular television
appearances on a local children's show, instructing museum nature classes,
presenting lectures to various civic groups, care of the museum's live
animal collection, curating and preparing specimens for the museum's
collections, and participation in various research projects. While in
California he performed similar duties at the Palo Alto Children's
Museum.
While in Arizona he was employed part-time in the Biology Department
of N.A.U. curating the insect collection and preparing bird and mammal
specimens. He also worked part-time for the Geology Department of the
Museum of Northern Arizona. In the summer of 1971 he was employed by


PERCENTAGE OF INITIAL VALUE


260
TOTAL BIOMASS
220-
140-
100
f
0 0.2 0.4 0.6 0.8
INITIAL INSECT CONCENTRATION
1.0


4^
CT>


and S. gargantua from California. He also included S. pleostigma Dyar
and indicated that the description of this species was in a forthcoming
paper.
Barnes and McDunnough (1914) considered Sphida Grt. synonymous with
Arzama Wlk. thus making obliqua, densa, gargantua, and anoa all species
of Arzama. They synonymized S. oecogenes Dyar with A. densa Wlk. but
made no mention of S. pleostigma Dyar. They considered A. densa Wlk.
distinct from B. gortynoides Wlk. by virtue of a frontal protuberance.
Later they described another species of Arzama from New Jersey and named
it A. brehmei in honor of its discoverer (Barnes and McDunnough 1916).
Grossbeck (1917) in a list of the insects of Florida recognized
B. gortynoides Wlk., B. melanopyga Grt., S. obliqua Wlk., and S. anoa
Dyar. Barnes and McDunnough (1917), apparently having identified S. obliqua
for Grossbeck, noted that they made their determination before the pub
lication of Dyar's S. anoa and indicated that the specimens they identified
were probably S. anoa Dyar. This is confusing, however, because here
they are recognizing S. obliqua Wlk. which they had earlier combined with
Arzama.
Dyar (1922) re-evaluated the status of the genera Arzama and Bellura.
He noted that Hampson (1910) placed A. densa Wlk. as a synonym of B.
gortynoides on the assumption that both have a smooth clypeus. He also
noted that Barnes and McDunnough (1914) found that the type specimen of
A. densa Wlk. did have a tubercle on the clypeus and resurrected the genus
Arzama making Sphida a synonym of it but considered B. gortynoides Wlk.
distinct. Dyar examined several specimens identified as B. gortynoides
Wlk. and found that they all had tubercles on the clypeus and suspected
that Walker's types would also. He felt this would probably synonymize


45
Seed Production and Dispersal
The degree of seed set seems to be extremely variable. Agharkar
and Banerji (1930) found that 10 hours after anthesis through natural
pollination (autogamous or allogamous not distinguished) 35% of the
flowers were fertilized (30% with actively growing pollen tubes and
another 15% with pollen grains present). Through artificial pollination,
up to 71.3% of the fertilized flowers set fruit. McLean (1922) in Bengal
found that only 1% of the flowers set any seed. Haigh (1936) found in
Ceylon that 36 to 71% of the capsules produced may be empty. Backer (1951)
found no seed set in Java and Misra (1969) found up to 48% of the fruits
bear seed in India.
The conditions for seed set have been investigated but the results
are confusing. Parija (1934) indicated that temperatures between 24C
and 29C were necessary. Agharkar and Banerji (1930) felt that relative
humidities above 90% were required. Bock (1966) found that seed was set
when the relative humidity was never greater than 72%. Haigh (1936)
found the number of seeds per inflorescence to be 86, 28, and 91 when
the relative humidity was 90%, 70%, and 67% respectively. Tag el Seed
and Obeid (1975) concluded that seed set was favored if pollination
occurred immediately after the flowers opened. Thereafter, successful
pollination was hindered by high temperature and low humidity which
affected the stickiness and receptivity of the stigma.
Data on the quantity of seeds set per fruit or inflorescence also
indicate a great deal of variability. Haigh (1936) artificially pollinated
flowers and found an average of 24 seeds/capsule with a maximum of 72.
Bock (1966) reported the average in California to be 4.2 with a range of


166
Table 8. (Continued)
Potassium
Maqnesium
Alkalinity
pH
Sul fates
Lake
Level
Solar Radiation
-0.249
0.175
0.091
-0.279
0.380
0.591
(0.0101)
(0.0731)
(0.3549)
(0.0037)
(0.0001)
(0.0001)
Min. Air
-0.188
0.044
0.083
-0.160
0.186
0.404
(0.0537)
(0.6510)
(0.3979)
(0.1016)
(0.0564)
(0.0001)
Max. Air
-0.224
-0.043
0.127
-0.041
0.189
0.493
(0.0207)
(0.6594)
(0.1935)
(0.6730)
(0.0523)
(0.0001)
Min. H^O
-0.266
0.022
-0.042
-0.158
0.254
0.522
(0.0058)
(0.8197)
(0.6667)
(0.1060)
(0.0087)
(0.0001)
Max. H,0
-0.279
0.152
0.308
0.080
0.308
0.473
(0.0038)
(0.1198)
(0.0013)
(0.4157)
(0.0013)
(0.0001)
Rhizome Damage
-0.037
0.165
0.003
0.381
0.058
-0.376
(0.7060)
(0.0912)
(0.9747)
(0.0001)
(0.5538)
(0.0001)
Leaf Damage
-0.029
0.135
-0.045
0.363
0.063
-0.319
(0.7675)
(0.1675)
(0.6501)
(0.0001)
(0.5227)
(0.0009)
Phosphorus
0.221
0.019
-0.004
-0.291
-0.130
-0.382
(0.0227)
(0.8444)
(0.9674)
(0.0025)
(0.1841)
(0.0001)
Nitrogen
0.416)
-0.066
-0.133
0.133
-0.237
-0.613
(0.0001)
(0.5003)
(0.2506
(0.2502)
(0.0145)
(0.0001)
Iron
-0.247
0.122
0.094
-0.195
0.165
0.092
(0.0108)
(0.2536)
(0.3380)
(0.0454)
(0.0911)
(0.3504)
Conductivity
-0.415
0.394
0.602
-0.087
0.762
0.237
(0.0001)
(0.0001)
(0.0001)
(0.3758)
(0.0001)
(0.0144)
Potassium
1.000
-0.146
-0.157
-0.052
-0.422
-0.267
(0.0000)
(0.1343)
(0.1077)
(0.5939)
(0.0001)
(0.0057)
Magnesium
1.000
0.233
-0.077
0.522
-0.024

(0.000)
(0.0161)
(0.4333)
(0.0001)
(0.8066)
Alkalinity
1.000
0.214
0.466
0.089
--
--
(0.0000)
(0.0275)
(0.0001)
(0.3632)
pH
1.000
-0.094
-0.141
--
--

(0.0000)
(0.3360)
(0.1485)
Sulfates
1.000
0.189
--
--

(0.0000)
(0.0522)
Lake Level
_ _
_ _
_ _
_ _
_ _
1.000
--

--
--

(0.0000)


Figure 32. The effects of varying insect concentrations on the total
waterhyacinth biomass (expressed as both detritus and living
plant material). Data was taken only from the fall experi
ment. Legend as in Figure 28.
Regression: Y = 180.75 26.58X, r = -0.1828


Figure 9. Solar radiation data from the University of Florida campus
from May 1974 through April 1975. Each point represents a
daily average spanning a seven-day period. Solar radiation
begins to decline before the summer solstice as a result of
afternoon cloud cover associated with summer rains. The
unit of energy (Langleys) represents calories per square
centimeter per day. (Data from Dr. E. A. Farber.)


159


POPULATION ESTIMATES (LAMV.'M')
!
I\D
VO


CHAPTER IV
NOTES ON THE BIONOMICS AND POPULATION DYNAMICS OF ARZAMA DENSA WLK.
Introduction
Arzama densa Walker (Noctuidae: Amphipyrinae) is a large moth whose
larvae are semi-aquatic in habit. The taxonomy of its species group is
poorly understood and deserves further attention (see literature review
section). It is obviously closely allied to the species generally included
in the genus Bellura as both adult morphology and larval habits show
striking similarities. The separation of these species into two genera is
based largely on the armature of the frons and characteristics of the
antennae. Todd (pers. com.) feels that the validity of these characters is
questionable and two are probably congeners. He therefore proposes the com
bination Bellura densa (Wlk.) as proper for this species and further sug
gests that it may be conspecific with B. gortynoides Wlk. as proposed by
Smith (1893). Because the status of this group is questionable I have used
the name Arzama densa Wlk. throughout this dissertation. It must be pointed
out that this may not be a valid name and future taxonomic studies are
needed. While I agree that Bellura and Arzama are probably congeneric, I do
not feel that A. densa and B. gortynoides are conspecific. I have had the
opportunity to observe both species in the immature stages and from the
viewpoint of a nontaxonomist they certainly appear to be distinct. In order
to resolve the taxonomic relationships within this group larval characters,
oviposition behavior, host plant preferences, and other aspects of the bio
logy and immature stages should be considered in a bio-systematic approach.
This study was not designed to investigate the life history of Arzama
densa as it was originally assumed that this had been adequately investi
gated by Vogel and Oliver (1969). In the course of my experiments
271


123
LEAVES PETIOLES TOT. PHOT. NON-PHOT.


Figure 38. The effects of varying insect feeding activity on waterhyacinth
green mass (pseudo-laminae and petioles). Data only from the
fall experiment. Legend as in Figure 28.
Regression: Y = 173.27 128.00X, r = -0.6501


327
Ornes, W. H., and D. L. Sutton. 1975. Removal of phosphorus from static
sewage effluent by waterhyacinth. Hyacinth Control J. 13: 56-58
Parija, P. 1930. A preliminary note on the physiology of the seedlings
of the water-hyacinth (Eiohhomia speoiosa). Agr. J. India 25(5):
386-391.
Parija, P. 1934. Physiological investigations on waterhyacinth (Eiohhomia
cvassipes) in Orissa with notes on some other aquatic weeds. Indian
J. Agr. Sci. 4: 399-429.
Parsons, W. T. 1963. Water hyacinth, a pest of world waterways. J. Agr.
Victoria 61: 23-27.
Penfound, W. T. 1956. Primary production of vascular aquatic plants.
Limnol. Oceanogr. 1: 92-101.
Penfound, W. T., and T. T. Earle. 1948. The biology of the water hyacinth.
Ecol. Monogr. 18(4): 448-472.
Penfound, W. T., and E. S. Hathaway. 1938. Plant communities in the
marshlands of southeastern Louisiana. Ecol. Monogr. 8(1): 1-56.
Perkins, B. D. 1972. Potential for waterhyacinth management with
biological agents. Proc. Annu. Tall Timbers Conf. on Ecol. Anim.
Control by Habitat Manage., Feb. 24-25, 1972: 53-64.
Perkins, B. D. 1973. Preliminary studies of a strain of the waterhyacinth
mite from Argentina. C.I.B.C. Mise. Publ. 6: 179-184.
Perkins, B. D. 1974. Arthropods that stress waterhyacinth. PANS 20(3):
304-314.
Pettet, A. 1964. Seedlings of Eiohhomia orassipes: a possible complication
to control measures in the Sudan. Nature 201(4918): 516-517.
Pieterse, A. H. 1974. The water hyacinth. Trop. Abstr. 29(2): X263-X483.
Pirie, N. W. 1960. Water hyacinth: a curse or a crop? Nature 185(4706): 116.
Rao, P. S. J. 1920a. Note on the geotropic curvature of the inflorescene
in Eiohhomia speoiosa Kunth (water hyacinth). Indian Bot. Soc. J.
1: 217-218.
Rao, P. S. J. 1920b. The formation of leaf-bladders in Eiohhomia
speoiosa Kunth (water hyacinth). Indian Bot. Soc. J. 1: 219-225.
Raynes, J. J. 1964. Aquatic plant control. Hyacinth Control J. 3: 2-4.
Riley, C. V. 1883a. in Meeting of the Entomological Club of the American
Association for the Advancement of Science. Can. Entomol. 15(9): 174, 176.


80
block. The water temperature thermometer was placed vertically on the
bottom of the block. The air temperature thermometer was placed vertically
in the concave side. The block was mounted on a rider which slid over a
piece of pipe which extended into the lake bottom. This allowed the
thermometer to move up or down as the water level changed. The bulb of
the underwater thermometer was about 4 cm below the surface and measured
the conditions the submersed plant portions were subjected to. The air
thermometer bulb was about 30 cm above the water surface and measured
conditions under the leaf canopy. The concave side of the block was
oriented towards the north so as to avoid direct exposure to the sun.
The overhang on the block also helped prevent this.
Water level was measured at the northwest corner of the study area
from a depth gauge established there previously by other investigators.
Solar radiation data was obtained from Dr. E.A. Farber of the solar
energy laboratory at the University of Florida.


260
Total biomass (living and dead material) was only 11% less at the
release site than at the control site. This appears to indicate that net
primary production was not dramatically reduced as a result of insect
activity. The amount of this represented as detritus (dead material)
more than doubled at the release site. Hence, living plant material
decreased, dead plant material increased, but the total of the two showed
little change.
Insect infestation did not significantly reduce the plant density,
in fact, a small increase was evident. The yield in biomass per sq. m.,
however, decreased. This indicates that the weight per plant decreased
more than is apparent from the total standing crop. Figure 45 illustrates
the biomass per plant for the various plant parts. In all cases a greater
change is noted when parameterized in this manner. The change in standing
crop, then, is not the result of the insects killing a portion of the
plants and leaving a portion intact. This would result in a smaller
standing crop but the weight per plant change would be less than or equal
to the standing crop change. Rather, insect attack resulted in a popu
lation of smaller plants. These were probably offsets produced in
response to an increase in available space as leaves from neighboring
plants died thus reducing the amount of shading in the mat.
Since the degree of change as a result of insect attack varied with
the plant parts the plant proportions must have changed. Table 13 lists
the ratios of the various plant parts at both sites. In general, the
plant proportions at the release site were typical of small plants. The
ratio of leaf weight to plant weight was similar at both sites. The
ratio of petiole weight to plant weight was less at the release site
probably due to the reduction in intraspecific competition resulting


LIST OF TABLES
Table
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
Page
Standing crop and productivity of waterhyacinths as estimated
by various authors 52
Host plants of the Bellura-Arzama complex listed from various
literature sources.
Hydrological budget for Lake Alice (March September 1973)
61
85
A comparison of water quality measurements from Lake Alice,
Gainesville, Florida with previous reports 92
Metabolic and morphometric comparisons of the two morphological
types of waterhyacinth studied 120
Average daily rates of change in biomass from initial and
final monthly values 156
Summary of multivariate regression analyses for annual varia
tion plant characteristics 152
Correlation coefficients (r) between independent variables. .
164
Correlation coefficients (r) between dependent variables. .
.167
Correlation coefficients (r) between dependent and independent
variables and probabilities for a greater ¡r | 172
Ratios of the various plant parts and the percent change in
the final values as compared to the initial values 236
Comparison of the samples from the release site with the
control site based on various estimates of the plant and
insect populations
vii
251


194
Figure 28 represents the response curve for plant height to one
generation of feeding by Arsama densa with varying initial insect
concentrations in the summer and in the fall. At the time the plants
were infested the stands had approached a steady-state as reflected by
the small increase in the controls (11% and 5% respectively). As feeding
intensity increased a nearly linear decline in both experiments was
observed. The 0.33 level of infestation appeared to be unduly low in
the summer experiment but this was the consequence of a severe attack
by spider mites (Tetranyahus gloveri Banks) which occured on that table.
A somewhat greater response to herbivore activity was observed in the
fall than in the summer. This is as expected and appears to be due to
an interaction effect between decreased solar radiation in the fall and
herbivore activity.


INTRODUCTION
At the onset of this study various state and federal agencies were
preparing for the release of exotic insects for the biological control
of waterhyacinth (Eichhomia cvassipes (Mart.) Solms) in Florida and
the Southeastern U.S. In order to evaluate the effect of these insects
it was apparent that prior information on the ecology of waterhyacinth
at the release sites, particularly with regard to annual variability,
and the effects of indigenous insects would be needed. The purpose of
this dissertation was to provide part of this information for Lake
Alice, the primary study site in the Gainesville area.
To achieve this end the annual sequence of events in the water
hyacinth population was studied as well as the actual and potential
effects of natural and introduced populations of Arzama densa Wlk., a
native insect which feeds on waterhyacinth. This dissertation is or
ganized into five sections. The first section is a literature review
organized into two parts. The first part reviews the biology of water-
hyacinths and is organized in a manner similar to that suggest by
Cavers and Mulligan (1972). The second part reviews the taxonomy and
biology of A. densa Wlk. and related species.
The second section is a study of the waterhyacinth population on
Lake Alice. A fairly detailed description of the study site is provided.
The phenology of various morphometric features of the waterhyacinth
population is described with regard to the possible influence of various
physical and biological factors. A short study on the productivity of
waterhyacinth is also included in this section.
1


328
Riley, C. V. 1883b. in Meeting of the Entomological Club of the American
Association for the Advancement of Science. Entomol. Soc. Ont. Rep.
1883: 18, 21.
Riley, C. V. 1885. Notes: June 4, 1885. Proc. Entomol. Soc. Wash. 1: 30.
Robbins, W. W., T. E. Weier, and C. R. Stocking. 1964. Botany. An intro
duction of plant science. John Wiley & Sons, Inc., New York. 614pp.
Robertson, H. F., and B. A. Thein. 1932. The occurrence of water hyacinth
Eiohhomia arassipes Solms) seedlings under natural conditions in
Burma. Agr. and Live-stock in India 2(4): 383-390.
Robertson-Mi1ler, E. 1923. Observations on the Be llura. Ann. Entomol. Soc.
Amer. 16: 374-386.
Robson, T. 0. 1974. Mechanical control. Pages 72-84 in D. S. Mitchell,
ed. Aquatic vegetation and its use and control. UNESCO, Paris.
Rogers, H. H., Jr. 1971. Nutrient removal by waterhyacinth. M. S. Thesis,
Auburn University. 52 pp.
Rogers, H. H., Jr., and D. E. Davis. 1971. Nutrient absorption from sewage
effluent by aquatic weeds. Proc. 24th S. Weed Sci. Soc.: 352.
Rummel, C. 1919. Trapping for larvae of Apeen tesis. Bull. Brooklyn Entomol.
Soc. 14: 62-63.
Rushing, W. N. 1974. Waterhyacinth research in Puerto Rico. Hyacinth
Control J. 12: 48-52.
Ruttner, F. 1957. Fundamentals of limnology. University of Toronto
Press, Toronto. 242 pp.
Sabrosky, C. W. 1974. Eugaurax setigena, (Diptera: Chloropidae) a new
stem miner in water hyacinth. Fla. Entomol. 57(4): 347-348.
Sachs, J. 1860. Physiologische Untersuchungen ber die Abhngigkeit
der Keimung von der Temperatur. Jahrbiicher Wiss. Bot. 2: 338-377.
Sahai, R., and A. B. Sinha. 1970. Contribution to the ecology of Indian
aquatics. I. Seasonal changes in biomass of water hyacinth (Eiohhomia
arassipes (Mart.) Solms). Hydrobiologia 35(34): 376-382.
Salveson, R. E. 1971. Utilization of aquatic plants in steer diets:
voluntary intake and digestibility. M.S.A. Thesis, University of
Florida, Gainesville.
Sankaran, T., D. Srinath, and K. Krishna. 1966. Studies on Gesonula
punatifrons Stal (Orthoptera: Arcrididae: Crytacanthacridinae)
attacking water-hyacinth in India. Entomophaga 11(5): 433-440.


PERCENTAGE OF INITIAL VALUE


65
case. Moffatt (1888b) replied that there was no evidence that the
larvae had fed in the stumps and that all of the larvae and pupae they
collected were in similar situations, but admitted to not having looked
in the Typha reeds for want of a boat.
Kellicott (1889) referred back to the communications between Riley
and himself in 1883 and had decided that they were both right in that
A. obliquata G. & R. produces two broods in Washington and one in New York.
He also presented evidence confirming Moffatt1s contention they they
overwinter in stumps as larvae. Brehme (1889) later sent sections of
Typha stalk to Moffatt with numerous burrows and two pupae. Moffatt (1889)
subsequently reared a pair of the moths from this material. Brehme (1889)
felt that Kellicott and Moffatt were mistaken in their assertion that the
larvae overwinter in stumps because the specimens he sent to Moffatt were
collected in the winter below the water in cattail reeds and some were
even under ice. He also disagreed with Riley over the clustering mode of
oviposition. He noted that he had always found eggs laid singly and felt
that if it were otherwise it would be impossible for several larvae to
live in one reed. Johnston (1889) agreed with Brehme as he had found
abundant larvae and pupae in cattail in the winter in Ontario. He noted,
however, that he had also found them on shore in old wood. He proposed
that those on shore were merely wanderers. Beutenmilller (1889) described
the mature larvae of this species and indicated that he had found full
brown specimens under decaying stumps. He later (Beutenmller 1902)
described the larva again under the name of Bellura obliqua (Walk.).
Hampson (1910) repreated Comstock's (1881) description of the larva
of Avzama melanopyga Grt. under the name of Belluva melanopyga (Grt.).


Multivariate analyses failed to implicate A. densa as a factor in
accounting for seasonal variability in the plant characteristics. Climate
was considered the most important factor regulating variables estimating
standing crop. Water quality seemed to be more important in the variables
associated with density. Because changes in water quality (nutrient loads)
are as likely to be a result of changes in the plants as well as a cause
of those changes I am not satisfied that these models (plant and leaf
density) reflect dependent relationships even though statistically "good"
fits were obtained.
Intraspecific competition for light and space seems to be strongly
implicated in changes in plant density. A significant negative correlation
exists between plant density and plant height. Also, as the plant height
distribution becomes more strongly skewed towards large plants the number
of height classes important in the total distribution drops sharply.
Further, there appears to be an almost total loss of small plants during
the summer when plant height is maximum.
A third phase is the winter "no growth" or dormancy phase. During
the 2-3 mo. period little change occurred in most of the characteristics
observed.
Greenhouse experiments in which the levels of infestation by A. densa
were controlled more effectively brought out the relationship between the
various plant characteristics and feeding by this insect. Repetitions of
the experiment in the summer and the fall produced quite different results.
In general, the plants were much more sensitive to attack by insects in
the fall, and less so in the summer. Height declined in both experiments
but the slope of the decline was greater in the fall. The changes in the


Figure 31. The effects of varying levels of insect feeding activity on
the number of waterhyacinth plants per unit area expressed
as a percentage of predetermined means. Legend as in
Figure 28.
Summer: Y = 122.77 + 81.07X, r = 0.4209
Fall: Y = 123.95 35.53X, r = -0.2758


332
Vogel, E., and A. D. Oliver, Jr. 1969b. Life history and some factors
affecting the population of Arzama densa in Louisiana. Ann. Entolmol.
Soc. Amer. 62(4): 749-752.
Wahlquist, H. 1969a. Effects of waterhyacinth on fish production.
F.A.O. Fish Cult. Bull 2(1): 5.
Wahlquist, H. 1969b. Effect of water hyacinths and fertilization on
fish-food organisms and production of bluegill and redear sunfish
in experimental ponds. Proc. 23rd Annu. Conf. SE. Assoc. Game and
Fish Comm.: 373-384.
Walker, C. 1954. Water hyacinth is a dangerous weed. New Zeal. J. Agr.
89: 605-606.
Walker, F. 1864. List of the specimens of lepidopterousinsects in the
collection of the British Museum. Part 32. Suppl. part 2. Catalogue
of the Lepidoptera Heterocera. Seventh Ser. 35 Parts., London.
Wallwork, J. A. 1965. A leaf-boring galumnoid mite (Acari: Cryptostigmata)
from Uruguay. Acarologia 7(4): 758-764.
Warner, R. E. 1970. Neochetina eiohhormiae, a new species of weevil from
waterhyacinth, and biological notes on it and N. bruohi. Proc. Entomol.
Soc. Washington 72(4): 487-496.
Webber, H. J. 1897. The water hyacinth and its relation to navigation in
Florida. U.S.D.A. Bull. 18: 1-20.
Welch, P. S. 1914. Habits of the larva of Bellura melanopyga Grote
(Lepidoptera). Biol. Bull. 27: 97-114.
Wellington, W. G. 1960. Qualitative changes in natural populations
during changes in abundance. Can. J. Zool. 38: 289-314.
Westlake, D. F. 1963. Comparisons of plant productivity. Biol. Rev.
38: 385-425.
Worthington, C. E. 1878. Arzama obliquata. in Miscellaneous Memoranda.
Can. Entomol. 10: 15-16.
Wunderlich, W. E. 1964. Water hyacinth control in Louisiana. Hyacinth
Control J. 3: 4-7.
Yount, J. L. 1964. Aquatic nutrient reduction-potential and possible
methods. Rep. 35th Annu. Meeting, Florida Anti-Mosquite Assoc.: 83-85.
Yount, J. L., and R. A. Crossman, Jr. 1970. Eutrophication control by
plant harvesting. J. Water Poll. Control Fed. 42(5) Part 2: R173-183.
Zeiger, C. F. 1962. Hyacinth-obstruction to navigation. Hyacinth Control
J. 1: 16-17.
Zettler, F. W., and T. E. Freeman. 1972. Plant pathogens as biocontrols
of aquatic weeds. Annu. Rev. Phytopathol. 10: 455-470.


Figure 45. The mass represented by the various plant parts for an
average waterhyacinth plant at both the control and release
sites.


53
Arzama densa Wlk.
Taxonomy
Walker (1864) described three genera and three species of moths in
two families which are now known to be closely related. These were
Edema obliqua (Notodontidae), Bellura gortynoides (Notodontidae), and
Arzama densa (Gortynidae). His description of the latter genus and
species follows:
Genus Arzama
Male. Body stout. Head with thick-set porrect hairs. Proboscis
short, slender. Palpi stout, porrect, pilose, not extending
beyond the hairs of the head; third joint extremely small,
not more than one-tenth the length of the second. Antennae
moderately pectinated, rather short. Abdomen extending much
beyond the hind wings, tapering towards the tip, which has a
very small tuft. Legs stout, rather short; hind tibiae with
a short fringe; spurs long, stout. Wings rather short and
narrow. Fore wings acute; exterior border almost straight,
hardly oblique; second inferior vein almost as near to the
third as to the first; fourth not very remote from the third.
Arzama densa
Male. Reddish. Underside, abdomen and hind wings reddish cinereous.
Fore wings with an oblique very broad brownish band, which
contains the orbicular and reniform marks; the latter are red,
oblique, and narrow; a submarginal brown-bordered slightly
dentate band, which is rather brighter than the ground hue.
Hind wings beneath with a round brown spot in the disk, and


18
said to a Sun reporter, "and he thought that he did the State a favor.
I have it from his own lips, and I've known him since long before that
time, for I used to carry him up the river in a launch year after year
to his orange grove. He was Mr. Fuller, father of W.F. Fuller of
Brooklyn, owner of Edgewater Grove, a property which he bought and
improved, until now it is a beautiful place, seven miles above Palatka.
Five years ago there wasn't a water hyacinth in the St. John's River,
nor in the state, so far as I know. One season Mr. Fuller brought some
there and put them in a pond on his premises. I understand that he brought
them from Europe. They added very much to the beauty of the place, and
they thrived so that he took some and threw them into the river. There
they grew and blossomed abundantly, and they were greatly admired, and
Mr. Fuller said to me one day: "The people of Florida ought to thank me
for putting these plants here."
"But presently those in his pond had spread so that they covered
it over. Then he cleared them all out. But it was too late to stop them
from spreading all over the river. They worked their way and were blown
up and down it for miles, and into the bayous, and finally up the
Acklawaha [sic]. Two years ago they had become a serious menace to navi
gation, and protest after protest was sent to the Government. At last
the War Department sent an agent to investigate, but he got to us just
after the visitation of that heavy frost of two years ago, which killed
all our orange trees. The hyacinths were killed too, apparently, and so
the agent reported that nature had cleared the rivers and that there was
nothing requiring the department's attention. But the plants were only


Page
Figure
34.The effects of varying insect feeding activity on the amount
of dead waterhyacinth plant material (detritus) per unit
area 214
35. Detritus as a percentage of total waterhyacinth biomass as
a function of insect feeding activity 218
36. Net waterhyacinth production as a function of insect feeding
activity 221
37.. The ratio of conversion of living waterhyacinth plant material
into detritus as a function of insect feeding activity. . .223
38. The effects of varying insect feeding activity on waterhyacinth
green mass (pseudolaminae and petioles) 227
39. The effects of varying insect feeding activity on waterhyacinth
non-green mass (roots, rhizomes, and stolons) 229
40. The effects of varying insect feeding activity on waterhyacinth
root mass per unit area 231
41. The effects of varying insect feeding activity on the water
hyacinth rhizome mass present per unit area 233
42. The effects of varying insect feeding activity on the water
hyacinth mass represented as stolons per unit area 235
43. A photographic comparison of the waterhyacinth stands at
experimental and control sites at different times of the
year following the release of Avzama densa at the
former 253
44. A comparison of the standing crop of waterhyacinths at the
control site and the release site 259
xii


CO


Figure 52 illustrates estimates of Lydella radiis and Campoletis sp.
populations during the study period. The parasite numbers are based
on the number of pupae or puparia and the pupal exuvia found associated
with dead A. densa larvae. A single Campoletis pupa is usually found
in association with a 4th instar larva although on at least one occasion
two pupae were found in association with a single larva. Lydella, on
the other hand produces 1-5 puparia per 7th instar larva. The normal
range is more on the order of 2-3. This accounts for the more prominent
peaks observed in the parasite populations for Lydella. Although
parasite populations of both species appear to be more intense in the fall
and early winter, the effects of this intensity may be tempered by the
relatively high asynchronous host population present at this time. While
a large number of larvae are being removed by parasites they are replaced
by younger larvae. Initially the parasites cannot respond numerically
fast enough to take advantage of the newly recruited larvae.
Since the parasite populations are dependent upon thelevel of the
host populations they remain at a fairly low level. As the parasite
populations gradually increase a subsequent decline in the larval popu
lation begins to take place. As this continues the number of parasites
present relative to the number of host larvae present increases. As a
result the parasite populations are able to more fully exploit the avail
able host populations in the spring. Ultimately the recuitment of new
individuals is insufficient to maintain the host population and a
dramatic decline occurs. The low host populations in the summer results
in a decline in the parasite populations. This permits the subsequent
buildup of the larval population in the fall. It is not apparent whether
the source of this fall population is the low population present earlier


Table 9. Correlation coefficients (r) between dependent variables. Values
in parentheses represent the probability of a greater |r| under
the null hypothesis.
Standing
Crop
Plant
Height
Leaf
Area
Index
Leaves
Per
Plant
Leaf
Density
Plant
Density
Standing Crop
1.000
_ _
_ _
(0.0000)
--
--
--
--
--
Plant Height
0.846
1.000

--
--
--
(0.0001)
(0.0000)
--
--
--
--
Leaf Area Index
0.820
0.664
1.000
--

(0.0001)
(0.0001)
(0.0000)

--
Leaves Per Plant
0.599
0.790
0.630
1.000
(0.0001)
(0.0001)
(0.0001)
(0.0000)
--
--
Leaf Density
-0.027
-0.436
0.181
-0.143
1.000
(0.8447)
(0.0015)
(0.1922)
(0.3082)
(0.0000)
--
Plant Density
-0.325
-0.760
-0.187
-0.651
0.806
1.000
(0.0166)
(0.0001)
(0.1756)
(0.0001)
(0.0001)
(0.0000)


315
Azam, M. A. 1941. Utilisation of water hyacinth in the manufacture of
paper and pressed boards. Sci. Cult. 6(11): 656-661.
Backer, C. A. 1951. Pontedariaceae. Pages 255-261 in C. 6. G. J. Van
Steenia, ed. Flora Malesiana, vol. 4.
Bagnall, L. 0., J.A. Baldwin, and J. F. Hentges, Jr. 1974. Processing
and storage of waterhyacinth silage. Hyacinth Control J. 12: 73-79.
Bagnall, L. 0., T. de S. Furman, J. F. Hentges, Jr., W. J. Nolan, and
R. L. Shirley. 1974. Feed and fiber from effluent-grown water hyacinth.
Waste Water Use in the Production of Food and Fiber Proc. EPA-660/2-74-041
Bagnall, L. 0., J. F. Hentges, Jr., and R. L. Shirley. 1973. Processing,
chemical composition and nutritive value of aquatic weeds as affecting
the economics of their removal and utilization. Inst. Food Agr.
Sci., University of Florida, Gainesville.
Baldwin, J. A. 1973. Utilization of ensiled water hyacinths in ruminant
diets. M.S.A. Thesis. University of Florida, Gainesville.
Baldwin, J. A., J. F. Hentges, Jr., L. 0. Bagnall, and R. L. Shirley.
1975. Comparison of pangolagrass and water hyacinth silages as
diets for sheep. J. Anim. Sci. 40(5): 968-971.
Bancroft, K. 1913. The water-hyacinth. Agr. Bull. F. M. S. 1: 228.
Banerjee, M. 1974. Cytological studies on some members of the family
Ponterderiaceae. Cytol. 39: 483-491.
Banerji, I., and H. C. Gangulee. 1937. Spermatogenesis in Eiahhomia
orassipes Solms. Indian Bot. Soc. J. 16: 289-295.
Barber, M. A., and T. B. Hayne. 1925. Water hyacinth and the breeding
of Anopheles. Public Health Rep. 40(47): 2557-2562.
Barman, E. H., Jr. 1974. Gross growth, net growth, and assimilation
efficiencies of Diacrisia virginioa (Artiidae: Lepidoptera) fed
leaves of water-hyacinth. Assoc. SE. Biol. Bull. 21(2): 37.
Barnes, W., and F. H. Benjamin. 1924. Notes and new species. Contrib.
Natur. Hist. N. Amer. Lepidoptera. 5(3): 169.
Barnes, W., and J. H. McDunnough. 1914. Synonymic notes on North American
Lepidoptera. Contrib. Natur. Hist. N. Amer. Lepidoptera. 2(5): 200-201.
Barnes, W., and J. H. McDunnough. 1916. Synonymic notes on North American
Heterocera. Contrib. Natur. Hist. N. Amer. Lepidoptera. 3(3): 166-167.
Barnes, W., and J. H. McDunnough. 1917. Remarks on Grossbeck's list of
Florida Lepidoptera. Contrib. Natur. Hist. N. Amer. Lepidoptera.
3(4): 217.
Barton, L. V., and J. E. Hotchkiss. 1951. Germination of seeds of Eiahhornia
orassipes Solms. Contrib. Boyce Thompson Inst. 16: 215-220.


58
Forbes (1954) considered all species of Belluva, Avzama, and Sphida
to be in the single genus Avzama. He divided the genus into two groups
based on whether the front was flat or had a strong central bulge. In
the first group he included govtynoides Wlk., diffusa Grt., and vulnifica
Grt. and placed melanopyga Grt. as a synonym of vulnifica. In the second
group he included obliqua Wlk., bvehmei B. & McD., and densa Wlk.
Kimball (1965) listed the Lepidoptera of Florida and again recognized
both Avzama and Belluve. He included A. obliqua (Wlk.), A. [bvehmei B. &
McD.], A. densa Wlk., A. anoa (Dyar), B. govtynoides Wlk., and B.
melanopyga (Grt.). He noted that the two species of Belluva were
probably conspecific. He also referred to the specimen listed as B.
govtynoides Wlk. by Grossbeck (1917) and noted that it was actually
A. densa Wlk. making the former a synonym of the latter.
The only species listed by Tietz (1972) was Avzama gavgantua (Dyar).
Levine (1974) noted that B. vulnifica and B. govtynoides are separated
largely by the color of their anal tuft, the former being brown and the
latter white. He found that dark brown-tailed females (B. fulnifica Grt.)
may produce white-tailed daughters (b. govtynoides Wlk.). This indicates
that B. vulnifica is merely a form of B. govtynoides Wlk.
I received a personal correspondence from Dr. E. L. Todd from the
Systematic Entomology Laboratory of the U. S. Department of Agriculture
in April 1974. He explained the taxonomic situation with regard to these
species as follows:
I consider that Belluva Walker 1864, type-species
B. govtynoides Walker by monotypy is the valid generic name.
Avzama Walker 1864, type-species A. densa Walker by monotypy
and Sphida Grote 1878 [1879], type-species Avzama obliquata


39
Many authors have studied transpiration rates and found variations
due to such factors as solar energy, wind speed, temperature, and technique
through a waterhyacinth mat was as high
Misra (1969) found that water
as 65.5 kg/m /da. This represented a water requirement of 6.74 kg of
water-per gram (dry wgt.) of biomass produced. The ratio of evapotrans-
piration to open water evaporation (Ej:Eq) ranged from 5.92 to 9.84.
Knipling et al. (1970) measured the moisture content of a stream
of air before and after it had passed over a waterhyacinth leaf. They
found during the day the transpiration rate increased from 1520 to 2450
mg/dm2 (leaf surface)/hr. in response to increasing light intensities.
Dark transpiration values were also high, however, averaging 1430 mg/dm2/hr
In another experiment plants were grown in beakers in a variety of phos
phorus concentrations and measured for daily water loss. There was no
significant difference in evapotranspiration between the phosphorus con
centrations. The Ej'Eq ratio, however, was 305 g/da:100 g/da or 3:1.
Dry matter production was 0.27 g/da indicating a water use efficiency
ratio of 1129 gm H20/g plant dry wgt.
Other average values reported for the E :E ratio have been 3.2
(Penfound and Earle 1948), 7.8 (in India; Holm, et al. 1969), 3.7 (Timmer
and Weldon 1967), 1.02-1.36 (Brezny et al. 1973), and 1.46 (Van der Weert
and Kammerling 1974). The latter authors have found that 97% of evaporation
from waterhyacinth covered situations is the result of the process of
evapotranspiration.
Knipling et al. (1970) have also provided data on respiration and
photosynthesis by measuring the C02 concentration in an air stream passed
over a leaf. Net photosynthesis increased from 7.8 to 16.1 mg C02/dm2/hr


29
shade allowed them to survive longer. Misra (1969) found that when the
rhizomes are air dried they progressively lose the ability to resprout
as the moisture content decreases. They can tolerate a lower moisture
content when dried in mud, however, than when dried in air. This may
enable them to survive droughts in some instance.
As far as I have been able to determine Bock (1966) is the only
one to have investigated the effects of humidity on the growth of
waterhyacinth. She grew plants in a growth chamber both inside a plastic
enclosure with high humidity and outside the enclosure. She concluded
that high humidity favors growth.
With the recent interest in the utilization of waterhyacinth for
nutrient removal in sewage effluent increasing attention has been directed
towards the nutrient requirements of this plant. Dymond (1948) found that
the plants grew in both nutrient-rich and nutrient-poor water and that
the nutrient content of the plant was higher in nutrient-rich water.
Hitchcock et al. (1949) concluded that waterhyacinths have relatively
low nutrient requirements as good growth occurred in solutions 0.01 to
0.001 times as strong as normal in water cultures. They also found that
the growth response increased with added nutrients.
In Africa (Anonymous 1957) it has been noted that the lower limit
of "mineralization" is very low but little is known of the upper limits.
Chadwick and Obeid (1966) found that an increase in nitrogen levels
caused a linear increase in the total yield and plant number but had
little effect on the mean weight per plant. Knipling et al. (1970)
studied two sites with notably different levels of orthophosphate and
were surprised to find that the average standing crop yields were


41
maximum was reached in August and the period of maximum growth was
between May and August. They also followed the flowering cycle over a
period of years in Louisiana. In the years 1945-1947 anthesis began in
April. They felt that a definite flowering rhythym occurs in a given
colony of plants. Anthesis is maximum in June and declines through
September although this may vary from colony to colony and a second
period of flowering occurs in September and October and continues through
November into December. Buckman and Co. (1930) reported that the plant
is supposed to bloom every two to three months. In India flowering occurs
throughout the year but is most abundant in the post-monsoon months
(Sahai and Sinha 1970). A pre-monsoon flowering period (April and June)
has been reported in India (Pieterse 1974). Sahai and Sinha (1970)
further found that biomass accumulation was highest in January and
February, and the area occupied [% cover) was greatest in February and
March in India.
Reproduction
Floral Biology
A single flowering spike contains a variable number of flowers.
Bock (1966) found the average to range between 5 and 10 flowers per
inflorescence although she indicated that other authors have observed up
to 35 flowers per inflorescence. Small (1936) indicated that flowering
occurred on a daily cycle appearing as buds up to 7:30 AM and opening by
8:00 AM. The mode of pollination may be allogamous or autogamous. Bock
(1966) found that allogamous pollination may occur through the actions
of several insect pollinators. She listed Apis mellifera, Haliotus (1 sp.),
and Lasioglossum (2 sp.) as known pollinators and syrphid flies as possible


RHIZOMES
*\*
\
0.2 0.4 0.6 0.8
INITIAL INSECT CONCENTRATION
0
1.0


13
with other fodders in nitrogen content (0.97 to 2.57% D.W.), rich in
chlorine (3-4% D.W.), and richer than Napier and Guinea grass in lime
(3.5% D.W.) and magnesia (0.96% D.W.). They also noted that i.ts phosphate
content was low (0.36% D.W.) but that the digestible nutrients compared
well with other fodders. Taylor and Robbins (1968) analyzed the com
position of waterhyacinth and found the leaves to contain 15.8% dry
matter which in turn was composed of 14.7% ash, 1.7% nitrogen, 10.7%
crude protein and 17% crude fiber. The whole plants were 8.9% dry
matter, 1.5% nitrogen and 9.6% crude protein. They also analyzed the
plants for the amino acid composition. They concluded that the lysine
content of waterhyacinth was sufficient to serve as an effective grain
protein supplement.
Boyd (1968a) determined that waterhyacinth contained 12-18%
(D.W.) crude protein. He subsequently fully analyzed the nutritive
value of waterhyacinth and found the dry weight to be 5.9%, the crude
protein to be 0.94% of the fresh weight (ca. 16% D.W.), cellulose ca.
28% (D.W.) total available carbohydrate 7.8%, ash 17%, and caloric
t
content ca. 3.8 kcal/g. He further analyzed the inorganic nutrient
content and the amino acid composition. Taylor et al. (1971) extracted
protein from waterhyacinth and found that the percentage on a dry
weight basis varied between 7.4 to 18.1%. They also analyzed the
protein for the amino acid composition.
Knipling et al. (1970) compared the nutrient content of water
hyacinth from two different sites. They performed comparative analyses
of various plant parts from the two sites for nitrogen, phosphorus,
calcium, potassium, and magnesium as well as chlorophyll and water


206
would allow the equilibrium density to again be achieved and the available
space occupied according to the carrying capacity of the habitat.
If these assumptions are true, then one would expect the seasonal
consequences of insect attack to vary. The carrying capacity of the
habitat depends to a large extent on solar flux. In the summer when
solar energy is nearly maximum one would expect insect infestation to
ultimately result in an increase in the absolute density of the plant
since energy is high and space is available. In the late fall, when
solar energy is waning, these consequences may be much different.
Because of the reduced energy available the carrying capacity is reduced.
An increase in available space at this time may ultimately result in
little change or a decrease in plant density as other factors (i.e. solar
energy) become limiting. The data from this experiment (Figure 31) tends
to support this hypothesis. If insect damage to the apical bud was solely
responsible for these increases in density without the influence of other
factors, then it would be reasonable to expect the plants to react simi
larly in both experiments. Instead I observed an increase in density in
the summer experiment and a slight decrease in the fall experiment which
favors the seasonal interaction explanation for density changes.
I do not intend to dispute the fact that damage to the terminal bud
may cause a response in the direction of lateral growth. One could cite
many examples of similar phenomena in many different plants. From field
observations, however, it is worth noting that the offshoots produced
from plants with severe damage from A. densa do not appear to be as
vigorous as normal offshoots. These offshoots are often deformed and the
leaves often appear to be rather thick and leathery. I merely intend to
demonstrate that the evolutionary strategy of the plant is to produce


Figure 17. The average area of the pseudolaminae of waterhyacinth
leaves. Each point represents the mean leaf size derived
from measurements of all the leaves of 10 waterhyacinth
plants selected randomly at weekly intervals. Only one
side of the pseudolamina was measured.


241
Discussion
It is evident from these two experiments that a seasonal effect
interacts with insect activity and produces varying plant responses
depending upon the time the infestation occurs. The assumption that
insects directly cause an increase in plant density appears to be over
simplified. The canopy is reduced as is evidenced by reductions in
both height and photosynthetic material. This allows an increased pene
tration of available light and may accelerate the growth of the remaining
plants. This may be due to a cessation of rather than a direct result
of stress. Any stress factor that would reduce the effect of intraspecific
shading by reducing the canopy would probably result in a short term
increase in the number of plants present until a new equilibrium density
is achieved. Competitors may become increasingly important and suppress
this density increase by interspecific competition in the field. In
these greenhouse experiments competing species were not present so this
response may be exagerated.
The effect of insect damage to the terminal bud could possibly
contribute to an increase in the number of plants. I feel that the
contribution by insects to this process is minimal. Penfound and
Earl (1948) found that decapitated rhizomes failed to produce new
sprouts only when 4 cm of the rhizome tip was removed. The rhizomes
of the majority of the plants were thoroughly fragmented as a result
of attack by the larger larvae and the plants were dead. Fewer than 10%
of the surviving plants had rhizome damage. Those with rhizome damage
produced rather sickly offshoots. The offsets remaining after insect
attack were not the type that usually occur in open stands. The petioles
were not bulbous so as to function as floats so their ability to disperse




283
values may be underestimated. The addition of 6 da (egg developmental
period) may be added to these figures to bracket the estimate. Hence
pupation occurs 42-48 da post-poviposition, adult emergence 50-56 da
(males 48-54; females 52-58), and oviposition 52-58 da. This conforms
to the hypothesized developmental schedule given in Table 15 developed
from direct laboratory observations. By summing the developmental times
for each instar in Table 15 pupation is expected at 46 da post-oviposition
and adult emergence at 56 da. This is considerably less than Vogel
and Oliver's (1969) estimates but they reared their larvae on an arti
ficial medium and at a lower temperature. I suspect that these two
factors account for the difference.
In attempting to rear large numbers of A. densa larvae for other
experiments I have used this data to estimate the time at which I should
harvest the plants to obtain primarily pupae. Eggs were collected in the
field and neonates reared in the laboratory. The first instar larvae
were placed on waterhyacinth plants in ten large troughs in an airhouse.
The plants were harvested from each trough between 43 and 48 da after
the collection of the eggs. In all cases 50% or more of the insects
recovered were pupae. Hence, I feel that the developmental data pre
sented here is reasonably accurate.
The data for head capsule measurements and the larval age at
each molt is summarized in Figure 47. The cross bars represent the
standard deviation for each parameter and the point of intersection
represents the means. The figure to the right and the figure below
represent the number of observations in the mean for the head capsule
measurement and the age respectively. The figure above represents the
instar these figures are derived from. I have found the head capsule


246
strategy to break down in the second generation and abnormally high
parasite populations to ultimately cause a decline in the A. densa
population.
The second strategy employed was that of an inundative release.
Assuming that we could obtain a reasonable survival rate in the first
generation a sufficient number of larvae had to be released to severely
damage the plant population before parasite buildups caused a decline
in the,A. densa population. The minimum insect concentration to achieve
this was determined to be 0.3 larvae per plant based on field observation
of natural populations of A. densa as well as greenhouse studies using
various insect concentrations.


Figure 43. A photographic comparison of the waterhyacinth stands at
experimental and control sites at different times of the year following
the release of Arzama densa at the former.
a) Experimental site shortly after the release of Arzama densa
(August 1974).
b) Release site in October.
c) Release site after two generations of insect damage and after
the first winter freezes (January). The predominant plant is
Hydrocotyle. A patch of dead waterhyacinth is noticeable to
the right.
d) The release site in the spring (March 1975). Most of the water
is covered with Hydrocotyle.
e) The release site in July 1975 as the Hydrocotyle stand begins
to open up.
f) The release site one year after the initial release (August
1975). The site is dominated by cattail (Typha sp.) Notice
the small stand of waterhyacinth in the background.
g) The control site in the spring (March 1975). Compare this
with Figure 43d.
h) The control site one year after the initiation of this study.
Compare this with Figure 43f.


Figure 29. The effects of varying levels of insect feeding activity
on the average number of leaves per waterhyacinth plant
expressed as a percentage of predetermined means. Legend
as in Figure 28.
Summer: Y = 102.16 40.83X, r = -0.76
Fall: Y = 112.22 53.60X, r = -0.69


73
Methods And Materials
Duirnal Waterhyacinth Productivity
Waterhyacinths of two distinct morphological types from the "open"
side of the catwalk on Lake Alice were selected for in situ metabolism
studies. Large plants, approximately 90 cm tall, with elongate petioles
were measured for C02 uptake on 11-12 August 1973. Small plants (<50 cm)
with bulbous petioles were measured on 12-13 August 1973. A section of
the mat approximately 0.5 m2 of each type was placed under a chamber
constructed of a PVC pipe frame covered with clear poly-acetate. The
base of the chamber was 71 cm x 71 cm (ca 0.5 m2).
Air was passed through the chamber with a blower and duct system.
The duct entered the chamber at the base on one side. Air was supplied
to the blower intake through a tube opening approximately 3 m above the
water surface so the C02 concentration would not be influenced by the
plants surrounding the chamber. The rate of air flow was determined
with a Hastings hot wire anemometer. The air was discharged from the
chamber through a duct located on top.
Carbon dioxide concentrations were monitored at the chamber air
intake duct and at the exhaust duct. The air at each location was col
lected through tubes which extended to a Beckman infra-red C02 gas analyzer.
Air flow was also measured at the intake and exhaust. This enabled the
determination of the ppm C02/unit of air/time entering and leaving the
chamber. The differential is the amount of CO2 produced or consumed
within the chamber.
The CO2 analyzer readings had to be calibrated against a standard


98
but a curious sharp rise occurred in April where the maximum concentration
of 0.40 mg/1 occurred. Concentrations dropped in May and returned to
initial levels in June.
Magnesium is an important structural component of the chlorophyll
molecule and is also used by plants in the metabolism of carbohydrates.
It is therefore also an essential element for plant growth. Magnesium
levels remained fairly constant throughout the year usually ranging
between 9 and 17 mg/1 (Fig. 4).
Nitrogen is a major nutrient for plants and becomes available in
the form of nitrates. Figure 5 illustrates the values for the sum of
nitrates for the study period. Since nitrites are usually fairly low
the curve probably gives a fair representation of relative nitrate values.
The range of total nitrates and nitrites was between 0.3 and 3.5. The
lower values occurred in July through October and increased through the
winter. A decline began in January and continued through March. Concen
trations began to increase gradually in the spring and by June the values
were higher than those from the previous year. Nitrate concentrations
may be inversely related to water level as similar but opposite trends
are noted in Figure 2.
Figure 6 shows the annual variation in pH measured over the study
period. Values did not vary much usually ranging between 7.0 and 7.7. A
decrease was noted in late December and again in April and May. At these
times the pH value became as low as 6.9. Maximum values occurred in
November and January when they reached 7.8. Except for the spring data
pH seems to parallel total alkalinity.


225
Plant Parts and Proportions
The response curves for the weights of various plant parts to
varying levels of insect infestations show very similar trends and
tend towards exponential declines (Figures 38-42). Regression anal
ysis, however, showed that the log response curves did not improve
the correlation coefficients when compared to a linear response curve.
The results, therefore, are plotted as straight line relationships.
If further data were available beyond the levels of infestation tested
in this experiment a curvilinear response might be more evident.
Of the four plant parts rhizomes showed the greatest increase in
the control treatment at approximately 228% of the initial value. The
0.33, 0.67, and 1.00 levels showed responses of 128%, 65%, and 48%
respectively. This indicates that in a situation without an insect
infestation and greatest proportion of the carbon fixed is stored in
the rhizome. The proportion of rhizome weight to total living plant
weight tends to support this. The rhizome represented 8.7% of the ini
tial plant weight and increased to 11.6%, this represents an increase
of 133% (see Table 11). Penfound and Earle (1948) stated that the
rhizome was the main organ of storage. It is apparent in this study
that a great deal of the energy assimilated by the plant is stored in
the rhizome as carbohydrates. Insects, by causing a decrease in the
ability of the plant to create this storage, cause a depletion in the
carbohydrate reserves. This directly affects the ability of the plants
to survive periods of further stress and to resprout from the rhizome
if the leaves are killed. This may reduce the ability of the plant to
survive periods of cold, herbicide treatments, pathogens, and further
insect attacks.


Figure 25. Percentage of the leaves and rhizomes of the waterhyacinth
population damaged through feeding activity of Arzama densa
at Lake A1ice.




Figure 34. The effects of varying insect feeding activity on the
amount of dead waterhyacinth plant material (detritus)
per unit area. Data was taken only from the fall experi
ment. Legend as in Figure 28.
Regression: Y = 269.15 3.76X + 2980.51X2 -
2456.33X3, r = -0.7984


and placed in petri dishes until all of the egg parasites emerged. By
17 September (13 da. post-oviposition) 80 adult T. arzamae had issued
resulting in 100% egg parasitizm. This leads me to believe that the
layered conformation of the typical egg mass protects the lower layers
of eggs from this parasite. Subsequent preliminary examinations of
parasitized egg clusters indicate that only the outside layer of eggs
are parasitized. The thick coating over the eggs probably prevents the
parasites from working their way down in between layers. Because of the
short ovipositor of this parasite only the outermost layer of eggs is
vulnerable to attack. Since only about one third of the eggs are so
protected I would expect the maximum egg mortality due to this parasite
to be about 67%. I have collected eggs for rearing purposes from various
locations at all times of the year and have found T. arzamae continually
present.
Vogel and Oliver (1969b) indicated that a second egg parasite,
Anastatus sp., was present in Louisiana. I have not found this in
Florida. I have found the ladybird larva, Coleomegilla imaculata DeGeer,
commonly feeding on the eggs on Pontederia sp. Vogel and Oliver
(1969b) also list this species. I have further found an unidentified
cecidomyiid larva commonly attacking the eggs on Pontederia.
Figure 49 illustrates the structure of the A. densa population based
on larval and pupal instars. The susceptibility of a particular instar
to sampling is dependent upon the length of the stadium and the prominence
of the damage. The duration of the seventh larval stadium is approximately
three times as long as the first larval stadium. Damage by the seventh
instar larva is very conspicuous and easily detected while damage by
the yaungerlarvae is less so. I believe that this accounts for the


114
Waterhyacinth Productivity
Figures 10 and 11 illustrate diurnal curves for the productivity of
small and large waterhyacinths. Incident solar radiation and ambient
temperature curves for the two days in which this study was done are also
given. This data is the result of infrared C02 gas analysis described
in the methods section. This study was carried out cooperatively with
Sandra Brown, Ken Dugger, and Bill Mitsch and it was agreed that each
investigator would use the results freely as his research dictated. Even
though this agreement was made I do wish to point out that this is not
entirely my own material.
Gross primary production of the large plants (Fig. 10) was determined
to be 19.3 g Carbon/m2/da. Respiration was estimated at 13.2 g C/m2/da.
With the assumption that 1 g carbon = 10 kcal these figures are trans
formed into 193 kcal/m2/da and 132 kcal/m2/da. This indicates a value
of 61 kcal/m2/da for net primary production. The ratio' of this value
and the incident solar energy indicates a net efficiency of 1.6%. This
translates into a net gain of 13.55 gm organic matter (assuming 1 gm 0M =
4.5 kcal). Since the standing crop was 2140 gm/sq. m. a net gain of 0.63%
for the 24 hour period is estimated (0.83% standardizing to the 4900 kcal
solar radiation measurement of the small plants).
The gross primary production of the small plants (Fig. 11) was 15.6
g C/m2/da (156 kcal/m2/da). Respiration was estimated at 7.6 gm C/m2/da
(76 kcal/m2/da) and net primary productivity at 8.0 gm C/m2/da (80 kcal/m2/da)
The net efficiency for the small plants then is also 1.6%. Since the
standing crop is smaller this represents a relatively larger organic
matter gain. The net productivity of 80 gm C equals an organic matter gain


67
although these should probably be called the leaf mining phase and the
stem phase (rather than the petiole phase). He noted that the length
of the pupal period averaged 17.6 days and described the egg, first
instar larva, full grown larva, pupa, and adult (from Walker).
Robertson-Miller (1923) published many observations on the biology
of Bellura gortynoides Wlk. and B. melanopyga Grt. Her information did
not differ much from that of Welch (1914). She described the larvae of
each and indicated that they did not appear to be much different. She
described the egg masses and indicated that those of B. gortynoides
were deposited in flat mats of about 20 eggs each. She noted that
some of the eggs were covered with silvery white threads. The masses
of B. melanopyga were similar to those of B. gortynoides. She found that
B. gortynoides may pupate in the petiole, in soil, or in wood. When
in the petiole the pupae of B. melanopyga was at the top of the burrow
while those of B. gortynoides were lower down. She also found that
B. gortynoides would feed of pickerel weed (Pontederia cordata) in
captivity.
Needham et al. (1928) repeated the observations of Claassen (1921)
on Arzama (= Bellura) obliqua (Wlk.) and of Welch (1914) on Bellura
melanopyga Grt.
Comstock and Dammers (1944) described the full grown larva and pupa
of Arzama gargantua Dyar. I see no distinction between this description
and previous authors' descriptions of these stages of Bellura obliqua (Wlk.).
Guppy (1948) described the habits of Arzama (=Bellura) obliqua
attacking skunk cabbage (Lysichiton kamtschatcense) on Vancouver Island,
B. C. He also indicated that they overwintered under loose bark on
fallen logs.


109
air temperatures had photosynthetic rates 15-30% lower during the day.
They also found that starch accumulation in the cholorplasts of 10C plants
was 2.5 times greater than the 30C plants. They attributed this to the
failure of the plant to translocate the previous day's starch accumulation
from the chloroplasts.
During the period of this study water temperature ranged between 10C
and 32C except for two weeks in January where the weekly minima were
recorded as about 4C. The median water temperature was never less than
12C nor more than 31C. The highest maxima occurred in late August and
early September and the lowest minima occurred in December and January.
The optimum median temperatures (28-30C as determined by Knipling et al.)
occurred in June through October but median temperatures above 20C occurred
from mid-March through late November. It may be concluded that water temp
eratures were generally favorable for waterhyacinth growth most of the
year although winter lows in December and January probably hindered growth.
The lethal limit (maximum) of 39-40C was never approached so it is not
reasonable to expect a summer decline resulting from high water temperatures.
Air temperature ranged from a minimum of 28 in December and January
to a maximum 41C in August. The median weekly temperature was never less
than 10C nor greater than 30C. Minimum temperatures, however, were con
sistently less than 10C from early November to late April. Summer temp
eratures were relatively constant and averaged about 27C. The winter was
comparatively mild with only 8 days having minimum temperatures of 0C or
less. Six of these freezes occurred in December and 2 occurred in January.
None were serious enough to severely damage the waterhyacinths. The only
detrimental effect noticed was a browning of the leaf tips on some of the
larger plants.


269
sufficient energy for regrowth in the spring. By reducing these reserves
the effect of the insect infestation was that of decreasing the winter
hardiness of the plants and preventing regrowth in the spring.
While total biomass showed little change between the two sites the
turnover rate obviously was greater at the release site and accounted for
the decrease in standing crop. This is manifest in the increased detrital
production. The ratio of dead:living plant material was almost nine
times greater at the release site than at the control site.
Following the decline in the waterhyacinth population several changes
in the species composition were noted at the release site. In August,
at the time of release, the pond was covered with a pure stand of water-
hyacinth (Figure 43(a)). By October a large proportion of the leaves were
beginning to wilt and become brown and Hydrocotyle had begun to appear
amongst the waterhyacinth plants (Figure 43(b)). In January only a few
patches of dead waterhyacinth were evident (Figure 43(c)) and hydrocotyle
dominated the surface although there were some areas of open water.
Hydrocotyle increased and by March (Figure 43(d)) it covered the entire
surface of the pond. In July the surface had again begun to open and
hydrocotyle was less dominant. A mixture of Hydrocotyle, Polygonum,
Bidens, and Ludwigia was present and the small stand of Typha had begun
to expand (Figure 43(e)). Waterhyacinth was again present but only in a
small patch on the southeast side. By mid-August Hydrocotyle was present
only in small patches and Typha was dominant. Waterhyacinth was present
in a pure stand the whole year at the control site with the exception
of a small fringe of Hydrocotyle which appeared in the spring.
I had expected waterhyacinth to reoccupy and dominate the pond at


321
Gay, P. A. 1960. Ecological studies of Eichhomia arassipes Solms in
the Sudan. I. Analysis of spread in the Nile. J. Ecol. 48: 183-191.
Goin, C. J. 1943. The lower vertebrate fauna of the water hyacinth
community in northern Florida. Proc. Fla. Acad. Sci. 6(3-4): 143-153.
Gordon, R. D., and J. R. Coulson. 1969. Report on field observations
of arthropods on water hyacinth in Florida, Louisiana, and Texas,
July 1969. Aquatic Plant Control Prog., Tech. Rep. 6: B3-B37,
U.S. Army Eng., WES, Vicksburg, Mississippi.
Gowanloch, J. N. 1944. The economic status of the water hyacinth in
Louisiana. La. Conserv. 2(9): 3,6,8.
Gowanloch, J. N., and A. D. Bajkov. 1948. Water hyacinth program.
La. Dept. Wildlife and Fish., Biennial Rep. (1946/1947) 2: 66-124.
Grossbeck, J. A. 1917. Insects of Florida. IV. Lepidoptera. (F. E.
Watson, ed.). Bull. Amer. Mus. Natur. Hist. 37: 1-147.
Grote, A. R. 1873. Descriptions of North American Noctuidae No. 3.
Trans. Amer. Entorno!. Soc. 4: 293-310.
Grote, A. R. 1874. List of the Noctuidae of North America. Bull.
Buffalo Soc. Natur. Sci. 2: 1-77.
Grote, A. R. 1878 [1879]. Descriptions of Noctuidae chiefly from
California. Bull. U.S. Geol. and Geogr. Surv. 4(1): 169-188,
Washington, 1878. Cited from: New Publications, Bull. Brooklyn
Entomol. Soc. 2: 36, 1879.
Grote, A. R. 1882. An illustrated essay on the Noctuidae of North
America with a colony of butterflies. E. W. Classey, Ltd. Middlesex,
England.
Grote, A. R. 1883. Introduction to a study of the North American
Noctuidae. Proc. Amer. Phil. Soc. 21: 157.
Grote, A. R. 1889. The Noctuidae of North America and Europe compared.
Can. Entomol. 21: 226-230.
Grote, A. R. 1890. North American Lepidoptera. Revised check list of
North American Noctuidae. Part I. Thyatirinae Noctuinae.
Bremen, Germany. 52 pp.
Grote, A. R., and C. T. Robinson. 1867-8. Descriptions of American
Lepidoptera No. 3. Trans. Amer. Entomol. Soc. 1: 323-360.
Gundlach, J. 1881. Contribucin la Entomologia Cubana. Havana, Cuba.
Guppy, R. 1948. Some notes on the habits of Avzoana obliqua on Vancouver
Island (Lepidoptera: Phalaenidae). Proc. Entomol. Soc. B.C. 44: 17-18.


147
values indicate that the distribution is skewed left (towards low values).
Positive values for kurtosis indicate a high degree of peaking where a few
classes contain most of the individuals in the distribution. Negative
values indicate a broad distribution with less distinct peaks. Both
skewness and kurtosis are approximately 0 in a normal distribution. The
regression indicates that when the population is not strongly skewed to
wards the larger size classes the distribution tends to be normal or de
pressed with several classes well represented. As the degree of skewness
towards the large size classes increases, however, the population shows
much stronger peaks indicating the increased predominance of a few size
classes. This supports the contention that the increased dominance by
the large plants results in a loss of the smaller size classes and a
decrease in density.
In general, then, as the predominant size class becomes larger there
appears to be a loss of plants in the smaller size classes and plant den
sity decreases. This is further illustrated in Figure 22 where each
weekly frequency distribution is plotted from January through December
in a three dimensional manner. This is particularly true in the summer
when the largest plants are also the predominant class. As the leaves
from the larger plants die the small plants become better represented
and the density increases. Density, then, appears to be auto-regulatory
and responds to the changes in the canopy. Even though the photophase
is decreasing the amount of light may increase in the lower canopy as
the leaves of the larger plants die off. This may stimulate offset pro
duction as is evidenced by the close inverse association between plant
height and plant density. There appears to be an optimum, however, and
this may have occur ed in April. At this time the degree of intraspecific


Table 8. Correlation coefficients (r) between independent variables.
Statistics inparentheses represent the probability of a
greater |r| under the null hypothesis.
Solar Min. Air Max. Air Min. H20 Max. H^O
Radiation Temp. Temp. Temp. Temp.
Solar Radiation
1.000
(0.0000)
0.615
(0.0001)
0.558
(0.0001 )
0.608
(0.0001)
0.612
(0.0001)
Min. Air
1.000 0.815 0.848 0.828
(0.0000) (0.0001) (0.0001) (0.0001)
Max. Air
1.000 0.796 0.830
(0.0000) (0.0001) (0.0001)
Min. H20
1.000 0.813
(0.0000) (0.0001)
Max. H20
1.000
(0.0000)
Rhizome Damage
Leaf Damage
Phosphorus
Nitrogen
Iron
Conductivity
Potassium
Magnesium
Alkalinity
pH
Sulfates
Lake Level


CHAPTER II
THE CONSEQUENCES OF ATTACK BY ARZAMA DENSA WLK. ON SOME ECOLOGICAL
CHARACTERISTICS AND MORPHORMETRIC FEATURES OF WATERHYACINTH.
Introduction
The effects of an insect attack on a host plant depends not only
on the biology of the insect but also on the ecological response of the
plant. Harris (1962) noted that insect attacks may decrease plant abun
dance, have no effect on plant abundance, or actually stimulate plant
growth. He further stated that insects which feed on aquatic plants
may cause sectioning of the stems from which propagation occurs and
increase the spread of the plant. Bennett (in Harris 1972, apparently
referring to Vogel and Oliver (1969a)) stated that it has been demon
strated that a large noctuid (probably Arsama densa) which attacks
waterhyacinth (Eichhomia erassipes) may create more plants and spread
the weed. Vogel and Oliver (1969a) attributed this increase in the
number of plants with increasing insect concentrations to a reduction
in dominance of the apical bud. Their hypothesis was that by feeding
on the apical meristem the insect caused the expression of the lateral
buds thereby increasing the number of offsets produced.
If herbivory can cause the spread of a weed and thus increase the
probelm, it is imperative to determine the mechanism by which this
occurs. A number of explanations other than a reduction in apical
dominance are possible for this increase in offsets. A reduction in
intraspecific competition or an increase in nutrient due to an accelerated
turnover rate may contribute to this. Seasonal effects must also be
considered. The purpose of this study is to examine in detail the
185


ao-
PLANT
DENSITY
OllltVIO 0*TA
raioicTie
N D


Figure 52. The number of parasites of 4th instar (Campoletis sp.) and 7th instar
[Lydella radiis) Arsama densa larvae as estimated from the number of
pupae, or pupal exuviae found in A. densa bores per square meter of
waterhyacinth mat.


42
pollinators. Penfound and Earle (1948) observed honeybees, bumblebees,
black unidentified bees, and sulfur butterflies visiting the flowers.
They described three patterns of behavior of honeybees in visiting the
flowers; visiting distal anthers only, alighting with the head among the
proximal anthers and the abdomen on the stigma, and visiting the proximal
anthers after alighting on the banner petal. They questioned the import
ance of insect pollinators in accounting for the production of seed in
this species. Bock (1966) noted, however, that honeybees crawl down the
floral tube to retrieve the nectar and in so doing receive pollen from
both sets of anthers. She observed a great deal of cross-pollination.
Autogamous pollination occurs when the flower wilts and the stamens
are twisted against the stigma (Penfound and Earle 1948; Bock 1966; Tag
el Seed and Obeid 1975). Penfound and Earle (1948) found much more pollen
on the stigma after the flowers had completely wilted than at any other
time thus stressing the prevalance of autogamous pollination.
Since waterhyacinth flowers are at least dimorphic with regard to
style length either legitimate (styles pollinated by anthers not of equi
valent length) or illegitimate (styles pollinated by anthers not of equi
valent length) crosses are possible (Ornduff 1966; Bock 1966; Frangois
1964-63). Both legitimate and illegitimate crosses result in seed production
(Bock 1966). Frangois (1964-63) reported that self-incompatibility was
stronger in long styled forms than in short styled forms. Ornduff (1966)
studied the breeding system of Pontederia cordata and compared it with
E. crassipes. He concluded, as did Bock (1966), that both species exhibit
relatively weak self-incompatability.
An interesting aspect of the floral biology of waterhyacinth is the


o
00


304
or if it is the result of immigration from other populations and
possibly from the other host plant [Pontederia).
A parasite of A. densa pupa has been encountered occasionally
although never from the Lake Alice population. It is an ichneumonid
(Chasmias seelestus Cr.) and in two years of collecting larvae and
pupae has only been found twice. I doubt that this species has a
serious impact on the A. densa population. Vogel and Oliver (1969b)
listed two other species, Ichneumon n. sp. (Ichneumonidae) and
Eupteromalus viridescens (Walsh), as pupal parasites of this host in
Louisiana. Table 17 is a list of the various parasites attacking A.
densa. At least seven parasites have been associated with A. densa
representing 5 different families of insects. At least four of these
occur in Florida.
Discussion
Arsama densa Walker (1865) occurs naturally in Florida feeding on
pickerelweed (Pontederia sp.) and has extended its host range to include
the introduced waterhyacinth (Eichomia crassipes (Mart.) Solms). The
larger larvae are capable of causing severe damage to waterhyacinth when
populations reach high levels. This has been observed in the field but
these outbreaks are generally very localized and of short duration.
Severe pressure by a diverse parasite complex appears to restrict such
outbreaks and maintain A. densa populations at low levels on waterhyacinth.
The results of the studies from Lake Alice indicate that many of
the larvae escape parasitism in the fall but since host stages are con
tinuously available the parasite populations eventually build up and
suppress the A. densa population. This suggests the possibility of the
manipulation of A. densa populations for the control of waterhyacinth.


Figure 5
Phosphorus concentrations present as phosphates and nitroqen
concentrations as total nitrate and nitrites from water
samples taken from Lake Alice. The lines represent 5-point
moving averages. w


326
Moffat, J. A. 1889. Arzama obliquata. In Correspondence Can. Entomol.
21: 99.
Mohamed, B. F., and F. F. Bebawi. 1975. The distribution of Eiohhomia
orassipes (Mart.) Solms in the White Nile, Sudan. MS. 13 pp.
Morton, J. F. 1962. Eiohhomia orassipes Solms. Hyacinth Control J.
1: 31. Reprinted from Wild plants for survival in South Florida.
Hurricane House, Miami, 1962. Page 18.
Mosher, E. 1919. Notes on lepidopterous borers found in plants with special
reference to the European corn borer. J. Econ. Entomol. 12: 258-268.
Muenscher, W. C. 1967. Aquatic plants of the United States. Comstock
Publ. Co., Inc., Ithaca, New York. 374 pp.
Mukherjee, R. K., A. Bhanja, P. R. Burman, and S. M. Sircar. 1964
Presence of bound auxin in the roots of water hyacinth (Eichhomia
orassipes). Bull. Bot. Soc. Bengal 18(1): 87-90.
Mulrennan, J. A. 1962. The relationship of mosquito breeding to aquatic
plant production Hyacinth Control J. 1: 6-7.
Mue, T. L., and J. W. Parham. 1954. Water hyacinth Bekabekairaga
(Eiohhomia orassipes, Solms). Fiji Agr. J. 25: 82-83.
Munz, P. A. 1965. A California flora. University of California Press,
Berkeley. 1681 pp.
National Science Research Council of Guyana and National Academy of Sciences,
U.S.A. 1973. Some prospects for aquatic weed management in Guyana.
Workshop on Aquatic Weed Management and Utilization, Georgetown,
Guyana. 39 pp.
Needham, J. G., S. W. Frost, and B. H. Tothill. 1928. Leaf-mining
insects. The Williams and Wilkins Co.. 351 pp.
Nolan, W. J., and D. W. Kirmse. 1974. The papermaking properties of
waterhyacinth. Hyacinth Control. 12: 90-97.
Odum, E. P. 1969. The strategy of ecosystem development. Science 164:
262-270.
Odum, H. T. 1953. Dissolved phosphorus in Florida waters. Fla. Geol.
Surv., Rep. of Invest. No. 9. 40 pp.
Odum, H. T. 1974. Energy basis for man and nature, or energy: crisis
to steady state. McGraw-Hill, Co., publ. pending.
O'Hara, J. 1967(1968). Invertebrates found in water hyacinth mats.
Quart. J. Fla. Acad. Sci. 30(1): 73-80.
Olive, E. W. 1894. Contribution to the histology of the Pontederiaceae.
Bot. Gaz. 19: 178-184.
Ornduff, R. 1966. The breeding system of Pontederia cordata L. Bull.
Torr. Bot. Club 93(6): 407-416.


This dissertation was submitted to the Graduate Faculty of the College
of Agriculture and to the Graduate Council, and was accepted as partial
fulfillment of the requirements for the degree of Doctor of Philosophy.
June 1976
2 oJk .
Deal/,/College of Agriculture
Dean, Graduate School


312
nearby natural population. This suggests that in some situations local
populations can escape parasitism. The mechanism for this is not clear
but it may have been due to synchronization of the age distribution in
the released populations.
About one half way through this research it became apparent that
A. densa was almost always more abundant on pickerelweed (Pontederia
cordata L.) than on waterhyacinth. These observations were casual,
however, and no data has been obtained from these populations. Some im
portant questions that should be answered could be derived from con
current studies of populations on both of these host plants. It would be
interesting to determine if the two populations are in phase or out of
phase. Is the population on waterhyacinth the result of dispersing in
dividuals from pickerelweed after the population from the latter host
builds up to a high level? Does waterhyacinth represent a secondary host
allowing the population to continue while pickerelweed is scarce? Do
A. densa populations control P. cordata to any extent? Is there any dif
ference in rates of parasitism of A. densa on its two host plants? To
what extend do populations from both hosts interact? Are these two popu
lations temporally separated? These and many other questions would greatly
enhance our understanding of the observations on the A. densa populations
reported here.
Life table studies of Arzama densa would be helpful in evaluating
methods of population augmentation. Such studies of natural populations
would be difficult if not impossible, however, because of the problems
associated with obtaining adequate samples discussed above. A possible
approach to this might be mass releases with subsequent life table studies
of the released population.


219
Productivity and Turnover Estimates
The relationship between biomass and detritus can be more graphically
illustrated if net productivity and turnover rates are considered (Figs.
36 and 37). Net productivity, as used here, is defined as the final
quantity of organic material present (total biomass present at the end of
the experiment detritus present at the beginning) per quantity of living
material initially present and is expressed as a percentage. A value of
100 for net productivity would indicate no change in biomass and is the
minimum value possible. Statistically Figure 36 indicates that there was
no significant change in net productivity as a result of insect feeding
activity. Intuitively there does appear to be some effect, however, as
productivity was approximately 170% at all levels of infestation whereas
it was approximately 205% for the control.
Assuming that productivity was not affected by the insect population
and with the knowledge that standing crop was significantly reduced, it
can be deduced that the important function of the insects was that of
accelerating turnover. The effects of insect feeding activity on the
relative turnover ratio is illustrated in Figure 37. The turnover ratio
as used here refers to the change in detritus per unit area per gram of
living plant material initially present. It is inversely related to
turnover time or the time required for the initial living material to be
converted to detritus. An approximately linear relationship is apparent
between insect activity and the turnover ratios. These ratios translate
into a range of turnover times of approximately 153 days for the control
to 40 days for the 1.00 infestation level (where T.0. time = j-qX
47.5 da (mean duration)). It is obvious, then, that the turnover time
decreases exponentially with increasing insect concentrations. These


Head Capsule Diameter iirnn)
285
Larval Age At Molt (da post eclosin)


ALKALINITY (ppm CO, HC03)
cn
CONDUCTIVITY (fnmhos/cm)


36


318
Burton G. J. I960. Studies on the bionomics of mosquito vectors which
transmit filariasis in India. II. The role of water hyacinth (Eichhomia
speciosa Kunth) as an important host plant in the life cycle of Mansonia
uniformis (Theobald) with notes on the differentiation of the late
embryonic and newly hatched stages of Mansonia uniformis (Theobald)
and Mansonia annulifera (Theobald). Indian J. Malario!. 14(2): 81-106.
Caldwell, S. Y. 1942. Orchid of the lily pool. Flower Grower 29: 269.
Carter, M. R., L. A. Burns, T. R. Cavinder, K. R. Dugger, P. L. Fore,
D. B. Hicks, H. L. Revells, and T. W. Schmidt. 1973. Ecosystems
analysis of the big cypress swamp and estuaries. U.S. E.P.A. Region
IV, Atlanta.
Cason, J. H. 1970. Lake Alice -- A study of potential pollution of the
Floridan Aquifer. Compass 47(4): 206-210.
Castellanos, A. 1958. Ntula sobre el gneroPontederia en Brasil.
Arquivos do Jardim Botnico do Rio de Janeiro 15: 59-67.
Castellanos, A. 1959. Las Pontederiaceae de Brasil. Arquivos do
Jardim Botnico do Rio de Janeiro 16: 149-216.
Cavers, P. B., and G. A. Mulligan. 1972. A new series -- The biology of
Canadian weeds. Can. J. Plant Sci. 52: 651-654.
Center, T. D. and J. Balciunas. 1975. The effects of water quality on
the distribution of alligatorweed and waterhyacinth. Aquatic Plant
Control Prog., Tech. Rep. 10: B3-B13, U.S. Army Eng., WES, Vicksburg,
Mississippi.
Chadwick, M. J. and M. Obeid. 1966. A comparative study of the growth of
Eiehhomia orassipes and Pistia stratiotes in water culture. J. Ecol.
54: 563-575.
Charudattan, R. 1975. Use of plant pathogens for control of aquatic weeds.
Ecol. Res. Ser., U.S. E.P.A. 660-3-75-001: 127-153.
Chatterjee, I., and M. A. Hye. 1938. Can water hyacinth be used as a
cattle feed? Agr. and Live-stock in India 8(5): 547-553.
Chhibbar, S. S., and G. D. Singh. 1971. Paddy straw and water hyacinth
silage. Indian Farming 20: 24-26, 32.
Claassen, P. W. 1921. Typha insects: their ecological relationships.
Cornell University Agr. Exp. Sta. Mem. 47: 458-531.
Coker, W. C. 1907. The development of the seed in the Pontederiaceae.
Bot. Gaz. 44: 293-301.
Combs, G. E. 1970. Aquatic plants for swine feeding. Proc. Aquatic
Plant Res. Conf., University of Florida, Gainesville.
Comstock, J. A. 1944. Four California moths associated with cat-tails.
Bull. S. Calif. Acad. Sci. 43(2): 81-83.


76
Annual Cycles and Insect Damage
Estimates of various plant characteristics, of the Arzama densa
population, and of plant damage by A. densa were taken on a weekly basis
from May 1974 to 30 April 1975. Sampling was done on a plot system
using a rubber ring enclosing an inside area of 0.316 m2. The samples
were taken each week in a pseudo-random manner. I have not been able
to devise a satisfactory system of pinpointing a previously randomly
selected point on a mat of waterhyacinths and then finding that point
while trying to maneuver through the dense stand of plants. To elim
inate the additional variables of seasonal plant species composition
changes along the shoreline and different waterhyacinth growth charac
teristics only the central area of the lake was studied. The area in
which samples were taken was defined by the catwalk on the west side
and extended 25 m to the north and 25 m to the south of the central
point on the catwalk. The eastern boundary was established by a small
row of bushed 50-60 m from the catwalk that extended into the lake from
the north shore. The study area, then, was 2500-3000 m2 in the central
more or less homogenous region of the waterhyacinth mat on the marsh
side of the catwalk. Sampling points were selected by throwing the ring
in a high arc. After it fell into the mat it was reached using two Dow
styrofoam billets (3 m X 0.5 m), one placed in front of the other sequen
tially. This allowed me to move (with some effort) over the mat on the
water surface. Once the ring was reached the billets were used as plat
forms for counting and recording data.
The ring was manipulated down over the waterhyacinths until it was
on the water surface. This involved making subjective decisions as to


84
As mentioned previously Lake Alice receives effluent from the campus
sewage treatment facility and cooling water from the heating plant. The
nutrient enriched water from the former and the warm water from the lat
ter have probably contributed significantly to the eutrophication of
this lake. Other sources include overflow from Hume Pond, also located
on the university campus, runoff from the local watershed, and direct
rainfall. Discharge of the system is through the wells mentioned earlier.
Water loss also occurs through surface evaporation and evapotranspir-
ation. Mitsch (1975) estimated the hydrological budget for the lake in
terms of flows and storage (see Table 3). The water storage at a stage
of 69 feet above mean sea level is estimated to be 254 x 103 m3. Water
retention is low due to the high input-volume ratio. Brezonik et aZ.(1969)
suggests that this may have a flushing effect causing low phytoplankton
populations noted in the lake.
Discharge through the wells is regulated by valves and is frequently
altered by campus personnel. The water level is often raised to facili
tate mechanical removal of the waterhyacinth. During the hurricane season
the water level is dropped to prevent flooding. Fluctuations are also
caused when the discharge screens over the wells become clogged with
debris. Water level appears to be correlated with seasonal precipitation
patterns (Fig. 2) except for the months of December and January. During
this time an oil spill occurred in the canal from the heating plant and
sewage treatment facility. Flows from these two sources were minimized
so the oil could be cleaned up. This resulted in a sharp drop in the
lake level. Normal flow was restored the first part of February and a
sharp increase in the lake level followed.


37
and Penfound and Earle (1948). The lacunar system may enable the plant
to utilize internal carbon dioxide (Billings and Godfrey 1967).
The inflorescence is displayed on a long peduncle (Penfound and
Earle 1948) and is usually elevated a few centimeters above the leaves
(du Toit 1938). Two unlike spathes subtend the inflorescence the lower
being leaf-like and bearing a pseudolamina and the upper bract-like
(Cook 1974). The inflorescence is a spike (du Toit 1938; Buckman and
Co. 1930; Penfound and Earle 1948; Bock 1966) or may be considered
spike-like or paniculate (Cook 1974; Bock 1966). The spike is 15-30 cm
long (du Toit 1938; Mune and Parham 1954) and contains numerous flowers
(6-20, du Toit 1938; 8, Parsons 1963; 10-12, McLean 1922; 6-12, Mune and
Parham 1954; 4-29, Misra 1969) borne on a rachis with a flowerless sub-
rachis below the inflorescence and above the spathes (Penfound and
Earle 1948). The individual flowers consist of a hypanthium about 1.4-
1.8 cm long (Misra 1969), 3 sepals, 3 petals, 6 stamens and a tricarpel late
ovary (Penfound and Earle 1948). The petals and sepals are lavender in
color (Bock 1966) and united at the base to form a 6-lobed tube (Cook
1974). The color of the flower is due to the anthocyanin, eichornin
(Shibata et at. 1965). The tube is curved, glandular and pubescent near
the base (Mune and Parham 1954). The perianth is slightly irregular with
all 3 sepals and 2 petals similar in size and shape but the upper petal
is somewhat wider and bears a distinctive yellow spot in the center
bordered by a darker blue or violet area (Bock 1966; Buckman and Co. 1930).
Buckman and Co. (1930) indicate that the function of this spot is obscure
but others have indicated that it may function as a nectar guide to
visiting bees (Sculthorpe 1967). The six stamens are arranged in two


275
dissected a 10 day old female pupa and found fully formed eggs. It
appears then that oviposition can occur almost immediately after
emergence and mating.
Fecundity
Several pupae were collected from waterhyacinths in July 1973.
These were individually placed in baby food jars which were held in sealed
aquaria lined with damp paper towelling. Upon emergence a single male
was paired with a single female in a one gallon ice cream carton. The
adults and pupae were held in an environmental chamber at 25 C, 16:8 L:D
photoperiod. In all, five pairs of adults were obtained from the field
collected pupae. The cartons were checked daily for eggs. The eggs were
removed and counted, held in baby food jars in the manner described in
Section 2, and allowed to eclose. The neonates were removed each day and
the egg developmental time noted. These results are summarized in
Table 14. The average fecundity was 225 eggs/female, 72.6% were laid on
the first day although oviposition generally continued for three days.
Viability was 80.7% and the average developmental time was 5.6 da.
Fecundity was checked again when larvae were reared for a field
release (Section 3). The pupa were placed in vermiculiteand held
in a cage constructed of hardware cloth and lined with wax paper. In
stead of isolating pairs of adults all were kept in the cage and fed
a sucrose solution. A total of 18 males and 17 females were reared
(C?:sex ratio = 1.06:1). Three females emerged after the death of
the last male and did not contribute fertile eggs. From the 14 females
that did mate 2872 eggs were obtained. This represents an average of
205 eggs per female. Only 4.49% of the eggs were sterile, 17.06% were
fertile but failed to eclose, and 78.45% eclosed. Both of these estimates


Figure 43(f)


CHAPTER I
THE RELATIONSHIP BETWEEN THE PHENOLOGY AND PRODUCTIVITY OF
WATERHYACINTHS AND VARIOUS PHYSICAL AND BIOLOGICAL FACTORS.
Introduction
To evaulate the effects of insects for the biological control of
weeds, a basic understanding of the ecology of the plant is essential.
In realization of this the Canada Weed Committee has instituted a series
on the biology of Canadian weeds (Cavers and Mulligan 1972). This is
an attempt to pull together all the available knowledge on the biology
of Canadian weeds that can be used in weed control efforts. Within this
framework the phenology of the plant (annual variation), and the response
of the plant to limiting factors and damage by indigenous insects is of
special interest for the evaluation of biological control attempts.
Omission of these considerations could result in the misinterpretation
of pertinent data. For example, natural seasonal declines in the plant
population could mistakenly be attributed to insect releases when the
insects are also seasonal if patterns of seasonal variation of the plant
are not known. Also, releases of insects may be more effective when
correlated with critical periods in the annual cycle of the plant.
Judgements for the timing of these releases can be made only on the
basis of what is known about the plant.
Limiting factors can be defined as the necessary components of
the organism's environment which are least available and thereby con
trol the life processes of the population. Liebig (1840) stated that a
process is limited by the quantity of a single component present in
minimal amounts relative to its optimal amounts. Sachs (1860) felt that
biological processes required a certain minimal level of a limiting
70


296


145
The latter two classes comprised 45% of the population and 40% were
smaller. Approximately 16% of the population was larger which was similar
to the previous month. By June this subpopulation of small plants began
to disappear and the distribution was skewed strongly toward the larger
size classes. The 91-100 and 100-110 cm classes were co-dominant with
54% of the population. The smaller classes comprises 34% of the total
and the larger only 10%. This continued in July and the same two size
classes represented 65% of the population. The contribution of the small
plants was minor with only a 23% representation. The predominant size
class was 101-110 cm in July and was the maximum height achieved by a
dominant class. The distribution was similar in August and September
but the predominant class was 91-100 cm in both months.
In October the predominance of the larger size classes was beginning
to decrease and the smaller plants were becoming more important. A net
increase in density occurred which was apparently responsible for the
increase in the smaller size classes. The two predominant classes (91-100,
101-110) comprised only 35% of the population. This trend continued in
November but the two predominant classes were smaller (71-80, 81-90) and
53% of the plants were smaller. By December the predominant class was 31-40
cm and the frequency distribution was broad. Six classes (21-80 cm) were
co-dominant.
Statistics for skewness (assymetry) and kurtosis (peakedness) were
determined for each weekly frequency distribution according to the methods
described by Sokal and Rohlf (1969). Kurtosis was plotted as a dependent
function of skewness and a hyperbolic regression fitted to the data (Fig.
21). Positive values for skewness indicate that the distribution is
skewed towards the ight (high values in the distribution). Negative


331
Taylor, K. G., R. P. Bates, and R. C. Robbins. 1971. Extraction of
protein from water hyacinth. Hyacinth Control J. 9(1): 20-22.
Taylor, K. G., and R. C. Robbins. 1968. The amino acid composition of
water hyacinth (Eiohhomia orassipes) and its value as a protein
supplement. Hyacinth Control J. 7: 24-25.
Thompson, W. R. 1944. A catalogue of the parasites and predators of
insect pests. Sect. 1. Parasite host catalogue. Pt. 5. Parasites of
the Lepidoptera. The Imperial Parasite Service, Belleville,
Ontario, Canada. 130pp.
Tietz, H. M. 1952. The Lepidoptera of Pennsylvania. Pensylvania State
College, PA. 194 pp.
Tietz, H. M. 1972. An index to the described life history, early stages
and hosts of the Macrolepidoptera of the Continental United States
and Canada. A.C. Allyn, publ., 2 Vols.. 1041 pp.
Tilghman, N. J. 1962. The value of water hyacinth in the propogation
of fish. Hyacinth Control J. 1:8.
Tilghman, N. J. 1963. The St. Johns River hyacinth story. Hyacinth
Control J. 2: 13-14.
Timmer, C. E., and L. W. Weldon. 1967. Evapotranspiration and pollution
of water by water hyacinth. Hyacinth Control J. 6: 34-37.
Timmons, F. L. 1960. Weed control in western irrigation and drainage
systems. U.S. Dept. Agr., Agr. Res. Serv., ARS 34-14. 22 pp.
UTtsch, G. R. 1973. The effects of water hyacinths (Eiohhomia orassipes)
on the microenvironment of aquatic communities. Arch. Hydrobiol.
72(4): 460-473.
Ultsch, G. R., and D. S. Anthony. 1973. The role of the aquatic exchange
of carbon dioxide in the ecology of the water hyacinth (Eiohhomia
orassipes). Florida Sci. 36(1): 16-22.
U.S.D.A. 1970. Selected weeds of the United States. Agr. Handbook 366,
Agr. Res. Serv., U.S. Gov't Printing Office, Washington, D. C., 463 pp.
Van der Weert, R., and G. E. Kamerling. 1974. Evapotranspiration of water
hyacinth (Eiohhomia orassipes). J. Hydrol. 22(34): 201-212.
Vernon, R. 0. 1969. Review of Florida's water pollution problems. Ground
water. Proc. Fla. Envir. Eng. Conf. Water Poll. Control. Gainesville,
Fla. Eng. & Ind. Exp. Sta. 24(3), Bull. Ser. 135: 28-43.
Vogel, E., and A. D. Oliver, Jr. 1969a. Evaluation of Arzama densa as
an aid in the control of water hyacinth in Louisiana. J. Econ.
Entolmol. 62(1): 142-145.


163
Iron appears to be an important nutrient in at least three models.
The coefficient for iron was negative in the expression for leaves per
plant. The number of leaves per plant is correlated with standing crop
(see Table 9), hence, the negative coefficient for iron can be taken to
indicate uptake of iron as the plant biomass increases. The positive
coefficients in the models for leaf and plant density indicate that
iron concentrations are high when these two variables are high. Since
maximum plant and leaf densities occur early in the growing season iron
levels may be high because the plants have not yet affected it. The
peak for plant density occurs immediately after the peak for iron con
centration (see Figs. 4 and 19). This indicates that a causal relation
ship may exist between the two.
Potassium was included in the expression for plant height, leaf
density, and plant density. A decrease in the potassium concentration
between late February and late April (Fig. 3) corresponds to the peaks
for leaf and plant density and accounts for the negative potassium co
efficient for these two variables. The positive relationship between
potassium and plant height (Fig. 14) is not obvious by mere inspection
of the data. The indication is that as potassium concentrations increase,
plant height also increases. The effects of more important variables
probably obscure this relationship between potassium and plant height.
Magnesium was significant in the model for leaf density. This was
somewhat surprising since magnesium concentrations were relatively con
stant through the year (see Fig. 4). The coefficient for leaf density
was negative but this inverse association is not obvious.
Hydrogen ion concentration (pH) was included in the model for leaf
area index with a negative coefficient. Values of pH had a narrow range


64
large dorsally oriented last pair of spiracles. He stated that there
were two annual broods, the second of which hibernated in stumps or
moss near the water.
Kellicott (1883a, b), however, felt sure that in New York this
species was single brooded and pupated in May. He also noted that they
overwinter in the soil or old wood.
Riley (1883a, b), in reply to Kellicott's comments, stated that
there could be no doubt as to the digoneutic (=double brooded) nature
of A. obliquata at Washinton (D. C.?).
Comstock (1888) referred to the habits of Arzama (misspelled Argoma)
that infest the leaves of pond lilies. He distinguished these from
truly aquatic larvae in that they "are obliged to come to the surface"
for air.
Brehme (1888a) described the eggs, larva, and pupa of A. obliquata
(G. & R.). He noted a developmental period of about 15 days for the
eggs which were laid on cattail between the long leaves. He found the
larval period to be 161 days and the pupal period to be 16 days making
a total egg to adult span of 190 days.
Brehme (1888a) also stated that the larva returns to the top of the
reed in its later larval stages and forms it pupa there. This sparked
a series of correspondence in the Canadian Entomologist. Moffatt (1888a)
stated that this was not its invariable habit in nature and he had found
the pupa beneath the bark of a decaying stump some distance away from
where the cattails grew. Brehme (1888b) replied that this may not be
invariable but that the majority of them pupate in the reed. He cited
a friend of his who had found the pupa in a stump but indicated that the
larva had been feeding there and wondered if that wasn't true in Moffatt's


Figure 21. Statistics of skewness and kurtosis (peaking) derived from
each weekly frequency distribution of plant density by
height classes. This figure indicates that as the frequency
of plants becomes skewed towards the larger height classes
the degree of peaking in the distribution increases sharply
(i.e. the diversity of height classes represented decreases


Figure 4. Magnesium and total iron from Lake Alice water samples.
Note the sharp increase in iron concentrations in April.
The lines represent 5-point moving averages.


72
water quality and water level), intraspecific conditions (plant density,
canopy effects, available space, etc.)* and biotic stress (insect damage).
These will be evaluated with possible interactions between them considered.
These concepts, possible interactions and all factors which control
the plant must be considered and investigated. Attempts to evaluate the
attack of an insect by studying only the insect or with only a super-
ficial knowledge of the target plant are subject to erroneous conclusions
and misinterpretation. Not only must the plant and the insect be studied
but the insect-plant interrelationships must be established. This field
has received increasing attention lately and may provide a basis for
future biological control efforts.


266
ovipositional success. Hence, the parasite build-up would be slower
than expected as long as the host population remained synchronized. This
assumes negligible recruitment to the parasite population from outside
the release area. This is probably not a valid assumption but the
parasites did fail to control A. densa before sufficient damage was done
to the waterhyacinth population. I plant to conduct a similar study
including a life table analysis of a field released population of A. densa
in the future.


I certify that I
conforms to acceptable
adequate, in scope and
Doctor of Philosophy.
I certify that I
conforms to acceptable
adequate, in scope and
Doctor of Philosophy.
I certify that I
conforms to acceptable
adequate, in scope and
Doctor of Philosophy.
I certify that I
conforms to acceptable
adequate, in scope and
Doctor of Philosophy.
I certify that I
conforms to acceptable
adequate, in scope and
Doctor of Philosophy.
have read this study and that in my opinion it
standards of scholarly presentation and is fully
quality, as a dissertation for the degree of
Dale H. Habeck, Chairman
Professor of Entomology
have read this study and that in my opinion it
standards of scholarly presentation and is fully
quality, as a dissertation for the degree of
er
Professor of Entomology
have read this study and that in my opinion it
standards of scholarly presentation and is fully
quality, as a dissertation for the degree of
George E/ Allen
Professor of Entomology
have read this study and that in my opinion it
standards of scholarly presentation and is fully
quality, as a dissertation for the degree of
Reece I. Sai
have read this study and that in my opinion it
standards of scholarly presentation and is fully
quality, as a dissertation for the degree of
John Reiskind
Associate Professor Zoology


5
Plants floating or sometimes fixed to the substrate,
the leaves in the form of a rosette with the stem
reduced and the plants connected by an elongated
horizontal rhizome; numerous plumose roots issue
from each plant. The aerial leaves are variable
in shape; petioles of 2 to 30 cm long are more or
less inflated; stipules 2-15 cm long with a small
apical orbicular-reniform lamina with a lacerate
[serrate ?] margin; submerged leaves never evident.
Inflorescense variable, internodes between the
spathes nearly absent; inferior spathe with lamina
1-5 cm long, the sheath 3.5-7 cm long. Flowers
4-6 cm long; perianth light purple or rarely white,
tube 1.5-2.0 cm long, lobes 2.5-4.5 cm long, with
entire margins. Stamens all exserted, filaments
villous-gladular. Capsule elliptical, trigonous,
12-15 mm long; seeds oblong-elliptical 1.2-1.5 x
0.5-0.6 mm with 10 longitudinal ridges. (Agostini 1974,P. 305)
The leaves of waterhyacinth are variable in the shape of the blade
and the extent of development of the float. This variability is appar
ently dueto the plants' responseto environmental conditions. It was
shown prior to 1930 that the size of the floats depends on such factors
as light, temperature, and water quality (La Garde 1930). Rao (1920b)
was convinced that increased water uptake (in a hypotonic medium)
promoted the development of the float. This is discussed further by
Bock (1966), Penfound and Earle (1948), and in an exceptionally good
account by Misra (1969).
The waterhyacinth flowers are possibly trimorphic with regard to
the length of the style. Although medium and long styled forms are known,
the existence of short styled forms is thought to be possible (Bock 1966).
The midstylous form is normally predominant and Bock (1966) feels that
only the mid and long styled forms exist. The flowers possess two whorls
of anthers which Bock (1966) notes are long and short in the mid-styled
form and short and mid-length in the long-styled form. This dimorphism
is apparently regulated by a two gene system, one exerting an epistatic


52
Table 1. Standing crop and productivity of waterhyacinths as estimated by various authors.
Source
Standing Crop
1.6-2.7 kg DW*/m2
Productivity
Penfound and Earle (1948)
Dymond (1949)
1.6 kg DW/m2
13-20 kg/m2/yr
Penfound (1956)
0.4-1.3 kg DW/m2
12.7-14.6 g DW/m2/da
5.7-6.5 g C/m2/da
Westlake (1963)
----
1.1-3.3 kg/m2/yr average
15 kg/m2/yr max. (19 g C/m2/da)
Yount (1964)

28 g C/m2/da
Bock (1966)
----
2.5/ per day (Calif, average)
Misra (1970)

9.4-9.6 g 0M*/m2/da (Aug. 1967)
3.48-8.98 g 0M/m2/da (Apr.-Feb. 1968)
Sahai and Sinha (1969)
0.46-0.72 kg DW/m2
3.8 g 0M/m2/da max.
103.0 g 0M/r:2/yr
247.0 g OM/m2 net production to max.
biomass
Knipling et al. (1970)
2.4-2.5 kg DW/m2
7.8-16.1 mg CO?/dm2 leaf/hr net
2.6-2.8 mg C02/dm2/hr respiration
Sinha and Sahai (1972)

1.43 g 0M/m2 leaf/hr net
0.56 g OM/m2 leaf/hr respiration
1.99 g OM/m2 leaf/hr gross
Ornes and Sutton (1975) 9.7 gm/m2 max.
*DW = dry weight; 0M = organic matter.
1.05X per day (= 30 gm 0M/r,;2/week) max.


62
In general there seems to be three families of plants attacked,
the Typhaceae, the Nymphaeaceae, the Pontederiaceae. The Typhaceae
are infested by Bellura obliqua (Wlk.) (in Typha latifolia L. and T.
angustifolia L.) and Bellura gargantua (Dyar) (in T. latifolia L.).
The Nymphaeaceae are infested by Bellura gortynoides Wlk. (Nuphar
advena Ait. and N. variegatum Engelm.) and B. diffusa ("water lily").
The Pontederiaceae are infested by Arsama densa (Wlk.) (Pontederia
cordata L. and Eichhomia crassipes (Mart.) Solms). This supports Todd's
(pers. comm.) contention that possibly only three species are involved.
A fourth family, the Araceae, is strongly implicated within the
host range of this family. Guppy's (1948) record (Table 2, No. lc)
of Arzama obliqua (Walk.) from skunk cabbage (Lysichiton kamtschatcense=
L. camtschatcense = L. americanum Hult. & St. John, see Munz 1965) seems
to be well founded. Tietz's (1952) citation of Symplocarpus foetidus
L. (Table 2, No. lc) probably refers to Guppy's paper. Habeck's (1974)
record of Arzama densa Wlk. from dasheen (Calocasia esculenta L. =
Colocasia esculenta (L.) Schott; Table 2, No. 4a) also seems substan
tiated. These represent two instances of the infestation of two
different species of the Araceae from two widely separted regions
(British Columbia, Guppy 1948 and Florida, Habeck 1974) by apparently
two species of the Bellura complex. Takhtajan (1969) indicates
that there is a close affinity between the Liliales (Pontederiaceae),
Arales, and Typhales and they all are represented along a line of
evolution in common with the Nymphaeales.
Mosher (1919) stated that Arzama obliqua has been reported from
corn. She did not cite any references to these reports, however, and I
have not been able to substantiate this claim. Because crop plants such


294
apparent predominance of the latter stages in the life cycle. This
together with the large degree of overlap between generations makes
analyses of age-specific mortality factors extremely difficult.
Figure 50 illustrates the annual curves for the density of larvae
and pupae collectively and population mortality. The percent mortality
is derived from the number of individuals found dead in the field
relative to the total living and dead. The mortality curve shows a
configuration very similar to the population curve but lags slightly
behind it. This would be expected when most of the mortality is due to
parasitoids. Higher parasitism occurs when the population is high but
mortality as a result of this parasitism does not occur until somewhat
later.
The data for the total annual population counts and the proportion
of each instar found dead is summarized in Table 16. It is apparent
that most of the mortality observed in the field occurred during the
fourth and seventh larval stadia. This was primarily the result of
parasitism by Campoletis sp. oxylus grp. (Ichneumonidae) to the fourth
instar and by Lydella sp. (Tachinidae) to the seventh instar.
Figure 51 summarizes the population data for the 4th and 7th instar
larvae. Illustrated for each instar is the percentage of the total found
dead in the field and the percentage parasitized but still living. The
dashed lines indicating the number escaping is an estimate of the number
of living larvae free of parasites and pathogens. Parasitism by Lydella
and Campoletis is high throughout the year. A few 7th instar larvae
die as a result of infection by the microsporidian Nosema neoatrix
and other causes but the majority of the total observed mortality is a
/
result of parasitoids.


o
BIOMASS [kgDW/W*]
O
PETIOLES
LEAVES
ROOTS
RHIZOMES
mn
x
m
r*
m
>
m
O
5
# *
- z¡ *
m m
STOLONS
DETRITUS
TOTAL
LIVING
TOTAL PHOTO
SYNTHETIC
TOTAL NON
PHOTOSYNTHETIC
TOTAL
BIOMASS
rv>
c_n
U3


134
leaf density occurs in the spring when the plants are at their maximum
density. Leaf density appeared to return very near its May 1974 level
in May 1975.
Plant height, plant density, leaf density, and leaves per plant
appear to be interrelated. It is confusing to consider any one of these
as an indicator of photosynthetic production. A more valuable index is
the leaf area index which takes into account the average area per leaf and
the leaf density. Figure 17 represents the curve for the average area per
leaf. This appears to be similar to the plant height curve but does not
show the brief period of decline in the late summer. A single distinct
peak occurs in July. The leaf area index is the product of the average
area per leaf and the number of leaves per square meter and is represented
in Figure 18. The leaf area index shows a strong increase in the spring
actually beginning as early as February. A peak occurs in May as a result
of both increasing area per leaf and a high leaf density. A secondary