Citation
Comparative studies of the morphology and ecology of sexual reproduction of Eichhornia crassipes (Pontederiaceae)

Material Information

Title:
Comparative studies of the morphology and ecology of sexual reproduction of Eichhornia crassipes (Pontederiaceae)
Creator:
Anderson, Robert Gerald, 1948-
Copyright Date:
1976
Language:
English

Subjects

Subjects / Keywords:
City of Gainesville ( local )
Eichhornia Crassipes ( jstor )
Inflorescences ( jstor )
Nutrients ( jstor )

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright Robert Gerald Anderson. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Resource Identifier:
AAB1500 ( LTUF )
1015738906 ( OCLC )

Downloads

This item has the following downloads:

comparativestudi00ande.pdf

comparativestudi00ande_0097.txt

comparativestudi00ande_0123.txt

comparativestudi00ande_0065.txt

comparativestudi00ande_0134.txt

comparativestudi00ande_0163.txt

comparativestudi00ande_0042.txt

comparativestudi00ande_0121.txt

comparativestudi00ande_0002.txt

comparativestudi00ande_0175.txt

comparativestudi00ande_0049.txt

comparativestudi00ande_0041.txt

comparativestudi00ande_0200.txt

comparativestudi00ande_0173.txt

comparativestudi00ande_0057.txt

comparativestudi00ande_0063.txt

comparativestudi00ande_0137.txt

comparativestudi00ande_0000.txt

comparativestudi00ande_0038.txt

comparativestudi00ande_0003.txt

comparativestudi00ande_0060.txt

comparativestudi00ande_0130.txt

comparativestudi00ande_0155.txt

comparativestudi00ande_0138.txt

comparativestudi00ande_0135.txt

comparativestudi00ande_0028.txt

comparativestudi00ande_0192.txt

comparativestudi00ande_0153.txt

comparativestudi00ande_0048.txt

comparativestudi00ande_0066.txt

comparativestudi00ande_0014.txt

comparativestudi00ande_0129.txt

comparativestudi00ande_0096.txt

comparativestudi00ande_0026.txt

comparativestudi00ande_0125.txt

comparativestudi00ande_0171.txt

comparativestudi00ande_0101.txt

comparativestudi00ande_0116.txt

comparativestudi00ande_0032.txt

comparativestudi00ande_0073.txt

comparativestudi00ande_0183.txt

comparativestudi00ande_0043.txt

comparativestudi00ande_0054.txt

comparativestudi00ande_0083.txt

comparativestudi00ande_0098.txt

comparativestudi00ande_0122.txt

comparativestudi00ande_0095.txt

comparativestudi00ande_0068.txt

comparativestudi00ande_0162.txt

comparativestudi00ande_0085.txt

comparativestudi00ande_0203.txt

comparativestudi00ande_0047.txt

comparativestudi00ande_0033.txt

comparativestudi00ande_0167.txt

comparativestudi00ande_0005.txt

comparativestudi00ande_0114.txt

comparativestudi00ande_0179.txt

comparativestudi00ande_0108.txt

comparativestudi00ande_0188.txt

comparativestudi00ande_0092.txt

comparativestudi00ande_0081.txt

comparativestudi00ande_0069.txt

comparativestudi00ande_0177.txt

comparativestudi00ande_0170.txt

comparativestudi00ande_0151.txt

comparativestudi00ande_0045.txt

comparativestudi00ande_0015.txt

comparativestudi00ande_0142.txt

comparativestudi00ande_0136.txt

comparativestudi00ande_0046.txt

comparativestudi00ande_0071.txt

comparativestudi00ande_0181.txt

comparativestudi00ande_0185.txt

comparativestudi00ande_0027.txt

comparativestudi00ande_0141.txt

comparativestudi00ande_0105.txt

comparativestudi00ande_0072.txt

comparativestudi00ande_0102.txt

comparativestudi00ande_0036.txt

comparativestudi00ande_0031.txt

comparativestudi00ande_0058.txt

comparativestudi00ande_0079.txt

comparativestudi00ande_0034.txt

comparativestudi00ande_0094.txt

comparativestudi00ande_0172.txt

comparativestudi00ande_0040.txt

comparativestudi00ande_0064.txt

comparativestudi00ande_0178.txt

comparativestudi00ande_0021.txt

comparativestudi00ande_0124.txt

comparativestudi00ande_0165.txt

comparativestudi00ande_0197.txt

comparativestudi00ande_0044.txt

comparativestudi00ande_0161.txt

comparativestudi00ande_0111.txt

comparativestudi00ande_0053.txt

comparativestudi00ande_0194.txt

comparativestudi00ande_pdf.txt

comparativestudi00ande_0024.txt

comparativestudi00ande_0012.txt

comparativestudi00ande_0080.txt

comparativestudi00ande_0187.txt

comparativestudi00ande_0139.txt

comparativestudi00ande_0067.txt

comparativestudi00ande_0144.txt

comparativestudi00ande_0143.txt

comparativestudi00ande_0190.txt

comparativestudi00ande_0147.txt

comparativestudi00ande_0037.txt

comparativestudi00ande_0199.txt

comparativestudi00ande_0010.txt

comparativestudi00ande_0025.txt

comparativestudi00ande_0149.txt

comparativestudi00ande_0016.txt

comparativestudi00ande_0001.txt

comparativestudi00ande_0140.txt

comparativestudi00ande_0050.txt

comparativestudi00ande_0184.txt

comparativestudi00ande_0131.txt

comparativestudi00ande_0052.txt

comparativestudi00ande_0023.txt

comparativestudi00ande_0084.txt

comparativestudi00ande_0146.txt

comparativestudi00ande_0008.txt

comparativestudi00ande_0062.txt

comparativestudi00ande_0029.txt

comparativestudi00ande_0019.txt

comparativestudi00ande_0145.txt

comparativestudi00ande_0006.txt

comparativestudi00ande_0088.txt

comparativestudi00ande_0159.txt

comparativestudi00ande_0196.txt

comparativestudi00ande_0110.txt

comparativestudi00ande_0160.txt

comparativestudi00ande_0004.txt

comparativestudi00ande_0017.txt

comparativestudi00ande_0078.txt

comparativestudi00ande_0166.txt

comparativestudi00ande_0055.txt

comparativestudi00ande_0202.txt

comparativestudi00ande_0201.txt

comparativestudi00ande_0119.txt

comparativestudi00ande_0035.txt

comparativestudi00ande_0158.txt

comparativestudi00ande_0030.txt

comparativestudi00ande_0176.txt

E2HKGPUQO_O8JV2Q_xml.txt

comparativestudi00ande_0104.txt

comparativestudi00ande_0007.txt

comparativestudi00ande_0195.txt

comparativestudi00ande_0089.txt

comparativestudi00ande_0180.txt

comparativestudi00ande_0076.txt

comparativestudi00ande_0133.txt

comparativestudi00ande_0148.txt

comparativestudi00ande_0164.txt

comparativestudi00ande_0022.txt

comparativestudi00ande_0118.txt

comparativestudi00ande_0011.txt

comparativestudi00ande_0091.txt

comparativestudi00ande_0059.txt

comparativestudi00ande_0077.txt

comparativestudi00ande_0120.txt

comparativestudi00ande_0074.txt

comparativestudi00ande_0112.txt

comparativestudi00ande_0115.txt

comparativestudi00ande_0193.txt

comparativestudi00ande_0109.txt

comparativestudi00ande_0093.txt

comparativestudi00ande_0174.txt

comparativestudi00ande_0169.txt

comparativestudi00ande_0106.txt

comparativestudi00ande_0126.txt

comparativestudi00ande_0107.txt

comparativestudi00ande_0157.txt

comparativestudi00ande_0100.txt

comparativestudi00ande_0082.txt

ECWZOXEC0_JFEGJL_xml.txt

comparativestudi00ande_0168.txt

comparativestudi00ande_0182.txt

comparativestudi00ande_0051.txt

comparativestudi00ande_0018.txt

comparativestudi00ande_0127.txt

comparativestudi00ande_0117.txt

comparativestudi00ande_0070.txt

comparativestudi00ande_0056.txt

comparativestudi00ande_0156.txt

comparativestudi00ande_0039.txt

comparativestudi00ande_0103.txt

comparativestudi00ande_0075.txt

comparativestudi00ande_0128.txt

comparativestudi00ande_0020.txt

comparativestudi00ande_0152.txt

E913XNTSG_MJXEI5_xml.txt

comparativestudi00ande_0198.txt

comparativestudi00ande_0191.txt

comparativestudi00ande_0013.txt

comparativestudi00ande_0099.txt

comparativestudi00ande_0086.txt

comparativestudi00ande_0154.txt

comparativestudi00ande_0087.txt

comparativestudi00ande_0090.txt

comparativestudi00ande_0009.txt

comparativestudi00ande_0113.txt

comparativestudi00ande_0132.txt

comparativestudi00ande_0150.txt

comparativestudi00ande_0186.txt

comparativestudi00ande_0189.txt

comparativestudi00ande_0061.txt


Full Text














COMPARATIVE STUDIES OF THE
MORPHOLOGY AND ECOLOGY OF SEXUAL REPRODUCTION
OF Eichhornia crassipes (PONTEDERIACEAE)















By

ROBERT GERALD ANDERSON


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



























to my father,

for introducing me to plants and for his demonstration of personal sacrifice, hard work, and personal satisfaction

















ACKNOWLEDGEMENTS


The author wishes to express his appreciation to Dr, Terry W, Lucansky, Chairman of the Supervisory Committee, for his guidance and assistance in the conduct and fulfillment of this investigation and his assistance during the preparation of this manuscript, Sincere appreciation is expressed to the members of the supervisory committee, Drs. Mildred M. Griffith, John J. Ewel, Dana G. Griffin, III, and David L, Sutton for advice during the progress of this investigation and critical evaluation of the manuscript, I also thank Drs. Ariel Lugo and William T. Haller for evaluation of the manuscript.

Appreciation is expressed to the Florida Department of Natural Resources for permission to complete portions of this investigation on Payne's Prairie State Preserve and for funding the project under grant number 77785, The Botanical and Ecological Aspects of Aquatic Weed Control. Gratitude is also expressed to Dr, Donald A. Graetz, Soils Department, for the use of his laboratory facilities.

My greatest thanks go to my wife, Joanne, for the many rough copies and final copy of this dissertation which she gladly typed, for her financial support of my education, and for her understanding during the struggle of this dissertation.


iii
















TABLE OF CONTENTS

Page

ACKNOWLEDGEMENTS . . . . . . . . . . .iii

LIST OF TABLES . . . . . . . . . . . v

LIST OF FIGURES . . . . . . . . . . vii

ABSTRACT . . . . . . . . . . . . x

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

METHODS . . . . . . . . . . . . 9

RESULTS . . . . . . . . . . . . 24

Sexual Reproduction in Waterhyacinth Populations . 24

Phenology of Waterhyacinth Populations . . 24 Initiation of Flowering . . . . . . 30
Inflorescence Density . . . . . . 44
Inflorescence Development . . . . . 55
Pollination and Seed Set . . . . . 75

Anatomy of the Inflorescence of Eichhornia crassipes
and its appendages . . . . . . . . 80

DISCUSSION . . . . . . . . . . . . 123

Reproductive and Vegetative Characteristics of
Waterhyacinth Populations . . . . . . . 123

Anatomical Aspects of Eichhornia crassipes . . 155 SUMMARY . . . . . . . . . . . . 179

LITERATURE CITED . . . . . . . . . . 184

BIOGRAPHICAL SKETCH . . . . . . . . . 190


iv
















LIST OF TABLES


Table Page

1 Mean concentrations of nitrate nitrogen, ammonia
nitrogen, total organic nitrogen, and total
phosphorous at the field sites . . . . . 28

2 Mean temperatures measured at Melton's Pond and the
Gainesville Power Plant during the flowering of
waterhyacinths at the study sites . . . . 32

3 Precipitation, solar insolation, and daylength means
during the flowering of waterhyacinths at the study
sites . . . . . . . . . . . 34

4 Mean total leaf length of the third mature leaf of
established waterhyacinths . . . . . . 38

5 The number of senesced inflorescences on waterhyacinths and their ramets at Melton's Pond and the
Nelumbo Area . . . . . . . . . . 50

6 Mean sizes of inflorescences and their appendages of
introduced or established waterhyacinths at high,
intermediate, and low nutrient sites . . . . 60

7 Growth in length of a developing inflorescence as a
percentage of total inflorescence length of
introduced or established waterhyacinths at high,
intermediate, or low nutrient sites . . . . 64

8 Length of peduncle as a percentage of the total
inflorescence length of established or introduced
waterhyacinths at high, intermediate, or low nutrient
sites . . . . . . . . . . . 66

9 Comparative data of floral appendages of lower
flowers of the inflorescence of introduced or
established waterhyacinths at high, intermediate,
or low nutrient sites . . . . . . . 69

10 Comparative data of floral appendages of middle flowers of the inflorescence of introduced or
established waterhyacinths at high, intermediate,
or low nutrient sites . . . . . . . 70


v











Table _Page

11 Comparative data of floral appendages of upper flowers of the inflorescence of introduced or
established waterhyacinths at high, intermediate,
or low nutrient sites . . . . . . . 71

12 Mean length of the banner petal of waterhyacinth flowers borne in the lower, middle, or upper
positions on the inflorescences at high, intermediate, and low nutrient sites . . . . . 73

13 Means of the lamina length to lamina width ratios
of the third mature leaf from introduced waterhyacinths at high, intermediate, and low nutrient
sites . . . . . . . . . . . 144

14 Length of the petiole (as a percentage of the total leaf length) from the third mature leaf of introduced waterhyacinths at high, intermediate, and low
nutrient sites . . . . . . . .. ... 145


vi














LIST OF FIGURES

Figure Page

1 The location of study sites in the north-central
region of Payne's Prairie State Preserve . . . 11

2 Portable dock apparatus . . . . . . . 16

3 Mean and standard deviation of total leaf length of
introduced waterhyacinths at high, intermediate,
and low nutrient sites . . . . . . . 37

4 Mean and standard deviation of petiole width of the
third mature leaf of introduced waterhyacinths at
high, intermediate, and low nutrient sites . . 41

5 The number of inflorescences per m2 with open flowers
during anthesis at two low nutrient sites . . 49

6 Total number of stem apices and number of
inflorescences of waterhyacinths in the nonfertilized
culture tank . . . . . . . . . . 52

7 Total number of stem apices and number of
inflorescences of waterhyacintlis in the fertilized
culture tank . . . . . . . . . . 54

8 Inflorescence structure before anthesis, at anthesis,
and during postanthetic curvature . . . . 59

9 An inflorescence of Eichhornia crassipes 18 hours
before anthesis . . . . . . . . . 63

10 Flower of E. crassipes at anthesis . . . . 68 11 Comparative data on the orientation of the capital bend to the orientation of the outer bract . . 77 12 Structure of the outer and inner bracts of an inflorescence of E. crassipes . . . . . 32

13 Stamen filament of E. crassipes . . . . . 86

14 Transection of a peduncle distal to its insertion
on the stem . . . . . . . . . . 86

15 Transection of a vascular bundle of a peduncle . 86


vii












Figure Page

16 Transection of a peduncle proximal to the bases
of the inflorescence bracts . . . . . 86

17 Transection of the sheathing base of the outer inflorescence bract . . . . . . . 90

18 Transections of the fused margins of the sheathing
base of the inner inflorescence bract . . . 90 19 Transection of the aborted lamina of the inner
inflorescence bract . . . . . . . 90

20 Transection of the functional lamina of the outer
inflorescence bract . . . . . . . 90

21 Transection of the vascular plexus of a mature
flower of E. crassipes . . . . . . . 97

22 Transection of the origin of the vascular traces
of a mature flower of E. crassipes . . . . 97 23 Diagrammatic representation of the vascular traces
of the floral appendages of E. crassipes 99 24 Transection of the base of an ovary of a mature
flower . . . . . . . . . . . 97

25 Transection of an ovary between the base of the
carpels and the placentae . . . . . . 97

26 Transection of the placental region of an ovary 102 27 Transection of a flower distal to the apex of the
ovary . . . . . . . . . . . 102

28 Longisection of the convex side of the capital
bend in the peduncle . . . . . . . 102

29 Longisection of a peduncle before postanthetic
curvature . . . . . . . . . . 102

30 Longisection of a fertilized ovule three days
after pollination . . . . . . . . 109

31 Transection of a fertilized ovule twelve days
after pollination . . . . . . . . 109

32 Transection of a stem of E. crassipes . . . 109


viii












Figure Page 33 Transection of an adventitious root of E.
crassipes . . . . . . . . . . 109

34 Transection of a stem proximal to the insertion of an axillary bud . . . . . . . 114

35 Transection of a stolon of E. crassipes . . 114 36 Transection of a stem at the insertion of an inflorescence . . . . . . . . . 114

37 Transection of a stem distal to the insertion of an inflorescence . . . . . . . 114

38 Transection of a tertiary stem apex . . . 119 39 Transection of a tertiary stem apex . . . 119 40 Transection of an aborted lamina and stipule-flap of a prophyll . . . . . . . . . 119

41 Morphological variation in leaves of waterhyacinth 142


ix












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







COMPARATIVE STUDIES OF THE
MORPHOLOGY AND ECOLOGY OF SEXUAL REPRODUCTION
OF Eichhornia crassipes (PONTEDERIACEAE) By

Robert Gerald Anderson

December, 1976

Chairman: Terry W. Lucansky
Major Department: Botany


Ecological and morphological aspects of sexual reproduction of Eichhornia crassipes (Mart.) Solms were investigated at six, north Florida aquatic ecosystems on Payne's Prairie State Preserve. Phenological changes in the populations of established and introduced waterhyacinths were observed at regular intervals throughout a calendar year.

Nutrient concentrations, herbivory, and temperature were the

major influences upon both the growth and development of the waterhyacinth populations and the time scale of phenological changes. Increases in nutrient concentrations increased leaf size and inflorescence size of the waterhyacinth plants. Stress to the waterhyacinth population by cold temperatures or by herbivore infestations (Arzama densa) resulted in increased stolon numbers per unit area and


x











caused decreases in the size of leaves of waterhyacinth plants in a given population.

Floral initiation in the waterhyacinth occurred during periods of vigorous growth and relatively low nitrogen concentrations in natural and artificial conditions. Floral initiation in E. crassipes was not correlated with mean temperatures, photoperiod, total solar insolation, or rainfall. Densities of inflorescences fluctuated at both the study sites and in the greenhouse.

Stems and stolons (elongated first internode of axillary bud) of E. crassipes possess collateral vascular bundles in both the parenchymatous central cylinder and the cortex with both aerencymatous and parenchymatous zones. Stems have a distinct parenchymatous layer between the cortex and the central cylinder, whereas the cortex and central cylinder of stolons are not separated by a distinct parenchymatous layer. The adventitious roots have an aerenchymatous cortex with large, radially-arranged intercellular spaces and a polyarch stele.

The peduncle of the inflorescence is composed of aerenchymatous ground tissue with scattered collateral vascular bundles and a large central lacuna. The sheathing base of the outer inflorescence bract terminates in a well-developed lamina and a membranous stipule-flap,

whereas the sheathing base of the inner inflorescence bract terminates in an aborted lamina and a membranous stipule-flap. Three postanthetic bends are found on the peduncle of the inflorescence of E. crassipes, and differential cell elongation occurs on opposite sides of the peduncle.


xi











Eichhornia crassipes exhibits a sympodial growth pattern. The inflorescence terminates the primary shoot and vegetative growth continues from the ultimate axillary bud (secondary shoot). The anatomy, prophyll morphology, and development of a secondary shoot (after floral initiation) is dissimilar from that of a typical axillary shoot (during vegetative growth).

Sepals, petals, and carpels of the flower of E. crassipes are similar anatomically, and each floral appendage possesses three vascular traces that originate from the vascular plexus. The stamens have a single trace that arises from the median vascular trace of each perianth member.


xii















INTRODUCTION


Eichhornia crassipes (Mart.) Solms, waterhyacinth, has created economic problems in most tropical and subtropical regions of the world (Bock, 1966; Sculthorpe, 1967; Holm, 1969). The plant is well adapted to a freshwater habitat and typically forms an impenetrable mat that congests slow-moving riverine systems, reservoirs, lakes, ponds, and canals (Bruhl and Dutta, 1923; Penfound and Earle, 1948; Gay, 1960; Sculthorpe, 1967; Batanouny and El-Fiky, 1975). The mat can reduce commercial and recreational water traffic, clog sluice gates in dams, overtake rice paddies, reduce flow in irrigation, sewage and drainage canals, and cause flooding (Penfound and Earle, 1948; Sculthorpe, 1967). Floating mats of waterhyacinths also retard gaseous exchange between the water and the air and reduce oxygen concentrations and available nutrients (Penfound and Earle, 1948; Zeiger, 1962; Sculthorpe, 1967). Many methods of control chemical, biological, and ecological have been attempted without worldwide success because of the expense and inefficiency of control measures and the reproductive capabilities of the plant (Sculthorpe, 1967).

The waterhyacinth has many of the ecological adaptions for a

colonizing species and is considered one of the ten worst weeds in the tropics (Sculthorpe, 1967; Holm, 1969). The waterhyacinth population can maintain itself in various aquatic situations. When an


1







2


area is filled by waterhyacinths, competition between individuals becomesintense and the high growth rate is translated to morphological plasticity and allows individual plants to change their form for the specific environmental conditions (Rao, 1920a; Penfound and Earle, 1948). Hydrogen ion concentration has little effect upon the growth of waterhyacinths (Bock,1966; Gay, 1960; Haller and Sutton, 1973), although various concentrations of nitrate and phosphorous increase the vigor of waterhyacinths (Haller and Sutton, 1973; Morris, 1974). Morris (1974) and Boyd and Scarsbrook (1975) demonstrated that differences in total biomass, productivity, plant density, and root:shoot biomass ratios of plants from different sites were directly proportional to nitrate and phosphorous concentrations and the degree of development of native flora. Waterhyacinths require relatively high light intensities for optimal growth, whereas low light intensities result in the death of plants (Penfound and Earle, 1948). The plants can withstand low temperatures although a minimum temperature of -5'C for 24 hours has been reported to kill plants (Penfound and Earle, 1948). Penfound and Earle (1948) reported the doubling time for numbers of individuals from 11-15 days, depending upon the weather conditions, allows waterhyacinths to invade an open area very quickly. Although E. crassipes does not utilize dicarboxylic acid (C4) metabolism in photosynthesis, photosynthetic rates of the plant are comparable to those for C4 plants (Knipling et al., 1970).







3


The genus Eichhornia is one of seven perennial, freshwater genera in the family Pontederiaceae and occurs natively in the riverine systems of tropical South America (Castellano, 1958; Sculthorpe, 1967; Agostini, 1974). All Eichhornia species have a floating-emergent habit with erect or prostrate stems (Castellano, 1958). The stem (rhizome) of E. crassipes is erect in the water with a vertically oriented apex slightly below the water surface (Penfound and Earle, 1948).

Eichhornia crassipes typically has inflated petioles that

allow flotation on the water surface,whereas the remaining species have linear petioles and occur in shallow or transient freshwater systems (Castellano, 1958; Agostini, 1974). Nonswollen linear petioles may also occur in E. crassipes during high population densities, high temperatures, and low light intensities, whereas the swollen petioles are correlated with low population densities, low temperatures, and high light intensities (Rao, 1920a; LaGarde, 1930; Penfound and Earle, 1948). Arber (1918, 1920) termed the lamina of waterhyacinth and other members of the Pontederiaceae a pseudo-lamina and reported it was unifacial and represented an expanded flattened section of the cylindrical unifacial petiole. However, Kaplan

(1973) demonstrated that the "phyllode concept" is not valid for monocotyledons. le confirmed that the lamina and petiole of certain dorsiventral, laminar leaves are distinct and separate elaboration:; of a basal meristem that partially encircles the leaf primordium and that the distal leaf zone remains rudimentary. Laminas of






4


waterhyacinth leaves are probably not pseudo-lamina. Leaves of waterhyacinth appear to be morphologically similar to leaves of species described by Kaplan (1973). Bruhl and Dutta (1923) have described the anatomy of the rudimentary distal leaf zone in a young leaf of waterhyacinth. A stipule or proliferating leaf base originates proximal to the base of the leaf as a membranous ochrea that surrounds the younger leaves and apex (Weber, 1950). A membranous lamina (stipule-flap), which is only two cell layers thick and has a network of vascular tissue, occurs at the apex of the stipule and has mucilage cells along its margin (Weber, 1950).

The roots of the waterhyacinth are adventitious, nonbranched, and possess numerous, multicellular lateral roots (Penfound and Earle, 1948; Weber, 1950; Couch, 1971). Two forms of roots may occur on plants that are partially or wholly rooted in soil (Weber, 1950). The typical purple roots have a small diameter and possess the normal complement of side roots, whereas the atypical white roots are larger in diameter and have few lateral roots. Weber (1950) calculated the total length of a typical root plus its lateral roots to be 73 m

(surface area of 0.05 m2).

Asexual reproduction is the primary means of reproduction for the waterhyacinth (Bruhl and Dutta, 1923; Penfound and Earle, 1948; Sculthorpe, 1967; Das, 1969). Populations have been reported to double every 10-15 days through stolon production (Penfound and Earle, 1948; Batanouny and El-Fiky, 1975). An axillary bud undergoes development and the elongation of its first internode (between







5


prophyll and first leaf) carries the newly developed leaves, rooLs, and shoot apex into the water. This elongated internode (stolon) varies in length according to the population density (Das, 1967).

Agharkar and Banerji (1930) describe the inflorescence of

Eichhornia crassipes as terminal and growth of the individual plant continues from an axillary bud of the terminal leaf. The inflorescence is a spike (Penfound and Earle, 1948; Bock, 1966; Tag el Seed, 1972) although Singh (1962) termed it a contracted panicle. The individual, zygomorphic flowers are borne within two bracts that originate on the peduncle. The outer bract has a sheathing leaf base and a developed lamina at its apex, whereas the inner bract has a sheathing leaf base and an aborted lamina at its apex (Weber, 1950). A variable number of flowers are borne on the floral axis of the inflorescence (Penfound and Earle, 1948; Tag el Seed and Obeid, 1975). Each individual lavender flower has three sepals and three petals which are adnate for one-third of their length and the sepals are typically narrower than the petals. The banner (upper) petal differs slightly from the other two petals in size and has a bright yellow spot in the middle of a broad blue-purple field (Penfound and Earle, 1948; Singh, 1962). Stamens arise from the perianth tube and separate from it near the sinuses of the perianth. Three stamens are short in length and occur deep within the throat of the perianth, whereas the other stamens will be longer than the pistil or intermediate in length (Penfound and Earle, 1948; Singh, 1962; Tag el Seed and Obeid, 1975). The gynoecium consists of a conical ovary with many ovules






6


and a white to lavender style that varies in length (Bock, 1966) and terminates in a white, capitate stigma (Penfound and Earle, 1948).

The vascular anatomy of the flowers has been described by

Singh (1962). A broken ring of vascular tissue occurs at the base of each flower and each perianth lobe receives three traces. Each stamen has a single trace and three vascular traces proceed into each carpel. Median traces from the carpellary wall pass into the style and proceed to the six-lobed papillate stigma.

All flowers of the inflorescence usually open at the same time shortly after sunrise and close at or after dusk. After anthesis the inflorescence bends to the surface of the water (Rao, 1920; Agharkar and Banerji, 1930; Earle, 1947; Penfound and Earle, 1948; Weber, 1950). A bend begins to form below the base of the two bracts on the inflorescence and is completed 48 hours after anthesis. Additional bends occur in the floral axis between the flowers and at the base of the peduncle and force the closed flowers into the water. Variable-sized cells are reported to occur on the convex and concave sides of the bend (Weber, 1950; Das, 1967).

Sexual reproduction is not responsible for the rapid population development of E. crassipes but is a major factor in site recolonization (Penfound and Earle, 1948; Sculthorpe, 1967). Although large amounts of seed (1.1 x 108/ha, Zeiger, 1962) may be produced, the specific and complex germination requirements for the seeds are not common in perennial aquatic systems (Hitchcock et al, 1949; Tag el Seed, 1972).







7


After natural or artificial pollination the fruit of the waterhyacinth usually develops beneath the surface of the water, although development of the fruit can occur in the air (Agharkar and Banerji, 1930; Das, 1967). All flowers do not set seed under natural conditions. The highest percentage of fruit set is approximately 35% (Agharkar and Banerji, 1930), although the percentage is frequently much lower (Bock, 1966; Tag el Seed, 1972). Bock (1966) and Tag el Seed and Obeid (1975) have shown that pollen grains are viable and will germinate and grow in vitro. Natural pollinators and selfpollination account for natural seed set (Penfound and Earle, 1948; Bock, 1966; Tag el Seed, 1972; Tag el Seed and Obeid, 1975), although artificial pollination can be successful (Francois, 1964).

Seed germination occurs naturally on reflooded ground (Parija, 1930; Weber, 1950). Laboratory germination is successful when wet or dry seeds are placed in shallow water on a peaty soil with high light intensity and water temperatures of 28*-36'C (Hitchcock et al, 1949; Tag el Seed, 1972) or after scarification (Penfound and Earle, 1948; Hitchcock et al, 1949; Das, 1969). As the water level increases and the seedlings are submerged, the primary root abscises and the rootless seedling floats to the surface of the water (Parija, 1930). Swollen petioles and adventitious roots soon develop on the seedling (Parija, 1930; Penfound and Earle, 1948).

The objectives of this study are to determine the comparative morphology and anatomy of sexual reproduction of Eichhornia crassipes at ecologically different sites through a growing season. Aspects of











floral initiation and continuously flowering waterhyacinth populations are described in relation to changes in the vegetative morphology of individual plants and populations of waterhyacinths and environmental parameters. These data can be used to determine and analyze the consequences of sexual reproduction to the management and control of waterhyacinths in freshwater aquatic systems.
















METHODS


Observations of specific morphological and phenological

characteristics of Eichhornia crassipes Mart. (Solms) were made on Payne's Prairie State Preserve located 3 km south of Gainesville, Florida, and under controlled conditions in a greenhouse at the University of Florida.

Payne's Prairie State Preserve is the bottom of a shallow lake that drained naturally near 1900 and was subsequently diked and channelized to maintain high water levels. Great diversity of aquatic plant species occur in the deep, shallow, and transient water systems that intergrade throughout the prairie. However, waterhyacinths alone dominate the drainage canals. The following study sites differed in nutrient concentrations, water flow, and development of native aquatic species:

1. Melton's Pond. Melton's Pond is a large spring-fed pond

on the north ridge of Payne's Prairie (Figure 1). Its

oligotrophic water covers approximately one hectare with

a center depth of 3-4 m. Native aquatic macrophytes

Typha latifolia L., Pontederia cordata L., Panicum hemitomon Schult., Limnobium spongia (Bose) Rich., Nuphar

luteum (L.) Sibthorp and Smith subsp. macrophyllum

show a typical peripheral zonation around the pond.

Ceratophyllum demersum L. fills the remaining open water


9

























The location of study sites in the north-central region of Payne's Prairie State Preserve three km south of Gainesville, Florida. Arrows indicate direction of water flow. A., Melton's Pond; B., Nelumbo Area; C., New Canal; D., Main Canal; E., Biven's Marsh; F., Highway 441 Canal.


Figure 1














brundary
EN








Canal Alachua


C.


Hwy. B. A.

441 Canal







Canal



F Canal







12


of the pond, although during Spring and Summer the floating

plants Wolffiella floridana (Smith) Thompson and Utricularia spp. are also conspicuous in the open water.

Waterhyacinths do not occur naturally in this site. 2. New Canal. During January 1975, this canal (8-10 m

wide and 2-3 m deep) was completed to carry nutrient rich waters at a rate of 3.0-4.5 x 104 m3.day-l from

Sweetwater Branch and the Gainesville sewage treatment plant directly to the Alachua Sink, the drainage point

for Payne's Prairie (Figure 1). Vigorous waterhyacinths (80-100 cm tall) invaded and totally dominate this canal with the exception of scattered plants of Amaranthus spp, and Polygonum spp. which may grow upon the waterhyacinth

mat.

3. Nelumbo Area. In the center of Payne's Prairie (0.8-1.0

km south of the Alachua Sink) is a large area of mixed marsh and shallow (1-1.5 m) open water (Figure 1). The

American lotus, Nelumbo lutea (Willd.) Pers., is the dominant macrophyte in this area, but Typha latifolia L. and

Pontederia cordata L. invade the shallow water during times of drought. The submerged aquatics Najas flexilis (Willd.)

Rostk. and Schmidt and CeratophylJluim demersuin L. are also

distributed in the deeper water of this community. aterhyacinths (60-80 cm tall) are found along the eastern edge

of the community along a dike and form only a minor component of this ecosystem.






13


4. Main Canal. This canal, populated solely by vigorous

waterhyacinths (80-100 cm tall), carries much of the

drainage water from the marshes of the Prairie to the

Alachua Sink (Figure 1). After the sewage effluent and

water discharge from the Gainesville sewage treatment plant was diverted into the New Canal in January 1975,

the amount of water carried by the Main Canal decreased.

As of June 1975, the Main Canal was controlled by flood

gates and became a water storage area of the Prairie.

5. Biven's Marsh. On the north central boundary of Payne's

Prairie was a mixed marsh whose main channels were dominated only by moderately vigorous waterhyacinths (40-60 cm

tall)(Figure 1). This marsh had received the sewage

effluent from the Gainesville sewage treatment plant until the diversion of effluent into the New Canal which greatly

decreased the size of Biven's Marsh. Flow of water from

Biven's Arm in south Gainesville into Biven's Marsh

occurred at low volumes.

6. Highway 441 Canal. Along the west side of U.S. Highway

441, which traverses Payne's Prairie State Preserve from

north to south, was a canal 8-10 m wide and 2-3 m deep

(Figure 1). This canal was in a separate water flow system than those previously described and remained

relatively slow moving and stagnant except in periods of

heavy rainfall. Moderately vigorous waterhyacinths






14


(40-60 m tall) dominated the canal and were the only

species in the deeper water.

A portable dock apparatus was constructed and allowed an

individual to sample a given study site regardless of water depth. Repeated measurements were possible and the competitive environment of the mat remained unchanged. The portable dock apparatus (Figure 2) was constructed in two sections composed of a 3.7 m ladder and one or two blocks (1.2 m by 0.6 m by 0.4 m or 0.6 m by 0.6 m by 0.4 m) of flotation polystyrene. The polystyrene blocks weighed 0.013 kg.m-3 (1.1 lb-ft-3) and supported 0.763 kg.m-3 (60 lb.ft-3), were enclosed in canvas, waterproofed with a coat of latex paint, and attached to the end of the ladder with wide (5 cm) woven canvas strapping. The portable dock was lowered onto and raised from the waterhyacinth mat with a 1.8 m steel pipe as a lever and a 5-rope block and tackle. Two sections of the portable dock apparatus, when connected in series, allowed an area of 10 m2 of waterhyacinths to be measured in a transect. When set parallel to each other and joined by a 6.1 m aluminum ladder, the portable docks allowed a total area of 18 m2 to be observed with minimal damage to the waterhyacinth mat.

On 4 January 1975, ten plants of Eichhornia crassipes were introduced into pens (3 m x 1.5 in) of polyvinyl chloride (PVC) drainage pipe (10 cm diameter) at Melton's Pond, Nelumbo Area, and Biven's Marsh. These plants were located at each study site for replicate measurements. The introduced plants were propagated from
































Figure 2 Portable dock apparatus. A. Diagram of portable
dock composed of an aluminum ladder (3.7 m long), two polystyrene blocks, and lever with block-andtackle. B. Photograph of portable dock above the water surface and waterhyacinth population at the
New Canal.





16


A


B







17


young stolons (1-2 leaves and roots 1-5 cm long) obtained from the waterhyacinth population at Lake Alice on the University of Florida campus in November 1974, and cultivated inside the greenhouse in tanks that contained soil, peat moss, and 3-40 cm of water. After introduction into the study sites, the plants were protected from frost damage at night by a polyethylene cover. The dredging of the New Canal in January 1975, caused a change in water flow patterns on Payne's Prairie State Preserve and dictated that the plants from Biven's Marsh be moved to the New Canal 15 March 1975. The introduced plants were measured at biweekly intervals for the period between 4 January 1975 and 10 December 1975. Nightly freezing temperatures began 19 December and terminated the field experiments. The following morphological measurements and observations were made on six labeled introduced plants at each site:

1. Leaf size length and width of the lamina of the third

mature leaf from the apex per plant, length of the third mature petiole from origin of the stipule to the base of

the lamina per plant, and the width of the petiole at

its widest point per plant.

2. Number of leaves per plant produced between dates of

observation.

3. Inflorescence number and flower number per inflorescence

per plant.







18


The following morphological measurements and observations were made on the succeeding generations of ramets of the six labeled plants at each site:

1. Number of leaves per plant produced between dates of

observation.

2. Length of third mature petiole from origin of the stipule

to the base of the lamina per plant.

3. Inflorescence number per plant and flower number per

inflorescence.

At intervals of 4-6 weeks from January to December 1975,

established waterhyacinths were extracted in three samples of onesixth m2 from the New Canal, Nelumbo Area, Main Canal, Biven's Marsh, and Highway 441 Canal. The following morphological measurements and observations were made from six established waterhyacinth plants at each site:

1. Number of leaves with intact lamina and petiole or leaves

with senesced lamina and intact petiole, or leaves with senesced lamina and petiole and number of leaf bases on

the stem.

2. Leaf size length and width of lamina of third mature

leaf per plant, length of third mature petiole from origin

of stipule to base of lamina per plant, and the width

of petiole at its widest point per plant.

3. Length of the longest root from its origin to its tip.

4. Inflorescence number at pre- and postanthesis.







19


5. Flower and capsule number per inflorescences per plant.

These additional measurements were made on a per area basis at each site:

1. Number of full-sized plants (plants whose leaves formed

the canopy of the mat).

2. Number of small plants (plants whose leaves were below

the canopy).

3. Inflorescence number at pre- and postanthesis.

4. Dry mass mass of total sample of waterhyacinths after

drying 4-5 days at 700C in an oven.

Specific areas of Biven's Marsh, Highway 441 Canal, and

Nelumbo Area were marked with white stakes before flowering commenced and were photographed at 3-4 day intervals during flowering to obtain the number of flowers produced per area. Leaves of waterhyacinths typically grew taller than inflorescences at New Canal, Main Canal, and Nelumbo Area and prevented accurate counts at these sites.

The size of the inflorescence and its parts were measured at Melton's Pond and Nelumbo Area in the Spring and Fall, at Biven's Marsh in the Spring, and at the New Canal in the Fall of the year. Twenty-five labeled inflorescences along a transect were measured for five consecutive days through preanthetic development, anthesis,

and postanthetic curvature. The following inflorescence parts were measured before anthesis:







20


1. Length of peduncle from origin of stipule of the

terminal leaf to the bases of the inner and outer bracts

and width of peduncle immediately below the bases of

the bracts.

2. Length of outer bract lamina and length of sheathing

base of outer bract from the base of the bract to the

base of the lamina.

3. Length of sheathing base of inner bract from the base

of the outer bract to the base of the aborted lamina. At anthesis the following characters were measured:

1. Length of peduncle from origin of stipule of the

terminal leaf to the bases of the inner and outer bracts and width of peduncle immediately below the

bases of the bracts.

2. Length of outer bract lamina and length of sheathing

base of outer bract from the base of the bract to the

base of the lamina.

3. Length of sheathing base of inner bract from the base

of the outer bract to the base of the aborted lamina.

4. Length of the subfloral peduncle from the bases of the

inner and outer bracts to the base of the first flower and length of the interfloral peduncle from the base of the first flower to the apex of the inflorescence axis.

5. Total flower number and number of flowers open per

plant.






21


6. Flower size three flowers per inflorescence per

plant were measured at the base, middle, and top of

the interfloral peduncle. Length of perianth members

was measured from the point where petals and sepals separate. The length and width of the banner petal and lower sepal, the length of the pistil, and the

longest set of stamens were also measured in each

flower.

Following anthesis and postanthetic curvature, the following measurements were made:

1. Length of the peduncle from the origin of stipule of

the terminal leaf to the base of the bend below the

bracts and the length of the peduncle from the bases of

inner and outer bracts to the top of the bend.

2. Orientation of the bend relative to the lamina of the

outer bract.

Ambient air temperature was measured with a recording

thermograph that was mounted in a specially-built instrument box 0.4 m above the water level at Melton's Pond. Daily temperatures were recorded from 14 January 1975 to 20 December 1975. These data were supplemented by temperatures published in Climatological Data by the National Oceanic and Atmospheric Administration.

Rainfall measurements were taken from the Climatological Data by the National Oceanic and Atmospheric Administration. The standard daylength calculations of Dr. J.P. Oliver, Department of Physics and Astronomy, University of Florida, were utilized in







22


this study and weekly totals of solar insolation were averaged from daily totals summarized in the Environmental Data Summary produced by the Environment-Plant Interaction Studies group of the Fruit Crops Department, University of Florida.

A 250 ml water sample was extracted from the top 10 cm of surface water at each site at 4-6 week intervals throughout 1975. Samples were preserved immediately with phenyl mercuric acetate, put on ice while in the field, and stored at -20'C. Concentrations of nitrate nitrogen (NO3), ammonia nitrogen (NH3), total Kjeldahl organic nitrogen, and phosphorous (P) were measured after samples had been filtered through 201 filters. Nitrate nitrogen was measured colorimetrically to 0.1 mg/l by its reaction to chromotropic acid in the method described by West and Ramachandran (1966). The ammonia specific ion electrode with the Orion Pontentiometer, model 801, was utilized for NH3 determinations as low as 0.1 mg/l. A standard micro-Kjeldahl method for water samples was utilized to determine total organic nitrogen concentrations. Phosphorous was measured colorimetrically to 0.1 mg/l by the ascorbic acid method described in Standard Methods for the Examination of Water and Wastewater (1971).

Waterhyacinths used in greenhouse studies and introduced into the study sites were propagated from young stolons with 1-2 leaves and roots 1-5 cm long. These stolon cuttings were placed into wet soil in a shallow tank 50 cm deep and lined with polyethylene film to prevent desiccation. The soil (5-15 cm) was used as a nutrient







23


source and sink in the bottom of the tank and covered with peat moss for support of the cuttings.

On 22 February 1976, ten waterhyacinth plants were introduced into each of two shallow polyethylene-lined tanks (30 cm deep) that held approximately 300 1 of water and 8 cm of nutrient-poor soil. After introduction of the plants, two tablespoons soluble 20-20-20 commercial fertilizer with trace elements (Rapid Gro) was placed in each tank. When flowering commenced, two additional tablespoons of fertilizer were added to one tank, whereas no additional fertilizer was given to the other tank. The total number of plants and total number of inflorescences produced per plant were counted at weekly and 3-4 day intervals, respectively.

Inflorescences at preanthesis and postanthesis were collected from four field sites at times of intensive flowering during the Spring and Fall and preserved in a formalin-acetic acid-ethanol (FAA) solution. Stems, stolons, and adventitious roots were collected from a large culture tank maintained in the greenhouse. Infiltration and embedding were done under vacuum with a standard tertiary butyl alcohol series and 56.5*C wax (Johansen, 1940). Sections were cut at 10-12,j on a rotary microtome and stained with a safranin-fast green series. Preserved flower buds were dissected and mounted in floyer's solution for clearing (Anderson, 1954). A Zeiss M35 automatic camera was used to photograph the sections.

Duncan's new multiple-range test (Steele and Torre, 1960) was used to compare treatment means at a 95% confidence interval.















RESULTS


Sexual Reproduction in Waterhyacinth Populations Phenology of Waterhyacinth Populations

Small vigorous waterhyacinths with float leaves 10. cm long were introduced into Biven's Marsh, Nelumbo Area, and Melton's Pond on 4 January 1975. Original leaf margins and the swollen petioles appeared necrotic after two weeks and these leaves arched into the water and senesced by 19 February. New leaves (6.-8. cm long) with intense red coloration distally on the adaxial surface of short, bulbous petioles were produced in 2-4 weeks. All individuals at the study sites produced many stolons and the nine pens were 1/6-1/3 full (160-350 plants) by 19 March.

Changes in plant size were first noted at the New Canal (introduced waterhyacinths were moved to this site from Biven's Marsh after diversion of nutrient-rich water from Biven's Marsh to the New Canal) two weeks after waterhyacinths were introduced and the pens were filled by plants with float leaves by 16 April. By 15 May the elongated leaves of introduced plants at New Canal were slightly shorter (65 cm vs. 85 cm) than established plants at the site. After June 1 all plants at the New Canal remained large and vigorous throughout the year, and both introduced and established populations fluctuated similarly with environmental stresses.


24







25


The water level dropped rapidly during March at the Nelumbo Area because water flow patterns were changed and water and sewage input from Biven's Marsh and the central prairie were diverted to New Canal. By 8 April water depth at the Nelumbo Area had decreased

1 m, and sediments were exposed throughout the area by 17 April. Introduced waterhyacinths with small leaves and swollen petioles were rooted into moist sediments. In early May terrestrial species (especially Cyperus strigosus and Amaranthus spp.) invaded the open sediments and during May and June these terrestrial species formed a canopy over the introduced waterhyacinths. The return of standing water to the Nelumbo Area during June and July and the availability of additional nutrients from the oxidation of the sediments allowed the introduced waterhyacinths to increase rapidly in numbers and size through late June and July. Water depth at the site returned to 0.7 m by 8 August and introduced and established waterhyacinths assumed a floating habit again. The introduced waterhyacinths had been rooted in sediments in a single pen since March and on 8 August these free-floating plants were separated into three pens and formed large vigorous plants by 10 September that filled the pens until 10 December.

Leaf length of introduced waterhyacinths was less than 15.0 cm from the time of introduction until 6 August at Melton's Pond. All

plants had short leaves with bulbous petioles until 10 July. Afterwards all introduced plants produced long leaves (35.-45. cm) with narrow petioles.







26


Established populations of waterhyacinths at the Main Canal, the Nelumbo Area, Biven's Marsh, and the 441 Canal consisted of small individuals with leaf lengths of 20.-30. cm between 22 January and 14 February. Individuals at the Main Canal grew rapidly during March and April and established maximum leaf lengths (75. cm) by 24 April. Populations at the Nelumbo Area, Biven's Marsh, and 441 Canal reached maximum leaf lengths of 73. cm on 24 July, 53. cm on 28 October, and 51. cm on 29 July, respectively. Fluctuations in leaf size of the populations at all sites occurred at various times throughout the summer because of herbivory by Arzama densa or changes in the water flow patterns. Decreases of 5.-10. cm in leaf length were noted at all study sites during late November and early December when day and night temperatures were cool. Frost on 19 December damaged laminas and 1/4-2/3 of the petioles at all sites.

The larvae of the noctuid moth, Arzama densa, caused periodic and devastating damage to introduced and established waterhyacinth populations at all sites by devouring the stem apex and 1-5 cm of stem tissue. Plants at the Main Canal and the 441 Canal were infested in January when the plants were initially sampled. Heavy infestations in the waterhyacinth plants occurred in late May and June in New Canal and in late June and July in the Main Canal; only

dead or dying leaves were observed in the population's canopy at these times. Heavy infestations of A. densa also occurred at Biven's Marsh during July and August and at the 441 Canal and Melton's Pond during late August and September. During heavy infestations of






27


A. densa, all individuals of a plant population would be killed, but rapid vegetative reproduction by stolons allowed the population to remain intact. Introduced individuals lived only 4-6 weeks during heavy infestations at the New Canal but lived 6-12 weeks during lighter infestations at Nelumbo Area. Recovery from A. densa infestations and reestablishment of maximum sized plants required 3-5 weeks at the New Canal and Main Canal, and 6-10 weeks at Melton's Pond, Biven's Marsh and the 441 Canal. Lesser damage from the insect was noted at the Nelumbo Area, New Canal, Main Canal, and 441 Canal at other times of the year. Mites periodically damaged the foliage with light to heavy infestations, and grasshopper-like orthopterans grazed heavily on open flowers of introduced plants at Melton's Pond and Nelumbo Area.

Concentrations of nutrients in the water at the study sites differed significantly between the study sites during 1975. The mean concentrations of ammonia nitrogen (NH3-N) and total organic nitrogen (org-N) were significantly greater at the New Canal than the other study sites and the mean concentration of total phosphorous (P) was significantly greater at the New Canal than at the 441 Canal, Biven's Marsh, and Melton's Pond (Table 1). Mean concentrations of nitrate nitrogen (N03-N) were similar at all study sites. Levels of org-N (2.0-7.0 mg/l) and total P (2.2-4.4 mg/l) remained high throughout the year at the New Canal, although NH3-N and N03-N reached maximum concentrations of 20.0 and 0.7 mg/l, respectively, early in the year and remained constant at low levels after 6 August.






28


Table 1


Mean concentrations of nitrate nitrogen, ammonia nitrogen, total organic nitrogen (Kjeldahl), and total phosphorous at field sites on Payne's Prairie State Preserve from 22 January 1975 to 7 January 1976.


Nutrient Concentrations (mg/i)


Nitrogen Phosphorous Study Sites Nitrate Ammonia Organic Total New Canal 0.2 aX 7.4 c 4.9 c 3.3 b Main Canal 0.4 a 3.0 b 2.2 b 3.0 b Nelumbo Area 0.4 a 0.1 a 0.8 a 2.4 b 441 Canal 0.3 a 0.1 a 0.5 a 0.5 a Biven's Marsh 0.1 a 0.0 a 0.3 a 0.9 a Melton's Pond 0.2 a 0.0 a 0.1 a 0.2 a


same letter


x Means of nutrient concentrations not followed by the
are significantly different at the 5% level.







29


Mean concentrations of NH3-N and org-N were significantly greater at the Main Canal than at the Nelumbo Area, 441 Canal, Biven's Marsh, or Melton's Pond (Table 1). Mean concentrations of total P at the Main Canal were similar to mean concentrations of total P at the New Canal and the Nelumbo Area, but were significantly greater than total P mean concentrations at the 441 Canal, Biven's Marsh, and Melton's Pond. The maximum concentrations of 1.7 mg/l N03-N, 7.0 mg/l NH3-N, 6.0 mg/i org-N, and 5.7 mg/1 total P were measured in the Main Canal before the nutrient-rich water from the New Canal was diverted from the Main Canal in early July; only low concentrations of each nutrient were measured after July. Fluctuating water levels and exposed sediments at the Nelumbo Area caused varying amounts of N03-N (0.0-2.0 mg/1), org-N (0.1-2.2 mg/l), and total P (0.4-4.4 mg/i) before and after the period of drawdown. Mean concentrations of dissolved nitrogen NO3, NH3, and organic at the Nelumbo Area were similar to mean concentrations of dissolved nitrogen at the 441 Canal, Biven's Marsh, and Melton's Pond but mean concentrations of total P were significantly greater at the Nelumbo Area than these study sites (Table 1). Mean nutrient concentrations at the 441 Canal, Biven's Marsh, and Melton's Pond were similar during 1975 (Table 1). At the 441 Canal, N03-N and total P concentrations ranged from 0.0-0.9 mg/1 and 0.1-2.0 mg/l, respectively, but concentrations of NH3-N and org-N remained relatively constant throughout the year. Only the lowest detectable amounts of nutrients were measured at Biven's Marsh and Melton's Pond during the year, although slight increases in N03-N and total P occurred at both







30


sites early in the Spring. Because of the significant differences in nutrient concentrations among the study sites and the size of the plants at the sites, the study sites have been categorized arbitrarily into high, intermediate, and low nutrient sites. According to this classification, the New Canal and Main Canal were the high nutrient sites, Nelumbo Area was the intermediate nutrient site, and the 441 Canal, Biven's Marsh, and Melton's Pond were the low nutrient sites. Initiation of Flowering

Several characteristics of flowering in waterhyacinths were initially noted in this study. Waterhyacinths flowered most abundantly in the Spring and Fall in the vicinity of Gainesville, Florida. Flowering exhibited periodicity and zonation patterns at a single site or among sites. Large numbers of flowers opened each day at a given site and were replaced by new flowers the next day. Waves of extensive flowering moved slowly along a canal or formed a peripheral zone around a body of water.

The time of initial flowering and cessation of flowering was variable between and within the study sites. Flowers were first noted in early April in the established waterhyacinth populations at the low nutrient sites (Biven's Marsh and 441 Canal). However, flowers did not occur at the intermediate nutrient site (Nelumbo Area) until mid-May or at the high nutrient sites (New Canal and Main Canal) until late August. Introduced waterhyacinth populations at the low nutrient site (Melton's Pond) initiated flowering in late April, whereas flowers were first observed in the introduced







31


waterhyacinth populations at the intermediate nutrient site in late May and at the high nutrient site in late August. Flowering continued in the observation areas of established waterhyacinths at the low nutrient sites until mid-June (Biven's Marsh) or late July (441 Canal) although continued flowering occurred at these sites outside the observation areas until late August. Established waterhyacinths at the intermediate nutrient site flowered at varying densities throughout the summer and flowering ended in early November. Flowers were observed in the established waterhyacinthsat the high nutrient sites until mid-October (New Canal) or mid-November (Main Canal). Cessation of flowering in the introduced waterhyacinths at the low and intermediate nutrient sites occurred in early November after periodic flowering had been observed throughout the Summer and Fall. Flowers were observed only until mid-October in the introduced waterhyacinths at the high nutrient site.

Environmental parameters influenced flowering at all study sites. Mean air temperatures at Melton's Pond ranged from 14-22*C in early Spring and late Fall, 24-26'C from mid-April to late June, and remained between 22-24*C during the Summer and early Fall (Table 2). High and low temperatures at Melton's Pond were mediated by the temperature stabilization effects of the water because air temperatures measured at the Gainesville Power Plant were lower in the Winter and higher in the Summer than the temperatures at Melton's Pond. Total precipitation was intermittant and light (0-45 mm) from February to late May, normal (9-116 mm) from late May to late September, and intermittant and light (0-25 mm) through October







3 2


Table 2 Mean temperatures measured at Melton's Pond and the
Gainesville Power Plant during the flowering of waterhyacinths at the study sites.


Mean Temperature 0C
Melton's Pond NOAA

Date High Low Mean Mean
Feb. 7 23 10 17 15
14 27 17 22 21 21 23 12 18 15 28 20 7 14 13 Mar. 7 26 15 21 20
14 25 12 19 20 21 26 15 21 23 28 28 16 22 18 Apr. 4 27 13 20 18
11 26 14 20 20 18 28 17 23 23 25 30 17 24 25 May 2 29 15 22 25
9 28 18 23 24
16 29 17 23 26 23 31 18 25 26 30 31 19 25 27 June 6 32 20 26 27
13 32 20 26 27 20 28 19 24 27 27 29 18 24 25 July 4 29 19 23 27
11 26 19 23 27 18 29 19 24 26 25 28 20 24 27 Aug. 1 29 18 24 28
8 28 18 23 27 15 29 18 24 27 22 28 18 24 18 29 27 17 22 28 Sept.5 28 18 23 28
12 27 17 22 27 19 26 17 22 26 26 26 17 22 24 Oct. 3 29 20 25 26
10 29 17 23 24 17 -- -- -- 21


24 31
Nov. 7
14 21 28


13 18 17
9
8
13


19 23
21 16
14 15


23
22 22 15 13







33


and November (Table 3). The official total precipitation for Gainesville (1250 mm) was approximately 100 mm below normal and the water levels were observed to decrease slowly at all study sites during the year. Weekly totals of solar insolation ranged from 2.85-3.95 Kcal-cm-2 day in April and June, 2.32-3.25 Kcal. cm-2 day-1 in March, July, August, and September, and 1.33-3.09 Kcal-cm-2 day-l in February, October, and November (Table 3). Daylengths ranged from 11 hours in early February to 14 hours

3 minutes at the summer soltice in June and returned to 10 hours 24 minutes in late November (Table 3).

Mean temperatures ranged from 20-240C and mean high temperatures ranged from 27-31*C on the dates that flowers were first noted at the low nutrient sites (Biven's Marsh and 441 Canal 2 April, Melton's Pond 30 April, the intermediate nutrient site (the Nelumbo Area 14 May, and the high nutrient sites (Main Canal and New Canal 20 August. Mean low temperatures (13-14*C) at Melton's Pond were lower in early April when flowering was first observed than the mean low temperatures (17-18'C) on dates of floral appearance at the other study sites (Table 2). Very low amounts of precipitation (0-9 mm) were recorded for four weeks before initial flowering at Biven's Marsh and 441 Canal, although rainfall was normal (12-37 mm) prior to initiation of flowering at the other study sites (Table 3). Amounts of total solar insolation were similar (approximately 3.0 Kcal.cm-2 day-1) were recorded when flowers first appeared at the intermediate nutrient site (Table 3). Daylengths were different






34


Table 3


Precipitation, solar insolation, and weekly daylength means during the flowering of waterhyacinths at the study sites.


NOAA Total Total solar Mean
precipitation Insolation Daylength
Date mm Kcal-cm-2.wk-1 hours: minutes


Feb. 7
14 21 28
Mar. 7
14 21
28
Apr. 4
11 18 25
May 2
9
16 23 30
June 6
13
20 27
July 4
11 18 25
Aug. 1
8
15
22 29
Sept. 5
12
19 26
Oct. 3
10 17
24 31
Nov. 7
14 21 28


38.
2.
40.
11. 0.
9.
4.
0.
0.
44. 12. 37. 3.
44. 18. 1.
27. 27.
21. 20. 29.
42. 76. 38. 9.
49. 22. 15. 15.
12. 98.
28. 116.
9.
25. 0.
26.
20. 0.
25. 0.
2.
16.


1.66 2.26 2.80 3.13 2.90 3.25 2.86 3.12 2.85 3.05 3.85 3.53 3.26 3.13 3.96 3.08 2.89 3.01 3.51 3.22 2.98 3.09 3.22 3.39 2.32 2.58 2.96 2.89
3.48 2.65
2.43 2.17 2.70 1.91 3.09 2.80 1.76
2.40 1.33 2.63 2.38 2.06


11: 01 11:12
11:24 11:36 11:49
12:02 12: 14 12:26 12:38
12:51 13 : 02 13:14 13: 25 13:34 13:43 13:50 13:56 14:00 14: 03 13:03
14: 02 13:59 13:54 13:47 13:40 13:32 13:20 13: 08 13:57
12:45 12:34
12:21 12:09 11: 56 11:45 11: 30 11:18 11:07 10:56 10:48 10:38 10:30 10:24







35


and ranged from 12 hours 30 minutes to 13 hours 48 minutes at the time of initiation of flowering at the study sites (Table 3).

Inflorescence production occurred in waterhyacinth populations

at all study sites regardless of the distinct differences in vegetative characteristics that occurred among the study sites. Introduced waterhyacinth plants at the high nutrient site, New Canal, possessed mean leaf lengths of the third mature leaf (maximum length 97. cm) that were significantly greater than mean leaf lengths (maximum length 51. cm) at the low nutrient site, Melton's Pond (Figure 3). Mean leaf lengths of introduced waterhyacinths at the Nelumbo Area (intermediate nutrient site) were highly variable (maximum lengths ranged from 45-85. cm) and intermediate between the mean leaf lengths measured at the New Canal and Melton's Pond (Figure 3). Waterhyacinth leaves produced by established plants were significantly longer at the high nutrient sites, New Canal and Main Canal, (maxima of 102 cm and 91 cm, respectively) than at the low nutrient sites, 441 Canal and Biven's Marsh, (maxima of 55 cm and 59 cm, respectively) (Table 4). Means of total leaf lengths of established waterhyacinths at the intermediate nutrient site (Nelumbo Area) were similar to the low nutrient sites from 22 January to 15 May but were similar to the high nutrient sites from 24 July to 5 December (Table 4). Minimum

leaf lengths occurred in the populations of established waterhyacinths from 22 January to 24 April when the canopy of leaves was disrupted by frost or cool temperatures. At various times during the year the canopy of both introduced and established waterhyacinth populations was disrupted by extensive damage from Arzama densa which also



























Figure 3 Mean and standard deviation of total leaf length (petiole plus lamina
of the third mature leaf) of introduced waterhyacinths at high, intermediate, and low nutrient study sites at Payne's Prairie State Preserve
from 7 January 1975 to 10 December 1976. Dates when maximum damage
from Arzama densa occurred to the plants are indicated by X.










90


80- 0 Melton's Pond (low)
0 Nelumbo Area (inter) x 70. -O New Canal (high) 70


x

60 X 300



40.



30



20



10 ...-100



8 22 5 19 5 19 2 16 30 14 28 11 25 9 23 6 20 3 17 1 15 29 12 26 10 J F M A M J J A S 0 N D
Date of Observation







38


Table 4 Mean total leaf length (lamina plus petiole) of the
the third mature leaf of established waterhyacinths
at five sites on Payne's Prairie State Preserve.


Date

Jan. 22, 29 Feb. 7, 14 Feb. 21, 28 Mar. 20, 27 Apr. 17, 24 May 15, 22 July 24, 29 Aug. 24, Sept. 9 Sept. 29 Nov. 6, Oct. 28 Dec. 5, Nov. 24


Leaf Length (cm)

New Main Nelumbo 441 Biven's Canal Canal Area Canal Marsh


61.3 65.2 93.3 56.7 72.6 90.9 81.3 67.9


cY

b

d

bcX

bX

b

d

b


26.7 32.2

42.4 57.6

74.3 75.4 64.1 66.3



73.5

64.9


ax

b

c

c

b

c

bcX

b



cd

b


19.5 a 15.0 a 23.6 ab 9.6 a 18.7 a 45.9 b 72.9 c 70.3 b 65.3 a 66.5 bc 55.0 b


29.3 19.9 27.2 13.8

19.4 34.4 50.6

40.0



35.6 32.9


aX ab

b

ab

a

a

b

aX



aX aX


26.8 ab 14.6 a 24.2 b 23.4 a 33.7 a 34.1 ax 39.8 ax



52.8 b

40.1 a


v Means of each date not followed by the same letter are signicantly different at 5% level.

x Dates when maximum damage caused by Arzama densa occurred to
the plants.


'A


--







39


decreased mean total leaf lengths. Maximum total leaf lengths occurred at all sites after canopy continuity was reestablished following frost or A. densa damage. Additionally, mean total leaf lengths of introduced waterhyacinths were at maxima 6-8 weeks after pens at the study sites New Canal (17 April), Nelumbo Area (15 May), and Melton's Pond (24 July) were filled completely with plants (Figure 3).

Mean petiole and lamina length and lamina width of the third mature leaf of introduced waterhyacinths also varied according to the nutrient concentration of the sites. Mean petiole lengths (48.9-64.0 cm), lamina lengths (12.9-18.5 cm), and lamina widths (12.7-16.1 cm) were significantly larger at the high nutrient site, New Canal, than mean petiole lengths (3.5-34.3 cm), lamina lengths (3.0-9.1 cm), and lamina widths (4.0-7.3 cm) measured at the low nutrient site, Melton's Pond. Introduced waterhyacinths at the intermediate nutrient site possessed leaves whose means of petiole length (12.8-53.2 cm), lamina length (4.3-12.7 cm), and lamina width (4.5-10.6 cm) were intermediate between or similar to leaf dimensions at Melton's Pond or the New Canal.

Petiole width of leaves of introduced waterhyacinths remained constant or decreased during the year, whereas petiole lengths, lamina lengths, and lamina widths on the same plants increased during the year. Decreases in petiole width at the high nutrient site (New Canal) and the low nutrient site (Melton's Pond) occurred in April and July, respectively, concurrently with increases in petiole length at both sites (cf. Figure 3,4). Mean petiole widths



























Figure 4 Mean and standard deviation of petiole width of the third mature
leaf of introduced waterhyacinths at high, intermediate, and low
nutrient study sites at Payne's Prairie State Preserve.












-35
35- o 'Melton's Pond (low) 0 Nelumbo Area (inter)

30- 0 New Canal (high) 30




25- < 25


e --0--20- 0- 20

0


155
15- AI




10 -I--- .-0




5 5


I6 I I I I I
8 22 5 19 5 19 2 16 30 14 28 11 25 9 23 6 20 3 17 1 15 29 12 26 10 J F M A M J J A S 0 N D
Date of Observation







42


of introduced plants were similar at both high and low nutrient sites from January until July when petiole length increases occurred at the low nutrient site (Figure 4). Mean petiole widths were significantly greater at the high nutrient site (New Canal) than at the low nutrient site (Melton's Pond) after 6 August. At the intermediate nutrient site (Nelumbo Area), mean petiole widths of introduced waterhyacinths were similar to the other sites during the Spring, variable and small during the drawdown at that site, and intermediate between the high and low nutrient sites after water returned to the Nelumbo Area during June and July. Nutrient levels at each site and population changes caused by A. densa were responsible for variations in the mean petiole widths of established waterhyacinths. Statistical comparisons between high and low nutrient sites were difficult because the phenologys of the established waterhyacinth populations were not synchronous. Mean petiole widths of short, swollen petioles were similar at all sites, whereas elongated, nonswollen petioles typically were significantly larger at the high nutrient sites (New Canal and Main Canal) than at the low nutrient sites (441 Canal and Biven's Marsh).

The rate of leaf production varied greatly at each site

throughout the year although the number of functional leaves per plant remained constant. The production of leaves ranged from

0.0-0.5 leaves per plant per week when the introduced waterhyacinth populations were stressed by A. densa, cool temperatures, or floral development; however, rates of 1.5-2.0 leaves per plant per week were measured when the waterhyacinths were not stressed. Average







43


leaf production rates of 0.92, 0.81, and 0.75 leaves per week were measured at the New Canal, the Nelumbo Area, and Melton's Pond, respectively, during the year. Waterhyacinth plants growing under stable conditions typically possessed 6-7 leaves with intact petioles and laminas. This number remained constant in established waterhyacinth populations at high, intermediate, and low nutrient sites, but decreased during A. densa infestations and cool temperatures.

The maximum root lengths of waterhyacinth plants established

at each of the five study sites varied with changes in the percentage of stolons per m2 and the nutrient concentration at each site. Younger plants recruited into the waterhyacinth populations reduced maximum root lengths because the young plants had shorter root systems than the established plants. Maximum root lengths in stable waterhyacinth populations in the low nutrient sites (441 Canal and Biven's Marsh) ranged from 17-58 cm; however root lengths at the high nutrient sites (New Canal and Main Canal) were shorter with maximum lengths of 5-20 cm.

Parameters of established waterhyacinth populations, plant density and total dry mass,were not closely related and these parameters varied independently among the study sites because phenological changes at all sites were not simultaneous. Definite increases in the number of stem apices per m2 were noted during the Spring and during A. densa infestations at all sites because of increases in the stolon number per m2. Total dry mass also increased during the Spring but no increases were noted during A. densa infestations. Relatively low plant densities of established







44


waterhyacinths occurred at each site between A. densa infestations and these populations were characterized by low stolon numbers per m2 and large plant size. The plant densities noted during these periods of population stability and vigorous growth were low at the high nutrient sites (53-89 apices m-2), intermediate at the intermediate nutrient site (89-109 apices m-2), and high at the low nutrient sites (97-126 apices m-2). Although plant densities varied among sites with different nutrient concentrations, total dry mass (1.0-1.9 kg m-2) seemed similar at all study sites.

Variations in vegetative characteristics of a waterhyacinth population at a single site had no effect upon inflorescence production at that site. Inflorescences were initiated at Melton's Pond in late April when plants were small and plant density was low and inflorescence production continued through the Summer and into November when plants were large and plant density was high. Changes in vegetative characteristics because of A. densa damage was related, however, to cessation of flowering at both the 441 Canal and Biven's Marsh. Few changes in vegetative characteristics occurred in waterhyacinth populations at high and intermediate nutrient sites and these changes were not related to inflorescence initiation or cessation.


Inflorescence Density

The maximum number of inflorescences with open flowers per m2 of surface area of established waterhyacinths varied among the study sites. Correlation of inflorescence numbers with nutrient







45


levels was impossible because the inflorescences were below the canopy and not visible at the high nutrient sites (New Canal, Main Canal) and the intermediate nutrient site (Nelumbo Area) for all or most of the year. The density of inflorescences with open flowers from established waterhyacinths reached maxima of 4.1-4.8 inflorescences per m2 and 6.3-7.8 inflorescences per m2 at the low sites (Biven's Marsh and 441 Canal, respectively)(Figure 5) and 1.6 inflorescences per m2 at the intermediate nutrient site while inflorescences were visible in May. A maximum of 3.3 inflorescences per m2 was observed in the introduced waterhyacinths at the low nutrient site (Melton's Pond) during late July. During the late Spring and early Summer (25 April-10 June) the established waterhyacinths at Biven's Marsh produced 71 inflorescences per m2 of plants (approximately 120-140) and averaged 1.6 inflorescences per m2 per day. Established waterhyacinths at the 441 Canal produced 175 inflorescences per m2 of plants (120-160) for an average of 2.5 inflorescences per m2 per day during this same time (25 April to

1 July).

Periodicity and zonation patterns of flowering were observed in established waterhyacinths from Biven's Marsh and 441 Canal (Figure 5). Maximum density of inflorescences with open flowers occurred 14 and 18 days apart (8 May and 22 May at Biven's Marsh; 12 May and 30 May at 441 Canal) at these two sites. Intensive flowering moved slowly along the main channel of Biven's Marsh. A high density of inflorescences occurred in the observation area of Biven's Marsh during May, although a high density of inflorescences




























The number of inflorescences per m2 with open flowers during anthesis at two low nutrient sites, the 441 Canal and Biven's Marsh, on Payne's Prairie State Preserve.


Figure 5














8 Biven's Marsh

o 441 Canal

7- -7



6- -6

0)


4 5- -5











r- .- -4
25 29 3 1/ 1 1 3 2 1 4 8 1 6 2 42
Q) -rCL .0








A 0
1.Db


00

25 29 3 7 11 15 1'9 23 2 '73 31 4 8 12 16 20 24 28 1 5 9
A Mj
Dates of Observation







48


was not observed in the marsh itself until mid-June. The marsh was located 150-200 m to the east of the channel. Flowering at the observation site on the main channel of Biven's Marsh had ceased when the density of flowering was greatest in the Marsh, even though the waterhyacinth population and water flow were continuous between the two locations.

The density of inflorescences with open flowers in established waterhyacinths in the 441 Canal was lower to the north and to the south of the observation area in May. The density of flowering increased at various locations along the 441 Canal through June, July, and August, although few flowers were observed at the study area. Damage from A. densa was widespread along the 441 Canal during September and October and flowering ceased.

Introduced waterhyacinths did not flower synchronously in all pens at Melton's Pond. Flowering began in one pen shortly before

1 May and had produced 82 inflorescences by 29 May; however, the other two pens at the site produced only 55 and 12 flowers during May. From 1 May to 7 November totals of inflorescences with open flowers varied in each pen, and each pen had its own periodicity (ranging from 15-25 days) of inflorescence production. No periodicity for density of inflorescences was observed in the established waterhyacinths at the high and intermediate nutrient sites, although zones of high densities of inflorescences occurred during September among the smaller plants that occurred in the mat after A. densa damage. Scattered inflorescences were also observed from May until October on plants rooted in the soil along the dikes on the margins of the







49


established waterhyacinth populations at the high and intermediate nutrient site.

Production of inflorescences by waterhyacinth plants and their ramets followed a random and periodic pattern. Although the genotypes were identical and the plants were in close proximity of each other, similar numbers of inflorescences did not occur in all plants of the same clone during the same period (Table 5). Under the same conditions more flowers were produced by clone #6 of the introduced waterhyacinths at Melton's Pond than by plants of clone #1 during May and June (Table 5). Plants of clone #1, however, produced more flowers than clone #6 from 10 June to 29 September. Similarly, introduced waterhyacinths labeled #4-1 and #5-1 at the Nelumbo Area produced 5 and 4 inflorescences, respectively, from 4 September to 7 November whereas plant #1-1 produced only 1 inflorescence (Table 5)'. Plant #1-1 and its six ramets (Melton's Pond) produced 16 inflorescences between 10 July and 29 September of which plants #1-1, #1-5, and #1-6 produced only one inflorescence and plants #1-2 and #1-7 produced 4 inflorescences each during the same time period.

The initiation and periodicity of flowering were observed in populations growing under different nutrient regimes in a greenhouse experiment. After a month in the initially fertilized water (two tbsp. 20-20-20 fertilizer), the ten initial plants in each tank had produced approximately 100 plants that showed signs of chlorosis (Figures 6 and 7). After the addition of more fertilizer (2 tbsp. 20-20-20 on 29 March to one tank) chlorotic symptoms disappeared from the plants, the number of plants increased to 257, and no








50


Table 5


The number of senesced inflorescences on waterhyacinths and their ramets at Melton's Pond and the Nelumbo Area on Payne's Prairie State Preserve.


Plant number

Melton's Pond Nelumbo Area (low nutrients) (intermediate nutrients) Date 1 6 1 4 5

May 1 1,lx
15 2,2
29
June 12 1,2,2,3 July 10 1,2,3 1,2,3,3,3,4
24 3,4,5 2 1 Aug. 7 2,6,7 3,4 ly
22 7 1,1
Sept. 4 3,4 1,2 1,1,2,2,2 1,1,2,2
29 2,2,7,7y 2y 2 3
Oct. 15 1 1 1,2,2,3,3 Nov. 7 1,1,2,2 1,2,3


x Each number represents one inflorescence
plant (1) or its ramets (2-7).


produced by the primary


y The individuals present on this date were destroyed by Arzama by
the following date and new individuals were labeled.





























Figure 6 Total number of stem apices and number of inflorescences of
waterhyacinths in the nonfertilized culture tank.














Number of stem apices


0 0


0


per culture


o I-.


-A


C.4 -j4-


0:




Ln


U,


Ca


Percentage of plants


I-. a


producing inflorescences


tank

k)
0
0


oo













**













/ C









0.e~



.--- ~


h, a-


eT
(D
0





0 (0


'a











Ta 'a


)-


0 (D

(f)0 C)
(D)


0


(D E (0


r/i





























Figure 7 Total number of stem apices and number of inflorescences of
waterhyacinths in the fertilized culture tank.















Number of stem apices per culture tank


CD C)


C


C


C)
o


C


* 0%


















*%
I -s

0 0













*,
/




























I 0
0 0



























IV (
a p eSo
/ -j


e oj

Percentage of plants producing inflorescences



11 -


rJ-


t (D
0

0 (D




0


I.
0 -


0


rNj
t-n


i







55


inflorescences were formed until 15 May. Chlorotic symptoms continued in the waterhyacinth plants in the nonfertilized tank and maxima of 5-10% of the plants produced inflorescences periodically from 5 April to 7 June (Figure 6). After a reduction of the number of plants in the fertilized tank on 19 April, the number of plants increased initially to 125 on 3 May and then slowly decreased. The leaves of the plants in the fertilized tank became chlorotic and maxima of 22% and 12% of the plants produced inflorescences in late May and early June (Figure 7). A total of 119 inflorescences were produced periodically in the nonfertilized tank in 90 days, whereas a total of 135 inflorescences were produced periodically in only 20 days in the fertilized tank. Waterhyacinth plants in the fertilized tank typically produced two inflorescences per plant whereas plants in the nonfertilized tank produced only one inflorescence per plant. The same plants in the nonfertilized tank typically did not flower during successive periods of the maximum densities of inflorescence.

The size and density of the waterhyacinth plants increased with the amount of nutrients added to each tank. Leaf lengths of the waterhyacinth plants in both culture tanks ranged from 15-20 cm shortly before 29 March, but the leaf lengths in the fertilized tank increased to 30-40 cm after fertilizer was added 29 March.

Leaf lengths of the waterhyacinths in the nonfertilized tank decreased during May and ranged from 6-12 cm in early June. Inflorescence Development

The stages of development of the inflorescence of the







56


waterhyacinth occurred in the same sequence and over similar periods of time at all study sites. In this study the developing inflorescence was typically macroscopically visible 6-7 days before anthesis (Figure 8). The lamina of the outer bract appeared first and initially was tightly appressed to the stipule-flap of the previous leaf. The lamina of the outer bract did not increase in size prior to or after anthesis. Four days before anthesis the bases of the bracts were visible. Daily elongation of the peduncle proximal to the bases of the bracts was greater than the daily elongation of the outer bract sheath. The latter reached its maximum length 24-48 hours before anthesis, although the inner bract sheath and enclosed colorless flowers did not fill the cavity formed by the outer bract. The aborted lamina located terminally on the inner bract sheath

grew, through a distal slit in the outer bract sheath 24 hours before anthesis. The peduncle length proximal to the bases of the bracts was 80-90% of its total length at anthesis, whereas the diameter of the peduncle 1.0 cm proximal to the bases of the bracts remained constant prior to and after anthesis.

Between 8 am and 4 pm on the day before anthesis the inner bract and enclosed flowers pushed through the slit in the outer bract sheath. Between 4 and 6 pm the inner bract sheath reached its maximum size and colored flowers were visible through the transparent

membranous portion of the sheath. The peduncle from the bases of the bracts to the apex of the inflorescence elongated after 7 pm, pierced the inner bract sheath, and reached its maximum length approximately midnight. The maximum length of the subfloral and





























Figure 8 Inflorescence structure before anthesis, at anthesis, and during
postanthetic curvature.

bb basal bend; cb capital bend; ifb interfloral bend















- 1 Day


- 5 Days


cb



+ 3 Days








ifb


Anthesis


+ 12 Hr.b








b b


- 3 Days








59


interfloral sections of the peduncle was attaLned unless all flowers on the inflorescence were not set to open the following morning. The distal 1-10 flowers of the inflorescence frequently did not open on the same day as the remaining flowers of the inflorescence, although no relationship existed between the number of flowers present and the number of flowers that open the first day. The terminal 1-3 flowers usually did not open with the remaining flowers of the inflorescence regardless of the number of flowers per inflorescence. At midnight the flowers were oriented vertically and tightly appressed to the interfloral peduncle. By sunrise on the day of anthesis all flowers that would open that morning were separated from each other and oriented at a 45-60* angle from the axis of the inflorescence. All flowers at a site opened synchronously 1-3 hours after sunrise, although shade or cloud cover delayed flower opening 1-2 hours.

Mean sizes of the inflorescences and their appendages and

flower number per inflorescence varied greatly among the study sites and were separated into four size classes. Mean sizes of the inflorescences and their appendages and the mean flower number per inflorescence at the high nutrient site (New Canal) were significantly larger than at the low nutrient sites, and similar to or significantly larger than the intermediate nutrient site (Table 6). The mean sizes of the inflorescences and their appendages and the mean flower number per inflorescence were generally smaller at the low nutrient sites (Melton's Pond and Biven's Marsh) than at the intermediate nutrient site (the Nelumbo Area)(Table 6). Similar mean sizes of the inflorescences and their appendages were obtained at Nelumbo Area during










Table 6


Mean sizes of inflorescences and their appendages of introduced or established waterhyacinths at high, intermediate, and low nutrient study sites on Payne's Prairie State Preserve.


Total Length Total Length Length Length flower of Length of of of number floral at outer- outer- innerper inter- anthesis bract bract bract Peduncle inflor- node sheath lamina sheath diameter Site Date escence mn cm cm cm cm cm

Melton's May 19 5. ax 7.0 a 13.5 a 3.3 a 1.0 a 4.4 a .6 a
Pond
(low) Oct. 1 9. ab 9.7 b 47.9 bc 7.8 bc 2.4 b -- .8 b

Biven's Apr. 29 14. bc 6.6 a 40.4 b 7.0 b 2.4 b 9.3 b 1.0 bc Marsh
(low)

Nelumbo May 30 19. cd 6.5 a 53.5 cd 9.8 c 4.3 c 11.9 c 1.1 bc Area
(intermediate) Aug. 25 16. bcd 6.4 a 47.7 bc 9.5 c 4.0 c 11.2 bc 1.1 bc

New Canal Sept. 25 22. d 6.5 a 60.4 d 13.2 d 7.9 d 14.4 d 1.4 d (high)


X Means in each
level.


column not followed by the same letter are significantly different at 5%







61


periods of maximum flowering (30 May and 25 August). However, these characteristics were significantly larger at Melton's Pond during maximum flowering in the Fall (1 October) than during the Spring 19 May). Plants at all sites typically had a mean floral internodal length of 6.4-7.0 cm, although this length was significantly less than at Melton's Pond on 1 October (Table 6). Total inflorescence length ranged from a minimum of 7.5 cm at the low nutrient site (Melton's Pond) on 19 May to maxima of 58. cm and 70. cm at the intermediate nutrient site (Nelumbo Area) on 30 May and the high nutrient site (New Canal) on 25 September, respectively. Dimensions of the outer bract sheath and outer bract lamina ranged from 2.5-17.7 cm in length and 0.7-15.6 cm in length, respectively, between the high and low nutrient sites (Figure 9). The length of the inner bract sheath varied from 3.0-25.2 cm between the high and low nutrient sites whereas the diameter of the peduncle ranged from 0.4-1.7 cm.

Inflorescences visible five days (114 hours) before anthesis had attained, typically, 26% of their maximum lengths (Table 7). Young inflorescences were not measured at the intermediate and high nutrient sites during maximum flowering because of low inflorescence density, large plant size, and the herbivory on young inflorescences. Inflorescences at Biven's Marsh and Nelumbo Area typically grew 11% of their total length daily until 18 hours before anthesis (Table 7). At Melton's Pond inflorescences did not demonstrate a regular pattern of growth but they exhibited increasingly higher percentages of growth (5-25%) from minus 114 hours to minus 18 hours before anthesis (Table 7). Inflorescences at all sites grew approximately 28% of




































Figure 9 An inflorescence of Eichhornia crassipes
18 hours before anthesis.



















































peduncle di

.4 1.7


inner bract
sheath

3.0 25.2 cm


outer bract lamina

0.7 15.6 cm



















outer bract sheath

2.5 17.7 cm


ameter cm


63







64


Growth in length of a a percentage of total duced and established intermediate, and low Payne's Prairie State


developing inflorescence as inflorescence length of introwaterhyacinths at high, nutrient study sites on Preserve.


Melton's Biven's Nelumbo New Pond Marsh Area Canal (low) (low) (intermediate) (high) May 19 Apr. 29 May 30 Aug. 25 Sept. 25

Minus 18 hr.
to anthesis 26% 28.% 27.% 29.% 28.%

Minus 42 hr.
to Minus 18 hr. 25. 12. 14. 11. 10.

Minus 66 hr.
to Minus 42 hr. 11. 10. -- 10. -Minus 90 hr.
to Minus 66 hr. 7. 11. -- 12. -Minus 114 hr.
to Minus 90 hr. 5. 12. -- 11. -Before 114 hr. 25. 27. -- 26. --


Table 7








65


their total length during the last 18 hours before anthesis, although this increase in length ranged from 1.5 cm at the low nutrient site (Melton's Pond) to 35. cm at the high nutrient site (New Canal)(Table 7). The distance from the bases of the bracts to the apex of the peduncle ranged from 39-50% of the total inflorescence length at all sites (Table 8). This distance was equally split between the length of the subfloral peduncle (from bases of the bracts to the lowermost flower) and the length of the interfloral peduncle (from the lowermost flower to the peduncle apex)(Table 8).

Length and width of the lower sepal and the banner petal varied among sites and within an inflorescence, but no relationship to nutrient levels at a site was noted. In all flowers the lower sepal length and width ranged from 27-37 mm and 10-17 mm, respectively (Figure 10). The smallest mean sepal lengths in lower, middle, and upper flowers were found at a low nutrient site (Melton's Pond), whereas the longest mean lengths occurred at another low nutrient site (Biven's Marsh) (Tables 9,10,11). Mean sepal lengths at the higher nutrient sites (Nelumbo Area and New Canal) were similar and intermediate between the two low nutrient sites. Length and width of the banner petal of all flowers ranged from 30-43 mm and 19-29 mm, respectively (FigurelO). The smallest mean lengths of banner petals of the lower and middle flowers were found at the intermediate nutrient site (Nelumbo Area)(Tables 9,10,11). The other sites, Biven's Marsh, Nelumbo Area, and New Canal had intermediate mean lengths of banner petals. The minimum and maximum banner petal lengths occurred at the low nutrient sites, whereas banner petals from the intermediate








66


Length of peduncle as a percentage of the total inflorescence length of established and introduced waterhyacinths at high, intermediate, and low nutrient study sites on Payne's Prairie State Preserve.


Base of bracts Base of lowerto base of most flower Site Date lowermost flower to peduncle tip


May 19 Oct. 1 Apr. 29


24.%


26.%


20. 19.


May 30 Aug. 25


New Canal Sept. 25


18.


22.


23.


22. 25.


23.


22. 24.


Table 8


Melton's Pond (low)

Biven's Marsh (low)


Nelumbo Area (intermediate)































Flower of E. crassipes at anthesis. A face view; B longitudinal section; bp banner petal; Is lower sepal.


Figure 10















'.o


dq


a


V







69


Table 9


Comparative data of floral appendages of lower flowers of the inflorescence of introduced and established waterhyacinths at high, intermediate, and low nutrient study sites on Payne's Prairie State Preserve.


Lower Sepal Banner Petal Long stamens Style
Site Date length/width length/width length length

Melton's May 19 32.a 13. 35.ax 22. 16.b 11. Pond
(low) Oct. 1 33.a 14. 37.ab 22. 16.b 10.

Biven's Apr. 29 36.b 14. 39.bc 23. 17.b 10. Marsh
(low)

Nelumbo May 30 34.ab 16. 41.c 26. 17.b 12. Area 9.aY 22. (intermediate) Aug. 25 33.a 15. 38.be 24. 18.b 11.

New Canal Sept. 25 34.ab 15. 39.bc 25. 18.b 10. (high)


X Means not followed by the same letter in each column are


significantly different at y Long-styled flower form.


5% level.







70


Table 10


Comparative data of floral appendages of middle flowers of the inflorescence of introduced and established waterhyacinths at high, intermediate, and low nutrient study sites on Payne's Prairie State Preserve.


Lower Sepal Banner Petal Long
Stamen Style
Site Date length/width length/width length length

Melton's May 19 30.aX 12. 35.a 22. 16.b 11. Pond
(low) Oct. 1 31.a 13. 36.a 24. 16.b 10.

Biven's Apr. 29 34.b 12. 38.bc 22. 17.b 10. Marsh
(low)

Nelumbo May 30 33.b 16. 40.c 25. 17.b 12. Area 9.aY 22. (inter- Aug. 25 33.b 14. 37.ab 23. 18.b 10. mediate)

New Canal Sept. 25 34.b 14. 38.bc 25. 18.b 10. (high)


X Means in each column not followed by the same letter are
significantly different at 5% level.


Y Long-styled flower form.







71


Table 11


Comparative data of floral appendages of upper flowers of the inflorescence of introduced and established waterhyacinths at high, intermediate, and low nutrient study sites on Payne's Prairie State Preserve.


Lower Sepal Banner Petal Long stamen Style
Site Date length/width length/width length length

Melton's May 19 30.aX 11. 34.a 22. 16.b 10. Pond
(low) Oct. 1 30.a 12. 34,a 22. 16.b 10.

Biven's Apr. 19 34.b 12. 38.b 22. 17.b 10. Marsh
(low)

Nelumbo May 30 30.a 14. 37.ab 23. 17.b 11. Area 9.aY 22. (inter- Aug. 25 31.a 13. 35.ab 22. 17.b 10. mediate)

New Canal Sept. 25 30.a 13. 35.ab 23. 17.b 9. (high)


X Means in each column not followed by the same letter are significantly different at 5% level.


Y Long-styled flower form.







72


and high nutrient sites exhibited intermediate mean lengths (Table 11) Mean banner petal lengths of lower and middle flowers from low, intermediate, and high nutrient sites were significantly larger than or similar to mean banner petal lengths of the upper flowers at the same site (Table 12).

Mean style lengths (9-12 mm) and mean length of the long.stamens (16-18 mm) were similar in the mid-styled flowers at all study sites (Tables 9,10,11). Long-styled flowers were only observed at the Nelumbo Area and exhibited a mean style length of 22 mm and a mean length of long stamens of 9 mm. Short stamens were the same length (4-5 mm) in all flowers.

All open flowers on an inflorescence began to close after 5-6 pm and were completely closed by sunset (7-9 pm). After anthesis the perianth lobes and style slowly folded into an amorphous mass distal to the developing ovary and persisted until fruit abscission and dehiscence. The perianth tube remained intact and covered the developing fruit until dehiscence.

Twelve to eighteen hours after anthesis two bends appeared in the peduncle (Figure 8). The first or capital bend was proximal to the bases of the bracts and was initiated at a different distance from the bases of the bracts on small inflorescences (1.5-2.0 cm) than on large inflorescences (2.5-3.0 cm). The capital bend developed whether or not all flowers on the inflorescence had opened and lowered the floral axis 70-120' from its original vertical orientation. The unopened flowers in an inflorescence opened the following day with the subfloral and interfloral portions of the







73


Table 12 Mean length of the banner petal of waterhyacinth
flowers borne in the lower, middle, or upper
positions on the inflorescences at high, intermediate, and low nutrient study sites on Paynets
Prairie State Preserve.


Banner petal length mm flower position Site Date lower middle upper


Melton's Pond May 19 35.ax 35.a 34.a
(low)

Nelumbo Area May 30 41.c 40.c 37.b (intermediate)

New Canal September 25 39.bc 38.b 35.a (high)

X Means of all sites not followed by the same letter are significantly different at 5% level.







74


peduncle in a horizontal position. The interfloral peduncle tissue also exhibited a second bend by the morning of the following day. This interfloral bend formed above the lowermost flower and had maximum bending (arc-shaped) from the middle flowers to the peduncle apex, although the bend was not as acute as the capital bend.

The day after anthesis the capital bend reoriented the floral axis from 150"-180' from its vertical orientation at anthesis (Figure 8). Slight elongation (less than 10t of total length) of the peduncle occurred after anthesis until the capital bend was complete. The bending process in the capital bend occurred acropetally. The distance from the bases of the bracts to the upper portion of the capital bend was 2.5-3.0 cm shortly after anthesis and decreased to 0.3-0.7 cm after 48 hours. A third or basal bend occurred in the base of the peduncle (3-10 mm distal to its insertion in the stem) and reoriented the peduncle 10-30' from its normal vertical orientation the second day after anthesis. This bend reoriented the peduncle 900 or more from its vertical orientation 4-10 days after anthesis. The basal bend of the peduncle together with the capital and interfloral bends carried the inflorescence from the central crown of leaves and oriented the developing fruit horizontally below the water surface. The curvature of the peduncle was completed 72 hours past anthesis, although cooler temperatures decrease the rate of curvature.

Orientation of the three bends in the peduncle during

postanthetic curvature was generally related to the location of the bracts on the inflorescence. The orientation of the capital







75


bend also determined the orientation of the interfloral and basal bends. The inflorescences (53%) were reoriented to the water by the postanthetic bends at right angles to the stem axis with some degree of variance (Figure 11), whereas the remaining inflorescences were reoriented across adjacent leaves (38%) or the stem apex (9%). The orientation of the postanthetic bends among the low, intermediate, and high nutrient sites varied and no correlations were noted (Figure 11).


Pollination and Seed Set

Various insects (Hymenopterans) pollinated waterhyacinth flowers at both Melton's Pond and the Nelumbo Area. Each of the insect species approached and landed on the waterhyacinth flowers (both mid- and long-styled flower forms) in different ways. Each species visited many flowers on the same inflorescence and large amounts of pollen were observed on the stigmas of the flowers.

Developing fruits were easily observed 8-10 days after pollination because of the increased size and firmness of the perianth tube and because nonpollinated flowers always abscised from nonpollinated inflorescences. Nonpollinated flowers on pollinated inflorescences typically did not abscise until the mature fruits abscised. All stages of fruit development and dehiscence usually occurred at or below the water surface, although fruits occasionally matured totally out of the water. When the fruits were mature the brown perianth tube split longitudinally and loculicidal dehiscence of the capsule occurred. Seeds were







































Comparative data on the orientation of the capital bend to the origin of the outer bract.


Figure 11







77


900
1350 450





1800 00
Side View
of Inflorescence


Date


Angle of bend relative to origin
of the outer bract

00 00-45* 45*-900 900-135*


1350


Melton's May 19 4 1 10 2 1 Pond Oct. 1 1 4 6 4 4 Nelumbo May 30 6 6 4 0 0 Area Aug. 25 3 8 3 2 0 New Canal Sept. 25 10 1 2 0 3


28% 25% 29%


Site


Percent of Total


9% 9%








78


released through a split in the perianth tube or after agitation or decomposition of the fruit, and sank immediately to the substrate below the water.

Development of the fruit required 18-22 days during the Summer after artificial pollination of flowers in the field and in the greenhouse, and 24-28 days during the Fall and Winter after artificial pollination in the greenhouse. Artificial pollination (self and cross) produced seed set (50-150) seeds per capsule in approximately 150 inflorescences during preliminary studies in 1974, whereas no appreciable seed set was found in 50 nonpollinated inflorescences. No relationships between fruit formation and either time of the day (9:00 and 11:00 am, 1:00 and 3:00 pm) or time of the year (February through August) for pollination were noted.

Fruits that resulted from natural pollination were only

found in established waterhyacinth populations at the intermediate nutrient site (Nelumbo Area) and at a low nutrient site (Biven's Marsh). Approximately thirty and fifteen fruits per m2 were collected on 24 July and 24 August, respectively, at Nelumbo Area and only six fruits per m2 were found only on 22 May at Biven's Marsh. No additional fruits were found in established waterhyacinth populations at the low nutrient sites (Biven's Marsh, 441 Canal), intermediate nutrient site (Nelumbo Area), or the high nutrient sites (New Canal, Main Canal) at any other time during the year. Although the seed number per capsule of the fruits collected at the Nelumbo Area and Biven's Marsh varied from 12-175, most capsules contained 25-65 seeds. Inflorescences








79


of introduced waterhyacinths (6 per site) were artificially pollinated (selfed) at the low, intermediate, and high nutrient sites in the Summer and early Fall. At the low nutrient site (Melton's Pond) 165 seeds per inflorescence were collected from the four inflorescences that produced seed. At the intermediate nutrient site (Nelumbo Area) 380 seeds per inflorescence were collected from each of the six inflorescences, although no correlation to the number of seeds per capsule could be made. Herbivores destroyed four of the pollinated inflorescences at the New Canal and the remaining two inflorescences were not relocated.

Natural seed germination of Eichhornia crassipes was observed at two locations on Payne's Prairie State Preserve near the study site at the intermediate nutrient site (Nelumbo Area). During May seedlings of E. crassipes were observed with seedlings of Pontederia cordata and Hydrocotyle ranunculoides on moist sediments. During July thousands of seedlings were found floating in 30-40 cm of water among Polygonum and Amaranthus plants 2-3 weeks after water had returned to the Nelumbo Area study site; however, no seedlings were noted in the sediments below the surface of the water. The youngest seedlings had only 3-5 linear leaves (2 cm long and 0.5 cm wide) and no adventitious roots, whereas the older seedlings had adventitious roots, 5-9 leaves, but no stolons. By late July only

a few young seedlings (5-10 leaves) were observed and the remaining seedlings had produced vigorous root systems and secondary and tertiary stolons. Seedlings produced flowers 8-10 weeks after seed germination occurred and continued until early November.







80


Twenty-five floating seedlings of various ages (5-20 leaves) were collected from a site near the Nelumbo Area study site in midJuly and placed in a greenhouse culture tank. Development of the younger and older seedlings was normal and the culture tank was filled with ramets (approximately 125 individuals) after six weeks. Flowering occurred eight weeks after introduction of the plants into the greenhouse tank.


Anatomy of the Inflorescence of Eichhornia Crassipes And Its Appendages


At anthesis the inflorescence of E. crassipes is fully developed and its appendages are clearly evident (Figure 8). Distally the mature peduncle is terete (.5-1.7 cm wide) although it is slightly compressed proximal to and at its point of insertion on the stem. The bases of the outer and inner bracts are located at the midpoint of the inflorescence and 5-35 cm distal to the insertion of the inflorescence. The sheathing base of the outer bract (3.-18. cm long) surrounds both the subfloral peduncle and the inner bract for one-half the length of the sheathing base. The body of the sheathing base of the outer bract terminates in a short petiole that connects the sheathing base to the subcordate lamina (0.7-15. cm long), whereas the membranous lateral portions of the sheathing base terminate in a small stipule-flap (Figure 12). The inner bract sheath (3.5-25. cm

long) occurs opposite to the outer bract and surrounds the subIloral peduncle for one-half the length of the sheathing base. The body of the sheathing base of the inner bract terminates in a short petiole
































Structure of the outer and inner bracts of an inflorescence of E. crassipes.

1 lamina; sf stipule-flap; sb sheathing base


Figure 12







82


-4
U) U)























4
U)







83


that connects the sheathing base to the aborted lamina (0.3-2.0 cm long), whereas the membranous lateral portions of the sheathing base of the inner bract terminate in a small stipule-flap (Figure 12).

The lavender flowers (5-32 per inflorescence) are arranged in a 3/8 phyllotaxy similar to that of the leaves. The lowermost flower is located between the upper portions of the sheathing bracts, and approximately one-half the distance from the bases of the bracts to the tip of the interfloral axis. Each succeeding flower is inset into the interfloral peduncle at regular intervals. The shape of the interfloral peduncle is terete with a single large indentation at the level of the lowest flower, forms an irregular polygon at the level of the middle and upper flowers and terminates typically in a small triangular extension of peduncle tissue (0.3-2.0 cm long) that continues beyond the base of the terminal flower.

Individual flowers have three sepals and three petals which are fused for one-third of their length into a perianth tube. The sepals are lancolate, shorter than the petals, and possess a dense layer of glandular trichomes on their abaxial surface. The ovate petals are relatively broad and possess small amounts of glandular trichomes on their abaxial surface. All the perianth members are similar in color (lavendar) and the banner petal has a rhomboidal yellow-gold spot centered on a large purple area. The short stamens are adnate to the banner petal and the upper two sepals, whereas the long or mid-length stamens are adnate to the lower petals and the lower sepal. Anthers of the six stamens are dorsifixed with pollen discharge from longitudinal slits; however, pollen discharge






84


is extrorse from the short stamens and introrse from the long stamens. The filaments of short and mid-length stamens are white, whereas the filaments of long stamens exhibit purple coloration distally; all filaments have glandular hairs along their length (Figure 13). The conical ovary at the base of the flower has a long or mid-length style that terminates in a vertically oriented papillate stigma. Both long and mid-length styles have glandular hairs along their length; mid-length styles are white, and the long styles exhibit a purple coloration distally.

Transections of the mature peduncle of E. crassipes show

several characteristic features (Figure 14). The epidermis is composed of rectangular parenchyma cells whose external periclinal wall is covered by a thin cuticle. Chlorenchyma cells (3-5 layers) of the outer zone of ground tissue occur internal to the epidermis and are more frequent external to the peripheral series of vascular bundles that occur in the outer zone of ground tissue. Larger thinwalled parenchyma cells may be interspersed with the chlorenchyma cells between the peripheral series of vascular bundles. Large idioblasts with raphid crystals or tanniferous substances frequently are found interspersed in the chlorenchyma and parenchyma cells of the outer zone of ground tissue. Large parenchyma cells (2-4 layers) occur internal to the chlorenchyma cells of the outer zone of ground tissue. Aerenchyma cells comprise the remaining inner zone of the ground tissue of the peduncle. Large intercellular spaces (0.6 mm in diameter and 2.0 mm in length) are interspersed throughout the ground tissue. These intercellular spaces are delineated by elongated






















Stamen filament of E. crassipes. X 100.

Transection of a peduncle distal to its insertion on the stem. X 11.

Transection of a vascular bundle of a peduncle. X 167.

Transection of a peduncle proximal to the bases of the inflorescence bracts. X 11.

a aerenchyma; ch chlorenchyma; cl central lacuna; e epidermis; id idioblasts; ob outer bract; pf perivascular fibers; pp primary phloem; px primary xylem; t trichomes; vb vascular bundles.


Figure 13 Figure 14 Figure 15 Figure 16








86


Jw~


id


C. v
4- --


1 a


/ (


4,~.


A


A


"'S
4


c h vb









S..,



16-, -


11


*4. 'r


41~ d I w~A~
.j.I* (


D


a







87


parenchyma cells that are penetrated occasionally by large styloid crystals. Vascular bundles occur randomly in both the parenchymatous outer zone and the aerenchymatous inner zone of the ground tissue. These collateral vascular bundles (approximately 150 per transection) are surrounded by a layer of large parenchyma cells and are composed of primary phloem which is external to primary xylem. The collateral vascular bundles have 1-2 tracheids with annular to helical wall thickenings, a conspicuous protoxylem lacuna, and 3-6 sieve tube members with companion cells (Figure 15). Perivascular fibers with poorly developed secondary walls occur external to the primary phloem, especially in the peripheral bundles. Large idioblasts with tanniniferous substances frequently are found parallel to the vascular bundles among the surrounding parenchyma cells. A central lacuna (2.-5. mm in diameter) occurs in the center of the peduncle.

Anatomical features of the peduncle remain constant throughout its length, although changes do occur at the bases of the inner and outer bracts (Figure 16). Proximal to the bases of the bracts, the intercellular spaces decrease in length and diameter and the central lacuna disappears. Vascular bundles in the ground tissue increase in number at the bases of the bracts. The collateral vascular bundles are surrounded by a layer of large parenchyma cells and are composed of primary xylem (1-2 tracheids with helical thickenings and a protoxylem lacuna) and primary phloem (3-6 sieve tube members and companion cells). Perivascular fibers occur external to the primary phloem. The peripheral vascular bundles of the peduncle continue directly

into the sheathing base of the outer bract. Some vascular bundles








88


move laterally into the median portion of the sheathing base of the outer bract. Distally, other vascular bundles move laterally into the median portion of the sheathing base of the inner bract.

A single-layered epidermis with a cuticle is external to the ground tissue of the sheathing base of the outer bract (Figure 17). Chlorenchyma cells (3-4 layers) occur internal to the exterior (abaxial) epidermis although only a single layer of parenchyma cells occur internal to the interior (adaxial) epidermis. Aerenchymatous ground tissue occurs between the chlorenchyma and parenchyma cells of the outer zone of ground tissue (Figure 17). Large intercellular spaces separate the 3-4 layers of aerenchyma cells in the median portion of the sheathing base, whereas only 1-2 layers of aerenchyma cells separated by large intercellular spaces occur in the fused margins of the sheathing base. Vascular bundles are distributed throughout the aerenchymatous ground tissue with one series of bundles interior to the outer (abaxial) layers of ground tissue, and one series of bundles interior to the inner (adaxial) layers of ground tissue (Figure 17). Some vascular bundles are also between the outer and inner series of bundles in the thickest portion of the sheathing base. These collateral vascular bundles possess primary xylem (1-2 tracheids with annular thickenings and a protoxylem lacuna) which is interior (adaxial) to the primary phloem (2-4 sieve tube members with companion cells). Perivascular fibers with poorly developed secondary walls also occur within the series of large parenchyma cells that surround the vascular bundle.




Full Text

PAGE 1

COMPARATIVE STUDIES OF THE MORPHOLOGY AND ECOLOGY OF SEXUAL REPRODUCTION OF Eichhornia crasslpes (PONTEDERIACEAE) By ROBERT GERALD ANDERSON 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

PAGE 2

to my father, for introducing me to plants and for his demonstration of personal sacrifice, hard work, and personal satisfaction

PAGE 3

ACKNOWLEDGEMENTS The author wishes to express his appreciation to Dr, Terry W, Lucansky, Chairman of the Supervisory Committee, for his guidance and assistance in the conduct and fulfillment of this investigation and his assistance during the preparation of this manuscript. Sincere appreciation is expressed to the members of the supervisory committee, Drs. Mildred M. Griffith, John J. Ewel Dana G. Griffin, III, and David L, Sutton for advice during the progress of this investigation and critical evaluation of the manuscript, I also thank Drs. Ariel Lugo and William T. Haller for evaluation of the manuscript Appreciation is expressed to the Florida Department of Natural Resources for permission to complete portions of this investigation on Payne's Prairie State Preserve and for funding the project under grant number 77785, The Botanical and Ecological Aspects of Aquatic Weed Control. Gratitude is also expressed to Dr, Donald A. Graetz, Soils Department, for the use of his laboratory facilities. My greatest thanks go to my wife, Joanne, for the many rough copies and final copy of this dissertation which she gladly typed, for her financial support of my education, and for her understanding during the struggle of this dissertation.

PAGE 4

TABLE OF CONTENTS Page ACKNOWLEDGEMENTS iii LIST OF TABLES v LIST OF FIGURES vii ABSTRACT x INTRODUCTION 1 METHODS 9 RESULTS 24 Sexual Reproduction in Waterhyacinth Populations . 24 Phenology of Waterhyacinth Populations 24 Initiation of Flowering 30 Inflorescence Density 44 Inflorescence Development 55 Pollination and Seed Set 75 Anatomy of the Inflorescence of Eichhornia crassipes and its appendages 80 DISCUSSION 123 Reproductive and Vegetative Characteristics of Waterhyacinth Populations 123 Anatomical Aspects of Eichhornia crassipes 155 SUMMARY 179 LITERATURE CITED 184 BIOGRAPHICAL SKETCH 190 iv

PAGE 5

LIST OF TABLES Table Page 1 Mean concentrations of nitrate nitrogen, ammonia nitrogen, total organic nitrogen, and total phosphorous at the field sites 28 2 Mean temperatures measured at Melton's Fond and the Gainesville Power Plant during the flowering of waterhyacinths at the study sites 32 3 Precipitation, solar insolation, and daylength means during the flowering of waterhyacinths at the study sites 34 4 Mean total leaf length of the third mature leaf of established waterhyacinths 38 5 The number of senesced inflorescences on waterhyacinths and their ramets at Melton's Pond and the Nelumbo Area 50 6 Mean sizes of inflorescences and their appendages of introduced or established waterhyacinths at high, intermediate, and low nutrient sites 60 7 Growth in length of a developing inflorescence as a percentage of total inflorescence length of introduced or established waterhyacinths at high, intermediate, or low nutrient sites 64 8 Length of peduncle as a percentage of the total inflorescence length of established or introduced waterhyacinths at high, intermediate, or low nutrient sites 66 9 Comparative data of floral appendages of lower flowers of the inflorescence of introduced or established waterhyacinths at high, intermediate, or low nutrient sites 69 10 Comparative data of floral appendages of middle flowers of the inflorescence of introduced or established waterhyacinths at high, intermediate, or low nutrient sites 70

PAGE 6

Table Page 11 Comparative data of floral appendages of upper flowers of the inflorescence of introduced or established waterhyacinths at high, intermediate, or low nutrient sites 71 12 Mean length of the banner petal of waterhyacinth flowers borne in the lower, middle, or upper positions on the inflorescences at high, intermediate, and low nutrient sites 73 13 Means of the lamina length to lamina width ratios of the third mature leaf from introduced waterhyacinths at high, intermediate, and low nutrient sites 144 14 Length of the petiole (as a percentage of the total leaf length) from the third mature leaf of introduced waterhyacinths at high, intermediate, and low nutrient sites 145

PAGE 7

LIST OF FIGURES Figure Page 1 The location of study sites in the north-central region of Payne's Prairie State Preserve 11 2 Portable dock apparatus 16 3 Mean and standard deviation of total leaf length of introduced waterhyacinths at high, intermediate, and low nutrient sites 37 4 Mean and standard deviation of petiole width of the third mature leaf of introduced waterhyacinths at high, intermediate, and low nutrient sites 41 5 The number of inflorescences per m 2 with open flowers during anthesis at two low nutrient sites 49 6 Total number of stem apices and number of inflorescences of waterhyacinths in the nonfertilized culture tank 52 7 Total number of stem apices and number of inflorescences of waterhyacinths in the fertilized culture tank 54 8 Inflorescence structure before anthesis, at anthesis, and during postanthetic curvature 59 9 An inflorescence of Eichhornia crassipes 18 hours before anthesis 63 10 Flower of _E. crassipes at anthesis 68 11 Comparative data on the orientation of the capital bend to the orientation of the outer bract 77 12 Structure of the outer and inner bracts of an inflorescence of E. crassipes 13 Stamen filament of _E. crassipes ....... 32 86 14 Transection of a peduncle distal to its insertion on the stem 86 15 Transection of a vascular bundle of a peduncle ... 86 vii

PAGE 8

Figure Page 16 Transection of a peduncle proximal to the bases of the inflorescence bracts 86 17 Transection of the sheathing base of the outer inflorescence bract 90 18 Transections of the fused margins of the sheathing base of the inner inflorescence bract 90 19 Transection of the aborted lamina of the inner inflorescence bract 90 20 Transection of the functional lamina of the outer inflorescence bract 90 21 Transection of the vascular plexus of a mature flower of E. crassipes 97 22 Transection of the origin of the vascular traces of a mature flower of _E. crassipes 97 23 Diagrammatic representation of the vascular traces of the floral appendages of Z. crassipes 99 24 Transection of the base of an ovary of a mature flower 97 25 Transection of an ovary between the base of the carpels and the placentae 97 26 Transection of the placental region of an ovary 102 27 Transection of a flower distal to the apex of the ovary 102 28 Longisection of the convex side of the capital bend in the peduncle 102 29 Longisection of a peduncle before postanthetic curvature 102 30 Longisection of a fertilized ovule three days after pollination 109 31 Transection of a fertilized ovule twelve days after pollination 109 32 Transection of a stem of _E. crassipes 109

PAGE 9

Figure Page ... 109 33 Transection of an adventitious root of V. crassipes 34 Transection of a stem proximal to the insertion of an axillary bud 114 35 Transection of a stolon of E. crassipes 114 36 Transection of a stem at the insertion of an inflorescence 114 37 Transection of a stem distal to the insertion of an inflorescence 114 38 Transection of a tertiary stem apex 119 39 Transection of a tertiary stem apex 119 40 Transection of an aborted lamina and stipule-flap of a prophyll 119 41 Morphological variation in leaves of waterhyacinth 142 ix

PAGE 10

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 COMPARATIVE STUDIES OF THE MORPHOLOGY AND ECOLOGY OF SEXUAL REPRODUCTION OF Eichhornia crassipes (PONTEDERIACEAE) By Robert Gerald Anderson December, 1976 Chairman: Terry W. Lucansky Major Department: Botany Ecological and morphological aspects of sexual reproduction of Eichhornia crassipes (Mart.) Solms were investigated at six, north Florida aquatic ecosystems on Payne's Prairie State Preserve. Phenological changes in the populations of established and introduced waterhyacinths were observed at regular intervals throughout a calendar year. Nutrient concentrations, herbivory, and temperature were the major influences upon both the growth and development of the waterhyacinth populations and the time scale of phenological changes. Increases in nutrient concentrations increased leaf size and inflorescence size of the waterhyacinth plants. Stress to the waterhyacinth population by cold temperatures or by herbivore infestations (Arzama densa) resulted in increased stolon numbers per unit area and

PAGE 11

caused decreases in the size of leaves of waterhyacinth plants in a given population. Floral initiation in the waterhyacinth occurred during periods of vigorous growth and relatively low nitrogen concentrations in natural and artificial conditions. Floral initiation in E. crassipes was not correlated with mean temperatures, photoperiod, total solar insolation, or rainfall. Densities of inflorescences fluctuated at both the study sites and in the greenhouse. Stems and stolons (elongated first internode of axillary bud) of E. crassipes possess collateral vascular bundles in both the parenchymatous central cylinder and the cortex with both aerencymatous and parenchymatous zones. Stems have a distinct parenchymatous layer between the cortex and the central cylinder, whereas the cortex and central cylinder of stolons are not separated by a distinct parenchymatous layer. The adventitious roots have an aerenchymatous cortex with large, radially-arranged intercellular spaces and a polyarch stele. The peduncle of the inflorescence is composed of aerenchymatous ground tissue with scattered collateral vascular bundles and a large central lacuna. The sheathing base of the outer inflorescence bract terminates in a well-developed lamina and a membranous stipule-flap, whereas the sheathing base of the inner inflorescence bract terminates in an aborted lamina and a membranous stipule-flap. Three postanthetic bends are found on the peduncle of the inflorescence of E. crassipes and differential cell elongation occurs on opposite sides of the peduncle. xi

PAGE 12

Eichhornia crassipes exhibits a sympodial growth pattern. The inflorescence terminates the primary shoot and vegetative growth continues from the ultimate axillary bud (secondary shoot) The anatomy, prophyll morphology, and development of a secondary shoot (after floral initiation) is dissimilar from that of a typical axillary shoot (during vegetative growth) Sepals, petals, and carpels of the flower of E_. crassipes are similar anatomically, and each floral appendage possesses three vascular traces that originate from the vascular plexus. The stamens have a single trace that arises from the median vascular trace of each perianth member. xii

PAGE 13

INTRODUCTION Eichhornia crassipes (Mart.) Solms, waterhyacinth, has created economic problems in most tropical and subtropical regions of the world (Bock, 1966; Sculthorpe, 1967; Holm, 1969). The plant is well adapted to a freshwater habitat and typically forms an impenetrable mat that congests slow-moving riverine systems, reservoirs, lakes, ponds, and canals (Bruhl and Dutta, 1923; Penfound and Earle, 1948; Gay, 1960; Sculthorpe, 1967; Batanouny and El-Fiky, 1975). The mat can reduce commercial and recreational water traffic, clog sluice gates in dams, overtake rice paddies, reduce flow in irrigation, sewage and drainage canals, and cause flooding (Penfound and Earle, 1948; Sculthorpe, 1967). Floating mats of waterhyacinths also retard gaseous exchange between the water and the air and reduce oxygen concentrations and available nutrients (Penfound and Earle, 1948; Zeiger, 1962; Sculthorpe, 1967). Many methods of control chemical, biological, and ecological have been attempted without worldwide success because of the expense and inefficiency of control measures and the reproductive capabilities of the plant (Sculthorpe, 1967). The waterhyacinth has many of the ecological adaptions for a colonizing species and is considered one of the ten worst weeds in the tropics (Sculthorpe, 1967; Holm, 1969). The waterhyacinth population can maintain itself in various aquatic situations. When an

PAGE 14

area is filled by waterhyacinths, competition between individuals becomes intense and the high growth rate is translated to morphological plasticity and allows individual plants to change their form for the specific environmental conditions (Rao, 1920a; Penfound and Earle, 1948). Hydrogen ion concentration has little effect upon the growth of waterhyacinths (Bock, 1966; Gay, 1960; Haller and Sutton, 1973), although various concentrations of nitrate and phosphorous increase the vigor of waterhyacinths (Haller and Sutton, 1973; Morris, 1974). Morris (1974) and Boyd and Scarsbrook (1975) demonstrated that differences in total biomass, productivity, plant density, and root: shoot biomass ratios of plants from different sites were directly proportional to nitrate and phosphorous concentrations and the degree of development of native flora. Waterhyacinths require relatively high light intensities for optimal growth, whereas low light intensities result in the death of plants (Penfound and Earle, 1948). The plants can withstand low temperatures although a minimum temperature of -5C for 24 hours has been reported to kill plants (Penfound and Earle, 1948). Penfound and Earle (1948) reported the doubling time for numbers of individuals from 11-15 days, depending upon the weather conditions, allows waterhyacinths to invade an open area very quickly. Although E. crassipes does not utilize dicarboxylic acid (C4) metabolism in photosynthesis, photosynthetic rates of the plant are comparable to those for C4 plants (Knipling et al. 1970).

PAGE 15

The genus Eichhornia is one of seven perennial, freshwater genera in the family Pontederiaceae and occurs natively in the riverine systems of tropical South America (Castellano, 1958; Sculthorpe, 1967; Agostini, 1974). All Eichhornia species have a floating-emergent habit with erect or prostrate stems (Castellano, 1958). The stem (rhizome) of E. crassipes is erect in the water with a vertically oriented apex slightly below the water surface (Penf ound and Earle, 1948) Eichhornia crassipes typically has inflated petioles that allow flotation on the water surf ace, whereas the remaining species have linear petioles and occur in shallow or transient freshwater systems (Castellano, 1958; Agostini, 1974). Nonswollen linear petioles may also occur in E. crassipes during high population densities, high temperatures, and low light intensities, whereas the swollen petioles are correlated with low population densities, low temperatures, and high light intensities (Rao, 1920a; LaGarde, 1930; Penfound and Earle, 1948). Arber (1918, 1920) termed the lamina of waterhyacinth and other members of the Pontederiaceae a pseudo-lamina and reported it was unifacial and represented an expanded flattened section of the cylindrical unifacial petiole. However, Kaplan (1973) demonstrated that the "phyllode concept" is not valid for monocotyledons. He confirmed that the lamina and petiole of certain dorsiventral, laminar leaves are distinct and separate elaborations of a basal meristem that partially encircles the leaf primordium and that the distal leaf zone remains rudimentary. Laminas of

PAGE 16

waterhyacinth leaves are probably not pseudo-lamina. Leaves of waterhyacinth appear to be morphologically similar to leaves of species described by Kaplan (1973). Bruhl and Dutta (1923) have described the anatomy of the rudimentary distal leaf zone in a young leaf of waterhyacinth. A stipule or proliferating leaf base originates proximal to the base of the leaf as a membranous ochrea that surrounds the younger leaves and apex (Weber, 1950). A membranous lamina (stipule-flap), which is only two cell layers thick and has a network of vascular tissue, occurs at the apex of the stipule and has mucilage cells along its margin (Weber, 1950). The roots of the waterhyacinth are adventitious, nonbranched, and possess numerous, multicellular lateral roots (Penfound and Earle, 1948; Weber, 1950; Couch, 1971). Two forms of roots may occur on plants that are partially or wholly rooted in soil (Weber, 1950). The typical purple roots have a small diameter and possess the normal complement of side roots, whereas the atypical white roots are larger in diameter and have few lateral roots. Weber (1950) calculated the total length of a typical root plus its lateral roots to be 73 m (surface area of 0.05 m^) . Asexual reproduction is the primary means of reproduction for the waterhyacinth (Bruhl and Dutta, 1923; Penfound and Earle, 1948; Sculthorpe, 1967; Das, 1969). Populations have been reported to double every 10-15 days through stolon production (Penfound and Earle, 1948; Batanouny and El-Fiky, 1975). An axillary bud undergoes development and the elongation of its first internode (between

PAGE 17

prophyll and first leaf) carries the newly developed leaves, roots, and shoot apex into the water. This elongated internode (stolon) varies in length according to the population density (Das, 1967). Agharkar and Banerji (1930) describe the inflorescence of Eichhornia crassipes as terminal and growth of the individual plant continues from an axillary bud of the terminal leaf. The inflorescence is a spike (Penfound and Earle, 1948; Bock, 1966; Tag el Seed, 1972) although Singh (1962) termed it a contracted panicle. The individual, zygomorphic flowers are borne within two bracts that originate on the peduncle. The outer bract has a sheathing leaf base and a developed lamina at its apex, whereas the inner bract has a sheathing leaf base and an aborted lamina at its apex (Weber, 1950). A variable number of flowers are borne on the floral axis of the inflorescence (Penfound and Earle, 1948; Tag el Seed and Obeid, 1975). Each individual lavender flower has three sepals and three petals which are adnate for one-third of their length and the sepals are typically narrower than the petals. The banner (upper) petal differs slightly from the other two petals in size and has a bright yellow spot in the middle of a broad blue-purple field (Penfound and Earle, 1948; Singh, 1962). Stamens arise from the perianth tube and separate from it near the sinuses of the perianth. Three stamens are short in length and occur deep within the throat of the perianth, whereas the other stamens will be longer than the pistil or intermediate in length (Penfound and Earle, 1948; Singh, 1962; Tag el Seed and Obeid, 1975). The gynoecium consists of a conical ovary with many ovules

PAGE 18

and a white to lavender style that varies in length (Bock, 1966) and terminates in a white, capitate stigma (Penfound and Earle, 1948). The vascular anatomy of the flowers has been described by Singh (1962) A broken ring of vascular tissue occurs at the base of each flower and each perianth lobe receives three traces. Each stamen has a single trace and three vascular traces proceed into each carpel. Median traces from the carpellary wall pass into the style and proceed to the six-lobed papillate stigma. All flowers of the inflorescence usually open at the same time shortly after sunrise and close at or after dusk. After anthesis the inflorescence bends to the surface of the water (Rao, 1920; Agharkar and Banerji, 1930; Earle, 1947; Penfound and Earle, 1948; Weber, 1950) A bend begins to form below the base of the two bracts on the inflorescence and is completed 48 hours after anthesis. Additional bends occur in the floral axis between the flowers and at the base of the peduncle and force the closed flowers into the water. Variable-sized cells are reported to occur on the convex and concave sides of the bend (Weber, 1950; Das, 1967). Sexual reproduction is not responsible for the rapid population development of E_. crassipes but is a major factor in site recolonization (Penfound and Earle, 1948; Sculthorpe, 1967). Although large amounts of seed (1.1 x 108/ha, Zeiger, 1962) may be produced, the specific and complex germination requirements for the seeds are not common in perennial aquatic systems (Hitchcock et al, 1949; Tag el Seed, 1972).

PAGE 19

After natural or artificial pollination the fruit of the waterhyacinth usually develops beneath the surface of the water, although development of the fruit can occur in the air (Agharkar and Banerji, 1930; Das, 1967). All flowers do not set seed under natural conditions. The highest percentage of fruit set is approximately 35% (Agharkar and Banerji, 1930), although the percentage is frequently much lower (Bock, 1966; Tag el Seed, 1972). Bock (1966) and Tag el Seed and Obeid (1975) have shown that pollen grains are viable and will germinate and grow in vitro. Natural pollinators and selfpollination account for natural seed set (Penfound and Earle, 1948; Bock, 1966; Tag el Seed, 1972; Tag el Seed and Obeid, 1975), although artificial pollination can be successful (Francois, 1964). Seed germination occurs naturally on reflooded ground (Parija, 1930; Weber, 1950). Laboratory germination is successful when wet or dry seeds are placed in shallow water on a peaty soil with high light intensity and water temperatures of 28-36C (Hitchcock et al, 1949; Tag el Seed, 1972) or after scarification (Penfound and Earle, 1948; Hitchcock et al., 1949; Das, 1969). As the water level increases and the seedlings are submerged, the primary root abscises and the rootless seedling floats to the surface of the water (Parija, 1930). Swollen petioles and adventitious roots soon develop on the seedling (Parija, 1930; Penfound and Earle, 1948). The objectives of this study are to determine the comparative morphology and anatomy of sexual reproduction of Eichhornia crassipes at ecologically different sites through a growing season. Aspects of

PAGE 20

floral initiation and continuously flowering waterhyacinth populations are described in relation to changes in the vegetative morphology of individual plants and populations of waterhyacinths and environmental parameters. These data can be used to determine and analyze the consequences of sexual reproduction to the management and control of waterhyacinths in freshwater aquatic systems.

PAGE 21

METHODS Observations of specific morphological and phenological characteristics of Eichhornia crassipes Mart. (Solms) were made on Payne's Prairie State Preserve located 3 km south of Gainesville, Florida, and under controlled conditions in a greenhouse at the University of Florida. Payne's Prairie State Preserve is the bottom of a shallow lake that drained naturally near 1900 and was subsequently diked and channelized to maintain high water levels. Great diversity of aquatic plant species occur in the deep, shallow, and transient water systems that intergrade throughout the prairie. However, waterhyacinths alone dominate the drainage canals. The following study sites differed in nutrient concentrations, water flow, and development of native aquatic species: 1. Melton's Pond. Melton's Pond is a large spring-fed pond on the north ridge of Payne's Prairie (Figure 1). Its oligotrophic water covers approximately one hectare with a center depth of 3-4 m. Native aquatic macrophytes Typha latifolia L Pontederia cordata L Panicum hemi tomon Schult., Limnobium spongia (Bose) Rich., Nuphar luteum (L.) Sib thorp and Smith subsp. macrophyllum show a typical peripheral zonation around the pond. Ceratophyllum demersum L. fills the remaining open water

PAGE 22

u

PAGE 23

11

PAGE 24

12 of the pond, although during Spring and Summer the floating plants Wolff iell a f loridana (Smith) Thompson and Utri cularia spp. are also conspicuous in the open water. Waterhyacinths do not occur naturally in this site. 2. New Canal. During January 1975, this canal (8-10 m wide and 2-3 m deep) was completed to carry nutrient rich waters at a rate of 3.0-4.5 x 10^ m^.day from S\-7eetwater Branch and the Gainesville sewage treatment plant directly to the Alachua Sink, the drainage point for Payne's Prairie (Figure 1). Vigorous waterhyacinths (80-100 cm tall) invaded and totally dominate this canal with the exception of scattered plants of Amaranthus spp, and Polygonum spp. which may grow upon the waterhyacinth mat. 3. Nelumbo Area. In the center of Payne's Prairie (0.8-1.0 km south of the Alachua Sink) is a large area of mixed marsh and shallow (1-1.5 m) open water (Figure 1). The American lotus, Nelumbo lutea (Willd.) Pers. is the dominant macrophyte in this area, but Typha latif olia L. and Pontederia cordata L. invade the shallow water during times of drought. The submerged aquatics Najas flexilis (Willd.) Rostk. and Schmidt and Ceratophyllum de mers um L. are also distributed in the deeper water of this community. Waterhyacinths (60-80 cm tall) are found along the eastern edge of the community along a dike and form only a minor component of this ecosystem.

PAGE 25

13 4. Main Canal. This canal, populated solely by vigorous waterhyacinths (80-100 cm tall) carries much of the drainage water from the marshes of the Prairie to the Alachua Sink (Figure 1). After the sewage effluent and water discharge from the Gainesville sewage treatment plant was diverted into the New Canal in January 1975, the amount of water carried by the Main Canal decreased. As of June 1975, the Main Canal was controlled by flood gates and became a water storage area of the Prairie. 5. Biven's Marsh. On the north central boundary of Payne's Prairie was a mixed marsh whose main channels were dominated only by moderately vigorous waterhyacinths (40-60 cm tall) (Figure 1) This marsh had received the sewage effluent from the Gainesville sewage treatment plant until the diversion of effluent into the New Canal which greatly decreased the size of Biven's Marsh. Flow of water from Biven's Arm in south Gainesville into Biven's Marsh occurred at low volumes. 6. Highway 441 Canal. Along the west side of U.S. Highway 441, which traverses Payne's Prairie State Preserve from north to south, was a canal 8-10 m wide and 2-3 m deep (Figure 1) This canal was in a separate water flow system than those previously described and remained relatively slow moving and stagnant except in periods of heavy rainfall. Moderately vigorous waterhyacinths

PAGE 26

14 (40-60 m tall) dominated the canal and were the only species in the deeper water. A portable dock apparatus was constructed and allowed an individual to sample a given study site regardless of water depth. Repeated measurements were possible and the competitive environment of the mat remained unchanged. The portable dock apparatus (Figure 2) was constructed in two sections composed of a 3.7 m ladder and one or two blocks (1.2 m by 0.6 m by 0.4 m or 0.6 m by 0.6 m by 0.4 m) of flotation polystyrene. The polystyrene blocks weighed 0.013 kg-m-3 (1.1 lb-ft -3 ) and supported 0.763 kg-m~3 (60 lb-ft~3), were enclosed in canvas, waterproofed with a coat of latex paint, and attached to the end of the ladder with wide (5 cm) woven canvas strapping. The portable dock was lowered onto and raised from the waterhyacinth mat with a 1.8 m steel pipe as a lever and a 5-rope block and tackle. Two sections of the portable dock apparatus, when connected in series, allowed an area of 10 m 2 of waterhyacinths to be measured in a transect. When set parallel to each other and joined by a 6.1 m aluminum ladder, the portable docks allowed a total area of 18 m 2 to be observed with minimal damage to the waterhyacinth mat. On 4 January 1975, ten plants of Eichhornia crassipes were introduced into pens (3 m x 1.5 m) of polyvinyl chloride (PVC) drainage pipe (10 cm diameter) at Melton's Pond, Nelumbo Area, and Biven's Marsh. These plants were located at each study site for replicate measurements. The introduced plants were propagated from

PAGE 27

Figure 2 Portable dock apparatus. A. Diagram of portable dock composed of an aluminum ladder (3.7 m long), two polystyrene blocks, and lever with block-andtackle. B. Photograph of portable dock above the water surface and waterhyacinth population at the New Canal.

PAGE 28

16 B

PAGE 29

17 young stolons (1-2 leaves and roots 1-5 cm long) obtained from the waterhyacinth population at Lake Alice on the University of Florida campus in November 1974, and cultivated inside the greenhouse in tanks that contained soil, peat moss, and 3-40 cm of water. After introduction into the study sites, the plants were protected from frost damage at night by a polyethylene cover. The dredging of the New Canal in January 1975, caused a change in water flow patterns on Payne's Prairie State Preserve and dictated that the plants from Biven's Marsh be moved to the New Canal 15 March 1975. The introduced plants were measured at biweekly intervals for the period between 4 January 1975 and 10 December 1975. Nightly freezing temperatures began 19 December and terminated the field experiments. The following morphological measurements and observations were made on six labeled introduced plants at each site: 1. Leaf size length and width of the lamina of the third mature leaf from the apex per plant, length of the third mature petiole from origin of the stipule to the base of the lamina per plant, and the width of the petiole at its widest point per plant. 2. Number of leaves per plant produced between dates of observation. 3. Inflorescence number and flower number per inflorescence per plant.

PAGE 30

18 The following morphological measurements and observations were made on the succeeding generations of ramets of the six labeled plants at each site: 1. Number of leaves per plant produced between dates of observation. 2. Length of third mature petiole from origin of the stipule to the base of the lamina per plant. 3. Inflorescence number per plant and flower number per inflorescence. At intervals of 4-6 weeks from January to December 1975, established waterhyacinths were extracted in three samples of onesixth m^ from the New Canal, Nelumbo Area, Main Canal, Biven's Marsh, and Highway 441 Canal. The following morphological measurements and observations were made from six established waterhyacinth plants at each site: 1. Number of leaves with intact lamina and petiole or leaves with senesced lamina and intact petiole, or leaves with senesced lamina and petiole and number of leaf bases on the stem. 2. Leaf size length and width of lamina of third mature leaf per plant, length of third mature petiole from origin of stipule to base of lamina per plant, and the width of petiole at its widest point per plant. 3. Length of the longest root from its origin to its tip. 4. Inflorescence number at preand postanthesis

PAGE 31

19 5. Flower and capsule number per inflorescences per plant. These additional measurements were made on a per area basis at each site: 1. Number of full-sized plants (plants whose leaves formed the canopy of the mat) 2. Number of small plants (plants whose leaves were below the canopy) 3. Inflorescence number at preand postanthesis 4. Dry mass mass of total sample of waterhyacinths after drying 4-5 days at 70 C in an oven. Specific areas of Biven's Marsh, Highway 441 Canal, and Nelumbo Area were marked with white stakes before flowering commenced and were photographed at 3-4 day intervals during flowering to obtain the number of flowers produced per area. Leaves of waterhyacinths typically grew taller than inflorescences at New Canal, Main Canal, and Nelumbo Area and prevented accurate counts at these sites. The size of the inflorescence and its parts were measured at Melton's Pond and Nelumbo Area in the Spring and Fall, at Biven's Marsh in the Spring, and at the New Canal in the Fall of the year. Twenty-five labeled inflorescences along a transect were measured for five consecutive days through preanthetic development, anthesis, and postanthetic curvature. The following inflorescence parts were measured before anthesis:

PAGE 32

20 1. Length of peduncle from origin of stipule of the terminal leaf to the bases of the inner and outer bracts and width of peduncle immediately below the bases of the bracts. 2. Length of outer bract lamina and length of sheathing base of outer bract from the base of the bract to the base of the lamina. 3. Length of sheathing base of inner bract from the base of the outer bract to the base of the aborted lamina. At anthesis the following characters were measured: 1. Length of peduncle from origin of stipule of the terminal leaf to the bases of the inner and outer bracts and width of peduncle immediately below the bases of the bracts. 2. Length of outer bract lamina and length of sheathing base of outer bract from the base of the bract to the base of the lamina. 3. Length of sheathing base of inner bract from the base of the outer bract to the base of the aborted lamina. 4. Length of the subfloral peduncle from the bases of the inner and outer bracts to the base of the first flower and length of the interfloral peduncle from the base of the first flower to the apex of the inflorescence axis. 5. Total flower number and number of flowers open per plant.

PAGE 33

21 6. Flower size three flowers per inflorescence per plant were measured at the base, middle, and top of the interfloral peduncle. Length of perianth members was measured from the point where petals and sepals separate. The length and width of the banner petal and lower sepal, the length of the pistil, and the longest set of stamens were also measured in each flower. Following anthesis and postanthetic curvature, the following measurements were made: 1. Length of the peduncle from the origin of stipule of the terminal leaf to the base of the bend below the bracts and the length of the peduncle from the bases of inner and outer bracts to the top of the bend. 2. Orientation of the bend relative to the lamina of the outer bract. Ambient air temperature was measured with a recording thermograph that was mounted in a specially-built instrument box 0.4m above the water level at Melton's Pond. Daily temperatures were recorded from 14 January 1975 to 20 December 1975. These data were supplemented by temperatures published in Climatological Data by the National Oceanic and Atmospheric Administration. Rainfall measurements were taken from the Climatological Data by the National Oceanic and Atmospheric Administration. The standard daylength calculations of Dr. J. P. Oliver, Department of Physics and Astronomy, University of Florida, were utilized in

PAGE 34

22 this study and weekly totals of solar insolation were averaged from daily totals summarized in the Environmental Data Summary produced by the Environment-Plant Interaction Studies group of the Fruit Crops Department, University of Florida. A 250 ml water sample was extracted from the top 10 cm of surface water at each site at 4-6 week intervals throughout 1975. Samples were preserved immediately with phenyl mercuric acetate, put on ice while in the field, and stored at -20C. Concentrations of nitrate nitrogen (NO.,) ammonia nitrogen (NHU) total Kjeldahl organic nitrogen, and phosphorous (P) were measured after samples had been filtered through 20ju. filters. Nitrate nitrogen was measured colorimetrically to 0.1 mg/1 by its reaction to chromotropic acid in the method described by West and Ramachandran (1966) The ammonia specific ion electrode with the Orion Pontentiometer model 801, was utilized for NH3 determinations as low as 0.1 mg/1. A standard micro-Kj eldahl method for water samples was utilized to determine total organic nitrogen concentrations. Phosphorous was measured colorimetrically to 0.1 mg/1 by the ascorbic acid method described in Standard Methods for the Examination of Water and Wastewater (1971). Waterhyacinths used in greenhouse studies and introduced into the study sites were propagated from young stolons with 1-2 leaves and roots 1-5 cm long. These stolon cuttings were placed into wet soil in a shallow tank 50 cm deep and lined with polyethylene film to prevent desiccation. The soil (5-15 cm) was used as a nutrient

PAGE 35

23 source and sink in the bottom of the tank and covered with peat moss for support of the cuttings. On 22 February 1976, ten waterhyacinth plants were introduced into each of two shallow polyethylene-lined tanks (30 cm deep) that held approximately 300 1 of water and 8 cm of nutrient-poor soil. After introduction of the plants, two tablespoons soluble 20-20-20 commercial fertilizer with trace elements (Rapid Gro) was placed in each tank. When flowering commenced, two additional tablespoons of fertilizer were added to one tank, whereas no additional fertilizer was given to the other tank. The total number of plants and total number of inflorescences produced per plant were counted at weekly and 3-4 day intervals, respectively. Inflorescences at preanthesis and postanthesis were collected from four field sites at times of intensive flowering during the Spring and Fall and preserved in a formalin-acetic acid-ethanol (FAA) solution. Stems, stolons, and adventitious roots were collected from a large culture tank maintained in the greenhouse. Infiltration and embedding were done under vacuum with a standard tertiary butyl alcohol series and 56.5C wax (Johansen, 1940). Sections were cut at 10-12u. on a rotary microtome and stained with a safranin-fast green series. Preserved flower buds were dissected and mounted in Hoyer's solution for clearing (Anderson, 1954). A Zeiss M35 automatic camera was used to photograph the sections. Duncan's new multiple-range test (Steele and Torre, 1960) was used to compare treatment means at a 95% confidence interval.

PAGE 36

RESULTS Sexual Reproduction in Waterhyacinth Populations Phenology of Waterhyacinth Populations Small vigorous waterhyacinths with float leaves 10. cm long were introduced into Biven's Marsh, Nelumbo Area, and Melton's Pond on 4 January 1975. Original leaf margins and the swollen petioles appeared necrotic after two weeks and these leaves arched into the water and senesced by 19 February. New leaves (6.-8. cm long) with intense red coloration distally on the adaxial surface of short, bulbous petioles were produced in 2-4 weeks. All individuals at the study sites produced many stolons and the nine pens were 1/6-1/3 full (160-350 plants) by 19 March. Changes in plant size were first noted at the New Canal (introduced waterhyacinths were moved to this site from Biven's Marsh after diversion of nutrient-rich water from Biven's Marsh to the New Canal) two weeks after waterhyacinths were introduced and the pens were filled by plants with float leaves by 16 April. By 15 May the elongated leaves of introduced plants at New Canal were slightly shorter (65 cm vs. 85 cm) than established plants at the site. After June 1 all plants at the New Canal remained large and vigorous throughout the year, and both introduced and established populations fluctuated similarly with environmental stresses. 24

PAGE 37

25 The water level dropped rapidly during March at the Nelumbo Area because water flow patterns were changed and water and sewage input from Biven's Marsh and the central prairie were diverted to New Canal. By 8 April water depth at the Nelumbo Area had decreased 1 m, and sediments were exposed throughout the area by 17 April. Introduced waterhyacinths with small leaves and swollen petioles were rooted into moist sediments. In early May terrestrial species (especially Cyperus strigosus and Amaranthus spp ) invaded the open sediments and during May and June these terrestrial species formed a canopy over the introduced waterhyacinths. The return of standing water to the Nelumbo Area during June and July and the availability of additional nutrients from the oxidation of the sediments allowed the introduced waterhyacinths to increase rapidly in numbers and size through late June and July. Water depth at the site returned to 0.7 m by 8 August and introduced and established waterhyacinths assumed a floating habit again. The introduced waterhyacinths had been rooted in sediments in a single pen since March and on 8 August these free-floating plants were separated into three pens and formed large vigorous plants by 10 September that filled the pens until 10 December. Leaf length of introduced waterhyacinths was less than 15.0 cm from the time of introduction until 6 August at Melton's Pond. All plants had short leaves with bulbous petioles until 10 July. Afterwards all introduced plants produced long leaves (35.-45. cm) with narrow petioles.

PAGE 38

26 Established populations of waterhyacinths at the Main Canal, the Nelumbo Area, Biven's Marsh, and the 441 Canal consisted of small individuals with leaf lengths of 20.-30. cm between 22 January and 14 February. Individuals at the Main Canal grew rapidly during March and April and established maximum leaf lengths (75. cm) by 24 April. Populations at the Nelumbo Area, Biven's Marsh, and 441 Canal reached maximum leaf lengths of 73. cm on 24 July, 53. cm on 28 October, and 51. cm on 29 July, respectively. Fluctuations in leaf size of the populations at all sites occurred at various times throughout the summer because of herbivory by Arzama dens a or changes in the water flow patterns. Decreases of 5.-10. cm in leaf length were noted at all study sites during late November and early December when day and night temperatures were cool. Frost on 19 December damaged laminas and 1/4-2/3 of the petioles at all sites. The larvae of the noctuid moth, Arzama densa caused periodic and devastating damage to introduced and established waterhyacinth populations at all sites by devouring the stem apex and 1-5 cm of stem tissue. Plants at the Main Canal and the 441 Canal were infested in January when the plants were initially sampled. Heavy infestations in the waterhyacinth plants occurred in late May and June in New Canal and in late June and July in the Main Canal; only dead or dying leaves were observed in the population's canopy at these times. Heavy infestations of A. densa also occurred at Biven's Marsh during July and August and at the 441 Canal and Melton's Pond during late August and September. During heavy infestations of

PAGE 39

27 A. densa all individuals of a plant population would be killed, but rapid vegetative reproduction by stolons allowed the population to remain intact. Introduced individuals lived only 4-6 weeks during heavy infestations at the New Canal but lived 6-12 weeks during lighter infestations at Nelumbo Area. Recovery from A. densa infestations and reestablishment of maximum sized plants required 3-5 weeks at the New Canal and Main Canal, and 6-10 weeks at Melton's Pond, Biven's Marsh and the 441 Canal. Lesser damage from the insect was noted at the Nelumbo Area, New Canal, Main Canal, and 441 Canal at other times of the year. Mites periodically damaged the foliage with light to heavy infestations, and grasshopper-like orthopterans grazed heavily on open flowers of introduced plants at Melton's Pond and Nelumbo Area. Concentrations of nutrients in the water at the study sites differed significantly between the study sites during 1975. The mean concentrations of ammonia nitrogen (NH3-N) and total organic nitrogen (org-N) were significantly greater at the New Canal than the other study sites and the mean concentration of total phosphorous (P) was significantly greater at the New Canal than at the 441 Canal, Biven's Marsh, and Melton's Pond (Table 1). Mean concentrations of nitrate nitrogen (NO3-N) were similar at all study sites. Levels of org-N (2.0-7.0 mg/1) and total P (2.2-4.4 tng/1) remained high throughout the year at the New Canal, although NH3-N and NO3-N reached maximum concentrations of 20.0 and 0.7 mg/1, respectively, early in the year and remained constant at low levels after 6 August.

PAGE 40

28 Table 1 Mean concentrations of nitrate nitrogen, ammonia nitrogen, total organic nitrogen (Kjeldahl), and total phosphorous at field sites on Payne's Prairie State Preserve from 22 January 1975 to 7 January 1976.

PAGE 41

29 Mean concentrations of NH3-N and org-N were significantly greater at the Main Canal than at the Nelumbo Area, 441 Canal, Biven's Marsh, or Melton's Pond (Table 1). Mean concentrations of total P at the Main Canal were similar to mean concentrations of total P at the New Canal and the Nelumbo Area, but were significantly greater than total P mean concentrations at the 441 Canal, Biven's Marsh, and Melton's Pond. The maximum concentrations of 1.7 mg/1 NO3-N, 7.0 mg/1 NH3-N, 6.0 mg/1 org-N, and 5.7 mg/1 total P were measured in the Main Canal before the nutrient-rich water from the New Canal was diverted from the Main Canal in early July; only low concentrations of each nutrient were measured after July. Fluctuating water levels and exposed sediments at the Nelumbo Area caused varying amounts of NO3-N (0.0-2.0 mg/1), org-N (0.1-2.2 mg/1), and total P (0.4-4.4 mg/1) before and after the period of drawdown. Mean concentrations of dissolved nitrogen NO3, NH3, and organic at the Nelumbo Area were similar to mean concentrations of dissolved nitrogen at the 441 Canal, Biven's Marsh, and Melton's Pond but mean concentrations of total P were significantly greater at the Nelumbo Area than these study sites (Table 1) Mean nutrient concentrations at the 441 Canal, Biven's Marsh, and Melton's Pond were similar during 1975 (Table 1). At the 441 Canal, NO3-N and total P concentrations ranged from 0.0-0.9 mg/1 and 0.1-2.0 mg/1, respectively, but concentrations of NH3-N and org-N remained relatively constant throughout the year. Only the lowest detectable amounts of nutrients were measured at Biven's Marsh and Melton's Pond during the year, although slight increases in NO3-N and total P occurred at both

PAGE 42

30 sites early in the Spring. Because of the significant differences in nutrient concentrations among the study sites and the size of the plants at the sites, the study sites have been categorized arbitrarily into high, intermediate, and low nutrient sites. According to this classification, the New Canal and Main Canal were the high nutrient sites, Nelumbo Area was the intermediate nutrient site, and the 441 Canal, Biven's Marsh, and Melton's Pond were the low nutrient sites. Initiation of Flowering Several characteristics of flowering in waterhyacinths were initially noted in this study. Waterhyacinths flowered most abundantly in the Spring and Fall in the vicinity of Gainesville, Florida. Flowering exhibited periodicity and zonation patterns at a single site or among sites. Large numbers of flowers opened each day at a given site and were replaced by new flowers the next day. Waves of extensive flowering moved slowly along a canal or formed a peripheral zone around a body of water. The time of initial flowering and cessation of flowering was variable between and within the study sites. Flowers were first noted in early April in the established waterhyacinth populations at the low nutrient sites (Biven's Marsh and 441 Canal). However, flowers did not occur at the intermediate nutrient site (Nelumbo Area) until mid-May or at the high nutrient sites (New Canal and Main Canal) until late August. Introduced waterhyacinth populations at the low nutrient site (Melton's Pond) initiated flowering in late April, whereas flowers were first observed in the introduced

PAGE 43

31 waterhyacinth populations at the intermediate nutrient site in late May and at the high nutrient site in late August. Flowering continued in the observation areas of established waterhyacinths at the low nutrient sites until mid-June (Biven's Marsh) or late July (441 Canal) although continued flowering occurred at these sites outside the observation areas until late August. Established waterhyacinths at the intermediate nutrient site flowered at varying densities throughout the summer and flowering ended in early November. Flowers were observed in the established waterhyacinths at the high nutrient sites until mid-October (New Canal) or mid-November (Main Canal) Cessation of flowering in the introduced waterhyacinths at the low and intermediate nutrient sites occurred in early November after periodic flowering had been observed throughout the Summer and Fall. Flowers were observed only until mid-October in the introduced waterhyacinths at the high nutrient site. Environmental parameters influenced flowering at all study sites. Mean air temperatures at Melton's Pond ranged from 14-22C in early Spring and late Fall, 24-26C from mid-April to late June, and remained between 22-24C during the Summer and early Fall (Table 2). High and low temperatures at Melton's Pond were mediated by the temperature stabilization effects of the water because air temperatures measured at the Gainesville Power Plant were lower in the Winter and higher in the Summer than the temperatures at Melton's Pond. Total precipitation was intermittant and light (0-45 mm) from February to late May, normal (9-116 mm) from late May to late September, and intermittant and light (0-25 mm) through October

PAGE 44

Table 2 Mean temperatures measured at Melton's Pond and the Gainesville Power Plant during the flowering of waterhyacinths at the study sites. Mean Temperature C Date Feb. 7 14 21 28 Mar. 7 14 21 28 Apr. 4 11 18 25 May 2 9 16 23 30 June 6 13 20 27 July 4 11 18 25 Aug. 1 8 15 22 29 Sept. 5 12 19 26 Oct. 3 10 17 24 31 Nov. 7 14 21 28 High 23 27 23 20 26 25 26 28 27 26 28 30 29 28 29 31 31 32 32 28 29 29 26 29 28 29 28 29 28 27 28 27 26 26 29 29 25 27 25 22 19 16 Melton's Pond Low io 17 12 7 15 12 15 16 13 14 17 17 15 18 17 18 19 20 20 19 18 19 19 19 20 18 18 18 18 17 18 17 17 17 20 17 13 18 17 9 8 13 Mean 17 22 18 14 21 19 21 22 20 20 23 24 22 23 23 25 25 26 26 24 24 23 23 24 24 24 23 24 24 22 23 22 22 22 25 23 19 23 21 16 14 15 NOAA Mean 15 21 15 13 20 20 23 18 18 20 23 25 25 24 26 26 27 27 27 27 25 27 27 26 27 28 27 27 18 28 28 27 26 24 26 24 21 23 22 22 15 13

PAGE 45

33 and November (Table 3) The official total precipitation for Gainesville (1250 mm) was approximately 100 mm below normal and the water levels were observed to decrease slowly at all study sites during the year. Weekly totals of solar insolation ranged from 2.85-3.95 Kcal-cm-2 day in April and June, 2.32-3.25 Kcalcm-2 • day-l in March, July, August, and September, and 1.33-3.09 Kcal-cm-2 day~l in February, October, and November (Table 3). Daylengths ranged from 11 hours in early February to 14 hours -' 3 minutes at the summer soltice in June and returned to 10 hours 24 minutes in late November (Table 3) Mean temperatures ranged from 20-24C and mean high temperatures ranged from 27-31C on the dates that flowers were first noted at the low nutrient sites (Biven's Marsh and 441 Canal 2 April, Melton's Pond 30 April, the intermediate nutrient site (the Nelumbo Area 14 May, and the high nutrient sites (Main Canal and New Canal 20 August. Mean low temperatures (13-14 C) at Melton's Pond were lower in early April when flowering was first observed than the mean low temperatures (17-18C) on dates of floral appearance at the other study sites (Table 2) Very low amounts of precipitation (0-9 mm) were recorded for four weeks before initial flowering at Biven's Marsh and 441 Canal, although rainfall was normal (12-37 mm) prior to initiation of flowering at the other study sites (Table 3). Amounts of total solar insolation were similar (approximately 3.0 Kcalcm 2 day-l) were recorded when flowers first appeared at the intermediate nutrient site (Table 3). Daylengths were different

PAGE 46

34 Table 3 Precipitation, solar insolation, and weekly daylength means during the flowering of waterhyacinths at the study sites.

PAGE 47

35 and ranged from 12 hours 30 minutes to 13 hours 48 minutes at the time of initiation of flowering at the study sites (Table 3) Inflorescence production occurred in waterhyacinth populations at all study sites regardless of the distinct differences in vegetative characteristics that occurred among the study sites. Introduced waterhyacinth plants at the high nutrient site, New Canal, possessed mean leaf lengths of the third mature leaf (maximum length 97. cm) that were significantly greater than mean leaf lengths (maximum length 51. cm) at the low nutrient site, Melton's Pond (Figure 3). Mean leaf lengths of introduced waterhyacinths at the Nelumbo Area (intermediate nutrient site) were highly variable (maximum lengths ranged from 45-85. cm) and intermediate between the mean leaf lengths measured at the New Canal and Melton's Pond (Figure 3). Waterhyacinth leaves produced by established plants were significantly longer at the high nutrient sites, New Canal and Main Canal, (maxima of 102 cm and 91 cm, respectively) than at the low nutrient sites, 441 Canal and Biven's Marsh, (maxima of 55 cm and 59 cm, respectively) (Table 4) Means of total leaf lengths of established waterhyacinths at the intermediate nutrient site (Nelumbo Area) were similar to the low nutrient sites from 22 January to 15 May but were similar to the high nutrient sites from 24 July to 5 December (Table 4) Minimum leaf lengths occurred in the populations of established waterhyacinths from 22 January to 24 April when the canopy of leaves was disrupted by frost or cool temperatures. At various times during the year the canopy of both introduced and established waterhyacinth populations was disrupted by extensive damage from Arzama dens a which also

PAGE 48

H 4-1

PAGE 49

37 (mo)

PAGE 50

38 Table 4 Mean total leaf length (lamina plus petiole) of the the third mature leaf of established waterhyacinths at five sites on Payne's Prairie State Preserve. Leaf Length (cm) Date New Canal

PAGE 51

39 decreased mean total leaf lengths. Maximum total leaf lengths occurred at all sites after canopy continuity was reestablished following frost or A. dens a damage. Additionally, mean total leaf lengths of introduced waterhyacinths were at maxima 6-8 weeks after pens at the study sites New Canal (17 April), Nelumbo Area (15 May), and Melton's Pond (24 July) were filled completely with plants (Figure 3) Mean petiole and lamina length and lamina width of the third mature leaf of introduced waterhyacinths also varied according to the nutrient concentration of the sites. Mean petiole lengths (48.9-64.0 cm), lamina lengths (12.9-18.5 cm), and lamina widths (12.7-16.1 cm) were significantly larger at the high nutrient site, New Canal, than mean petiole lengths (3.5-34.3 cm), lamina lengths (3.0-9.1 cm), and lamina widths (4.0-7.3 cm) measured at the low nutrient site, Melton's Pond. Introduced waterhyacinths at the intermediate nutrient site possessed leaves whose means of petiole length (12.8-53.2 cm), lamina length (4.3-12.7 cm), and lamina width (4.5-10.6 cm) were intermediate between or similar to leaf dimensions at Melton's Pond or the New Canal. Petiole width of leaves of introduced waterhyacinths remained constant or decreased during the year, whereas petiole lengths, lamina lengths, and lamina widths on the same plants increased during the year. Decreases in petiole width at the high nutrient site (New Canal) and the low nutrient site (Melton's Pond) occurred in April and July, respectively, concurrently with increases in petiole length at both sites (cf. Figure 3,4). Mean petiole widths

PAGE 52


PAGE 53

41 (uiui) M3PTM 3101:33^

PAGE 54

42 of introduced plants were similar at both high and low nutrient sites from January until July when petiole length increases occurred at the low nutrient site (Figure 4) Mean petiole widths were significantly greater at the high nutrient site (New Canal) than at the low nutrient site (Melton's Pond) after 6 August. At the intermediate nutrient site (Nelumbo Area) mean petiole widths of introduced waterhyacinths were similar to the other sites during the Spring, variable and small during the drawdown at that site, and intermediate between the high and low nutrient sites after water returned to the Nelumbo Area during June and July. Nutrient levels at each site and population changes caused by A. densa were responsible for variations in the mean petiole widths of established waterhyacinths. Statistical comparisons between high and low nutrient sites were difficult because the phenologys of the established waterhyacinth populations were not synchronous. Mean petiole widths of short, swollen petioles were similar at all sites, whereas elongated, nonswollen petioles typically were significantly larger at the high nutrient sites (New Canal and Main Canal) than at the low nutrient sites (441 Canal and Biven's Marsh). The rate of leaf production varied greatly at each site throughout the year although the number of functional leaves per plant remained constant. The production of leaves ranged from 0.0-0.5 leaves per plant per week when the introduced waterhyacinth populations were stressed by A. densa cool temperatures, or floral development; however, rates of 1.5-2.0 leaves per plant per week were measured when the waterhyacinths were not stressed. Average

PAGE 55

43 leaf production rates of 0.92, 0.81, and 0.75 leaves per week were measured at the New Canal, the Nelumbo Area, and Melton's Pond, respectively, during the year. Waterhyacinth plants growing under stable conditions typically possessed 6—7 leaves with intact petioles and laminas. This number remained constant in established waterhyacinth populations at high, intermediate, and low nutrient sites, but decreased during A. dens a infestations and cool temperatures. The maximum root lengths of waterhyacinth plants established at each of the five study sites varied with changes in the percentage of stolons per m2 and the nutrient concentration at each site. Younger plants recruited into the waterhyacinth populations reduced maximum root lengths because the young plants had shorter root systems than the established plants. Maximum root lengths in stable waterhyacinth populations in the low nutrient sites (441 Canal and Biven's Marsh) ranged from 17-58 cm; however root lengths at the high nutrient sites (New Canal and Main Canal) were shorter with maximum lengths of 5-20 cm. Parameters of established waterhyacinth populations, plant density and total dry mass, were not closely related and these parameters varied independently among the study sites because phenological changes at all sites were not simultaneous. Definite increases in the number of stem apices per m2 were noted during the Spring and during A. densa infestations at all sites because of increases in the stolon number per m2. Total dry mass also increased during the Spring but no increases were noted during A. densa infestations. Relatively low plant densities of established

PAGE 56

44 waterhyacinths occurred at each site between A. densa infestations and these populations were characterized by low stolon numbers per m2 and large plant size. The plant densities noted during these periods of population stability and vigorous growth were low at the high nutrient sites (53-89 apices • m ^) intermediate at the intermediate nutrient site (89-109 apices • m~2) and high at the low nutrient sites (97-126 apices • m~2) Although plant densities varied among sites with different nutrient concentrations, total dry mass (1.0-1.9 kg • m~2) seemed similar at all study sites. Variations in vegetative characteristics of a waterhyacinth population at a single site had no effect upon inflorescence production at that site. Inflorescences were initiated at Melton's Pond in late April when plants were small and plant density was low and inflorescence production continued through the Summer and into November when plants were large and plant density was high. Changes in vegetative characteristics because of A. densa damage was related, however, to cessation of flowering at both the 441 Canal and Biven's Marsh. Few changes in vegetative characteristics occurred in waterhyacinth populations at high and intermediate nutrient sites and these changes were not related to inflorescence initiation or cessation. Inflorescence Density The maximum number of inflorescences with open flowers per m2 of surface area of established waterhyacinths varied among the study sites. Correlation of inflorescence numbers with nutrient

PAGE 57

45 levels was impossible because the inflorescences were below the canopy and not visible at the high nutrient sites (New Canal, Main Canal) and the intermediate nutrient site (Nelumbo Area) for all or most of the year. The density of inflorescences with open flowers from established waterhyacinths reached maxima of 4.1-4.8 inflorescences per m2 and 6.3-7.8 inflorescences per m2 at the low sites (Biven's Marsh and 441 Canal, respectively) (Figure 5) and 1.6 inflorescences per m2 at the intermediate nutrient site while inflorescences were visible in May. A maximum of 3.3 inflorescences per m2 was observed in the introduced waterhyacinths at the low nutrient site (Melton's Pond) during late July. During the late Spring and early Summer (25 April-10 June) the established waterhyacinths at Biven's Marsh produced 71 inflorescences per m2 of plants (approximately 120-140) and averaged 1.6 inflorescences per m2 per day. Established waterhyacinths at the 441 Canal produced 175 inflorescences per m 2 of plants (120-160) for an average of 2.5 inflorescences per m2 per day during this same time (25 April to 1 July). Periodicity and zonation patterns of flowering were observed in established waterhyacinths from Biven's Marsh and 441 Canal (Figure 5). Maximum density of inflorescences with open flowers occurred 14 and 18 days apart (8 May and 22 May at Biven's Marsh; 12 May and 30 May at 441 Canal) at these two sites. Intensive flowering moved slowly along the main channel of Biven's Marsh. A high density of inflorescences occurred in the observation area of Biven's Marsh during May, although a high density of inflorescences

PAGE 58

Ml

PAGE 59

hi j\)i aad saDuaDsaaojjuj uado go aaquinj^

PAGE 60

48 was not observed in the marsh itself until raid-June. The marsh was located 150-200 m to the east of the channel. Flowering at the observation site on the main channel of Biven's Marsh had ceased when the density of flowering was greatest in the Marsh, even though the waterhyacinth population and water flow were continuous between the two locations. The density of inflorescences with open flowers in established waterhyacinths in the 441 Canal was lower to the north and to the south of the observation area in May. The density of flowering increased at various locations along the 441 Canal through June, July, and August although few flowers were observed at the study area. Damage from A. densa was widespread along the 441 Canal during September and October and flowering ceased. Introduced waterhyacinths did not flower synchronously in all pens at Melton's Pond. Flowering began in one pen shortly before 1 May and had produced 82 inflorescences by 29 May; however, the other two pens at the site produced only 55 and 12 flowers during May. From 1 May to 7 November totals of inflorescences with open flowers varied in each pen, and each pen had its own periodicity (ranging from 15-25 days) of inflorescence production. No periodicity for density of inflorescences was observed in the established waterhyacinths at the high and intermediate nutrient sites, although zones of high densities of inflorescences occurred during September among the smaller plants that occurred in the mat after A. densa damage. Scattered inflorescences were also observed from May until October on plants rooted in the soil along the dikes on the margins of the

PAGE 61

49 established waterhyacinth populations at the high and intermediate nutrient site. Production of inflorescences by waterhyacinth plants and their ramets followed a random and periodic pattern. Although the genotypes were identical and the plants were in close proximity of each other, similar numbers of inflorescences did not occur in all plants of the same clone during the same period (Table 5) Under the same conditions more flowers were produced by clone #6 of the introduced waterhyacinths at Melton's Pond than by plants of clone #1 during May and June (Table 5). Plants of clone //l, however, produced more flowers than clone #6 from 10 June to 29 September. Similarly, introduced waterhyacinths labeled #4-1 and #5-1 at the Nelumbo Area produced 5 and 4 inflorescences, respectively, from 4 September to 7 November whereas plant #1-1 produced only 1 inflorescence (Table 5) 1 Plant #1-1 and its six ramets (Melton's Pond) produced 16 inflorescences between 10 July and 29 September of which plants #1-1, #1-5, and #1-6 produced only one inflorescence and plants #1-2 and #1-7 produced 4 inflorescences each during the same time period. The initiation and periodicity of flowering were observed in populations growing under different nutrient regimes in a greenhouse experiment. After a month in the initially fertilized water (two tbsp. 20-20-20 fertilizer), the ten initial plants in each tank had produced approximately 100 plants that showed signs of chlorosis (Figures 6 and 7). After the addition of more fertilizer (2 tbsp. 20-20-20 on 29 March to one tank) chlorotic symptoms disappeared from the plants, the number of plants increased to 257, and no

PAGE 62

50 Table 5 The number of senesced inflorescences on waterhyacinths and their ramets at Melton's Pond and the Nelumbo Area on Payne's Prairie State Preserve.

PAGE 63

o a a)

PAGE 64

52 saouaosaaoxjuj gujonpoad suuBxd jo aSemiaoaad ^ m o .„ LH

PAGE 65

a c XI

PAGE 66

5 A S8ouaosaao"[juj Sujonpoad s}ue"[d jo aamueoaaj ^ut?:} oanqx n 3 • I d eoojdu uiorjs jo joquinN

PAGE 67

55 inflorescences were formed until 15 May. Chlorotic symptoms continued in the waterhyacinth plants in the nonf ertilized tank and maxima of 5-10% of the plants produced inflorescences periodically from 5 April to 7 June (Figure 6). After a reduction of the number of plants in the fertilized tank on 19 April, the number of plants increased initially to 125 on 3 May and then slowly decreased. The leaves of the plants in the fertilized tank became chlorotic and maxima of 22% and 12% of the plants produced inflorescences in late May and early June (Figure 7). A total of 119 inflorescences were produced periodically in the nonf ertilized tank in 90 days, whereas a total of 135 inflorescences were produced periodically in only 20 days in the fertilized tank. Waterhyacinth plants in the fertilized tank typically produced two inflorescences per plant whereas plants in the nonf ertilized tank produced only one inflorescence per plant. The same plants in the nonf ertilized tank typically did not flower during successive periods of the maximum densities of inflorescence. The size and density of the waterhyacinth plants increased with the amount of nutrients added to each tank. Leaf lengths of the waterhyacinth plants in both culture tanks ranged from 15-20 cm shortly before 29 March, but the leaf lengths in the fertilized tank increased to 30-40 cm after fertilizer was added 29 March. Leaf lengths of the waterhyacinths in the nonf ertilized tank decreased during May and ranged from 6-12 cm in early June. Inflorescence Development The stages of development of the inflorescence of the

PAGE 68

56 waterhyacinth occurred in the same sequence and over similar periods of time at all study sites. In this study the developing inflorescence was typically macroscopically visible 6-7 days before anthesis (Figure 8) The lamina of the outer bract appeared first and initially was tightly appressed to the stipule-flap of the previous leaf. The lamina of the outer bract did not increase in size prior to or after anthesis. Four days before anthesis the bases of the bracts were visible. Daily elongation of the peduncle proximal to the bases of the bracts was greater than the daily elongation of the outer bract sheath. The latter reached its maximum length 24-48 hours before anthesis, although the inner bract sheath and enclosed colorless flowers did not fill the cavity formed by the outer bract. The aborted lamina located terminally on the inner bract sheath grew, through a distal slit in the outer bract sheath 24 hours before anthesis. The peduncle length proximal to the bases of the bracts was 80-90% of its total length at anthesis, whereas the diameter of the peduncle 1.0 cm proximal to the bases of the bracts remained constant prior to and after anthesis. Between 8 am and 4 pm on the day before anthesis the inner bract and enclosed flowers pushed through the slit in the outer bract sheath. Between 4 and 6 pm the inner bract sheath reached its maximum size and colored flowers were visible through the transparent membranous portion of the sheath. The peduncle from the bases of the bracts to the apex of the inflorescence elongated after 7 pm, pierced the inner bract sheath, and reached its maximum length approximately midnight. The maximum length of the subfloral and

PAGE 69

m

PAGE 71

59 interfloral sections of the peduncle was attained unless all flowers on the inflorescence were not set to open the following morning. The distal 1-10 flowers of the inflorescence frequently did not open on the same day as the remaining flowers of the inflorescence, although no relationship existed between the number of flowers present and the number of flowers that open the first day. The terminal 1-3 flowers usually did not open with the remaining flowers of the inflorescence regardless of the number of flowers per inflorescence. At midnight the flowers were oriented vertically and tightly appressed to the interfloral peduncle. By sunrise on the day of an thesis all flowers that would open that morning were separated from each other and oriented at a 45-60 angle from the axis of the inflorescence. All flowers at a site opened synchronously 1-3 hours after sunrise, although shade or cloud cover delayed flower opening 1-2 hours. Mean sizes of the inflorescences and their appendages and flower number per inflorescence varied greatly among the study sites and were separated into four size classes. Mean sizes of the inflorescences and their appendages and the mean flower number per inflorescence at the high nutrient site (New Canal) were significantly larger than at the low nutrient sites, and similar to or significantly larger than the intermediate nutrient site (Table 6) The mean sizes of the inflorescences and their appendages and the mean flower number per inflorescence were generally smaller at the low nutrient sites (Melton's Pond and Biven's Marsh) than at the intermediate nutrient site (the Nelumbo Area) (Table 6). Similar mean sizes of the inflorescences and their appendages were obtained at Nelumbo Area during

PAGE 72

60 oj en x W QJ a 3 (B E OJ

PAGE 73

61 periods of maximum flowering (30 May and 25 August) However, these characteristics were significantly larger at Melton's Pond during maximum flowering in the Fall (1 October) than during the Spring 19 May). Plants at all sites typically had a mean floral internodal length of 6.4-7.0 cm, although this length was significantly less than at Melton's Pond on 1 October (Table 6). Total inflorescence length ranged from a minimum of 7.5 cm at the low nutrient site (Melton's Pond) on 19 May to maxima of 58. cm and 70. cm at the intermediate nutrient site (Nelumbo Area) on 30 May and the high nutrient site (New Canal) on 25 September, respectively. Dimensions of the outer bract sheath and outer bract lamina ranged from 2.5-17.7 cm in length and 0.7-15.6 cm in length, respectively, between the high and low nutrient sites (Figure 9). The length of the inner bract sheath varied from 3.0-25.2 cm between the high and low nutrient sites whereas the diameter of the peduncle ranged from 0.4-1.7 cm. Inflorescences visible five days (114 hours) before anthesis had attained, typically, 26% of their maximum lengths (Table 7) Young inflorescences were not measured at the intermediate and high nutrient sites during maximum flowering because of low inflorescence density, large plant size, and the herbivory on young inflorescences. Inflorescences at Biven's Marsh and Nelumbo Area typically grew 11% of their total length daily until 18 hours before anthesis (Table 7) At Melton's Pond inflorescences did not demonstrate a regular pattern of growth but they exhibited increasingly higher percentages of growth (5-25%) from minus 114 hours to minus 18 hours before anthesis (Table 7). Inflorescences at all sites grew approximately 28% of

PAGE 74

Figure 9 An inflorescence of Eichhornia crassipes 18 hours before anthesis.

PAGE 75

63 inner bract sheath 3.0 25.2 cm / outer bract lamina 0.7 15.6 cm outer bract sheath 2.5 17.7 cm peduncle diameter .4 1.7 cm

PAGE 76

64 Table 7 Growth in length of a developing inflorescence as a percentage of total inflorescence length of introduced and established waterhyacinths at high, intermediate, and low nutrient study sites on Payne's Prairie State Preserve.

PAGE 77

65 their total length during the last 18 hours before anthesis, although this increase in length ranged from 1.5 cm at the low nutrient site (Melton's Pond) to 35. cm at the high nutrient site (New Canal) (Table 7). The distance from the bases of the bracts to the apex of the peduncle ranged from 39-50% of the total inflorescence length at all sites (Table 8). This distance was equally split between the length of the subfloral peduncle (from bases of the bracts to the lowermost flower) and the length of the interfloral peduncle (from the lowermost flower to the peduncle apex) (Table 8). Length and width of the lower sepal and the banner petal varied among sites and within an inflorescence, but no relationship to nutrient levels at a site was noted. In all flowers the lower sepal length and width ranged from 27-37 mm and 10-17 mm, respectively (Figure 10). The smallest mean sepal lengths in lower, middle, and upper flowers were found at a low nutrient site (Melton's Pond), whereas the longest mean lengths occurred at another low nutrient site (Biven's Marsh) (Tables 9,10,11). Mean sepal lengths at the higher nutrient sites (Nelumbo Area and New Canal) were similar and intermediate between the two low nutrient sites. Length and width of the banner petal of all flowers ranged from 30-43 mm and 19-29 mm, respectively (FigurelO). The smallest mean lengths of banner petals of the lower and middle flowers were found at the intermediate nutrient site (Nelumbo Area) (Tables 9,10,11). The other sites, Biven's Marsh, Nelumbo Area, and New Canal had intermediate mean lengths of banner petals. The minimum and maximum banner petal lengths occurred at the low nutrient sites, whereas banner petals from the intermediate

PAGE 78

66 Table 8 Length of peduncle as a percentage of the total inflorescence length of established and introduced waterhyacinths at high, intermediate, and low nutrient study sites on Payne's Prairie State Preserve.

PAGE 79

d

PAGE 80

68 GO

PAGE 81

69 Table 9 Comparative data of floral appendages of lower flowers of the inflorescence of introduced and established waterhyacinths at high, intermediate, and low nutrient study sites on Payne's Prairie State Preserve. Lower Sepal Banner Petal Site Date length/width length/width Long stamens Style length length Melton' s Pond (low) Biven' s Marsh (low) Nelumbo Area (intermediate) New Canal (high) May 19 32. a 13. Oct. 1 33. a 14. Apr. 29 36. b 14. May 30 34. ab 16. Aug. 25 33. a 15. Sept. 25 34. ab 15. 35. a x 22. 37. ab 22. 39. be 23. 16. b 16. b 17. b 11. 10. 10. 41. c

PAGE 82

70 Table 10 Comparative data of floral appendages of middle flowers of the inflorescence of introduced and established waterhyacinths at high, intermediate, and low nutrient study sites on Payne's Prairie State Preserve. Site Lower Sepal Banner Petal Date length/width length/width Long Stamen length Style length Melton' s Pond (low) Biven' s Marsh (low) Nelumbo Area (intermediate) New Canal (high) May 19 30. a x 12. 35. a 22. Oct. 1 31. a 13. 36. a 24. Apr. 29 34. b 12. 38. be 22. May 30 33. b 16. 40. c 25. Aug. 25 33. b 14. 37. ab 23. Sept. 25 34. b 14. 38.be 25. 16. b 16. b 17. b 17. b 9. ay 18. b 18. b X Means in each column not followed by the same letter are significantly different at 5% level. 11. 10. 10. 12. 22. 10. 10. Y Long-styled flower form.

PAGE 83

71 Table 11 Comparative data of floral appendages of upper flowers of the inflorescence of introduced and established waterhyacinths at high, intermediate, and low nutrient study sites on Payne's Prairie State Preserve.

PAGE 84

72 and high nutrient sites exhibited intermediate mean lengths (Table 11). Mean banner petal lengths of lower and middle flowers from low, intermediate, and high nutrient sites were significantly larger than or similar to mean banner petal lengths of the upper flowers at the same site (Table 12). Mean style lengths (9-12 mm) and mean length of the long. stamens (16-18 mm) were similar in the mid-styled flowers at all study sites (Tables 9,10,11). Long-styled flowers were only observed at the Nelumbo Area and exhibited a mean style length of 22 mm and a mean length of long stamens of 9 mm. Short stamens were the same length (4-5 mm) in all flowers. All open flowers on an inflorescence began to close after 5-6 pm and were completely closed by sunset (7-9 pm) After anthesis the perianth lobes and style slowly folded into an amorphous mass distal to the developing ovary and persisted until fruit abscission and dehiscence. The perianth tube remained intact and covered the developing fruit until dehiscence. Twelve to eighteen hours after anthesis two bends appeared in the peduncle (Figure 8) The first or capital bend was proximal to the bases of the bracts and was initiated at a different distance from the bases of the bracts on small inflorescences (1.5-2.0 cm) than on large inflorescences (2.5-3.0 cm). The capital bend developed whether or not all flowers on the inflorescence had opened and lowered the floral axis 70-120 from its original vertical orientation. The unopened flowers in an inflorescence opened the following day with the subfloral and interfloral portions of the

PAGE 85

73 Table 12 Mean length of the banner petal of waterhyacinth flowers borne in the lower, middle, or upper positions on the inflorescences at high, intermediate, and low nutrient study sites on Payne's Prairie State Preserve. Site Melton's Pond (low) Nelumbo Area (intermediate) New Canal (high) Date May 19 May 30 September 25 Banner petal length mm flower position lower middle upper 35. a x 35. a 34. a 41. c 40. c 37. b 39. be 38. b 35. a X Means of all sites not followed by the same letter are significantly different at 5% level.

PAGE 86

74 peduncle in a horizontal position. The interfloral peduncle tissue also exhibited a second bend by the morning of the following day. This interfloral bend formed above the lowermost flower and had maximum bending (arc-shaped) from the middle flowers to the peduncle apex, although the bend was not as acute as the capital bend. The day after anthesis the capital bend reoriented the floral axis from 150-180 from its vertical orientation at anthesis (Figure 8). Slight elongation (less than 10% of total length) of the peduncle occurred after anthesis until the capital bend was complete. The bending process in the capital bend occurred acropetally. The distance from the bases of the bracts to the upper portion of the capital bend was 2.5-3.0 cm shortly after anthesis and decreased to 0.3-0.7 cm after 48 hours. A third or basal bend occurred in the base of the peduncle (3-10 mm distal to its insertion in the stem) and reoriented the peduncle 10-30 from its normal vertical orientation the second day after anthesis. This bend reoriented the peduncle 90 or more from its vertical orientation 4-10 days after anthesis. The basal bend of the peduncle together with the capital and interfloral bends carried the inflorescence from the central crown of leaves and oriented the developing fruit horizontally below the water surface. The curvature of the peduncle was completed 72 hours past anthesis, although cooler temperatures decrease the rate of curvature. Orientation of the three bends in the peduncle during postanthetic curvature was generally related to the location of the bracts on the inflorescence. The orientation of the capital

PAGE 87

75 bend also determined the orientation of the interfloral and basal bends. The inflorescences (53%) were reoriented to the water by the postanthetic bends at right angles to the stem axis with some degree of variance (Figure 11) whereas the remaining inflorescences were reoriented across adjacent leaves (38%) or the stem apex (9%). The orientation of the postanthetic bends among the low, intermediate, and high nutrient sites varied and no correlations were noted (Figure 11) Pollination and Seed Set Various insects (Hymenopterans) pollinated waterhyacinth flowers at both Melton's Pond and the Nelumbo Area. Each of the insect species approached and landed on the waterhyacinth flowers (both midand long-styled flower forms) in different ways. Each species visited many flowers on the same inflorescence and large amounts of pollen were observed on the stigmas of the flowers. Developing fruits were easily observed 8-10 days after pollination because of the increased size and firmness of the perianth tube and because nonpollinated flowers always abscised from nonpollinated inflorescences. Nonpollinated flowers on pollinated inflorescences typically did not abscise until the mature fruits abscised. All stages of fruit development and dehiscence usually occurred at or below the water surface, although fruits occasionally matured totally out of the water. When the fruits were mature the brown perianth tube split longitudinally and loculicidal dehiscence of the capsule occurred. Seeds were

PAGE 88

Figure 11 Comparative data on the orientation of the capital bend to the origin of the outer bract.

PAGE 89

77 Side View of Inflorescence Angle of bend relative to origin of the outer bract Site

PAGE 90

78 released through a split in the perianth tube or after agitation or decomposition of the fruit, and sank immediately to the substrate below the water. Development of the fruit required 18-22 days during the Summer after artificial pollination of flowers in the field and in the greenhouse, and 24-28 days during the Fall and Winter after artificial pollination in the greenhouse. Artificial pollination (self and cross) produced seed set (50-150) seeds per capsule in approximately 150 inflorescences during preliminary studies in 1974, whereas no appreciable seed set was found in 50 nonpollinated inflorescences. No relationships between fruit formation and either time of the day (9:00 and 11:00 am, 1:00 and 3:00 pm) or time of the year (February through August) for pollination were noted. Fruits that resulted from natural pollination were only found in established waterhyacinth populations at the intermediate nutrient site (Nelumbo Area) and at a low nutrient site (Biven's Marsh). Approximately thirty and fifteen fruits per m2 were collected on 24 July and 24 August, respectively, at Nelumbo Area and only six fruits per m2 were found only on 22 May at Biven's Marsh. No additional fruits were found in established waterhyacinth populations at the low nutrient sites (Biven's Marsh, 441 Canal), intermediate nutrient site (Nelumbo Area), or the high nutrient sites (New Canal, Main Canal) at any other time during the year. Although the seed number per capsule of the fruits collected at the Nelumbo Area and Biven's Marsh varied from 12-175, most capsules contained 25-65 seeds. Inflorescences

PAGE 91

79 of introduced waterhyacinths (6 per site) were artificially pollinated (selfed) at the low, intermediate, and high nutrient sites in the Summer and early Fall. At the low nutrient site (Melton's Pond) 165 seeds per inflorescence were collected from the four inflorescences that produced seed. At the intermediate nutrient site (Nelumbo Area) 380 seeds per inflorescence were collected from each of the six inflorescences, although no correlation to the number of seeds per capsule could be made. Herbivores destroyed four of the pollinated inflorescences at the New Canal and the remaining two inflorescences were not relocated. Natural seed germination of Eichhornia crassipes was observed at two locations on Payne's Prairie State Preserve near the study site at the intermediate nutrient site (Nelumbo Area). During May seedlings of E. crassipes were observed with seedlings of Pontederia cordata and Hydrocotyle ranunculoides on moist sediments. During July thousands of seedlings were found floating in 30-40 cm of water among Polygonum and Amaranthus plants 2-3 weeks after water had returned to the Nelumbo Area study site; however, no seedlings were noted in the sediments below the surface of the water. The youngest seedlings had only 3-5 linear leaves (2 cm long and 0.5 cm wide) and no adventitious roots, whereas the older seedlings had adventitious roots, 5-9 leaves, but no stolons. By late July only a few young seedlings (5-10 leaves) were observed and the remaining seedlings had produced vigorous root systems and secondary and tertiary stolons. Seedlings produced flowers 8-10 weeks after seed germination occurred and continued until early November.

PAGE 92

no Twenty— five floating seedlings of various ages (5-20 leaves) were collected from a site near the Nelumbo Area study site in midJuly and placed in a greenhouse culture tank. Development of the younger and older seedlings was normal and the culture tank was filled with ramets (approximately 125 individuals) after six weeks. Flowering occurred eight weeks after introduction of the plants into the greenhouse tank. Anatomy of the Inflorescence of Eichhornia Crassipes And Its Appendages At anthesis the inflorescence of E. crassipes is fully developed and its appendages are clearly evident (Figure 8) Distally the mature peduncle is terete (.5-1.7 cm wide) although it is slightly compressed proximal to and at its point of insertion on the stem. The bases of the outer and inner bracts are located at the midpoint of the inflorescence and 5-35 cm distal to the insertion of the inflorescence. The sheathing base of the outer bract (3.-18. cm long) surrounds both the subfloral peduncle and the inner bract for one-half the length of the sheathing base. The body of the sheathing base of the outer bract terminates in a short petiole that connects the sheathing base to the subcordate lamina (0.7-15. cm long), whereas the membranous lateral portions of the sheathing base terminate in a small stipule-flap (Figure ] 2) The inner bract sheath (3.5-25. cm long) occurs opposite to the outer bract and surrounds the subfloral peduncle for one-half the length of the sheathing base. The body of the sheathing base of the inner bract terminates in a short petiole

PAGE 93

14-1

PAGE 94

32 (/> SI

PAGE 95

83 that connects the sheathing base to the aborted lamina (0.3-2.0 cm long) whereas the membranous lateral portions of the sheathing base of the inner bract terminate in a small stipule-flap (Figure 12) The lavender flowers (5-32 per inflorescence) are arranged in a 3/8 phyllotaxy similar to that of the leaves. The lowermost flower is located between the upper portions of the sheathing bracts, and approximately one-half the distance from the bases of the bracts to the tip of the interfloral axis. Each succeeding flower is inset into the interfloral peduncle at regular intervals. The shape of the interfloral peduncle is terete with a single large indentation at the level of the lowest flower, forms an irregular polygon at the level of the middle and upper flowers and terminates typically in a small triangular extension of peduncle tissue (0.3-2.0 cm long) that continues beyond the base of the terminal flower. Individual flowers have three sepals and three petals which are fused for one-third of their length into a perianth tube. The sepals are lancolate, shorter than the petals, and possess a dense layer of glandular trichomes on their abaxial surface. The ovate petals are relatively broad and possess small amounts of glandular trichomes on their abaxial surface. All the perianth members are similar in color (lavendar) and the banner petal has a rhomboidal yellow-gold spot centered on a large purple area. The short stamens are adnate to the banner petal and the upper two sepals, whereas the long or mid-length stamens are adnate to the lower petals and the lower sepal. Anthers of the six stamens are dorsifixed with pollen discharge from longitudinal slits; however, pollen discharge

PAGE 96

84 is extrorse from the short stamens and introrse from the long stamens. The filaments of short and mid-length stamens are white, whereas the filaments of long stamens exhibit purple coloration distally; all filaments have giandular hairs along their length (Figure 13) The conical ovary at the base of the flower has a long or mid-length style that terminates in a vertically oriented papillate stigma. Both long and mid-length styles have glandular hairs along their length; mid-length styles are white, and the long styles exhibit a purple coloration distally. Transections of the mature peduncle of E. crassipes show several characteristic features (Figure 14). The epidermis is composed of rectangular parenchyma cells whose external periclinal wall is covered by a thin cuticle. Chlorenchyma cells (3-5 layers) of the outer zone of ground tissue occur internal to the epidermis and are more frequent external to the peripheral series of vascular bundles that occur in the outer zone of ground tissue. Larger thinwalled parenchyma cells may be interspersed with the chlorenchyma cells between the peripheral series of vascular bundles. Large idioblasts with raphid crystals or tanniferous substances frequently are found interspersed in the chlorenchyma and parenchyma cells of the outer zone of ground tissue. Large parenchyma cells (2-4 layers) occur internal to the chlorenchyma cells of the outer zone of ground tissue. Aerenchyma cells comprise the remaining inner zone of the ground tissue of the peduncle. Large intercellular spaces (0.6 mm in diameter and 2.0 mm in length) are interspersed throughout the ground tissue. These intercellular spaces are delineated by elongated

PAGE 97

Figure 13 Stamen filament of _E. crassipes X 100. Figure 14 Transection of a peduncle distal to its insertion on the stem. X 11. Figure 15 Transection of a vascular bundle of a peduncle. X 167. Figure 16 Transection of a peduncle proximal to the bases of the inflorescence bracts. X 11. a aerenchyma; ch chlorenchyma; cl central lacuna; e epidermis; id idioblasts; ob outer bract; pf perivascular fibers; pp primary phloem; px primary xylem; t trichomes; vb vascular bundles.

PAGE 98

o( —*7T7 1*0 i ch — ^ P vb Ilr5fi!>^7^ <%&*i&? j* 16 Ok '"^^3^ •-N.. <

PAGE 99

87 parenchyma cells that are penetrated occasionally by large styloid crystals. Vascular bundles occur randomly in both the parenchymatous outer zone and the aerenchymatous inner zone of the ground tissue. These collateral vascular bundles (approximately 150 per transection) are surrounded by a layer of large parenchyma cells and are composed of primary phloem which is external to primary xylem. The collateral vascular bundles have 1-2 tracheids with annular to helical wall thickenings, a conspicuous protoxylem lacuna, and 3-6 sieve tube members with companion cells (Figure 15) Perivascular fibers with poorly developed secondary walls occur external to the primary phloem, especially in the peripheral bundles. Large idioblasts with tanniniferous substances frequently are found parallel to the vascular bundles among the surrounding parenchyma cells. A central lacuna (2.-5. mm in diameter) occurs in the center of the peduncle. Anatomical features of the peduncle remain constant throughout its length, although changes do occur at the bases of the inner and outer bracts (Figure 16). Proximal to the bases of the bracts, the intercellular spaces decrease in length and diameter and the central lacuna disappears. Vascular bundles in the ground tissue increase in number at the bases of the bracts. The collateral vascular bundles are surrounded by a layer of large parenchyma cells and are composed of primary xylem (1-2 tracheids with helical thickenings and a protoxylem lacuna) and primary phloem (3-6 sieve tube members and companion cells). Perivascular fibers occur external to the primary phloem. The peripheral vascular bundles of the peduncle continue directly into the sheathing base of the outer bract. Some vascular bundles

PAGE 100

move laterally into the median portion of the sheathing base of the outer bract. Distally, other vascular bundles move laterally into the median portion of the sheathing base of the inner bract. A single-layered epidermis with a cuticle is external to the ground tissue of the sheathing base of the outer bract (Figure 17) Chlorenchyma cells (3-4 layers) occur internal to the exterior (abaxial) epidermis although only a single layer of parenchyma cells occur internal to the interior (adaxial) epidermis. Aerenchymatous ground tissue occurs between the chlorenchyma and parenchyma cells of the outer zone of ground tissue (Figure 17) Large intercellular spaces separate the 3-4 layers of aerenchyma cells in the median portion of the sheathing base, whereas only 1-2 layers of aerenchyma cells separated by large intercellular spaces occur in the fused margins of the sheathing base. Vascular bundles are distributed throughout the aerenchymatous ground tissue with one series of bundles interior to the outer (abaxial) layers of ground tissue, and one series of bundles interior to the inner (adaxial) layers of ground tissue (Figure 17) Some vascular bundles are also between the outer and inner series of bundles in the thickest portion of the sheathing base. These collateral vascular bundles possess primary xylem (1-2 tracheids with annular thickenings and a protoxylem lacuna) which is interior (adaxial) to the primary phloem (2-4 sieve tube members with companion cells) Perivascular fibers with poorly developed secondary walls also occur within the series of large parenchyma cells that surround the vascular bundle.

PAGE 101

Figure 17 Transection of the sheathing base of the outer inflorescence bract. X 47. Figure 18 Transection of the fused margins of the sheathing base of the inner inflorescence bract. X 113. Figure 19 Transection of the aborted lamina of the inner inflorescence bract. X 113. Figure 20 Transection of the functional lamina of the outer inflorescence bract. X 150. a aerenchyma; ch chlorenchyma; e epidermis; g ground tissue; ie inner epidermis; le lower epidermis; m mesophyll; oe outer epidermis; pf perivascular fibers; pm palisade mesophyll; pp primary phloem; px primary xylem; st stomate; ue upper epidermis; vb vascular bundle.

PAGE 102

0 1 7 fm f -Mi 1*3* a -• a 19 '<&&#>* ( -j m is 20 \ v <*. j s\ !. > 1 .*

PAGE 103

91 Anatomical features of the sheathing base of the inner bract are similar to features of the outer bract, although the sheathing base of the inner bract is composed of fewer cell layers. A single-layered epidermis with cuticle is external to the ground tissue of the sheathing base of the inner bract. Internal to the exterior (abaxial) epidermis 2-3 layers of chlorenchyma cells occur and 1-2 layers of parenchyma cells occur internal to the inner (adaxial) epidermis. The aerenchymatous ground tissue of the sheathing base of the inner bract is separated by a single layer of large intercellular spaces in the median portion of the base although the aerenchymatous ground tissue is lacking in the fused lateral membranous portions of this structure (Figure 18). Vascular bundles are located in a single series at regular intervals throughout the ground tissue of the sheathing base. These collateral vascular bundles are composed of primary xylem (1-2 tracheids with annular thickening and a protoxylem lacuna) which is adaxial to the primary phloem (2-4 sieve tube members with companion cells) The vascular bundles are surrounded by a layer of large parenchyma cells The median regions of the sheathing bases of both the inner and outer bracts become thicker andmore cylindrical proximal to the apex of the sheathing bases. The reflexed margins of the previously fused lateral regions of the sheathing bases of the inner and outer bracts decrease in size acropetally (Figure 12) The vascular bundles of the lateral membranous portions of the sheathing base of both the inner and outer bracts continue acropetally and terminate in a small stipule-flap at the apex of the lateral sheathing portion

PAGE 104

92 of the sheathing bases. The vascular bundles of the median region of the sheathing base of both bracts move acropetally through the cylindrical petiole and continue either into the aborted lamina that terminates the inner bract or into the functional lamina of the outer bract. Anatomical features of the functional lamina of the outer bract demonstrate that it has differentiated cell layers similar to lamina of leaves of E. crassipes The aborted lamina of the inner bract has anatomical characteristics similar to the sheathing base. A single layer of parenchyma cells form the epidermal layer that surrounds the ground tissue of the aborted lamina (Figure 19) The ground tissue has an exterior zone of 3-5 layers of chlorenchyma cells and an interior zone of aerenchyma with a few large intercellular spaces. Vascular bundles (15-25) are randomly dispersed in the ground tissue and are surrounded by a layer of large parenchyma cells (Figure 19) The vascular bundles contain 1-2 primary xylem tracheids with annular thickening, a protoxylem lacuna, and 2-3 sieve tube members with companion cells. A single layer of parenchyma cells comprises the upper (adaxial) and lower (abaxial) epidermal layers of the functional lamina of the outer bract (Figure 20) Palisade mesophyll cells occur in 2-3 layers interior to the upper epidermis and in 1-2 layers interior to the lower epidermis. These layers of palisade mesophyll occur in longitudinal rows and are separated by the veins of the lamina (Figure 20). A layer of large intercellular spaces separates the adaxial layer of palisade mesophyll from the abaxial layer of

PAGE 105

93 palisade mesophyll. The layer of intercellular spaces forms a single series at the margins of the lamina, but increases to 2-4 layers in the median region of the lamina. Vascular bundles are randomly dispersed throughout the aerenchyma ground tissue in the median portion of the lamina and are located within the upper and lower palisade mesophyll regions throughout the lamina. These collateral vascular bundles are composed of primary xylem (1-2 tracheids with annular to helical thickening and 1-2 protoxylem lacunnae) and primary phloem (2-4 sieve tube members and companion cells) Perivascular fibers occur external to the primary phloem and the vascular bundle is surrounded by a series of large parenchyma cells. Vascular bundles located in the adaxial palisade mesophyll or in the aerenchymatous ground tissue have primary xylem adaxial to primary phloem. Vascular bundles located in the abaxial palisade mesophyll have an inverted orientation of primary xylem abaxial to primary phloem. The anatomy of the subfloral and interfloral peduncle is similar although the interfloral peduncle decreases in diameter acropetally. A singlelayered epidermis and cuticle forms the outer boundary of both the suband interfloral peduncle. Internal to the epidermis, small, thin-walled parenchyma cells (4-5 layers) comprise the outer zone of the ground tissue. This outer zone of ground tissue is interspersed with raphide and tanniniferous idioblasts. The inner zone of ground tissue is composed of aerenchyma cells that surround large intercellular spaces. The center of the subfloral peduncle and the proximal half of the interfloral peduncle is occupied by a central lacuna.

PAGE 106

94 Collateral vascular bundles are randomly distributed throughout the inner zone of aerenchymatous ground tissue of the sub— and interfloral peduncles. The vascular bundles contain 1-3 primary xylem tracheids with annular to helical wall thickenings, a protoxylem lacuna, and 3-7 sieve tube members with companion cells in the primary phloem. Primary xylem is internal to primary phloem in all vascular bundles. Perivascular fibers with poorly developed walls occur peripheral to the primary phloem. These fibers are distinct in regions of the subfloral peduncle, but are not well developed dis tally in the interfloral peduncle. A single layer of large, thin-walled parenchyma cells surrounds each vascular bundle. Some of the vascular bundles of the subfloral peduncle terminate in the vascular plexus of the first flower or simply branch into the first flower. The remaining vascular bundles of the interfloral peduncle continue acropetally and branch into or terminate in the vascular plexus of succeeding flowers. The number of vascular bundles per transection remains constant through 1/3—1/2 of the bases of the proximal flowers and decreases dis tally with each succeeding flower. Although most vascular bundles terminate in the vascular plexus of the terminal flower, 1-2 vascular bundles typically proceed into the terminal portion of the interfloral peduncle and terminate at the apex of the interfloral peduncle. The interfloral peduncle proximal to the base of a flower is characterized anatomically by small intercellular spaces and an increased frequency of diaphragm cells. Proximal to the vascular plexus of the flower the interfloral tissue becomes parenchymatous

PAGE 107

95 with no large intercellular spaces. Vascular bundles (16-20) from the interfloral peduncle proceed to the vascular plexus of the mature flower. The vascular plexus of a mature flower is a three-dimensional ring of vascular tissue (Figure 21) Primary xylem with annular and helically-thickened tracheids and primary phloem are noted within the plexus. A few vascular bundles (4-6) located on the periphery of the interfloral peduncle proceed directly into the perianth tube and are integrated into sepal or petal vasculature. A median and two marginal sepal traces proceed radially from the vascular plexus at a proximal level into each of the three sepals (Figure 22,23). A median and two marginal bundles also proceed radially from the plexus distal to and between the sepal traces into each of these petals (Figure 22,23). Stamen traces do not arise independently from the vascular plexus, but rather are conjoint with the median traces of each of the perianth members. The six marginal bundles of the carpels separate dis tally from an ill-defined vascular plexus between the median bundles of the perianth and proceed inward to the center of the flower (Figure 22, 23) Median carpellary traces separate and proceed slightly inward from the vascular plexus at the locus of separation of median sepal traces. The separation and movement of the marginal carpellary bundles from the plexus occurs concurrently with the separation and movement of the median carpellary bundles. Separation of the locule begins between the median carpellary trace and the marginal traces at the level of vertical orientation of

PAGE 108

Figure 21 Transection of the vascular plexus of a mature flower of E. crassipes X 49. Figure 22 Transection of the origin of the vascular traces of a mature flower of E_. crassipes X 49 Figure 24 Transection of the base of an ovary of a mature flower. X 52. Figure 25 Transection of an ovary between the base of the carpels and the placentae. X 75. 1 locule; lc marginal carpel trace; lp marginal petal trace; Is marginal sepal trace; mc median carpel trace; mp median petal trace; ms median sepal trace; ov ovule; s stamen trace; sc stylar canal.

PAGE 109

97

PAGE 110

c

PAGE 111

99 c3 -H 01 O

PAGE 112

100 the carpellary traces (Figure 24) The gynoecium is separated from the fused perianth tube by separation of the compact layers of parenchyma cells exterior to the carpellary wall. The external carpellary wall has an external, single-layered epidermis, 4-7 layers of compact, thin-walled, and isodiametric parenchyma cells with small infrequent intercellular spaces interspersed with tanninif erous idioblasts and an internal, single— layered locular epidermis (Figure 26). The intercarpellary wall also has 4-7 layers of compact thinwalled parenchyma cells between the locular epidermal layers of each carpel, although these parenchyma cells deteriorate distally and only the locular epidermis remains intact. Stylar canals are first formed in the intercarpellary wall below the placental region and are lined with a single layer of transmitting tissue composed of large cells with densely staining contents (Figure 25). These stylar canals do not occur beyond the base of the placentae because all the parenchyma cells of the intercarpellary wall have deteriorated. A two-lobed placenta originates in each carpel and bears anatropous ovules. A deep division then occurs distally within the placenta of each carpel and forms two distinct placentae (Figure 26). The epidermal layer that divides the two placentae of each carpel is composed of large, rectangular parenchyma cells with densely staining contents, and occurs from the placentae through the hollow style to the stigma. The two marginal vascular bundles of each carpel proceed acropetally from the base of the carpel, undergo multiple divisions, and form large vascular bundles (Figure 25). As the marginal carpel bundles proceed distally, the number of

PAGE 113

Figure 26 Transection of the placental region of an ovary. X 47. Figure 27 Transection of a flower distal to the apex of the ovary. X 56. Figure 28 Longisection of the convex side of the capital bend in the peduncle. X 277. Figure 29 Longisection of a peduncle before postanthetic curvature. X 275. ch chlorenchyma; e epidermis; f filament; lc marginal carpel trace; mc median carpel trace; ov ovule; pi placenta; s stamen trace; sy style; tt transmitting tissue.

PAGE 114

102 MM sy • i ft £*?V mc tt

PAGE 115

103 primary xylem tracheids with helical wall thickening in the vascular bundle increases and 20-40 tracheids occur in a single bundle proximal to the origin of the placentae. The tracheids are randomly dispersed in the bundle among 30-60 densely staining immature sieve tube members. The two placentae separate distally and the two placentae from adjoining carpels appear as a single placenta. However, the remaining locular epidermal layers of the intercarpellary wall maintain separation of the locules and placentae. Branches of the marginal carpellary bundles proceed into the placentae distally. The marginal bundles of each carpel decrease in size distally, fuse into three bundles (from the fusion of marginal bundles from adjoining carpels) proximal to the apex of the carpels, and terminate at the apex of the carpels. The median vascular bundles of each carpel remain distinct, continue acropetally from the base of the carpel through the carpellary wall (Figure 27) into the three-lobed, threelocule, hollow style, and divide at the base of the stigma. Each branch of the median vascular bundle terminates in a papillate stigmatic lobe. Six stamen trases separate from the median trace of the sepals and petals at the level at which the gynoecium is free from the corolla tube (Figure 24). These traces remain within the fused perianth distally to the apex of the carpels. At the apex of the carpels a filament separates from the banner petal and is followed shortly by the separation of filaments from the two upper sepals (Figure 27). Distally a filament separates from the lower sepal followed by the separation of the remaining filaments from the

PAGE 116

104 lower petals. The first three stamens form the short-lengthed group, whereas the latter three stamens represent the mid-length or long stamens. All filaments terminate in two-lobed, four-locular anthers with a well-developed endothecium. Tanninif erous and raphide idioblasts occur frequently in the parenchymatous ground tissue of the filament (Figure 13). A single vascular bundle occurs in the center of a filament and is collateral with primary xylem interior to primary phloem throughout most of its length. The vascular bundle of the filament becomes weakly bicollateral to concentric only near its distal termination. Primary xylem tracheids (6-14) with helical wall thickening and primary phloem sieve tube members (5-20) with companion cells comprise the vascular bundle in a typical anther filament. The perianth tube remains similar anatomically from its base to the point of separation of the sepals and petals. Each perianth member has a cuticle and epidermis of a single layer of parenchyma cells as an outer boundary on its adaxial and abaxial surface (Figure 27). An external zone of the ground tissue of the perianth member is internal to the epidermis and is composed of thin-walled, isodiametric parenchyma cells in 4-6 layers interior to the abaxial surface and in 3-4 layers interior to the adaxial surface (Figure 27) Tanninif erous idioblasts are randomly dispersed throughout the external zone of ground tissue, although idioblasts occur at a greater frequency in the abaxial zone than in the adaxial zone. The inner zone of ground tissue of the perianth members is composed of aerenchyma cells separated by a single series of intercellular

PAGE 117

105 spaces (0.025-0.10 mm in diameter). The median vascular bundles of each perianth member and the adnate stamen traces are surrounded by 2-3 layers of large thin-walled parenchyma cells, whereas the remaining lateral and branch vascular bundles are surrounded by a single layer of large, thin-walled parenchyma cells. Vascular bundles with primary xylem adaxial to primary phloem occur in a single series throughout the outer zone of ground tissue that is interior to the abaxial surface. Smaller vascular bundles with an inverted orientation of primary xylem abaxial to primary phloem occur in a single series throughout the outer zone of ground tissue that is interior to the adaxial surface. These collateral vascular bundles are composed of 1-2 primary xylem tracheids with annular to helical wall thickening and 2-4 sieve tube members with companion cells in the primary phloem. At the level of separation of the perianth members the cells of abaxial ground parenchyma have decreased to 2-3 layers, although the number of layers of adaxial ground parenchyma remain constant. Intercellular spaces occur proximal to the median vascular bundle of each perianth member, although the intercellular spaces decrease in size and are lacking in the lateral regions of each perianth member. Distally, the thin-walled parenchyma cells of the ground tissue interior to the adaxial and abaxial surfaces decrease to one layer of cells and no large intercellular spaces are present. The apices of the perianth members are 3-4 layers of small, isodiametric parenchyma cells thick. Postanthetic inflorescence development in the waterhyacinth involves growth phenomena that bring the flowers to the surface of

PAGE 118

106 the water for fruit deyelopraent. The initial capital bend determines the orientation of the interfloral and basal bends but has no specific orientation itself (cf. Figure 11). The rectangular, thin-walled parenchyma and chlorenchyma cells of the outer zone of ground tissue and the epidermis on the convex side of the peduncle are longer (0.050-0.135 mm) than on the concave side of the bend (0.035-0.090 mm) (Figure 28) The relative or actual differences in cell size are determined by the degree of bending because an erect peduncle has typical cell lengths of 0.030-0.100 mm (Figure 29). Increases in cell size on the convex side of the basal bend (0.040-0.110 mm) are not as great as increases in cell size on the convex side of the capital bend. Only slight variation in cell size accompanied the interfloral bend because the arc of the bend is gradual. No morphological or anatomical changes occur in the outer and inner bracts during postanthetic development, and these structures remain intact until the inflorescence deteriorates and decomposes basipetally. At dusk on the day of anthesis all open flowers begin to wilt and close by a random inrolling and fusion of the perianth lobes, and form an amophous mass distal to the ovary and perianth tube. This amorphous mass includes the stigma, the distal portion of the style, and the perianth lobes, but does not contain either the stamens or the perianth tube. The former floral appendages dry and deteriorate, whereas the latter structure remains intact throughout fruit development. If pollination occurs in most flowers on the inflorescence, all flowers remain attached to the interfloral peduncle for 18-28 days. If pollination does not occur

PAGE 119

107 in any of the flowers all flowers abscise from the interfloral peduncle 4^5 days after anthesis. Three daysafter pollination the parenchyma cell layers and the large intercellular spaces of the perianth tube and carpellary walls are still intact. Twelve days after pollination cells of the parenchymatous ground tissue of the perianth tube and carpellary wall become crushed. The inner (adaxial) layer of ground parenchyma is obliterated in the carpellary wall, and the inner (locular) epidermis separates from the carpellary wall. Secondary wall formation is evident in both the outer and inner layer of parenchyma cells of the inner integument of the developing seed three days after pollination (Figure 30) Twelve days after pollination the seed coat is well formed and possesses characteristic external longitudinal ridges that originate from the outer integument. Internally a doublelayer of macrosclerids originates from the inner integument and is oriented transversely to the axis of the seed (Figure 31). Nineteen days after pollination the seeds appear mature, placentae are no longer connected to the seeds, and loculicidal dehiscence of the carpels occurs at the location of the median carpellary vascular bundle. Abscission of the flower or fruit occurs in the parenchymatous region of the interfloral peduncle proximal to the vascular plexus of the flower, although no broken cells or abscission layers occur in this region seventeen days after pollination or 2-3 days after fruit abscission. The inflorescence of Eichhornia crassipes terminates the primary stem apex. Vegetative or reproductive growth is initiated

PAGE 120

Figure 30 Longisection of a fertilized ovule three days after pollination, X 243. Note secondary wall formation in cells of the inner integument. Figure 31 Transection of a fertilized ovule twelve days after pollination. X 86. Note macrosclerids in inner integument Figure 32 Transection of a stem of E_. crassipes X 11. Figure 33 Transection of an adventitious root of E. crassipes X 91. c cortex; cc central cylinder; e epidermis; em embryo; en endosperm; h hypodermi6; ii inner integument; lr lateral root; oi outer integument; p parenchymatous layer; v vessel (metaxylem) ; vb vascular bundle.

PAGE 121

109 ^TV^em ^^Qj^y; *£ r^-V v jv ** ^i ^Ssfcfl 33

PAGE 122

110 from an axillary bud of the leaf and produces the secondary stem apex Csecondary shoot). The secondary. stem apex produces a -membranous prophyll and either leaves or another terminal inflorescence. If the secondary stem apex initially forms leaves and then terminates in an inflorescence, the axillary bud of the terminal leaf becomes a tertiary stem apex. If the secondary stem apex terminates in an inflorescence, the axillary bud subtended by the prophyll forms a tertiary stem apex which may form leaves or another terminal inflorescence. The mature stem of 15. crassipes is surrounded by overlapping aerenchymatous leaf bases (Figure 32) A single-layered epidermis composed of isodiametric parenchyma cells is external to the cortical region that is composed of three distinct zones. The outer zone of the cortex consists of 4-6 compact layers of large, thin-walled parenchyma cells, whereas the inner cortical zone is composed of 3-5 compact layers of smaller, thin-walled parenchyma cells. The middle cortical zone comprises the bulk of the cortex and consists of small aerenchyma cells separated by intercellular spaces, typically 0.08-0.12 mm in diameter, organized into 6-10 layers. Vascular bundles are dispersed at low density in the middle cortical zone and are oriented vertically or progress laterally. These collateral vascular bundles (xylem interior to phloem) are composed of 4-8 primary xylem tracheids with helical wall thickening and 4-6 sieve tube members with companion cells in the primary phloem. Perivascular fibers with poorly developed secondary walls are included with each vascular bundle (exterior to the primary phloem) within

PAGE 123

Ill a single series of large, thin-walled parenchyma cells. A distinct layer of parenchyma cells separatesthe cortex of the stem from the central cylinder of the stem. This layer is composed of small, rectangular, thin-walled parenchyma cells (1-3 layers) that are elongated periclinally and may be continuous or interrupted by root, leaf, or bud traces or inflorescence development (Figure 32). Vascular bundles are randomly distributed throughout the compact ground tissue of the central cylinder of the stem which is composed of small, thin-walled parenchyma cells. These collateral vascular bundles typically have one vessel (helical to scalariforra wall thickenings) and 1-4 tracheids (helical wall thickening) in the primary xylem which is interior to the primary phloem composed of 2-4 sieve tube members and companion cells. Neither perivascular fibers nor a distinct layer of parenchyma around the vascular bundle are typically associated with the vascular bundles in the central cylinder of the stems. Most of the vascular bundles proceed vertically in the central cylinder of the stem, although some of the vascular bundles proceed laterally into adventitious roots, axillary buds, or leaves. Tanninif erous and raphide idioblasts occur occasionally among the parenchyma cells of the stem. Adventitious roots arise in the distinct parenchymatous layer of the stem and proceed obliquely through the stem cortex and the aerenchymatous leaf bases. These roots have an epidermis composed of isodiametric parenchyma cells and a slightly sclerified hypodermis composed of thick-walled parenchyma cells. Internal to the hypodermis is a cortex composed of three distinct zones (Figure 33).

PAGE 124

112 The outer zone of the cortex is composed of large, thin-walled parenchyma cells (2-3 compact layers). The middle cortical zone is composed of elongate parenchyma cells of the radially oriented diaphragms that separate the intercellular spaces ( 20 u. wide by 150 /t long) that comprise the bulk of the cortex. The inner zone of the cortex is composed of 5-7 layers of cylindrical, somewhat thickwalled parenchyma cells that are separated by small, but uniformly sized intercellular spaces. The innermost layer of the inner cortical zone, the endodermis, is composed of a single series of parenchyma cells with a slightly thickened exterior wall and a thin interior wall where the endodermis borders the pericycle. The pericycle of the root is composed of small, thin-walled parenchyma cells and is not differentiated from the exterior layers of parenchyma cells of the pith. The polyarch stele is composed of 1-3 sieve tube members with companion cells in the primary phloem and one metaxylem vessel with helical to scalariform wall thickening and 1-2 protoxylem tracheids with helical wall thickening in the primary xylem at each pole. Tanninif erous idioblasts are interspersed among the thin-walled parenchyma cells that separate the poles of vascular tissue. Thick-walled angular parenchyma cells comprise the pith region. The parenchymatous layer of the stem that separates the cortex from the central cylinder is also the origin of the four bud traces that proceed to an axillary bud (Figure 34) A pair of bud traces arises on either side of the stem below the insertion of the axillary bud, girdles and diverges from the parenchymatous

PAGE 125

Figure 34 Transection of a stem proximal to the insertion of an axillary bud. X 48. Note axillary bud traces in the cortex of the stem. Figure 35 Transection of a stolon of _E. crassipes X 23. Figure 36 Transection of a stem at the insertion of an inflorescence. X 18. Note dispersed parenchymatous layer and expanded central cylinder of the stem. Figure 37 Transection of a stem distal to the insertion of an inflorescence. X 18. Note restored central cylinder of the stem (secondary shoot) and separate peduncle. a aerenchyma; aerenchymatous cortex; c cortex; cc central cylinder; cl central lacuna; e epidermis; lbt lateral bud trace; p parenchymatous layer; pc parenchymatous cortex; pd peduncle; vb vascular bundle; vbt vertical bud trace.

PAGE 126

114 cc vbt, Ibt 34* HBHHI . . : '* 'u C cc !fc? •.' : ••• .•'*• 36 ^ •' # • ..... **? r • :\'.-. 37 a

PAGE 127

115 layer and proceeds vertically toward the axillary bud. A second pair of bud traces arises on either side of the stem proximal to the insertion of the axillary bud, girdles and diverges from the parenchymatous layer, and proceeds laterally toward the axillary bud. Single bud traces from the vertical pair and the lateral pair of bud traces fuse on the same side of the stem and form two larger bud traces in the cortex of the stem. The two bud traces fuse proximal to the insertion of the axillary bud and proceed as a single trace into the axillary bud. Transections of the mature stolon the elongated first internode of an axillary bud reveal anatomical characteristics that are similar to the mature stem (Figure 35) A single-layered epidermis composed of longitudinally elongated parenchyma cells and a welldeveloped cuticle form the outer boundary of the stolon. Internal to the epidermis are 5-7 compact layers of large, thin-walled parenchyma cells that comprise the outer cortex. The inner cortex is composed of small, thin-walled aerenchyma cells separated by intercellular spaces (typically 0.1 mm in diameter) that are distributed in 3-5 layers. Vascular bundles found in the cortex are surrounded by a single layer of large parenchyma cells and are distributed in two concentric series, one series in each layer of the cortex. These collateral vascular bundles have primary xylem (1-3 protoxylem lacunae and possible 1-2 tracheids with annular thickenings) internal to primary phloem (2—5 sieve tube members with companion cells). Perivascular fibers are associated with each vascular

PAGE 128

116 bundle and occur external to the phloem. The central cylinder of the stolon is differentiated from the cortex because the compact ground tissue of the central cylinder is composed of large, thinwalled parenchyma cells, whereas the cortex is primarily aerenchymatous. Vascular bundles are randomly distributed in the central cylinder of the stolon. These collateral vascular bundles are composed of 1-2 large protoxylem lacuna (tracheids with annular wall thickenings are visible only in longisection) in the primary xylem which is interior to the 2-4 sieve tube members with companion cells of the primary phloem. Transections at the insertion of a mature inflorescence in the primary stem reveal distinct anatomical features (Figure 36). Approximately 5/8 of the parenchymatous layer of the stem is dispersed and no longer distinct. The parenchymatous ground tissue and vascular bundles of the central cylinder of the stem expand through the region vacated by the distinct parenchymatous layer and become continuous with the aerenchyma cells in the former middle cortical zone of the stem. The remainder of the parenchymatous layer remains intact and separates the remainder (3/8) of parenchymatous ground tissue and vascular bundles of the central cylinder of the stem from the inner cortical zone. Distally the open ends of the intact parenchymatous layer advance inward and fuse in the center of the stem to form a crescent-shaped boundary around the remainder (3/8) of parenchymatous ground tissue and vascular bundles of the central cylinder (Figure 37). The distinct anatomical features of a stem are present in this crescent-shaped sector of the transection, whereas the distinct

PAGE 129

117 anatomical features of a peduncle of an inflorescence occur in the remaining large, circular region of the transection. The crescentshaped portion of the transection is the base of the secondary shoot, although the interior flank of the secondary shoot is not separated totally from the peduncle. A single-layered epidermis is the outer boundary of the secondary shoot and is external to a cortex of distinct parenchymatous zones (inner and outer) and a middle aerencymatous zone. Periclinally elongated parenchyma cells in 1-2 layers surround the parenchymatous ground tissue and the scattered collateral bundles of the central cylinder of the secondary shoot. A single-layered epidermis is also the outer boundary of the peduncle and is external to an outer parenchymatous zone and an inner aerenchymatous zone of ground tissue. Collateral vascular bundles are scattered throughout the aerenchymatous ground tissue which surrounds the central lacuna. A typical primary shoot apex with two inflorescences separated by a leaf demonstrates many of the components involved in the sympodial growth pattern of the waterhyacinth (Figure 38). An ensheathing leaf base (L4) originates from the primary shoot and surrounds a mature peduncle and a secondary shoot. The mature peduncle does not occur in the genetic spiral of the primary shoot because the inflorescence has terminated the primary shoot. Vegetative growth of the plant continues, however, through the development of the secondary shoot which is located in the axil of the terminal leaf of the primary shoot. The prophyll of the secondary shoot surrounds the axillary bud of the prophyll and the first leaf (L3) of the secondary shoot. A young

PAGE 130

Figure 38 Transection of a tertiary stem apex. X 11. Note young leaf (1-0 that is inserted on the secondary shoot between the peduncles of the two inflorescences. Figure 39 Transection of a tertiary stem apex. X 11. Note consecutive inflorescences that are not separated by a young leaf. Figure 40 Transection of an aborted lamina and stipuleflap of a prophyll. X 53. ab axillary bud; al aborted lamina; I3, I4, third and fourth leaf; pd peduncle pr prophyll; sf stipule-flap; va vegetative apex; yl young leaf.

PAGE 131

119 39 :

PAGE 132

120 inflorescence terminates the secondary shoot; vegetative growth of the plant continues, however, through the development of a tertiary shoot in the axil of the terminal leaf of the secondary shoot (Figure 38). The developing prophyll of the tertiary shoot surrounds the leaf primordia (P]^, P2) of the tertiary shoot and the axillary bud of the prophyll (Figure 38). Consequently, two sympodial branches form and each sympodial branch develops in conjunction with the development of an inflorescence. When two inflorescences are initiated consecutively the organization of the shoot apex varies (Figure 39) The ensheathing leaf base of the terminal leaf (L3) of the primary shoot surrounds both the older inflorescence that terminates the primary shoot and the secondary shoot that appears in an axillary position. The prophyll of the secondary shoot surrounds both the young inflorescence that has terminated that shoot and the tertiary shoot that subsequently occurs in the axillary position of the secondary prophyll. The tertiary shoot apex is vegetative and produces a prophyll with an axillary bud and two leaf primordia (Figure 39) Although a plant with only two consecutive inflorescences is described above, plants with three and four consecutive inflorescences were observed occasionally and growth of the plant in each case continued through the development of axillary buds of prophylls. Sympodial branching in _E. crassipe s occurs in conjunction with a single floral initiation and reoccurs with each successive floral initiation even if the floral initiations occur consecutively.

PAGE 133

121 The insertion of the prophyll of a secondary shoot occurs 0.5-2. mm distal to the separation of the secondary shoot and the inflorescence. This distance between the terminal leaf of the primary shoot and the insertion of the prophyll is similar to the typical internodal length of the primary shoot. In live or preserved whole material, the base of the secondary shoot appears to be another internode of the primary shoot and the inflorescence seems to be lateral. Characteristic anatomical features are visible in a transection of a prophyll of a secondary shoot (Figure 39) A single-layered epidermis of small, rectangular, thin-walled parenchyma cells and a cuticle are the outer boundary of the prophyll on both the adaxial (outer) and abaxial (inner) surfaces. Internal to the epidermis on the adaxial and abaxial surfaces one and 2-3 compact layers, respectively, of small, isodiametric, thin-walled parenchyma cells form an outer zone of parenchymatous ground tissue. Aerenchymatous cells separated by large intercellular spaces in a single series occur between the outer layers of parenchymatous ground tissue. The intercellular spaces and ground parenchyma are lacking in the fused lateral margins of the sheathing prophyll, and consequently these margins consist of only the abaxial and adaxial epidermal layers. The vascular bundles are distributed in a single series in the ground tissue but occur more frequently in the median regions than in the lateral regions of the sheathing prophyll. These collateral vascular bundles consist of 1-4 primary xylem tracheids (helical wall thickening) which occur adaxial to 1-5 sieve tube members with companion cells in the primary

PAGE 134

122 phloem. Large, thin-walled parenchyma cells (1-2 layers) surround the vascular bundles In the thickened median regions, whereas a specific layer of parenchyma cells does not occur around vascular bundles in the lateral fused margins of the prophyll. The apical stipule-flap represents a membranous extension of the lateral regions of the sheathing prophyll and surrounds either the lamina of the next leaf or the lamina of the outer bract of the next inflorescence (Figure 40). An aborted lamina arises abaxially to the origin of the stipule-flap. This cylindrical aborted lamina consists of an external layer of thin-walled parenchyma cells that form the epidermis and 2-3 layers of ground tissue composed of small, thin-walled parenchyma cells. Tanniniferous idioblasts occur frequently among the parenchyma cells of the ground tissue. The terminal portion of the median vascular bundle of the sheathing prophyll also is found in this aborted lamina.

PAGE 135

DISCUSSION Reproductive and Vegetative Characteristics of Waterhyacinth Populations The production of inflorescences was a prominent feature of the phenology of both introduced and established plants of Eich hornia crassipes at all study sites. Nutrient concentrations, herbivory, and temperature were the major influences upon vegetative characteristics of the waterhyacinth populations and upon the time scale of phenological changes. Nutrient concentrations also affected floral initiation and inflorescence development. Floral Initiation Development of the inflorescence of E_. crassipes required 5-6 weeks from initiation to anthesis during this study. Penfound and Earle (1948) previously reported inflorescence development from initiation to anthesis required only 2-3 weeks. At floral initiation the primary stem apex was surrounded by 3-4 immature leaves within the ochrea-like stipule of the mature fourth or fifth leaf, Each of the leaf primordia of introduced waterhyacinths developed in succession at a rate of approximately one leaf per week at each study site. The inflorescence also developed to maturity at the same rate as the leaves, Mean temperatures, mean low temperatures, weekly totals of precipitation, and mean weekly daylengths varied at the time of 123

PAGE 136

124 floral initiation at all the study sites on Payne's Prairie State Preserve (cf. Tables 2,3), Searle (1965) and Chailakhyan (1968) have reported that specific temperatures were responsible for floral initiation in many plant species, especially plants of the north temperate regions. However, specific temperatures probably do not stimulate floral initiation in the waterhyacinth because mean temperatures and mean low temperatures varied at the time of floral initiation at all study sites. During this study cool mean temperatures (12-16C) from mid-November through March prevented floral initiation, unless exceptionally warm mean temperatures (20-24C) occurred during this period (as during January, 1974). A threshold mean temperature or mean low temperature may be necessary before flowering is initiated, and flowering cannot be initiated below this temperature. Hitchcock ej^ al. (1949) reported flowering in waterhyacinths does not occur at or below mean low temperatures of 16C. Mean weekly daylengths varied from 35-140 minutes between dates of floral initiation at the study sites (Table 3) and inflorescences were produced by waterhyacinths in the greenhouse during all months of the year. The daily photoperiod had no effect upon floral initiation in _E. crassipes during this study. Bock (1966) and Das (1967) also reported that no correlation occurred between flowering in E. crassipes and daylength. Previous authors (Door enbos and Wellensiek, 1959; Chailakhyan, 1968; Evans, 1969; Zeevaart, 1976) reported that daylength was the major stimulus for floral induction in most plant species, although Doorenbos and Wellensiek and Chailakhyan indicated that tropical or day— neutral plant

PAGE 137

125 species may react to other environmental stimuli. In the present study mean weekly totals of solar insolation were similar (2.83.2 Kcal-cnf^-wk-l) during floral initiation at the study sites. However, total solar insolation ranged from 2.3-3.9 Real* cm 2. w k-l during the two weeks before and after floral initiation (Table 3). The similarities of amounts of total solar insolation on dates of floral initiation are probably coincidental. No correlation was noted between floral initiation and the amounts of precipitation recorded at the study sites (Table 3). Chailakhyan (1968) indicated that environmental parameters other than day length and temperature could be responsible for floral induction, although previous reports of floral induction caused by light intensity or soil moisture and humidity were usually based upon observations. Environmental parameters temperature, daylength, light intensity and precipitation were not correlated with initiation of flowers in waterhyacinth populations at the study sites or the greenhouse. The nutrient status of the plants or their habitat was correlated with floral initiation of waterhyacinths. Flowers were first observed at the low nutrient sites (Biven's Marsh, 441 Canal, and Melton's Pond) in early Spring, whereas flowers were not observed until much later (mid-Summer) at the high nutrient sites (New Canal and Main Canal) Flowering occurred in late Spring at the intermediate nutrient site (Nelumbo Area) Waterhyacinths in the greenhouse did not flower until the nutrients in the water were depleted and the plants appeared chlorotic. Only vegetative

PAGE 138

126 development was noted initially in the plants in the fertilized tank, although after five weeks the plants appeared chlorotic and inflorescences were noted. Low nutrient levels were present during floral initiation in both established waterhyacinths at the study sites and cultured waterhyacinths in the greenhouse. Nutrient stress of the waterhyacinth plants could be involved with floral initiation of the plants. The waterhyacinth populations at all the study sites, however, had undergone rapid growth and population development immediately before flowering occurred at each study site. Established waterhyacinths at the low nutrient sites initiated inflorescences in late Winter and early Spring when reactivation and redevelopment of the population began. Introduced waterhyacinths at the low nutrient site were observed to be vigorous during early Spring when inflorescences were initiated at that site. Waterhyacinth plant size increased and plant densities were stable (not affected by herbivore or density stress) at the high nutrient sites during early and mid-Summer when floral initiation occurred at these study sites (cf. Table 4, Figure 4). Plant size and plant densities of cultured waterhyacinths were also at maxima when these plants appeared chlorotic and floral initiation occurred (cf. Figures 6,7). Although the vigorous plants at the 441 Canal and in the greenhouse had chlorotic symptoms shortly before the inflorescences appeared, floral initiation in waterhyacinth plants at all study sites and in the greenhouse occurred when the waterhyacinth populations were both stable and composed of large, vigorous plants. The waterhyacinth populations at all study sites and in the greenhouses were also growing in water with relatively low nitrogen

PAGE 139

127 concentrations (0.1-0.5 rag/1) at the time of floral initiation. The ratio of nitrogen in the water to the stored carbohydrate in the vigorous waterhyacinth plants may have been the stimulus for floral initiation. Previous authors (Miller, 1938; Meyer and Anderson, 1952; Doorenbos and Wellensiek, 1959; Chailakhyan, 1968) have described a similar phenomena in other plant species. Other authors (Murneek, 1948; Searle, 1965; Evans, 1969) indicated, however, that no major element had any special role in flowering and that only indirect effects of mineral nutrition upon flowering were observed in experiments specifically designed to measure nutrition effects. Nitrogen concentrations were low at the low nutrient sites in the Winter and early Spring and floral initiation occurred in the vigorous plants (not affected by herbivore or temperature stress) at these sites after the threshold mean temperature for flowering was attained. Correlations of flowering to nitrogen concentrations and vigor of the waterhyacinth plants at the intermediate site are difficult because of the fluctuations in the nutrient levels caused by the drawdown at that site. Introduced waterhyacinths at the intermediate nutrient site produced many inflorescences during September and October when the waterhyacinth populations were rot affected by drought, herbivore, or temperature stresses and nitrogen levels were low. Nitrogen levels decreased at the high nutrient sites after July because of dilution of the water in the New Canal by rainfall and because the Main Canal no longer received nutrient— rich water from the New Canal. Inflorescences occurred with regular frequency 5-6 weeks

PAGE 140

128 after nitrogen levels decreased at both high nutrient sites. Floral initiation in introduced and established waterhyacinth populations at all sites was correlated with the vigor (carbohydrate status) of the plants and the concentration of nitrogen in the water. The nitrogen-carbohydrate status of the waterhyacinth may be the only periodic environmental cue for floral initiation in these plants. Doorenbos and Wellensiek (1959) and Chailakhyan (1968) indicated that both daylength and temperature are reliable and accurate signals for floral initiation in temperate plant species, although they also mentioned that plant species from tropical climates may utilize other reliable environmental cues to induce sexual reproduction. These authors state that production of the seed is of primary importance to the temperate plant species so the seed may act as a perennating organ to maintain the species through adverse (low temperature) conditions. Although .E. crassipes is native to tropical aquatic ecosystems without temperature stesses, the species still faces periodic and strong population pressures from its riverine habitat. Mohamed (1975) indicated that populations of waterhyacinths decreased drastically during the rainy season and subsequent periods of high water levels along the White Nile in Sudan. Photoperiod and temperature variations do not occur in the native ecosystems of the waterhyacinth. Annual fluctuations in water levels occur inevitably, however, in these riverine systems (Mohamed, 1975). In order to produce seeds to maintain the waterhyacinth populations through adverse conditions of river flooding, the species may utilize the changes in nutrient concentrations in

PAGE 141

129 the water at the beginning of the rainy season as a stimulus for floral initiation. Brinson (1973) and Mohamed (1975) reported great increases in nutrients occurred in a tropical lake-riverine ecosystem and in the White Nile, respectively, early in the rainy season from the runoff from the watershed. The waterhyacinth plants that remained on the floodplain during the dry season and the new waterhyacinth seedlings are able to expand and develop their population rapidly with influx of high nutrient water at the beginning of the rainy season. In the present study expansion and development of the waterhyacinth population occurred after the water level increased at the intermediate nutrient site and when waterhyacinths were introduced into the study sites. Mohamed (1975) and Hestand and Carter (1975) also indicated that waterhyacinth populations increased rapidly in reflooded tropical and subtropical aquatic ecosystems, respectively. Continued rainfall during the rainy season dilutes and decreases the nutrient concentrations in the river water (Brinson, 1973). This change in nutrient levels in the water of the riverine system may stimulate sexual reproduction in the waterhyacinth population. Floral initiation is caused by the favorable nitrogen-carbohydrate status of the vigorous plants in the natural waterhyacinth population. In this study expansion and development of the waterhyacinth population (after temperature, herbivory, or drought stress) always preceded the appearance of flowers. The concentrations of nitrogen at the study sites was also low relative to previous levels, at the time of floral initiation. The utilization of changing nutrient levels as a stimulus for floral initiation allows waterhyacinth plants to maintain their population in very dynamic aquatic riverine systems.

PAGE 142

130 High density of inflorescences in waterhyacinth populations in low nutrient sites and in the greenhouse are correlated with the vigor of the waterhyacinth plants and the nitrogen concentrations in the water at these locations. All waterhyacinth plants in a population do not flower during a peak period of flowering and some plants may not flower during the entire period of anthesis. The number of plants that do produce inflorescences at a given site is determined by the carbohydrate status of the individual plants and low nitrogen levels in the water. No correlation in flowering was noted among the introduced waterhyacinth plants and their raraets at either Melton's Pond or the Nelumbo Area (Table 5). Each individual plant reacts to the nutrient status of the water independently, and flowering is affected by the plants own carbohydrate status. Mean concentrations of all nutrients were somewhat higher at the 441 Canal than at Biven's Marsh (Table 1) and higher densities of inflorescences occurred at the 441 Canal than at Biven's Marsh (Figure 5). More waterhyacinth plants at the 441 Canal had an adequate carbohydrate status at the time of low nitrogen levels, and more plants at the 441 Canal produced inflorescences than at Biven's Marsh. Higher densities of inflorescences also occurred in the fertilized tank in the greenhouse than in the nonf ertilized tank (cf. Figures 6,7). The waterhyacinth plants in the fertilized tank were more vigorous at the time of floral initiation than plants in the nonf ertilized tank and produced twice as many inflorescences as the plants in the nonf ertilized tank. In a preliminary experiment an optimal nitrogencarbohydrate ratio occurred and cultured waterhyacinths produced 585

PAGE 143

131 inflorescences per 120—135 plants during a 60-day period. The latter data are 3-7 times higher than the highest densities of inflorescences observed at the study sites. High densities of inflorescences occurred at the study sites and in the greenhouse when the plants in the waterhyacinth population had a similar carbohydrate status at the time low nitrogen levels occurred in the water. The pattern of inflorescence development on individual waterhyacinth plants may also be related to the vigor (carbohydrate status) of the individual plant and the nitrogen concentration of the water. No pattern of initiation of inflorescences was noted for any individual waterhyacinth plants observed at either the low or the intermediate nutrient sites (Table 5) Although all introduced waterhyacinth plants at both sites were in close proximity and groups of the plants had the same genotype, each individual plant at a given site produced inflorescences in a different sequence. Because the inflorescence of E. crassipes is terminal, a specific morphogenetic process must occur each time sexual reproduction is initiated in this species. A single initiation could be responsible for two or more successive inflorescences, but two or more inflorescences intermittant with leaves could not be produced by a single stimulus. If each inflorescence or set of inflorescences produced by a waterhyacinth plant is initiated by the nitrogen— carbohydrate status of the plant, this ratio for an individual plant must also vary irregularly. The waterhyacinth plants may produce leaves periodically during conditions of floral initiation to replenish resources used in flower production. The drain of resources (nutrients) during

PAGE 144

132 flowering of the waterhyacinths is noted by the chlorotic symptoms of the leaves 1-2 weeks before inflorescences appear. Under an optimal nitrogen regime (i.e., preliminary greenhouse experiment) waterhyacinth plants may produce many inflorescences successively and produce few leaves. Under unfavorable nitrogen regimes (Le. high nutrient sites until July) few or no inflorescences will be produced. Cessation of flowering at the study sites was correlated with either increases in nitrogen content of the water or the return of cool temperatures in the Fall. The concentration of nitrate nitrogen increased during late April at the low nutrient sites and the density of inflorescences decreased greatly 5-6 weeks later (the period of development of an inflorescence) at the three low nutrient sites (cf. Figure 5). The concentrations of ammonia and total organic nitrogen increased during late August (nitrate concentrations remained low) at a high nutrient site (New Canal) and flowers were not observed at that site 5-6 weeks later. Sparse flowering continued at a low, intermediate, and high nutrient site until 5-6 weeks after mean low temperatures of 13— 15C (early November) occurred at the study sites (cf. Table 2). If the mean low temperature of 16C is the threshold temperature for floral initiation in waterhyacinths as reported by Hitchcock, et al(1949), the response to this low temperature by the waterhyacinth appears to be absolute. No flowering was noted in late November or early December 5-6 weeks after mean low temperatures of 17-18C occurred at the study sites.

PAGE 145

133 Nutrient concentration of the water at the study sites was the primary factor that affected flowering, growth, and development of the waterhyacinth populations. In order to compare the relative differences in the flowering and growth of waterhyacinths at study sites with differing nutrient levels, the study sites were categorized as high (New Canal and Main Canal) intermediate (Nelumbo Area) and low nutrient sites (441 Canal, Biven's Marsh, and Melton's Pond) (Table 1). A similar categorization of the same or analogous study sites on Payne's Prairie State Preserve was reported by Morris (1974). Morris (1974) reported a higher mean nitrate nitrogen value (2.2 mg/1) and a lower mean phosphorous value (0.9 mg/1) for water samples taken at his high nutrient study site than were measured during this study. Means of nitrate, ammonia, and total organic nitrogen and total phosphorous concentrations measured at New Canal were similar or somewhat lower than the mean concentrations (NOv^N-l, 7 mg/1; NHt-n-5.4 mg/1; org-N 2.3 mg/1; total P-5.9 mg/1) from effluent from four sewage treatment plants in central Florida (Wodzinski, 1975). Nitrate values for the high nutrient site reported in this study may have been lower than previously reported values by Morris (1974) and Wodzinski (1975) because the effluent had been conducted through a mile-long canal that was filled with waterhyacinths before it reached the study site. Concentrations of N0n~N and org-N and total P measured at a low nutrient site (441 Canal) were similar to the low values measured at the same site by Knipling et jal. (1970). Only the lowest detectable amounts of NO3-N and total P were measured at Melton's Pond during this study and these values were similar to the concentrations reported by

PAGE 146

134 Morris (1974) for the same site and similar to concentrations of the same nutrients obtained by Stevens (1976) for a similar small pond in north Florida. Vegetative characteristics of the waterhyacinth populations varied with changes in the amount of herbivory by A. dens a and changes in temperature as well as nutrient concentrations. The herbivore, A. densa, that disrupted waterhyacinth populations at all study sites, has been described and utilized for biological control of waterhyacinths (Vogel and Oliver, 1969; Center, 1976). Heavy insect infestations were noted in the waterhyacinths only during April and September to November by earlier workers (Vogel and Oliver, 1969; Center, 1976) whereas in the present study insect damage to waterhyacinth populations was noted throughout the year with heavy infestations in January, May-June, July-August, and August-September. The amount of damage done to the waterhyacinths by the insect and the nutrient availability at the site determined the rate of recovery for the waterhyacinth populations. When major damage occurred at the high nutrient sites, the recovery of the population was rapid (4-6 weeks) because excessive nutrients available in the water were absorbed quickly and assimilated to replace the amount of standing crop lost to the herbivore. Replacement of the standing crop and its nutrient storage was much slower (6-10 weeks) at the low nutrient sites. On a few occasions A. densa damage was responsible for the death and submersion of small areas (25-5Q m2) of the waterhyacinth population. However, these areas were rapidly filled by young raraets from vegetative reproduction. Although the insect can damage and

PAGE 147

135 stress the waterhyacinth populations, the insect alone cannot remove a waterhyacinth population because it has many parasites and predators (Center, 1976). Damage by A. densa in conjunction with other population pressures (competition, low temperatures, etc) may result in control of a waterhyacinth population as was observed by Center (1976). Cool winter temperatures and frost caused a decrease in plant size in the established waterhyacinth populations at all study sites (Table 4). Reductions in plant size of waterhyacinth during cool weather were also observed by LaGarde (1930) Penfound and Earle (1948) and Center (1976). The cool temperatures and low light intensities that occur during the winter probably decrease the metabolism of the plants. However, frost damage is the major cause of reduced plant size in winter. The canopy of the waterhyacinth population is interrupted and opened by frost damage to the leaves. Elongation of the subsequent leaves is reduced because they are not shaded by the canopy and plant size is reduced until the canopy becomes continuous again. An increase in stolon frequency was observed in the established waterhyacinth populations during the Spring and was a major feature (with increased plant densities) of the reactivation and growth of the waterhyacinth population. Center (1976) reported the highest densities of a waterhyacinth population occurred in April, although the number of stolons was not distinguished from the number of canopy plants in his study. Warmer temperatures and higher light intensities in the Spring seem to stimulate the production of stolons which increase the plant density and is responsible for the development of a vigorous waterhyacinth population. Increases in the number of

PAGE 148

136 stolons per m^ were noted at all field sites during and after heavy infestations of the herbivore, A. densa The removal of the stem apex by the insect and the subsequent loss of apical dominance is primarily responsible for the increased production of stolons per plant although increased stolon production could also be a response of the population to herbivore damage. Vickery (1972) has shown increased primary productivity of temperate grasslands when optimal grazing occurs. Constant plant densities in established waterhyacinths were maintained for short periods of time hetween A-.densa infestations at all study sites. The stable plant densities were related to the nutrient concentrations of each study site because the plant size was larger at the high nutrient sites than at the lower nutrient sites and fewer large plants than smaller plants could fit into a m^ of surface area. Populations at the high nutrient sites stabilized at densities of 53-89 apices/m^; Center (1976) also reported plant densities of 60-90 apices/m2 at a high nutrient site. Stable densities at the intermediate nutrient site (89-110 apices/m^) were higher than the high nutrient sites and somewhat lower than the lower nutrient sites (97-126 apices/m ) Boyd and Scarsbrook (1975) measured plant densities of 110-115 apices/m^ in a fertilized pond from a waterhyacinth population composed of plants similar in size to plants at the intermediate nutrient sites in this study. Although plant densities varied greatly throughout the year at the study sites, low nutrient sites typically had higher plant densities than high nutrient sites in this study.

PAGE 149

137 Values of total dry mass Cstanding crop) of established waterhyacinth. populations of both large and small plants at all study sites reached maxima (1.2-1.7 kg/m 2 ) in the Spring after maximum plant density was reached. Increased standing crop after increased plant density was also reported for Pistia stratiotes by Hall and Okali (1974) and for waterhyacinths by Center (1976). Maximum amounts of dry mass (1.2-1.7 kg/n)2) in this study were lower than previously measured values of 2.7 kg/m2 (Penfound and Earle, 1948), 2.5 kg/m2 (Knipling _et al. 1970), 2.0 kg/m 2 (Boyd and Scarsbrook, 1975), 2.5 kg/m2 (Morris, 1974), and 2.4 kg/ra2 (Center, 1976). Of the previous authors only Center (1976) mentioned herbivory to the observed waterhyacinth populations, although in his study damage from A. dens a was not as extensive as occurred in this study. The presence of A. dens a populations during this study may account for each population's inability to reach the maximum standing crop that has been reported by other authors. Maximum root length of established waterhyacinths varied with both the Arzama infestations and the nutrient concentrations of the water at each study site. Waterhyacinths at the low nutrient sites had longer and more extensive root systems than those waterhyacinths at the high nutrient sites as previously reported by Morris (1974) and Boyd and Scarsbrook (1975). The length and extent of the root system of waterhyacinth plants seemed to be a good indicator of high or low nutrient concentrations in the water during this study. Established waterhyacinths in stahle populations at all study sites maintained a constant number of functional larainas per plant

PAGE 150

138 (5-7 leaves with intact laminas) Center (1976) also observed a constant leaf density of 5-7 functional leaves per plant in a stable waterhyacinth. population. The amount of space per individual plant is limited in the stable waterhyacinth populations because of intense intraspecif ic competition; thus waterhyacinth plants do not possess more than 5-7 functional leaves per plant. The rate of leaf production was the most variable parameter of the introduced waterhyacinths throughout the year and fluctuated with every change in the waterhyacinth's environment. Decreases in the rate of leaf production were noted at the intermediate and low nutrient sites during February and March because of the cool winter temperatures and "transplanting shock." Infestations and damage by the herbivore, Arzama densa reduced leaf production rates at the high nutrient site, New Canal, and at the low nutrient site, Melton's Pond. Leaf production ceased in damaged plants from the time the insect destroyed the stem apex until new stolons were formed. Leaves and inflorescences are not formed simultaneously by waterhyacinths and reduced leaf production rates that occurred at the low nutrient site during the early Summer were correlated with the period when profuse inflorescence production occurred. Increased leaf production rates in the introduced waterhyacinths were common at the high, intermediate, and low nutrient sites during the higher temperatures and high light intensities of the Spring. At all sites leaf production rates typically increased as the waterhyacinth populations stablized after one of the environmental stresses had dissipated. Leaf size of established and introduced waterhyacinths varied with nutrient concentrations. Petiole length and width and lamina

PAGE 151

139 length and width were significantly larger in plants at high nutrient sites than in plants at low nutrient sites, whereas these dimensions were intermediate in plants at the intermediate nutrient site (Table 4; Figures 3 and 4) Mean leaf lengths at the high nutrient sites (75-93 cm) were similar to mean leaf lengths of 70-90 cm from waterhyacinths at high nutrient sites reported by Morris (1974) Boyd and Scarsbrook (1975), and Center (1976). Mean leaf lengths measured at the intermediate nutrient site (65-73 cm) were similar to a mean leaf length (70 cm) reported by Morris (1974) at the same site. A mean length (18 cm) of waterhyacinths at a low nutrient site (Melton's Pond) reported by Morris (1974) was much smaller than mean leaf lengths (35-42 cm) recorded in this study at Melton's Pond because Morris' study terminated before the introduced waterhyacinth population had stabilized. Maximum size of morphological parameters of waterhyacinth leaves were determined primarily by the nutrient concentration of the site (Table 4; Figures 3 and 4), although waterhyacinth leaves do have the ability to vary leaf morphology and size in response to intraor interspecific competitive stress. When waterhyacinths invade an open site, density and competition (intra-, interspecific) are low, and a certain leaf form (invasion form) occurs on all waterhyacinths in the population. A separate, distinct and different leaf form (competition form) is found on all plants during high population density and intense competition among waterhyacinths or between waterhyacinths and other species. The invasion leaf form is characterized by petiole lengths which are 40-60% of total leaf length,

PAGE 152

140 petiole widths that are larger than the width of the leaf base, and laminas whose ratio of mean length to mean width is less than one CFigure 41) The competition leaf form is characterized by petiole lengths which are 70-81% of the total leaf length, petiole widths that are similar to the width of the leaf base, and laminas whose ratio of mean length to mean width is greater than one (Figure 41) Shading of the stem apex of a waterhyacinth plant during competition (intraor interspecific) may result in the formation of the competition leaf form. For example, a mixed community of waterhyacinths and Nelumbo lutea was observed where the canopy of the waterhyacinths was irregular because some waterhyacinth leaves were taller than the leaves of adjacent plants. The larger leaves were produced to exceed the height of the Nelumbo leaves that had overtopped them. LaGarde (1930) and Penfound and Earle (1948) also observed elongated nonswollen petioles on waterhyacinth plants at high plant densities or those that occur in the shade. Penfound and Earle (1948) determined that light intensities less than 500 ft.c. stimulated petiole elongation, and light intensities greater than 500 ft.c. formed swollen petioles. In this study maximum leaf sizes of waterhyacinth plants occurred in competing waterhyacinth populations regardless of the nutrient concentration in the water at the study site. The change from the invasion leaf form of waterhyacinths to the competition leaf form occurred whenever plant density increased sufficiently and caused a continuous canopy that shaded individual plants. Open water was available in the pens of introduced waterhyacinths at the low nutrient site (Melton's Pond) until 12 June,

PAGE 153

Figure 41 Morphological variation in leaves of waterhyacinth. A. Invasion form. Note swollen petiole and reniform lamina. B. Competition form. Note elongated petiole and ovate lamina,

PAGE 154

142 A

PAGE 155

143 and the petiole width of the plants remained between 20 and 28 ram (Figure 4) Petiole width at this site decreased through July and averaged only 10 mm after the first week in August because of the concoramitant petiole elongation that occurred when maximum plant density was reached. The ratio of mean lamina length to width of waterhyacinth leaves at Melton's Pond averaged 0.6 until early May, increased to 1.0 in late June, and averaged 1.2 the remainder of the year (Table 13). The petiole comprised only 60% of the total leaf lengthuntil the second week in J*une but increased to 75-80% the first week in August and remained constant the remainder of the year (Table 14) The change in the form of the lamina from reniforra to ovate and the increased length of the petiole demonstrate that whole leaves, lamina and petiole, from waterhyacinth plants at Melton's Pond elongated greatly under the stress of intra— specific competition (Figure 41). Similarly, decreases in the petiole width, increases in the lamina length to width ratios, and increases in the petiole length as a percentage of total leaf length were observed in introduced waterhyacinths at intermediate (Nelumbo Area) and high nutrient sites (New Canal) (Figure 4; Tables 13,14). These changes in the morphology of the leaves were observed two weeks after the pens were filled with plants at each site (Nelumbo Area 12 June, New Canal 1 May). Increased intraspecif ic competition at the intermediate and high nutrient sites also caused elongation of the lamina and petiole of the waterhyacinth leaves.

PAGE 156

144 Table 13 Means of the lamina length to lamina width ratios of the third mature leaf from introduced waterhyacinths into high, intermediate, and low nutrient study sites on Payne's Prairie State Preserve.

PAGE 157

145 Table 14 Length of the petiole (as a percentage of the total leaf length) from the third mature leaf of introduced waterhyacinths at high, intermediate, and low nutrient study sites on Payne's Prairie State Preserve.

PAGE 159

147 inflorescence development reported by these authors was longer than observed in this study because Penfound and Earle and Das observed transplanted waterhyacinth plants under inadequate, shortterm, artificial conditions. The size of the outer Bract lamina of the inflorescence typically remained constant after its appearance 5-6 days before anthesis, although the total length of the inflorescence continued to elongate at regular daily increments (13-60 mm per day) until 18 hours before anthesis. Penfound and Earle (1948) observed that the total length of the inflorescence increased in daily increments of 15 mm per day. Elongation of the interfloral and subfloral peduncle began 18-24 hours before anthesis as the inner bract sheath and enclosed flowers emerged through a slit in the outer bract sheath. The interfloral and subfloral peduncle continued to elongate prior to anthesis, broke through the membranous region of the inner bract sheath 12-14 hours before anthesis, and attained their maximum length by midnight (8 hours before anthesis). Earle (1947) and Penfound and Earle (1948) also noted opening of the inner bract and peduncle elongation approximately 15 hours before anthesis. In this study individual flowers of the inflorescence attained their normal 45-60 orientation to the axis by sunrise, and anthesis occurred 1-3 hours later. Earle (1947) and Penfound and Earle (1948) reported that movement of the flowers to their normal orientation of 30-45 began shortly after midnight and proceeded until

PAGE 160

148 anthesis which_ was completed 1—3 hours after dawn. Individual flowers of the inflorescence began to wilt and close between 5 and 6 pra on the day of anthesis. Simultaneously cell elongation occurred on one side of the peduncle 1.5-3.0 cm distal to the bases of the bracts and slowly formed the capital bend which carried the subfloral and interfloral axis to a horizontal position (70-120) from the vertical by the following morning (Figure 8). Initiation of postanthetic curvature was also observed at 5-7 pm by Rao (1920b), Earle (1947), and Penfound and Earle (1948). In this study the capital bend developed acropetally. The distance from the base of the bend to the insertion of the inflorescence remained constant, whereas the distance from the top of the bend to the bases of the bracts decreased during the 36-48 hours of the bend's development. Rao (1920b) found that the capital bend usually orients the interfloral and subfloral peduncle directly away from the stem apex, although in this study the orientation of the interfloral and subfloral peduncle varied greatly within a study site and among study sites (cf. Figure 11). Penfound and Earle (1948) reported that the capital bend did not occur until all flowers of the inflorescence had opened and might require two days. The capital bend in all inflorescences in this study typically occurred during the night following initial anthesis, regardless of the number of flowers that had opened. The interfloral bend was usually complete by dawn of the day following anthesis, as

PAGE 161

149 reported by Earle (1947) and Penfound and Earle (1948).. Initiation of the basal bend was not apparent simultaneously with the initiation of the capital and interfloral bends in this study, and did not occur until 18-36 hours past anthesis. Earle (1947) and Penfound and Earle (1948) reported, however, that the capital, interfloral, and basal bends are initiated simultaneously 10-12 hours past anthesis. These authors noted that the basal bend was complete 48 hours past anthesis; however in this study, development of the basal bend continued for 5-7 days after anthesis. The orientation of the interfloral and basal bends was identical to the orientation .of the capital bend, and the stimulus that oriented the capital bend also directed the bending of the interfloral axis and the basal portion. The general sequence of events of preanthetic and postanthetic development of the inflorescence observed during this study required more time during cooler temperatures. Earle (1947) and Penfound and Earle (1948) indicated that postanthetic curvature required 1-4 extra days during cooler temperatures. The numbers of flowers and the dimensions of the inflorescence of introduced and established waterhyacinths were correlated with the nutrient status of the water at the study sites. Waterhyacinths at the low nutrient sites (Melton's Pond and Biven's Marsh) produced inflorescences that were generally smaller than those at the intermediate nutrient site (Nelumbo Area) (Table 6). The inflorescences at the former sites were

PAGE 162

150 always smaller than the inflorescences at the high nutrient site (New Canal). The amount of growth of the inflorescence per day before anthesis (as a percentage of the total length at anthesis) was similar at all study sites (Table 7) Maximum sizes of inflorescences and appendages were determined by the nutrient status of the water at the study site, whereas the rate of growth of inflorescences was independent of the nutrient concentrations of the water. Length and width of the lower sepal and the banner petal of flowers of E. crassipes varied within an inflorescence and among sites, and no correlations with nutrient concentration of the water at the study sites were noted (cf. Tables 9,10,11). The smallest mean sepal lengths and mean banner petal lengths occurred at a low nutrient site (Melton's Pond), whereas the largest mean sepal lengths occurred at the other low nutrient site (Biven's Marsh). The largest mean banner petal lengths occurred at an intermediate nutrient site (Nelumbo Area). Dimensions of the lower sepal and banner petal from waterhyacinth flowers at a high nutrient site (New Canal) were intermediate or similar to those dimensions at the intermediate nutrient site. Flower number and the dimensions of the inflorescences were greater at one low nutrient site (Biven's Marsh) than at the other low nutrient site (Melton's Pond) during the Spring and these differences coincide with the observed differences in plant vigor at these two sites. Biven's Marsh had been the site of effluent input for Payne's Prairie until December, 1974, and the increased

PAGE 163

151 vigor of these plants may have been caused by nutrient storage in the marsh sedimenta or the plant biomass. Lower and middle flowers of the waterhyacinth inflorescence at all sites typically possessed larger lower sepals and banner petals than the upper flowers on the same inflorescence. In this study sepals were 27-37 mm long and 10-17 mm wide, whereas Castellano (1958) reported smaller sizes (30 mm long, 10 mm wide) for this floral appendage. The banner petal of waterhyacinth flowers in this study were 30-43 mm long and 19-29 mm wide, and other workers (Castellano, 1958; Agostini, 1974) reported similar lengths (25-45 mm) for the banner petals. In this study the distance between the stigma and the long stamens of midstyled flowers varied from 5-7 mm, and these data are similar to the range of 4.5-6.1 mm reported by Tag el Seed and Obeid (1975). Pollination Seed Set and Seed Germination During preliminary studies artificial pollination was qualitatively successful among selfand cross-pollinated waterhyacinth flowers grown in the greenhouse from February to August. These pollinations were successful throughout the day (from 9 am to 3 pm) although previous workers (Bruhl and Gupta, 1920; Agharkar and Banerji, 1930; Tag el Seed and Obeid, 1975) reported successful selfand cross-pollinations occurred only in the morning hours. These authors indicated that high temperatures and low humidities prevented pollination, however, Bock (1966) reported that there was little correlation between humidity and successful pollination of flowers of E,. crassipes Because humidity was always high in

PAGE 164

152 Florida and especially in the greenhouse, no conclusions concerning the effect of humidity upon successful pollination of flowers of E. crassipes can be drawn from this study. Hymenopteran insect pollinators were observed among the waterhyacinth flowers during this study at both a low and an intermediate nutrient site. Penfound and Earle (1948) and Bock (1966) also noted Hymenopteran pollinators (honeybees, bumblebees) as well as syrphid flies and yellow sulphur butterflies on waterhyacinth flowers, Both Penfound and Earle (1948) and Bock (1966) observed that each insect species landed and moved into the flower in specific ways. During this study each insect species was observed to approach, land and move into the flower in a different manner. Pollen was transferred from anthers to stigma by two insects observed in this study, although Penfound and Earle (1948) noted only insignificant amounts of pollen on stigmas after insects had visited the flowers. Penfound and Earle (1948) suggested that most natural pollination of waterhyacinth flowers was caused by the senescing perianth. Although this natural selfing was observed during this study and by Bock (1966), only low numbers of seeds occurred in nonpollinated (control) flowers utilized in this study and reported by Bock (1966) and Tag el Seed (1972). The large amounts of seed formation observed by Robertson and Ba Thein (1932), Penfound and Earle (1948), and Tag el Seed (1972) probably resulted from both natural selfing and insect pollination.

PAGE 165

153 The amount of seeds produced per capsule and per inflorescence varied irregularly among the study sites. Both natural pollination and artificial pollination were successful at these sites. The percentage of flowers that possessed fruit with seed varied from 621% at the low and intermediate nutrient sites after natural pollination of flowers, whereas Tag el Seed and Obeid (1975) indicated that 12% of the flowers observed in Sudan had been naturally pollinated. In the present study a range of 4-110 seeds per capsule were observed from the low and intermediate nutrient sites after natural pollination, whereas previous workers (Robertson and Ba Thein, 1932; Zeiger, 1962; Tag el Seed, 1972) observed a range of 3-542 seeds per capsule after natural pollination. Given the previous data, approximately 3.1 X 10 seeds would be produced per hectare in this study (441 Canal flower densities). Champness and Morris (1948) indicated that a range of 0.01-1.43 X 10 seeds/ha was produced by terrestrial plant species. Similar estimates of seed production by waterhyacinths .22 X 10 and 1.1 X 10, were reported by Penfound and Earle (1948) and Zeiger (1962), respectively. Germination requirements of waterhyacinth seeds are complex and include specific water levels, soil types, and light and temperature regimes (Hitchcock e_t aT. 1949; Tag el Seed, 1972). In this study seed germination was observed near the intermediate nutrient site on peaty sediments during conditions of high light intensities and high temperatures in early Summer. No appreciable amounts of standing water were present. Other workers (Robertson and Ba Thein, 1930;

PAGE 166

154 Haigh, 1936; Penfound and Earle, 1948) observed natural seed germination on reflooded sediments during high temperatures and high light conditions. Germination of waterhyacinth seed was also successful at high light intensities, high temperatures, and in shallow (1-3 cm) water in the laboratory (Hitchcock et al. 1949; Barton and Hotchkiss, 1951; Tag el Seed, 1972). Artificial or natural drawdowns of aquatic systems will not control waterhyacinth populations if the drawdown occurs in warm weather because the return of water to the peaty substrates will stimulate seed germination. Widespread seed germination was observed after reflooding of the sediments at the intermediate nutrient site during this study. Richardson (1975) also indicated that reflooding of an aquatic system after a drawdown stimulates waterhyacinth seed germination. Drawdowns may be successful as a control measure for waterhyacinth populations if completed during cool weather, because Hitchcock et al., (.1949) and Barton and Hotchkiss (1951) have reported that germination of waterhyacinth seed requires high temperatures (25-35C) Rapid development of waterhyacinth seedlings was observed at the study sites and in the greenhouse during this study. Extensive stolon production occurred 3-5 weeks after germination at the intermediate nutrient site. The waterhyacinth seedlings produced flowers 8-10 weeks after seed germination at this site. Seedlings introduced into the greenhouse also flowered 8-10 weeks after seed germination. The population of waterhyacinth seedlings and their ramets continued to flower throughout the summer and ceased flowering in early November. In this study waterhyacinth plants completed a life cycle (seed to seed)

PAGE 167

155 in 3-4 months (June-September). Das (1967) also reported the reproductive life cycle of _E. c rass ipes was completed in 3-4 months in India, although. Penfound and Earle (1948) reported the life cycle of the waterhyacinth required more than one growing season. Anatomical Aspects of Eichhornia crassipes Aquatic vascular plants typically demonstrate specialized anatomical characteristics. Arber (1920) and Sculthorpe (1967) reported that increases in aerenchymatous ground tissues throughout all plant organs and a general decrease in vascular and lignified tissues are anatomical features of aquatic plants. Aquatic plants also exhibit other anatomical adaptions (specialized diaphragms, epidermal chlorenchyma, absence of cuticle and stomates) which are related to their specific aquatic habitat. Eichhornia crassipes has both anatomical features of land plants and anatomical characteristics of aquatic plants. The presence of distinct cortical and stelar regions is characteristic of the adventitious roots of the waterhyacinth (Figure 33). A distinct singlelayered epidermis of isodiametric, thinwalled, parenchyma cells forms the outer boundary of the root. Although Olive (1894) and Bruhl and Dutta (1923) reported no hypodermal layer, a singlelayered hypodermis of thick-walled parenchyma cells does occur in the root of the waterhyacinth. A similarly structured epidermis and hypodermis also occur in the emergent aquatic plant, Abolboda (Carlquist, 1960). The outer zone of the cortex of the waterhyacinth root typically has 2-3 compact layers of thin-walled, isodiametric, parenchyma cells, whereas the inner

PAGE 168

156 zone is loosely-arranged and composed of 5-7 layers of slightly thick-walled, cylindrical parenchyma cells. The middle cortical zone is composed of radially-arranged aerenchyma cells separated by intercellular spaces of varying sizes. Previous authors (Olive, 1894; Bruhl and Dutta, 1923; Weber, 1950) also noted the middle cortical zone of the root was composed of radially oriented cells and inner and outer cortical zones with 5-7 and 2-4 cell layers, respectively. Although these workers reported that the parenchyma cells of the inner cortex were thin-walled, the present study indicates that these cells are slightly thick-walled. Roots of Abolboda also have parenchymatous inner and outer cortical zones as well as a middle cortical zone of radially elongated aerenchyma cells (Carlquist, 1960). The radial orientation of aerenchyma and intercellular spaces that occurs in the roots of waterhyacinth also occurs in stems (e.g., Jussiaea Myriophyllum ) and roots (e.g., Vallisneria Sagittaria Myriophyllum ) of many monocotyledons and dicotyledons (Arber, 1920; Sculthorpe, 1967; Stant, 1964; Sutton and Bingham, 1973). The stele of the waterhyacinth root is delimited by a pericycle of thin-walled parenchyma cells. A pith with slightly thick-walled parenchyma cells occurs in the center of the root. Parenchyma cells of the pith or pericycle of the waterhyacinth root were not described by Olive (1894) or Bruhl and Dutta (1923) although Carlquist (1960) noted a distinct parenchymatous pericycle and sclerenchymatous pith in the roots of Abolboda Stant (1964) reported, however, that the pericycle was often composed of thick-walled parenchyma cells in the

PAGE 169

157 roots of species of the Alismataceae. The stele of the watcrhyacinth root is polyarchand possesses endarch primary xylem and primary phloem at alternating poles. Olive (1894), Bruhl and Dutta (1923), and Couch (1971) also observed that the stele was polyarch, and the former authors noted xylem maturation was endarch. The primary phloem is composed of 1-3 sieve tube members with companion cells and the primary xylem has one metaxylem vessel and 1-2 protoxylera tracheids. Cheadle (1970) considers the metaxylem vessels of the root of _E. crassipes the most advanced vessels in the Pontederiaceae. Although each xylem pole of the root of E_. crassipes has only a single vessel, Stant (1964) reported that 1-3 vessels occur at each xylem pole of roots of Sagittaria spp. and Echinodorus spp. Stems and stolons (first internode of the axillary bud) of the waterhyacinth have similar anatomical features (Figures 32,35). A single-layered epidermis of thin-walled parenchyma cells forms the outer boundary of both the stem and the stolon, although neither Olive (1894) nor Bruhl and Dutta (1923) noted an epidermal layer in either the stolon or the stem. Stems of species of the Alismataceae, however, have no epidermis and a multilayered periderm is the only outer boundary (Stant, 1964). The outer cortical zones of the stem and stolon consist of 3-6 compact layers of large, thin-walled parenchyma cells, although neither Olive (1894) nor Bruhl and Dutta (1923) mentioned this layer. The middle cortical zone of the stolon and stem are anatomically similar and composed of aerenchyma tissue separated by intercellular spaces (0.08-0.12 mm in diameter). Olive (1894)

PAGE 170

.158 noted a distinct aerenchymatous middle cortical zone in the stolon, although Bruhl and Dutta (1923) described the cortex of the stem as a pseudo-cortex comprised of fused leaf bases. Small waterhyacinth plants observed in this study typically had very short internodes and the leaf bases were nearly continuous and appeared to form a pseudo-cortex as Bruhl and Dutta described. The epidermis and cortical layers of the stem were distinct, however, in the longer interrtodes of larger waterhyacinth plants also observed in this study. The cortex of stems and stolons of waterhyacinths have a distinct aerenchymatous zone, although the cortex of stems of members of the Alismataceae were typically parenchymatous (Stant, 1964). In the present study the cortical vascular bundles of the stem contain 4-8 primary xylem tracheids, whereas the cortical bundles of the stolon may have 1-2 primary xylem tracheids and 1-3 conspicuous protoxylem lacunae. The number of sieve tube members and companion cells of the primary phloem of cortical vascular bundles is somewhat higher in the stems (4-6) than in the stolons (2-5) Perivascular fibers occur external to the primary phloem within a single layer of large, thinwalled parenchyma cells that surrounds the cortical vascular bundles in both organs. Stant (1964) reported that fibers and large, thinwalled parenchyma cells were always associated with vascular bundles in stems of species of the Alismataceae. The central cylinder of parenchymatous ground tissue and vascular bundles is distinct in both the stem and stolon (cf. Figures 32,35). Small, rectangular, thin-walled parenchyma cells in 1-3

PAGE 171

159 layers clearly delimit the central cylinder of the stem, but do not occur in the stolon. The compact ground tissue of the central cylinder of the stem is composed of small, thin-walled parenchyma cells, whereas the compact ground tissue of the central cylinder of the stolon is composed of large, thin-walled parenchyma cells. Olive (1894) and Bruhl and Dutta (1923) noted the compact parenchymatous nature of the central cylinder of both stolons and stems, respectively, although the latter authors did not identify a distinct parenchymatous layer in the stem. Collateral vascular bundles are randomly distributed throughout the central cylinder of the stem and stolon, and Olive (1894) and Bruhl and Dutta (1923) also noted collateral vascular bundles in these organs. The vascular bundles of the central cylinder of the stem are typically composed of one metaxylem vessel and 1-4 protoxylem tracheids, whereas vascular bundles of the central cylinder of the stolon have 1-3 conspicuous protoxylem lacunae and occasionally have 1-2 protoxylem tracheids. The primary phloem is composed of 2-4 sieve tube members and companion cells in the vascular bundles of the central cylinder of both stems and stolons. The vascular systems of the central cylinder and the cortex of the stems and the stolons of the waterhyacinth are distinct and separate. Tomlinson (1970) and Zimmerman and Tomlinson (1972) reported the occurrence of distinct and separate vascular systems in both the central cylinder and the cortical regions of plants from five unrelated monocot families. By analyzing the course and development

PAGE 172

160 of vascular bundles in monocotyledons, these authors have shown the existence of both an inner (central cylinder) and an outer (cortical) axial vascular system. The vascular bundles of the central cylinder of the stems and stolons of E_. crassipes are complex three-dimensionally. In the central cylinder of stems the vascular bundles proceed vertically and obliquely in typical transections. Zimmerman and Tomlinson (1972) indicated three-dimensional complexity is a major characteristic of the course of the vascular bundles in the inner axial system. Although the course of cortical bundles in stems and stolons of E. crassipes was not analyzed in this study, the low number of vascular bundles in the cortex compared to the central cylinder indicates that the cortical axial vascular system is much less developed than the axial vascular system of the central cylinder. Zimmerman and Tomlinson (1972) reported that outer axial vascular systems were typically less developed than inner axial vascular systems in the monocot species they studied. Cortical vascular bundles of E. crassipes differ anatomically from the vascular bundles of the central cylinder, and Zimmerman and Tomlinson (1972) also described anatomical differences between vascular bundles in the inner and outer vascular systems of monocot stems. The vascular bundles of the cortex and the central cylinder of stems and stolons of 12. crassipes appear distinct. Consequently, vascular development in _E. crassipes is probably organized into inner and outer vascular systems in accordance with suggestions by Tomlinson (1970) and Zimmerman and Tomlinson (1972) that all monocotyledons have inner and outer vascular systems.

PAGE 173

161 The primary vegetative shoot of E. crassipes terminates in an inflorescence. Vegetative growth of the primary shoot continues from the axillary bud (secondary shoot) of the terminal leaf of the primary shoot shortly after floral initiation. The growth pattern of the waterhyacinth is sympodial.. The sympodial growth pattern of the waterhyacinth was also described by Warming (1871) who reported that an inflorescence terminated the primary axis of the plant (Agharkar and Banerji, 1930). Warming indicated that the axillary bud of the terminal leaf underwent development and produced vegetative growth (Agharkar and Banerji, 1930). Sympodial growth patterns with terminal or hapaxanthic inflorescences are common in the monocotyledons and predominate in the Bromeliaceae, Haemodoraceae, Heliconiaceae, Zingiberaceae, and Alismataceae (Tomlinson, 1970; Charlton, 1973). Preliminary studies also indicate that sympodial growth also predominates in the Pontederiaceae. Although the secondary shoot arises from the ultimate axillary bud of the primary shoot, the vascular supply to the secondary shoot, the anatomy of the cortex, and the prophyll of a secondary shoot are dissimilar from comparable structures in an axillary shoot. The vascular supply of an axillary shoot originates in the parenchymatous layer of the stem from two separate pairs of bud traces. Manis (unpublished data, Botany Department, University of Florida) also noted that axillary bud traces arise in separate pairs from the parenchymatous layer of the stem. In this study one pair of bud traces was observed below the insertion of the axillary bud and proceeded vertically toward the axillary bud, whereas the other pair of bud traces arose below and proximal to the insertion

PAGE 174

162 of the axillary bud and proceeded laterally toward the axillary bud. The four bud traces finally fused in the cortex of the stem to form a single large vascular trace that proceeded into the axillary bud. The vertical and lateral progression of axillary bud traces and their fusion into a single trace proximal to the insertion of an axillary bud has also been reported by Manis (unpublished data) Specific bud traces cannot be identified in transections through the insertion of the secondary shoot in this study. The separate vascular bundles of the central cylinder of the stem continue directly into the secondary shoot. The secondary shoot receives the vascular tissue of the central cylinder directly, whereas an axillary shoot receives only lateral bud traces from the vascular tissue of the central cylinder. The secondary shoot (ultimate axillary bud of the primary shoot) of E. crassipes is the functional stem apex of the plant after the primary stem apex has terminated in an inflorescence. The anatomical features of the secondary shoot are identical with anatomical features of the stem (cf. Figures 32,37), whereas the anatomical features of the axillary shoot are identical with the anatomical features of the stolon (Figure 35). The secondary shoot has a distinct parenchymatous layer composed of 1-2 layers of small, rectangular, thin-walled parenchyma cells that separate the parenchymatous ground tissue of the central cylinder from the inner cortex. A distinct parenchymatous layer is absent in the axillary shoot. The axillary shoot has an outer compact cortical zone of large, thin-walled parenchyma cells and inner cortical

PAGE 175

163 zone of aerenchyma cells separated by large intercellular spaces, In addition to a compact outer cortical zone of large, thin-walled parenchyma cells and a middle cortical zone of aerenchyma cells separated by large intercellular spaces, the secondary shoot possesses an inner cortical zone of 3-5 layers of small, thin-walled parenchyma cells. Vascular bundles that occur in the cortex of central cylinder of an axillary shoot typically possess 1-3 conspicuous protoxylem lacunae and occasionally possess 1-2 primary xylem tracheids. Vascular bundles in the central cylinder or cortex of the secondary shoot, however, possess one metaxylem vessel and 4-8 primary xylem tracheids. Transections through the insertion of the secondary shoot on the primary shoot reveal that some anatomical features typical of a secondary shoot never occur in an axillary shoot. The prophyll and the first leaf of the secondary shoot are not morphologically similar to the prophyll and first leaf of a typical axillary shoot. The membranous prophyll of a secondary shoot arises distal (0.5-2.0 cm) to the insertion of the secondary shoot on the primary stem, whereas the prophyll of an axillary shoot arises at the insertion of the axillary bud on the primary stem. The membranous body of a prophyll of both an axillary shoot and a secondary shoot terminates ina stipule-flap and aborted lamina. The stipule-flap of the prophyll of a secondary shoot is broad, well-developed, and the same size as the stipule-flap that terminates the stipule of a typical leaf, however, Weber (1950) reported that the stipule-flap of the prophyll of an axillary shoot is small, poorly developed, and much

PAGE 176

164 smaller than the stipule-flap of the stipule of a typical leaf. The prophyll of a secondary shoot has an elongate, cylindrical aborted lamina, whereas the aborted lamina of the prophyll of an axillary shoot is small and flattened (Weber, 1950). After the prophyll the initial leaf formed by the secondary shoot is similar in size and morphology to the leaves produced by the primary shoot. If the primary shoot produces short leaves with reniform laminas and swollen petioles, the first leaf of a secondary shoot also has short leaves with reniform laminas and swollen petioles. The initial leaf of a secondary shoot has an ovate lamina and elongated petiole if the primary shoot produced elongated leaves before floral initiation. A prophyll-like leaf is initially formed by the axillary shoot after the prophyll. This first leaf is modified to protect the young ramet and is composed of a well-developed stipule with a broad stipule-flap. Weber (1950) demonstrated the well-developed, broad stipule, and broad stipule-flap of the first leaf of the axillary shoot. He also indicated that the first leaf had a long, narrow aborted petiole with an aborted lamina that originated at the midpoint of the stipule. This characteristic aborted petiole and lamina was evident in the present study and was noted to be distinctly different from a typical leaf with a distinct lamina and petiole. Although the second leaves of an axillary shoot and a secondary shoot are morphologically similar to a typical leaf of the primary plant, the first leaf of an axillary shoot is quite atypical and modified. The first leaf of a secondary shoot, however, is similar to a typical leaf of the primary shoot.

PAGE 177

165 A secondary shoot and an axillary shoot vary developmentally The time of development of these organs also varies. Axillary shoots produce an elongated internode (stolon) between the prophyll and the first leaf of the ramet that carries the new ramet horizontally to the water surface. The secondary shoot has no elongated internode and is oriented vertically. The secondary shoot initiates development shortly after floral initiation. Development of the secondary shoot continues regularly and the first mature leaf of the secondary shoot matures 2-5 days after anthesis of the terminal inflorescence of the primary shoot. Apical dominance arrests the development of the axillary shoot after the 6th-8th plastochron (Manis, unpublished data) and the axillary bud remains dormant until apical dominance is released. The axillary shoot dimorphism of jE. crassipes is caused primarily by the change in the primary stem apex from vegetative growth to reproductive growth. An axillary bud primordium formed on the apex of the stem of E. crassipes may become a secondary shoot or remain as an axillary bud. The development of the axillary bud primordium is determined by the developmental status of the stem apex. When the primary stem apex becomes reproductive and an inflorescence terminates the primary shoot, the ultimate axillary bud primordium becomes the new stem apex. When the primary stem apex is destroyed (i.e., by the herbivore Arzama densa ) the ultimate axillary bud or bud primordium develops into an axillary shoot or a secondary shoot, respectively. During Arzama infestations dimorphic axillary shoots (typical elongated lateral stolons or atypical nonelongated, vertical stolons)were observed.

PAGE 178

166 The typical, elongated, lateral stolons develop from axillary buds because the whole apex of the plant and apical dominance was removed by the insect. The atypical, nonelongated vertical stolons probably develop because only a portion of the apical meristem was removed and a proximal axillary bud primordium developed as a secondary shoot. Consequently, the proximity of an axillary bud or bud primordium to the apex possibly determines the type of development the bud will possess. The extensive and complete vascular attachment of the secondary stem apex to the axial vascular systems of the primary stem is determined by the proximal location of the bud primordium to the primary stem apex. Zimmerman and Tomlinson (1972) stated that the type of vascular attachment of a lateral organ to a monocot stem depends upon the proximity of the developing lateral organ to the origin of the axial vascular system. Tomlinson (1970) indicated that the axial vascular system in monocotyledons originated in the meristematic cap that is proximal to the leaf primordia. The ultimate axillary bud primordium of E. crassipes becomes the functional stem apex shortly after floral initiation. Because the ultimate axillary bud primordium is proximal to the primary apex, this bud primordium is able to redirect the uncommitted vascular bundles from the developing inflorescence into the secondary shoot. The inner (central cylinder) and the outer (cortical) axial vascular systems of _E. crassipes develop fully in the secondary shoot distal to the insertion of the secondary shoot in the primary stem because the secondary apex has assumed all morphogenetic and physiological functions of the primary apex.

PAGE 179

167 A small, lateral vascular attachment (vertical and lateral bud traces) occurs between a typical axillary shoot and the primary stem.. The primary stem apex remains intact throughout the development of the axillary bud primordium and only a few vascular bundles develop toward the axillary shoot apex. If a proximal axillary bud primordium becomes the vegetative apex (after destruction of the primary stem apex) the proximal axillary bud primordium receives a complete vascular attachment to the primary shoot. The peduncle of an inflorescence of E: crassipes is anatomically dissimilar to that of either the stem or stolon (cf. Figures 14 32,35). A distinct single-layered epidermis of isodiametric, thin— walled parenchyma cells and cuticle form the outer boundary of the peduncle. The outer zone of parenchymatous ground tissue of the peduncle is 5-7 cell layers thick and composed of small chlorenchyma cells and larger thin-walled parenchyma cells. The outer cortical zone of both a stem and a stolon are composed of 5-7 layers of large, thin-walled parenchyma cells, although no chlorenchyma cells are present. Stant (1964) reported that peduncles of members of the Alismataceae also possess a chlorenchymatous outer zone of ground tissue, whereas outer zones of ground tissue of stems of the same species are nonchlorenchymatous. Distinct cortical and central cylinder regions of ground tissue are not present in the peduncle, whereas these regions are discernible in both stems and stolons. Weber (1950) also reported that cortical and central cylinder regions were not distinct in the peduncle of

PAGE 180

168 E. crassipes and indicated that the aerenchymatous ground tissue of the peduncle strikingly resembled the aerenchymatous ground tissue of petioles of the leaves. In the present study a central lacuna is present in allthe peduncles of inflorescences, whereas a central lacuna is lacking in the leaf petioles. This characteristic is also observed in peduncles and petioles of E. azurea and Pontederia cordata Peduncles of Alisma plantago and Damasonium stellatum (Alismataceae) also have a central lacuna whereas leaf petioles of the same species are wholly aerenchymatous (Stant, 1964). The peduncle of the inflorescence of E. crassipes has only aerenchymatous ground tissue around the central lacuna. Stems and stolons of E. crassipes have a distinct parenchymatous central cylinder and a cortical region of both aerenchymatous and parenchymatous tissues. Vascular bundles are randomly distributed in both the central cylinder and aerenchymatous cortical zone of the stem and stolon, whereas vascular bundles are randomly distributed in the aerenchymatous ground tissue of the peduncle. Vascular bundles are typically arranged in concentric rings in peduncles of members of the Alismataceae (Stant, 1964), whereas vascular bundles in the peduncle of E. crassipes are randomly distributed. In this study the collateral vascular bundles of the peduncles of E. crassipes have primary xylem composed of 1-3 protoxylem tracheids and a protoxylem lacuna. The primary xylem of the collateral vascular bundles of the stolon is composed of 1-3 conspicuous protoxylem lacunae and occasionally 1-2 primary xylem tracheids, whereas the primary xylem of the collateral

PAGE 181

169 vascular bundles of the stem is. composed of 1—8 primary xylem tracheids and a single metaxylem vessel. The primary phloem in the peduncle consists of 2—6 sieve tube members and companion cells. Perivascular fibers are external to the primary phloem and occur within the layer of large, thin-walled parenchyma cells that surround each vascular bundle of the peduncle. The vascular bundles in the central cylinder of both stem and stolon are surrounded only by ground parenchyma. Perivascular fibers do not occur in these bundles. Anatomical changes occur within the peduncle of the inflorescence of E. crassipes as postanthetic bends form and lower the flowers to the water surface for fruit development and seed dispersal. Postanthetic bends in the peduncle of both E. crassipes and Pontederia cordata are distinct and necessary to bring the emergent inflorescences to the water surface. The postanthetic bend in the peduncle of E_. azurea is small and not distinct, however, because the inflorescence is borne near the water surface. Arber (1920) noted that numerous species of aquatic plants (Aponogetonaceae, Hydrocharitaceae, Pontederiaceae) also have peduncle bends or submerse entirely in order to bring their fruit to the water surface. The thin-walled, rectangular parenchyma cells of both the epidermis and the outer layer of ground tissue on the convex side of the peduncle increase in length 1.3-1.6 times (.050-. 135 mra) during postanthetic curvature. The thinwalled rectangular parenchyma cells of both the epidermis and outer layer of ground tissue on the concave side of the peduncle remain relatively constant in size (.035-. 090 mm) during postanthetic

PAGE 182

170 curvature (Figures 28, 29)Weber (1950) demonstrated that the length of epidermal and cortical cells on the convex side of the peduncle increased approximately 1.4-1.9 times compared to cells on the concave side, whereas Das (1967) reported that the hypodermal cell lengths on the convex side are 2,1-2,6 times longer than cells on the concave side. Weber (1950) correlated the bending phenomena of postanthetic inflorescence development of E. crassipes with the high density of stomates and associated chlorophyllous tissue that occur in the region of the capital bend. He indicated that the chlorophyllous tissues were functionally responsible for the bending phenomena. Differential localization of auxin, however, may be responsible for the bending phenomena. Differential cell elongation in coleoptiles is caused by auxin activity (Davies, 1973) and auxin concentration is an important factor in cell and stem elongation in most plant species (Galston and Purves, 1960). Altered auxin levels caused the bending phenomenon in the gynophore that directs the fruit of the peanut ( Arachis hypogea ) underground (Jacobs, 1951). Differential cell elongation by localized auxin activity may cause the postanthetic curvature of the inflorescence of the waterhyacinth, although the origin of the auxin is unknown. Salisbury and Ross (1969) indicated that high auxin concentrations typically occur in pollinated flowers of most plant species. Previous workers (Rao, 1920b; Bock, 1966; Das, 1967) have reported that postanthetic bending of the inflorescence of E. crassipes occurs if pollination was prevented or after removal of the flowers o r the whole inflorescence. If high amounts of auxin

PAGE 183

171 occur in the pollinated flowers of E. crassipes this auxin probably has no effect on postanthetic curvature. The characteristic inner and outer bracts of the inflorescence are inserted opposite each other at the midpoint of the inflorescence. (Figure 9). The outer bract hasa well-developed sheathing base that terminates adaxially in a stipule-flap and adaxially in a subcordate lamina. The inner bract has a sheathing base that is thinner and more membranous than the sheathing base of the outer bract and terminates adaxially in a stipule-flap and abaxially in an aborted lamina. A membranous inner bract that terminates in an aborted lamina and an outer bract that terminates in a leaf-like lamina have also been observed in preliminary studies on the inflorescences of E. azurea and Pontederia cordata In the Pontederiaceae flowers are subtended by two bracts with sheathing bases and the outer bract is terminated by a leaf-like lamina (Castellano, 1958). The lamina of the outer bract of F,. crassipes has a subcordate base, whereas the lamina of a typical leaf of this species is larger and has an obtuse to truncate base (cf. Figures 12,41). No differences in gross morphology between the lamina of the outer bract and the lamina of a typical leaf of E_. azurea and Pontederia cordata were observed in this study. The lamina of the outer bract of the inflorescence of E. crassipes is similar anatomically to the lamina of a typical leaf. The lamina of an outer bract has two layers of palisade mesophyll near the adaxial surface. The palisade mesophyll is interrupted

PAGE 184

172 by a series of collateral bundles wiLh xylem adaxial Lo phloem. One layer of palisade mesophyll occurs near the abaxial surface of the lamina. This layer of palisade mesophyll is interrupted by a series of small collateral bundles with xylem abaxial to phloem (inverted orientation) (Figure 20). Previous workers (Olive, 1894; Arber, 1920; Bruhl and Dutta, 1923) also reported that a lamina of a leaf of _E. crassipes was composed of an adaxial and an abaxial layer of palisade mesophyll separated by collateral vascular bundles with normal and inverted orientations. Arber (1920) also demonstrated that the leaf of _E. crassipes is anatomically similar to the unifacial leaf of Pontederia cordata and Heteranthera zos teraef olla The aborted lamina of the inner bract has a unifacial anatomy. Chlorenchyma cells (3—5 layers) occur near both the abaxial and adaxial surfaces, and collateral vascular bundles with primary xylem interior to primary phloem are dispersed in the ground tissue (Figure 19). The unifacial anatomy characteristic of the laminas of the bracts and leaves of E. crassipes and other members of the Pontederiaceae is a major factor in the "phyllode theory" (Arber, 1918). Arber (1918, 1920, 1925) reported that leaves of many species of monocotyledons had unifacial anatomical characteristics. She described vascular bundles with an inverted orientation from leaves of many species with and without laminas. She theorized the laminas of monocot leaves were lost evolutionarily when monocotyledons were derived from dicotyledons. She interpreted the lamina of the leaf of E. crassipes as a pseudolamina because it appeared to be an

PAGE 185

173 expanded and flattened petiole. Kaplan (1973) demonstrated that the lamina of dorsiventral leaves (unifacial) of some monocotyledons is not positionally equivalent to the lamina of dicotyledonous leaves. He stated that these dorsiventral laminar leaves are developmental elaborations of the basal, meristem-encircling part of the leaf primordium, and the distal upper leaf zone (homologous to the lamina of dicot leaves) remains rudimentary. Previous work (Bruhl and Dutta, 1923; Roth, 1949) indicates that the leaves of E. crassipes may be developmentally similar to other monocotyledon species with dorsiventral laminar leaves. In this study the rudimentary upper leaf zone (precursor tip) is an apical brown portion on the laminas of both the outer bract and leaves of E_. crassipes Bruhl and Dutta (1923) also noted the aborted brown apical portion of the laminas of the leaves and demonstrated that the vascular bundles in the precursor tip of young leaves arose from the laminar portion of the leaf. The precursor tip also has a cylindrical shape (Bruhl and Dutta, 1923), as shown for the precursor tip of other monocotyledonous leaves (Kaplan, 1973). Kaplan (1973) reported that monocot leaves with distinct morphological differentiation between lamina and petiole had small precursor tips. Bruhl and Dutta (1923) indicated that the precursor tip of leaves of E. crassipes had a length of 0.2 mm, whereas the precursor tip of leaves of Ornitholgalum caudatum was typically 25 mm long (Kaplan, 1973). The lamina and petiole of 0. caudatum were poorly differentiated (Kaplan, 1973), whereas in this study the lamina and petiole of leaves of E. crassipes are morphologically distinct. Kaplan (1973) also reported that monocot leaves

PAGE 186

174 with distinct morphological differentiation hetween lamina and petiole demonstrated this differentiation very early in the development of the leaf, Roth (1949) reported that leaf primordia of JE, crassipes (0.20 mm) and j?. cordata (0,38 mm) are very small when the lamina is morphologically distinct from the petiole. Thus, differentiated leaf primordia of E. crassipes are small and similar in size to the differentiated leaf primordia (0.25 mm) of Zantedeschia aethiopica (Kaplan, 1973). Kaplan analyzed the ontogeny of the leaves of this species and determined that this species possesses a dorsiventral laminar leaf. Leaves of E_. crassipes possess a very small but distinct precursor tip, are differentiated into lamina and petiole, and exhibit this differentiation early in their development. These morphological features are characteristic of dorsiventral laminar leaves (Kaplan, 1973). Because the leaves of E_, crassipes appear to be dorsiventral laminar leaves, the leaves of this species would possess a true lamina. Although the laminas of leaves of E_. crassipes have unifacial anatomical features, the leaf is probably morphologically similar to typical bifacial leaves with true laminas. Consequently, the term pseudolamina (Arber 1918, 1920) is not valid for the lamina of the waterhyacinth leaf. The vascular anatomy of the mature flowers of E. crassipes reveals a sequential progression of vascular traces from a vascular plexus into each succeedingly distal group of floral appendages (Figure 22). Each perianth member has three vascular traces that originate from a vascular plexus at the base of the flower, although the petal traces separate from the vascular plexus distal to the

PAGE 187

175 separation of sepal traces., Singh (1962) reported vascular traces from sepals and petals separated simultaneously from the vascular plexus in flowers of 15. crassipes and Monochoria spp. (Pontederiaceae) .. Stamen traces of E_. crassipes occur conjoint to the median perianth trace until the distal level of carpel-perianth tube separation, although Singh (1962) indicated that separation of a stamen trace from a median perianth trace occurs at the level of the separation of the carpel from the perianth tube. Singh (1962) also reported that the conjoint stamen-median perianth trace of Monochoria spp. separates proximal to its origin in the vascular plexus. Stamen filaments are epiphyllous only at the base of the perianth members of Monochoria (Singh, 1962), whereas stamen filaments of E. crassipes are epiphyllous upon the perianth members for the entire length of the perianth tube. The three carpels of a flower of E_. crassipes have two marginal vascular bundles and a single median vascular bundle. Singh (1962) reported that three marginal carpellary traces proceeded from a vascular plexus, separated dis tally, and formed the marginal bundles of each carpel, whereas in the present study six marginal carpellary traces proceed from a vascular plexus and form the marginal carpellary bundles. The marginal carpellary bundles do not arise on the same radius as the median carpellary bundle, although Singh (1962) reported a common radius for the origin of these bundles. Although Singh (1962) indicated that the median carpellary bundles separated from the vascular plexus proximally to the origin of the marginal carpellary bundles, the latter carpellary bundles separated from the vascular plexus proximal to the origin of the median bundles in this

PAGE 188

176 study. Distal to the hase of the carpel the marginal vascular bundles of each carpel undergo multiple divisions to form large vascular bundles, whereas Singh (1962) indicated that a single large vascular bundle occurred in each carpel and was the single trace from the vascular plexus. In the flower of Monochoria spp. Singh (1962) observed six large marginal carpellary traces that occurred as a single ring of vascular tissue. He reported that primary xylem occurred internally in these large vascular bundles, whereas in this study primary xylem was interspersed with primary phloem in the large marginal bundles of the carpels of 12. crassipes The marginal carpellary bundles terminate in the apex of the carpels, whereas the median carpellary bundles continue into the style and terminate in the stigma. The marginal carpellary bundles and the median carpellary bundle of each carpel of the flowers of E. crassipes and Monochoria spp. also terminate in the apex of the carpel and the stigma, respectively (Singh, 1962). The intercarpellary walls of the flowers of E_. crassipes vary anatomically distal to the base of the carpels. At the base of the carpels the intercarpellary wall is composed of 4-7 layers of small, isodiaraetric, thin-walled parenchyma cells, although these cells deteriorate distally. This deterioration of cells in the intercarpellary walls is more pronounced in the flowers of Monochoria vaginalis (Singh, 1962). A canal lined witli large densely stained parenchyma cells appears within each intercarpellary wall and continues from the base of the carpel to the level of the placentae.

PAGE 189

177 These blind canals may be stylar canals lined with transmitting tissue, although theyare located in the proximal end of the pistil. Because transmitting tissue is evident in the distal region of the pistil and it is not connected to the canals that occur in the base of the pistil, no function for these basal canals is known. The functional transmitting tissue of the carpels is located between the divided placentae in the carpels, continues through the style, and terminates in the stigma (cf. Figures 26,27). Singh (1962) did not observe these canals in the intercarpellary walls of flowers of either E_. crassipes or Monochoria spp. Characteristic features of the seed coat of mature seeds are visible in the carpels of the flower of E_. crassipes twelve days after pollination (Figure 31) Longitudinal ridges derived from the outer integument of the ovule are apparent. Previous workers (Coker, 1907; Parija, 1934; Penfound and Earle, 1948) noted these external ridges on seeds of E. crassipes and the latter authors indicated that 12-15 ridges typically occur on each mature seed. Castellano (1958) has noted longitudinal ridges on the seed coat are typical in the Pontederiaceae. In the present study a double layer of macrosclerids occurs in the seed coat of E_. crassipes and is derived from the inner integument of the ovule. Coker (1907) reported that the sclerified seed coat of both E. crassipes and Heteranthera limosa (Pontederiaceae) is derived from the inner integument. The perianth members, petals and sepals, of the flower of E. crassipes are anatomically similar (Figure 27) The perianth

PAGE 190

178 members of 12. crassipes are fused for 1/3 of their length into a perianth, tube, whereas the perianth members of the flowers of Monochoria spp. are free (Singh, 1962). The ground tissue of the fused petals and sepals of _E.. crassipes is composed of an outer zone of 4-6 layers of thin-walled, isodiametric parenchyma cells, a middle zone of aerenchyma cells separated by large intercellular spaces in a single series, and an inner zone of 3-4 layers of thin-walled, isodiametric parenchyma cells. The vascular traces from the vascular plexus proceed into each perianth member, divide repeatedly, and form two rows of unequal-sized bundles. The larger vascular bundles occur in a single row in the outer zone of ground tissue and possess primary xylem adaxial to primary phloem. The smaller vascular bundles possess primary xylem abaxial to primary phloem and occur in a single row in the inner zone of ground tissue of each perianth member. Although inverted bundles occur in the perianth members of E_. crassipes inverted bundles do not appear in perianth members of Monochoria (Singh, 1962). He concluded that the unifacial nature (inverted vascular bundles) of the perianth members of E. crassipes indicated the perianth were phyllodic in nature.

PAGE 191

SUMMARY The following statements summarize this comparative ecological and morphological study of sexual reproduction in Eichhornia crassipes (Mart.) Solms with emphasis on anatomical features of sexual reproduction. 1. Increases in nutrient concentrations increased leaf size of waterhyacinth plants and decreased maximum rootlength and the overall root system. The number of stolons per unit area and the amount of total mass (standing crop) was not affected by changes in nutrient concentrations. 2. Stress by cold temperatures or herbivore infestations caused increased stolon number per unit area. Increased numbers of stolons resulted in decreases in maximum rootlengths, leaf size, and number of leaves in a given waterhyacinth population. 3. Waterhyacinth plants exhibited a distinct morphological leaf type when the waterhyacinth population was invading an open site. Under conditions of low plant density and low intraand interspecific competition, leaves of waterhyacinth plants had short, swollen petioles that comprised 40-60% of the total leaf length and reniform laminas that were wider than long. 179

PAGE 192

180 4. Waterhyacinth plants exhibited a distinct morphological leaf type in stable, competing waterhyacinth populations. Leaves of waterhyacinth plants had elongated, narrow petioles that comprised 70-80% of the total leaf length and ovate laminas that were longer than wide under conditions of high population density and intense intraor interspecific competition. 5. Floral initiation in waterhyacinths was not correlated with weekly means of mean temperature, high temperatures, low temperatures, photoperiod, total solar insolation, or rainfall. Mean low temperatures of 6-15C prevented floral initiation and indicated that a threshold low temperature may be necessary for flowering in E_. crassipes 6. The nutrient status of the waterhyacinth plants and of their habitat was correlated to floral initiation. Floral initiation occurred in waterhyacinths during periods of vigorous growth and relatively low nitrogen concentrations. Low nutrient conditions occur annually in the native habitat of _E. crassipes and may stimulate sexual reproduction and seed production to maintain the waterhyacinth population through the periods of catastrophic mortality caused by the floods of the annual rainy season. 7. Increases in nutrient concentration increased the flower number and the size of both the inflorescence and the

PAGE 193

131 individual bracts but had no effect on the size of floral appendages. The growth rate of the inflorescence was not affected by nutrient concentrations and inflorescence development required 6-9 days. 8. Natural pollination and seed set were observed infrequently at the study sites, although successful artificial pollination and seed set was affected at both the study sites and in the greenhouse. 9. Widespread seed germination occurred on reflooded peaty sediments during high temperatures and high light intensities of Summer after a controlled drawdown and reflooding at one study site. 10. The adventitious roots of E. crassipes have an aerenchymatous cortex with large, radially arranged intercellular spaces and a polyarch stele. 11. Stems and stolons of _E. crassipes have similar anatomical characteristics. The central cylinder of both the stem and stolon is parenchymatous, whereas the cortex of both organs has an aerenchymatous zone. Collateral vascular bundles occur in both the central cylinder and the cortex of the stem and stolon. These vascular bundles occur more frequently in the central cylinder than in the cortex of these organs. 12. The peduncle of the inflorescence is composed of aerenchymatous ground tissue and a large central lacuna. Collateral vascular bundles are scattered throughout the

PAGE 194

182 aerenchymatous ground tissue of the peduncle. 13. Two bracts with sheathing bases are borne on the inflorescence of JE. crassipes The outer bract has a well-developed sheathing base that terminates in a membranous stipule-flap and a well-developed lamina. The sheathing base of the inner bract is thinner, however, and terminates in a membranous stipule-flap and an aborted lamina. 14. Eichhornia crassipes exhibits a sympodial growth pattern. The inflorescence terminates the primary shoot and vegetative growth continues from the ultimate axillary bud (secondary shoot). The anatomy, prophyll morphology, and development of the secondary shoot (after floral initiation) is dissimilar from that of a typical axillary shoot (during vegetative growth) 15. Postanthetic bends occur at the base (basal), proximal to the bases of the inflorescence bracts (capital) and in the interfloral region (interf loral) of the peduncle of the inflorescence of E. crassipes Differential cell elongation on opposite sides of the peduncle occurs at the location of each postanthetic bend. 16. Sepals, petals, and carpels of the flower of E_. crassipes have three vascular traces that originate from the vascular plexus, respectively, whereas the stamens have a single vascular trace that arises from the median vascular trace of each perianth member. Sepals, petals, and carpels are

PAGE 195

183 similar anatomically and possess an aerenchymatous middle zone of ground tissue with narrow parenchymatous zones near the adaxial and abaxial surfaces.

PAGE 196

LITERATURE CITED Agharkar, S.P., and I. Banerji. 1930. Studies in the pollination and seed formation of waterhyacinth (Eichhornia speciosa Kunth). Agr. J. India. 25:286-296. Agostini, G. 1974. El genero Eichhornia (Pontederiaceae) Acta Bot. Venez. 9 (1-4): 303-310. American Public Health Association. 1971. Standard. Methods for the Examination of Water and Wastewater. 13th ed. APHA 769 pp. Anderson, L.E. 1954. Hoyer's solution as a rapid permanent mounting medium for bryophytes. Bryologist 57:242-244. Arber, A. 1918. The phyllode theory of the monocotyledonous leaf with special reference to anatomical evidence. Ann. Bot. 32:465-501. Arber, A. 1920. Water Plants; A Study of Aquatic Angiosperms University Press, Cambridge. 436 pp. Arber, A. 1925. Monocotyledons A Morphological Study. University Press, Cambridge. 258 pp. Barton, L.V., and J.E. Hotchkiss. 1951. Germination of seeds of Eichhornia crassipes Solms. Contr. Boyce Thompson Inst. 16:215-220. Batanouny, K.H., and A.M. El-Fiky. 1975. The waterhyacinth (Eichhornia crassipes Solms) in the Nile System, Egypt. Aquatic Bot. 1(3) : 243-252. Bock, J.H. 1966. An ecological study of Eichhornia crassipes with special emphasis on its reproductive biology. PhD thesis. Univ. of California, Berkeley. Boyd, C.E., and E. Scarsbrook. 1975. Influence of nutrient additions and initial density of plants on production of waterhyacinth, Eichhornia crassipes Aquatic Bot. 1(3): 253-261. Brinson, M.M. 1973. The organic matter budget and energy flow of a tropical aquatic ecosystem. PhD thesis. Univ. of Florida, Gainesville. 184

PAGE 197

185 Bruhl, P., and A. Dutta. 1923. Commentationes phytomorphologicae et phytophysiologicae. II. Elchhornia studies. Jour. Dept. Sci., Univ. Calcutta 4:1-9. Bruhl, P., and J.S. Gupta. 1927. Commentationes phytomorphologicae. IV. Eichhornia studies. III. On the production of ripe seeds by artificial pollination of Eichhornia s peciosa Jour. Dept. Sci., Univ. of Calcutta 8: 1-8. Carlquist, S. 1960. Anatomy of Guayana Xyridaceae, Abolboda Orectanthe and Achtyphila Mem. N.Y. Bot. Gard. 10(2): 65-117. Castellano, A. 1958. Las Pontederiaceae de Brasil. Arq. Jard. Bot. Rio de J. 16:147-236. Center, T.D. 1976. The potential of Arzama densa for the control of waterhyacinth with special reference to the ecology of waterhyacinth (Eichhornia crassipes ). PhD thesis. Univ. of Florida, Gainesville. Chailakhyan, M.K. 1968. Internal factors of plant flowering. Ann. Rev. Plant Physiol. 19:1-36. Charlton, W.A. 1973. Studies in the Alismataceae. II. Inflorescences of Alismataceae. Can. J. Bot. 51:775-789. Champness, S.S., and K. Morris. 1948. The population of buried viable seeds in relation to contrasting pasture and soil types. J. Ecol. 36:149-173. Cheadle, V.I. 1970. Vessels in Pontederiaceae, Ruscaceae, Smilaceae, and Trilliaceae. New Research in Plant Anatomy. J. Linn. Soc. Lond. Bot. 63(Suppl. 1): 45-50. Coker, W.C. 1907. The development of the seed in the Pontederiaceae. Bot. Gaz. 44:293-301. Couch, R. 1971. Morphology and anatomy of waterhyacinths. Proc. 24th S. Weed Sci. Soc. p. 332. Das, R.R. 1967. Growth and distribution of Eichhor nia crassipes (Mart.) Solms and Spirodela polyrrhiza (L. ) Schleid. PhD thesis. Banaras Hindu Univ. Das, R.R. 1969. A study of reproduction in Eichhornia c rassipes Trop. Ecol. 10(2): 195-198. Davies, P.J. 1973. Current theories on the mode of action of auxin. Bot. Rev. 39:139-171.

PAGE 198

186 Doorenbos, J., and S.J. Wellensiek. 1959. Photoperiodic control of floral initiation. Ann. Rev. Plant Physiol. 10:147-184. Earle, T.T. 1947. The flowering cycle of waterhyacinth. Proc. La. Acad. Sci. 10:27-29. Evans, L.T. 1969. The Induction of Flowering. Some Case Histories. Cornell Univ. Press, Ithaca, N.Y. 488 pp. Francois, J. 1964. Observations sur 1'heterostylie ches Eichhornia crassipes (Mart.) Solms. Acad. Roy. Sci. d'Outre-Mer Seances. 1964:501-519. Galston, A.W. and W.K. Purves. 1960. The mechanism of action of auxin. Ann. Rev. Plant Physiol. 11:239-276. Gay, P. A. 1960. Ecological studies of Eichhornia crassipes Solms in the Sudan. J. Ecol. 48:133-191. Haigh, J.C. 1936. Notes on the waterhyacinth (Eichhornia crassipes Solms) in Ceylon. Ceylon J. Sci. Sect A 12(2) : 97-108. Hall, J.B., and D.U.U. Okali. 1974. Phenology and productivity of Pistia stratiotes L. on the Volta Lake, Ghana. J. App. Ecol. 11:59-61. Haller, W.T., and D.L. Sutton. 1973. Effect of pH and high phosphorous concentrations on the growth of waterhyacinths. Hyacinth Control J. 11:59-61. Hestand, R.S., and C.C. Carter. 1975. Succession of aquatic vegetation in Lake Ocklawaha two growing seasons following a winter drawdown. Hyacinth Control J. 13:43-47. Hitchcock, A.E., P.W. Zimmerman, H. Kirkpatrick, and T.T. Earle. 1949. Waterhyacinth its growth, reproduction, and practical control by 2,4-D. Contr. Boyce Thompson Inst. 15:363-401. Holm. C. 1969. Weed problems in developing countries. Weed Sci. 17(1):118-133. Jacobs, W.P. 1951. Auxin relationships in an intercalary meristem: further studies of the gynophore of Arachis hypogea L. Amer. J. Bot. 38:307-310. Johansen, D.A. 1940. Plant Microtechnique. McGraw-Hill, New York. Kaplan, D.R. 1973. The monocotyledons: their evolution and comparative biology. VII. The problems of leaf morphology and evolution in the monocotyledons. Q. Rev. Biol. 48:437-457.

PAGE 199

187 Knipling, E.B., S.H. West, and W.T. Haller. 1970. Growth characteristics, yield potential, and nutritive content of waterhyacinths. Soil Crop Sci. Soc. Fla. Proc. 30:51-63. LaGarde, R.V. 1930. A plant that stopped navigation. Missouri Bot. Gard. Bull. 18:48-51. Meyer, B.S., and D.B. Anderson. 1952. Plant Physiology. 2nd ed. D. Van Nostrand, New York. 284 pp. Miller, E.G. 1938. Plant Physiology. McGraw-Hill, New York. 1200 pp. Mohamed, B.F. 1975. The ecology of waterhyacinth in the White Nile, Sudan. Hyacinth Control J. 13:39-43. Morris, T.L. 1974. Waterhyacinth, Eichhornia crassipes (Mart.) Solms: its ability to invade aquatic ecosystems of Payne's Prairie Preserve. MS thesis. Univ. of Florida, Gainesville. Murneek, A.E. 1948. Nutrition and metabolism as related to photoperiodism. pp 83-90. in Murneek, A.E., and R.O. Whyte ed. Vernalisation and Photoperiodism. Olive, E.W. 1894. Contributions to the histology of the Pontederiaceae. Bot. Gaz. 19:178-184. Parija, P. 1930. A preliminary note on the physiology of the seedlings of the waterhyacinth (Eichhornia speciosa ) Agr. J. India 25:386-391. Parija, P. 1934. A note on the reappearance of waterhyacinth seedlings in cleared tanks. Indian J. Agr. Sci. 4:1049. Penfound, W.T., and T.T. Earle. 1948. The biology of the waterhyacinth. Ecol. Monogr. 13:447-472. Rao, P.S.J. 1920a. The formation of leaf bladders in Eichhornia speciosa Kunth (waterhyacinth). Jour. Ind. Bot. 1: 219-225. Rao, P.S.J. 1920b. Note on the geotropic curvature of the inflorescence of Eichhornia speciosa Kunth. Jour. Ind. Bot. 1:217-213. Richardson, L.V. 1975. Water level manipulation: a tool for aquatic weed control. Hyacinth Control J. 13:8-11. Robertson, H.F., and Ba Thein. 1932. The occurrence of waterhyacinth (Eichhornia crassipes ) seedlings under natural conditions in Burma. Agric. and Livestock, India. 2(4) : 383-390.

PAGE 200

183 Roth, I. 1949. Zur Entwicklungsgeschichte des Blattes, mit besonderer Berucksichtigung von Stipularund Ligularbildungen. Planta 37(3) : 299-336. Salisbury, F.B., and C. Ross. 1969. Plant Physiology. Wadsworth Publishing. Belmont, Calif. 740 pp. Sculthorpe, CD. 1967. The. Biology of Aquatic Vascular Plants. St. Martens Press, New York. 610 pp. Searle, N. E. 1965. Physiology of flowering. Ann. Rev. Plant Physiol. 16:97-113. Singh, V. 1962. Vascular anatomy of the flower of some species of the Pontederiaceae. Proc. Indian Acad. Sci., Ser. B 56:339-353. Stant, M.Y. 1964. Anatomy of the Alismataceae. J. Linn. Soc. Lond. Bot. 59:1-42. Stevens, M.L. 19 76. The relationships of seasonal changes in plants, sediments, nutrients, and waterlevel in a small pond in north Florida. MS thesis. Univ. of Florida, Gainesville. Steele, R.G.D., and J.H. Torre. 1960. Principles and Procedures of Statistics. McGraw-Hill, New York. 481 pp. Sutton, D.L., and S.W. Bingham. 1973. Anatomy of emersed parrotfeather. Hyacinth Control J. 11:49-54. Tag el Seed, M. 1972. Some aspects of the biology of Eichhornia crassipes PhD thesis. Univ. of Khartoum, Sudan. Tag el Seed, M. and M. Obeid. 1975. Sexual reproduction of Eichhornia crassipes (Mart.) Solms in the Nile. Weed Res. 15:7-12. Tomlinson, P.B. 1970. Monocotyledons towards an understanding of their morphology and anatomy. Adv. Bot. Res. 3:207-292. Vickery, P.J. 1972. Grazing and net primary productivity of a temperate grassland. J. App. Ecol. 9:307-314. Vogel, E. and A.D. Oliver, Jr. 1969. Evaluation of Arzama densa as an aid in the control of waterhyacinths in Louisianna. J. Econ. Entomol. 62 (1) : 142-145. Weber, H. 1950. Morphologische und anatomische studien uber Eichhornia crassipes. Abhandl. Akad. Wissensch. Lit. Mainz Math. Nat. Kl. 1950(6) : 135-161.

PAGE 201

189 West, P.W., and T.P. Ramachandran. 1966. Spectrophotometry determination of nitrate using chromotropic acid. Anal. Chim. Acta 35:317-324. Wodzinski, R.J. 1975. Chemical, physical, and biological composition of "typical" effluents. Fla. Sci. 33:194-201. Zeevaart, J.A.D. 1976. Physiology of flower formation. Ann. Rev. Plant Physiol. 27:321-348. Zeiger, C.F. 1962. Hyacinth obstruction to navigation. Hyacinth Control J. 1:16-17. Zimmerman, M.H., and P.B. Tomlinson. 1972. The vascular system of monocotyledonous stems. Bot. Gaz. 133:141-155.

PAGE 202

190 BIOGRAPHICAL SKETCH Robert Gerald Anderson was born March 12, 1948 in Duluth, Minnesota, He graduated from Two Harbors High School, Two Harbors, Minnesota in 1966. In September 1966, he began his undergraduate education at the University of Minnesota, Duluth and transferred to the University of Minnesota, Minneapolis in September, 1967. He was awarded the Bachelor of Science degree with a biology major in June, 1970. In June, 1970, he entered graduate studies in the Department of Horticultural Science, University of Minnesota, St. Paul. He received the degree Master of Science in Horticulture in March, 1973. He entered graduate studies at the University of Florida in September, 1972, where he is presently a candidate in the Department of Botany for the degree Doctor of Philosophy, Robert G. Anderson is married to Joanne Marie Erno Anderson and is the father of one boy and one girl. He is a member of Pi Alpha Xi honorary ornamental horticulture fraternity.

PAGE 203

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. ftp, ypc-iw^ Terry W. Lucansky s Chairman Associate Professor of Botany I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Mildred M. Griffith Professor of Botany
PAGE 204

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. rhn J. Ewe'l Kc John J. Ewe'l ^J'"Associate Professor of Botany I certify that I have read this study and that in my opinion it conforms to acceptable standarts of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. David L. Sutton Associate Professor of Agronomy This dissertation was submitted to the Graduate Faculty of the Department of Botany in the College of Arts and Sciences and to the Graduate Council, and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy, December, 1976 V Dean, .Graduate School

PAGE 205

UNIVERSITY OF FLORIDA 3 1262 08666 417 3


xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID E2HKGPUQO_O8JV2Q INGEST_TIME 2017-10-26T17:39:00Z PACKAGE UF00102836_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES


xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID ECWZOXEC0_JFEGJL INGEST_TIME 2017-07-11T21:22:04Z PACKAGE UF00102836_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES


xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID E913XNTSG_MJXEI5 INGEST_TIME 2017-11-14T18:07:55Z PACKAGE UF00102836_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES