SPECIES ABUNDANCE RELATIONSHIPS OF AQUATIC INSECTS
IN MONOTYPIC WATERHYACINTH COMMUNITIES IN FLORIDA,
WITH SPECIAL EMPHASIS ON FACTORS AFFECTING DIVERSITY
JOSEPH KESTUTIS BALCIUNAS
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
Dedicated to my father. Jurgis,
whose support made this possible.
This study would not have been possible without the
contributions from many authorities. I would like to
acknowledge the following for verifying my identifications
of aquatic insect groups for which they are the acknowl-
edged authorities: Dr. Lewis Berner, Dr. Oliver Flint,
Dr. Dale Habeck, Dr. John Hellman, Dr. Jon Herring, Dr.
A. S. Menke, Dr. Paul J. Spangler, Dr. 11inter J. Westfall,,
and Dr. Frank Young. My special thanks to Mlr. William
Beck for identifying all thousand-odd chironomid larvae.
I would like to give special thanks to Drs. Minter J.
Westfall and Archie Carr, during whose courses this project
was conceived, and to my major professor, Dr. Habeck, whose
support and perseverance allowed completion of the study.
I would like to thank Dr. Ramon Littell and Mlr. Walter
Offen of the University of Florida Department of Statistics
for their help w~ith- statistical analyses of my data. I would
also like to thank my former roommate, Mlr. Roger Jones, pres-
ently at the Dartmouth Department of Physics, for his guid-
ance concerning the mathematical portions of this study.
I also acknowledge Florida Collection of Arthropods
andi the Northeast Regional Dat~a Cent-er (NE1RDC) for use of
their facilities and Florida Department of Natural Resources
for Lheir partial support of this study.
Also wish to acknowledge a special debt of gratitude
to L~inda M. Barber for her editing and typing of the manu-
script and for her general support during the analyses and
documentation stages of this project.
TABLE OF CONTENTS
ACKN.~OW~EIGLEDGENT . . . . . iii~~ii
LIST OF TABLES . . vii
LIST OF FIGURES .. . . . .viii
ABSTRACT . . .. . .~~~ ix
INTRODUCTION . . 1
LITERATURE REVIEW .. . . . 3
Eichhornia cr-assipes (M~artius) Solms-Laubach . . 3
Description .. . . ~ 3
Distribution--Flori~da .. 5
Productivity and Reproduction . . . .. 7
Environmental Requirements . . . .. 8
WJater Quality . . . 10
Economic Importance . . . 12
Habitat for Aquatic Insects . .. . 15
HIETHIODS . .. ... 19
Samnple Site Selection . . . . 19
Collection Methods .. . .. .. . 21
Water Quality . . . . . . 21
Identification . ... . .. 24
Analyses . . . . .. 28
RESI!LTS AND DISCUSSION . .. . .. 41
General . . .. 41
Comlments on Species List . . . . 41
Physical and Chemical Variable
In terre lat ions hips . . . 119
Regression Analyses . . . .. 120
TABLE OF CONTENTS (Con~tinued)
RESULTS AND DISCUSSION (Continued)
Comparison of Study Sites .. .. . 120
Description . .. . . . 120
Species List Comparisons . . .. . .. 123
Estimation of Total Number of Species . .. 128
Factors Affecting the Numnber of Species
and Genera . .. . . 140
Diversity Studies . . . .. . 141
General . . .. . .. . . 141
Dependence on Sample Size . . . . .. 142
Relations to Plant Part Size, Depth~and
Time of Year ... .... 144
Relationships to Water Quality Parameters . 146
Conclusions--Choice of Indices . . . . 147
SUMMIARY ..... ..~~. 150
REFERENCES CITED . .. . . .. 152
APPENDIX A . . . . 164
Fortran program for Sh~annon's and Simpson's
diversity indices (and Brillouin's, whefn possible)
APPENJDIX B . . . 165
Fortran program for rarefaction diversity index
BIOGRAPHICAL SKE~TCH . .. . ... . 166
LIST OF TABLES
Table 1. List of collection sites .
Table 2. Annotated list of insects
Table 3. Ten most abundant insects
Table 4. Ten most frequent insects
Table 5. WJater quality correlation
Table 6. TIen most abundant insects,
Table 7. Ten most frequent insects,
Table 8. Ten most abundant insects,
Table 9. Ten m7ost frequent insects,
Table 10. Ten most abundant insects,
Table 11. Ten most frequent insects,
. . . 20
. . . 42
. . 56
. . 67
. . 68
. . 69
. . 101
. . 117
. . . 117
. . . 121
Camps Canal . . 124
Camps Canal . . 124
Lake Alice . . 125
Lake Alice . . 125
I-75 ditch . - 127
I-75 ditch . . 127
Mean values for diversity indices ....
Order of entry of variables into diversity
models . .. .
. . 143
LIST OF FIGURES
Figure 1. Generalized species accumulation curve . .. 34
Figure 2. Preston's species abundance curve . ... 34
Figure 3. Species accumulation curve, Camps Canal .. 130
Figure 4. Species accumulation curve, Lake Alice . 131
Figure 5. Species accumulation curve, I-75 ditch . .. 132
Figure 6. Species abundance curve, Camps Canal . .. 136
Figure 7. Species abundance curve, Lake Alice . . 136
Figure 8. Species abundance curve, I-75 ditch .. .. 137
Abstract of Dissertation Presented to the Graduate Council
of thle University of Florida in Partial Fulfillment of
the Requiremnents for the Degree of Doctor of Philosophy
SPECIES ABUNDANCE RELATIONSHIPS OF AQUATIC INSECTS
IN MODNOTYPIC WAtTERR1YACINTH~ COMMUNITIES IN FLORIDA,
WITHI SPECIAL EMPHASIS ON FACTORS AFFECTING DIVERSITY
Joseph Kestutis Balciunas
Chairman: Dale H. Habeck
Mlaj or Department: Entomology and Nematology
Collections of aquatic insects beneath mionotypic water-
hyacinth communities, initially standardized by collecting
effort, later based on a standard sampling area, were made at
37 different sites in 18 Florida counties. Identifications,
verified by authorities for respective groups of the 5485
specimens collected, indicated 147 species of aquatic insects
were present. Comparison of this species list with those from
2 other studies more limited in scope indicated several mis-
identifications by previous workers and a much greater range
of aquatic entomofauna. Although the low level or absence
of dissolved oxygen (DO) has been frequently reported, DO-breathing
forms were abundant and frequent in my collections.
The importance of the relative abundances of each species
at the collection sites was demonstrated by discriminant
analysis which showed that the 3 repetitively sampled sites
were significantly different bansed on the abundances of only
the 10 most frequently collected insects. Species abundances
were also used in 2 methods of estimating the total number of
species present at each of the 3 study sites. Statistical fit-
ting of an exponential species accumulation curve revealed
that approximately all species present at each site had been
collected. The fitting of a lognormal distribution curve to
the plot of the species abundance data indicated that approxi-
mately 70% of the total species present at each site had been
Reducing the species abundance distribution for each
collection to a single statistic often helps in elucidating
the effects of plant morphometric, water quality and other
parameters. Diversity indices, especially those which combine
species richness and species evenness, are the most common method
of reducing the species distribution data to one statistic,
and 3 different diversity indices were calculated for all my
collections. All 3 indices indicated an increase in diver-
sity with decreasing values of alkalinity or some parameters
strongly correlated with it. Higher levels of iron in the
water increased the diversity at least for the Camps Canal
study site. However, none of the diversity indices were
able to distinguish between the 3 study sites, indicating a
loss of informlation due to the reduction to a single statis-
tic. TLhe use of diversity indices would thus appear to have
more limitations than might b~e inferred from their popular-
ity in ecological literature.
Although waterhyacinths cover an estimated 200,000 acres
of water in Florida, the aquatic entomofauna beneath them has
scarcely been studied. Some researchers believe that the
water beneath waterbyacinth mats is relatively devoid of life
due to a reduction of dissolved oxygen. Of the two previous
studies of aquatic insects beneath waterhvacinths, one was
restricted to canals in South Florida, the other to a single
reservoir in Central Florida; neither identified the aquatic
insects to species level. By collecting and identifying
aquatic insects beneath waterhyacinth communities at 37 loca-
tions in 18 Florida counties, I hoped to provide a better
picture of the variety and extent of the aquatic life in
this common Florida habital-.
MIany studies of a particular habitat are surveys, re-
porting only the presence of a species. By standardizing my
collecting methods, I hoped to determine if the abundance of
individuals in each species was also important. By taking
measurements of the waterhyacinths and a number of water
quality parameters, I hoped to elucidate the environmental
factors affectinge species composition and abundance. Repeti-
tive sampling of 3 study sites also would allow detection of
seasonal changes in the co~mposition of the aquatic entomofauna
and determination of whether the abundances of different
species could be used to differentiate between the sites.
The utility of species abundances for determining the
estimated number of species present at a sampling site was
demonstrated by 2 different methods. By comparing the re-
sults of the estimated total number of species, as calculated
by fitting an exponential species accumulation curve, with
the estimate obtained by fitting a loginormal distribution
curve to the species abundance plot, I gained a knowledge of
these techniques and their relative value.
Several diversity indices which reduce the number of
species and their relative distributions to a single statis-
tic were calculated. Their utility in describing a collec-
tion could then be tested, as could the effects of environ-
mental factors on these indices.
The use of various multivariate statistical techniques
and numerical ecological methods not only illustrated their
utility for ecological research but also helped define the
important parameters and their interactions in these water-
hyacinth cormmunities and the methods which could be used to
Eichhornia crassipes (Martius) Solms-Laubach
The waterbyacinth, Eichihornia crassipes, is a widespread
aquatic weed. The most recent botanical treatment of this
species is probably by Agostini (1974, p. 305), whose descrip-
tion of the species, as translated by Center (1976, p. 5), is:
Plants floating or sometimes fixed to the
substrate, the leaves in the form of a
rosette with the stem reduced and the plants
connected by an elongated horizontal rhizome;
numerous plumose roots issue from each plant.
The aerial leaves are variable in shape;
petioles of 2 to 30 cm long are more or less
inflated; stipules 2-15 cmi long with a small
apical orbicular-reniform lamina with a
lacerate [serrate?] margin; submcered leaves
never evident. Inflorescence variable,
internodes between the spathes nearly ab-
sent; inferior spathe within lamina 1-5 cm
long, thle sheath 3.5-7 cm long. Flower 4-6
em long; perianth light purple or rarely
white, tube 1.5-2.0 em long, lobes 2.5-4.5
cm long, with entire margins. Stamens all
exserted, filaments villous-gladular. Cap-
sule elliptical, trigonous, 12-15 mm long;
seeds oblong-elliptical 1.2-1.5 x 0.6-0.6
mnm with 10 longitudinal ridges.
Eichhornia is one of 9 genera, all aquatic, of the
pickerelweed family, Pontederiaceae, all but 2 of which are
considered endemic to the NowJ Uorld (Cook et al. 1974). The
waterbyacint~h is the most widespread of the 7 species of
this mainly neotropical genus. E. crassipes is the only
member of the genus found in the United States, although
there are reports of E. azurea (Sw.) Kunth in South Flori-
ida (Burkhalter 1974).
Waterbyacinth first received the attention of European
taxonomists during the beginning of the nineteenth century,
and Bock (1966) has an excellent historical review of the
taxonomy of this species. Center (1976, p. 3) cites Agostini
(1974) for the following synonomy for Eichhornia crassipes:
Eichhornia crassipes (MTart.) Salms in DC., Monogr.
Pontederia8 cassipes Mart., Nov. Gen. 1:9.
Parops crass pes (Mart.) Raf., Fl. Tell. 2:
Eichhornia speciosa Kunth, Enum. Pl. 4: 131.
Eichhornia cordifolia Gandog., Bull. Soc.
Bot. France TREETT 1920.
The scientific synonyms E. spcis and Piaropus cras-
sipes are especially common in the literature even 50 years
after Solms-Laubach's revision.
WJhile the binomial Eichhornia crassipes has gained wide-
spread acceptance in the last 30 or 40 years, the authorship
of the name is often confused, sometimes being mistakenly
attributed to Kunth. The correct form for the scientific
name with the authors abbreviated is "Eichhornia crassipes
E. crassies_ has a variety of common names in different
parts of the world, with Bock (1966) listing 48 names from
18 countries. In the English language the name waterhyacinth
is commonly used. The structure of the name varies, however,
sometimes being written as one word (waterhyacinth), as a
hyphenated word (water-hyacinth), or most frequently as two
words (water hyacinth). Although most modern authoritative
botanical works (e.g. Cook et al. 1974, Muenscher 1967,
Stodola 1967) use the two-word form "water hyacinth," I will
use the single-word form "waterhyacinth," as recommended in
Kelsey and Dayton's (1942) list of standardized plant names
and in the Composite List of Weeds by the Subcommittee on
Standardization of Common and Botanical Names of Weeds (Anon-
Waterhyacinth, now generally believed to be a native of
South America, is widely distributed throughout the tropical,
subtropical and, occasionally, temperate regions of the world
(Holm et al. 1969, Bock 1966, Center 1976). Although some
earlier workers consider waterhyacinth a native of Florida
(Buckman 1930, Small 1933, Muenscher 1967), most agree that
it was introduced. Goin (1943) cites a University of Florida
botany professor with crediting the U. S. introduction of
w~aterhyacinth to the Venezuelan delegation to the 1835 Cen-
tennial Exposition, with waterhyacinth subsequently being
introduced about 1840 into Florida by cattle-growers. There
is an amazing reference (Gowanloch 1944) that waterhyacinth
was introduced to South Almerica from Japan. This report is
probably due more to wartime emotionalism than scientific
fact. Penfound and Earle (1948) mention that it may have
been cultivated as a greenhouse plant as early as shortly
after the Civil War. As Bock (1966) points out, it is dif-
ficult to believe that a species as large and showy as water-
hyacinth was overlooked until the late 1880's by all the
early botanists in the state. M~ost authorities agree that,
directly or indirectly, the waterbyacinths in Florida came
from those brought from Venezuela by the Japanese delegation
for distribution as souvenirs at the 1884 International
Cotton Exhibition (sometimes referred to as the 1884 Cotton
Centennial Exhibition) in New Orleans (Klorer 1909, Buckman
1930, Penfound and Earle 1948, Tabita and wJoods 1962).
Although the precise time and area where waterhyacinths
were introduced into Florida is not surely known, the earli-
ast reports place it in the St. Johns River near Palatka in
1890. A New York newspaper account (Anonyvmous 1896) quotes
a N~r. J. E. Lucas of Palatkca that th~e waterhvacinths were
introduced by a M~r. Fuller [in 18911 into the St. Johns
River- seven miles north of Falatka at EdgewJater Grove from
plants brought originally from Europe. Webber (1897) places
the point of introduction ". . about 1890, at Edgiewater,
shout four miles north of Palatka" (p.11), while Tilghman
(1962), an old resident of Falacka, states "Florida's first
water byacinthl was placedl in the St. Johns River by a winter
visitor, Mlrs. W. F. Fuller, at San Ilateo, five miles south of
Falatka" (p7. 8). Most subsequent workers (Buckman 1930, Penfound
and Earle 1948, Tabita and Woods 1962, Seabrook 1962, Zeiger
1962, Raynes 1964, and many others) give similar versions of
thle introduction based directly or indirectly on the above
In 1897 waterbyacinth distribution in Florida was thought
to be confined to the St. Johns River and its tributaries and
a f~ew landlocked lakes (Webber 1897). In 1930 waterhyacinths
seemed still to be confined to the 42-square-mile (26,800
acres) St. Johns River drainage (Buckman 1930). By 1947 the
estimated area of infestation increased to 63,000 acres and
to approximately 80,000 acres by 1962 (Tabita and Woods 1962).
In 1964 an estimated 90,000 acres were infested (Ingersoll
1964), while by 1972 the estimated area of infestation had
increased to 200,000 acres (Perkins 1973). The 1975 estimate
is also 200,000 acres (Center 1976).
P'ro ductivity and Regraduction"
This increase in the area infested, despite massive con-
trol measures dating backed over 75 years, points out the re-
markanble reproductive ability of this plant. Considered
one of the world's most productive photosynthetic organisms
(WJestlake 1963), waterbyacinth-s under optimal conditions
may~ double in number in 11 to 18 days (Penfound and Earle
1948). Center (1976) presents a table on waterhyacinth pro-
ductivity an-d SLanding crop comp~iledl from many different
Such rapid increase in new plants is due primarily to
vegetative r-eprodluction (Hitchcock et al. 1950), with one
plant producing many offsets, or suckers, on stolons (Pen-
found and Earle 1948). New plants are also regenerated
from broken portions of the rhizome (ibid.).
Unlike California and some other parts of the world
(Bocke 1966), in Florida reproduction from seed definitely
occurs, with about 5% germinating under normal conditions
(Zeiger 1962). Seeds are considered important in Florida
chiefly as propagules for infesting new areas or for rein-
festing areas where waterhyacinths had been controlled.
Although it is frequently believed that the absence of
waterbyacinths in a suitable body of water is due to the lack
of introduction of a propagule (Penfound and Earle 1948),
this is probably an over-simplification As Morris (1974)
points out, an invading organism's success is due not only
to favorable physical factors such as nutrients, light and
temperature, but also to the competitive ability of plants
already present in the area. He demonstrated that waterhy-
acinths would not become established when introduced to an area
with abundant native vegetation. However, man frequently
alters these natural, balanced, aquatic systems, sometimes
by increasing the nutrient level. A good colonizing species
lilke w~aterhyacinth can easily become established and out-
compete the native flora in such disturbed aquatic systems.
L~ight. Waterbyacinths require reasonably high light
intensities for growth, at least 60% full sunlight according
to Bock (1966). Penfound and Earle (1948) found that the
light intensity in July above a waterhyacinth mat was 420
footcandles and noted that plants at 130 footcandles were
dying. Knipling, et al. (1970) foundl that photosynthesis
increased fromm 7.8 mg CO2/dm" leaf surface/hr to 16.1 mg/
dm2/hr when light intensity increased from 1450 to 8000
Air temperature. W~aterbyacinths can easily survive
freezing temperatures for short periods of times, Webber
(1897), Buck (1930) and many subsequent workers observed re-
growth of shoots from the submerged rhizome after the tops
of the plants had been killed by frost. Survival at 21oF
for 12 hrs has been recorded, with lower temperatures killing
the rhizome and preventing regrowth (Penfound and Earle 1948).
I could find no literature on optimum or maximum air tempera-
tures for waterhyacinth; however, it is undoubtedly fairly
high,, as most prolific growLh usually occurs during summer
when daytime temperatures are frequently in the uppier 80's
and low 90's oF. Balciunas (unpublished data) observed lux-
uriant growth in a greenhouse where daytime summer tempera-
ture wJas consistently about 1000,.
later temperature. Knipling et al. (1970) determined
that the optimum water temperature for waterhyacinth was
28-30"C (82.4-86.00F) but that growth was relatively high
over the range of 22-350C (71.6-95.0"F'). Waterhyacinths
w~ill survive a water temperature of 34"C (93.2oF) for 4 or
5 weeks (Penfouind anid Earle 1948), but higher temperatures
are detrimental, with negative growth occurring at 40"C (1040F)
Wat-er dyth.l' Waterb!yacinths can g:rowd on land; P'enfound
and Earle (1948) noted survival of plants for up to 18 days
out of water. A high soil moisture content seems necessary
for prolonged survival (Webber 1897, Bock 1966).
There seems to be no good correlation between increasing
water depth and waterbyacinth growth (Mlorris 1974).
Hydrogen ion concentration. w~aterbyacinth seems tolerant
to pH values normally encountered in aquatic systems. Haller
and Suitton (1973) found that optimal growth occurred in acid
to slightly alkaline conditions (pH 4-8) and some growth oc-
curred from pH 8.0 to 10.0. Bock (1966), citing various
sources, gives a pH range of 4 to 9. Ponfound and Earle (1948)
found that the pH of water beneath waterbyacinths usually
ranged from 6.2-6.8 but th~at wvaterhyacinths could tolerate
extremes of 4-5 and 9-10. Chadwick and Obeid (1966), in com-
paring growth of waterhyacinths and wjatertectuce (Pistia
stratiotes L.), found that optimal growth for waterhyacinths
occurred at a pH of 7.0 while 4.0 was optimal for waterlettuce.
Center and Balciunas (1975), comparing water quality at vari-
ous locations having waterbyacinths, alligator weed (Alter-
nanlthera ILhiloxeroides (Mart.) Griseb.), or neither weed,
found there was little difference in the pHI preferences of
the plants and that the pHI of areas with waterbyacinth was
slightly lower (7.06 + 0.84t), though not significantly, than
the pH of areas having no arlu~atic weeds (7.55 + 1.06).
NTutrients. Dymond (1948a) and Hlitchcock et al. (1949)
found that wYaterhyacinths grow well in nutrient-poor as well
as in nutrient-rich water but that added nutrients favor
growth. Haller et al. (1970) found that less than 0.01 ppm
phosphorus was limiting to waterbyacinth growth and that above
this level phosphorus was absorbed in luxury amounts. Haller
and Sutton (1973) found that maximum growJth occurred in water
with a phosphorus concentration of 20 ppm and that levels
greater than 40 ppm were toxic. Boyd and Scarsbrooke (1975)
added fertilizer at 4 different levels to waterbyacinth ponds
and found that the lowest waterhyacinth biomass yield was
from unfertilized ponds while the highest yields came from
ponds with the intermediate level of fertilization.
Wahlquist (1972) compared yields of waterhyacinths grown
with no fertilizer, with fertilizer containing phosphorus,
and with fertilizer containing both nitrogen and phosphorus.
He found that fertilized ponds had much higher yields (550.4
and 590.9 metric tons/ha) than unfertilized ponds (174.5 metric
tons/ha) and that ponds fertilized with nitrogen and phos-
phorus had a slightly higher (but statistically insignificant)
yield than ponds fertilized with phosphorus only.
Salinity. Although plants can survive up to 13 days in
100i: seawater (Bocke 1966), waterbyacinth~ is intolerant to salt
water, with Buckman (1930) listing a survival time of only
24 brs. Penfound and Earle (1948) found that waterhyacinths
did not tolerate more than slightly brackish water and were
not found in lakes or streams with an average salinity greater
than 15% seawater.
Alkalinity. Center and Balciunas (1975) found the alka-
linity of water containing waterhyacinths or alligatorweed was
higher than that: of water without either species.
Metallic ions. Sutton and Blackburn (1971a, 1971b)
found that 3.5 ppm copper for 2 weeks inhibited waterbyacinth
Center and Balciunas (1975) found waterbyacinths more
tolerant of low fron levels than alligatorweed. Morris (1974)
found no correlation between waterbyacinth growth and the
levels of copper and iron at his study sites.
Problems. The explosive growth of waterbyacinth has
caused it to be ranked as one of the 10O most important weeds
and the most important aquatic weed (Holm et al. 1969). The
massive amount of literature on the problems caused by water-
hyacinth is well reviewed by Del Fosse (1975) and Center
(1976). A list of the maini categories of problems is:
(1) Interference with navigation;
(2) Clogging of water drains, irrigation canals, spray
equipment and pumps;
(3) Interference with fishing, swimming and other aqua-
tic recreational activities;
(4) Oxygen depletion caused by heavy infestations,
making water inhospitable to many aquatic organisms;
(5) Increased evapotranspiration rates in an infested
area (1.5 to 5 times higher than evapotranspiration from ad-
jacent open water);
(6) Reduction of fish populations by destruction of
spawning beds, competition for nutrients and space, depletion
of dissolved oxygen, and by preventing predators from finding
(7) Creation of deep beds of organic sediment;
(8) Reduction of open water available to waterfowl;
(9) Creation of ideal breeding places for certain mos-
quitoes, some of them disease vectors;
(10) Increased flooding due to obstruction of waterways;
(11) Occasional destruction of bridges, trestles and
other structures during flooding;
(12) Shading out and otherwise out-competing beneficial
(13) Hionetary and ecological costs of control;
(14) Rendering unsightly and aesthetically unpleasant
the water surfaces which they completely cover.
Control. The U. S. Army Corps of Engineers estimates
that a total of $76 million was allocated for aquatic weed
control in Florida during fiscal year 1976, with almost
$5 million being allocated Eor waterbyacinth control (Center
1976). Mlorris (1974) cites a USDA source for a $12-$16 mil-
lion estimate of the costs of waterhvacinth control in Flor-
ida in 1973.
The literature on waterbyacinth control is enormous.
There is even a Hyacinth Control Journal (renamed in 1975
the A~quatic Weed Management Journal) which- began publication
in 1962. Pieterse (1974) provides a good review of the most
important literature on waterhyacinth control. Del Fosse
(1975) has a more detailed review of the various aspects of
There are also good, recent review articles for each
particular aspect of waterhyacinth control: methods of
mechanical control have been reviewed by Robson (1974); chem-
ical control and the various compounds available were re-
viewed by Blackburn (1974); biological control of aquatic
weeds has been reviewed by Bennett (1974) and by Andres and
Bennett (1975). The use of plant pathogens was reviewed by
Zettler and Freeman (1972), Freeman et al. (1974) and by
Charudatt~an (1975). Mlitchell (1974) reviewed habitat manage-
ment as a means of aquatic weed control.
Utilization. Possible beneficial uses of waterhyacinths
have been a concern of even the earliest reports (Webber 1897,
Buckman 1930, Penfound and Earle 1948). Bock (1966), Pieterse
(1974), Del Fosse (1975), and Center (1976) all review the
abundant literature regarding the beneficial aspects of water-
h-yacinths, th-e main ones of which are:
(1) Removal of nutrients from water, including use of
w~aterbyacinths in sewage treatment;
(2) Protection of shorelines from erosion;
(3) Use as mulch and fertilizer;
(4) Use as a source of production of natural gas;
(5) Increase in aquatic organisms utilizable as fish
(6) Use as fodder for cattle, pigs, catfish or other
(7) Shading out of nuisance submerged plants like
(8) Decrease in breeding habitat for certain mosquito
species (Barber and Hayne 1925);
(9) Use in construction of a large variety of objects,
e.g., chair bottoms, cigar wrappers, ice chests, paper;
(10) Aesthetic appeal of beautiful blossoms and luxuri-
ant green foliage.
Habitat for Aquatic Insects
The insects found beneath waterbyacinths have received
little attention. Most of the literature about insects assoc-
iated with waterhyacinths deals with those which might have
potential as control agents through their feeding activity
or transmission of pathogens. Fred Bennett of the Common-
wealth Institute for Biological Control in Trinidad has
published many papers on insects and mites on waterhyacinth
and their possible use as biocontrol agents (Bennett 1967,
1968a, 1968b, 1970, 1972, Bennett and Zwolfer 1968). Others
who have providedl lists of insects attacking waterhyacinth
are Gordon and Coulson (1969), Coulson (1971), Perkins (1972,
1974), and Spencer (1973, 1974). None of these investigators
mentions any aquatic insects except the weevils, Neochetina
spp., which build their pupal case in the root hairs of water-
hyacinths (DeLoach 1975).
Waterbyacinths' extensive root systems, sometimes over a
meter in length, create a vast new habitat for a variety of
aquatic insects where previously a few larger, predaceous,
open-water forms dominated. Weber (1950) estimates the area
of the roots of one small waterhyacinth is 7.31 square meters.
O'Hlara (1968) believes that wYaterbyacinth has a greater inter-
face area than any other aquatic plant.
Although Goin (1943) thought that the root length of
waterhyacinth plants was dependent on the depth of the water
beneath them, it has been demonstrated that root length is
a function of the nutrient content of the water: the longest
roots occur in nutrient-poor waters, the shortest roots in
nutrient-rich water (W~akefield and Beck 1962, Haller and Sut-
ton 1973, M~orris 1974). The blue color frequently seen in the
roots, often cited as a diagnostic aid, is also due to water
quality. Plants grownI in phosphorus-deficient water have
iridescent blue roots, indicative of anthocyanin production
and phosphorus deficiency, while roots of plants grown in
nutritive waters are a normal, gray-black color (Haller and
While fishermen (Tilghman 1962) frequently praise the
waterhyacinth for producing an abundance of aquatic insects
and other organisms desirable as fish food, most investiga-
tors disagree. They believe the apparent abundance of
aquatic organisms occurs only on~ the edge of a mat, with the
water under the center of the mat being ". . unsuitable for
the existence of most forms of plankton and aquatic insect
life." (Lynch et al. 1947, p. 64). Many agree, citing low
levels of dissolved oxygen beneath a solid mat (Lynch 1947,
Ultsch 1971, 1973). Wahlquist (1969) believes the reduction
of fish yield in waterhyacinth-infested ponds is due to
shading out of phytoplankton by waterbyacinths.
Few workers have actually surveyed the aquatic entomo-
fauna of waterbyIacinths. Goin (1943) surveyed the lower ver-
tebrate forms of waterbyacinth communities but he mentions
only one insect, the midge Chironomus. O'Hara (1961, 1968)
surveyed the invertebrate fauna of waterhyacinth-covered
canals in South Florida and lists the insects collected, most
of which are identified only to family. Katz (1967), in her
study of the effects of chemical eradication of waterhyacinths
on the associated aquatic fauna, provides lists of insects
on and below waterhyacinths. Most of the insects are identi-
fied to the generic level, but I suspect there are many
erroneous identifications. Hansen et al. (1971), in studying
the food chains of aquatic organisms beneath waterbyacinths,
mentions several different aquatic insects. Lynch et al.
(1947) collected a relatively small number of aquatic insects
from beneath waterhyacinths, most of which were identified
to family or order. Wahlquist (1969) also mecntions some in-
sects, identified only to Eamil~y or order, which serve as
food for fishes living in wzaterbyacinth-covered ponds.
The only group of aquatic insects whose relationship
to waterhyacinth has been investigated more thoroughly is
the mosquitoes (Culicidae). Barber and Hayne (1925) list
4 species of Anopheles collected among waterhyacinths, A.
crucians being the most common. Seabrook (1962), in his
study of correlation of mosquito breeding to waterhyacinths,
mentions 2 species of Aniopheles and all 3 species of Mansonia
as being associated with waterhyacinths. He, along with
Barber and Hayne (1925) and Lynch et al. (1947), believes
that waterbyacinths increase mosquito production. However,
Viosca (1924; cited by Barber and Hayne 1925), Mulrennan
(1962) and Ferguson (1968) believe that waterbyacinths re-
duce mosquito production by shading out planktonic food and
the submerged plants which serve as refuges.
The only other references I found to aquatic insects
associated with waterbyacinths are casual references by ento-
mologists to the habitat where a certain species was col-
lected, e.g., Blatchley (1914, 1925), Young (1954).
Aquatic insects were collected 88 times in 37 different
wzaterbyacinth communities in 18 Florida counties. In addi-
tion, 3 study sites in Alachua County (Camps Canal, Lake
Alice, and a drainage ditch at Interstate 75) were sampled
repeatedly in order to ascertain seasonal and other temporal
changes. Twenty-nine collections were made at Camps Canal,
including regular samples at 2- to 3-week intervals during
all of 1974. The first collection was made 12 Aug. 1972,
the last on 5 Dec. 1974. Table 1 presents a list of collec-
tion sites and dates.
Sample Site Selection
The exact sampling point at a given collection site was
selected on the basis of several considerations. An area
containing an essentially pure stand of waterhyacinths was
chosen in order to eliminate complicating effects which
other aquatic or successional plants might have on the biota
beneath a waterhyacinth mat. The w\aterhyacinth communities
sampled were essentially monotypic, although almost all con-
tained small amounts of duckweed, Lemna and Spirodella spp.
and/or some water fern, Salvinia and Azolla spp. Watermeal,
Wolfia spp. and Wolfiella spp., was sometimes present mixed
with the Lemna. These floating macrophytes were seldom
4" lliII Ill 17 ii
11 I2~ r I n ~I
Ill L. I 1'1 Ill
U: "' ~ I~ L)
I rl ~ r.r.l
i~' Ini I ;,,~~~..,,
LniF I r* i r
nr ,;I ~~r I err
C1- ~ (r-.\(
Llr . (;: i
L1r 41 1
I.l"r~ rc- r~
~~I I~. i ~
Llh~ ii I-~
CL'I- ~, r q:.i
In'Ii '. I-r
(:IV(i : Tn~.ni
Clnll rl-r ~u
II- I rd).II
r-4CI I lyi i
i I*II I- iii !~
ri. ,i...~r vllii.ni;iiii
nr r. 111.
"1"I 'Ir I\
:''; ';" i,,
~ ~ ~ ~,
I~I!- .~~~.~ ,.r Ei:-I~i
~' i'. ''~ i-. rl.r ~, ~ r ir.
!''`' .. r.~i .;II
'" I-r..*r. r. r I1a.1r. ri
"1'Cl' ~ir.; 5i:~I il , Iri
J~rv: 4 I I ~i' ;. II L I 9 i. \ i I
I I: i- ':1:I. 1( ~;IIII1 r,
"~ '' ~:~. -~ ~, ~~.I~ r:, r,
"~ i;. "~ `lb nC III.I L b-ll)n -I ~
'* ' r)'. "n I;1I1. ~, Ellll: F... II
I" IV II
`'' L I~1 I~I,I I-; LI? I i I ~I~ I~;i
rri; ii Lil. ?I' ~L. ~
,~ 'r. ,1. , I rr~ . ~ ~I~; :I ..
::: ;1:iD; u(, - :, , ,,..1,~ .Lniil.ll Cn.
r -; ~I
1L ni r ~ :
L~' :1 17 1'1 :''''L T.I:illlr r-1
1--I 1i UI "1 "IFi~i`' 1 blilliUZ
i~ C1''9'1~1";: \I.*L.i
"'1P' r I ?I II C~
CI- -. -.~ II C ;-ilil. -I ~-ln C1I.
(llr yr 'Ld.. .r:. r'll' ln~,I~
"r ci ;, "':1~L' 1 "i(i Ilr
7ii I "' '-I (I '* Ti~
I- 1~. In
" "I II (p;l-. (( i[, I'
i i 'I : I I I 1 i 1' I I I
Il~r. I "i - '-"i i'. L.r ( I: ~I:
II ill I~I *i
Ill. 1 I irl~~ii I 1 111 11-~
r II 1~
'' ' '7-
71 Ilii "
(i JIII ~1
h JII( Ir
II J"I "
11 J1II "'
11 .1'11 II
Il rr I,
7- 'Irn ,1
7' rlCr II
17 ii r
1' I'CT I~
ii Inli ~n
~5 Fit. I-,
I- F"i )~
Is rr 7n
o r r 7a
IT unr rr
I~' JII -)
7'r JIIi -1
1) Jllr -4
7 411h I*
I~ 4LI; )I
11 1111 '74
~' il iO
~I Liiy II
2* 'L1: -1
I" i'~- "
I)Y I :d
~I -i"Y -d
II 'I `r
~I Y 1
Table 1. List of collection sites
abundant except in a few places where waterhyacinth growth
was poor and the waterbyacinth plants widely separated from
each other. Emergent, rooted aquatic vegetation such as
cattails, Typhus sp., and grasses was sometimes present along
the shoreline, but submerged aquatic plants were almost never
present beneath the waterbyacinth mat, although they were
sometimes within the sampling area.
Within a mat a sampling point which was some distance
(2-20 meters) from the shoreline was usually chosen in order
to eliminate ecotone and emergent vegetation effects. The
sampling point was reached by wading if water depth and bot-
tom configuration permitted; otherwise, 2 Styrofoam billets
allowed me to reach and sample the interior of the mat by
alternately standing on one billet and moving th~e other.
Accessability was another consideration. Since a con-
siderable amount of equipment, e~g. Hach test kit, Styrofoam
billets, sampler, collecting and recording equipment, had to
be transported, suitable sampling areas near roads were usu-
The collections made in 1972 (#1 #'7) were strictly
qualitative, with no attempt at standardization.
Standardized collection time. The 1973 collections,
along with those from the early part of 1974 (#11 #58),
were made on a standard-time basis, i.e., a waterbyacinth
mat was sampled for approximately 1 hour, usually divided
into 2 half-hour subsanpl~es.
If the waterbyacinth mat was not too thick, a triangular
dip net was placed well under the plants and a clump of water-
hyacinths was lifted inside the net out of the water. The
contents of the net and the waterbyacinths were washed in a
pail of water, then small portions of wrater from the pail
were poured into a white enamel pan and searched for insects,
which were placed into a vial of 70"% alcohol. Each vial had
a collection number placed inside and also written on the
stopper. Once all the water in the pail had been searched
for insects, the entire procedure was repeated until the
time period elapsed.
The entire contents of the pail were searched even if
the time period ended before all the water had been examined.
Actual time spent collecting the insects, as wjell as number
of plants and petioles searched, their height above the water,
root length and water depth were recorded. Offsets were con-
sidered plants if a root system had developed; petioles were
counted only if they had a portion of a green leaf.
If a mat was too thick, i.e., the numerous stolons and
petioles prevented passage of the dip net, then a clump of
waterhyacinths would be quickly pulled out of the water into
the pail, the roots washed and the water examined as previously
outlined. This standardized collection time method or "catch
per unit effort" is a valid way of estimating insect popula-
tion density (Morris 1960), however, preliminary analysis of
my data showed that extension of these type data could lead
to spurious conclusions.
Standardized area. A~ sampler which would cut through
a waterbyacinth mat and capture the organisms underneath was
designed and constructed. It consisted of a rectangular
aluminum box, 80 x 40 x 50 cm, open at both ends, reinforced
with an external frame and with strips of sharpened stainless
steel attached to the bottom edge. The sampler, when brought
down vigorously in a vertical position on top of a waterbya-
cinth mat, cut a 1/5-square-meter (0.2 m2) section out of the
mat. A 10-cm high aluminum drawer with a double mesh bottom
was then quickly inserted into a slot 15 cm from the bottom
of the sampler, thus enclosing the waterbyacinth mat sample
and aquatic organisms beneath it in the sampler.
The sampler was then lifted out of the water, the water
was drained through the screen drawer, and the drawer was
searched for aquatic organisms, washed and searched again.
Roots of waterbyacinths trapped inside the sampler were washed
vigorously in a bucket of water and the water was searched
for aquatic organisms. Although some active organisms might
have eluded capture by swimming straight dlown before the
drawer could be inserted in the sampler, the presence of
numerous fish, crayfish and other active swimmers inside the
sampler indicated that it was reasonably effective in cap-
turing the organisms beneath the waterbyacinth mat. As with
the standard-time sampling method, plant height, root length
and water depth as well as number of plants and petioles in-
side the sampler were recorded. Water temperature at a depth
of approximately 15 cm was also recorded.
Water quality data for each collection site were taken
in the following manner. Three water samples were drawn
from a depth of about 15 cm in pint-sized Whirl-pakR plastic
bags labelled with the collection number. Total alkalinity,
total hardness, pH and specific conductance of one sample
were tested immediately, using a portable H~ach DR-EL/2 test
kit. At the laboratory, the second sample was tested with
the Hach kit for chlorides, total nitrates and nitrites,
total phosphates, and sulphates. The third water sample was
refrigerated until delivered to the University of Florida
Soils Laboratory, which conducted a variety of tests, of
which those for copper, potassium and iron were the most im-
The methods applied to the water samples were:
chloride--titration, mercuric nitrate;
cop~per--atomic absorption spectrophotometer;
hardness (total)--titration, Titra Ver;
iron--atomic absorption spectrophotometer;
nitrates and nitrites (total)--cadmium reduction;
pH--calorimetric, wide-range ;
phosphates--ascorbic acid, P'hos Ver III;
potassium--flame emission spectrophotometer;
specific conductance--direct measurement, conduc-
sulphlates--turbidimetric, Sulfa Ver IV.
Identification, labelling and cataloging of the thou-
sands of specimens collected took place in the laboratory.
Of the organisms collected, only aquatic insects were iden-
tified. Aquatic insects were defined as those having at
least one life stage which spent time in or on the water.
This definition excluded insects living inside the water-
hyacinth plant and those found on the emergent portion of
I identified the specimens using taxonomic keys and
referring to identified specimens at the Florida State Col-
lection of Arthropods and to the original species description
and other taxonomic papers. Although most groups lacked keys
to nymphal stages, I was able to identify the nymphs of al-
most all groups after examining many specimens and construct-
ing life series.
Since there are few keys for larval forms (especially
to species level), I originally attempted to rear many of the
larvae collected. However, the lack of adequate facilities
and time to rear these many aquatic larvae with their diverse
requirements, andi the subsequent loss of some specimens and
poor quality of others caused the termination of this approach.
Subsequently all larvae were preserved in the field along with
the other insects. Some larvae were identified by experts on
I identified the mayfly (Ephemeroptera) nymphs using
Berner's (1950, 1968) keys, and Dr. Lewis Berner of the
University of Florida kindly checked the identifications of
some representative specimlens.
Dragonfly (0donata:An~isoptera) nymphs were identified
using the keys of Wright and Peterson (1944) and Needham and
Westfall (1955). Damselflie~s (0donata:Zygoptera) were
especially difficult to identify since no comprehensive key
to species exists, but by using Walker's (1953) descriptions
and other species descriptions as well as characters provided
by Dr. Mlinter J. Westfall of the University of Florida, and
by constructing life series, I was able to identify all speci-
mens, including very early instars. Dr. Westfall examined
and verified all my odonate specimens.
Aquatic Hemiptera were identified using the keys of
Herring (1950a, 1950b, 1951a, 1951b) for most families.
Velvet water bugs (Hebridae) were identified using Chapman's
(1958) key; water-crawling b~ugs (Naucoridae) were identified
using La Rivers' (1948, 1970) keys and by construction of
life series of nymphal stages. Dr. Jon Herring, USDA Syste-
matic Entomnology Laboratory, confirmed my identifications of
representative specimens of different species of aquatic and
semi-aquatic Hemiptera except for the giant water bugs
(Belostomatidae), which were checked by Dr. Mencke at the
Caddisfly (Trichoptera) larvae were determined using
W~allace's (1968) kev. Dr. Oliver Flin= at the Smithsonian
checked all my caddisfly identifications. Species level
determination was frequently impossible.
Dobsonfly (M~egaloptera: Corydalidae) larvae were iden-
tified with the keys of Chandler (1956) and Cuyler (1958).
Adult water beetles (Coleoptera), except Hydrochidae,
were identified by using Young's keys (1954, 1956, 1963).
Hydrochidae were identified by Dr. John L. Hellman of the
University of Mlaryland. The larvae of aquatic beetles could
not be easily identified, though Leech and Chandler's (1956)
keys in Aquatic Insects of California were of some value in
identifying larvae of some families to the generic level.
Species level determination was possible only for genera
monotypic in Florida. Dr. Paul J. Spangler of the Smithson-
ian Institution graciously identified or confirmed my iden-
tifications of all aquatic beetle larvae except Helodidae.
Helodidae larvae are very poorly known and even the identi-
fication of adults was very difficult. I was successful in
rearing some larvae to adult stage. Dr. Dale Habeck, Univer-
sity of Florida, and I wJere then able to find characters to
discriminate the different genera. Dr. Frank Young at the
University of Indiana checked all of my aquatic beetle adults
except the adult: Hydrophilidae, which were checked by Dr.
Dr. Dale Habecke also identified the aquatic Lepidoptera
Mosquito larvae (Culicidlae) were readily identified to
species using Carpenter and LaCasse (1955). Dr. William Beck
of Florida A & M graciously identified all thousand-odd of
my midge (Chironomidae) larvae specimens. Other families of
Diptera larvae could be identified to the generic level using
keys of Wirth and Stone (1956), while a few (17 specimens) of
Diptera larvae and pupae could be placed only at the family
After all specimens had been identified and recorded,
comparison of the species composition of different collec-
tions was desired. With the number of species in a collec-
tion ranging from 1-31 and the number of specimens in a col-
lection from 3-373, direct comparisons of species composition
of collections was difficult. In ecological studies it is
common to represent the numbers of species and their relative
distribution by a single statistic, an index of species
The choice of diversity indices is enormous, with much
confusion about terms and applicability (Hlurlbert 1971, Peet
1974). Mluch of the confusion results from the dual concepts
which~ most, but by no means all, authorities believe an in-
dext of diversity should embody. Almost all agree that species
richness, the number of species in a sample (sometimes re-
ferred to as species number or species count), should be re-
flected in the species diversity index used. Many research-
ers, especially outside the field of ecology, in fact tend to
equate diversity with species richness, i.e., a collection
has "high diversity" because many species are present. How-
ever, this reliance exclusively on species richness does not
take into account the relative abundances of each species.
T'he species evenness or equitability concept of diversity
stresses these relative abundances, equating high diversity
(more properly, high equitability) with~ an even distribution
of individuals among the species present, while low equit-
ability implies a few abundant species, other species being
relatively rare. MIost authorities in this area agree that
a proper measure of species diversity includes both species
richness and species equitability components (Margalef 1969,
Pielou 1969, Peet 1974). H~urlbert (1971), among others, would
restrict the term diversity to this dual concept.
Because of the preponderance of authoritative opinion
favoring a diversity index with both richness and evenness
components, classed by Peet (1974) as heterogeneity indices,
my investigations were limited to this type.
Peet (1974), in his excellent review article on diversity
indices, lists 4 commonly used heterogeneity indices. E. H.
Simpson proposed a diversity index in 1949 which recognized
the dual concept of diversity and which bears his name. Simp-
son's index, with slight modifications, is extensively used
and is recommended, at least for certain applications, by many
authorities (W~illiams 1964, Whittaker 1965, 1972, Sanders
1968, Pielou 1969). Pielou's (1969) restatement of Simpson's
index as adjusted for finite sample size is frequently used:
D = 1 E ng1
where N is the total number of specimens in the collection,
n. is the number of specimoens in the ith species, and s is
the number of species in the sample.
MIcIntosh (1967) suggested another index:
Mc~ntosh's index, while receiving attention in review arti-
cles on diversity indices (Pielou 1969, Peet 1974), does not
appear to be frequently used. Peet (1974) considers it a
variation of Simpson's index, while Bullock (1971) criticizes
its applicability and sensitivity.
By far the most popular and widely used index is Shannon-
Weaver's diversity index:
H' = -.1l p. Log p ,
where pi is the proportion of the total specimens comprised
by- the ith species of p.=n./N. Based on information theory,
diversity is equated to uncertainty. As Pielou puts it:
Diversity in this connexion means the
degree of uncertainty attached to the
specific identity of any randomly sel-
ected individual. The greater th~e
number of species and the more nearly
equal their proportions, the greater
the uncertainty and hence the diversity.
(1966a, p. 131)
Although the most popular diversity index, the Shannon-
Weaver~ index is also the most widely criticized. Monk (1967)
and Sager and H~asler (1969) believe this index to be insensi-
tive to rare species whereas Peet (1974) suggests it is most
sensitive to rare species. Fager (1972), Whittaker (1972)
and Poole (1974) believe Shannon's index most sensitive to
species of intermediate importance. Pielou (1966b) believes
it is used frequently in situations where it is not appli-
cable and questions the validity of equating uncertainty with
diversity (1969). Both Pielou (1966a, 1966b) and Peet (1974)
agree that the Shannon-Weaver index is applicable only when
the collection is a random sample drawn from an infinitely
large population pool. For finite collections, such as
light-trap collections, Pielou (1966a) suggests the use of
Brillouin's (1960) index:
H = 1/N log (N!/n,!n,!..ns '
Unfortunately, this involves computing very large numbers.
Any integer factorial greater than 69! results in a number
larger than 10'oo, which exceeds the capacity of even large,
modern computers. Thus, direct calculation of this index
for larger collections is laborious unless a simplifying
method such as Sterling's approximation to the factorial is
employed. Howrever, the substitution of the Sterling's approx-
imation for the factorial results in this index becoming
equivalent to the Shannon-Weaver index (Peet 1974).
Hlurlbert believes that ". . the recent literature on
species diversity contains many semantic, conceptual and
technical problems", so many problems, in fact, that he con-
cludes ". . species diversity has become a nonconcept."
(1971, p. 57 ) He would possibly retain the term if the
meaning of species diversity were restricted to those terms
which combine both species richness and species evenness,
I.e., heterogeneity indices. He believes that there are 2
useful indices, one of which is modification of the rarefac-
tion index E(Sn) proposed by Sanders (1968). This index
calculates the expected number of species, E(S ), the sample
would contain if the number of specimens were scaled down,
i.e., rarefied, to some common number which would allow com-
parison with other samples. The scaling, which was done in-
correctly by Sanders (Hurlbert 1971, Fager 1972, Simberloff
1972), is necessary because larger samples would contain more
species than a smaller one even if they were drawn from the
same community. Hurlbert (1971) and Simberloff (1972) pro-
vide similar, correctly "scaled", calculating formulae for
the rarefaction index:
s~~ -1 n."j "
E(S ) = [ /
WThile this results in a scaled estimate of species richness,
and Feet (1974) classifies this as a richness index, I be-
lieve it can be classified under the heterogeneity indices
since the species composition, i.e. the evenness, was used in
I chose the following species diversity indices:
Shannon-Wveaver index, HI', since it is the most commonly used
index in recent ecological literature; Simpson's index, D,
since it overcomes some of the sh-ortcomings of H'; and the
rarefaction diversity index, E(S _), because of its inherent
rationality and ease of interpretation. I wrote a small
Fortran program, shown in Appendix A, which calculates
Shannon's H', Simpson's D, and also Brillouin's H when the
collection is small enough. For these diversity indices,
the choice of the base for the logarithm is left up to the
researcher. Since no particular base seems to have become
standardized, I chose to use natural logarithms, although
the base 2 and base 10 logarithms also are frequently used.
For calculating E(Sn) I modified slightly a program cited
in Simberloff (1972) and Heck et al. (1975) and provided by
Dr. Heck of Florida State University. As a check on the
sensitivity of these 3 indices, I also used a fourth, ex-
tremely simple index, the number of specimens per species,
in all my analyses of diversity.
Efficiency of sampling method and determination of the
total number of species in the community sampled are usually
unanswered questions in ecological studies. A relatively
crude, frequently employed estimate of sample effort (Wilhm
1972, Heck et al. 1975) is the graphical plotting of a species
accumulation curve. The cumulative number of new species is
plotted versus the cumulative number of specimens, with each
collection being added sequentially, hopefully resulting in
a curve such as that shown in Figure 1.
The resultant points initially form a straight line,
as each collection adds a relatively constant increment of
species per specimens collected. However, with increased
sampling only rare species remain, and the line rapidly
curves to become asymptotic at the value of the total number
of species in the community and produces the typical exponential
Figure i. Generalized species accumulation curve.
1 (ICIS~~11 Per 5venesr
) ii II
Figure 2. Lognormal species abundance curve
(from Preston, 19148)
sp~ecies accumulation curve. Unfortunately, this asymptote is
frequently attained only when the total number of individuals
sampled is very large, sometim-es only when nearly every in-
dividual in the community has been collected and identified.
In general practice, most researchers use this method to deter-
mine if future sampling will be worthwhile, i.e., whether the
crest or some other previously determined point on a species
accumulation curve has been reached. This is usually done
through simple visual inspection of a graph. The asymptote
level, i.e. total number of species, is not usually calculated.
For my data from my 3 repetitively sampled study sites, I
refined this procedure. A curve such as the one shown in
Fig. 1 can be represented by the general exponential equation:
y = a(1 e-bn)
where a is the asymptote for the curve. Rew~ritten in terms
of thie species notation previously used, this equation be-
s = S* (1 e-bn)
The estimated total number of species in the collecting area,
S*, and the constant, b,, can be determined from a curve fitted
to theP data. Although the computer program SAS PROC NLIN
will fit this general curve to the data, it requires estimates
of both S* and b. While a very general approximation of the
value for S*' can be obtained fromn inspection of the data, the
approximate value for b is not readily apparent. However, I
was able to devise a graphical method for obtaining an esti-
mate for the value of b, which- in turn allowed an estimation
of S*. First rewriting the equation as:
s = S* S~e-b
thien taking the derivative results in:
ds = bS~e-bndn
or, in terms of the slope, ds/dn, of the curve:
ds/dn = bS~e-b
Taking the natural logarithm of both sides results in the
In(ds/dn) = In(bS*) + (-bn).
Thus the logarithm of the slope fits the general equation for
a straight line:
y = ~o + ti,x,
where the intercept 00 equals In(bS*), and the slope B, equals
-b. The slope ds/dn of the original equation can be approxi-
mated by AS/An, the increase in the n-umber of cumulative
species over the number of specimens added with each additional
collection. Tihus, plotting In(As/an) against the cumulative
number of specimens, ni, approximates plotting In(ds/dn) versus
n. Fitting a straight line to the rLesultant points by using
the least squares procedure provides the value B,, which is a
good approximation of b. I'ugging this value of b and the
values of a pair of s and n back in to the original equation
results in a point estimate of S*. Using these crude esti-
mates of S* and b allows the use of PROC NLIN to fit the curve
to the data point, which then gives precise estimation of S*
and b. The goodness of fit of the curve to the data points
can be demonstrated by noting the level of significance for
the F' value, calculated by dividing the mean square of the re-
gression by the mean square of the residual. Sampling effi-
ciency was then simply the actual number of species collected
(s) over the expected number of species (S*') or:
sampling efficiency = s/S*,
Another semigraphical method for determining the total
number of species (S*) is by plotting what is termed a species
abundance curve. F. Wi. Preston first presented this method
in 1948 and elaborated on it in 1958 andi 1962. The data are
first arranged according to abundance of individuals in each
species. The species are then grouped, each successive group
representing species with twice as many individuals as the
preceding group or, in other words, on a base 2 logarithmic
scale. The third "octave," Preston's term for the groups,
represents species which have between 4 and 8 individuals.
Any species which falls on an interval boundary is split
equally between both ocen7ves, thus a species containing 4
specimens is counted as contributing one-half to the second
octave (2-4 individuals) anid one-half to the third octave
(4-8 individuals). Thus thle first octave contains half of
the species having a single specimen and half of the species
having 2 specimens. The number of species per octave is then
plotted against the octave, resulting in a curve such as that
shown in Fig:. 2, representing, the species and abundances of
moths caught in a light trap, as presented by Preston (1948).
Note that the left-hand portion of the curve ends abruptly
at the y-axis, called the "veil line" by Preston. The species
to the left of the veil line are those which would be repre-
sented by less than a single individual if collected in the
same proportion as they exist in the sampling area, and they
are, therefore, "hidden by the veil line," to use Preston's
terminology. The general equation of the Gaussian curve
which Preston fits to this and other data is:
s= soe BR)2
where so is the number of species in the modal octave, Ro
(the octave containing the most species), s is the number of
species in an octave which is R octaves from the modal octave,
and a is ". . a constant calculated from- the experimental
evidence" (p. 258). In practice, a is extremely difficult to
determine. It is derived by solving the equation for the
curve which has been fitted to the data. Unfortunately, fit-
ting a truncated Gaussian curve is a very difficult task. The
statisticians in the Departmlent of Statistics at the Univer-
sity of Florida are awaiting arrival of some special statis-
tical. tables to enable themm to do this. Although most texts
dealing with mathematical ecology (Cody and Diamond 1975,
Pielou 1969, Price 1975, Poole 1974) mention Preston and
present his graphs, I have found very few authors (Good 1953,
Patrick 1954) who are able to apply his techniques to their
own data. However, the value of a has been found to be very
close to 0.2 for all data analyzed by Preston (1948, 1958,
1962). The total number of species in the sampling area S*
can then be found from the relation:
S* = so"F/la
As Preston (1948) carefully points out, this theoretical
total number of species is really the total number of species
in the sampling universe, which differs from the total number
of species in the sample area. It represents the total number
of species which would be found if all individuals collectible
by the sampling methods used were collected during the sample
period. For example, the number of species estimated from
light-trap data represents only those species which are in the
vicinity of the light trap and are attracted or blunder into
it. It does not represent the total number of species of all
moths in the collecting area, just the collectible species.
For purposes of comparison, species abundance curves were
fitted by eye to the data for my 3 study sites and the value
of a was assumed to equal 0.2. While extremely rough, the S*
derived by this method can be compared to S* derived from the
species accumulation curves.
For other analyses of my data 1 used the 76.4 version
of a variety of statistical computer programs which collec-
tively are knownm as SAS (Barr et al. 1976). The most commonly
used statistical procedures were standard descriptive sta-
tistics (PROC MZEANr), correlation (PROC CORR), linear regres-
sion (PROC GLM), stepwise multiple regression (PROC STEPWISE),
nonliniear regression (PROC N~LLIN), and discriminant analysis
(PROC DISCRIM). Standard printing and plotting programs
(PROC PRINT and PROC SCATTER) were also used. All computer
analyses were done by the IBMI 370/165 computer (later replaced
by an AMIDAHL 470 V/6) at the Njortheast Regional Data Center
(NERDC) on the University of Florida campus.
Because of the great variety of analyses, the results
will be discussed under three separate headings: general,
study site comparison, and diversity studies.
RESULTS AND DISCUSSION
Comments on Species List
A total of 5485 aquatic insects were found in 88 collec-
tions from 37 sites in 18 Florida counties. Table 1 presents
a list of collections, locations and dates. Some immature
forms could be identified only to genus level, while 17 speci-
mens of immature Diptera could be placed only to family level.
Represented in the collections were at least 147 species of
aquatic insects belonging to 44 families in 8 orders. Of
these 147 species, 38 were represented by more than one life
stage. These immature frorms were treated as separate classes
in all analyses except those involving species counts. An
annotated list of all species collected is presented in Table
2. The species are arranged according to taxonomic groups;
the orders are arranged in the same evolutionary sequence as
in Usinger (1956). Within the orders the families are ar-
ranged alphabetically, as are the genera and species within
the families. For each species the number of specimens col-
lected, the number of sites and counties, as well as a list
of collection numbers is presented. Any significant correla-
tions with physical or chemical parameters as well as associ-
ations with other insect species are noted. Maximum densities
(text continued p. 116)
Table 2. An annotated list of insects collected in
wjaterbyacinth roots in Florida.
Note: All correlations were computed using normalized
(log transformed) data. For each species under the cate-
gories "water quality preference" and "insect associations"
is listed (1) the factor or species with which the species
being discussed is correlated; (2) the correlation coeffi-
cient, r; (3) p, the probability that the correlation coeffi-
cient observed would occur by chance alone; and (4) n, the
number of times the pair of factors or species being con-
sidered occurred together. Only associations with a prob-
ability less than 0.05 (52) of occurring by chance are men-
tioned. WLhile the probability p takes into account the
number of observations, it is overly sensitive at low values
of n. Correlations which have less than 5 Daired observa-
tions (n<5) therefore have not been included. Maximum den-
sity per square meter is given for species collected using
0.2 square meter sampler.
Ephemeroptera (Mlayflies) --N;ymphs
Only 150 mayfly nymphs of 4 different species were
found. Constituting only 2.7%/ of the total insects col-
lected, they were never very numerous except in collection
#87, where 31 specimens of 3 species made up 41.9% of the
specimens collected there. Katz (1967) mentions collecting
the genera Brachycerus, Cloeon, and Hexaeia in addition to
Caenia from h~yacinths in the Withlacoochee River. The first
two genera are probably misidentifications due to their dis-
tribution and ecological requirements. O'Hara (1961) col-
lected Callibaetis floridanus and Caenis diminuta from
waterbyacinth-covered canals in South Florida.
1. Callibaetis floridanus Banks--45 specimens from
12 collections at 8 different locations in 6 counties.
Collections: 2, 20, 21, 45, 48, 50, 75, 82, 87, 88,
Haximumm density: .6/plant or 30/m2 at #87.
Water quality preference: nitrates, r=-.859, p=.028,
Nrymphs of C. floridanus, known to inhabit brackish water,
". . have the widest limits of toleration of any mayfly
nymph in North A~merica" (B~erner 1950, p. 196). Although
this species was collected numerous times in waterhyacinth
roots by myself and by O'Hlara (1961), Berner (1950) does not
believe them common inhabitants of waterh~vacinths due to low
I collected the majority of nymphs of C floridanuss
during the summer, especially August. Some were also col-
lected in November through January. The absence of nymphs
during the other six months may indicate a bivoltine life
C. floridanus' high correlation with Shannon-weaver's
diversity index (r=.666, p=-.018, n=6) may indicate that it
is especially sensitive to conditions which help increase
the diversity of species at the site. This is supported
by the highly significant correlation to number of species
(r=.745, p=.006, n=12). The high negative correlation to
the nitrate level, one of the chief causes of eutrophication,
and the high negative correlation to number of plants per
square meter (r=-.963, p-.002, n=6) also points this out.
The relative numbers of this species might, therefore, find
use as an indicator of water quality.
2. Callibaetis pretiosus Banks--14 specimens in 4
collections from 2 different locations in Alachua Co.
Collections: 36, 42, 50, 87.
Maximum density: .25/plant or 12.5/m2 at #87.
Berner (1950) believes C. pretiosus to have similar
ecological requirements as C. floridanus except that it is
not quite as tolerant to high and low pH and to brackish
3. Caenis diminuta W~alker--75 specimens from 24 col-
lections at 13 different locations in 10 counties.
Collections : 2, 3, 5, 26, 27, 31, 45, 48, 52, 75, 80,
82, G5, 87, 88, 89, 92, 93, 94, 97, 98, 100, 102, 103.
Hajximum density: .7/plant or 35/m~' at #87.
Insect associations: Hydroanthu oblongus Sharp,
r=.841, p=.002, n=10.
Mly most commonly collected mayfly, it is also recorded
by O'Hara (1961) and is probably the Caenis sp. collected
by Katz (1967). According to Berner (1930), this species
prefers small ponds, especially with emergent vegetation,
but is usually not found in water covered by waterhyacinths.
There is an extremely significant correlation with the
noterid beetle Hydrocanthus _oblngus which probably indicates
that these species prefer the same set of environmental con-
Nymphs were collected throughout the year.
4. Stenacron intergu~nc~tau (Say)--16 specimens from
8 collections at 2 locations in Alachua Co.
Collections: 31, 35, 38, 39, 50, 53, 56, 73.
Maximum density: .129/plant at #35; 2/m2 at #73.
Formerly known as Stenonema proximum, S. interpunctatum
nymphs are usually inhabitants of large streams and sand-
bottomed lakes with little vegetation. Dr. Berner (personal
communication) finds their existence in waterhyacinth roots
to indicate a possible broadening of the species' ecological
MIost of my specimens were collected during the fall and
Odonata (DraonflieE~s and Damselflies)--Nymphs
Odonata were well represented with 1,027 specimens
(18.7% of the total specimens) in 18 species. Odonates
made up at least half of the specimens in many collections
and constituted 75;/ of the specimens in collection #102.
Odonat-e nymphs were the most common arthropod group in
W~ahlquist's (1969) study of fish food organisms in experi-
mnental ponds in Auburn, Alabama.
Since all odonate nymphs are predaceous, their abundance
implies a latrge prey population. The most probable prey item,
at least in rnumbiers,, would b~e the anliphipod Hyalella azteca
(Saussure). Hansen et al. (1971) reports 66 amphipods per
waterbyacinth plant, but in the field they were probably more
numerous, with a bucketful of water frequently containing
literally thousands of amphipods. O'Hara (1961) recorded
as many as 81,000 per square meter at one site.
Zygo~ptera (Damselflies) Nymphs
This suborder, with 752 specimens, comprised 13.7% of
all specimens and 73.2"/ of all odonates. Damselfly nymphs
were found in most collections, with one species, Ischnura
posita, being found in 53.4% of my collections, making it
the most frequently represented species. While only 7
species and the E~. signatum-pollutum complex were found, 5
of these were collected over 8 times as compared to only 3
out of the 11 species of Anisoptera nymphs. O'Hara (1961)
found Ischnura ramburii (Selys) and Enallagma sp. nymphs
beneath waterhlyacinths in canals, while K~atz (1967) records
4 gienera of damselflies from waterhvacinth roots at the
Of the 3 families of damselflies known from Florida,
this was the only one which was found beneath waterbya-
cinths. Of the 35 species in this family that occur in
Florida (Johnson and WJestfall 1970) 7 were collected in
5. Argia apicalis (Say)--1 specimen collected at Lake
Talquin, Gadsden Co.
Density: .05/plant; 5/m2
Due to the rarity of this species in my collections,
no significant conclusions can be drawn concerning its
preferences or associations. While usually considered a
stream species (Walker 1953), it has been recorded from ponds
in Indiana (Montgomery 19414).
6. Argia sedula (Hagen)--39 specimens from 3 locations
in 3 counties.
Collections: 48, 102, 103.
MIaximum; density: .769/plant at #-48; 35/m2 at #,103.
This species was rarely collected. Walker (1953) states
that the habitat of this species is "streams with gentle cur-
rent and rich vegetation on the banks" (p. 144). All 3 times I
collected this species were in streams and rivers of South
Florida during mid December of 1973 and 1974.
7. Enallagma pollutum (Hagen)--2b specimens from 8
collections at 2 places in Alachua Co.
Collections: 17, 18, 31, 38, 72, 75, 82, 94.
Maximum density: .364/plant at #17; 7.5/m2 at #~75.
Nymphss of this species are extremely difficult to dis-
tinguish from E. signatum nymphs. Only larger, last instar
nymphs could be separated with confidence by counting the
number of spines on the caudal lamellae. Smaller nymphs
wer~1e classed no belonging to thle EE1Em signatum-pollutum
complex. No water quality preferences for this species were
noted. All but 3 specimiens were collected from Camps Canal
east of Mlicanopy. All were found between mid-Mlay and
mid-November. Since this species has a significant, nega-
tive correlation with Simpson's diversity index (r=-.826,
p=.012, n=8), it may be a low diversity indicator.
8.Enallagma signatum (Hagen)--25 specimens from 9
collections at 4 locations in 2 counties.
Collections: 38, 42, 50, 53, 56, 75, 87, 96, 99.
Miaximnum density: .35/plant and 35/m2 at #96.
With- only last instar nymphs distinguishable from E.
pol1utum, this species was collected mainly from November
through February. A strong negative correlation (r=.774,
p=.022, n=9) with the number of species may indicate that
older nymphs of this species would not be expected at a
place which has numerous species.
Enallagma signatum-eollutum complex--119 specimens
from 22 collections at locations in 6 counties.
Collections: 2, 3, 17, 18, 27, 31, 35, 38, 42, 45, 48,
50, 56, 72, 73, 75, 78, 83, 94, 96, 102, 103.
Maximum density: .515/plant at #',17; 35/m2 at #102.
WJater quality preferences: pHl, r=.740, p=.023, n=9;
temperature, r=.755, p=,050, n=7, depth, r=-.519, p=.019,
n=20; chloride. r=.880, p=.004, n=8: potassium, r=.866,
Tnsecr as-.ociations: nIlvdrocanthus ob~longus Sharp,
r=.655, p=.040, n=10.
Young Enal~lam nymphrs which could be determined as
belongiing Lo either E~. pallutum or E. signatum but whose
exact species remained unktnown due to their small size and
undeveloped dliagnostic characters were placed in this group.
This species complex was found in shallower water which was
high in chlorides and potassium and had a high pH. It was
correlated with the presence of H_. oblongus These nymphs
were collected throughout the year.
9. Ischnura pos ta (H-agen)--319 specimens from 47 col-
lections at 17 locations in 9 counties.
Collections: 2, 3, 4, 7, 11, 13, 17, 20, 21, 27, 32,
35, 38, 40, 41, 42, 43, 45, 47, 48, 49, 50, 51, 52, 53, 54,
55, 56, 57, 69, 70, 71, 72, 73, 74, 75, 78, 85, 86, 87, 88,
90, 94, 98, 101, 102, 103.
Maximum density: .575/plant at #70; 80/m2 at #69.
Water quality preference: nitrates, r=-.477, p=.045,
Insect associations: Belostoma spp. nymphs, r=.779,
p=.008, n=10: Pelocoris femoratus (Palisa~t-B~eauvois), r=.806,
p=.0001, n=22; Ranatra australis Hungerford, r=-.774, p=.041,
While it was only the fourth most numerous in specimens
collected, i. posita was found in 53.4Z of my collections,
more than any other species. It was present throughout the
vear but reached peak density in the springs,. It appears to
prefer less outrophic waters, i.e., those with less nitrates.
It wans generally found when Ranatra australis was present in
low numbers and when Belostoma spp. nymphs were abundant.
There was an extremely significant correlation with Pelocoris
femoratus, with P. femoraLus being collected only 7 times in
the absence of I. posita. The strength of this association
suggests to me that I. posita may be a preferred prey item
for P. femoratus and/or they both share a common food re-
source such as the amphipod Hyalella azteca (as opposed to
sharing an uncommon food resource). It is possible that
this correlation reflects some unknown symbiotic relation-
sh-ip between th-ese two species. The Ischnura ramburii re-
corded by O'Hara (1961) may have been misidentified I.
posita since the nymphs are very similar and the pigmentation
of the caudal lamellae, which is used as a diagnostic char-
acter, is quite variable or absent in smaller nymphs.
10. Telebasis byersi Westfall--220 specimens from 26
collections at 8 locations in 4 counties.
Collections: 2, 21, 24, 27, 32, 36, 37, 39, 40, 41, 42,
43, 47, 49, 51, 52, 54, 55, 70, 71, 75, 85, 87, 88, 90, 97.
lMaximum density: 1.67/plant at #39; 70/m2 at #97.
Insect associations: Brachyvatus seminulum LeConte,
r=-.918, p=.028, n=5; Scirtes larvae, r=-.886, p=.045, n=5.
The second m~ost numerous odonate in my collections,
T. bversi w~as found in 29.5% of my collections, which ranked
it n:; l~enth in nui~mber of collections in whiich it was present
and sixth in Lotal number of specimens. It was found through-
out the year. Florida specimens of this species were thought
to be Telebasis salva until W~estfall described T. byersi in
1957. 10m~ile there are significant negative correlations
with 2 species of aquatic beetles, no water quality prefer-
ences were noted. Since T. byersi keys out as belonging to
the genus Nehalenia in most keys, it is probable that the
Nehalenia recorded by Katz, (1967) from waterhyacinth roots
was really T. byersi.
Anisoptera (Dragonflies) NymI~Phs
With 275 specimens (5% of the total specimens collected),
this suborder was not as numerous as the damselflies and made
up only 26.7'i of the odonates collected. However, while dam-
self lies were represented by 1 family and 7 species, 11 spe-
cies in 5 different families of anisopterans were collected.
Of these anisopteran species, 5 are represented by only one
specimen and only 3 species were collected more than 3 times.
While usually not numerous, they were occasionally abundant,
comprising 66.4% of the 122 specimens found at collection
O'H~ara (1961) collected 5 species of A~nisoptera from
waterhyacinth-covered canals while Katz (1967) records 8
different species. I collected 4 species in common with
O'H~ara but did not find any Brachymesia (=Cannacria) gravida
(Calvert). While I collected only 3 species in common with
Katz, several of her determinations are suspect, e.g. Dythemis
sp. and Ndannothemis bella are both thought to be found in
Florida only in the western panhandle region.
Hith only 5 specimens in 3 species, this family was rarely
found beneath w~aterhyacinthls and neither O'Hara (1961) nor Katz
(1967) record any aeshnid species.
11. Anax junius (Drury)--1 specimen collected at Camps
Canal in December 1973.
Col lec tion : 42 .
12, Boyeria vinosa (Say)--1 specimen collected in a
stream in December 1973.
13. Coryphaeschna ingens (Rambur)--1 specimen collected
at Camps Canal in December 1973.
Byers (1930) records this species from w~aterhyacinth.
14. Nasiaeschna pentacantha (Rambur)--3 specimens
from 3 collections at 2 locations in 2 counties.
Collections: 35, 45, 50.
Maximumm density: .077/plant at #45.
15. ITe~tragoeuri cynosura (Say)--2 specimens from 2
collections at a small stream in Alachua Co.
Collections: 27, 87.
H~aximium density: .05/plant or 2.5/m2 at #87.
This genus is now~ considered by some authorities to be
part of Epiptheca. Both O'Hiara (1961) and Katz: (1967) col-
lected Tetragioneuria sp. from waterhyacinths.
Most members of this family are burrowing forms and
are rarely found beneath waterbyacinths.
16. Aphyla williamsoni (Gloyd)--1 specimen collected
at Camps Canal on 29 Mlay 1974.
Density: .997/plant or 1/m2
The one specimen I found of this species was a fully
developed nymph from which the adult was starting to emerge,
indicating that it had been knocked into the water from the
aerial portion of the plant where emergence was taking place.
This species lives on the bottom and may be present on water-
hyacinth roots only when moving out of the water for emer-
gence. Katz (1967), however, also recorded this species in
one of her collections.
With 266 specimens in 5 species, libellulids were well-
represented. Two of the species were my tenth and eleventh
most abundant overall. Of the Anisoptera I collected, 96.7%
were members of this family.
17. Erythemis simplicollis (Say)--36 specimens in 9
collections at 9 locations in 4 counties.
Collections: 2, 3, 13, 21, 45, 47, 83, 86, 87.
Maximum density: .615/plant at //45; 40/m2 at #85.
This species was not usually very abundant in my col-
lections, but 18 of the 29 odonate nymphs found by O'Hara
(1961) in waterbyacinth mats were E. simplicollis. My
specimens were all collected in late summer or during Decem-
18. Miathvria marcella (Selys)--116 specimens from 10
collections at 9 locations in 6 counties.
Collections: 5, 7, 20, 30, 31, 47, 75, 88, 100, 102.
Hlaximum density: 5/plant or 400/m2 at #102.
While this w~as my most abundant Anisoptera nymph, 69%!
of all the specimens of this species were found in collection
#'102. O)'Hara (1961) recorded th-is species. Katz (1967) did
not report this species, possibly due to its absence from the
keys she used. Byers (1930) did not find this tropical spe-
cies in Florida, while today this is one of our more common
species. It would be interesting to compare the overlap of
distribution of this recently introduced species with that of
waterhyacinth. I collected nymphs of this species only be-
tween June and Decemtber.
19 Pachyd~iplax longipennis (Burmeister)--111 specimens
in 30 collections from 15 locations in 8 counties.
Collections: 2, 4, 7, 13, 29, 30, 31, 35, 36, 38, 42,
4-5, 4~7, 51, 53, 54, 56, 57, 6s9, 70, 75, 79, 83, 85, 87, 88,
94, 95, 97, 103.
H aximum density: .391/plant at #51; 37.5/m2 at #85.
!Jater quality preference: iron, r=.900, p=.014, n=6.
Insect- associations: Pelocoris femoratus, r=-.554,
While not overly abundant at any location, this species
w~as wJidely distributed, and among the odlonates, only Isch-nura
piosita was found in more collections. Among all my specimens,
P. 10n~gagenngs ranked seventh in the number of collections in
which it was found and eleventh in overall abundance. Nymphs
were collected throughout the year, usually at locations
which had a high iron content and harbored low populations of
Pelocoris femoratus. Not surprisingly, both O'Hara (1961) and
Katz (1967) report this species from their waterhyacinth col-
20. Perithemis tenera (Say)--1 specimen collected at
Lake Lawne in Orlando on 13 July 1973.
Frequently considered as a subfamily of the family
Libellulidae, nymphs of this family are bottom sprawlers
and were infrequent in my collection.
21. Mlacromia taeniolata Rambur--2 specimens from 2
collections in 2 counties.
Collections: 102, 103.
Maximum density: .062/plant or 5/m2 at #102.
Both specimens were found in mid-December.
Hemiptera (True Bugs)--Adults and Nymphs
While aquatic hemipterans, with 504 specimens, com-
prised only 9.2% of my total specimens, they represent 10
different families and at least 22 species. The contribu-
tion of aquatic Hemiptera to the diversity (at least the
richness component of diversity) of a collection was rela-
tively great. Only the aquatic beetles and Diptera could
equal the large number of different families in my collec-
tions. While sometimes locally abundant, they were usually
O'Hara (1961) reported only 2 kinds of aquatic Hemiptera:
Belostomatidae and Naucoridae. Katz (1967) lists 9 families
and 10 genera of aquatic H~emiptera, with Notonecta sp. the
only one which I did not also collect.
Since all aquatic HIemiptera (except the Corixidae) are
predaceous, their presence should be associated with that of
a suitable prey organism, most probably the amphipod Hyalella
Belostomatidae (Giant W~ater Bugs)
H~ith 82 specimens, the 3 species of Belostomatidae com-
prised 16.3% of the total Hemiptera collected and were my
second most abundant hemipteran family. They were usually
not particularly abundant at any one site, averaging less
than 4 specimens whenever found. However, they did consti-
tute 26.6%/ of the 60 specimens in collection #54. Both O'Hara
(1961) and Katz (1967) as well as Hansen et al. (1971) men-
tion collecting Belostoma spp. in waterhyacinth roots.
22. Belostoma lutarium (Stal)--12 specimens in 11 col-
lections from 4 locations in 3 counties.
Collections: 18, 36, 70, 71, 72, 75, 88, 94, 95, 96,
Maximum density: .18/plant or 10/m2 at #J94.
Adults of this species were never abundant, more than
one specimen being collected only once. My specimens were
all collected between May and December.
This species was strongly correlated with all the
diversity indices except Simpson's index. The Shannon-
Weaver index for collections in which this species was found
averaged at a relatively high 2.26. The presence of this
species might be used as an indicator of high diversity of
insect fauna in w~aterhyacinth- covered wa~iers.
23. Belostoma testaceum (Leidy)--17 specimens in 10
collections at 5 locations in 2 counties.
Collections: 7, 11, 19, 22, 33, 37, 40, 54, 90, 94.
Maximum density: .333/plant at #137, 10/m2 at #94.
Most of my specimens (65%) came from a drainage culvert
along Interstate 75 north of MZicanopy. Except for 2 speci-
mens found in February, all were collected from June through
Nlovelb er .
Belostoma spp. nymphs--52 specimens in 16 collections
at 5 locations in 3 counties.
Collections: 3, 22, 33, 35, 52, 54, 57, 71, 75, 78, 79,
86, 88, 89, 91, 92.
H~aximum dlensity: .83j/plant at #i22; 15/m2 at #75.
Insect associations: Ischnura posita, r=.779, p=.008,
n=10; Chironomus attenuatus Walker, r=-.661, p=.052, n=9.
Since only the adults of this genus could accurately
be placed at species level, all nymphs of the genera were
placed in this category. Nymphs were collected from Febru-
ary through October.
24r. Lethocerus uhleri (Montandon)--1 specimen col-
lected at Camps Canal in October 1973.
Corixidae (1Jater Boatmen)
All specimens were collected at Otter Creek, Levy Co.,
in July of 1973.
Collections: 20, 21.
Maximum density: .154/plant at #"21.
While the 2 adults appear to be Trichorixa sp., the
other specimen is a nymph and cannot be keyed. Members of
this family are rare among waterbyacinths and were all col-
lected at one location. They were not reported by either
O'Hara (1961) or Kiatz (1967).
Gerridae (Water Striders)
These surface-film forms were rare in my collections.
Although present at the collection sites, they generally
seem to prefer more open water than that found between water-
hyacinth plants. Only a total of 6 specimens of this family
26. Gerris canaliculau (Say)--only 1 specimen col-
lected at a small stream near Gainesville in August 1974.
Density: .05/plant or 2.5/m2
27. Limnogionus hesione (Kirkaldy)--only 1 specimen
collected at Otter Creek in Levy Co. in July of 1973.
Katz (1967) reports collecting Liminogonus sp. in water-
28. Trepobates sp.--only 1 specimens collected with
L. hesione (above).
Unidentified gierrid nymphs--3 specimens from 3 locations
in 2 counties.
Collections: 20, 38, 75.
Maximum density: .045/plant at #~38 and 2.5/m2 at #75.
Hebridae (Velvet Wjater Bugs)
These small insects, less than 2.5mm long, are incon-
spicuous and frequently overlooked since, unless they are
moving, they appear to be pieces of debris. A total of 21
specimens in 4 species were collected, all of them in Alachua
Co. Katz (1967) records Hebrus sp. from waterhyacinths in
the Withlacoochee River.
29. Hebrus burmeisteri (L. & S.)--4 specimens from 3
locations in Alachiua Co.
Collections: 30, 54, 72.
Maximum density: .971/plant at #~54; 1.0/m2 at #72.
30. H~ebrus consolidus Uhler--5 specimens from 5 loca-
tions in Alachua Co.
Collections: 15, 16, 49, 54, 70.
H~aximum density: .048/plant at #4-9; 1.67/m2 at #70.
All my specimens of this species were collected between
January and July.
31. Merragataa brevis Champion--only 1 specimen from
Camps Canal in Mlay of 1974.
Density: .014/plant or 1.66/m2
32. Mlerragata brunnea Drake--11 specimens in 4 collec-
tions from 3 locations in Alachua Co.
Collections: 15, 29, 75, 85.
Maximum density: .083/plant at #15, 7.5/m2 at #85.
All my specimens of this species were collected during
the summer months.
Hydrometridae (HJater Measurers)
With only 18 specimens distributed between 2 species,
this family was never particularly abundant. Katz (1967) re-
ports finding them in 2 of her collections.
33. Hydrometra myvrae Bueno--11 specimens from 7 collec-
tions, all at Camps Canal.
Collections: 31, 69, 70, 72, 73, 95, 99.
Maximum density: .235/plant or 20/m2 at #99.
MIy most common hydrometrid, II. myrae is considered
"... the most prevalent Hydroetra~ in the state" (H~erring 1948,
p.113). My specimens were collected between April and Novem-
34. Ilydromtr wilevi Hungerford--3 specimens in 2 col-
lections from 2 counties.
Collections: 27, 107.
Maximum density: .125/plant or 10/m2 at #126.
Herring (1948) records this species from only one site
in the state. My records add 2 new counties to the distribu-
tion of this species.
Hydrometra sp. nymphs--4 specimens from 4 collections
at 4 locations in 3 counties.
Collections: 1, 2, 21, 47.
Mlaximumn density: .077/plant at #;21.
H~esoveliidae (Ndater Treaders)
Never very abundant and with only 59 specimens, meso-
veliids were found in 30 of my collections, making them the
second best hemipteran family for representation in differ-
ent collections. Katz (1967) also collected some Mesovelia
in wsaterhya cinths .
35. Hlesovelia mulsanti White--15 specimens in 10 col-
lections from 7 locations in 4 counties.
Collections: 5, 17, 54, 70, 75, 85, 88, 94, 96, 100.
H~aximum density: .091/plant at #94, 10/m2 at #185.
All the adult mesoveliids I collected belonged to this
species. Hterringi (1951b,) records this species from
wa~terhyac inth~s. There is a strong correlation with root
length (r=.813, p=.008, n=9) and high root:shoot ratio
(r=.807, p=.009, n=9) for this species. Adults were col-
lected in February and May through April.
MIesovelia sp. nymphs--44 specimens in 23 collections
at 10 locations in 4 counties.
Collections: 1, 2, 11, 41, 21, 22, 23, 24, 33, 36, 38,
49, 53, 70, 72, 73, 78, 89, 91, 93, 94, 95, 96.
Maximum density: .227/plant at d38; 15/m2 at #95.
Insect associations: yxi~osagu spp., r=-.746, p=.013,
Although all mesoveliid nymphs key out to M. amoena
Uhler, I classed all these nymphs simply as Mlesovelia sp.
Dr. Jon Herring at the National Museum considered most (if
not all1) of my mesoveliid nymphs to be HI. amoena, but I con-
sider this unlikely since I collected no0 M. amoena adults
but many N. mulsanti adults. There is a strong, positive cor-
relation with the presence of stratiomyid larvae, M~yxosargus.
These nymphs were collected throughout th-e year.
Hlaucoridae (Creeping Water Bues)
Inaucorids wJere not only the best-represented hemipteran
family,, being present inl 43 collections, b~ut it was also the
most numerous hemipteran family, with 194 specimens making up
38.53: of hemipterans collected and 3.5% of the total insects.
Bothi O'Hara (1961) andi Katz; (1967) recordl this family, which
is Irepresenited in the eastern Ui. S. by 1 genus and 3 species.
36. Pelocoris balius La Rivers--107 specimens from 31
collections at 9 locations in 3 counties.
Collections: 7, 11, 14, 15, 17, 19, 23, 24, 31, 32,
35, 36, 38, 39, 42, 50, 55, 56, 69, 70, 71, 72, 73, 75, 76,
89, 91, 94, 95, 97, 99.
Maximum density: .405/plant at #35; 16/m2 at #72.
Water quality preference: potassium, r=-.739, p=.009,
Insect associations: Pelocoris femoratus, r=.490, p=.047,
n=17; Hydrovatus larvae, r=.881, p=.048, n=5; Suphisellus in-
sularis (Sharp), r=-.549, p=.052, n=13; Suphisellus puncticol-
lis Crotch, r=.816, p=.048, n=6.
Considered as a subspecies of P. femoratus by La Rivers,
P. balius has been elevated to species rank by me since the 2
species were collected together 17 times.
P. balius is associated with low levels of potassium,
with P. femoratus, Hydrovatus larvae, and with Suphisellus
puncticallis. It is generally found in waters where the num-
bers of S. in~sularis are low. It was collected throughout
the year, with over 79% of the specimens coming from Camps
37. Pelocoris ~femoratus (Palisot-Beauvois)--87 speci-
mens in 29 collections from 8 locations in 2 counties.
Collections: 3, 11, 31. 35, 38, 40, 42, 50, 53, 55, 56,
69, 70, 71, 72, 73, 74, 75, 78, 81, 83, 85, 86, 88, 89, 93,
95, 09, 102.
Maximum density: .205/plant or 25/m2 at #70.
Insect associations: 1schnura p~osit, r=.806, p=.0001,
n=22; Pachydiplax longipennis, r=-.554, p=.050, n=13; Peloc-
oris balius, r=.471, p=.057, n=17; Ranatra australis Hunger-
ford, r=-.920, p=.010, n=6.
This species also was collected throughout the year,
frequently along with P. balius, with which it is marginally
correlated. There is an extremely significant correlation
(p=.0001) with the damselfly nymph Ischnura posita, with P.
femora~tus being found only 7 times (out of 29) in the absence
of I. Roita P. femoratus is negatively correlated with the
anisapteran Pachydiplax 10ngipennis and with the nepid Ranatra
australis. There is a strong correlation (r=-.601, p=.01,
n=17) with plants having fewer leaves. Over 83% of the speci-
mens were collected at Camps Canal.
Nepidae (Nater Scorpions)
With 52 specimens in 9 collections, the 3 species of
water scorpions collected made up only 10.3% of the hemip-
terans collected and less than 1% of the total specimens.
Hlowvever, th~ey were sometimes extremely abundant locally,
with collection #75 averaging 75 specimens/m2. Neither O'Hara
(1961) nor Katz (1967) records any water scorpions from their
38. R~anatra australis Hungerford--35 specimens in 10
collections from 4 locations in 3 counties.
Collections : 3, 17, 73, 75, 76, 94, 95, 99, 102, 103.
Maximum density: .486/plant or 75/m2 at #75.
Water quality preference: nitrates, r=.747, p=.033,
Insect associations: Ischnura posi~ta, r=-.774, p=.041,
n=7; Pelocoris femoratus, r=-.920, p=.010, n=6.
Herring (1951a) records this ". . the most prevalent
Ranatra in Florida" (p. 18) fromt waterhyacinths. This species
was gJenerally found in shallow water which was high in nitrates.
It is negatively associated with the presence of either I.
posita or P. femnoratus. Almost 43% of my specimens were col-
Tected at one time at Prairie Creek in DeSoto Co. Of the re-
maining specimens, 80% came from Camps Canal. My specimens
were all collected from May through December.
39. Ranatra buenoi Hungerford--6 specimens collected
only at Camps Canal, Alachua Co.
Collections: 70, 73, 94, 95.
Maximum density: .20/plant or 15/m2 at #95.
This relatively rare species was collected only at
Camps Canal in May, October, and November.
40. R~anatra nigr Herrich-Schaeffer--11 specimens from
2 collections in 2 counties.
Collections: 102, 103.
Maximum density: .625/plant or 50/m2 at #~102.
Although it was collected only in December at 2 loca-
tions in South~ Florida, it was quite abunldant at Prairie
Creek, averaging 50 water scorpions per square meter.
Pleidae (PygmY Back Swimmers)
Although members of this family, which is represented
by a single species in the U. S., prefer dense, tangled vege-
tation, they were never common in my collections.
41. Ifeoplea seriola (Fieber)--22 specimens in 9 col-
lections from 4 locations from 2 counties.
Collections: 3, 18, 36, 53, 71, 75, 87, 88, 97.
Mlaximumn density: .50/plant: or 25/m2 at #97.
Water quality preference: potassium, r=.910, p=.032,
Formerly known as Plea striola, this species was assoc-
iated with waters high in potassium and having a low Simp-
son's diversity index, with 59% coming from Lake Alice. Katz
(1967) recorded this species from 2 of her collections. Al-
thiough they were found throughout the year, 64% of the speci-
mens were collected during the summer months.
Veliidae (Broad-Shouldered Water Striders)
With 47 specimens distributed between 2 species, this
family was never common, averaging less than 2 specimens
when found, but it was well-represented, being present in
over 30% of my collections. Thus, although they comprise
less than 1% of the total insects collected, their contribu-
tion to the diversity of the collection is important.
42. M~icrovelia borealis Bueno--25 specimens in 15 col-
lections from 12 locations in 4I counties.
Collections : 1, 2, 15, 16, 19, 21, 23, 24, 25, 34, 43,
44, 53, 72, 75.
Hlaximum density: .095i/plant at #16h; 2.5/m2 at #~75.
Katz (1967) recorded this genus in her collections.
43. Velia brachialis Stal--22 specimens in 13 collec-
tions from 5 locations in 4 counties.
Collections: 17, 18, 21, 31, 35, 38, 42, 50, 87, 90,
95, 102, 103.
Maxirmum density: .182/plant at #138; 10/m2 at #103.
Velia is negatively correlated with number of leaves
per plant (r=-.933, p=.020, n=5).
Almost 73% of the specimens came from Camps Canal. Katz
(1967) reports this genus from her collections.
Trichoptera (Caddisflies) Larvae
Only 16 caddisfly larvae belonging to 2 species were
found during this study anid 15 of these came from the same
location. Although from m~y collections Tlrichoptera would
appear to be rare, O'Hiara (1961) reports a probable 3 differ-
ent, unidentified species while Katz (1967) recorded 5 dif-
feren~t groups in her survey.
44. 0ecotis prob. cinerascens (Hagi.)--6 specimens in 4
collections from 2 locations in Alachua Co.
Collections: 3, 35, 50, 87.
Hlaximum density: .05/plant or 2.5/m2 at #87.
All but one of my specimens came from Camps Canal and
all were collected in August and September except for one
specimen collected in January.
45. Polycentropus spp. larvae--10 specimens in 7 col-
lections, all from Camps Canal.
Collections: 70, 72, 74, 76, 80, 82, 91.
Maximum density: .059/plant at #82; 4/m2 at #7;4.
Insect associations: Chironomus attenuatus, r=-.873,
All my specimens were collected between May and Septem-
ber at Camps Canal. Katz (1967) recorded this genus from
all her collections.
A total of 17 specimens, all belonging to the same
species, were collected. I have found no references for any-
one reporting this group from waterbyacinths.
46. Chauliodes rasticornis Rambur--17 specimens from
10 collections at 2 locations in Alachua Co.
Collections: 11, 42, 69, 70, 72, 73, 74, 89, 91, 95.
Maximum density: .059/plant at #:92; 10/m2 at #69.
WaJter quality preference: nitrates, r=-.697, p=.055,
The larvae of this species can be identified by the
yellow stripe on trhe dorsulm, as noted by Cuyler (1958). The
larvae I collected were correlated with low levels of nitrates
in the w~ater. All but one specimen came from Camps Canal.
With 2,498 specimens (45.5%/ of the total specimens),
the beetles were my most abundant order and were also the
best represented, being found in all but 3 of my collections.
Collections averaged 33.9 beetles and beetle larvae per col-
lection (excluding the 3 beetle-less collections) At least
64 different species belonging to 42 genera and 10 aquatic
families were present. Of the 2,498 beetle specimens, 14.8%
were larvae. Most beetle larvae could be identified only
to genus except when the genus was monotypic.
O'Hara (1961) reports the following beetles from his
waterhyacinth root collection: Dytiscidae, Omophronidae,
Haliplidae, Gyrinidae, Carabidae and Hydrophilidae. Carabi-
dae, of which he had only one specimen, are not considered
aquatic. Omophronidae, whJlich are considered part of Carabi-
dae by some authorities, have not been reported frequently
from Florida and are semi-aquatic, living in the moist soil
of riverbanks and lakes. I believe, since these were his
most common beetles (and I collected none), that he probably
has misidentified a noterid such as Colpius inflatus or a
dytiscid such as Hiydrovatus sp.
Katz (1967) presents 45 different genera of beetles.
At least 10 of her genera, however, belong to families such
aIs Alnthicidae, Curculionidiae, Scaphidiidae, Scolytidae and
Staphylinidae. which are not known to have truly aquatic mem-
bers in Florida. By truly aquatic, I mean insects which have
at least one life stage in or on the water. She also lists
Limnichus sp., which is only known from southwestern U. S.,
and Heogors which has not, to my knowledge, been other-
wise recorded from Florida.
Dryopidae (Long-Toed Water Beetles)
With only 3 specimens of one species, dryopids were
relatively rare in waterhyacinth roots. Since they are
usually found in riffle areas of small, rapid streams, few
would be expected in the lentic situations in which water-
hyacinths are found.
47. Pe~olonm obscurus gracilipes Chevrolat--3 speci-
mens from 3 collections at Camps Canal, Alachua Co.
Collections: 89, 91, 94.
Maximum density: .091/plant or 5/m2 at #94.
Katz (1967) reports collecting Pelonomus sp.
Dytiscidae (Predaceous Water Beetles)
WJith 752 specimens, the dytiscids, along with the
Coenagrionidae, were tied for third most abundant family,
preceded only by the noterids with 1,432 specimens and the
chironomids with 1,046 specimens. At least 30 different
species of dytiscids were collected, by far surpassing the
16 species of chironomids, which ranked second in terms of
number of species per family. Thus, the contribution of
dytiscids to the diversity of the sysLem w~as very great,
especially since they comprised only 13.7% of the total
specimens collected and 30.11/ of the beetles collected.
Of the dytiscid subfamilies, the H-ydroporinae are
definitely a characteristic component of waterhyacinth root
beetle fauna. The Hydroporinae comprised 89.6% of my dytis-
cid specimens and of the 12 genera known from Florida, I
Of the dytiscid specimens, 132 (17.6%) were larvae. Of
the beetle families, only the helodids had more larvae.
Katz, (1967) lists the following dytiscid genera from her
waterbyacinth root collections: Bidessus, Laccodytes, Hydro-
vatus, Laccophilus, Orcodytes (=Hydro orus i at) eia
Cogelatus, Derovatellus, Hygrrotus, Coptotomus and Cybister.
Of these genera I collected all but Laccodytes, Derovatellus
and ~1gygots O'HIara (1961) found no dytiscid larvae and
found adults were relatively rare.
48. Anodochiilus exiguus (Aube)--1 specimen collected at
Otter Creek, Levy Co., inl July.
49. Bidessonotus longavalis (Blatchley)--1 specimen
collected at Camps Canal, Alachua Co., in July.
50. Bidessonotus pulicarius (Aube)--7 specimens in 5
collections at 2 locations in Allachua Co.
Collections: 36, 75, 93, 94, 95.
M-aximumn density: .182/plant or 10/m2 at #~94.
Five of my specimens were collected in October while
one each were found in June and November.
Bidessonotus spp. females--4 specimens in 4 collections.
Collections : 2, 34, 71, 74.
Maximum density: .026/plant or 5/m12 at #71.
Since some species of Bidessonotus can be identified
only by the shape of the aedeagus, some unidentifiable fe-
males of the genera are placed in this classification.
51. Bidessus flavicollis (LeConte)--4 specimens in 2
collections in Alachua and Orange Cos.
Collections: 3, 96.
M~aximumn density: .15/plant or 15/m2 at #96.
My specimens were collected in August and November.
This species is usually found in algal mats.
52. Brachyvatus seminulus (LeConte)--79 specimens in
10 collections from 6 locations in 2 counties.
Collections: 3, 25, 35, 71, 75, 79, 85, 87, 88, 94.
Maximum density: 1.163/plant or 142.5/m2 at #85.
Insect associations: Telebasis bversi, r=-.918, p=.028,
This genus was once considered as a subgenus of Bidessus.
A single collection at the Styx River in August accounted for
72% of my specimens of this species, which reached a density
of over 140 per square meter. All of m~y specimens were col-
lected from May through October, with the majority (85%)
being collected in August.
53. Ce~lina anustata Aube--6 specimens in 5 collections
from 5 locations in 3 counties.
Collections: 15, 19, 29, 77, 79.
Maximum density: .133/plant and 5/m2 at #~77 and #79.
All my specimens were collected in June or July.
54. Celina grossula LeConte--19 specimens in 10 col-
lections from 6 locations in 2 counties.
Collections: 13, 15, 23, 30, 44, 80, 81, 89, 91, 92.
Maximum density: 3.84/plant or 25/m2 at #92.
Insect associations: Colpius inflatus (LeConte),
r=-.934, p=.020, n=5.
This species was significantly negatively correlated
with almost all the measures of diversity, indicating that
it is generally found in low diversity locations. All my
specimens were collected June through September except 2
collected in December in South Florida.
55. Celina slossonni Mlutchler--3 specimens from 3 loca-
tions in Alachua Co.
Collections: 29, 34, 92.
Maximum density: .077/plant or 5/m2 at #~92.
COLlected in July, September and October.
Celina spp. larvae--44 specimens in 22 collections from
9 locations in 2 counties.
Collections: 12, 13, 19, 22, 24, 26, 29, 32, 34, 36,
37, 39, 52, 53, 54, 73, 81, 82, 85, 86. 89, 91.
Maximum density: .333/plant at #37, 15/m2 at #86.
Insect associations: Hydrocanthus larvae, r=.840,
p=.009, n=8; P'oyp~e~dilm illinoense (Mlalloch), r=.878,
My second most abundant beetle larvae, Celina larvae
were found in more different collections than any other
Celina spp. larvae were collected throughout the year
but appear to be most commIon during late summer and fall.
These larvae were found to be strongly associated with a
dytiscid larva and significantly associated with a chiro-
56. Copelatus caelatipennis princeps Young--11 speci-
mens in 8 collections from 5 locations in 2 counties.
Collections: 1, 3, 7, 37, 52, 74, 81, 89.
Maximum density: 1.667/plant at #37; 3.33/m2 at #8s9.
All my specimens of this species were collected from
June through October except for one specimen collected at
Lake Alice in February.
57. Copeflatus chevrolati chevrolati Aube--13 specimens
in 5 collections from 3 locations in 2 counties.
Collections: 1, 21, 42, 93, 94.
Maximum density: .183/plant or 10/m2 at #194.
Found from July through December, this species is known
to prefer areas with accumulations of organic debris.
58. Coptotomus inppyoggyp g obscurus Sharp--3 specimens
froml 3 locations in 2 counties.
Collections: 17, 27, 103.
Maximum density: .032/plant or 5/m2 at #~103.
Both of my Alachua Co. specimens w~ere collected in July
while the one from South Florida was found in December.
59. Cybister frimlbiolatus crotchi Wilks.--1 specimen
from a drainage ditch in Alachua Co. in February.
While only a single adult of this species was collected,
11 larvae were found. This may indicate that the larvae are
most attracted to waterhyacinth roots or perhaps that the
adults, among the largest of Florida's water beetles, were
able to elude capture.
Cybister sp. larvae--11 specimens in 7 collections from
6 locations in 3 counties.
Collections: 33, 47, 75, 78, 85, 88, 99.
Maximum density: .073/plant: or 7.5/m2 at #~75.
Cybister sp. larvae were collected from June through
December with most (45.4%/) being found in June.
60. Desmopachria grana (LeConte) complex--27 specimens
in 16 collections from 11 locations in Alachua Co.
Collections: 1, 11, 14, 15, 19, 23, 27, 34, 37, 39, 41,
52, 57, 91, 94, 95.
MIaximum density: .654/plant or 25/m2 at #94.
Insect associations: My~xosargus sp., r=.845, p=.034,
This small dytiscid, characteristic of detritus pond
condiitions, w~as well-distributed throughout Alachua Co.,
being present at 11 of the 16 different areas I collected.
I collected this species throughout most of the year, with
the exception of the spring months. It w~as most abundant in
myr October collections, with 38% of the total Desmopachria
61. Ilydroporus dix~ianus Fall--2 specimens collected
together at Lake Talquin, Gadsden Co.
Density: .1/plant or 10/mn2
Only one other specimen of this rare species has been
recorded from Florida (Young 1955).
62. Hy~droou lobatus Sharp--ll specimens from 6 col-
lections at 3 locations in Alachua Co.
Collections: 17, 18, 51, 74, 76, 83.
Maximum density: .087/plant at #51; 5/m2 at #83.
Of my specimens, 10 were collected during the summer and
one in Jaunary. Most (73%) were collected at Camps Canal in
63. Hydroporus lynceus Sharp--1 specimen collected in
November at Lake Talquin, G~adsden Co.
Density: .05/plant or 5/m2
64. Hiydroporus v~ittatipennis Gemminger & Von Harold--
1 specimen collected in November at Lake Talquin, Gadsden Co.
Density: .05/plant or 5/m2
65. Hlydrovatus compr~essus Sharp--276 specimens in 36
collections from 15 locations in 6 counties.
Collections : 12, 23, 24, 25, 26, 29, 30, 32, 33, 36, 37,
39, 41, 44, 46, 49, 52, 55, 57, 71, 75, 76, 78, 79, 80, 81,
83, 84, 85, 86, 88, 89, 91, 92, 93, 97.
Maximum density: 1.095/plant at #~49; 110/m2 at #~71.
Water quality preferences: sulphur, r=.785, p=.021, n=8;
magnesium, r=.604, p=.010, n=17; sodium, r=.554, p=.021, n=17.
Insect associations: Suphisellus insularis (Sharp),
r=.548, p=.006, n=24.
This species was extremely well represented, being found
in 41% of my collections. Only one species, the damselfly
Ischnura posita, was found in more collections. In sheer
numbers, it was my most common dytiscid, my second most abun-
dant beetle, and fifth in abundance out of all the species.
Comprising 36.7% of the dytiscids collected, this beetle has
significant correlation coefficients indicating that it occurs
most commonly in waters that have relatively high values of
sulphur, magnesium and sodium. It is negatively correlated
with leaves per plant (r=-.486, p=.048, n=17). It is also
extremely strongly associated with the presence of the noterid
Suphisellus insularis, the only beetle monre abundant than H.
compressus. If one of these species were present, in 67%
of the cases the other would be present also. Lake Alice
accounted for 63.6% of the specimens, whiiLe Camps Canal pro-
vided an additional 17%/. This species was abundant especially
during late w~inter and was collected throughout the year.
66. Hydrovatus inexpec~tatu Young--1 specimen collected
in July at Camps Canal, Alachua Co.
Collection : 80.
Density: .013/plant or 1/m2
This species is very similar to H. compressus and more
specimens may be mixed in with that species. However, Dr.
Frank Young, whio described the species (Young 1963), examined
all my specimens and believes that my determinations are cor-
67. Hydrovatus peninsularis Young--104 specimens in 25
collections from 12 locations in 3 counties.
Collections: 3, 11, 15, 16, 19, 22, 23, 24, 25, 29,
33, 37, 39, 40, 44, 57, 74, 76, 80, 85, 86, 89, 81, 93, 97.
Maximum density: .667/plant at #37; 40/m2 at #86.
Insect associations: Phaenotumn excstriatus Say, r=.781,
p=.013, n=9; Hesonoterus addendus (Blatchley), r=-.837, p=.038,
Accounting for 14% of the dytiscid specimens, this was my
second most abundant and my second best represented species
of dytiscid. Strongly associated with taller waterhyacinth
plants (r=.447, p=.018, n=24), H. peninsularis is also associ-
ated w~ith the presence of the hydrophilid Phaenotum exstria-
tus. The conditions which favor the presence of the noterid
Mesonoteruss addendus would seem to inhibit: the presence of
H. peninsularis. 11. peninsularis appeared to be most abun-
dant in July, during which 30.8% of my specimens were collected.
Hpdrovatus sp. larvae--20 specimens in 12 collections
from 5 locations in 2 counties.
Collections: 23, 27, 30, 39, 41, 44, 71, 74, 87, 94,
95, 97, 99.
Maximum density: .194/plant at #;55, 5/m2 at #~75.
Although Hydrovatus larvae are correlated (r=.912,
p=.031, n=5) with higher water temperatures, they were found
throughout the year.
68. Laccophilus gentilis LeConte--17 specimens in 13
collections from 7 locations in 3 counties.
Collections: 23, 27, 30, 39, 41, 44, 71, 74, 87, 94,
95, 97, 99.
Maximum density: .2/plant or 10/m2 at #97.
Insect associations: Suphisellus insularis, r=-.832,
Collected from May through December, over half the
specimens were found during the fall. A greater number of
leaves per plant was correlated to this species (r=.823,
69. Laccophilus proximum Say--14 specimens in 9 col-
lections from 4 locations in 3 counties.
Collections: 38, 45, 75, 91, 92, 94, 95, 101, 102.
Maximum density: .188/plant or 15/m2 at #~102.
Water quality preference. pH, r=.798, p=.031, n=7.
This species seems to prefer waters high in pH. It was
collected from June through December.
LaccoIly lus sp. larvae--4 specimens from 4 locations in
Collections: 12, 27, 29, 33.
Maximum density: .062/plant or 3.33/m2 at #:88.
All the Laccophilus larvae were collected during the
70. Liodessus affinis (Say)--2 specimens from 2 loca-
tions in 2 counties.
Collections : 18, 4;8.
Maximum density: .026/plan-t at #4-8.
Liodessus was previously considered a subgenus of
71. Liodessus fuscatus (Crotch) -- 1 specimen collected
in July at Otter Creek, Levy Co.
This species is usually associated with sphagnum moss.
72. Matus ovatus blatchleyi Leech--1 specimen collected
in June at a pond in Alachua Co.
M~atus sp. larvae--2 specimens collected fromi 2 locations
in Alachua Co. during June and July.
Collections: 11, 23.
Maximum density: .017/plant at #23.
73. Neobidessus pullus floridanus (Fall)--1 specimen
from Lake Allice, Alachua Co., in June.
Density: .024/plant or 2.5/m'.
This genus was also part of the old genus Bidessus.
74. Pachydrus obniger Chevrolat--3 specimens from col-
lections at a drainage ditchr in Alachua Co.
Collections: 33, 40.
Hiaximum density: .038/plant.
Florida specimens of this species were called P. princeps
(Blatchley), but Dr. F. N. Young (personal communication) does
not believe them to be distinct from the P. obnige from Cuba.
Pachydrus obniger larvae--45 larvae in 18 collections
from 8 locations in 3 counties.
Collections: 14, 24, 29, 33, 37, 52, 71, 74, 75, 78,
79, 83, 84, 86, 89, 91, 93, 94.
H~aximum density: .469/plant or 25/m2 at #89.
Water quality preference: iron, r=.912, p=.031, n=5.
BLatchley (1914) recorded this species (as P. princeps,
a new species) from beneath dead waterbyacinths along the
shore of Lake Okeechobee.
Insect associations: Hydrovatus Eeninsularis, r=.681,
p=.043, n=9; Phaenotumr exstriatus, r=.875, p=.010, n=7;
Suphisellus g~ibbulus (Aube), r=.744, p=.021, n=9; M~yxosargus
sp., r=.889, p=.017, n=6.
Pachydrus larvae were found in waters having high iron
content and -frequently havinge populations of the dytiscid H~.
peninsularis, the hydrophilid P. exstriatus, the noterid S.
gibutus, and the stratiomyid Myxasargus. Over 55% of the
specimens came from Camps Canal, and all but 2 specimens were
collected between May and October.
75. Thermonoctuls basillaris (H~arris)--1 specimen col-
lected in July at a drainlage- ditch in Alachua Co.
76. Uvarus falli (Young)--1 specimen collected in July
at Otter Creek, Levy Co.
This species is thought to be restricted to sand-
bottomed streams such as Otter Creek.
Unidentifiable Bidessini larvae--10 larvae in 7 collec-
tions from 4 locations in Alachua Co.
Collections: 19, 24, 25, 49, 52, 86, 91.
Maximum density: .143/plant at #~49; 5/m2 at #86.
Some dytiscid larvae could be identified only to the
tribe level. The tribe Bidessini has 5 g~enera: Bidessus
(sensu latu), Bidessonotus, Brachyvatus, Desmopachria and
Elmidae (Riffle Beetles)
It is not surprising that I collected only 3 specimens
from this family since its members are usually restricted to
riffle areas of streams. Katz (1967) found 3 genera of
elmids in her waterbyacinth root collections.
77. Dubiraphia quadrinotata (Say)--3 specimens in 2
collections from 2 counties.
Collections: 18, 103.
Maximum density: .032/plant: or 5/m2 at #~103.
Gyrinidae (Whirligig Beetles)
Only 3 specimens from 3 different species were collected
from w~aterhyacinth roots. While members of this family were
frequently present at the collecting site, they seem to pre-
fer patches of open water larger than that found between
waterhyacinth plants in a mat. O'Hara (1961) reports col-
lecting gyrinid larvae but no adults. Katz (1967) records
one genus of gyrinid but does not distinguish whether larvae
or adults or both were found.
78. Dineutes sp. larvae--1 specimen at Otter Creek,
Levy Co., in July.
Katz (1967) reports this genus from several of her col-
79. Gyrinus elevatus LeConte--single specimen collected
at Peace River, Hardee Co., in December.
Density: .063/plant or 5/m2
80. Gvrinus woodruf~fi Fall--single specimen collected
in October at Camps Canal, Alachua Co.
This species is considered to be primarily a stream
species, thus it would not be expected in the lentic habitats
that waterbvacinths refer.
Halip~lidae (Crawling Water Beetles)
Only 12 specimens in 3 different species in one genus
were found. While not common, they were not truly rare in
waterbyacinth roots. Since adults crawl and swim along the
bottom, they probably only occasionally hide in waterhyacinth
roots. Katz (1967) recorded Peltodyte sp. from 2 of her
collections. O'Hara (1968) found haliplids locally abundant
but recorded only the genus HIaliplus, which was not collected
either by Katz or myself.
81. Peltodytes dietrichi Young--3 specimens in a single
collection in July from Otter Creek, Levy Co.
This recently described species (Young 1961) was col-
lected only once.
82. Pel~todvtes floridecnsis Mlathesonl--G specimens from
3 collections at 2 locations in 2 counties.
Collections: 20, 21, 87.
Maximum density: .385/plant at #:21; 2.5/m2 at #87.
83. Peltodytes o~ppoiu Roberts--3 specimens from 3
locations in 2 counties.
Collections: 38, 71, 102.
Maximum density: .063/plant or 5/m2 at #102.
Hlelodidae (=Cyphionidae) (Ma~rsh Beetles) Larvae
The Helodidae, somnetimies also ktnownm as the Cyphonidae,
consisted of 138 Jarvae belonging to 2 genera and at least
3 species. In older literature, usually considered a part
of Dascillidae, the American helodid fauna is very poorly
known, with the best key to the species of adults being al-
most 100 years old (Horn 1880). Existing keys to larvae,
such as Leech and Chandler (1968), do not include all the
genera. My specimens were identified by rearing the larvae
and comparing the adults to old species descriptions and
identified material at The F~lorida State Collection of
Arthropods. While the adults enter the water only for ovi-
position or escape, the larvae were sometimes abundant in
closely packed, detritus-filled mats of wYaterbyacinths.
Katz (1967) collected specimens from this family but reports
Cyglgon, a genus which I did not find, and does not mention
the genera I collected.
84. Ora hvacintha Blatchley (adults reared from lar-
vae)--3 specimens reared from larvae from 2 collections at
a drainage ditch, Alachua Co.
Collections: 33, 37.
Maximum density: .333/plant.
Described by Blatchley in 1914 from adults found among
dead waterbyacinths on the shore, this species could be dis-
tinguished from my other species of Ora only in the adult
stage. The species name seems especially appropriate in
viewJ of the abundance of what are probably larvae of this
species in some waterbyacinth habitats.
85. Ora troberti (Guer.) (adults roared from larvae)--
3 specimens from 2 locations at a drainlage ditch in Alachua
Collections: 33, 37.
Maximum density: .333/plant.
The adults of this species were also obtained from
rearing collected larvae. While my specimens appear to fit
the description for this species, they may represent another
similar, yet undescribeed species.
Ora sp. larvae--90 specimens from 8 collections at a
drainage ditch in Alachua Co.
Collections: 22, 33, 37, 40, 43, 51, 54, 57.
Maximum density: .879/plant.
I could not distinguish the larvae OE 0. hyacinth
from the larvae of 0. trobert-i due to the overlap of almost
all characteristics studied. All Ora larvae were lumped to-
gether; even though the largest larvae were almost surely
0. Igacnta the smaller larvae could be either species.
Ora sp. larvae were associated with waterhyacinths growing
in deeper water. All my Ora larvae came from the same loca-
tion, a drainage ditch off of I-75 north of Micanopy in
Alachua Co., where they were the most frequently collected
and second most numerous insects found. All but 10 speci-
mens were collected in the fall and early winter months.
86. Scirtes spp. adults--single specimen from drainage
ditch- in Alachua Co. in September.
Collection : 33 .
A single Scirtes adult was found. Although adults are
not truly aquatic, since thley sometimes enter the water this
specimen was not excluded from my collection.
Scirtes sp. larvae--41 specimens from 13 collections
in 9 locations in 2 counties.
Collections: 11, 12, 13, 14, 51, 22, 33, 51, 54, 75,
89, 90, 97.
Maximum density: .545/plant or 30/m2 at #~90.
Insect associations: Telebasis byersi, r=-.886,
Scirtes larvae were found at many more different loca-
lions than Ora larvae although not as many specimens were
collected. Unlike the Ora larvae, most were collected
during the summer, although some were found during the winter
also. Scirtes larvae also appear to be associated with
waterhyacinths plants having shorter roots.
Hydraenidae (=Limnebiidae) (M~inute Moss Beetles)
These tiny beetles, less than 14- mm long, were never
abundant in my collections, and their size may have contri-
buted to their being overlooked in some collections. Only
one of the two species known from Florida was found and it
was represented by 12 specimens. Neither O'Hlara (1961) nor
Katz (1967) records any members of this family.
87. Hydraena marginiecllis Kiesenwetter--12 specimens
fromt 7 collections at 5 locations in 2 counties.
Collections: 2, 3, 21, 33, 34, 42, 53.
Maximum density: .15lt/plant.
Considered by many to be a subfamily of the hydro-
philid-s, I felt that Drs. Young, Hellman and others were
justified in considering the hydrochids a separate family.
Only 9 specimens belonging to 4 different species were col-
lected. All were identified by Dr. Hellman of the University
88. Hydrochus ainaqualis LeConte--5 specimens from 4
collections at Camps Canal, Alachua Co.
Collections: 38, 53, 91, 94.
Maximum density: .091/plant or 5/m2 a 94
89. Hydrochus prolatus Hellmian--single specimen col-
lected in July at Otter Creek, Levy Co.
90. Hydrochus simplex Leconte--single specimen collected
in July at Otter Creek, Levy Co.
91. Hydrochu woodi Hellman--2 specimens in 1 collec-
tion from Camps Canal in July.
Hydrophilidae (W~ater Scavenger Beetles)
With only 131 specimens, 17 of which were larvae,
hydrophilids were not one of my most abundant families.
While only 3 of the 13 species were represented by 10 or
more individuals, they were fairly well distributed, with
hydrophilids being found in 40 collections. O'Hara (1961)
reports collecting no larvae and that hydrophilid adults
were rare. Of the Hydrophilidae genera listed by Katz (1967)
from waterbyacinth roots, I collected: Paracymus, Enochrus,
Tropisternus, Phaenotum, Berosus, Helobata, Helochares and
Cren-itulus. I additionally collected Cercyon, Derrallus and
Hydrobiomorpha, while she also found Heloghorus and Laccobius.
92. B~erosus exiguus Say--2 specimens collected in July
at Otter Creek, Levy Co.
This species is usually collected from sand-bottomed
streams such as Otter Creek.
93. Cercyon praetextatus Say--single specimen collected
in Miay at Camps Canal, Alachua Co.
Density: .008/plant or 1/m2.
94. Crenitulus suturalis (LeConte)--18 adults from 4
collections at 3 locations in 2 counties.
Collections: 1, 21, 22, 33.
Maximum density: .225/plant.
This was my third most numerous hydrophilid, but two-thirds
of my specimens were in collection #:33, with the I-75 drain-
age ditch supplying 78%/ of the specimens. All were collected
from Junie tirouigh July.
Crenitulus suturalis (?!) larvae--1 larva collected in
August at Camps Canal, Alachua Co.
Density: .031/plant or 1.66/m2
This larva is different from any of the described
genera from Florida. Since Crenitulus, a monotypic genus,
is the only known genus of hydrophilid from Florida and
whose larva is not known, Dr. Paul Spangler at the Smith-
sonian believes it highly probable that the larva is of
95. Derrallus altus (LeConte) adults--6 specimens in
5 collections from 2 locations in 2 counties.
Collections: 17, 72, 79, 91, 94.
Maximum density: .182/plant or 10/m2 at #94.
Derrallus altus larvae--3 specimens in 3 collections
from 2 locations in Alachua Co.
Collections: 49, 50, 70.
Maximum density: .048/plant at #49; 1.67/m2 at #70.
96. Enochrus blatchleyi (Fall)--2 specimens in 2
collections from a drainage ditch in Alachua Co.
Collections: 40, 51.
Maximum density: .043/plant at #~51.
97. Enochrus ochraceus Melsheimer--29 specimens in 16
collections from 6 locations in 3 counties.
Collections: 3, 18, 30, 33, 41, 42, 43, 45, 70, 71,
86, 89, 91, 94, 95, 98.
Maximum density: .353/plant or 30/m2 at #70.
This species, a detritus pond inhabitant, was my second
most abundant hydrop~hilid and wras tied in the number of col-
lections in which it was found. All my specimens were col-
lected between M~ay and December, with over half coming from