Citation
Tadpoles and predators

Material Information

Title:
Tadpoles and predators patterns in space and time and the influence of habitat complexity on their interactions
Creator:
Babbitt, Kimberly Jane, 1962-
Publication Date:
Language:
English
Physical Description:
xvii, 135 leaves : ill. ; 29 cm.

Subjects

Subjects / Keywords:
Amphibians ( jstor )
Biomass ( jstor )
Breeding ( jstor )
Breeding sites ( jstor )
Ecology ( jstor )
Predation ( jstor )
Predators ( jstor )
Species ( jstor )
Tadpoles ( jstor )
Wetlands ( jstor )
Anura -- Habitat -- Florida ( lcsh )
Dissertations, Academic -- Wildlife Ecology and Conservation -- UF
Tadpoles -- Habitat ( lcsh )
Wetland ecology -- Florida ( lcsh )
Wildlife Ecology and Conservation thesis, Ph. D
Archbold Biological Station ( local )
Genre:
bibliography ( marcgt )
theses ( marcgt )
government publication (state, provincial, terriorial, dependent) ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1996.
Bibliography:
Includes bibliographical references (leaves 123-134).
Additional Physical Form:
Also available online.
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Kimberly Jane Babbitt.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
The University of Florida George A. Smathers Libraries respect the intellectual property rights of others and do not claim any copyright interest in this item. This item may be protected by copyright but is made available here under a claim of fair use (17 U.S.C. §107) for non-profit research and educational purposes. Users of this work have responsibility for determining copyright status prior to reusing, publishing or reproducing this item for purposes other than what is allowed by fair use or other copyright exemptions. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder. The Smathers Libraries would like to learn more about this item and invite individuals or organizations to contact the RDS coordinator (ufdissertations@uflib.ufl.edu) with any additional information they can provide.
Resource Identifier:
023779314 ( ALEPH )
35777920 ( OCLC )

Downloads

This item has the following downloads:


Full Text











TADPOLES AND PREDATORS: PATTERNS IN SPACE
AND TIME AND THE INFLUENCE OF HABITAT
COMPLEXITY ON THEIR INTERACTIONS


















By

KIMBERLY JANE BABBITT

















A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA

1996

UNIVERSITY OF FLORIDA LIBRARIES
































Copyright 1996

by

Kimberly Jane Babbitt


































To my father.














ACKNOWLEDGMENTS



Funding for this project was provided by the Institute

of Food and Agricultural Sciences at the University of

Florida, Archbold Biological Station, and a private donation

by Frank "Sonny" Williamson. Archbold Biological Station

provided housing, laboratory space, and use of a four-wheel

drive vehicle. The cooperation of the staff at Archbold

Biological Station and the MacArthur Agro-Ecology Research

Station is greatly appreciated.

I thank my committee members, Lyn Branch, Dick Franz,

Carmine Lanciani, and Carole McIvor for their help and

support throughout my project. I chose my committee members

with specific purpose in mind, hoping that their differing

specialties would provide a broad and balanced assessment of

my work. They fulfilled that purpose. I particularly want to

thank George Tanner for serving as my advisor. George

provided a supportive base from which to work. His easy-

going nature made working on a complicated project

relatively painless.

Natalie Hardman was hired as an assistant during my

first sampling season. She had a strong desire to learn

about amphibians. Unfortunately, the weather did not


iv









cooperate and we spent much of the summer wondering when the

rain would come. It never did. Natalie helped me with non-

field activities and kept her spirits up through the

disappointing field season. Her flexibility and good humor

were appreciated. Sam Gibbs assisted me at the beginning of

my first rainy season and was particularly helpful with trap

building and setting up initial wading pool experiments.

Vicki Drietz and Joan Morrison were very helpful in watching

for metamorphs from my experiments when other obligations

took me away from the ranch. Frank Jordan, Ken Portier, and

Christy Steible provided valuable statistical advice.

Jerrell Daigle was particularly helpful in verifying odonate

identification.

Ranch managers Dan Childs and Gene Lollis, and the rest

of the staff at Buck Island were a pleasure to work with.

They were particularly helpful in building a wetland

enclosure and a fence around my experimental arrays.

I particularly want to acknowledge the friendship of

Frank Jordan. I could not have asked for a better colleague

or friend. I hope we continue to collaborate in the future.

I want to thank my family for their support. In

particular, I want to thank my parents and grandparents for

introducing me to the great outdoors and encouraging my

interest in nature. Finally, I want to thank Leslie, whose

support and friendship over the past thirteen years has made

all of this possible.


v









This is contribution No. 26 from the MacArthur Agro-

ecology Research Center.




















































vi


















TABLE OF CONTENTS
page
ACKNOWLEDGMENTS... .......................................iv

LIST OF TABLES................................... ......... x

LIST OF FIGURES.............................................. xii

ABSTRACT ..................................... .............. xv

CHAPTERS

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

Overview ........................................... 1
Patterns of Distribution Among Wetlands............ 3
Patterns of Distribution Within Wetlands...........5
Predation, Habitat Structure, and Behavior..........8
Dissertation Structure.............................10

2 USE OF TEMPORARY WETLANDS BY ANURANS IN A
MODIFIED LANDSCAPE: EFFECTS OF HYDROLOGY
AND WETLAND SIZE ................................12

Introduction........................................... 12
Methods................................................. 14
Study Site.................................... 14
Sampling Methods ...............................15
Statistical Analyses............................16
Results............................................17
Species Richness and Abundance.................17
Breeding Phenology.............................. 18
Comparison of Breeding Activity Among Three
Summers.................................... 19
Discussion......................................... .20

3 SPATIAL AND TEMPORAL DYNAMICS OF TADPOLES AND
AQUATIC INSECT PREDATORS DEVELOPING IN
TEMPORARY WETLANDS .............................40

Introduction .......................................40
Methods......................................... 43
Study Site.................................... 43
Sampling Methods ...............................44

vii









Statistical Analyses............................45
Results...........................................46
Physical Parameters............................46
Tadpole and Predator Distribution.............. 48
Species-Specific Responses.....................49
Relationship to Physical Parameters............ 50
Relationship to Macroinvertebrates............. 51
Discussion.........................................51


4 EFFECTS OF COVER AND PREDATOR SIZE ON SURVIVAL
AND DEVELOPMENT OF RANA UTRICULARIA
TADPOLES.......................................74

Introduction......................................74
Methods........................................... 76
Experimental Design.............................76
Statistical Analysis............................79
Results ...........................................79
Survival.......................................79
Mass at Metamorphosis............................80
Age at Metamorphosis.............................81
Discussion.........................................81

5 EFFECTS OF FOOD AVAILABILITY AND RISK OF
PREDATION ON BEHAVIOR AND GROWTH OF RANA
UTRICULARIA TADPOLES........................... 88

Introduction.....................................88
Model Predictions................................91
Growth and Development..........................91
Activity.......................................91
Methods.......................................... 92
Behavioral Observations........................94
Response Variables and Statistics ..............94
Results..........................................95
Effects on Survival.............................95
Effects on Growth and Development.............. 95
Effects on Tadpole Activity and
Distribution.............................. 96
Discussion........................................97
Behavioral Responses .............................97
Growth and Development.........................100
Ecological Consequences of Responses..........102

6 EFFECTS OF COVER AND PREDATOR IDENTITY ON
PREDATION OF HYLA SQUIRELLA TADPOLES..........110

Introduction.................................. ... 110
Methods..........................................112
Results..........................................114
Discussion........................................114

viii









7 SUMMARY AND CONCLUSIONS ...........................118

LITERATURE CITED............... ..............*........... 123

BIOGRAPHICAL SKETCH...................................... 135

















































i















LIST 01F TIABLES
Table page

2-1 Occurrence and abundance of tadpoles at 12
temporary wetlands at MAERC .........................29

3-1 Split-plot analyses of variance testing for the
effects of wetland, zone, sampling period (time),
zone x wetland and zone x sampling period on
estimates of depth, temperature, pH, cover, and
plant biomass. The main effects of zone and wetland
were tested over the split-plot error (zone x
wetland), whereas other factors were tested over
the whole plot error (mean square error) ............58

3-2 Split-plot analyses of variance testing for the
effects of wetland, zone, sampling period (time),
zone x wetland and zone x sampling period on
transformed estimates of tadpole and predator
densities and biomass. The main effects of zone
and wetland were tested over the split-plot error
(zone x wetland), whereas other factors were
tested over the whole plot error (mean square
error) ..............................................59

3-3 Split-plot analyses of variance testing for the
effects of wetland, zone, sampling period (time),
zone x wetland and zone x sampling period on
transformed estimates of H. squirella, R.
utricularia, and P. ocularis densities. The main
effects of zone and wetland were tested over the
split-plot error (zone x wetland), whereas other
factors were tested over the whole plot error
(mean square error) ................................. 60

3-4 Correlations between all species, Hyla squirella,
Rana utricularia, and Pseudacris ocularis with
water depth, temperature, pH, plant cover, total
plant biomass, and macroinvertebrate density and
biomass. ............................................ 61

4-1 Summary of ANOVA for responses of Rana utricularia
tadpoles to cover and predator treatments. CD is
the coefficient of determination, which is the
percentage of variation in the response variable
attributable to the treatment effect................86








5-1 Summary of ANOVA for growth and larval period
responses of Rana utricularia tadpoles to food and
predator treatments. CD is the coefficient of
determination, which is the percentage of
variation in the response variable attributable to
the treatment effect............................... 106

5-2 Summary of ANOVA for activity and distribution
responses of Rana utricularia tadpoles to food and
predator treatments. CD is the coefficient of
determination, which is the percentage of
variation in the response variable attributable to
the treatment effect............................... 107








































xi














LIST OF FIGURES

Figure page

2-1 Map of the MacArthur Agro-Ecology Research Center
showing the system of ditches that enhance
drainage............................................ 30

2-2 Location of wetland study sites at MAERC............ 31

2-3 Abundance of tadpoles of anuran species at 12
temporary wetlands at MAERC. Bars are mean + 1 SE
and are based on total combined captures in all
0.25m2 traps at each wetland ........................ 32

2-4 Relationship between wetland size and anuran
species richness (top) and abundance (bottom).
Abundance and richness based on total captures
from all 0.25m2 traps within each wetland........... 33

2-5 Relationship between wetland hydroperiod and
anuran species richness (top) and abundance
(bottom). Abundance and richness based on total
captures from all 0.25m2 traps within each wetland..34

2-6 Pattern of tadpole occurrence during 17 months of
continuous sampling at 12 temporary wetlands at
MAERC ...............................................35

2-7 Monthly rainfall totals at MAERC for 1993-1995.
Total rainfall for each year was: 121.6 cm in
1993, 137.6 cm in 1994, and 157.6 cm in 1995........36

2-8 Mean (+ 1 SE) water depth at sampled wetlands.
The number above each bar indicates the number
of wetlands that had water during the sampling
period. (n=12 through 5/23/95 and 11 thereafter).... 37

2-9 Comparison of mean (+ 1 SE) tadpole abundance at
temporary wetlands during the summer breeding
season (June-September) during 1994 and 1995.
Abundance represents the combined captures in all
0.25m2 traps at each wetland ........................ 38




.ii








2-10 Comparison of total sample composition between the
1994 and 1995 summer breeding seasons at wetlands
that were (n=6) and were not (n=5)impacted by
flooding from ditches containing fish predators.
Species categorized as temporary site breeders
were species that do not breed in sites with fish
predators. Rana grylio, Rana utricularia, Acris
gryllus, and Hyla cinerea will breed with fish
predators and were categorized as permanent site
breeders............................................ 39

3-1 Variation in water depth (mean +1 SE) among
sampling dates (top), wetland (middle), and
wetland zone (bottom). Zone means with different
letters are significantly different..................62

3-2 Variation in water temperature (mean +1 SE) among
sampling dates (top), wetland (middle), and
wetland zone (bottom).................................63

3-3 Variation in pH (mean +1 SE) among sampling dates
(top), wetland (middle), and wetland zone
(bottom). Zone means with different letters are
significantly different........................... .. 64

3-4 Variation in plant cover (mean +1 SE) among
sampling dates (top), wetland (middle), and
wetland zone (bottom). Zone means with different
letters are significantly different................. 65

3-5 Variation in plant biomass (mean +1 SE) among
sampling dates (top), wetland (middle), and
wetland zone (bottom). Zone means with different
letters are significantly different..................66

3-6 Variation in tadpole numbers (mean +1 SE) among
sampling date (top), wetland (middle), and wetland
zone (bottom). Zone means with different letters
are significantly different .........................67

3-7 Variation in tadpole biomass (mean +1 SE) among
sampling date (top), wetland (middle), and wetland
zone (bottom)....................................... 68

3-8 Variation in macroinvertebrate predator numbers
(mean +1 SE)among sampling date (top), wetland
(middle), and wetland zone (bottom)................. 69

3-9 Variation in macroinvertebrate predator biomass
(mean +1 SE) among sampling date (top), wetland
(middle), and wetland zone (bottom). Zone means
with different letters are significantly
different............................................ 70

xiii








3-10 Variation in Hyla squirella numbers (mean +1 SE)
among sampling date (top), wetland (middle), and
wetland zone (bottom)................................71

3-11 Variation in Rana utricularia numbers (mean +1 SE)
among sampling date (top), wetland (middle), and
wetland zone (bottom)................................72

3-12 Variation in Pseudacris ocularis numbers
(mean +1 SE) among sampling date (top), wetland
(middle), and wetland zone (bottom) ................73

4-1 Mean (+ 1 SE) responses of Rana utricularia
tadpoles to cover and predator treatments: percent
survival (top), wet mass at metamorphosis (middle),
and age at metamorphosis (bottom). Squares indicate
high cover, circles indicate low cover..............87

5-1 Mean (+ 1 SD) responses of Rana utricularia
tadpoles to food level and predator treatments: age
at metamorphosis (top) and wet mass at
metamorphosis (bottom). Circles indicate predator
absent and squares indicate predator present. (n=6
for all treatments)................................108

5-2 Mean (+ 1 SD) responses of Rana utricularia
tadpoles to food level and predator treatments:
percentage of tadpole on the side of the container
opposite the predator (top) and percentage of
tadpoles that were active (bottom). Circles
indicate predator absent and squares indicate
predator present. (n=6 for all treatments).........109

6-1 Number of Hyla squirella tadpoles that survived
predation under different cover levels. Bars are
means + 1 SE (n=4 for each bar).................... 117
















X-iv















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

TADPOLES AND PREDATORS: PATTERNS IN SPACE
AND TIME AND THE INFLUENCE OF HABITAT
COMPLEXITY ON THEIR INTERACTIONS

By

Kimberly Jane Babbitt

August, 1996

Chairman: George W. Tanner
Major Department: Wildlife Ecology and Conservation


I examined the spatial and temporal dynamics of

tadpoles and aquatic insect predators in 12 temporary

wetlands in south-central Florida. I also conducted

experiments on the role of habitat structural complexity in

mediating predator-prey interactions. Eleven species of

anurans bred in the wetland sites. Rana utricularia

accounted for 44% of all captures. Species diversity was

positively correlated with wetland size but was unrelated to

hydroperiod. Varied meteorological conditions, coupled with

the presence of a system of drainage ditches, resulted in

annual variation in hydrologic conditions that had a large

effect on species composition. Flooding resulted in the

intrusion of water from ditches that contained fish into

otherwise isolated wetlands. Species that breed exclusively


xv








in sites without fish avoided these sites. Hydrologic

disturbance resulted in a system that is both spatially and

temporally dynamic resulting in variable assemblage

structure and composition at affected sites. Distribution

patterns of tadpoles within wetlands were related weakly to

habitat features. Both tadpole and aquatic insect predator

numbers varied among wetlands and across time, but were

distributed relatively evenly within wetlands. Thus,

selection for microhabitat at the within-wetland scale did

not appear to be strong. Distribution patterns of tadpoles

appear to be more strongly related to macrohabitat selection

of breeding sites by adults than by microhabitat selection

within wetland sites. Experiments demonstrated that

increased cover can reduce predation rates by aquatic

insects on tadpoles that reduce activity in response to

predators (R. utricularia), and those that do not (Hyla

squirella). Thus, although distribution patterns in the

field were not correlated strongly to habitat features,

tadpole success still may be influenced strongly by them.

Finally, reductions in activity in response to predators had

growth and developmental costs in R. utricularia that were

dependent on background resource level. Tadpoles on low food

metamorphosed at later dates but at larger sizes when

exposed to predators. Tadpoles on high food also had

prolonged development but were smaller compared to tadpoles

not exposed. Thus, assessment of responses to both resource


xvi









level and threat of predation may be necessary for

understanding the consequences of cost-benefit trade-offs

mediated by behavior.


















































xvii











CHAPTER 1

INTRODUCTION



Overview

One of the major goals in ecology is to determine the

factors that influence patterns of distribution and

abundance of species. This task is not a simple one because

many factors impinge upon the ecological and evolutionary

processes that result in observed relationships among

species. Resource availability and distribution change

along many spatial and temporal scales, and the biotic

interactions among species change similarly. The resulting

heterogeneous environment offers choices to the organisms

living within it. How an individual responds to habitat

heterogeneity may influence food availability and predation

pressure, and thus can have important influence on

population dynamics and ultimately evolutionary processes

(Holt 1987).

Examination of the factors influencing distribution

patterns of many anurans is complicated by their comple::

life cycles, which include a terrestrial adult stage and an

aquatic larval stage (Wilbur 1980). E::aminations of habitat

use and resource partitioning of anuran assemblages have

focused mostly on the adult stage (Bowker and Bowker 1979;

Crump 1971, 1982; Lizana et al. 1990; Toft 1985). Analyses


1











of larval assemblages usually address temporal partitioning

(Di::on and Heyer 1968; Weist 1981) or partitioning relative

to permanent versus temporary pond sites (Gascon 1991;

Woodward 1983). Most authors come to similar conclusions

regarding habitat partitioning of larval assemblages:

resources are partitioned along spatial (temporary versus

permanent ponds) and temporal (breeding season) a::es. Less

research has been done on larval food partitioning; however,

evidence suggests that partitioning along the food a::is is

not as strong as that along temporal and spatial a::es

(Duellman and Trueb 1986; Heyer 1974, 1976; Seale and

Beckvar 1980).

These studies provide important information regarding

possible selection pressures relative to habitat use and

resource partitioning (Crump 1982; Magnussan and Hero 1991).

However, they do not provide a complete picture of the

spatial ecology of tadpoles because they focus on one scale.

Distribution patterns of organisms can be e::amined on a

coarse-grained, macrohabi Lat scale u a fLiIne-grai nid,

microhabitat scale. Factors that can influence tadpole

success, such as competitor or predator abundance and

composition, vary in space not only among wetland sites

(Gascon 1992; Smith 1983; Werner and McPeek 1994; Woodward

1983) but also among patches within wetlands (Banks and

Beebee 1988; Diehl 1988, 1992). Therefore, the distribution

patterns of tadpoles at both the macrohabitat (breeding











site) and microhabitat (patch site) scale can have a large

influence on tadpole growth and survival, and thereby

assemblage structure.

Although the importance of considering scale has been

widely discussed (Wiens 1989, 1992), few studies of tadpole

ecology have addressed this issue. Processes that are

important population regulatory factors at one scale of

investigation may fail to operate at other scales (Wiens

1992). For e::ample, predation has been identified as an

important process influencing anuran breeding site selection

(Magnusson and Hero 1991). However, how predators and

tadpoles interact at the within-wetland scale is poorly

understood. Because ecological patterns are not independent

of scale, examination of spatial and temporal distribution

patterns at several scales should provide a more complete

understanding of the importance of a process, such as

predation, in regulating tadpole populations (Wiens 1989,

1992).



Patterns of Distribution Among Wetlands

Patterns of distribution of larval anurans among

aquatic sites are largely determined by two related factors;

the permanence of the site (i.e., hydrology) and the

composition of predators (Skelly 1995; Werner and McPeek

1994). These factors are strongly interrelated because

predator composition changes along the hydrologic gradient.









4

Fish predators dominate in permanent sites; however, systems

that have a cycle of filling and drying prevent predatory

fish populations from becoming established. A switch from

fish-dominated to invertebrate or salamander-dominated

predator systems occurs along the gradient so that tadpoles

developing in temporary and ephemeral sites are subject to a

different suite of predators.

Because biotic and abiotic conditions change along a

hydrologic gradient, tadpoles developing in different

wetlands face a different set of challenges. Traits that

increase developmental success at one end of the hydrologic

gradient may not be suitable at the other end. For e::ample,

interspecific variation in anti-predator mechanisms is

related to differences in breeding habitat (Azevedo-Ramos et

al. 1992; Kats et al. 1988; Woodward 1983). The larvae of

several species that breed in permanent sites are

unpalatable to fish, whereas species that breed in ephemeral

sites without fish do not possess this anti-predator

mechanism (Formanowicz and Brodie 1982; Kats et al. 1988).

In addition, behavioral adjustments, such as reduced

activity, are often absent in species that breed in

ephemeral sites (Kats eL al. 1988; Woodward 1983).

Decreased foraging activity may reduce growth or

developmental rates (Skelly 1992, 1995; Skelly and Werner

1990). At ephemeral sites, mortality threats from predators

are often not as significant as the threat from habitat











drying (Newman 1987; Rowe and Dunson 1993, 1995). Thus,

behavioral traits for reducing predation risk, while

beneficial in permanent habitats, can lead to increased

mortality risk at ephemeral sites.

The trade-off between rapid development and anti-

predator responses appears to be a major factor limiting the

distribution of anurans to certain types of breeding sites

(Skelly 1995; Werner and McPeek 1994). Weak anti-predator

responses that may increase developmental rates in ephemeral

sites would usually result in mortality in permanent sites

(Skelly 1994; Woodward 1983). As a broad generalization,

species that breed in permanent sites are passively e::cluded

from ephemeral sites by long larval developmental periods,

whereas species that breed in ephemeral sites are e::cluded

from permanent sites by predation.



Patterns of Distribution Within Wetlands

Few studies have e:amined tadpole distribution patterns

within breeding sites. Diaz-Paniagua (1987) e::amined the

spatial distribution of six species occurring in 16

temporary ponds. Dip-net samples were taken in 5 depth

zones. Vegetation structure varied along the depth gradient

from shallow zones characterized by dense grass cover to

deeper areas with little or no vegetation. Along the depth

qradient, ve(,etation chancqed from omerquecnt in slal l1ow water

to submergent in deep water. Few tadpoles of any species









6

were found in the inner non-vegetated zones. Although most

species were found in several of the vegetated zones,

interspecific differences in spatial distribution were

detected. Three species were most abundant in the shallow

zones, two species were most abundant in the emergent zone

and the remaining species was most common in the submergent

zone.

Larval assemblages also may separate along thermal

(Dupre and Petranka 1985) and water column (Heyer 1973,

1976) strata. Ontogenetic changes in microhabitat

association have been found relative to size class in Rana

utricularia (Alford and Crump 1982). Alford (1986) found

evidence for both intraspecific (i.e., size class) and

interspecific spatial partitioning among four anuran species

within a single temporary pond site. Two species, Pseudacris

ocularis and Pseudacris ornata, separated relative to cover

and water temperature, and their responses to these

environmental features changed over time. Alford (1986)

suggested that these responses were probably largely due to

competition (i.e., avoidance relative to species or size

class). However, he did not examine the potential influence

ou pLedation.

Differences in predator abundance or composition also

may exist within a breeding site. Although poorly studied in

the tadpole-predator system, the potential for spatial

segregation between predators and prey to be an important








7

factor within wetlands sites was demonstrated by results

from Banks and Beebee (1988) who found low spatial overlap

between Bufo calamita tadpoles and the invertebrate

predators Notonecta sp. and Dytiscus sp. Tadpoles were

concentrated in shallow water, whereas predators were

collected most frequently in deeper areas. Because few

studies have ex:amined spatial distribution of tadpoles

within breeding sites, it is not surprising that there is

limited information about the spatial relationships of

tadpole and predator distributions.

Spatial overlap at the within-wetland scale among

tadpoles and aquatic insect predators may be high if both

groups select habitats within the wetland based on similar

abiotic features such as water depth, temperature, or plant

cover. In contrast, differential selection criteria could

lead to segregation of tadpoles and predators. For e::ample,

females of some odonates require emergent vegetation for egg

laying (Corbet 1980). In this circumstance, tadpole survival

may be largely influenced by patterns of microhabitat use

and factors such as habitat structure. A better

understanding of how patterns of tadpoles and predators vary

in space and time, and how these patterns are related to

abiotic and biotic (e.g., vegetation structure) factors,

should provide valuable insight into the potential for such

patterns to generate differences in species richness and

assemblage structure among various sites.










Predation, Habitat Structure, and Hehavior

Predation can be a significant regulating lactul in

many systems (reviewed by Sih ct al. 1985). ::perimental

studies on tadpole assemblages have shown that difierential

predation on competitively domiinact specie; can alter

assembIlage st ructure (Morin I 1 l, 1, 1 1 Hi() (;nera I ly,

competitive ability in tadpoles is related to activity

levels: actively foraging tadpoles out-compete tadpoles that

are less active (Werner 1991, 1992a; Woodward 1982).

However, because prey activity is an Lmporl anLt compone]nt of

predator detection (Lima and DlI 19839; Weinrc and AniolL

1993), active tadpoles are more susceptible to predators

(Azevedo-Ramos et al. 1992; Skelly 1994; Woodward 1983).

Interspecific difterences in survival due Lo predati-on, and

the factors that influence predation rates are important in

explaining between-site or wi Lih n-siLe di leIt:i:nces in

species abundance patterns.

A plethora of studies on various aqudtic La::a i as

demonstrated the important role that habitat comple::ity

plays in mediating predator-prey interactions (reviewed by

Heck and Crowder 1991). Increased habitat comple:ity can

.tduce predat ion rates by prLovidilli C:uv(: Ml p I l i l i ltluge

areas for prey (e.g, Folsom and Collins 1984; Rozas and Odum

1988), or by decreasing foraging success of predators

because of decreased maneuverabil iLy or visual range (e.g.,

Crowder and Cooper 1982; Savino and Stein 1982; Werner et








9

al. 1983). Considering the important effect predators have

on tadpole survival and assemblage structuie, surprisingly

little research has focused on how habiLtat stLucture affects

Ladpole-predator interactions. Mu:rin (1981, 193, 19HG)

suggested that Pseudacris crucifer tadpoles survived better

than Scaphiopus holbrooki tadpoles wlhen sul ject to predation

by Notophthalmus viridescens by iestricting activity to the

bottom leaf litter. In contrast, S. holbrooki actively

foraged in the water column. Whether differences in

microhabitat use or activity levels (or both) were nure

important in determining vulnerability is not clear and has

not been demonstrated.

The results of the few studies that have e::amined the

effects of habitat structure on tadpole-predator

interactions are equivocal. In relatively simple

environments, Banks and Beebee (1908) found Llat even low

levels of habitat structure increased predation on tadpoles

by providing perch sites for odonate larvae. i.'igiel and

Semlitsch (1991) found that hab t compie::ity did i ot

affect crayfish (Procambarus acutus acuLus) predation on

artificially injured Hyla chrysoscelis. In a rare field

study, Jrcdl and Cullins (199:)) luuind L I atL lib itait

comple;:ity alone did not significantly affect larval

performance; however, interactions among habitat complex;ity,

predaLtor dJen;i ty, and prey d l.;nsity i nu 1 I l in
1 inear relat ionslhi ps.








10

F'inally, di Iferences in mi croh iLaL t struc:Lure miay noL

affect only overall predation raLes, buL also susce pt i ility

of different size classes. Hews (1995) found that small

tadpoles were more susceptible to predaLion by a lish

predator in a vegetated microlabitat whereas larger tadpoles

were more susceptible on a gravei substraLe.

The question of how ant -i -preda (tor r espoises aiffeclt

predation levels in environmenits that diLltr iln labit-aL

comple;ity would provide valuable information on how

tehavior aind abiotic leatures i ni a t to 1 i r min

susceptibility to predation. Foir e::ample, does couple::

habitat structure lead to decreased predation pressure on

actively foraging species that show lit tle responise to

predation, or do high activity Ievels lea l to hiigher

predation levels regardless of habitat structute? Similarly,

for species that do respond to predators, does cover provide

enhanced protection over that due to behavioral responses?

Experimental e:amination of thelse questions should provide

additional insight into the signiificance of patterns of

distribution found in the field.





The research I conducted ex:amined spatial and temporal

dynamics of larval anurans and aquaLic insect predators and

the effects of microhabitat structure on t hei r predator-prey

relationships. My research h1as three oleneota componniil s: ( I








11

ex:amination of the distribution patterns of larvae and

predators at 3 spatial scales (amoni wetlands, between

wetland zones, and within zone) across 3 summer and 1 winter

breeding season, (2) e;:perimenLal mianipulatLullo; o c::amine

the effect of microhabitat on predation, and (3)

e;:aminations of the effect of Larval predatL -avoidance

behavior on growth and development under countrolled

conditions.

I e;:amined spatial and temporal relationships of

tadpoles and aquatic insect predatols in severuL tLemporary

wetlands to determine whether species composition and

distribution patterns differed among wetlands (Chapter 2)

and within wetlands relative to physical and biological

characteristics in different patclhe (Chaptr 3) To ex:amine

the potential consequences of physical features in the

environment to tadpole survival, I conduct ed conltrol led

e;:periments to determine the rule of habitat comiple::ity in

mediating predator-prey interactilons (Chapters 1 and I6. In

a subset of these e::periments I e::amined whether there were

different effects based on the identity of Lhie predator

(Chapt ers 6) I a so ::ami llin wlin l ir p l oi ; : ':: I id aI l



I'in al ly, I examined t.iae behav in l responsel of R;Pan

ut. iculaiia tadpoles to piCeda l and tlie ell cet ol Ltdpule

responses on growth and development (Chapter 5).














CHAPTER 2
USE OF TEMPORARY WETLANDS BY ANURANS IN A MODI F'I ED
LANDSCAPE: EFFECTS OF HYDROLOGY AND WETLAND SlIZE


Introduct i.on

Patterns of distribution of larval anurans among

aquatic sites are largely determined by two related factors:

the permanence of the site (i.e., hydrology) and the

composition of predators (Skelly 1995; Werner and McPeek

1994). These factors are strongly interrelated because

predator composition changes along hydrologic gradients.

Fish predators dominate in permanent sites; however, systems

that have an annual cycle of filling and drying prevent

predatory fish populations from becoming establisled.

Interspecific differences in life hlisLtoy LraiiLs among

anurans, such as length of the larval developmental period

and ability to breed successfully with fish predators, limit

the range of wetlands within which a species can breed

successfully. Species with long larval periods are excluded

passively from sites with shoLt hydroperiods, whereas most

species that breed in temporary sites do not possess the

necessary chemical or behavioral characteristics lot

avoiding the heavy predation pressure fouind in perLmanent

sites (Kats et al. 1988; Woodward 1983). Because of

iiitcLspccilic diff ccII s amtolig m1 11 mll ; i ':;I, .,'11 i viil I

1,2







13

that alter wetland hydroperiod (the amount of time a wetland

holds water) can have a large effect on Lhe use of weLlanids

as breeding sites and the ability of the wetlands to

successfully produce metamorphs (Pechmanll et al. 1989).

Ditching of wetlands to enhance drainage can have a

largje effect on anuran use of wel land si te; (Vickers (L al.

1985). In many parts of southern Florida, wetlands have

been ditched and drained to make the nlaLurally wet sites

more suitable for agricultural activities. Under this

management activity, large tracts of land are subjecL to an

extensive series of interconnected ditches. Changes to

wetlands within the altered landscape include wetland loss,

reduced hydroperiod, and in some cases periodic conniction

to deeper water bodies that contain fish predators. The

value of these wetlands for anurans, or how anuran

assemblages may have changed as a result of these activities

has, to my knowledge, not been examined previously.

A majority of anuran species in Florida breed

e;:clusively or facultatively inl temporary wetland sites

(LaClaire and Franz 1990; Moler and Franz 1987). Because

temporary wetlands provide critical habitat for so many

allnu anl ill Flul ida, I e:xamiii(ed use o1 I III
a landscape modified to increase drainage to determine the

value of these altered wetlands ats breedingj sites Lot

anurans. I examnined anuran breedinq activities by sampling

anuran larvae at several temporary wetlands and asked








14

whether wetland hydrology or area affected use by anurans.

I'urther, I compared breeding activity durinq surimmer

periods that varied in summer rainfall paLLevns.



Methods

Study Site

Research was conducted at L he HMacAt Lhur Agro-lu- cu logy

Research Center (MAERC) in Lake Placid, Highlands CouniLy,

Fl'orida. The site is an active cattle ranch c-onsisting of

4,800 ha of improved and semi-improved pasture interspersed

with emergent wetlands and oak (Quercus virginiana) and palm

(Sabal palmetto) hammock.

The ranch was ditched e;:Lenisively in Lhte 1 IG.U's Lo

enhance drainage, and most wetlands are connected to the

vast system of ditches that ultimately enLers inLo ldainey

Pond Canal, which flows into Lake Okeechobee. (Figure 2-1)

Most wetlands are properly cliaracel' i z-ed ,Is eme gelt

freshwater marsh. Thick-steimmed emergents, such as

Pontederia cordata and SagitL Larc Ia i _nci o ia, are dominant

species in deeper water with Panicum hemitromon, Polygonum

punctatum and AlLeranthera -hi .lo::eroides as sub-dominianL

species. Outer, shallower areas of t.emporary wetlands ate

characterized by lower-stature emer(gents such as Bacopa

carol iniana, Hydrochloa carolinensis and Diodia virgiiiaiana.







15

Sampling Methods

From May 1994 through GSeteimber 19'j5, sampled I'd

wetlands every three weeks (23 sampling periods). One

wetland was dropped during the lastL s: sai iiipl in pe i oIds due

to damage from cattle trampling. In addition, I monitored

sites from May 1993 through August 1993 during a drought

period when no breeding took place. At each site, I

collected nine samples using a stlaLiiied aindou slampling

protocol. Three replicate samples were collected in each of

three concentric belts correspondii ng alpplo:ima tely Lo one

third the radius of the wetland.

Tadpoles were collected with a 1/4-rm open-ended, bo;:

trap. Bo:: traps have been used to sample small fish and

imacroinvertebrates (Chick et al. 1i-t; KFeeman eL al. 19-4;

Kushlan 1981) and anuran larvae (Caldwell et al. 1980; Calef

1973; Pfenning 1990), and aie e l ( eeL ive lot s iampll ing ill

areas with comple:: vegetation structure. A bar seine fitting

the diameter of the bo:; trap and Litted with Liberglass

insect screening was swept through the water column (after

vegetation had been cleared) until 3 consecutive sweeps

yielded no additional captures. Samples were preserved in

Ith.l I e( Id w ith ) IJ 1, bu L ti ti I niul i 1 ii I i ,, I I<

species in the lab.

linlormatiulo un other cai t L acLeo isL is Llih pLuteiLtialy

influence tadpole distribution a nd abundance wit hin wetlands

also was collected and is reported elsewhere (Chapter 3).









For the purposes of this study, the nine samples provided a

simple measure of breeding use (presencce/abeice) elatLive

abundance, and species richness at each site.

WeLland area was esLimaLed from data lreviously eLntered

into a Geographic Information System. Wetland area varied

from 0.11 to 2.21 ha (Figure 2-2) Water depth was Laken

from a permanent PVC stake at the center of each wetland at

Lhe time of sampling. Rainfall data were obtailed from an

all-weather gage at the research center. flydroperiod is

usually measured as the total number of days a wetland held

standing water. I took measurements every three weeks;

therefore, I do not have actual dates of wetland filling and

drying. During dry-down periods, I checked wetlands between

sampling periods to determine whether wetlands actually

dried completely. Therefore, I defined hydroperiod as the

number of weeks a wetland held standing water.



Statistical Analyses

Data for each site were treated as independent samples,

as were data from each summer breeding period. I used

Pearson Product Moment correlation to determine whether

:;(ecies i clin i' u Lu La1 1 i ul I li wi,<* I I < I I

wetland hydroperiod or size.

During 1995, some wetlands were affected by spill-over

of water from ditches containinq fish predators (e.g.,

Lepomis sp.). To ex:amine the effect of this disturbance on







17

assemblage composition, I used a chi-square goodness of fit

test to examine yearly differences in the number of tadpoles

that were from species that will or will not breed in sites

with fish predators. Because not all wetlands were affected

by this disturbance, I used the proportion of tadpoles in

each of the two categories (i.e., temporary breeding species

and permanent breeding species) in unimpacted wetlands to

generate expected values for tadpole numbe:rs in wetlands

that were impacted.



PesulIts

Species Richness and Abundance

A total of 3, 67 Ladpuoes I. Lu I 11 sl ci ,: s w"S

collected. Of the 11 species caught, only the southern

leopard frog (Rana utricularia) was found at every wetland

(Table 2-1). The squirrel treefrog (Hyla squirella) was the

ne;:t most widely distributed. In contrast, the oak toad

(Bufo quercicus) and southern chorus frog (Pseudacris

nigrita) were collected from only two wetlands each.

R. utricularia was the most abundant species,

accounting for about 44"o of all captures (Table 2-1; ligure

2-j) 11. squ i l I1a compt I sed I / u 1 Il' In(11 l[S, I he I t l e

grass frog (Pseudacris ocularis), pig ILog (Riana grcylIo) and

pinewoods treefrog (Hyla femoralAis) were moderately

abundant. The remaining species each accounted for 5' or

less of the total sample.








1lb

Anuran species richness varied fromt si : Lo ninie and was

positively correlated with wetland size (Figure 2-4; r=0.65,

p=0.023). However, the number of tadpoles captured was not

significantly related to wetland si e (Figure 2-4; r=0 35,

p=0.29) .

Most wetlands (n=9) dried only once: d(uinri)l( the Y 17 month

sampling period. The number of weeks that wetlands held

water varied from 49 to 65. Heither species richness

(r=0.03, p=0.93) nor number of individuals (r=0.40, p=0.22)

was related to wetland hydrology (i'igure 2-5). Further,

hydroperiod was not related to wetland size (r=0.177,

p=0.602) .



Breeding Phenology

Although the e;:act date of breeding among anurans

varies from year to year due to variability in

meteorological conditions, many species have particular

breeding seasons. During the 17 moinths of continuous

sampling in this study, si:; species; H. quer i us, 11.

squirel l a, IH. femoralis, H. grat iosea, H. cinerea, and the

eastern narrow-mouth toad (Gastrophryne carol inensis), bred

onl y sdur is t e su r iy eas (Figui e 2-I) I'. 1 i i l

tadpoles were collected only during the winter season. I

did, however, hear sporadic ca I Jing dJur.i.n sIiumnmer iUmonths

indicat i nq t-hat this species lmay broedr on a 1 im ii t:d hasis

(ut ni e th le i i ia n wi ntter b 'edl i ,seasoi Wi wt r .)1 ,i leaist







19

ten of the species collected bred during the sununer, only

four species; the cricket frog (Acris gryll us), R.

utricularia, P. nigrita, and I. ecu laris bred durin tlhe

winter. A. gryllus, P. ocula is and 1P. uLi iL u aI( i ad

tadpole occurrence patterns cons si :;t-(t wit- h year-r-ouid

breeding. Finally, Rana grylio bied during spring and

summer. Occurrence of tadpoles during winter months reflects

the long larval developmental peiiod of this species.



Comparison of Breeding ActiviLy Amoing 'Thliree Summers

Rainfall patterns differed amonii g (i thle llre sulmmer

periods (Figure 2-7). Low rainfall during the 1993 sunmmer

breeding season resulted in drought conditions and dry

wetlands throughout the entire summer. As a result, no

breeding occurred at any sites during the summ er breeding

season in 1993. Although summer l iaiiinall patLeLI s were

similar in 1994 and 1995, differences in antecedent rainfall

in winter and spring between the Lwo years (and higilhei

levels in Harney Pond Canal) resulted in a higher water

table at the beginning of the raidjy season in 1995. Whereas

all wetlands were dry for several months prior to summer



did so only one to two weeks prior to the 1995 summer rains

(F.igure 2-8). The consequence of L he diiietiLence ini waLer

table levels was that heavy rainfall at the beginning of the

1995 summer season resulted in ditches overflowing their








20

banks and spilling water into adjacent wetlands. Comparisons

of anuran breeding activity between 1994 and 1995 indicaLed

that species composition was affected by this flooding event

(Figure 2-9). H1. squirella and G. carolinelsis, two species

that breed e;clusively in temporary sites, were less

abundant in 1995 compared to 1994. 11n contrasL, species that

associate more strongly with permanent sites, such as R.

grylio and A. gryllus were more abundant in 1995. Wetlands

that were unimpacted by flooding had similar proportions of

tadpoles from species that will ou will not breed in sites

with fish predators (Figure 2-10). In contrast, wetlands

that were impacted by spill-over from ditches differed

greatly in composition between 1994 and 1995 (Figure 2-10;

;:'=1008, df=3, p<0.0001).



Discuss.i on

Changes due to ditching at MAERC have resulted in

changes in wetland hydrology, and in a probable shift in

assemblage structure and species composition that favors

anurans that associate more strongly with wetlands lacking

fish predators. Although historical data are not available

LO L coi p}U l .l i;,un, ii .at s; u I 1 i, I i1i 1 tI.iv' Ii IiuL, J i

patterns that are probably similar to pre-di.tching

conditions at MAERC such as the Northern Everglades, R.

grylio, A. grylluis, and H. cineroa dominate the tadpole

community (Babbitt et al., unpubl. data). These species are







21

often associated with longer hydroperiod sites that contain

fish predators, and were probably much more abutidain Lhan

they are now (Carr 1940; Duellman and Schwartz 1958).

Although each of these species bred in some of the wetlands

I sampled, they concentrated most of their breeding activity

in the larger ditches and an atLificial lake (pers. obs.).

Most of the wetlands on site, including those sampled

during this study, are intermediate in hydroperiod and are

best described as annually or semi-annually drying temporary

wetlands. By viewing habitat drying as a d isLurbance, we

would predict that the temporary wetlands at MAERC would

have the highest species richness (compared to permanent or

very ephemeral sites) based on predictions from the

intermediate disturbance hypothesis (sensu Connell 1978)

Both Heyer et al. (1975) and Wilbur (1984) argued that the

highest species diversity would be at intermediate values

along a hydrologic gradient. Only one species that occurs at

MAERC, the gopher frog (Rana capito), was not collected

during the study. This species has an e;;tremely limited

distribution on-site, but it does breed in Lemporary

wetlands similar to those used in the study (pers. obs.).

T'hu.:;, 1 J ai Liit ;p (i.'; : es [ l i
wetlands that were intermediate in hydroperiods. In

comparison, only a sub-set of anuran species at MAERC breed

in permanent wetlands or wetlands with very short

hydroperiods.









The long-duration temporary wetlands used in this study

appear to provide suitable breeding habitat Lor all anurans

occurring at MAERC. What precludes most anurans from

successful development in permanent sites is predation by

fish. Because these temporary ponds dry on an annual (or

semi-annual) basis, the removal of the fish predatoj

assemblage makes these sites suitable for species that avoid

permanent ponds. At the same Lime, all species in south

Florida that breed in permanent ponds have larval

developmental periods that are less than a year, and

therefore could successfully reach metamorphic size in many

temporary wetlands (Ligas 1960; pers. obs.).

Although temporary wetlands fall within the range of

breeding habitats that all anuLran species at MAERC

considered suitable, they probably do not necessarily

provide the "best" breeding habitat for all species. For

e;xample, B. quercicus and P. nigrita were collected at only

two wetlands. Both species utilize ephemeral sites and

appeared to concentrate breeding activities at sites that

were of shorter duration than the wetlands I sampled (pers.

obs.). Considering the varying life history requirements of

lll- Lt dlpu l ul d ill lct L :;tj 'i i:;, I ,:; 'v I ,I

time, presence or absence of anti-predator mechanisms, and

competitive ability, long-duration temporary wetlands are

probably suitable, but marginal habitat, for some species.

In contrast, these wetlands may provide "primee" bltedinlg







12

habitat for R. utricularia. R. utricularia was the most

abundant species and had the broadest dist ibution. This

species also has been found to be numerically dominant or

abundant in censuses at other Lemporary weLland sites in

Florida (Dalyrumple 1988; Enge and Marion 1986; O'Neill

1995; Vickers et al. 1985) .

Viewed at the landscape level (i.e., MAERC) the

ditching of MAERC probably increased anuran species

richness. The effects of wetland alterations on anurans at

MAERC differ from those that occuL at many uLher sites

because this site was altered from a large (usually)

permanent aquatic mosaic similar to the Everglades to a

terrestrial system containing numerous wetlands of varying

hydroperiods. Reductions in hydroperiod at sites that are

already temporary can result in decreased species richness

or abundance because truncated hydroperiods preclude some

species from successfully metamorphosing (Pechmann et al.

1989; Rowe and Dunson 1995; Semilitsch 1987; Wilbur 1987)

The effects of ditching of wetlands on anuran species

richness and breeding success depend on botl the original

and resultant conditions of a site.

[In this sLudy, neitlie- spi' 'S I i11us) n 1 L (ipui(A

abundance was related to wetland hydroperiod. These results

contrast with those of Pechmann et al. (1989) who found that

hydroperiod was positively correlated with species richness

and the abundance of metamorphosing amphibians. In their









2/1

study, hydroperiods varied almost by a factor of five (i.e.,

5d to 263 days). Hydroperiods in tLhe curreiL study varied by

only a factor of 1.4, from 49 to 65 weeks. lurther, the

hydroperiod of these wetlands was much longIe than Lhe

minimum required for metamorphosis of all the anuran species

occ:urrrinq at MAERC (Iiqas 1)60; Wi lt r 1 ),"/) I i a ive-year

study, Dodd (1992) did not find a relationship between

habitat duration and species richness at a single temporary

wetland in north-central Florida. That study was conducted

during a drought, and visitation by adult aiuLrans was not

necessarily associated with breeding activity. Large

differences in hydrology among sites (or amu ng years) may be

required to generate significant differences in species

richness. Relevant differences may include sites with

hydroperiods that are too short for some species to

successfully produce metamorplis, and permaneiti sites with

fish predators. It is likely that the wetlands in the

current study provided similar habil t suit abi l ity, relative

to hydrology, and that any variation in use patterns by

adults was largely related to other factors.

Species richness was positively related to wetland

keL5 I. Kesult.s f rom studies thi, t hLave p::amini nc d t he

relationship between wetland size and species richness or

abundance are equivocal. Whereas some researchers have found

a positive relationshi p between wet land area and species

richness (Kutka and Bachmann 1990(; haanl alnd Ve, boom 1990),







) L

others have found no relation at all (Diaz-Paniagua 1990;

Laan and Verboom 1990; Richter and Azous 199) Again,

differences in these studies may represent dif ferences in

scale, as well as differences in other characteristics that

influence use by adults (Lann and Verboom 1990; Richter and

Azous 1995). Because species richness at a particular site

is a community metric and therefore a product of the varying

responses of community members, it may not be surprising

that the above studies provide differing results.

Within a particular class of wetlands, such as the

temporary wetlands with relatively long hydroperiods sampled

in this study, year-to-year cihanges in hydiology driven by

meteorological conditions may have a larger influence on

breeding use than many other factors. The largest influence

on anuran use of wetland sites at MAERC was amount and

pattern of rainfall. During Lthe summer droughL condi Lions of

1993, no breeding took place in any temporary wetlands

because they were dry. In 1994, essentially all non-

permanent sites provided suitable habitat for species that

breed in temporary wetlands because all sites dried down

prior to the summer rains. This eliminated any fish

predators that would have made these siLos unisui table Lor a

majority of species. These two years contrasted further with

1995, when some wetlands dried completely before the summer

rains but others did not completely dry. 'urther, over-spill

from adjacent fish-containing ditches apparently caused some








26

anurans to avoid using sites that otherwise would have been

suitable.

Although the mechanism through which adult amphibians

are able to identify sites that contain fish is not known,

experimental choice studies have demonstrated that adults

are capable of differentiating between sites that do and do

not contain fish predators (Hopey and Petranka 1994;

ResiLarits and Wilbur 1991). Differences in species

composition among sites with and without fish predators are

due (at least partly) to adult selection, not simply

predation on the tadpoles of certain species.

The impacts of over-spilt of water from F[ish-containing

ditches into otherwise isolated wetlands are transient

because wetlands usually dry on an annual basis. Further,

alternative sites are available for anurans that avoid

breeding in sites with predatory Lish. F'or e;:ample, ialtlough

species such as H. squirella and G. carolinensis avoided

breeding in wetlands affected by over-spill Irom fish-

containing ditches, these species bred in adjacent sites

that were unaffected (pers. obs.).

In contrast, low rainfall can result in missed breeding

uppul LuniiLie-s ad ( us Llie -imiL. ds La .i u:;< u w I I ,l i ,i Idi 11

e::acerbates the dry conditions. buring seasonal droughts,

the only species that can successlully breed are Lhose that

breed at permanent sites with Fish prtjedators. Whereas

periodic seasonal droughts may have only a small ilmpact on










population numbers, the effects of long-term drought may be

more profound. Few studies have examined amphibian

populations during drought, and the comple:: and erratic

patterns of population number of amphibians make assessment

of the impacts of drought difficult (Dodd 1992). However,

some species appear to be capable of surviving long-term

droughts (Dodd 1995). Evidence that at least some

individuals within a population may be opportunistic in

selecting breeding sites, rather than philopatric, suggests

that this may be an important mechanism for maintaining a

breeding population during drought conditions (Dodd 1995).

Opportunistic use of breeding sites also appeared to be

important in areas affected by flooding during 1995 (pers.

obs.). Particularly in landscapes such as Florida where

wetlands are numerous, opportunistic use of breeding sites

may be more important than previously recognized (also see

Dodd 1995).

These data suggest that temporary wetlands on this site

provide dynamic habitats that offer varying breeding

opportunities and larval developmental conditions that are

highly dependent on meteorological conditions. The effect

of Llis is a spatially-LLempo)iLli Iy dyinamic syseLm 1rsultiny

in differing assemblage structure and composition at

particular sites. Because MAERC contains numerous wetlands

and ditches that vary in hydrology from ephemeral to

permanent, most species can find suitable breeding habitat










except under extreme drought conditions. During such

conditions, ditching exacerbates the drought conditions by

increasing drainage, lowering the water table, and

prolonging dry conditions.

In conclusion, historical changes to the landscape have

probably increased local anuran species richness and altered

tadpole assemblage structure at the wetlands at MAERC. An

important aspect that was beyond the scope of this study is

the effects of past and current land management on the

amount and distribution of varying terrestrial habitats. The

surrounding upland matrix can have a large influence on

wetland use as breeding sites, particularly for amphibians

that have a terrestrial adult phase (Kutka and Bachmann

1990; Laan and Verboom 1990). For example, arboreal species

such as H. femoralis appeared to be locally abundant, but

limited in distribution. Temporary wetlands near forested

hammocks consistently contained H. femoralis tadpoles;

however, wetlands away from hammocks in large open pastures

did not. Consideration of upland characteristics is an

important aspect of overall conservation of anurans.

Research on the relationship between upland habitat patch

characteristics and wetland use would increase our

understanding of breeding use patterns, as well as our

ability to provide guidance for managing anuran populations.










Table 2-1. Occurrence and abundance* of tadpoles at 12 temporary wetlands at MAERC.
Wetland
1 2 3 4 5 6 7 8 9 10 11 12 Total


Bufo ouercicus 0 0 0 14 0 0 0 0 0 0 0 19 33
Hyla cratiosa 1 0 1 2 0 0 0 2 15 11 2 9 43
Acris gryllus 0 0 5 10 61 8 23 51 1 5 2 25 191
Hyla squirella 81 173 60 26 1 168 3 6 20 48 48 0 634
Hyla cinerea 52 12 7 0 0 0 3 1 14 18 22 0 129
Hyla femoralis 0 0 0 129 73 1 3 0 0 0 0 20 224
Pseudacris nigrita 0 0 0 0 0 44 10 0 0 0 0 0 54
Pseudacris ocularis 0 1 2 21 43 105 99 17 0 5 1 18 313
Gastrophryne carolinensis 16 73 12 0 0 0 4 1 6 1 26 0 139
Rana grylio 7 7 1 6 216 0 2 5 1 17 0 40 303
Rana uaricularia 398 160 35 92 38 19 315 348 68 7 101 22 1604
Uniden:ified Hyla 1 0 0 1 1 0 0 0 0 8 0 0 11

Total 556 426 123 300 432 345 461 431 126 120 202 153 3678

*Number of individuals captured in all 0.25m traps combined within each wetland









30













25MI ) TD 25(IK 75(x
CA 1









































Figure 2-1. Map of the MacArthur Agro-Ecology Research
Center showing the system of ditches that enhance drainage.









31










S-- -=*-*
A 'ALTE K ,1:50x

i -



i -llii iiii-: I







0


\ 0
S2dq










^===.=-------












Figure 2-2. Location of wetland study sites at MAERC.









180 -

160 -

140 -

S120 -
0


o
S100 -
4*-
0
a 80 -

:= 60 -
z w

40-

20 -



00


Species


Figure 2-3. Abundance of tadpoles of anuran species at 12 temporary wetlands at MAERC.
Bars are mean + 1 SE and are based on total combined captures in all 0.25m- traps at
each wetland.








33
10


9 -
a,


4-

.0 7 -
E
=3
Z
6- 0 *


5 ---------------------

0.0 0.5 1.0 1.5 2.0 2.5

Wetland Size (ha)


600 -

500 -


"0
cL 400 -
I-
S300-
0

n 200 *
z 0
100 -

0-
0 -
0.0 0.5 1.0 1.5 2.0 2.5

Wetland Size (ha)
Figure 2-4. Relationship between wetland size and anuran
species richness (top) and abundance (bottom). Abundance and
richness based on total captures from all 0.25m' traps
within each wetland.








34
10


9- *
O.


0
1-

S 7- *
E
z
6- 0* *


5 _- I -I- i I i
48 50 52 54 56 58 60 62 64 66

Wetland Hydroperiod (wks)

600 -

500 -

o. 400-

300 -

S200 -

z 0
100 -

0 I I I I
48 50 52 54 56 58 60 62 64 66

Wetland Hydroperiod (wks)
Figure 2-5. Relationship between wetland hydroperiod and
anuran species richness (top) and abundance (bottom).
Abundance and richness based on total captures from all
0.25m' traps within each wetland.













Summer Breeding Season Winter/Spring Breeding Season Summer Breeding Season



Bufo quercicus


Hyla gratiosa

Acris gryllus


Hyla squirella

Hyla cinerea


Hyla femoralis

Pseudacris nigrita -


Pseudacris ocularis


G. carolinensis


Rana grylio


Rana utricularia


6116 7T11 8/1 8/22 9/12 10/3 10/25 11/16 12/8 12/26 1/17 2/9 2/27 3/22 4/10 5/3 5/23 6/12 7/3 7/24 8/14 9/4 9/26
1994 1995






Figure 2-6. Pattern of tadpole occurrence during 17 months of continuous sampling at 12
temporary wetlands at MAERC.












40


35 -1993
c- 1994
+---1995
30 -


25


u 20
%4--


15


10


5




JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC

Month


Figure 2-7. Monthly rainfall totals at MAERC for 1993-1995. Total rainfall for each
year was: 121.6 cm in 1993, 137.6 cm in 1994, and 157.6 cm in 1995.









50-

12
12

40 11
12 12 T 121
12 12
12r 1 i 11
12 12 1 12 1
30 -



o 20 8


7 11

10 -


2
S 4/28 5/5 5/26 6/9 6/16 711 8/1 8/22 9/12 10/3 10125 1116 12/8 12/26 1/17 2/9 Z'27 3/22 4/10 5/3 5/23 6/12 6/19 7/3 7/14 8/14 914 9/26
1994 1995
Date




Figure 2-8. Mean (+ 1 SE) water depth at sampled wetlands. The number above each bar
indicates the number of wetlands that had water during the sampling period. (n=12
through 5/23/95 and 11 thereafter).









80 -
1994
70 1 1995


60 -
C'

o 50 -
50

40 -
0
E 30 -
=3 0
z
20 -


10- I


o cja -ET D--

Z, \00


Species
Figure 2-9. Comparison of mean (+ 1 SE) tadpole abundance at temporary wetlands during
the summer breeding season (June-September) during 1994 and 1995. Abundance represents
the combined captures in all 0.25m' traps at each wetland.








39



700 -
Temporary
I 1 Permanent
600 -



500 -

(.
0-
"3 400-
I-
'4- 1 |
O
300-
E
Z
200 -



100 -


0
1994 1995 1994 1995

Impacted Unimpacted






Figure 2-10. Comparison of total sample composition between
the 1994 and 1995 summer breeding seasons at wetlands that
were (n=6) and were not (n=5) impacted by flooding from
ditches containing fish predators. Species categorized as
temporary site breeders were species that do not breed in
sites with fish predators. Rana grylio, Rana utricularia,
Acris gryllus, and Hyla cinerea will breed with fish
predators and were categorized as permanent site breeders.















CHAPTER 3
SPATIAL AND TEMPORAL DYNAMICS OF TADPOLES AND AQUATIC INSECT
PREDATORS DEVELOPING IN TEMPORARY WETLANDS


Introduction

Predation is a major source of mortality in anuran

tadpoles (Calef 1973; Heyer et al. 1975; Smith 1983), and is

thought to have a large influence on breeding site selection

by adult anurans (Magnusson and Hero 1991). Fish are major

predators of tadpoles in permanent aquatic sites, and many

species appear to avoid breeding in sites that contain fish

predators (Hopey and Petranka 1994; Resetarits and Wilbur

1991). Although ephemeral wetland sites generally do not

contain fish predators, the short duration of these sites

makes them unsuitable for species with long larval

developmental periods (Wilbur 1980, 1984, 1987). Thus

differences in habitat duration and predator composition,

coupled with interspecific differences in the selection of

breeding sites by adults, can generate differences in

species composition of larval anurans along gradients of

habitat permanence (Werner and McPeek 1994).

However, factors that influence tadpole success, such

as abundance and composition of competitors or predators,

vary in space not only among wetland sites (Gascon 1992;

Smith 1983; Werner and McPeek 1994; Woodward 1983) but also

40









41

among patches within wetlands (Banks and Beebee 1988; Diehl

1988, 1992). Therefore, the distribution patterns of

tadpoles and predators at both the macrohabitat (breeding

site) and microhabitat (patch site) scale can have a large

influence on tadpole growth and survival, and thereby

assemblage structure.

Although the importance of considering scale has been

widely discussed (Wiens 1989, 1992), few studies of tadpole

ecology have addressed this issue. Processes that are

important in population regulation at one scale of

investigation may fail to operate at other scales (Wiens

1992). For example, whereas predator composition has a large

influence on anuran breeding site selection, the spatial

dynamics of predators and tadpoles at the within-wetland

scale are poorly understood. If tadpoles occupy patches

that provide increased protection from predators, such as

areas with high cover (Babbitt and Jordan 1996; Chapters 4

and 6), or areas with lower predator densities (Banks and

Beebee 1988), then differential habitat use within sites may

be an important feature influencing predation rates, and

possibly assemblage structure (Morin 1986).

Spatial overlap at the within-wetland scale among

tadpoles and predators may be high if both groups select

microhabitats based on similar abiotic features such as

water depth, temperature, or plant cover. Understanding

such patterns, and how they vary in space and time, should








42

provide valuable insight into the potential for differences

in microhabitat selection to generate differences in species

richness and success among various sites. For example,

occupation of vegetated shallow areas of permanent aquatic

sites that have deeper, poorly vegetated open-water areas

may provide increased protection from fish predators (Diehl

1992; McIvor and Odum 1988; Werner et al. 1983). How

distribution patterns influence predation rates on tadpoles

developing in temporary wetlands where aquatic insects are

the major predators is largely unknown.

Several studies have found interspecific differences in

tadpole distribution within wetland sites (e.g, Alford 1986;

Diaz-Paniagua 1987; Heyer 1976); however, few studies have

related tadpole distribution to predator distribution (Banks

and Beebee 1988). The lack of research on within-site

distribution patterns of tadpoles and aquatic insect

predators leaves open the question of whether differential

distribution patterns within wetlands is a potential

mechanism for decreasing predation pressure on tadpoles.

The goal of this study was to determine the spatial and

temporal patterns of anuran larval and aquatic insect

predators within several temporary wetland sites. I asked

whether tadpoles or predators were differentially

distributed in three zones based on water depth. Further, I

asked whether habitat features influenced tadpole

distribution to determine whether tadpoles differentially








43

occupied patches based on physical or chemical habitat

features. By examining distribution patterns within

wetlands, I was able to determine whether differences in

distribution could function as a mechanism, in addition to

adult breeding site selection, for decreasing predation

pressure on larval anurans.



Methods

Study Site

Research was conducted at the MacArthur Agro-Ecology

Research Center in Lake Placid, Highlands County, Florida.

The site is an active cattle ranch consisting of 4,800 ha of

improved and semi-improved pasture interspersed with

emergent wetlands and oak (Quercus virginiana) and palm

(Sabal palmetto) hammock.

The ranch was ditched extensively in the 1960'S to

enhance drainage, and most wetlands are connected to the

vast system of ditches that ultimately enters into Harney

Pond Canal, which flows into Lake Okeechobee. Most wetlands

are properly characterized as emergent freshwater marsh.

Thick-stemmed emergents such as Pontederia cordata and

Sagittaria lancifolia, are dominant species in deeper water

with Panicum hemitomon, Polygonum punctatum and Alteranthera

philoxeroides as sub-dominant species. Outer, shallower

areas of temporary wetlands are characterized by lower-

stature plants such as Bacopa caroliniana, Hydrochloa

carolinensis and Diodia virginiana.









44

Sampling Methods

From May 1994 through September 1995, I sampled 12

wetlands every three weeks (n=23 sampling periods). One

wetland was dropped during the last si:: sampling periods due

to damage from cattle trampling. At each site, I collected

9 samples using a stratified random sampling protocol. Three

replicate samples were collected in each of 3 concentric

belts corresponding approximately to one third the radius of

the wetland.

Tadpoles and macroinvertebrates were collected with a

1/4-mn' open-ended, box trap. Bo:: traps have been used to

sample small fish and macroinvertebrates (Chick et al. 1992;

Freeman et al. 1984; Kushlan 1981) and anuran larvae

(Caldwell et al. 1980; Calef 1973; Pfenning 1990), and are

effective for sampling in areas with complex vegetation

structure. A bar seine fitting the diameter of the box trap

and fitted with fiberglass insect screening was swept

through the water column (after vegetation had been cleared)

until 3 consecutive sweeps yielded no additional captures.

Samples were preserved in the field with 10% buffered

formalin. Macroinvertebrates were removed from formalin,

rinsed thoroughly in water and stored in alcohol in Lhe lab.

Tadpoles and macroinvertebrates were identified to species.

For the purposes of this study, the macroinvertebrate

assemblage was analyzed as a whole, rather than by species.

The assemblage consisted of 11 species of odonates from the








45

families Aeshnidae, Libellulidae and three species of

Hemiptera from the families Belostomatidae and Naucoridae.

Water depth was measured to the nearest centimeter

inside each trap prior to removal of plants. Foliar coverage

of above-ground vegetation was estimated ocularly, and then

plants were removed and shaken to dislodge any macrofauna.

Grass clippers were used to clip plants as necessary to

remove all vegetation. Vegetation was placed in a mesh bag,

which was spun in the air 20 times to remove e::cess water

prior to measuring biomass to the nearest 0.1 kg. Water

temperature ('C) and pH measurements were taken adjacent to

each trap at 0.1 m below the water surface.



Statistical Analyses

Split-plot analysis of variance (ANOVA) was used to

test for effects of wetland site, sampling period (time),

wetland zone and the wetland :: zone and sampling period ;:

zone interactions on the habitat features water depth, water

temperature, pH, cover, and plant biomass, and on tadpole

and macroinvertebrate density and biomass. In addition,

separate analyses were done on the three most abundant

anuran species. Values from the three samples taken within

each zone were averaged to calculate a mean value for each

zone for all parameters. Tadpole and macroinvertebrate

density and biomass values were log transformed (log (::+1))

prior to analysis to reduce skewness. Examination of








46

residuals for the measures of environmental parameters

indicated that transformations were not necessary (Sokal and

Rolf 1981). The main effects of wetland and zone were tested

over the split-plot error (i.e., wetland : zone), whereas

the time and time x zone effects were tested over the whole

plot error (i.e., mean square error). The Bonferonni-Dunn

post hoc multiple comparisons tests was used to test for

differences among the three zones when a significant

(p<0.05) zone effect was detected (Day and Quinn 1989).

Correlation analyses were performed to e::amine the

relationships between the abundance of all tadpoles

combined, as well as the abundance of the three most common

anurans (H. squirella, R. utricularia, and P. ocularis),

with habitat variables and macroinvertebrate abundance and

biomass (Sokal and Rolf 1981).



Results

Physical Parameters

Water depth, water temperature, pH, plant cover, and

plant biomass all varied significantly among wetlands and

across time (Table 3-1; Figures 3-1 through 3-5). The main

effect of time, which reflects seasonal di f ereices,

accounted for most of the variation in water depth (37%) and

water temperature (58%). In contrast, wetland explained a

majority of the variation in pH (73%), plant cover (36%),

and plant biomass (29%). Wetlands that were adjacent to








47

upland hammocks tended to have lower pH values compared to

wetlands that were surrounded by pasture. Wetland

differences in plant cover and biomass indicate differences

in plant communities, as well as differences in the areal

extent of coverage of thicker-stemmed emergents.

Significant differences among zones were detected for

water depth, pH, plant cover and plant biomass, but not for

water temperature (Figures 3-1 through 3-5). Both water

depth and pH increased progressively from the outer to the

inner zone. Whereas the zone effect explained 22% of the

variation in water depth, it accounted for only 1% of the

variation in pH.

Cover levels were higher in the outer wetland zones

compared to the middle and inner zones (Figure 3-4). Cover

in the outer zone was nearly 100% at most wetlands

throughout the study. Plant biomass also differed among

zones; however, in contrast to cover where highest levels

were in the outer zone, plant biomass was highest in the

inner zone (Figure 3-5). Whereas difference among zones

accounted for 20% of the variation in cover, it accounted

for only 5% of the variation in biomass.

Few interactions were significant, and most that were

explained very little of the variation in the ANOVA model.

The significant zone : wetland interaction for plant cover

and biomass was due to differences among wetlands in the

extent of the development of the inner thick-stemmed

emergent zone.








48

Tadpole and Predator Distribution

Tadpole densities varied significantly among wetland,

time, and wetland zone (Table 3-2, Figure 3-6). Wetland and

time explained 9% and 28% of the variation in tadpole

density, respectively. Densities were higher in the outer

zone compared to both the inner and middle zones, but did

not differ significantly between the inner and middle zones

(Figure 3-6). However, zone explained less than one percent

of the variation in tadpole densities. Such differences are

probably biologically insignificant. In contrast to density,

tadpole biomass did not differ significantly among zones

(Table 3-2, Figure 3-7). Differences among wetlands

explained 9% of the variation in tadpole biomass and time

explained 23%. Although tadpole numbers were highest at the

beginning of each summer breeding period, tadpole biomass

was highest during the winter months. This difference

reflects the dominance of the relatively larger tadpoles of

R. utricularia in winter samples.

The pattern for macroinvertebrate predators differed

from that of tadpoles. Whereas zone differences were found

for tadpole density but not biomass, the opposite was true

lor predators (Table 3-2). Macroilnvui Lli.daL, Iiumb IIis did n1ot

differ among zones; however, macroinvertebrate biomass was

higher in the middle zone compared to the outer zone.

(Figures 3-8 and 3-9). Again, although zone differences in

biomass were significant, they e:plained less than one








49

percent of the variation in biomass. In contrast to tadpole

density and biomass where time e::plained most of the

variation, wetland explained a higher percentage of the

variation in macroinvertebrate number and biomass than did

time. Wetland accounted for 34% and 24% of the variation in

macroinvertebrate density and biomass, respectively, whereas

time explained 11% and 17%. Although the numbers of

macroinvertebrates varied significantly among wetlands,

wetland size was not correlated with predator numbers

(R=0.258, p=0.44).



Species-Specific Responses

Three species, H. squirella, R. utricularia, and P.

ocularis, were abundant enough to e::amine separately.

Wetland and time had siqnificant pfffcts on density of each

species (Table 3-3; Figures 3-10 through 3-12). For H.

squirella, time accounted for a majority of variation in

tadpole numbers (40%). Wetland e::plained only 4% of the

variation in H. squirella numbers. The main effects of

wetland and time explained 18% and 16%, respectively, of the

variation in R. utricularia densities. These effects each

explained 12" of Lhe variation in P. ocular is densiiies. R.

utricularia and P. ocularis breed throughout the year, so it

is not surprising that time explained less of the variation

in density of these two species compared to H. squirella,

which is strictly a summer breeder. Differences in the








50

pattern of distribution among zones were not significant for

any of these species.



Relationship to Physical Parameters

Although significant relationships were found between

tadpole numbers and many environmental variables, these

relationships were very weak (Table 3-4). The strongest

relationship was a negative correlation with water depth.

More complex patterns emerge, and some stronger

relationships were detected, when the three most abundant

species are examined separately. H. squirella had a strong

negative relationship with water depth and a strong positive

relationship with temperature and pH (Table 3-4). In

addition, H. squirella was negatively correlated with plant

biomass. R. utricularia also had a negative relationship

with water depth but this relationship was weak. In contrast

to H. squirella, R. utricularia was negatively correlated

with temperature. The differences between these two species

reflect the interspecific differences in seasonality of

breeding activity. R. utricularia also was positively

correlated with plant cover. P. ocularis numbers were not

correlated significantly with any of the environnmcntal

variables measured.








51

Relationship to Macroinvertebrates

Tadpole numbers and macroinvertebrate numbers were not

significantly related, and only a weak negative relationship

was found between tadpole numbers and macroinvertebrate

biomass (Table 3-4). No significant relationships were found

between macroinvertebrate number or biomass and any of the

three species examined separately.



Discussion

Patterns of tadpole abundance within wetlands provided

little evidence that differential occupation within wetland

zones was an important anti-predator mechanism. Although

differences in tadpole abundance among wetland zones were

statistically significant, these differences explained only

one percent of the total variation in tadpole numbers.

Examination of the entire tadpole assemblage as a group may

obscure some species-specific patterns that may e;ist;

however, separate analyses of the three most abundant

species did not reveal any differences in distribution

relative to wetland zone. Because predator numbers and

biomass also were distributed relatively equally among

zones, differential use by tadpoles presumably would not

have been an effective mechanism for avoiding predators.

In addition to the lack of differences in distribution

based on wetland zone, tadpoles showed weak relationships

with habitat characteristics. Examination of the entire










tadpole assemblage together revealed non-significant or weak

correlations. However, patterns among the three species

e::amined separately indicated that the distribution of at

least one species, H. squirella, was related to some

environmental parameters. This species was negatively

correlated with depth and plant bionass, and positively

correlated with water temperature and pH. The positive

correlation with temperature was due to e::clusive summer

breeding rather than differences within wetlands (which were

not significant). Further, the positive correlation with pH

was due to differences among wetland sites, rather than

differences within wetlands. Both water depth and plant

biomass were significantly lower in outer wetland zones,

suggesting that the negative correlation between these

parameters and H. squirella numbers was due to use of outer

zones. However, ANOVA indicated that H. squirella numbers

were not related to zone. Because H. squirella breeds

shortly after wetland fills from rainfall, use of shallow

wetlands and wetlands with less-developed zones of thick-

stemmed emergents probably contributed to the negative

correlations. Thus, correlations of H. squirella with

habitat features appeared to be largely related to early

summer breeding and adult selection of wetland breeding

sites.

R. utricularia numbers were related only to three

variables, cover (positive), water depth (negative) and








53

temperature (negative). Similar to H. squirella, differences

may largely be related to differences among wetland sites,

since R. utricularia numbers also did not differ

significantly among zones. The negative correlation with

temperature reflects winter breeding activity. Finally, P.

ocularis numbers were not related to any variables measured.

The results suggest that within a class of wetland,

(i.e., temporary), predator avoidance may not be related to

microhabitat selection, at least at the scale examined in

this study. At the scale of wetland site, Gascon (1991)

found that species-specific responses to habitat features at

several temporary wetland sites were highly variable and

were not consistent from year to year. Thus, whereas species

may breed only in temporary sites, they may use sites that

are poorly vegetated or well vegetated (see also Heyer

1976).

Some studies that have e:xamined distribution patterns

of tadpoles have found differential distribution within

sites. Generally, the patterns observed have been higher

occupation of vegetated areas and lower occupation of deeper

open-water areas or areas with very low cover (Alford 1986;

Banks and Beebee 19UU; DidaZ-L'dliicyqud 91'/; luluimuiuki 19U9;

Loschenkohl 1986). Because wetland sites in the current

study were vegetated throughout, differences in results with

previous studies may indicate that large differences in

habitat features, such as cover or no cover, may be








54

necessary for generating differential distributions.

Although cover levels did vary significantly among zones,

these differences may not have been large enough to affect

microhabitat use.

Observed patterns of tadpole distribution may be

generated by a variety of processes, including differential

predation, tadpole selection of microhabitats, and adult

selection of oviposition sites. The relative role of these,

or alternative mechanisms, in generating observed patterns

is unclear. Unfortunately few studies that have examined

tadpole distribution have also examined predator

distribution. An exception is work by Banks and Beebee

(1988) who found that both predatory dytiscid beetle larvae

and Bufo calamita tadpoles were non-randomly distributed

relative to water depth and vegetation, with tadpoles

concentrating in shallow water and beetle larvae being more

common in non-vegetated, deeper water areas.

An aspect of microhabitat use not addressed in this

study is differentiation based on water column strata.

Interspecific differences in use of the water column have

been observed or inferred from tadpole morphology (e.g.,

lieyei 1S*/S3; l'LcLtsuii eL dl. lI( t 1;.;u V I (_W I I Wi t)ur

1980). Such interspecific differences often have been

interpreted as a mechanism for avoiding exploitation

competition; however, regardless of the mechanism generating

this use pattern, such differences could result in








55

differential predation pressure. For e:ample, although

aeshnid odonates such as Ana junius cling to vegetation,

many libellulid odonates are benthic (Pritchard 1965). Thus,

the foraging strategies of the macroinvertebrates could

generate interspecific differences in predation on tadpoles

in patches with differing predator composition.

Many odonates and anurans oviposit shortly after

wetlands fill with water; therefore, tadpoles developing in

varying parts of a wetland may face similar levels of

predation pressure. For these species, the susceptibility to

predators will largely be a function of growth rates.

Tadpoles that grow rapidly can effectively eliminate

predation threats from all but the largest predators (Crump

1984; Richards and Bull 1990; Travis et al. 1985a; but see

Crump and Vaira 1991). Such a mechanism is probably

important for species such as H. squirella. Thus, although

occupation of areas with high cover may decrease predation

rates on H. squirella tadpoles (Babbitt and Jordan 1996;

Chapter 6), the timing of breeding, avoidance of sites with



for successful tadpole development in this species.

Although breeding early may decrease predation pressure

on some species, and may explain why the tadpoles of these

species do not show strong microhabitat use, R. utricularia,

which is mainly a winter breeder, did not use this strategy.

It bred in sites that contained both abundant number of








56

predators as well as large predators (Figures 3-8 and 3-11;

pers. obs). However, like H. squirella, this species did not

show strong patterns of microhabitat use. Thus, lack of

microhabitat selection to avoid predation was not limited to

early breeding species. Therefore, early breeding offers

only a partial explanation for the lack of microhabitat

selection. More likely, the broad and even distribution of

the predators may make such a strategy ineffective in

wetlands that do not vary greatly in habitat features such

as plant cover.

Although tadpoles did not e::hibit strong correlation

with midiiy lldbiLdL ciLdLcLL L L ic::;, LhliL Iu;i nolt I ,i

occupation of patches that differ does not affect tadpole

growth and survival (Holomuzki 1986a; Sredl and Collins

1992; Travis and Trexler 1986). Experimental studies have

demonstrated the importance of increased cover in mediating

predator-prey interactions in aquatic systems (reviewed in

Heck and Crowder 1991), including iLteractions beLween

tadpoles and aquatic insect predators (Babbitt and Jordan

1996; Chapters 4 and 6).

The lack of strong responses to habitat features within

wetlands contrasts with the large change in composition

observed in responses to the introduction of water from

ditches containing fish (Chapter 2). The shift from an

assemblage dominated by species that breed in temporary

sites to one dominated by species that will breed in sites








57

containing fish predators was interpreted as being due to

adult breeding site selection (Hopey and Petranka 1994;

Resetarits and Wilbur 1991). Thus, predator composition is

important for determining breeding use at the macrohabitat

level. Although abiotic habitat characteristics may

determine use by some species, features within sites that

fall within a "suitable" range for breeding may not be

important in determining species use (Gascon 1991). In

addition, use patterns within sites may not be strongly

influenced by predators, e:xcept in sites with clear

partitioning of habitat features (i.e., cover versus not

cover) or predators (i.e, present/absent or

fish/macroinvertebrates).

This study indicated that avoidance of sites with fish

predators, breeding shortly after temporary ponds fill, and

rapid development, appear to be the major anti-predator

mechanisms used by anuran species that breed in temporary

ponds, and that within-wetland habitat selection was of

minimal importance. A better understanding of the potential

role of microhabitat use as an anti-predator strategy could

be gained by examining more closely the size distributions

of predator and prey (e.g., Werner et al. 1983), as well as

the foraging strategies of predators compared to the

positional use by tadpoles.











58


Table 3-1. Split-plot analyses of variance testing for the
effects of wetland, zone, sampling period (time), zone x
wetland and zone x sampling period on estimates of depth,
temperature, pH, cover, and plant biomass. The main effects
of zone and wetland were tested over the split-plot error
(zone x wetland), whereas other factors were tested over the
whole plot error (mean square error).




ource df SS M::


Wetl] and 11 123'2 4 1 1 0 1 .
none x Wetland 2 21s 5.7 8.0 41.88' 0 01
Time 2 34 00. l 3. b3.
2,-Tie :: T im4 / '4 .,-; '* 1 i 1 .': ,;i.
l ':dJuO l *-I 1] ,1l'. 7 1

TeImpe ra t re



1one 1 .3 3 i1 7
Wet land 1 1 ]588.8 5 1 P 11 i -i1
; ,. ] l :; W i t l in r: .

;,ie y Time 11 5 .. 1 3 1
R, :^i, ] [A, ,1. 8.



So urce d SS M3 F F


Wet land 1 1 35. 1 3 9 .
, e ;< Wetland 22 0.9 a. i 4.

:i dual 10..1 1
Ii (111a I


'over
0014 E



.one 8 53.3 4 H. 14 .1
Wet i d 11 1'6 85. 6 1 8."71 .1 l .. ,, i. ri-i p R
W. :: Itlan 1 1 1 11

',one '. Time 441 1 4. / 1 ^. : 1. 11 0.:5834
Residiual 94 89949.2 1 1.1

T ,t a l int PBi i

Source f SS F


Wetland 11 17054247.1 1 7 I. 3
one x Wetland 5075688 4 11 2 7. CnO]
T11 1 1.
ei ,, ". I l I I 1 .
Residuall] '4 21039148 .2 1 i11" 4









59

Table 3-2. Split-plot analyses of variance testing for the
effects of wetland, zone, sampling period (time), zone ::
wetland and zone :: sampling period on transformed estimates
of tadpole and predator densities and biomass. The main
effects of zone and wetland were tested over the split-plot
error (zone x wetland), whereas other factors were tested
over the whole plot error (mean square error).

Tadpole Density

Source df SS MS F P

Zone 2 0.540 0.270 7.010 0.0044
Wetland 11 6.849 0.623 16.162 0.0001
Zone x Wetland 22 0.848 0.039 0.529 0.9631
Time 22 21.351 0.971 13.314 0.0001
Zone x Time 44 3.308 0.075 1.031 0.4191
Residual 594 43.299 0.073

Tadpole Biomass

Source df SS MS F P

Zone 2 0.136 0.068 1.704 0.2051
Wetland 11 4.491 0.408 10.234 0.0001
Zone x Wetland 22 0.878 0.040 0.734 0.8060
Time 22 11.503 0.523 9.619 0.0001
Zone x Time 44 1.599 0.036 0.668 0.9511
Residual 594 32.288 0.054


Macroinvertebrate Density

Source df SS MS F P

Zone 2 0.078 0.039 0.379 0.6687
Wetland 11 21.740 1.976 19.257 0.0001
Zone x Wetland 22 2.258 0.103 1.920 0.0072
Time 22 7.332 0.333 6.234 0.0001
Zone x Time 44 1.282 0.029 0.545 0.9930
Residual 594 31.756 0.053


Macroinvertebrate Biomass

Source df SS MS F P

Zone 2 0.046 0.023 4.972 0.0165
Wetland 11 1.548 0.141 30.601 0.0001
Zone x Wetland 22 0.101 0.00U 0.806 0.7195
Time 22 1.123 0.051 8.951 0.0001
Zone x Time 44 0.217 0.005 0.864 0.7208
Residual 594 3.387 0.006








60

Table 3-3. Split-plot analyses of variance testing for the
effects of wetland, zone, sampling period (time), zone x
wetland and zone x sampling period on transformed estimates
of H. squirella, R. utricularia, and P. ocularis densities.
The main effects of zone and wetland were tested over the
split-plot error (zone x wetland), whereas other factors
were tested over the whole plot error (mean square error).


H. squirella

Source df SS MS F P

Zone 2 0.042 0.021 3.246 0.0582
Wetland 11 0.813 0.074 11.398 0.0001
Zone x Wetland 22 0.143 0.006 0.348 0.9978
Time 22 8.084 0.367 19.726 0.0001
Zone x Time 44 0.332 0.008 0.405 0.9998
Residual 594 11.065 0.019


R. utricularia

Source df SS MS F P

Zone 2 0.072 0.036 1.028 0.3743
Wetland 11 7.187 0.653 18.767 0.0001
Zone ; Wetland 22 0.766 0.035 0.915 0.5746
Time 22 6.248 0.284 7.468 0.0001
Zone x Time 44 3.272 0.074 1.955 0.0003
Residual 594 22.592 0.038

P. ocularis

Source df SS MS F P

Zone 2 0.044 0.022 2.011 0.1578
Wetland 11 1.240 0.113 10.243 0.0001
Zone x Wetland 22 0.242 0.011 0.928 0.5581
Time 22 1.205 0.055 4.619 0.0001
Zone : Time 44 0.161 0.004 0.308 1.0000
Residual 594 7.047 0.012








61

Table 3-4. Correlations between all species, Hyla squirella,
Rana utricularia, and Pseudacris ocularis with water depth,
temperature, pH, plant cover, total plant biomass, and
macroinvertebrate density and biomass.


r p

Water depth vs. H. squirella -0.432 0.0008
R. utricularia -0.197 0.0020
P. ocularis -0.051 0.6752
All species -0.230 <0.0001


Temperature vs. H. squirella 0.391 0.0027
R. utricularia -0.244 0.0002
P. ocularis 0.152 0.2254
All species 0.122 0.0017

pH vs. H. squirella 0.386 0.0030
R. utricularia 0.150 0.0557
P. ocularis -0.181 0.1948
All species 0.130 0.0025

Plant biomass vs. H. squirella -0.295 0.0257
R. utricularia -0.014 0.8218
P. ocularis -0.026 0.8292
All species -0.069 0.0691

Plant cover vs. H. squirella 0.132 0.3291
R. utricularia 0.254 <0.0001
P. ocularis 0.190 0.1155
All species 0.151 0.0001


Macroinvertebrate H. squirella -0.269 0.0668
density vs. R. utricularia -0.035 0.6014
P. ocularis -0.073 0.6139
All species -0.073 0.0530


Macroinvertebrate H. squirella -0.029 0.8477
biomass vs. R. utricularia -0.100 0.1344
P. ocularis -0.245 0.0838
All species -0.136 0.0003







62
40-
35-
S30 -
S25
0c
)- 20
15-




Sampling Period
35
30-
E 25 -

c. 15 -
S10 -
5 -
0 I I-- --- --- ----- L

1 2 3 4 5 6 7 8 9 10 11 12
Wetland

30- c
25 b
E 20-

Q.
a 10-
5
0
0 ------------------ --
Outer Middle Inner
Wetland Zone
Figure 3-1. Variation in water depth (mean +1 SE) among
sampling dates (top), wetland (middle), and wetland zone
(bottom). Zone means with different letters are
significantly different.







63
32-
& 30-
o, 28-
a 26-
4 24
L 22-
0 20
E 18
16-
14 I1 I I 1 i ll I I i i


Date

30
-0 25- -
^ 20 -
15 -
_0 10-
U, 5
0
1 2 3 4 5 6 7 8 9 10 11 12
Wetland

30
00 25
S20 -
S15 -
10
E
0 5-
I-
0
Outer Middle Inner
Wetland Zone

Figure 3-2. Variation in water temperature (mean +1 SE)
among sampling dates (top), wetland (middle), and wetland
zone (bottom).







64
7 -


6 -





4


Date

7 -
6
5 -
4 -
3 -
2-
1 -
0
0 ----------------U------
1 2 3 4 5 6 7 8 9 10 11 12
Wetland

6- a b c
5-
4-
I 3-
2-
1-
0
Outer Middle Inner
Wetland Zone
Fiqure 3-3. Variation in pll (mean -1 SE) among samp]li nq
dates (top), wetland (middle), and weLland zone (bottom)
Zone means with different letters are significantly
different.







65
100
S90-
0 80-
E 70-
i 60 -
0- 50
4 0 I I 1 I I I I I I I I I I I
4("b\ C'b c V

Date
120 -
100-
O 80-
S60 -
0)
2 40 -
0- 20 -
0
1 2 3 4 5 6 7 8 9 10 11 12
Wetland

100 a

,0 80 b
> r ~T b
0 60-

) 40-


0
a. 20
0- _
Outer Middle Inner
Wetland Zone
Figure 3-4. Variation in plant cover (mean +1 SE) among
sampling dates (top), wetland (middle), and wetland zone
(bottom). Zone means with different letters are
significantly different.







66
1000 -
900 -
NE 800-
N 700
0
o 600-
500 -
400 -


Date
1200 -
1000 -

E 800 -
cM 600 -
o) 400 -
200 -
0 ---
1 2 3 4 5 6 7 8 9 10 11 12
Wetland

800 b
700 ab
600 a
E 500-
e( 400-
o 300-
S200 -
100
0
Outer Middle Inner
Wetland Zone
Figure 3-5. Variation in plant biomass (mean +1 SE) among
sampling dates (top), wetland (middle), and wetland zone
(bottom). Zone means with different letters are
significantly different.








67
14
E 12-
(C 10 -
0 8
70 6
O)

Q. 4
-p- 2

110 9 < h \ '0 (

Date

5-
E 4-
LO
3

0 2-


0
1 2 3 4 5 6 7 8 9 10 11 12

Wetland

3-
i a
E b
S2
0 b

(- 1
"o
05--


0
Outer Middle Inner

Wetland Zone
Figure 3-6. Variation in tadpole numbers (mean +1 SE) among
sampling date (top), wetland (middle), and wetland zone
(bottom). Zone means with different letters are
significantly different.







68
4-

3-
E 2-


0-
I I I I I I I I I I I I
L(
(N






Date

3-


E 2- T


0)


0-
1 2 3 4 5 6 7 8 9 10 11 12

Wetland

1.5-

T T
"E 1.0- T
io

0.5-


0.0
Outer Middle Inner

Wetland Zone

Figure 3-7. Variation in tadpole biomass (mean +1 SE) among
sampling date (top), wetland (middle), and wetland zone
(bottom).








E 69
Lo 6

0 5
4
3

a 2


0
11 ry rV (1 .b (
Date

E 10

0)


82


o 0-

S1 2 3 4 5 6 7 8 9 10 11 12

Wetland

E
04-



S2-

:)
>1
I-




Wetland Zone

Figure 3-8. Variation in macroinverLebraLe predator numbers
(mean 41 SE) among sampling date (top), wetland (middle),
and wetland zone (bottom).








70
0.6-
0.5-
E 0.4 -
(N 0.3 -
S0.2 -
0.1-
0 .0 1 N 1 I I I I I I I I I 1 I I I I I I I 4


Date

1.0
0.9-
0.8-
(c 0.7-
E 0.6-
1 0.5
o 0.4-
M 0.3 -
0.2 -
0-1 F- r1
0.0
1 2 3 4 5 6 7 8 9 10 11 12
Wetland

0.35 b
0.30 -ab
a
0.25- a
E 0.20-
c 0.15-
oC 0.10-
0.05 -
0.00
Outer Middle Inner
Wetland Zone
Figure 3-9. Variation in macroinvertebrate predator biomass
(mean +1 SE) among sampling date (top), wetland (middle),
and wetland zone (bottom). Zone means with different letters
are significantly different.








71
6-


(0 4
0
n3

co 2 -


0-


Date

2-
E
in 1


0





1 2 3 4 5 6 7 8 9 10 11 12

Wetland

0.5-
E 0.4 -



0.0
I-




Outer Middle Inner
1 2 3 4 5 6 7 8 9 10 11 12
















Wetland Zone


Figure 3-10. Variation in Hyla squirella numbers (mean +1
SE) among sampling date (top), wetland (middle), and wetland
zone (bottom).
zone (bottom)








72
6
NE 5


0
(" 3







Date


4 -

E
i 3-
Oc
S2
0,
o



O.
-0 1 I





Wetland


2-
(1




o T
CN T


0 3

0
0 -----------------








Outer Middle Inner

Wetland Zone


Figure 3-11. Variation in Rana utricularia numbers (mean +1
SE) among sampling date (top), wetland (middle), and wetland
zone (bottom).








73
2.0-

u0 1.5 -
(N
O15-

1.0 -
0
C.
S0.5-

0.0 -


Date

1.00 -

L 0.75-
(N

S0.50 -
a)
0
- 0.25 -

0.00 F'
1 2 3 4 5 6 7 8 9 10 11 12

Wetland

0.3
E

S0.2 -
0


0o 0.1 -
"o
(-

0.0
Outer Middle Inner

Wetland Zone

Figure 3-12. Variation in Pseudacris ocularis numbers (mean
+1 SE) among sampling date (top), wetland (middle), and
wetland zone (bottom).














CHAPTER 4
EFFECTS OF COVER AND PREDATOR SIZE ON SURVIVAL AND
DEVELOPMENT OF RANA UTRICULARIA TADPOLES



Introduction

Size-limited predation is a particularly important

process during amphibian development. Predation levels in

systems with size-limited predators are largely dependent on

the relative sizes of predator and prey (Ebenman and Persson

1988). Numerous studies have shown that predation rates on

tadpoles are a function of tadpole body size, and a majority

of mortality due to predation occurs early in development

(e.g., Banks and Beebee 1988; Cronin and Travis 1986;

Richards and Bull 1990, Semlitsch 1990, Semlitsch and

Gibbons 1988; Travis et al. 1985a). Thus, it has been

suggested that selection for rapid growth during the larval

phase may, at least in part, be a response to predation by

size-limited predators (Travis 1983, Travis et al. 1985b).

For anurans breeding in temporary aquatic habitats,

oviposition shortly after the breeding site fills provides a

potentially important mechanism for decreasing predation

pressure. For example, because dragonflies, a major predator

of tadpoles in temporary breeding sites, also oviposit after

pond filling (Ward 1992), tadpoles that grow rapidly often


74








75

are able to reach a size refuge and decrease their

cumulative risk of predation. Because a majority of

predation on tadpoles occurs early in development, the

"growth race" between dragonfly naiads and tadpoles can have

a particularly large influence on the ability of predators

to regulate tadpole populations (Caldwell et al. 1980;

Tejado 1993; Travis et al. 1985a).

Not all species that use temporary wetlands employ this

reproductive strategy, however. The southern leopard frog

(Rana utricularia) breeds year round; however, the majority

of breeding is done during the winter-spring breeding

season. In addition, this species has very broad breeding

site associations. This can result in larvae being exposed

to an assemblage of aquatic insect predators that are

already established, and therefore large in size relative to

newly hatched tadpoles. Thus, early predation pressure by

large predators may be significant. Under such

circumstances, physical features of the environment, such as

complex habitat structure, may provide an important

mechanism for lowering predation rates.

Habitat structural comple:ity has been shown to be an

important factor mediating predator-prey interactions

(reviewed in Heck and Crowder 1991). Increased habitat

complexity can reduce predation rates by providing cover or

p rl.t ial j rerugc aroas Fos prey ( .1d, lI1 sum niid u Coll iis 1984;

Rozas and Odum 1988), or by decreasing foraging success of








76

predators because of decreased maneuverability or visual

range (e.g., Crowder and Cooper 1982; Savino and Stein 1982;

Werner et al. 1983). R. utricularia tadpoles can grow to a

large size relative to most aquatic insect predators, thus

mechanisms that decrease early predation pressure may have a

particularly important effect on overall survival rates.

In this study I exposed tadpoles of the southern

leopard frog to two levels of cover and two size classes of

the same predator. By doing so I could determine whether

increased habitat complexity (i.e., cover) provides a

mechanism for decreasing predation and further whether such

protection is dependent on predator size. I predicted that

predation rates would be lower at high cover levels and that

large predators would be more effective than small predators

regardless of cover level. In addition, because the thinning

effects of predators can release prey from competition

(Wilbur 1987, 1988), I predicted that tadpoles developing in

treatments with higher predation rates (i.e., low cover and

large predator) would have enhanced growth.



Methods

Experimental Design

I examined development and survival of tadpoles in a

factorial experiment in which I manipulated cover level

(high versus low) and predation (large, small, or no

predator) in a complete randomized block design replicated








77

four times. I established a rectangular array of 24, 1.14 m

diameter plastic wading pools with tight fitting fiberglass

screen tops at the MacArthur Agro-Ecology Research Center

(MAERC) in Lake Placid, Highlands County, Florida. On 8

March 1995, I filled pools with 95 1 of well water (depth =

12 cm) and added 1000 ml of well-mi::ed algae and zooplankton

collected from several wetlands. I added an additional 500

ml of algae and vegetation (see below) on 16 March 1995.

Water overflow was prevented by a drainage pipe. I checked

pools weekly to determine if water levels were maintained at

the proper level by rainfall. When rainfall was not adequate

to compensate for evaporation, I added well water (3 times

during the experiment).

I provided two levels of cover, 2000g or 500g (wet

weight), of the aquatic plant Hydrochloa carolinensis. H.

carolinensis is a thin-leafed grass that provides comple:

structure throughout the water column. It is abundant in

wetlands at MAERC. Both treatments provided cover throughout

the water column; however, the higher cover level provided

relatively dense cover whereas the low cover treatment

provided relatively sparse cover. Predator treatments were

large (mean I 1 SE: 4.18 4 0.26 cm; i = 6) or small (2.35 1

0.11 cm; n = 6) Tramea carolina (Odonata: Libellulidae)

naiads, or no predator.

On 23 March 1995, I haphazardly added 30 Gosner (1960)

stage 25 (0.04 + 0.01 g n = 15) tadpoles to each pool. I








78

added predators the following day. When metamorphosis

commenced, I checked pools twice a day for tadpoles with

emerged forelimbs (stage 42; Gosner 1960). I held tadpoles

individually in 500-ml plastic jars until tail resorption

(stage 46; Gosner 1960), and then recorded wet mass (to

0.001 g).

During the second week of May 1995, the area

experienced an unseasonal heat wave with daytime highs

e:xceeding 32"C for several days. At the end of the heat wave

air temperatures exceeded 38"C and water temperatures

reached 40"C. This event caused a massive die-off of

tadpoles on 15 May 1995. At this stage of the ex:periment 71%

of tadpoles surviving to that point already had

metamorphosed successfully. I collected the dead tadpoles

and sacrificed any tadpoles that remained alive and obtained

developmental stage. Thus, 15 May 1995 was considered the

end of the e::periment. Because of the die-off, response

variables were the mass of tadpoles that reached Gosner

(1960) stage 46 (stage 42 by 15 May 1995), age of metamorphs

(calculated as the start of the ex:periment to stage 42), and

percent survival. Because tadpoles were large compared to

predators, I assuitmed LlidL ay Ladpulcs ii the pouol onl 15

May 1995, whether dead or alive, would have survived

predation. Thus, survival was based on the number of

tadpoles that reached stage 46 plus pre-metamorphic tadpoles

that remained at the time of the die-off.








79

Statistical Analysis

Treatment effects were analyzed using two-way fi;:ed

effect analysis of variance (ANOVA). Mean values per pool

were analyzed because measurements of individuals within

pools are not independent. One replicate from the no

predator :: low cover treatment was dropped from the

experiment because the pool developed a leak causing rapid

water loss resulting in tadpole mortality. Percentage data

we e 1 ju I, 1 I I y tI r 111 uI K Fo ed Ilk ot I w I I I:> vi i I .idl k-; w(' I

log transformed (log(:: + 1) for mass values) prior to

analysis. I performed orthogonal contrasts to test for

predictions regarding effect of predators (high predation by

large predators) and cover (higher predation at lower

cover), as well as prediction regarding enhanced growth.

When interactions were significant (p<0.05), I conducted

separate comparisons within treatment levels.



Results

Survival

Predator treatment had a significant effect on total

survival, accounting for 71.6% of variation in survival.



(Figure 4-1; Table 4-1). Orthogonal contrasts indicted that

survival of tadpoles was lower in treatments with large

predators compared to treatments with no or small predators.

This was true at high (F=34.66, p<0.001) and low (F=11.80,








80

p=0.004) cover treatments. The interaction between predator

and cover was due to lower survival within the low cover

treatment compared to the high cover treatment for tadpoles

e::posed to large predators. However, the interaction

explained only 8.3% of variation. Small predators were

largely ineffective. Survival of tadpoles in pools without

predators did not differ from that of tadpoles in pools with

small predators, regardless of cover (contrasts: low cover

F=0.11, p=0.74; high cover F=0.07, p=0.80). The main effect

of cover was not significant. Overall, tadpole survival was

lowest under the large predator :: low cover treatment.



Mass at Metamorphosis

Mass at metamorphosis was affected by the main effects

of predator and cover, and by the interaction of cover and

predator (Figure 4-1; Table 4-1). Within the large predator

treatment, tadpole mass was higher within the low cover

treatment compared to the high cover treatment (F=24.42,

p=0.002). Effects of small ('=0.02, p=0.8O) and no predator

(F=1.16, p=0.30) treatments were not significantly

influenced by cover. Effects of predator treatment on

tadpole mass at metamorphosis were not significantly

different within the high cover treatment (F=2.748, p=0.12);

however, in low cover treatments, tadpoles within the large

predator treatment were significantly heavier than those in

the small (F=37.51, p<0.001) and no (F=24.5, p<0.001)









81

predator treatments. Tadpole mass did not differ between the

small predator and no predator treatments within the low

cover treatment (F=0.41, p=0.53).



Age at Metamorphosis

The main effect of predator and the interaction of

predator and cover had a significant effect on age at

metamorphosis; however, the main effect of cover was not

significant (Figure 4-1; Table 4-1). At high cover levels,

age at metamorphosis was not significantly affected by

predator treatment (F=0.01, p-0.92). HIowever, witLinii the low

cover treatment, tadpoles e::posed to large predators

metamorphosed earlier (F=11.34, p<0.001) compared to the

small and no predator treatments, which did not differ

(F=0.001, p=0.97).



Discussion

Predator size had a much larger effect on tadpole

performance than did level of habitat comple::ity. High cover

level did reduce predation levels compared to low cover when

large predators were present, suggesting that habitat

complexity does provide increased protectioni from predators.

The most parsimonious interpretation for lack of effects by

small predators is that tadpoles were able to grow large

enough early in the experiment to reach a size refuge and

escape predation. Thus, early growth by R. utricularia








O 1


tadpoles developing in temporary environments can greatly

reduce predation pressure from aquatic insects that are

oviposited around the same time.

Similarly, growth probably decreased overall predation

rates by large predators; however, tadpoles were within the

gape limits of large predators for a longer time, allowing

increased predation by odonate naiads. Because tadpole

survival was higher at high cover levels when large

predators were present, this suggests that increased cover

decreased the foraging efficiency of T. carolina naiads. A

possible alternative hypothesis is that tadpoles raised

under higher cover grew faster and therefore decreased

predator success by surpassing the gape limits of predators.

This is unlikely, however, because growth of tadpoles on the

other treatments differed little relative to cover level.

Thus, the most likely enplanation for decreased predation at

high cover levels is interference in predator foraging.

Decreased foraging efficiency with increasing habitat

comple::ity has been found in other studies of aquatic insect

predators, including species that are active foragers (e.g.,

Ana:: junius) and sit and wait predators (e.g., Belostoma

sp.) (Chapter 6). Heck and Crowder (1991) predicted that

less mobile predators would actually have increased foraging

efficiency in comple:: habitats due to increased perch sites

for foraging. Banks and Beebee (1988) demonstrated that

increasing complex:ity from no vegetation to some vegetation








83

increases foraging success of odonate naiads on Bufo

calamita tadpoles. Their study provides some support for

Heck and Crowder's prediction. However, a majority of

research to date suggests that, like their mobile piscine

counterparts, aquatic insects often have decreased foraging

efficiency in structurally comple:: habitats.

R. utricularia tadpoles reduce activity in the presence

of Ana:: junius naiads (Chapter 5). Reduced activity has been

identified as a potentially important anti-predator

mechanism in many amphibian species, particularly among

species that breed in permanent aquatic habitats where fish

are the dominant predators (Werner and McPeek 1994; Werner

and Anholt 1993). Research suggests that the anti-predator

response is general, rather than predator specific (Stauffer

and Semlitsch 1993). Thus, reduced activity may have

afforded R. utricularia some protection from predation. The

results of this study, however, suggest that size

differences between predator and tadpole may be more

important than prey activity, at least when aquatic insects

are the predators. For example, increased cover also

decreased predation levels on Hyla squirella tadpoles, a

species thaL lhas high activity levels eveii ini Lh(i presence

of predators (Chapter 6).

The thinning effects of predators resulted in larger

tadpole size and more rapid development (earlier age at

metamorphosis). These results agree with other studies of




Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EBFEIUX9X_QFHCVD INGEST_TIME 2014-12-11T22:37:13Z PACKAGE AA00026485_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES



PAGE 1

TADPOLES AND PREDATORS: PATTERNS IN SPACE AND TIME AND THE INFLUENCE OF HABITAT COMPLEXITY ON THEIR INTERACTIONS By KIMBERLY JANE BABBITT A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 1996 UNIVERSITY OF FLORIDA LIBRARIES

PAGE 2

Copyright 1996 by Kimberly Jane Babbitt

PAGE 3

To my father.

PAGE 4

ACKNOWLEDGMENTS Funding for this project was provided by the Institute of Food and Agricultural Sciences at the University of Florida, Archbold Biological Station, and a private donation by Frank "Sonny" Williamson. Archbold Biological Station provided housing, laboratory space, and use of a four-wheel drive vehicle. The cooperation of the staff at Archbold Biological Station and the MacArthur Agro-Ecology Research Station is greatly appreciated. I thank my committee members, Lyn Branch, Dick Franz, Carmine Lanciani, and Carole Mclvor for their help and support throughout my project. I chose ray committee members with specific purpose in mind, hoping that their differing specialties would provide a broad and balanced assessment of my work. They fulfilled that purpose. I particularly want to thank George Tanner for serving as my advisor. George provided a supportive base from which to work. His easygoing nature made working on a complicated project relatively painless. Natalie Hardman was hired as an assistant during my first sampling season. She had a strong desire to learn about amphibians. Unfortunately, the weather did not iv

PAGE 5

cooperate and we spent much of the summer wondering when the rain would come. It never did. Natalie helped me with nonfield activities and kept her spirits up through the disappointing field season. Her flexibility and good humor were appreciated. Sam Gibbs assisted me at the beginning of my first rainy season and was particularly helpful with trap building and setting up initial wading pool experiments. Vicki Drietz and Joan Morrison were very helpful in watching for metamorphs from my experiments when other obligations took me away from the ranch. Frank Jordan, Ken Portier, and Christy Steible provided valuable statistical advice. Jerrell Daigle was particularly helpful in verifying odonate identification Ranch managers Dan Childs and Gene Lollis, and the rest of the staff at Buck Island were a pleasure to work with. They were particularly helpful in building a wetland enclosure and a fence around my experimental arrays. I particularly want to acknowledge the friendship of Frank Jordan. I could not have asked for a better colleague or friend. I hope we continue to collaborate in the future. I want to thank my family for their support. In particular, I want to thank my parents and grandparents for introducing me to the great outdoors and encouraging my interest in nature. Finally, I want to thank Leslie, whose support and friendship over the past thirteen years has made all of this possible. V

PAGE 6

This is contribution No. 2 6 from the MacArthur Agroecology Research Center. vi

PAGE 7

ACKNOWLEDGMENTS LIST OF TABLES. LIST OF FIGURES ABSTRACT CHAPTERS 1 INTRODUCTION Overview Patterns of Distribution Among Wetlands... Patterns of Distribution Within Wetlands.. Predation, Habitat Structure, and Behavior Dissertation Structure TABLE OF CONTENTS page iv X xii XV 2 USE OF TEMPORARY WETLANDS BY ANURANS IN A MODIFIED LANDSCAPE: EFFECTS OF HYDROLOGY AND WETLAND SIZE 12 Introduction j2 Methods ]^ Study Site |^ Sampling Methods 1^ Statistical Analyses 16 Results Species Richness and Abundance 1 Breeding Phenology 1^ Comparison of Breeding Activity Among Three Summers 1^ 9 n Discussion 3 SPATIAL AND TEMPORAL DYNAMICS OF TADPOLES AND AQUATIC INSECT PREDATORS DEVELOPING IN TEMPORARY WETLANDS Introduction Methods Study Site Sampling Methods vii

PAGE 8

statistical Analyses Results Physical Parameters Tadpole and Predator Distribution.. Species-Specific Responses Relationship to Physical Parameters Relationship to Macroinvertebrates Discussion 4 EFFECTS OF COVER AND PREDATOR SIZE ON SURVIVAL AND DEVELOPMENT OF RANA UTRICULARIA TADPOLES "^^ Introduction Methods Experimental Design Statistical Analysis "^9 Results Survival Mass at Metamorphosis 0 Age at Metamorphosis 81 Discussion 8-'5 EFFECTS OF FOOD AVAILABILITY AND RISK OF PREDATION ON BEHAVIOR AND GROWTH OF RANA UTRICULARIA TADPOLES 88 Introduction 88 Model Predictions 91 Growth and Development 91 Activity 91 Methods Behavioral Observations 94 Response Variables and Statistics 94 Results 95 Effects on Survival 9b Effects on Growth and Development 95 Effects on Tadpole Activity and Distribution 96 Discussion ^ Behavioral Responses 97 Growth and Development 100 Ecological Consequences of Responses 102 6 EFFECTS OF COVER AND PREDATOR IDENTITY ON PREDATION OF HYLA SQUIRELLA TADPOLES 110 Introduction HO Methods Results ll"^ Discussion viii

PAGE 9

7 SUMMARY AND CONCLUSIONS 118 LITERATURE CITED 123 BIOGRAPHICAL SKETCH 135 ix

PAGE 10

LIST Of T/VBLlillj Table page 21 Occurrence and abundance of tadpoles at 12 temporary wetlands at MAERC 29 31 Split-plot analyses of variance testing for the effects of wetland, zone, sampling period (time), zone X wetland and zone x sampling period on estimates of depth, temperature, pH, cover, and plant biomass. The main effects of zone and wetland were tested over the split-plot error (zone x wetland) whereas other factors were tested over the whole plot error (mean square error) 3-2 Split-plot analyses of variance testing for the effects of wetland, zone, sampling period (time), zone X wetland and zone x sampling period on transformed estimates of tadpole and predator densities and biomass. The main effects of zone and wetland were tested over the split-plot error (zone X wetland), whereas other factors were tested over the whole plot error (mean square error) 3-3 Split-plot analyses of variance testing for the effects of wetland, zone, sampling period (time), zone X wetland and zone x sampling period on transformed estimates of H. squirella R. utricularia and P. ocularis densities. The mam effects of zone and wetland were tested over the split-plot error (zone x wetland), whereas other factors were tested over the whole plot error (mean square error) 3-4 Correlations between all species, Hyla squirella Rana utricularia and Pseudacris ocularis with water depth, temperature, pH, plant cover, total plant biomass, and macroinvertebrate density and biomass 58 59 60 61 4-1 Summary of ANOVA for responses of Rana utricularia tadpoles to cover and predator treatments. CD is the coefficient of determination, which is the percentage of variation in the response variable attributable to the treatment effect 86 X

PAGE 11

5-1 Summary of ANOVA for growth and larval period responses of Rana utricularia tadpoles to food and predator treatments. CD is the coefficient of determination, which is the percentage of variation in the response variable attributable to the treatment effect 5-2 Summary of ANOVA for activity and distribution responses of Rana utricularia tadpoles to food and predator treatments. CD is the coefficient of determination, which is the percentage of variation in the response variable attributable to the treatment effect xi

PAGE 12

LIST OF FIGURES Figure EiS^ 2-1 Map of the MacArthur Agro-Ecology Research Center showing the system of ditches that enhance drainage -^^ 2-2 Location of wetland study sites at MAERC 31 2-3 Abundance of tadpoles of anuran species at 12 temporary wetlands at MAERC. Bars are mean + 1 SE and are based on total combined captures in all 0.25m^ traps at each wetland 32 2-4 Relationship between wetland size and anuran species richness (top) and abundance (bottom) Abundance and richness based on total captures from all 0.25m^ traps within each wetland 33 2-5 Relationship between wetland hydroperiod and anuran species richness (top) and abundance (bottom) Abundance and richness based on total captures from all 0.25m' traps within each wetland.. 34 2-6 Pattern of tadpole occurrence during 17 months of continuous sampling at 12 temporary wetlands at MAERC 35 2-7 Monthly rainfall totals at MAERC for 1993-1995. Total rainfall for each year was: 121.6 cm in 1993, 137.6 cm in 1994, and 157.6 cm m 1995 3b 2-8 Mean (+ 1 SE) water depth at sampled wetlands. The number above each bar indicates the number of wetlands that had water during the sampling period. (n=12 through 5/23/95 and 11 thereafter) 37 2-9 Comparison of mean (+ 1 SE) tadpole abundance at temporary wetlands during the summer breeding season (June-September) during 1994 and 1995. Abundance represents the combined captures m all 0.25m^ traps at each wetland 38 xii

PAGE 13

210 Comparison of total sample composition between the 1994 and 1995 summer breeding seasons at wetlands that were (n=6) and were not (n=5) impacted by flooding from ditches containing fish predators. Species categorized as temporary site breeders were species that do not breed in sites with fish predators. Rana grylio, Rana utricularia Acris gryllus and Hyla cinerea will breed with fish predators and were categorized as permanent site breeders -'^ 31 Variation in water depth (mean +1 SE) among sampling dates (top), wetland (middle), and wetland zone (bottom) Zone means with different letters are significantly different 62 3-2 Variation in water temperature (mean +1 SE) among sampling dates (top), wetland (middle), and wetland zone (bottom) 63 3-3 Variation in pH (mean +1 SE) among sampling dates (top), wetland (middle), and wetland zone (bottom) Zone means with different letters are significantly different 64 3-4 Variation in plant cover (mean +1 SE) among sampling dates (top), wetland (middle), and wetland zone (bottom) Zone means with different letters are significantly different 65 3-5 Variation in plant biomass (mean +1 SE) among sampling dates (top), wetland (middle), and wetland zone (bottom) Zone means with different letters are significantly different 66 3-6 Variation in tadpole numbers (mean +1 SE) among sampling date (top), wetland (middle), and wetland zone (bottom) Zone means with different letters are significantly different 67 3-7 Variation in tadpole biomass (mean +1 SE) among sampling date (top), wetland (middle), and wetland zone (bottom) 6b 3-8 Variation in macroinvertebrate predator numbers (mean +1 SE) among sampling date (top), wetland (middle), and wetland zone (bottom) 69 3-9 Variation in macroinvertebrate predator biomass (mean +1 SE) among sampling date (top), wetland (middle), and wetland zone (bottom). Zone means with different letters are significantly different '^^ xiii

PAGE 14

3-10 Variation in Hyla sguirella numbers (mean +1 SE) among sampling date (top), wetland (middle), and wetland zone (bottom) "^1 3-11 Variation in Rana utricularia numbers (mean +1 SE) among sampling date (top), wetland (middle), and wetland zone (bottom) "^2 312 Variation in Pseudacris ocularis numbers (mean +1 SE) among sampling date (top), wetland (middle), and wetland zone (bottom) 73 41 Mean (+ 1 SE) responses of Rana utricularia tadpoles to cover and predator treatments: percent survival (top), wet mass at metamorphosis (middle), and age at metamorphosis (bottom) Squares indicate high cover, circles indicate low cover 87 51 Mean (+ 1 SD) responses of Rana utricularia tadpoles to food level and predator treatments: age at metamorphosis (top) and wet mass at metamorphosis (bottom) Circles indicate predator absent and squares indicate predator present, (n-6 for all treatments) 52 Mean (+ 1 SD) responses of Rana utricularia tadpoles to food level and predator treatments: percentage of tadpole on the side of the container opposite the predator (top) and percentage of tadpoles that were active (bottom) Circles indicate predator absent and squares indicate predator present. (n=6 for all treatments) iuy 61 Number of Hyla sguirella tadpoles that survived predation under different cover levels. Bars are means + 1 SE (n=4 for each bar) xiv

PAGE 15

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy TADPOLES AND PREDATORS: PATTERNS IN SPACE AND TIME AND THE INFLUENCE OF HABITAT COMPLEXITY ON THEIR INTERACTIONS By Kimberly Jane Babbitt August, 1996 Chairman: George W. Tanner Major Department: Wildlife Ecology and Conservation I examined the spatial and temporal dynamics of tadpoles and aquatic insect predators in 12 temporary wetlands in south-central Florida. I also conducted experiments on the role of habitat structural complexity in mediating predator-prey interactions. Eleven species of anurans bred in the wetland sites. Rana utricularia accounted for 44% of all captures. Species diversity was positively correlated with wetland size but was unrelated to hydroperiod. Varied meteorological conditions, coupled with the presence of a system of drainage ditches, resulted in annual variation in hydrologic conditions that had a large effect on species composition. Flooding resulted in the intrusion of water from ditches that contained fish into otherwise isolated wetlands. Species that breed exclusively XV

PAGE 16

in sites without fish avoided these sites. Hydrologic disturbance resulted in a system that is both spatially and temporally dynamic resulting in variable assemblage structure and composition at affected sites. Distribution patterns of tadpoles within wetlands were related weakly to habitat features. Both tadpole and aquatic insect predator numbers varied among wetlands and across time, but were distributed relatively evenly within wetlands. Thus, selection for microhabitat at the within-wetland scale did not appear to be strong. Distribution patterns of tadpoles appear to be more strongly related to macrohabitat selection of breeding sites by adults than by microhabitat selection within wetland sites. Experiments demonstrated that increased cover can reduce predation rates by aquatic insects on tadpoles that reduce activity in response to predators (R. utricularia ) and those that do not (Hyla squirella ) Thus, although distribution patterns in the field were not correlated strongly to habitat features, tadpole success still may be influenced strongly by them. Finally, reductions in activity in response to predators had growth and developmental costs in R. utricularia that were dependent on background resource level. Tadpoles on low food metamorphosed at later dates but at larger sizes when exposed to predators. Tadpoles on high food also had prolonged development but were smaller compared to tadpoles not exposed. Thus, assessment of responses to both resource xvi

PAGE 17

level and threat of predation may be necessary for understanding the consequences of cost-benefit trade-offs mediated by behavior. xvii

PAGE 18

CHAPTER 1 INTRODUCTION Overview One of the major goals in ecology is to determine the factors that influence patterns of distribution and abundance of species. This task is not a simple one because many factors impinge upon the ecological and evolutionary processes that result in observed relationships among species. Resource availability and distribution change along many spatial and temporal scales, and the biotic interactions among species change similarly. The resulting heterogeneous environment offers choices to the organisms living within it. How an individual responds to habitat heterogeneity may influence food availability and predation pressure, and thus can have important influence on population dynamics and ultimately evolutionary processes (Holt 1987) Examination of the factors influencing distribution patterns of many anurans is complicated by their complex life cycles, which include a terrestrial adult stage and an aquatic larval stage (Wilbur 1980) Examinations of habitat use and resource partitioning of anuran assemblages have focused mostly on the adult stage (Bowker and Bowker 1979; Crump 1971, 1982; Lizana et al. 1990; Toft 1985). Analyses 1

PAGE 19

of larval assemblages usually address temporal partitioning (Dixon and Heyer 1968; Weist 1981) or partitioning relative to permanent versus temporary pond sites (Gascon 1991; Woodward 1983) Most authors come to similar conclusions regarding habitat partitioning of larval assemblages: resources are partitioned along spatial (temporary versus permanent ponds) and temporal (breeding season) axes. Less research has been done on larval food partitioning; however, evidence suggests that partitioning along the food axis is not as strong as that along temporal and spatial a;:es (Duellman and Trueb 1986; Heyer 1974, 1976; Seale and Beckvar 1980) These studies provide important information regarding possible selection pressures relative to habitat use and resource partitioning (Crump 1982; Magnussan and Hero 1991) However, they do not provide a complete picture of the spatial ecology of tadpoles because they focus on one scale. Distribution patterns of organisms can be examined on a coarse-grained, macrohabitat scale or a Cine-grained, microhabitat scale. Factors that can influence tadpole success, such as competitor or predator abundance and composition, vary in space not only among wetland sites (Gascon 1992; Smith 1983; Werner and McPeek 1994; Woodward 1983) but also among patches within wetlands (Banks and Beebee 1988; Diehl 1988, 1992) Therefore, the distribution patterns of tadpoles at both the macrohabitat (breeding

PAGE 20

3 site) and microhabitat (patch site) scale can have a large influence on tadpole growth and survival, and thereby assemblage structure. Although the importance of considering scale has been widely discussed (Wiens 1989, 1992), few studies of tadpole ecology have addressed this issue. Processes that are important population regulatory factors at one scale of investigation may fail to operate at other scales (Wiens 1992) For er.ample, predation has been identified as an important process influencing anuran breeding site selection (Magnusson and Hero 1991) However, how predators and tadpoles interact at the within-wetland scale is poorly understood. Because ecological patterns are not independent of scale, e::amination of spatial and temporal distribution patterns at several scales should provide a more complete understanding of the importance of a process, such as predation, in regulating tadpole populations (Wiens 1989, 1992) Patterns of Distribution Among Wetlands Patterns of distribution of larval anurans among aquatic sites are largely determined by two related factors; the permanence of the site (i.e., hydrology) and the composition of predators (Skelly 1995; Werner and McPeek 1994) These factors are strongly interrelated because predator composition changes along the hydrologic gradient.

PAGE 21

4 Fish predators dominate in permanent sites; however, systems that have a cycle of filling and drying prevent predatory fish populations from becoming established. A switch from fish-dominated to invertebrate or salamander-dominated predator systems occurs along the gradient so that tadpoles developing in temporary and ephemeral sites are subject to a different suite of predators. Because biotic and abiotic conditions change along a hydrologic gradient, tadpoles developing in different wetlands face a different set of challenges. Traits that increase developmental success at one end of the hydrologic gradient may not be suitable at the other end. For example, interspecific variation in ant i -predator mechanisms is related to differences in breeding habitat (Azevedo-Ramos et al. 1992; Kats et al. 1988; Woodward 1983). The larvae of several species that breed in permanent sites are unpalatable to fish, whereas species that breed in ephemeral sites without fish do not possess this anti-predator mechanism (Formanowicz and Brodie 1982; Kats et al. 1988). In addition, behavioral adjustments, such as reduced activity, are often absent in species that breed in ephemeral sites (Kats et al 1988; Woodward 1983). Decreased foraging activity may reduce growth or developmental rates (Skelly 1992, 1995; Skelly and Werner 1990) At ephemeral sites, mortality threats from predators are often not as significant as the threat from habitat

PAGE 22

5 drying (Newman 1987; Rowe and Dunson 1993, 1995) Thus, behavioral traits for reducing predation risk, while beneficial in permanent habitats, can lead to increased mortality risk at ephemeral sites. The trade-off between rapid development and antipredator responses appears to be a major factor limiting the distribution of anurans to certain types of breeding sites (Skelly 1995; Werner and McPeek 1994) Weak anti-predator responses that may increase developmental rates in ephemeral sites would usually result in mortality in permanent sites (Skelly 1994; Woodward 1983) As a broad generalization, species that breed in permanent sites are passively er.cluded from ephemeral sites by long larval developmental periods, whereas species that breed in ephemeral sites are excluded from permanent sites by predation. Patterns of Distribution Within Wetlands Few studies have examined tadpole distribution patterns within breeding sites. Diaz-Paniagua (1987) examined the spatial distribution of six species occurring in 16 temporary ponds. Dip-net samples were taken in 5 depth zones. Vegetation structure varied along the depth gradient from shallow zones characterized by dense grass cover to deeper areas with little or no vegetation. Along the depth gradient, veqetation chanqed from emerqent in shallow water to submergent in deep water. Few tadpoles of any species

PAGE 23

6 were found in the inner non-vegetated zones. Although most species were found in several of the vegetated zones, interspecific differences in spatial distribution were detected. Three species were most abundant in the shallow zones, two species were most abundant in the emergent zone and the remaining species was most common in the submergent zone Larval assemblages also may separate along thermal (Dupre and Petranka 1985) and water column (Heyer 1973, 1976) strata. Ontogenetic changes in microhabitat association have been found relative to size class in Rana utricularia (Alford and Crump 1982) Alford (1986) found evidence for both intraspeci f ic (i.e., size class) and interspecific spatial partitioning among four anuran species within a single temporary pond site. Two species, Pseudacris ocularis and Pseudacris ornata separated relative to cover and water temperature, and their responses to these environmental features changed over time. Alford (1986) suggested that these responses were probably largely due to competition (i.e., avoidance relative to species or size class) However, he did not er.amine the potential influence of predation. Differences in predator abundance or composition also may exist within a breeding site. Although poorly studied in the tadpole-predator system, the potential for spatial segregation between predators and prey to be an important

PAGE 24

7 factor within wetlands sites was demonstrated by results from Banks and Beebee (1988) who found low spatial overlap between Bufo calamita tadpoles and the invertebrate predators Notonecta sp. and Dytiscus sp. Tadpoles were concentrated in shallow water, whereas predators were collected most frequently in deeper areas. Because few studies have er.amined spatial distribution of tadpoles within breeding sites, it is not surprising that there is limited information about the spatial relationships of tadpole and predator distributions. Spatial overlap at the within-wetland scale among tadpoles and aquatic insect predators may be high if both groups select habitats within the wetland based on similar abiotic features such as water depth, temperature, or plant cover. In contrast, differential selection criteria could lead to segregation of tadpoles and predators. For example, females of some odonates require emergent vegetation for egg laying (Corbet 1980) In this circumstance, tadpole survival may be largely influenced by patterns of microhabitat use and factors such as habitat structure. A better understanding of how patterns of tadpoles and predators vary in space and time, and how these patterns are related to abiotic and biotic (e.g., vegetation structure) factors, should provide valuable insight into the potential for such patterns to generate differences in species richness and assemblage structure among various sites.

PAGE 25

8 Predation, Habitat Structure, and Behavior Predation can be a significant regulating factor in many systems (reviewed by Sih et al. 1985). E;:per imental studies on tadpole assemblages have shown that differential predation on competitively dominant species can alter assemblage structure (Morin 1M1, 1Q3, Generally, competitive ability in tadpoles is related to activity levels: actively foraging tadpoles out-compete tadpoles that are less active (Werner 1991, 1992a; Woodward 1982). However, because prey activity is an important component oE predator detection (Lima and Dill 1989; Werner and Anliolt 1993), active tadpoles are more susceptible to predators (Azevedo-Ramos et al. 1992; SkeUy 1994; Woodward 1983). Interspecific differences m survival due to predation, and the factors that influence predation rates are important in e;:plaining between-site or witlim-site differences m species abundance patterns. A plethora of studies on various aquatic ta;:a lias demonstrated the important role that habitat compler.ity plays in mediating predator-prey interactions (reviewed by Heck and Crowder 1991). Increased habitat comple;:ity can reduce predaLiun rates by providing cover ui. paitial lofuge areas for prey (e.g, Folsom and Collins 1984; Rozas and Odum 1988), or by decreasing foraging success of predators because of decreased maneuverability or visual range (e.g., Crowder and Cooper 1982; Savino and Stein 1982; Werner et

PAGE 26

9 al. 1983). Considering the important effect predators have on tadpole survival and assemblaye structure, surprisingly little research has focused on how habitat structure affects tadpole-predator interactions. Morin {1981, 1983, 198G) suggested that Pseudacris crueller tadpoles survived better than Scaphiop us hoibrooki tadpoles when subject to predation by Notophthalmus viridescens by restricting activity to the bottom leaf litter. In contrast, G. hoibrooki actively foraged in the water column. Whether differences in microhabitat use or activity levels (or both) were more important in determining vulnerability is not clear and has not been demonstrated. The results of the few studies that have examined the effects of habitat structure on tadpole-predator interactions are equivocal. In relatively simple environments. Banks and Beebee (1988) found that even low levels of habitat structure increased predation on tadpoles by providing perch sites for odonate larvae. Figiel and Semlitsch (1991) found tliat habitat comple;:ity did not affect crayfish (Procambarus acutus acutus) predation on artificially injured Hyla chrysoscel is In a rare field ;jtudy, :Jredl and Collins (199^) Lound that h.ibitat comple:city alone did not significantly affect larval performance; however, interactions among habitat comple;:ity, predator density, and prey density tosullo.l in comply.-., nonlinear relationships.

PAGE 27

iO Finally, dlElGrences in iii i crotial) i LaL .sLr ucLure may noL affecL only overall predaLion raLes, but also suscepL ibi 1 1 ty of different size classes. Hews (1995) found that small tadpoles were more susceptible to predaLion by a fish predator in a vegetated microhabitat whereas larger tadpoles were more susceptible on a gravel substrate. The question of how ant i -preda Lor responses affecL predaLion levels in enviroiimenLs LliaL differ in habiLaL comple;;ity would provide valuable information on how behavior and abiotic IcaLures inLeracL lo d.^Lefmine suscepLibiliLy to predation. For er.ample, does comple;: habitat structure lead to decreased predation pressure on actively foraging species that show little response to predation, or do high activity levels lead Lo higher predation levels regardless of habitat structure? Similarly, for species that do respond to predators, does cover provide enhanced protection over that due to behavioral responses? E;:perimental e::araination of these questions should provide additional insight into the significance of patterns of distribution found in the field. Uissertatiui i :: trucLui e The research I conducted e;:amined spatial and temporal dynamics of larval anurans and aquatic insect predators and the effects of microhabitat structure on their predator-prey relationships. My research has thrne general componenis: (1)

PAGE 28

11 e;:amination of the distribution patterns of larvae and predators at 3 spatial scales (among wetlands, between wetland zones, and within zone) across 3 suimner and 1 winter breeding season, (2) e;:perimenta 1 manipulations to e;:amine the effect of microhabitat on predation, and (3) e::aminations of the effect of larval predator-avoidance behavior on growth and development under controlled conditions I e;:amined spatial and temporal relationships of tadpoles and aquatic insect predators in several temporary wetlands to determine whether species composition and distribution patterns differed among wetlands (Chapter 2) and within wetlands relative to physical and biological characteristics m different patches (Chapter 3). To e::amine the potential consequences of physical features in the environment to tadpole survival, I ronducted controlled e::periments to determine the role of habitat comple::ity in mediating predator-prey interactions (Chapters '\ and G) In a subset of these er.periments 1 e;:amined whether there were different effects based on the identity of the predator (r^haplrrs f,). I 1 so o;:amino
PAGE 29

CHAPTER 2 USE OF TEMPORARY WETLANDS BY ANUIWIS IN A MODIFIED LANDSCAPE: EFFECTS OF HYDROLOGY AND WETLAND SIZE Introduction Patterns of distribution of larval anurans among aquatic sites are largely determined by two related factors: the permanence of the site (i.e., hydrology) and the composition of predators (Skeily 1995; Werner and McPeek 1994). These factors are strongly interrelated because predator composition changes along hydrologic gradients. Fish predators dominate in permanent sites; however, systems that have an annual cycle of fiinny and dryxng prevent predatory fish populations from becoming established. Interspecific differences m life history traits among anurans, such as length of the larval developmental period and ability to breed successfully with fish predators, limit the range of wetlands within which a species can breed successfully. Species with long larval periods are e:;cluded passively from sites with short hydroper rods whereas most species that breed in temporary sites do not possess the necessary chemical or behavioral characteristics for avoiding the heavy predation pressure found m permanent sites (Kats et al 1988; Woodward 1983). Because of intcrspccilic differences among .inui.ni i > ic I i v i I i 12

PAGE 30

13 that alter wetland hydroperiod (the amount of time a wetland holds water) can have a large ellect on the use ol wetlands as breeding sites and the ability of the wetlands to successfully produce metamorplis (t'echmann et al 1989). Ditching of wetlands to enhance drainage can have a large effect on anuran use of wetland sites (Vickers ct al 1985) In many parts of southern Florida, wetlands have been ditched and drained to make the naturally wet sites more suitable for agricultural activities. Under this management activity, large tracts of land are subject to an e;:tensive series of interconnected ditches. Changes to wetlands within the altered landscape include wetland loss, reduced hydroperiod, and m some cases periodic connection to deeper water bodies that contain fish predators. The value of these wetlands for anurans, or how anuran assemblages may have changed as a result of these activities has, to my knowledge, not been e;:amined previously. A majority of anuran species in Florida breed e::clusively or facultatively in temporary wetland sites (LaClaire and Franz 1990; Moler and Franz 1987). Because temporary wetlands provide critical habitat for so many auuians m Floiida, 1 e::amincd use oL temporary wcAlands in a landscape modified to increase drainage to determine the value of these altered wetlands as breeding sites for anurans. I er.amined anuran breeding activities by sampling anuran larvae at several temporary wetlands and asked

PAGE 31

14 whether wetland hydrology or area affected use by anurans. I'urther, 1 compared breeding aclivily duriiK) •! summer periods that varied in sunmier rainfall patterns. Methods Study Site Researcli was conducted at the MacAr thur Aq ro-f^colocjy Research Center (MAERC) in Lake Placid, Highlands County, Florida. The site is an active cattle ranch consisting of 4,800 ha of improved and semi-improved pasture interspersed with emergent wetlands and oak (Quercus virgiaiana) and palm (Sabal palmetto ) hammock. The ranch was ditched e;:tensively in tlie 1900 's to enhance drainage, and most wetlands are connected to the vast system of ditches that ultimately enters into Harney Pond Canal, which flows into Lake Okeechobee. (Figure 2-1). Most wetlands are properly characterized as emergent freshwater marsh. Thick-stenmied emergents, such as Pontederia cordata and SagitLaria lanci_folla, are dominant species in deeper water with Panicum hemUojnon, Polygonum punctatum and Alteranthera phUoiero^ as sub-dummant species. Outer, shallower areas of temporary wetlands are characterized by lower-stature emergents such as Bacopa caroliniana Hydrochloa carolinensis and Diodia v irginiana

PAGE 32

15 Sampling Methods From May 1994 through September 199Ij, 1 sampled 12 wetlands every three weeks (23 sampling periods) One wetland was dropped during the last si;: sampling periods due to damage from cattle trampling. In addition, I monitored sites from May 1993 through August 1993 during a drought period when no breeding took place. At each site, 1 collected nine samples using a stratilied landom sampling protocol. Three replicate samples were collected in eacli of three concentric belts corresponding appro:;imately to one third the radius of the wetland. Tadpoles were collected with a 1/4-m • open-ended, bo:: trap. Bo;: traps have been used to sample small fish and macroinvertebrates (Chick et al. 1992; Ficeman et al 1984; Kushlan 1981) and anuran larvae (Caldwell et al. 1980; Calef 1973; Pfenning 1990), and are effective Jor sampling In areas with comple:: vegetation structure. A bar seine fitting the diameter of the bo;: trap and fitted with fiberglass insect screening was swept through the water column (after vegetation had been cleared) until 3 consecutive sweeps yielded no additional captures. Samples were preserved m Ll.u Licld wiLI. bulLoi.Hl loiinaliM ,u,.l id.-nlil i-d l<. Species in the lab. information on other charac tei is t ics tliat puteiitially influence tadpole distribution and abundance within wetlands also was collected and is reported elsewhere (Chapter 3).

PAGE 33

1 1. For the purposes of this study, the nine samples provided a simple measure of breeding use (presence/absence), relative abundance, and species richness at each site. Wetland area was estimated from data previously entered into a Geographic Information System. Wetland area varied from 0.11 to 2.21 ha (Figure 2-2). Water depth was taken from a permanent PVC stake at the center of each wetland at the time of sampling. Rainfall data were obtained from an all-weather gage at the research center. Hydroperiod is usually measured as the total number of days a wetland held standing water. I took measurements every three weeks; therefore, 1 do not have actual dates of wetland filling and drying. During dry-down periods, I checked wetlands between sampling periods to determine whether wetlands actually dried completely. Therefore, I defined hydroperiod as the number of weeks a wetland held standing water. Statistical An a_lyses Data for each site were treated as independent samples, as were data from each suimner breeding period. I used Pearson Product Moment correlation to determine whether ,,pccies iichnc.;^ ur tutai tadpuL. ..bund,,.,,.. w.m
PAGE 34

17 assemblage composition, I used a chi-square goodness of fit test to e;:amine yearly differences in the number of tadpoles that were from species that will or will not breed in sites with fish predators. Because not all wetlands were affected by this disturbance, I used the proportion of tadpoles in each of the two categories (i.e., temporary breeding species and permanent breeding species) in unimpacted wetlands to generate e;:pected values for tadpole numbers rn wetlands that were impacted. Results Species Richness and Abundance A total of 3,G7y tadpoles fium 11 species was collected. Of the 11 species caught, only the southern leopard frog {Rana utricularia) was found at every wetland (Table 2-1). The squirrel treefrog (Hyla squirella) was the ne::t most widely distributed. In contrast, the oak toad (Bufo quercicus) and southern chorus frog (Pseudacris niqrita ) were collected from only two wetlands each. R. utricularia was the most abundant species, accountinq for about 44. of ail captures (Table 2-1; i'lqure Z-J). H. squ i re I la comprised ll'i of the s.implos, the little grass frog (Pseudacris ocularis ), pig frog (Hani SLD^) ^'^'^ pinewoods treefrog (H^la femoralis) were moderately abundant. The remaining species each accounted for 51 or less of the total sample.

PAGE 35

Anuran species richness varied from si;: to nine and was positively correiated with wetland size (Figure 2-4; r-0.65, p=0.023). However, the nuiiUjer of tadpoles captured was not significantly related to wetland size (Figure 2-4; r=0.35, p-0.29) Most wetlands (n=9) dried only once during the 17 month sampling period. The number of weeks that wetlands held water varied from 4 9 to 65. Neither species richness (r-0.03, p=0.93) nor number of individuals (r=0.40, p=0.22) was related to wetland hydrology (Figure 2-5). Further, hydroperiod was not related to wetland size (r=0.177, p-0.602) Breeding Phenology Although the e:cact date of breeding among anurans varies from year to year due to variability in meteorological conditions, many species have particular breeding seasons. During the 17 months of continuous sampling in this study, si:: species; B. quercicus, H. squirelj^, H. femoran^, H. gratjjosa, H. clnerea, and the eastern narrow-mouth toad ( Gastrophryne carolinensis) bred only during the sui.uuer rainy season (Figuie 2-b). F. rngMj^ tadpoles were collected only during the winter season. I did, however, hear sporadic calling during summer months indicating that this species may brood on a limited basis OMI si do Mioi, main winfor brood in
PAGE 36

19 ten of the species collected bred during the suimner, only four species; the cricket frog ( Acris gryllus) K. utricularia P. nigrita and P. o_cu]^ris bred during the Winter. A. gryllus P. ocularis and R. ulj_icuU£i liad tadpole occurrence patterns consistent with year-round breeding. Finally, Ran a grylio bred during spring and summer. Occurrence of tadpoles during winter months reflects the long larval developmental period of this species. comparison of Breeding Acti vi_tj^Among,Jl^^ Rainfall patterns differed among the three sunmier periods (Figure 2-7). Low rainfall during the 1993 suimuer breeding season resulted in drought conditions and dry wetlands throughout the entire sununer As a result, no breeding occurred at any sites during the suimner breeding season in 1993. Although suimuer lamfall patterns were similar in 1994 and 1995, differences in antecedent rainfall m winter and spring between the two years (and higher levels in Harney Pond Canal) resulted in a higher water table at the beginning of the rainy season in 1995. Whereas all wetlands were dry for several months prior to summer J,, I'lM/l, U:.> weU.UHls IImI -1, mmI wm, 1 1 d y in 1 did so only one to two weeks prior to the 1995 summer rains (Figure 2-8). The consequence of the difference m water table levels was that heavy rainfall at the beginning of the 1995 summer season resulted m ditches overflowing their

PAGE 37

20 banks and spilling water into adjacent wetlands. Comparisons of anuran breeding activity between 1994 and 199b indicated that species composition was affected by this flooding event (i'igure 2-9). H. squirella and G. carolinensis two species that breed e;:clusively in temporary sites, were less abundant in 1995 compared to 1994. In contrast, species that associate more strongly with permanent sites, such as R. gryiio and A. gryllus were more abundant in 1995. Wetlands that were unimpacted by flooding had similar proportions of tadpoles Irom species that will or will not breed in sites with fish predators (Figure 2-10). In contrast, wetlands that were impacted by spill-over from ditches differed greatly in composition between 1994 and 1995 (Figure 2-10; ;'/=1008, df=3, p<:0.0001) Discussion Changes due to ditching at MAERC have resulted in changes in wetland hydrology, and in a probable shift in assemblage structure and species composition that favors anurans that associate more strongly with wetlands lacking fish predators. Although historical data are not available LuL cumpaii:.un, ,n ...cas ul FI>Mi.l,, 11,. ,1 h.ivo 1 y, i r I o, ,i < : patterns that are probably similar to pre-ditching conditions at MAERC such as the Northern Everglades, R. gryl^o, A. gr yl lus and H. ci_ner-oa dominate the tadpole community (Babbitt et al., unpubl. data). These species are

PAGE 38

21 often associated with longer hydroperiod sites that contain lish predators, and were probably much more abundant than they are now (Carr 1940; Duellman and Schwartz 1958). Although each of these species bred in some of the wetlands I sampled, they concentrated most of their breeding activity in the larger ditches and an aLtiUciai lake (pers. obs ) Most of the wetlands on site, including those sampled during this study, are intermediate in hydroperiod and are best described as annually or semi-annually drying temporary wetlands. By viewing habitat drying as a disturbance, we would predict that the temporary wetlands at MAERC would have the highest species richness (compared to permanent or very ephemeral sites) based on predictions from the intermediate disturbance hypothesis (sensu Connell 197fJ). Both Heyer et al (1975) and Wilbur (1984) argued that the highest species diversity would be at intermediate values along a hydrologic gradient. Only one species that occurs at MAERC, the gopher frog (Rana capito) was not collected during the study. This species has an e;:tremely limited distribution on-site, but it does breed m temporary wetlands similar to those used in the study (pers. obs.). Thu.., ail an.n.an species that
PAGE 39

The long-duration temporary wetlands used in this study appear to provide suitable breeding habitat lor all anurans occurring at MAERC. What precludes most anurans from successful development in permanent sites is predation by fish. Because these temporary ponds dry on an annual (or semi-annual) basis, the removal of the fish predator assemblage makes these sites suitable for species that avoid permanent ponds. At the same time, all species in south Florida that breed in permanent ponds have larval developmental periods that are less than a year, and therefore could successfully reach metamorphic size in many temporary wetlands (Ligas 19G0; pers. obs ) Although temporary wetlands fall within the range of breeding habitats that all anuran species at MAEKC considered suitable, they probably do not necessarily provide the "best" breeding habitat for all species. For example, B. quercicus and P. nAgrUa were collected at only two wetlands. Both species utilize ephemeral sites and appeared to concentrate breeding activities at sites that were of shorter duration than the wetlands I sampled (pers. obs.). considering the varying life history requirements of Lh. tadpuic. ul dKLc.euL .p.M,,.:. snH, l.-v, 1 „ .hum. 1.,. 1 time, presence or absence of anti-predator mechanisms, and competitive ability, long-duration temporary wetlands are probably suitable, but marginal habitat, for some species. Tn contrast, these wetlands may provide "prime" breeding

PAGE 40

23 habitat for R. utricularia R. utricularia was the most abundant species and had the broadest drstribution. This species also has been found to be numerically dominant or abundant in censuses at other temporary wetland srtes in Florida (Dalyrumple 1988; Enge and Marion 1986; O'Neill 1995; Vickers et al. 198b). Viewed at the landscape level (i.e., MAERC) the ditching of MAERC probably increased anuran species richness. The effects of wetland alterations on anurans at MAERC differ from those that occur at many other sites because this site was altered from a large (usually) permanent aquatic mosaic srmilar to the Everglades to a terrestrial system containing numerous wetlands of varying hydroperiods. Reductions rn hydroperrod at sites that are already temporary can result in decreased species richness or abundance because truncated liydroper iods preclude some species from successfully metamorphosing (Pechmann et al 1989; Rowe and Dunson 1995; Semlrtsch 1987; Wrlbur 1987). The effects of ditching of wetlands on anuran species richness and breeding success depend on both the original and resultant conditions of a site. in this study, neither spocics i.j<:hne;;s noi tadpole abundance was related to wetland hydroperrod. These results contrast wrth those of Pechmann et al (1989) wlio found that hydroperiod was positively correlated with species richness and the abundance of metamorphosing amphibians. In their

PAGE 41

24 study, hydroperiods varied almost by a factor of five (i.e., 58 to 263 days) Hydroperiods in the current study varied by only a factor of 1.4, from 49 to G5 weeks. Further, the }iydroperiod of these wetlands was much longer Lhan the minimum required for metamorphosis of all the anuran species occurrinq at MAF.RC (Liqas l^GO; Wilbur l')7). In a (i.ve-year study, Dodd (1992) did not find a relationship between habitat duration and species richness at a single temporary wetland in north-central Florida. That study was conducted during a drought, and visitation by adult anurans was not necessarily associated with breeding activity. Large differences in hydrology among sites (or among years) may be required to generate significant differences in species richness. Relevant differences may include sites with hydroperiods that are too short for some species to successfully produce metamorphs, and permanent sites with fish predators. It is likely that the wetlands in the current study provided similar habilaL suiLability, relative to hydrology, and that any variation in use patterns by adults was largely related to other factors. Species richness was positively related to wetland size. ResuILs from studies tli.il have o::aininod l.he relationship between wetland size and species richness or abundance are equivocal. Whereas some researchers have found a positive relationship between wetland area and species richness (Kutka and Bachmann 1990; Laan and Veiboom 1990),

PAGE 42

2b others have found no relation at all (Diaz-Paniagua 1990; Laan and Verboom 1990; Richter and Azous 199b) Again, differences in these studies may represent differences in scale, as well as differences in other characteristics that influence use by adults (Lann and Verboom 1990; Richter and Azous 199b). Because species richness at a particular site is a community metric and therefore a product of the varying responses of coimnunity members, it may not be surprising that the above studies provide differing results. Within a particular class of wetlands, such as the temporary wetlands with relatively long hydroperiods sampled in this study, year-to-year changes in hydrology driven by meteorological conditions may have a larger influence on breeding use than many other factors. The largest influence on anuran use of wetland sites at MAERC was amount and pattern of rainfall. During the suimner drought conditions of 1993, no breeding took place in any temporary wetlands because they were dry. In 199/1, essentially all nonpermanent sites provided suitable habitat for species that breed in temporary wetlands because all sites dried down prior to the summer rains. This eliminated any fish predators that would have made tliese sites unsuitable for a majority of species. These two years contrasted further with 199b, when some wetlands dried completely before the suimner rains but others did not completely dry. Further, over-spill from adjacent fish-containing ditches apparently caused some

PAGE 43

anurans to avoid using sites that otherwise would have been suitable Although the mechanism through which adult amphibians are able to identify sites that contain frsh is not known, experimental choice studies have demonstrated that adults are capable of dif ferentrating between sites that do and do not contain fish predators (Hopey and Petranka 1994; Resitarits and Wilbur 1991). Differences in species composition among sites with and without frsh predators are due (at least partly) to adult selection, noL simply predation on the tadpoles of certain species. The impacts of over-spill ol water from f ish-conLaming ditches into otherwise isolated wetlands are transient because wetlands usually dry on an annual basis. Further, alternative sites are available for anurans that avoid breeding m sites with predatory fish, tor e::ampfe, aitliough species such as H. squirella and G. carolinensi^ avoided breeding m wetlands affected by ovcr-spill [rom fishcontaining ditches, these species bred m adjacent sites that were unaffected (pers. obs.). in contrast, tow rainfall can result m missed breeding oppo.tuniLics across Lhu iand.c.,u. l-ccau.. w..ll,iu.l d,„lMng er.acerbates the dry conditions. During seasonal droughts, the only species that can successfuffy breed are those fhat breed at permanent sites with fish predators. Whereas periodic seasonal droughts may have only a small impact on

PAGE 44

27 population numbers, the effects of long-term drought may be more profound. Few studies have examined amphibian populations during drought, and the comple:; and erratic patterns of population number of amphibians make assessment of the impacts of drought difficult (Dodd 1992) However, some species appear to be capable of surviving long-term droughts (Dodd 1995) Evidence that at least some individuals within a population may be opportunistic in selecting breeding sites, rather than philopatric, suggests that this may be an important mechanism for maintaining a breeding population during drought conditions (Dodd 1995) Opportunistic use of breeding sites also appeared to be important in areas affected by flooding during 1995 (pers. obs.). Particularly in landscapes such as Florida where wetlands are numerous, opportunistic use of breeding sites may be more important than previously recognized (also see Dodd 1995) These data suggest that temporary wetlands on this site provide dynamic habitats that offer varying breeding opportunities and larval developmental conditions that are highly dependent on meteorological conditions. The effect ul this is a spaLially-temporaliy dynamic system resulting in differing assemblage structure and composition at particular sites. Because MAERC contains numerous wetlands and ditches that vary in hydrology from ephemeral to permanent, most species can find suitable breeding habitat

PAGE 45

28 except under extreme drought conditions. During such conditions, ditching exacerbates the drought conditions by increasing drainage, lowering the water table, and prolonging dry conditions. In conclusion, historical changes to the landscape have probably increased local anuran species richness and altered tadpole assemblage structure at the wetlands at MAERC. An important aspect that was beyond the scope of this study is the effects of past and current land management on the amount and distribution of varying terrestrial habitats. The surrounding upland matrix can have a large influence on wetland use as breeding sites, particularly for amphibians that have a terrestrial adult phase (Kutka and Bachmann 1990; Laan and Verboom 1990) For example, arboreal species such as H. femoralis appeared to be locally abundant, but limited in distribution. Temporary wetlands near forested hammocks consistently contained H. femoralis tadpoles; however, wetlands away from hammocks in large open pastures did not. Consideration of upland characteristics is an important aspect of overall conservation of anurans Research on the relationship between upland habitat patch characteristics and wetland use would increase our understanding of breeding use patterns, as well as our ability to provide guidance for managing anuran populations. J

PAGE 46

29 rH • ft) u 4-1 PC O Eh CM rH CO iH c rH 4J (L) O % iH >l fO o CD CD c; fO rH rH 4J [ — 4J (U fO C/) (U o -d^ -p M-1 Q /II r 1 r* Ut CO \J I— 1 o C\l a m iH XirHOOr~rHLr)CDO lO rH =3' oomcoroc^ocTi^cMLOO CM rH Cn rH oocDCoOrH'^LOOOcr^o O rH OOrHrHOPOOrOOVDOO ix> r~ ^ rH ro CM ^C^10>X>OC^lOrHO'XlC^I rH rH CM CNl Csi Cr> OrHLOOr-OOCMCNJrHLOO 1X1 rH 00 OOOCOCMOOrHPOI^OO rH rIXl OrHOrH C^iOOO 00 lD ^ r00 rH •H H rH -P M 0 H m u fO H U H (0 •H >1 m fO to 0^ 3 u M a: 3 (0 rH -H •H u to T3 u (0 3 -H fO H 0 0) -rH o -H 0) 0) ITS c 0 3 0) u •H H ^1 M M m m >i •H O •H U -P > H Q) O H -H u H d) (0 3 C g M M si >, M -H M DH a) u u a, M 1 cr CP (0 O M-l (0 o cn 3 c; m T5 TS M 0) o •H (0 (0 m 3 3 (0 T5 IM S-J -H .H rH 0) tt) C C H > u >, >i >, m m (TJ c 3: -a; CM ct; p

PAGE 47

Figure 2-1. Map of the MacArthur Agro-Ecology Research Center showing the system of ditches that enhance drainage

PAGE 48

31 Figure 2-2. Location of wetland study sites at MAERC.

PAGE 49

32 c/ O 0) CL CO O o O O O CO O CD O o o CM seiodpei \o jeqiunN

PAGE 50

33 10 -1 0) o 0) Q. E 3 8 0.0 0.5 1.0 1.5 Wetland Size (ha) 2.0 2.5 600 500 Q) 400 0) o Q. D 03 t 300 200 100 0 0.0 0.5 1.0 1.5 2.0 2.5 Wetland Size (ha) Figure 2-4, Relationship between wetland size and anuran species richness (top) and abundance (bottom) Abundance and richness based on total captures from all 0.25m' traps within each wetland.

PAGE 51

34 10 n *o 0) Q. CO 8 E 7 48 "n 1 r 50 52 54 -n T" 56 58 60 62 64 66 Wetland Hydroperiod (wks) 600 (0 0) o Q(0 Q) J3 500 400 300 200 100 0 48 50 52 54 56 58 60 62 64 66 Wetland Hydroperiod (wks) Figure 2-5. Relationship between wetland hydroperiod and anuran species richness (top) and abundance (bottom) Abundance and richness based on total captures from all 0.25m'' traps within each wetland.

PAGE 52

c a. tn s r VI c 1 I I in o ra n r 5. I "5. Z c c C -H rH & (15 m CO ^3 o n) C -H P c o u o m -p c o en d H M Ti a> o a 0) u o o o (U iH o • "•^ & 0 4-> M to (1) T! 4-) C -M m n3 <-\ (U • 3 1 >. 0) M M O ^ & -H 0)

PAGE 53

36 c o (LUO) IIBjUlBy u 0) M o m rH d -H fd M • LT) (M CTi -P rH O -rH Lo 6 CTi O CTi >-! IX) I 00 rCTi lo (Ti rH o H o CN] I 0) 2 M Hi M H QJ Uj >i

PAGE 54

37 o o CO o CNJ (Luo) Lijclaa 0) Q rH O II c Q) — (P • > -a o o X3 -H (0 i-l 0) • 0) CO TS -P C rH d -P -rH (U M X5 -d 0) M -p m 4:: p -P 03 ax; 0) -P • n m u Ti 0) d -P 05 03 5 -P 0) -~ 5 o (U -p 03 (D U (D -P Q) T3 03 03 0) a)\ X ro -P CM 00 1 V) lO CM (U p x; 0) 0) tJi M u d -H 0 -H d X H -P

PAGE 55

38 en en -H TS n c 03 0) CO Q) CD u cu cu o c cfl >i a M o & • g in cu CTi 4-J CTi O) CD seiodpei p jeqiunN Cf) cu M Hi tji CU H O O CU

PAGE 56

39 Impacted Unimpacted Fiqure 2-10. Comparison of total sample composition between the 1994 and 1995 summer breeding seasons at wetlands that were (n=6) and were not (n=5) impacted by flooding from ditches containing fish predators. Species categorized as temporary site breeders were species that do not breed in sites with fish predators. Rana grylio Rana utricularia, Acris gryllus, and Hyla cinerea will breed with fish predators and were categorized as permanent site breeders.

PAGE 57

CHAPTER 3 SPATIAL AND TEMPORAL DYNAMICS OF TADPOLES AND AQUATIC INSECT PREDATORS DEVELOPING IN TEMPORARY WETLANDS Introduction Predation is a major source of mortality in anuran tadpoles (Calef 1973; Heyer et al 1975; Smith 1983), and is thought to have a large influence on breeding site selection by adult anurans (Magnusson and Hero 1991) Fish are major predators of tadpoles in permanent aquatic sites, and many species appear to avoid breeding in sites that contain fish predators (Hopey and Petranka 1994; Resetarits and Wilbur 1991) Although ephemeral wetland sites generally do not contain fish predators, the short duration of these sites makes them unsuitable for species with long larval developmental periods (Wilbur 1980, 1984, 1987). Thus differences in habitat duration and predator composition, coupled with interspecific differences in the selection of breeding sites by adults, can generate differences in species composition of larval anurans along gradients of habitat permanence (Werner and McPeek 1994) However, factors that influence tadpole success, such as abundance and composition of competitors or predators, vary in space not only among wetland sites (Gascon 1992; Smith 1983; Werner and McPeek 1994; Woodward 1983) but also 40

PAGE 58

41 among patches within wetlands (Banks and Beebee 1988; Diehl 1988, 1992) Therefore, the distribution patterns of tadpoles and predators at both the macrohabitat (breeding site) and microhabitat (patch site) scale can have a large influence on tadpole growth and survival, and thereby assemblage structure. Although the importance of considering scale has been widely discussed (Wiens 1989, 1992), few studies of tadpole ecology have addressed this issue. Processes that are important in population regulation at one scale of investigation may fail to operate at other scales (Wiens 1992) For example, whereas predator composition has a large influence on anuran breeding site selection, the spatial dynamics of predators and tadpoles at the within-wetland scale are poorly understood. If tadpoles occupy patches that provide increased protection from predators, such as areas with high cover (Babbitt and Jordan 1996; Chapters 4 and 6), or areas with lower predator densities (Banks and Beebee 1988), then differential habitat use within sites may be an important feature influencing predation rates, and possibly assemblage structure (Morin 1986) Spatial overlap at the within-wetland scale among tadpoles and predators may be high if both groups select microhabitats based on similar abiotic features such as water depth, temperature, or plant cover. Understanding such patterns, and how they vary in space and time, should

PAGE 59

42 provide valuable insight into the potential for differences in microhabitat selection to generate differences in species richness and success among various sites. For example, occupation of vegetated shallow areas of permanent aquatic sites that have deeper, poorly vegetated open-water areas may provide increased protection from fish predators (Diehl 1992; Mclvor and Odum 1988; Werner et al 1983). How distribution patterns influence predation rates on tadpoles developing in temporary wetlands where aquatic insects are the major predators is largely unknown. Several studies have found interspecific differences in tadpole distribution within wetland sites (e.g, Alford 1986; Diaz-Paniagua 1987; Heyer 1976); however, few studies have related tadpole distribution to predator distribution (Banks and Beebee 1988) The lack of research on within-site distribution patterns of tadpoles and aquatic insect predators leaves open the question of whether differential distribution patterns within wetlands is a potential mechanism for decreasing predation pressure on tadpoles. The goal of this study was to determine the spatial and temporal patterns of anuran larval and aquatic insect predators within several temporary wetland sites. I asked whether tadpoles or predators were differentially distributed in three zones based on water depth. Further, I asked whether habitat features influenced tadpole distribution to determine whether tadpoles differentially

PAGE 60

43 occupied patches based on physical or chemical habitat features. By examining distribution patterns within wetlands, I was able to determine whether differences in distribution could function as a mechanism, in addition to adult breeding site selection, for decreasing predation pressure on larval anurans. Methods Study Site Research was conducted at the MacArthur Agro-Ecology Research Center in Lake Placid, Highlands County, Florida. The site is an active cattle ranch consisting of 4,800 ha of improved and semi-improved pasture interspersed with emergent wetlands and oak ( Quercus virginiana) and palm ( Sabal palmetto ) hammock. The ranch was ditched extensively in the 1960 'S to enhance drainage, and most wetlands are connected to the vast system of ditches that ultimately enters into Harney Pond canal, which flows into Lake Okeechobee. Most wetlands are properly characterized as emergent freshwater marsh. Thick-stemmed emergents such as Pontederia cordata and Saqittaria lancifoiia, are dominant species in deeper water with Panicum hemitomon. Polygonum punctatum and Alteranthera philoxeroides as sub-dominant species. Outer, shallower areas of temporary wetlands are characterized by lowerstature plants such as Bacopa caroliniana Hydrochloa carolinensis and Diodia virginiana

PAGE 61

44 Sampling Methods From May 1994 through September 1995, I sampled 12 wetlands every three weeks (n=23 sampling periods) One wetland was dropped during the last si:: sampling periods due to damage from cattle trampling. At each site, I collected 9 samples using a stratified random sampling protocol. Three replicate samples were collected in each of 3 concentric belts corresponding appro:iimately to one third the radius of the wetland. Tadpoles and macroinvertebrates were collected with a 1/4-m^' open-ended, bo:: trap. Bo;: traps have been used to sample small fish and macroinvertebrates (Chick et al 1992; Freeman et al 1984; Kushlan 1981) and anuran larvae (Caldwell et al. 1980; Calef 1973; Pfenning 1990), and are effective for sampling in areas with comple:: vegetation structure. A bar seine fitting the diameter of the bo:: trap and fitted with fiberglass insect screening was swept through the water column (after vegetation had been cleared) until 3 consecutive sweeps yielded no additional captures. Samples were preserved in the field with lO-o buffered formalin. Macroinvertebrates were removed from formalin, rinsed thoroughly in water and sLored in alcohol in the lab. Tadpoles and macroinvertebrates were identified to species. For the purposes of this study, the macroinvertebrate assemblage was analyzed as a whole, rather than by species. The assemblage consisted of 11 species of odonates from the

PAGE 62

45 families Aeshnidae, Libellulidae and three species of Hemiptera from the families Belostomatidae and Naucoridae. Water depth was measured to the nearest centimeter inside each trap prior to removal of plants. Foliar coverage of above-ground vegetation was estimated ocularly, and then plants were removed and shaken to dislodge any macrofauna. Grass clippers were used to clip plants as necessary to remove all vegetation. Vegetation was placed in a mesh bag, which was spun in the air 20 times to remove e::cess water prior to measuring biomass to the nearest 0.1 kg. Water temperature C'C) and pH measurements were taken adjacent to each trap at 0.1 m below the water surface. Statistical Analyses Split-plot analysis of variance (ANOVA) was used to test for effects of wetland site, sampling period (time), wetland zone and the wetland zone and sampling period zone interactions on the habitat features water depth, water temperature, pH, cover, and plant biomass, and on tadpole and macroinvertebrate density and biomass. In addition, separate analyses were done on the three most abundant anuran species. Values from the three samples taken within each zone were averaged to calculate a mean value for each zone for all parameters. Tadpole and macroinvertebrate density and biomass values were log transformed (log (:'+!)) prior to analysis to reduce skewness. Examination of

PAGE 63

46 residuals for the measures of environmental parameters indicated that transformations were not necessary (Sokal and Rolf 1981) The main effects of wetland and zone were tested over the split-plot error (i.e., wetland x zone), whereas the time and time x zone effects were tested over the whole plot error (i.e., mean square error). The Bonf eronni-Dunn post hoc multiple comparisons tests was used to test for differences among the three zones when a significant (p<0.05) zone effect was detected (Day and Quinn 1989). Correlation analyses were performed to examine the relationships between the abundance of all tadpoles combined, as well as the abundance of the three most common anurans (H. sguirella R. utricularia and P. ocularis ) with habitat variables and macroinvertebrate abundance and biomass (Sokal and Rolf 1981) Results Physical Parameters Water depth, water temperature, pH, plant cover, and plant biomass all varied significantly among wetlands and across time (Table 3-1; Figures 3-1 through 3-5) The main effect of time, whicli reflects seasonal differences, accounted for most of the variation in water depth (37%) and water temperature (58%) In contrast, wetland explained a majority of the variation in pH (73%), plant cover (36%), and plant biomass (29%). Wetlands that were adjacent to

PAGE 64

47 upland hammocks tended to have lower pH values compared to wetlands that were surrounded by pasture. Wetland differences in plant cover and biomass indicate differences in plant communities, as well as differences in the areal extent of coverage of thicker-stemmed emergents. Significant differences among zones were detected for water depth, pH, plant cover and plant biomass, but not for water temperature (Figures 3-1 through 3-5) Both water depth and pH increased progressively from the outer to the inner zone. Whereas the zone effect explained 22% of the variation in water depth, it accounted for only 1% of the variation in pH. Cover levels were higher in the outer wetland zones compared to the middle and inner zones (Figure 3-4). Cover in the outer zone was nearly 100% at most wetlands throughout the study. Plant biomass also differed among zones; however, in contrast to cover where highest levels were in the outer zone, plant biomass was highest in the inner zone (Figure 3-5) Whereas difference among zones accounted for 20% of the variation in cover, it accounted for only 5% of the variation in biomass. Few interactions were significant, and most that were explained very little of the variation in the ANOVA model. The significant zone x wetland interaction for plant cover and biomass was due to differences among wetlands in the er.tent of the development of the inner thick-stemmed emergent zone.

PAGE 65

48 Tadpole and Predator Distribution Tadpole densities varied significantly among wetland, time, and wetland zone (Table 3-2, Figure 3-6) Wetland and time explained 9% and 28% of the variation in tadpole density, respectively. Densities were higher in the outer zone compared to both the inner and middle zones, but did not differ significantly between the inner and middle zones (Figure 3-6) However, zone explained less than one percent of the variation in tadpole densities. Such differences are probably biologically insignificant. In contrast to density, tadpole biomass did not differ significantly among zones (Table 3-2, Figure 3-7) Differences among wetlands explained 9% of the variation in tadpole biomass and time explained 23%. Although tadpole numbers were highest at the beginning of each summer breeding period, tadpole biomass was highest during the winter months. This difference reflects the dominance of the relatively larger tadpoles of R. utricularia in winter samples. The pattern for macroinvertebrate predators differed from that of tadpoles. Whereas zone differences were found for tadpole density but not biomass, the opposite was true lor predators (Table 3-2) Macr oiiivur Lebi a Le numbers did not differ among zones; however, macroinvertebrate biomass was higher in the middle zone compared to the outer zone. (Figures 3-8 and 3-9) Again, although zone differences in biomass were significant, they er.plained less than one

PAGE 66

49 percent of the variation in biomass. In contrast to tadpole density and biomass where time e;;plained most of the variation, wetland explained a higher percentage of the variation in macroinvertebrate number and biomass than did time. Wetland accounted for 34o and 24% of the variation in macroinvertebrate density and biomass, respectively, whereas time explained 11% and 17%. Although the numbers of macroinvertebrates varied significantly among wetlands, wetland size was not correlated with predator numbers (R=0.258, p=0.44) Species-Specific Responses Three species, H. sguirella R. utricularia and P. ocularis were abundant enough to examine separately. Wetland and time had siqnifirant effects on donsity of each species (Table 3-3; Figures 3-10 through 3-12) For H. sguirella time accounted for a majority of variation in tadpole numbers (40%) Wetland explained only 4% of the variation in H. sguirella numbers. The main effects of wetland and time e::plained 18% and 16%, respectively, of Lhe variation in R. utricularia densities. These effects each explained 12% ol the vdriaLion in P. ocularis densities. R. utricularia and P. ocularis breed throughout the year, so it is not surprising that time explained less of the variation in density of these two species compared to H. sguirella which is strictly a summer breeder. Differences in the

PAGE 67

50 pattern of distribution among zones were not significant for any of these species. Relationship to Physical Parameters Although significant relationships were found between tadpole numbers and many environmental variables, these relationships were very weak (Table 3-4). The strongest relationship was a negative correlation with water depth. More complex patterns emerge, and some stronger relationships were detected, when the three most abundant species are examined separately. H. squirella had a strong negative relationship with water depth and a strong positive relationship with temperature and pH (Table 3-4) In addition, H. squirella was negatively correlated with plant biomass. R. utricularia also had a negative relationship with water depth but this relationship was weak. In contrast to H. squirella R. utricularia was negatively correlated with temperature. The differences between these two species reflect the interspecific differences in seasonality of breeding activity. R. utricularia also was positively correlated with plant cover. P. ocularis numbers were not correlated significantly with any of the environmental variables measured.

PAGE 68

51 Relationship to Macroinvertebrates Tadpole numbers and macroinvertebrate numbers were not significantly related, and only a weak negative relationship was found between tadpole numbers and macroinvertebrate biomass (Table 3-4) No significant relationships were found between macroinvertebrate number or biomass and any of the three species examined separately. Discussion Patterns of tadpole abundance within wetlands provided little evidence that differential occupation within wetland zones was an important anti-predator mechanism. Although differences in tadpole abundance among wetland zones were statistically significant, these differences explained only one percent of the total variation in tadpole numbers. Examination of the entire tadpole assemblage as a group may obscure some species-specific patterns that may ex.ist; however, separate analyses of the three most abundant species did not reveal any differences in distribution relative to wetland zone. Because predator numbers and biomass also were distributed relatively equally among zones, differential use by tadpoles presumably would not have been an effective mechanism for avoiding predators. In addition to the lack of differences in distribution based on wetland zone, tadpoles showed weak relationships with habitat characteristics. Examination of the entire

PAGE 69

52 tadpole assemblage together revealed non-significant or weak correlations. However, patterns among the three species e;:amined separately indicated that the distribution of at least one species, H. squirella was related to some environmental parameters. This species was negatively correlated with depth and plant biomass, and positively correlated with water temperature and pH. The positive correlation with temperature was due to er.clusive summer breeding rather than differences within wetlands (which were not significant) Further, the positive correlation with pH was due to differences among wetland sites, rather than differences within wetlands. Both water depth and plant biomass were significantly lower in outer wetland zones, suggesting that the negative correlation between these parameters and H. squirella numbers was due to use of outer zones. However, ANOVA indicated that K. squirella numbers were not related to zone. Because H. squirella breeds shortly after wetland fills from rainfall, use of shallow wetlands and wetlands with less-developed zones of thickstemmed emergents probably contributed to the negative correlations. Thus, correlations of H. squirella with habitat features appeared to be largely related to early summer breeding and adult selection of wetland breeding sites R. utricularia numbers were related only to three variables, cover (positive), water depth (negative) and

PAGE 70

53 temperature (negative) Similar to H. squirella, differences may largely be related to differences among wetland sites, since R. utricularia numbers also did not differ significantly among zones. The negative correlation with temperature reflects winter breeding activity. Finally, P. ocularis numbers were not related to any variables measured. The results suggest that within a class of wetland, (i.e., temporary), predator avoidance may not be related to microhabitat selection, at least at the scale examined in this study. At the scale of wetland site, Gascon (1991) found that species-specific responses to habitat features at several temporary wetland sites were highly variable and were not consistent from year to year. Thus, whereas species may breed only in temporary sites, they may use sites that are poorly vegetated or well vegetated (see also Heyer 1976) Some studies that have examined distribution patterns of tadpoles have found differential distribution within sites. Generally, the patterns observed have been higher occupation of vegetated areas and lower occupation of deeper open-water areas or areas with very low cover (Afford 1986/ Banks and Beebee l^^6^6 ; Uiaz-Baniay ua iyU7; Hulumuzki lyfjy; Loschenkohl 1986) Because wetland sites in the current study were vegetated throughout, differences in results with previous studies may indicate that large differences in habitat features, such as cover or no cover, may be

PAGE 71

54 necessary for generating differential distributions. Although cover levels did vary significantly among zones, these differences may not have been large enough to affect microhabitat use. Observed patterns of tadpole distribution may be generated by a variety of processes, including differential predation, tadpole selection of microhabitats and adult selection of oviposition sites. The relative role of these, or alternative mechanisms, in generating observed patterns is unclear. Unfortunately few studies that have examined tadpole distribution have also examined predator distribution. An exception is work by Banks and Beebee (1988) who found that both predatory dytiscid beetle larvae and Bufo calamita tadpoles were non-randomly distributed relative to water depth and vegetation, with tadpoles concentrating in shallow water and beetle larvae being more common in non-vegetated, deeper water areas. An aspect of microhabitat use not addressed in this study is differentiation based on water column strata. Interspecific differences in use of the water column have been observed or inferred from tadpole morphology (e.g., lleyer 197J; L'eLeisun el al ifi^ll; sou dl^o lovLow in WLlbur 1980) Such interspecific differences often have been interpreted as a mechanism for avoiding exploitation competition; however, regardless of the mechanism generating this use pattern, such differences could result in

PAGE 72

55 differential predation pressure. For example, although aeshnid odonates such as Anax junius cling to vegetation, many libellulid odonates are benthic (Pritchard 1965) Thus, the foraging strategies of the macroinvertebrates could generate interspecific differences in predation on tadpoles in patches with differing predator composition. Many odonates and anurans oviposit shortly after wetlands fill with water; therefore, tadpoles developing in varying parts of a wetland may face similar levels of predation pressure. For these species, the susceptibility to predators will largely be a function of growth rates. Tadpoles that grow rapidly can effectively eliminate predation threats from all but the largest predators (Crump 1984; Richards and Bull 1990; Travis et al. 1985a; but see Crump and Vaira 1991) Such a mechanism is probably important for species such as H. sguirella Thus, although occupation of areas with high cover may decrease predation rates on H. sguirella tadpoles (Babbitt and Jordan 1996; Chapter 6), the timing of breeding, avoidance of sites with ,iiul rci[)i(i (jtuwLh iiKiy \>v I h.< iiu);-.l i iiif x ) r I 1 1 1 1 in<>cli.iM i sirir. for successful tadpole development in this species. Although breeding early may decrease predation pressure on some species, and may explain why the tadpoles of these species do not show strong microhabitat use, R. utricularia which is mainly a winter breeder, did not use this strategy. It bred in sites that contained both abundant number of

PAGE 73

56 predators as well as large predators (Figures 3-8 and 3-11; pers. obs). However, like H. squireUa, this species did not show strong patterns of microhabitat use. Thus, lack of microhabitat selection to avoid predation was not limited to early breeding species. Therefore, early breeding offers only a partial explanation for the lack of microhabitat selection. More likely, the broad and even distribution of the predators may make such a strategy ineffective in wetlands that do not vary greatly in habitat features such as plant cover. Although tadpoles did not exhibit strong correlation iLh many habxLaL cliai ac Lcii s L icl; LiuiL due;.; iiuL in./aii ccupation of patches that differ does not affect tadpole growth and survival (Holomuzki 1986a; Sredl and Collins 1992; Travis and Trexler 1986) Experimental studies have demonstrated the importance of increased cover in mediating predator-prey interactions in aquatic systems (reviewed in Heck and Crowder 1991), including interactions between tadpoles and aquatic insect predators (Babbitt and Jordan 1996; Chapters 4 and 6) The lack of strong responses to habitat features within wetlands contrasts with the large change in composition observed in responses to the introduction of water from ditches containing fish (Chapter 2). The shift from an assemblage dominated by species that breed in temporary sites to one dominated by species that will breed in sites wi o

PAGE 74

57 containing fish predators was interpreted as being due to adult breeding site selection (Hopey and Petranka 1994; Resetarits and Wilbur 1991) Thus, predator composition is important for determining breeding use at the macrohabitat level. Although abiotic habitat characteristics may determine use by some species, features within sites that fall within a "suitable" range for breeding may not be important in determining species use (Gascon 1991) In addition, use patterns within sites may not be strongly influenced by predators, except in sites with clear partitioning of habitat features (i.e., cover versus not cover) or predators (i.e, present/absent or f ish/macroinvertebrates) This study indicated that avoidance of sites with fish predators, breeding shortly after temporary ponds fill, and rapid development, appear to be the major anti-predator mechanisms used by anuran species that breed in temporary ponds, and that within-wetland habitat selection was of minimal importance. A better understanding of the potential role of microhabitat use as an anti-predator strategy could be gained by examining more closely the size distributions of predator and prey (e.g., Werner et al. 1983), as well as the foraging strategies of predators compared to the positional use by tadpoles.

PAGE 75

58 Table 3-1. Split-plot analyses of variance testing for the effects of wetland, zone, sampling period (time), zone :i wetland and zone k sampling period on estimates of depth, temperature, pH, cover, and plant biomass. The main effects of zone and wetland were tested over the split-plot error (zone X wetland), whereas other factors were tested over the whole plot error (mean square error) Wpt 1 and ;',one X Wetland Ti me 1 < 1 1 I'l -1 i .1 1 '.S. 7 •i I 00. 1 1 : ri 1 ObJ 1 1 1 .BP/ S.* 01 'i 1 1 'irifri 0001 oOi.il Temperat.u re Wr>t. 1 and \b f i in 1 1 ime h 1 iu 1 1 1 1 2 1.1 '1 Sij'1 i 1 -n OOnl I'l;,WRt. I and ^ .n ; Wetland I 1 me n ; I'lmke:: 1 dua i 1 1 1 ib 0 10 1 ;t 3 0 ri '8'j1 ne Wt1 1 alKl IK ; Wi-tlan-J { mo Time Res idua I 1 1 4 4 ,94 H'jhbi 3 1 'ibOR^. 6 / 4 4 4 7 8 9 94 9. 2 4 3H30.7 1 4 37] 1 b 9 1 31 14 R',4 1 1 (I'll!] ; 8 3 4 1 til fl ant. Bi'.m df Wi t la rid ne X Wet 3 and F^^T idiia 1 11 ['I 9^> 1 ,,3 I 17034247 1 3f)73bR8 4 1 1 1 300 ^ 1 (I I bi OlIRR .( 21 03-914 8. ..2 1 3 31) MU, 37 1 b3 2 3'>4 10.4 0 0 0 n 3 ri 0 0 o 1

PAGE 76

59 Table 3-2. Split-plot analyses of variance testing for the effects of wetland, zone, sampling period (time), zone wetland and zone sampling period on transformed estimates of tadpole and predator densities and biomass. The main effects of zone and wetland were tested over the split-plot error (zone x wetland), whereas other factors were tested over the whole plot error (mean square error) Tadpole Density Source df SS MS Zone 2 0. ,540 0. ,270 Wetland 11 6. 849 0. ,623 Zone X Wetland 22 0, 848 0, ,039 Time 22 21. ,351 0, .971 Zone X Time 44 3, ,308 0, 075 Residual 594 43. ,299 0 073 7 010 16. 162 0.529 13.314 1.031 0. 0044 0.0001 0. 9631 0. 0001 0. 4191 Tadpole Biomass Source df SS Zone 2 0. 136 Wetland 11 4 ,491 Zone X Wetland 22 0. ,878 Time 22 11, 503 Zone X Time 44 1, 599 Residual 594 32, .288 MS F P 0.068 1.704 0.2051 0.408 10.234 0.0001 0.040 0.734 0.8060 0.523 9.619 O.OOOi 0.036 0.668 0.9511 0.054 Macroinvertebrate Density Source df SS Zone 2 0, 078 Wetland 11 21. ,740 Zone X Wetland 22 2. ,258 Time 22 7, .332 Zone X Time 44 1 .282 Residual 594 31, .756 MS F P 0.039 0.379 0.6687 1.976 19.257 0.0001 0.103 1.920 0.0072 0.333 6.234 0.0001 0.029 0.545 0.9930 0. 053 Macroinvertebrate Biomass Source df SS Zone 2 0. 046 Wetland 11 1 548 Zone X Wetland 22 0, 101 Time 22 1 123 Zone X Time 44 0. ,217 Residual 594 3, ,387 MS F P 0.023 4.972 0.0165 0.141 30.601 0.0001 0.00b 0.806 0.7195 0.051 8.951 0.0001 0.005 0.864 0.7208 0. 006

PAGE 77

60 Table 3-3. Split-plot analyses of variance testing for the effects of wetland, zone, sampling period (time), zone x wetland and zone k sampling period on transformed estimates of H. squirella R. utricularia and P. ocularis densities. The^main effects of zone and wetland were tested over the split-plot error (zone x wetland), whereas other factors were tested over the whole plot error (mean square error) H squirella Source df ss MS F P Zone 2 0, .042 0, .021 3. .246 0. .0582 Wetland 11 0. .813 0, .074 11, ,398 0. .0001 Zone X Wetland 22 0. .143 0, .006 0. ,348 0. ,9978 Time 22 8. .084 0. .367 19, ,726 0, .0001 Zone X Time 44 0, .332 0, .008 0, .405 0, ,9998 Residual 594 11 .065 0 .019 R. utricularia Source df SS Zone 2 0. 072 Wetland 11 7 187 Zone X Wetland 22 0. 766 Time 22 6. 248 Zone X Time 44 3. 272 Residual 594 22. 592 P. ocularis Source df SS Zone 2 0. 044 Wetland 11 1 240 Zone X Wetland 22 0. 242 Time 22 1 205 Zone X Time 44 0. 161 Residual 594 7 047 MS F P 0 .03 6 1 .028 0. 3743 0 653 18 .767 0. 0001 0 .035 0 .915 0. 5746 0 .284 7 .468 0. 0001 0 074 1 .955 0. 0003 0 .038 MS F P 0 022 2 .011 0. 1578 0 .113 10 .243 0. 0001 0 .011 0 .928 0. 5581 0 .055 4 .619 0. 0001 0 .004 0 .308 1. 0000 0 .012

PAGE 78

61 Table 3-4. Correlations between all species, Hyla squirella Rana utricularia and Pseudacris ocularis with water depth, temperature, pH, plant cover, total plant biomass, and macroinvertebrate density and biomass. Water depth vs H squirella R. utricularia P ocularis All species 0.432 0.197 0.051 0.230 0.0008 0.0020 0. 6752 <0.0001 Temperature vs pH vs Plant biomass vs Plant cover vs. H. squirella R, utricularia P ocularis All species H. squirella R. utricularia P ocularis All species H squirella R. utricularia P ocularis All species H squirella R. utricularia P ocularis All species 0. 391 0. 0027 -0. 244 0. 0002 0. 152 0. 2254 0. 122 0. 0017 0. 386 0. 0030 0. 150 0. 0557 -0. 181 0. 1948 0. 130 0. 0025 -0. 295 0. 0257 -0. 014 0. 8218 -0. 026 0. 8292 -0. 069 0. 0691 0. 132 0. 3291 0. 254 <0. 0001 0. 190 0. 1155 0. 151 0. 0001 Macroinvertebrate H. squirella -0.269 J'^^^^ density vs. R. utricularia "O-O^S O-^Jl^ P. ocularis -0.073 0.6139 All species -0.073 0.0530 Macroinvertebrate biomass vs. H squirella R. utricularia P ocularis All species 0.029 0. 100 0.245 -0.136 0.8477 0.1344 0.0838 0. 0003

PAGE 79

62 Sampling Period 35 n 30 E 25 20 •4—' Q. 15 0 Q 10 5 0 ^ 2 3 4 5 6 7 8 9 10 11 12 Wetland E o 30 n 25 20 15 H Q. 0) 10 Q 5 0 Outer Middle Wetland Zone Inner Figure 3-1. Variation in water depth (mean +1 SE) among sampling dates (top), wetland (middle), and wetland zone (bottom) Zone means with different letters are significantly different.

PAGE 80

63 32 o o 0) i_ 26 Z2 -t— ro i_ 22 ^ I I I I I I Date 30 g 25 9? 20 I 15 H Q. 10 Q) 5 I0 2 3 4 5 6 7 8 9 10 11 12 Wetland 30 n 0 25 9? 20 1 15 Q. 10 H E ^
PAGE 81

X Q. 7 n I I I I I 64 I I I I I I I I I I I I I I Date 7 6 5 I 4 ^ 3 2 1 0 1 2 3 4 5 6 7 8 9 10 11 12 Wetland X D. 6 5 4 3 H 2 1 0 Outer Middle Wetland Zone Inner Fiqure 3-3. VaciaLion in pll (mean +1 SF,) amonq sampllnq dates (top), wetland (middle), and wetland zone (bottom) Zone means with different letters are significantly different

PAGE 82

65 2 40 Q20 1 2 3 4 5 6 7 8 9 10 11 12 Wetland 0) > o O 0) o Q) CL 100 80 60 40 20 0 Outer Middle Wetland Zone Inner Figure 3-4. Variation in plant cover (mean +1 SE) among sampling dates (top), wetland (middle), and wetland zone (bottom) Zone means with different letters are significantly different.

PAGE 83

66 1000 -| 900 800 700 600 500 400 TH I I I \ I I I I I tC^ ^^. x*^
PAGE 84

14 n 1 2 3 4 5 6 7 8 9 10 11 12 Wetland E If) CN O W 0) O Q. D (0 0 Outer Middle Wetland Zone Inner Figure 3-6. Variation in tadpole numbers (mean +1 SE) among sampling date (top), wetland (middle), and wetland zone (bottom) Zone means with different letters are significantly different.

PAGE 85

o 1 0 68 I I I I I Date 3 n to CN 0 L J. X 1 i 1 2 3 4 5 6 7 8 9 10 11 12 Wetland 1.5 1.0
PAGE 86

E ID (N O 0) •*-> t: (U > c o i_ o CO 69 CM E o w t_ (U > c o o 6 n 5 i 1 0 10 8 6 4 2 0 I I I I I I 1 — m \ I I I I I \ I I I I to -sj A -?> A ,4V ^ .-c ,rv .•(^ > > > lo 4 o I ^ -Q 2 I ^ o x^V v*^ -0/ W Date 1 2 3 4 5 6 7 8 9 10 11 12 Wetland Outer Middle inner Wetland Zone Figure 3-8. Variation in macroinverLebraLe predator numbers (mean +1 SE) among sampling date (Lop), wetland (middle), and wetland zone (bottom)

PAGE 87

70 I I I I I I Date E
PAGE 88

71 E lO CM O 7n o Q. T3 ro 0 1 2 3 4 5 6 7 8 9 10 11 12 Wetland CM E in CM o 0) o Q. T3 (0 0.5 0.4 0.3 0.2 0.1 0.0 Outer Middle Inner Wetland Zone Figure 3-10. Variation in Hyla squirella numbers (mean +1 SE) among sampling date (top), wetland (middle), and wetland zone (bottom)

PAGE 89

72 4 n I 3 d o & 1 CO I0 X X i 2 3 4 5 6 7 8 Wetland 9 10 11 12 0 T Outer Middle Inner Wetland Zone Figure 3-11. Variation in Rana utriculari a nun\bers (mean +1 SE)' among sampling date (top), wetland (middle), and wetland zone (bottom)

PAGE 90

73 2.0 n <^ <^ ^
PAGE 91

CHAPTER 4 EFFECTS OF COVER AND PREDATOR SIZE ON SURVIVAL AND DEVELOPMENT OF RANA UTRICULARIA TADPOLES Introduction Size-limited predation is a particularly important process during amphibian development. Predation levels in systems with size-limited predators are largely dependent on the relative sizes of predator and prey (Ebenman and Persson 1988) Numerous studies have shown that predation rates on tadpoles are a function of tadpole body size, and a majority of mortality due to predation occurs early in development (e.g., Banks and Beebee 1988; Cronin and Travis 1986; Richards and Bull 1990, Semlitsch 1990, Semlitsch and Gibbons 1988; Travis et al. 1985a). Thus, it has been suggested that selection for rapid growth during the larval phase may, at least in part, be a response to predation by size-limited predators (Travis 1983, Travis et al. 1985b). For anurans breeding in temporary aquatic habitats, oviposition shortly after the breeding site fills provides a potentially important mechanism for decreasing predation pressure. For example, because dragonflies, a major predator of tadpoles in temporary breeding sites, also oviposit after pond filling (Ward 1992), tadpoles that grow rapidly often 74

PAGE 92

75 are able to reach a size refuge and decrease their cumulative risk of predation. Because a majority of predation on tadpoles occurs early in development, the "growth race" between dragonfly naiads and tadpoles can have a particularly large influence on the ability of predators to regulate tadpole populations (Caldwell et al. 1980; Tejado 1993; Travis et al 1985a). Not all species that use temporary wetlands employ this reproductive strategy, however. The southern leopard frog ( Rana utricularia ) breeds year round; however, the majority of breeding is done during the winter-spring breeding season. In addition, this species has very broad breeding site associations. This can result in larvae being exposed to an assemblage of aquatic insect predators that are already established, and therefore large in size relative to newly hatched tadpoles. Thus, early predation pressure by large predators may be significant. Under such circumstances, physical features of the environment, such as complex habitat structure, may provide an important mechanism for lowering predation rates. Habitat structural complexity has been shown to be an important factor mediating predator-prey interactions (reviewed in Heck and Crowder 1991) Increased habitat complexity can reduce predation rates by providing cover or f)atLLal rorugc ar'cas for pr'cy (o.q, P'olsom and Collins 190'!; Rozas and Odum 1988), or by decreasing foraging success of

PAGE 93

76 predators because of decreased maneuverability or visual range (e.g., Crowder and Cooper 1982; Savino and Stein 1982; Werner et al. 1983) R. utricularia tadpoles can grow to a large size relative to most aquatic insect predators, thus mechanisms that decrease early predation pressure may have a particularly important effect on overall survival rates. In this study I exposed tadpoles of the southern leopard frog to two levels of cover and two size classes of the same predator. By doing so I could determine whether increased habitat complexity (i.e., cover) provides a mechanism for decreasing predation and further whether such protection is dependent on predator size. I predicted that predation rates would be lower at high cover levels and that large predators would be more effective than small predators regardless of cover level. In addition, because the thinning effects of predators can release prey from competition (Wilbur 1987, 1988), I predicted that tadpoles developing in treatments with higher predation rates (i.e., low cover and large predator) would have enhanced growth. Methods Experimental Design I er.amined development and survival of tadpoles in a factorial experiment in which I manipulated cover level (high versus low) and predation (large, small, or no predator) in a complete randomized block design replicated

PAGE 94

77 four times. I established a rectangular array of 24, 1.14 m diameter plastic wading pools with tight fitting fiberglass screen tops at the MacArthur Agro-Ecology Research Center (MAERC) in Lake Placid, Highlands County, Florida. On 8 March 1995, I filled pools with 95 1 of well water (depth = 12 cm) and added 1000 ml of well-mixed algae and zooplankton collected from several wetlands. I added an additional 500 ml of algae and vegetation (see below) on 16 March 1995. Water overflow was prevented by a drainage pipe. I checked pools weekly to determine if water levels were maintained at the proper level by rainfall. When rainfall was not adequate to compensate for evaporation, I added well water (3 times during the experiment) I provided two levels of cover, 2000g or 500g (wet weight), of the aquatic plant Hydrochloa carolinensis H. carolinensis is a thin-leafed grass that provides complex structure throughout the water column. It is abundant in wetlands at MAERC. Both treatments provided cover throughout the water column; however, the higher cover level provided relatively dense cover whereas the low cover treatment provided relatively sparse cover. Predator treatments were large (mean + 1 SE : 4.18 + 0.26 cm; n = 6) or small (2.35 0.11 cm; n = 6) Tramea Carolina (Odonata: Libellulidae) naiads, or no predator. On 23 March 1995, I haphazardly added 30 Gosner (1960) stage 25 (0.04 + 0.01 g n = 15) tadpoles to each pool. I

PAGE 95

78 added predators the following day. When metamorphosis commenced, I checked pools twice a day for tadpoles with emerged forelimbs (stage 42; Gosner 1960) I held tadpoles individually in 500-ml plastic jars until tail resorption (stage 46; Gosner 1960), and then recorded wet mass (to 0.001 g) During the second week of May 1995, the area experienced an unseasonal heat wave with daytime highs exceeding 32"C for several days. At the end of the heat wave air temperatures exceeded 38"C and water temperatures reached 40"C. This event caused a massive die-off of tadpoles on 15 May 1995. At this stage of the experiment 71% of tadpoles surviving to that point already had metamorphosed successfully. I collected the dead tadpoles and sacrificed any tadpoles that remained alive and obtained developmental stage. Thus, 15 May 1995 was considered the end of the er.periment. Because of the die-off, response variables were the mass of tadpoles that reached Gosner (1960) stage 46 (stage 42 by 15 May 1995), age of metamorphs (calculated as the start of the experiment to stage 42), and percent survival. Because tadpoles were large compared to prcdaLors, I assumed LliaL any Ladpoles in the pools on 15 May 1995, whether dead or alive, would have survived predation. Thus, survival was based on the number of tadpoles that reached stage 46 plus pre-metamorphic tadpoles that remained at the time of the die-off.

PAGE 96

79 Statistical Analysis Treatment effects were analyzed using two-way fi::ed effect analysis of variance (ANOVA) Mean values per pool were analyzed because measurements of individuals within pools are not independent. One replicate from the no predator low cover treatment was dropped from the er.periment because the pool developed a leak causing rapid water loss resulting in tadpole mortality. Percentage data vjctc ciiujularly I t .uil; fui mod .ukI oIIum i c';;[)on;;o VcUi.ibU^;; wlm o log transformed (log(x + 1) for mass values) prior to analysis. I performed orthogonal contrasts to test for predictions regarding effect of predators (high predation by large predators) and cover (higher predation at lower cover), as well as prediction regarding enhanced growth. When interactions were significant (p<0.05), I conducted separate comparisons within treatment levels. Results Survival Predator treatment had a significant effect on total survival, accounting for 71.6% of variation in survival. I lowi .• VI.' I U K' ( • I I ( M : I u I | j i t m I, i I i ; i wi • i c i li [ i u I. i 1 1 iH i < ( j v. i (Figure 4-1; Table 4-1) Orthogonal contrasts indicted that survival of tadpoles was lower in treatments with large predators compared to treatments with no or small predators. This was true at high (F=34.66, p<0.001) and low (F=11.80,

PAGE 97

80 p=0.004) cover treatments. The interaction between predator and cover was due to lower survival within the low cover treatment compared to the high cover treatment for tadpoles e::posed to large predators. However, the interaction explained only 8.3% of variation. Small predators were largely ineffective. Survival of tadpoles in pools without predators did not differ from that of tadpoles in pools with small predators, regardless of cover (contrasts: low cover F=0.11, p-0.74; high cover F=0.07, p=0.8Q). The main effect of cover was not significant. Overall, tadpole survival was lowest under the large predator low cover treatment. Mass at Metamorphosis Mass at metamorphosis was affected by the main effects of predator and cover, and by the interaction of cover and predator (Figure 4-1; Table 4-1) Within the large predator treatment, tadpole mass was higher within the low cover treatment compared to the high cover treatment (F=24.42, p-0.002). Effects of small (F=0.02, p=0.88) and no predator (F=1.16, p=0.30) treatments were not significantly influenced by cover. Effects of predator treatment on tadpole mass at metamorphosis were not significantly different within the high cover treatment (F=2.748, p=0.12); however, in low cover treatments, tadpoles within the large predator treatment were significantly heavier than those in the small (F=37.51, p<0.001) and no (F=24.5, p
PAGE 98

81 predator treatments. Tadpole mass did not differ between the small predator and no predator treatments within the low cover treatment (F=0.41, p=0.53). Age at Metamorphosis The main effect of predator and the interaction of predator and cover had a significant effect on age at metamorphosis; however, the main effect of cover was not significant (Figure 4-1; Table 4-1). At high cover levels, age at metamorphosis was not significantly affected by predator treatment (F=0.01, p=0.92). However, within the low cover treatment, tadpoles exposed to large predators metamorphosed earlier (F=11.34, p<0.001) compared to the small and no predator treatments, which did not differ (F=0.001, p=0.97) Discussion Predator size had a much larger effect on tadpole performance than did level of habitat comple;:ity. High cover level did reduce predation levels compared to low cover when large predators were present, suggesting that habitat complexity does provide increased protection from predators. The most parsimonious interpretation for lack of effects by small predators is that tadpoles were able to grow large enough early in the experiment to reach a size refuge and escape predation. Thus, early growth by R. utricularia

PAGE 99

82 tadpoles developing in temporary environments can greatly reduce predation pressure from aquatic insects that are oviposited around the same time. Similarly, growth probably decreased overall predation rates by large predators; however, tadpoles were within the gape limits of large predators for a longer time, allowing increased predation by odonate naiads. Because tadpole survival was higher at high cover levels when large predators were present, this suggests that increased cover decreased the foraging efficiency of T. Carolina naiads. A possible alternative hypothesis is that tadpoles raised under higher cover grew faster and therefore decreased predator success by surpassing the gape limits of predators. This is unlikely, however, because growth of tadpoles on the other treatments differed little relative to cover level. Thus, the most likely er.planation for decreased predation at high cover levels is interference in predator foraging. Decreased foraging efficiency with increasing habitat complexity has been found in other studies of aquatic insect predators, including species that are active foragers (e.g., Anax junius ) and sit and wait predators (e.g., Belostoma sp.) (Chapter 6). Heck and Crowder (1991) predicted that less mobile predators would actually have increased foraging efficiency in complex habitats due to increased perch sites for foraging. Banks and Beebee (1988) demonstrated that increasing complexity from no vegetation to some vegetation

PAGE 100

83 increases foraging success of odonate naiads on Bufo calamita tadpoles. Their study provides some support for Heck and Crowder's prediction. However, a majority of research to date suggests that, like their mobile piscine counterparts, aquatic insects often have decreased foraging efficiency in structurally complex habitats. R. utricularia tadpoles reduce activity in the presence of An ax junius naiads (Chapter 5) Reduced activity has been identified as a potentially important anti-predator mechanism in many amphibian species, particularly among species that breed in permanent aquatic habitats where fish are the dominant predators (Werner and McPeek 1994; Werner and Anholt 1993) Research suggests that the anti-predator response is general, rather than predator specific (Stauffer and Semlitsch 1993) Thus, reduced activity may have afforded R. utricularia some protection from predation. The results of this study, however, suggest that size differences between predator and tadpole may be more important than prey activity, at least when aquatic insects are the predators. For example, increased cover also decreased predation levels on Hyla squirella tadpoles, a species that has liiyh activity levels even in tlie pieseuce of predators (Chapter 6) The thinning effects of predators resulted in larger tadpole size and more rapid development (earlier age at metamorphosis) These results agree with other studies of

PAGE 101

84 predation on tadpoles that have found enhanced growth of surviving tadpoles (Wilbur 1987, 1988) Under these circumstances, predation can provide an overall positive effect on population size, resulting in a mutualistic relationship between predators and tadpoles that escape predation. Particularly in environments where habitat desiccation is a significant mortality threat, the thinning effects of predators can lead to increased survival among individuals that escape predation (Wilbur 1987) For example, in cattle tank experiments where habitat drying was manipulated, Wilbur (1987) demonstrated that R. utricularia tadpoles subject to predation by newts actually had higher survival compared to tadpoles in tanks without predators. When predators were absent, competitive effects led to decreased growth and prolonged development, resulting in death due to desiccation. In this study, the enhanced growth of tadpoles in the large predator treatments suggests that tadpoles in treatments where predation was low experienced intraspecif ic competition. Even so, this had little effect on susceptibility to predation, which was determined more bLiungly by piedaLui size Llian any cuitipu L i L i vc eliecLs diuoiig tadpoles. This suggests that competitive effects were probably not significant until after tadpoles had already grown beyond the gape-limits of T. Carolina naiads.

PAGE 102

85 The importance of relative size differences between tadpoles and aquatic insect predators has been demonstrated many times (Brodie and Formanowicz 1983; Caldwell et al. 1980; Crump 1984; Formanowicz 1986; Richards and Bull 1990; Travis et al. 1985a). Although some studies have shown that larger tadpoles may be selected over small tadpoles if both size classes are within the gape-limits of predators (Crump and Vaira 1991; Tejedo 1993), predation on tadpoles that are capable of growing beyond the gape-limits of predators is concentrated on smaller, early-stage tadpoles. This study provides additional evidence of the importance of relative size differences between predators and tadpoles. Further, it demonstrates that habitat structural complexity can play an important role in mediating predator-prey interactions, even when tadpoles start out at a size disadvantage relative to predators

PAGE 103

86 Table 4-1. Summary of ANOVA for responses of Rana utricularia tadpoles to cover and predator treatments. CD is the coefficient of determination, which is the percentage of variation in the response variable attributable to the treatment effect. Percent Survival Source of variation df SS F P CD Block 3 0 .03834 0 68 0 578 2.3 Predator 2 1 19807 31 92 <0 001 71.6 Cover 1 n 0'^4q3 1 8 6 0 194 2 1 Predator ; Cover 9 U 1 ^ Q T 1 J J? D J> J 1? n 051 • ^ X 8 3 Error 1 A 1 H n Mass (g) Source of variation df SS F p CD Block 3 0 01888 1 .97 0 .165 7.6 Predator 2 0 .10474 16 .40 <0 .001 42.2 Cover 1 0 .03549 11 .11 0 .005 14.3 Predator : Cover 2 0 04429 6 .93 0 .008 17.9 Error 14 0 .04472 Age (days) Source of variation df SS F P CD Block 3 0 00009 0 .29 0 .835 2.5 Predator 2 0 .00088 4 .04 0 .041 24.2 Cover 1 0 00007 0 61 0 .450 1.9 Predator : Cover 2 0 .00107 4 .89 0 .025 29.4 Error 14 0 .00153

PAGE 104

87 E CD 3 w w (0 2.2 2.0 1.8 1.6 1.4 1.2 1.0 0.8 None Small Large 48 47 46 45 D) 44 < 43 42 (0 D None Small Large Predator Treatment Figure 4-1. Mean (+ 1 SE) responses of Rana utricularia tadpoles to cover and predator treatments: percent survival (top), wet mass at metamorphosis (middle), and age at metamorphosis (bottom) Squares indicate high cover, circles indicate low cover.

PAGE 105

CHAPTER 5 EFFECTS OF FOOD AVAILABILITY AND RISK OF PREDATION ON BEHAVIOR AND GROWTH OF RANA UTRICULARIA TADPOLES Introduction Investigations into the role that predation plays in the regulation of prey populations and assen)lage structure have led to an increased understanding that the effects of predators are broader than those associated with direct predator-induced mortality. Studies documenting keystone predator effects (Paine 1969) and trophic cascades (Carpenter et al. 1987; Power 1990) demonstrate the large effects predators can have on assemblage structure even when such effects are indirect. Direct, but non-lethal, effects (sensu Strauss 1991) can have a large impact on prey populations as well. Such effects are often manifested through a shift in the behavior of prey in response to the presence of a predator. Behavioral adjustments to avoid predators may involve a shift in microhabitat use or a change in timing or amount of activity (reviewed in Lima and Dill 1990). Facultative reductions in activity in response to the presence of a predator have been documented in a wide variety of ta::a including aquatic insects (McPeek 1990; Peckarsky et al. 1993), crustaceans (Stein and Magnuson 1976), fishes (Eraser 88

PAGE 106

89 and Gilliam 1987), and anuran larvae (Lawler 1989; Semlitsch and Reyer 1992; Skelly 1995; Skelly and Werner 1990; Werner 1991) Reduction in activity may be a widespread response to predators because prey activity is a large component of the process through which many predators detect prey (Lima and Dill 1990; Werner and Anholt 1993) Thus, there should be strong selection pressure against activity in the face of predation Reduction in activity, however, involves a trade-off between decreasing risk of predation and obtaining resources because both risk (predation) and reward (resources) are positively related to activity (Werner 1992b; Werner and Anholt 1993) The trade-off between predation risk and resource acguisition suggests that an important non-lethal effect of predators on prey is a decrease in growth (Skelly and Werner 1990) Both intraspeci f ic and interspecific differences in activity have been related to competitive ability in tadpoles: active tadpoles are often competitively superior (Lawler 1989; Morin 1983; Werner 1991, 1992a; Woodward 1983) However, active tadpoles are also more susceptible to predation (Anholt and Werner 1995; AzevedoRamos et al 1992; Skelly 1994). This inverse relationship between susceptibility to predation and competitive superiority associated with activity level can have a large influence on assemblage structure (Morin 1983, 1986)

PAGE 107

90 The strength of non-lethal predator effects, in turn, may be largely influenced by background resource levels. Theoretical models predict that activity should be inversely related to mortality risk (Abrams 1993; McNamara and Houston 1987, 1994; Werner and Anholt 1993) These models also predict that foraging activity should be inversely related to resource level when mortality risk is dependent on foraging activity (Abrams 1993; McNamara and Houston 1987, 1994; Werner and Anholt 1993) For example, Anholt and Werner (1995) found that activity rates of Rana catesbeiana tadpoles exposed to predators were inversely related to food level, and this resulted in lowered mortality for tadpoles on a high food level. However, when mortality risk is independent of activity, foraging activity should increase with increasing resource levels (McNamara and Houston 1994) Therefore, both predation risk and resource availability should be er.amined to understand the consequences of behavioral adjustments on individual growth. The purpose of this study was to examine the behavioral responses and growth performance of tadpoles of the southern leopard frog ( Rana utricularia ) raised on different resource levels both in the presence and absence of a non-lethal predator. By altering both the perceived risk of predation and resource levels, I was able to examine the relative importance of these factors and how they interacted during the development of anuran larvae.

PAGE 108

91 Model Predictions Growth and Development The Gilliam-Werner model (Gilliam 1982; Werner 1986) predicts that fitness in pre-reproductive individuals is ma::imized by minimizing the ratio of mortality (u) to growth rate (g) Based on this model, factors that increase mortality in the larval environment should result in decreased size at metamorphosis (see hypothetical u/g curves in Werner 1986) Increases in resource level increases growth rate and therefore results in the opposite effect; increased size at metamorphosis. Reduction in activity in response to predators decreases foraging rates and therefore should result in smaller size at metamorphosis. The Gilliam-Werner model makes no prediction about the length of the metamorphic period because reduced size could result from either early metamorphosis at smaller size or prolonged larval period under poorer growth opportunities resulting in smaller size (Wilbur and Collins 1973) Activity If tadpoles are responding to predators they should reduce activity in their presence (Abrams 1993; McNamara and Houston 1987, 1994; Werner and Anholt 1983) Further, tadpoles under a higher food regime should be less active than those under a poor regime of nutrition in the presence of a predator (Werner and Anholt 1993) When predators are

PAGE 109

92 absent, activity rates should be positively related to resource level (McNamara and Houston 1994). Based on these models, I predicted that tadpoles would reduce activity in response to the presence of predators and that this would result in a smaller size at metamorphosis. Further, I predicted that activity level would be positively related to food resource when predators were absent but negatively related to resource level when predators were present Methods I examined growth and behavior of tadpoles in a factorial experiment in which I manipulated resource levels (low versus high per capita food level) and non-lethal ex.posure to larval dragonflies An ax Junius (Odonata: Aeshnidae) On 19 February 1995 I collected a single R. utricularia egg mass from an ephemeral wetland at the MacArthur Agro-Ecology Research Center in Lake Placid, Florida. The egg mass was held in a plastic wading pool filled with rain water and allowed to hatch. Late instar A. junius naiads were collected from ephemeral wetlands at the research center and maintained in the laboratory on a diet of R. utricularia tadpoles. On 1 March 1995 I haphazardly added four Gosner (1960) stage 25 tadpoles (mean + 1 SE: SVL = 5.75 + 0.3 mm, preserved wet mass = 0.031 + 0.003 g) to each of 24 plastic containers (34x29x14 cm) Containers were filled with seven

PAGE 110

93 liters of aged well water and arranged on laboratory shelves in six blocks of four containers each. Fiberglass bags (8x12 cm) were suspended along one side of the container to hold predators. Bags prevented predators from preying on tadpoles but allowed tadpoles to chemically detect predators. Bags in predator treatments contained one late instar A. j unius naiad, whereas empty bags were used as controls in nopredator treatments. Once a week, following water changes, each predator was fed a single R. utricularia tadpole. Predators were replaced with new individuals as necessary throughout the experiment. Tadpoles were fed a finely ground mixture of Purina rabbit chow and pelleted fish food (3:1 by mass). Low food treatments received per capita rations (7.5o body mass) that were limiting, whereas high food treatments received per capita rations (15% body mass) that promoted growth and development (Alford and Harris 1988; Werner 1992a) The quantity of food added to each container was determined by the following formula: per capita food ration x mean mass of the tadpoles in container :i number of tadpoles surviving in container x number of days to next feeding. Within each tood treatment, the mean tadpole mass per container was determined once a week by weighing tadpoles from containers without predators. I changed container water every week at the beginning of the study, twice a week by week three, and every other

PAGE 111

94 day by week five to avoid water fouling. A light-dark cycle of 14:10 was provided by a series of florescent lights and natural lighting. Water temperature varied between 22-24"C during the experiment. Behavioral Observations I used scan sampling to measured activity rates and location of tadpoles within containers during the first four weeks of the experiment (n=18 observations) I made observations between 0900-1400 h on non-feeding days and terminated observations at 4 weeks because older tadpoles appeared to react to my presence. I recorded the number of tadpoles moving (swimming or feeding) and the number of tadpoles on the side of the container opposite the predator bag and converted counts to percentages for each bin averaged across the 18 observation periods. Response Variables and Statistics Metamorph mass (wet weight) was measured when at least one forelimb had emerged (stage 42, Gosner 1960) The larval period was defined as the start of the experiment to forelimb emergence. Container means were used as response variables for all statistical analyses. Mass and larval period data were log-transformed and activity and distribution data were angularly transformed to homogenize variances and normalize data prior to analysis. I used

PAGE 112

95 factorial ANOVA testing for the effects of laboratory block, predator presence, resource level, and the interaction of predator presence and resource level on response variables (Snedecor and Cochran 1980) I performed orthogonal contrasts to test my predictions regarding effects on mass and activity. Additional contrasts determined whether the presence of a predator affected the length of the larval period. Results Effects on Survival Survival to foreliiiib emergence was 98 o. One tadpole jumped out of its container and one died of unknown causes. Effects on Growth and Development Food treatment had a significant effect on tadpole mass at metamorphosis, accounting for 49% of the variation in tadpole mass (Table 5-1) Tadpoles on the high food treatment were 25.7% larger than tadpoles on the low food treatment. The main effect of predator was not significant; however, there was a significant interaction between food and predator treatment (Figure 5-1) Tadpoles on the low food treatment were 14.3c smaller when predators were absent, whereas tadpoles on the high food treatment were 8.2b larger in the absence of A. junius Orthogonal contrasts indicated that differences in mass within food

PAGE 113

96 treatment were significant for tadpoles on low food (F=4.88, p=0.043) but not for tadpoles on high food (F=1.659, p=0.217) Both food and predator treatments affected age at metamorphosis (Table 5-1) Food had the largest effect, explaining 77.7% of the variation in larval period, whereas predator treatment accounted for only 8.5% of the variation (Figure 5-1). Lowered food levels resulted in a 12. 3o increase in larval period, whereas non-lethal er.posure to a predator prolonged the larval period by 4.3% (about 2 days). The interaction between food and predator was marginally significant (Table 5-1) Orthogonal contrasts showed that tadpoles on low food reached metamorphosis significantly earlier when predators were absent (F-14.15, p=0.002); however, age at metamorphosis was not affected by predator treatment for tadpoles on high food (F=2.32, p=0.148). Effects on Tadpole Activity and Distribution The presence of A. junius had a significant effect on the spatial distribution of tadpoles, accounting for 45.2% of the variation in distribution (Table 5-2). Overall, 69.8% oi the tadpoles were on the side uL Lliu conLaiuci opposite the predator when A. junius was present compared to 54.8% when predators were absent (Figure 5-2) Neither food nor the interaction of food and predator was significant.

PAGE 114

97 The effects on activity were more complex. Predator presence had a significant effect; tadpoles in predator containers reduced activity by 60. 80 (Figure 5-2). Both food and the interaction of predator and food treatments were marginally significant (Table 5-2) When predators were present, activity levels were similar on both food treatments. Activity levels were higher when predators were absent for both the high food (F-29.57, p=0.0007) and low food (F=5.94, p=0.028) treatments; however, tadpoles on the high food treatment had a higher level of activity compared to tadpoles on the low food treatment when predators were absent (Figure 5-2) Discussion Behavioral Responses Tadpoles altered both spatial distribution and activity levels in response to the presence of predators. Because activity level is positively correlated with mortality risk (Skelly 1995; Werner and Anholt 1993), reduced activity can be an important mechanism for decreasing risk of predation. Reductions in activity in response to predators have been found in several oLlier anurau species including Pseudaci is crucifer (Skelly 1995), Bufo terrestris (Skelly and Werner 1990), Rana clamitans and Rana catesbeiana (Werner 1991, 1992b) Rana lessonae and Rana esculenta (Horat and Semlitsch 1994) Decreases in tadpole activity in response

PAGE 115

98 to predators appears to be a general, rather than a predator-specific response (Lawler 1989; Semlitsch and Gavasso 1992) Further, interspecific differences in response to predators appear to be strongly related to differences in breeding habitat association. Species that behaviorally respond to predators, such as those listed above, typically develop in sites that contain large predator populations (i.e., permanent sites or sites with long hydroperiods) In contrast, species using the most ephemeral habitats, which often lack predators, generally do not possess the same behavioral mechanisms for avoiding predation (Babbitt, unpubl. data; Woodward 1983). The costs and benefits of behavioral activity patterns in tadpoles, and therefore the e::pression of such patterns, appear to be strongly related to differences in the major source of mortality at permanent (i.e., predation) versus ephemeral sites (i.e., habitat desiccation) (Skelly 1995). In most situations, behavioral adjustments that are flexible and context-specific should be strongly selected for because individuals possessing such flexibility should be better able to balance costs and benefits associated with specific activities (Sih 1987) Thus, wlLli tlie exception ol sites that are too short in duration for larvae to complete development, species with tadpoles that reduce mortality risk through reductions in activity could potentially breed in a wide variety of sites containing different suites of

PAGE 116

99 predators. R. utricular ia provides a good example, because it breeds in a wide range of habitats, including permanent sites containing fish predators and temporary sites where macroinvertebrates are the main predators (pers. obs) R. utricularia tadpoles also altered spatial distribution in response to predators. In simple environments, such as the one used in this study, spatial shifts are limited to near versus away from the predator (e.g., Skelly and Werner 1990). Other studies have demonstrated shifts to refuge areas and benthic micohabitats (Holomuzki 1986b; Lawler 1989; Semlitsch and Gavasso 1992) The potential growth costs of shifts in spatial distribution are clear for species that occupy refuge areas that lack food, or for species that shift locations to areas with lower food availability. However, for individuals that are developing in areas that offer partial refuges from predators, such as areas with increased plant cover, movement away from predators may not always involve a tradeoff between mortality and growth (Diehl 1992) Increased cover may provide increased food resources, as well as increased protection from predators (Mclvor and Odum 1988) Overall, the largest costs associated with predator-induced habitat shifts are probably linked to differences in the spatial distribution of food resources, competitors, and predators. Thus, trade-offs between mortality risk and food acquisition related to habitat shifts may or may not be biologically significant.

PAGE 117

100 Growth and Development The presence of predators resulted in longer larval periods (i.e., delayed metamorphosis) on both food treatments. However, the effects on mass at metamorphosis were dependent on resource level. Tadpoles raised under the low food treatment metamorphosed at larger sizes when predators were present, whereas metamorphs on the high food treatments were larger when predators were absent. Although tadpoles had similar activity levels when predators were present, tadpoles on the high food treatment were more active when predators were absent compared to tadpoles on the low food treatment. This suggests that tadpoles subjected to predators on the high food treatment incurred a higher cost in terms of growth due to a greater decrease in activity. Thus, although high food availability resulted in larger size compared to low food availability regardless of predator treatment, the cost of reducing activity in response to predators was higher on high food. A possible e::planation for lack of a similar cost for tadpoles on the low food treatment is that because less food was available, even at lower activity levels Ladpoles were able Lo garner all resources in the container within each feeding period, whereas tadpoles on the high food predator present treatment were not. In a similar study, Skelly and Werner (1990) found no interaction of predator and food treatment

PAGE 118

101 for either age or mass at metamorphosis. In their study, decreased activity associated with predator presence did not affect larval period, but resulted in reduced size at metamorphosis. In field enclosures, Skelly (1992) found that Hyla versicolor tadpoles responding to predators had both decreased growth and developmental rates. Thus, although costs of predator-induced behavioral adjustments appear to be common, how these costs are reflected in growth versus development appear to vary among taxa or under varying e::perimental conditions (Werner and Anholt 1993) Several studies have found that increases in food level are associated with decreased activity (reviewed in Werner and Anholt 1993) Such patterns have been found in Rana sylvatica (Werner 1992a), Rana catesbeiana (Anholt and Werner 1995), and Pseudacris triseriata (Skelly 1995). In contrast, Bufo terrestris (Skelly and Werner 1990) and Pseudacris crucifer (Skelly 1995) showed little responses to food levels. Kohler and McPeek (1989) found higher activity levels at lower food levels in a Baetis mayfly and a Glossosma caddisfly. In this study, R. utricularia had hiylici ucLivLty IcvcIl; aL highui 1 uud Icvt^Uj. 'I'liuuc icl;uILl; (jonLradict Lliosc of Werner (1992a) for Rana pLpiens As is true for activity responses to predators, differences among anuran species in their responses to resource level may be related to breeding habitat associations, and the relative costs of reducing activity (Skelly 1995) Thus, I would

PAGE 119

102 predict that species breeding in permanent ponds would be more likely to decease activity at high resource levels because the costs of decreased growth are comparatively small. In contrast, for species developing in more ephemeral sites, reduced foraging activity could greatly increase the possibility of death due to habitat drying. For these species, we would predict either small adjustments or increased foraging activity at higher resource levels (Skelly 1995) Ecological Consequences of Responses The duration of the larval period has important consequences for species that use temporary habitats as breeding sites. Because the threat of mortality due to habitat drying is significant, factors that prolong the larval period can have negative consequences (Skelly 1995) In this study, tadpoles responded to predators by prolonging the larval period. R. utricularia breeds in a variety of habitats, ranging from temporary wetlands to permanent ponds. Depending on specific breeding habitat, the prolonged larval period may or may not increase risk of mortality due to habitat desiccation. Reductions in activity can have positive effects by decreasing risk of mortality regardless of habitat association because for most of the broad suite of predators that R. utricularia may face in different breeding sites, whether vertebrate or

PAGE 120

103 invertebrate, prey movement is an important component of predator detection (Lima and Dill 1990) Therefore, the costs of reduced activity in terms of prolonged larval period should be weighed against the benefit of decreased predation risk. Because R. utricularia breeds in temporary sites, it would be instructive to test the response of tadpoles to the presence of predators under varying drying regimes to determine whether R. utricularia can balance the costs and benefits of foraging versus predator detection with the costs of habitat drying. Reductions in size at metamorphosis can have potential costs regardless of breeding habitat association. Size at metamorphosis can have an important influence on adult survival and reproductive fitness (Berven 1990; Goater 1994; Semlitsch et al. 1988; Smith 1987) In this study, tadpoles on the higher plane of resources that responded to predators incurred relatively higher cost in terms of reduced size at metamorphosis, although they had absolutely larger masses compared to tadpoles on the lower food treatment. Food treatment had a greater effect on growth than did predator treatment. Increased size at metamorphosis as a response to increased ioud hay been docuiuun Lud immeiuus Liiiiuy (luviuwed in Wilbur 1980) That food availability had such a large effect on growth is not surprising; however, the differential costs of responding to predators on different food treatments is. These results demonstrate the need for

PAGE 121

104 assessing responses to both resource level and threat of predation in order to fully understand the consequences of cost-benefit trade-offs mediated by changes in behavior. Interspecific differences in behavioral responses to predators, and the consequences these differences have on growth and survival, are important considerations for eiiplaining assemblage structure at different breeding sites or habitat patches (Werner and Anholt 1996) Across breeding sites, interspecific differences in behavioral responses to predators largely may represent evolutionary responses to different costs associated with breeding habitats. For example, the relative mortality costs due to predation versus habitat drying, may largely explain interspecific differences in behavioral responses to predators (Skelly 1995; Woodward 1983) Within breeding sites, the effects of interspecific differences in behavioral responses on growth, coupled with patchy distribution patterns of predators and tadpoles, could influence local differences (i.e., among patches) in assemblage structure. Results of empirical studies (Anholt and Werner 1995; Werner and Anholt 1996) as well as theoretical models (MacNamara and Houston 1987) suggest that population regulation must be viewed as a multi-factor process. Behavioral responses made by individuals to changes in I lio onvironmoiil ,i f ToiM rink oT |uo(i,il ion, re^r.own-nr. acquisition, and therefore population-level responses.

PAGE 122

105 Whether behavioral adjustments have direct or indirect effects within food web dynamics, our understanding of these processes will be enhanced by considering behavior as an important factor in future experiments.

PAGE 123

106 Table 5-1. Summary of ANOVA for growth and larval period responses of Rana utricular ia tadpoles to food and predator treatments. CD is the coefficient of determination, which is the percentage of variation in the response variable attributable to the treatment effect. Mass (mg) Source of variation df ss F P Block 5 0, 0309 2 .10 0. 1221 Food 1 0, ,0904 30 .74 0. ,0001 Predator 1 0. ,0012 0 .42 0. 5248 Predator : : Food 1 0. 0180 6 12 0, 0259 Error 15 0. ,0441 CD 1 67 48.97 0. 65 9.75 Larval period (days) Source of variation df Block 5 Food 1 Predator 1 Predator :i Food 1 Error 15 SS F P CD 0.0008 1.01 0.4064 3.24 0.0192 130.28 <0.0001 77.73 0.0021 13.97 0.0020 8.50 0.0004 2.50 0.1344 1.62 0.0022

PAGE 124

107 Table 5-2. Summary of ANOVA for activity and distribution responses of Rana utricularia tadpoles to food and predator treatments. CD is the coefficient of determination, which is the percentage of variation in the response variable attributable to the treatment effect. Location (% away from predator) Source of variation df SS F P Block 5 0. 0356 0. 7 6 0. 5927 Food 1 0. 0084 0. 90 0. 3576 Predator 1 0 1597 17 05 0. 0009 Predator Food 1 0 0089 0. 95 0. 3453 Error 15 0. 1405 CD 10.08 2.38 45.23 2.52 Activity (b active) Source of variation df SS F P Block 5 0. 1439 o .99 0. .0453 Food 1 0. 0404 4 .20 0. .0582 Predator 1 0, ,2982 31 .01 0, .0001 Predator : : Food 1 0, .0433 4 .50 0, .0510 Error 15 0, 1442 CD 21.48 6.03 44 .51 6.46

PAGE 125

108 65 n 60 55 Low High Food Treatment Figure 5-1. Mean (+ 1 SD) responses of Rana utricular ia tadpoles to food level and predator treatments: age at metamorphosis (top) and wet mass at metamorphosis (bottom) Circles indicate predator absent and squares indicate predator present. (n=6 for all treatments)

PAGE 126

109 80 75 70 65 H I 60 CO 8 55 50 45 40 Low High 50 n 40 30 5 20 < 10 0 Low High Food Treatment Figure 5-2. Mean (+ 1 SD) responses of Rana utriculari a tadpoles to food level and predator treatments: percentage of tadpole on the side of the container opposite the predator (top) and percentage of tadpoles that were active (bottom) Circles indicate predator absent and squares indicate predator present. (n=6 for all treatments)

PAGE 127

CHAPTER 6 EFFECTS OF COVER AND PREDATOR IDENTITY ON PREDATION OF HYLA SQUIRELLA TADPOLES Introduction Habitat structural comple::ity can strongly affect biotic interactions. In particular, the intensity of predator-prey interactions can be mediated by habitat complexity because the effectiveness of predators often decreases in structurally comple:: habitats (Babbitt and Jordan 1996; Diehl 1988; Werner et al 1983). Typically, increased habitat compler.ity reduces predation rates by providing refuges for prey or by decreasing predator efficiency (Crowder and Cooper 1982; Savino and Stein 1982; Werner et al 1983) The role of habitat structure in mediating predatortadpole interactions is potentially important because predation is a significant source of mortality of anuran tadpoles (Calef 1973; Heyer et al. 1975; Smith 1983). Species breeding in permanent ponds display a wide array of anti-predator defenses for avoiding predation by fish (Formanowicz and Brodie 1982; Kats et al. 1988; Lawler 1989) In general, such defenses are lacking in species that breed in ephemeral ponds (Kats et al. 1988). In addition, activity rates of tadpoles in temporary ponds often are 110

PAGE 128

Ill higher compared to species in permanent ponds (Woodward 1983) Because many predators of tadpoles are gape-limited, rapid growth to a size-refuge is an important mechanism for escaping predation (Crump 1984; Richards and Bull 1990; Travis et al. 1985a; but see Crump and Vaira 1991). For species breeding in ephemeral environments, rapid growth is also important for decreasing risk of desiccation (Newman 1989; Wilbur 1987) However, trade-offs e::ist between rapid growth and predator avoidance because actively foraging tadpoles are more likely to be detected by predators (Skelly 1994; Werner and Anholt 1993; Woodward 1983) Although ephemeral ponds lack fish predators, these sites often contain aquatic insect predators that also are significant mortality agents (Smith 1983; Wilbur and Fauth 1990) Therefore, any mechanism that can reduce predator efficiency without decreasing foraging activity could enhance tadpole survival, particularly in ephemeral environments. Because structurally comple;: habitats can reduce predator foraging efficiency, compler. environments may provide tadpoles with partial protection from predators, even when tadpoles are actively foraging. ill this study 1 e;;amine how habitat structure aiiects predation rates on squirrel treefrog ( Hyla squirella ) tadpoles. H. squirella is a coimmon anuran in Florida, and is often numerically dominant in the ephemeral sites in which it breeds (Babbitt, unpubl. data). H. squirella tadpoles are

PAGE 129

112 active foragers even in the presence of insect predators (Babbitt, unpubl. data), and therefore they are a good model to test whether increased habitat structure provides protection for actively foraging tadpoles. Methods I conducted the experiment at the MacArthur AgroEcology Research Center in Lake Placid, Highlands County, Florida. I established a rectangular array of 16, 1.14 m diameter plastic wading pools with tight-fitting screen tops. On 3 September 1994 I filled pools with 95 1 of well water (depth = 12 cm) To allow normal feeding behavior, I added 2000 ml of pond water from a wading pool used for culturing algae. Pools received either 1000 g (high cover) or 500 g (low cover) of the wetland plant Hydrochloa carolinensis This plant is common in many wetlands where H. squirella breeds, and provides comple:: structure throughout the water column. I collected one gravid female and several males and placed them together in an aquarium with aged well water and sparse vegetation. The female layed a clutch of eggs on 1 Septeinber 1994, and on 15 September 1994 i [laphazardly assigned 30 Gosner (1960) stage 35-37 tadpoles (mean + 1 SD; SVL 12.2+0.47 mm, weight 0.301+0.04 g; n=25) to each pool. Prior to the experiment, I kept tadpoles together in a plastic wading pool and provided rabbit chow ad libitum. I

PAGE 130

113 used late-stage tadpoles because the predators I used are extremely effective predators of early stage tadpoles (Babbitt, unpubl. data), and therefore using early stage tadpoles would not provide a good test of habitat effects. I used larval common green darners ( Ana:: Junius Odonata: Aeshnidae) (total length 29.3+0.14 mm; n=8) or giant water bugs ( Lethocerus americanus Hemiptera: Belostoraatidae) (total length 17.5+0.09 ram; n=8) as predators. Both species are common inhabitants of the ephemeral wetlands in which H. squirella breed. I held predators individually without food for 48 hrs to standardize hunger within each species. I began the e;:periment at 1900 hrs on 16 September 1994 when I added a single predator to each pool. I terminated the e;:periment at 1900 hrs on 22 September and censused remaining tadpoles to determine survival. Diurnal high temperature was 31"C during the e::periment. The e::periment was a 2X2 factorial design arranged in a randomized complete block and replicated four times. Two species of predator (A. junuis or L. americanus ) were crossed with two plant densities (1000 g or 500 g) In addition to predator treatments, i also ran two control pools at each plant density without predators to determine whether plant treatment alone affected survival. I used fi;:ed effect analysis of variance (ANOVA) to determine whether predator identity, plant density, or their

PAGE 131

1 1 -1 interaction affected survival of H. squirella tadpoles. The number of tadpoles surviving was log transformed to reduce skewness (Snedecor and Cochran 1980) Controls were not included in the above analysis because all tadpoles survived in the control pools, and all mortality in treatment pools could be attributed to predation. Resul ts Tadpoles had higher survival in the high cover treatment compared to low cover (F5_,,=26. 15; p=0.003). Predator identity did not significantly affect survival (F, = 63; p=0.26), and the interaction between predator species and plant density was not significant (F,_i^=0.64; p=0.44). The number of tadpoles surviving under the higher density treatment was nearly double that of the low density treatment under both predator treatments (Figure 6-1) Discussion Compared to research on many other aquatic organisms, only limited empirical research has focused on effects of habitat comple;:ity in mediating relations between predators and anuran larvae. Sredl and Collins (1992) found that differences in habitat structure did not affect survival of mountain treefrog ( Hyla e:>imia ) tadpoles subject to predation by tiger salamanders ( Ambystoma tigrinum ) Similarly, Figiel and Semlitsch (1991) examined predation by

PAGE 132

115 crayfish (P. acutus ) on injured gray treefrog (H. chrysocelis ) tadpoles and found no effects of habitat comple;:ity. In contrast, increased plant density resulted in decreased predation by aquatic insects on southern toad ( Bufo terrestris ) tadpoles (Babbitt and Jordan 1996) Differences among these studies could be related to the relative degree of cover, or to differences in predator or prey behavior. Reduction in foraging success for the aquatic insects used in this study probably resulted from both reduced visual range and interference in stalking behavior. For er.ample, visual detection is important for prey capture by odonate naiads (Pritchard 1965), and increased vegetation cover can interfere with the predatory efficiency of these active foragers (Diehl 1992; Folsom and Collins 1984) However, foraging efficiency of L. americanus a sit and wait predator, also was reduced suggesting that increased habitat complexity can mediate predator-prey interactions between tadpoles and a broad suite of aquatic insect predators. For example, in a similar study, predation on B. terrestris tadpoles was lower at higher cover levels for both odonate naiads and hemipterans ( Belostoma f luminea ) (Babbitt and Jordan 1996) H. squirella tadpoles develop rapidly and are active foragers, even in the presence of aquatic insect predators (Babbitt, unpubl. data). Rapid growth and development.

PAGE 133

116 necessary for attaining metamorphic size prior to pond drying, may e::plain why many species developing in ephemeral sites do not respond to predators. Although active tadpoles are generally more susceptible to predators (Skelly 1994; Werner and Anholt 1993; Woodward 1983), species developing in ephemeral ponds presumably would be more restricted in reduced-activity responses if the threat of habitat drying poses a more signi f leant mortality threat. Rather, foraging in structurally complex environments may provide increased protection from predators, even for species that do not facultatively decrease activity in response to predators. Research that examines the relative role of habitat structure in mediating predation by different predator groups (e.g., fish versus aquatic insects) on species with different activity rates should provide valuable insight into the relative costs and benefits of different behavioral strategies and the degree to which habitat features can influence their effectiveness.

PAGE 134

117 Cover Level Figure 6-1. Number of Hyla squirella tadpoles that survived predation under different cover levels. Bars are means + 1 SE (n=4 for each bar)

PAGE 135

CHAPTER 7 SUMMARY AND CONCLUSIONS The introduction of water from fish-containing ditches into otherwise isolated temporary wetlands produced large changes in species composition of anurans Wetlands that e;:perienced the hydrologic disturbance were avoided by species that breed in fish-free sites; however, use by species that do breed in sites with fish increased. Species that avoided impacted sites bred in adjacent sites that were unaffected by water from ditches. These results suggest that temporary wetlands on this site provide dynamic habitats that offer varying breeding opportunities and conditions for larval developmental that are highly dependent on meteorological conditions. The effect of the hydrologic disturbance is a system that is both spatially and temporally dynamic resulting in variable assemblage structure and composition at affected sites. Because MAERC contains numerous wetlands and ditches that vary in hydrology from ephemeral to permanent, most species can find suitable breeding habitat except under e;:treme drought conditions. Ditching exacerbates the drought by increasing drainage, lowering the water table, and prolonging dry conditions 118

PAGE 136

119 Historical changes to the landscape have probably increased local anuran species richness and altered tadpole assemblage structure at the wetlands at MAERC. The surrounding upland matrix can have a large influence on wetland use as breeding sites. At least one species, Hyla femoralis appeared to be locally abundant but limited in distribution to forested hammocks. Research on the relationship between upland habitat patch characteristics and wetland use would increase our understanding of breeding use patterns, as well as our ability to provide guidance for managing anuran populations in altered systems. By examining spatial and temporal patterns of tadpoles and aquatic insect predators, I was able to determine that microhabitat use within wetlands does not appear to be an important mechanism for avoiding predation within the temporary wetlands that I studied. The invertebrate predators that are the dominant predators at these sites were relatively evenly distributed throughout these wetlands; therefore, differential use of wetland areas by tadpoles presumably would not have provided increased protection. Examinations of microhabitat use by Hyla squirella a species tliat breeds shortly alLer wetland filling, and Rana utricularia which is more plastic in breeding site use, revealed only weak microhabitat selection. Early breeding, rapid development, and avoidance of sites containing fish predators appear to be the major

PAGE 137

120 anti-predator mechanisms in H. squirella In contrast, R. utricularia may rely more heavily on early growth (as opposed to rapid development) to attain a size that is beyond the gape limits of predators, and behavioral adjustments in activity to avoid detection by predators. Such differences in life history strategies influence the range of breeding sites that each species can use successfully. Some studies have demonstrated spatial differences in tadpole distribution within wetlands. Most of Lhese studies have been conducted in wetlands with zonal differences that vary greatly in cover, such as wetlands with open-water centers. The differences in cover among zones in the wetlands I studied were relatively small, and may not have provided enough difference in protection or food (or other factors) to generate differences in microhabitat use. Tadpoles may exhibit plasticity in microhabitat selection relative to the degree of difference in microhabitat features such as cover. Interspecific differences in tadpole plasticity relative to cover could lead to differences in assemblage structure under varying habitat conditions (i.e., different levels of cover). Multi-factor experiments that examine microhabitat effects are needed to determine the relative importance of factors, such as cover, in structuring tadpole assemblage.

PAGE 138

121 Although tadpoles did not differentially occupy different areas of wetlands, experimental research identified the importance of increased cover in mediating predator-prey interactions. This was true for tadpoles that behaviorally adjust activity to decrease detection by predators (R. utricularia ) and those that do not (H. squirella ) Further, cover decreased foraging rates of aquatic insects that are sit-and-wait predators that cling to vegetation ( Lethocerus americanus ) those that cling to vegetation but are more active ( Ana:: junius ) and those that are benthic ( Tramea Carolina ) Thus, although tadpole microhabitat selection relative to cover did not appear to be important, occupation of areas with high cover may decrease predation rates. Higher densities of aquatic insects in areas with high cover could negate the benefits of cover; however, in this study predators were distributed evenly. One aspect not e::amined in this study was microhabitat use relative to the water column. Differences in positional use within the water column (i.e., mid-water versus benthic) could be an important anti-predator mechanism for some species. For e::ample, a large number of benthic predators may pose little threat to tadpoles foraging in the water column. Finally, tadpoles of some species, including R. utricularia exhibit shifts in spatial location and activity levels in response to the presence of predators. Reduction

PAGE 139

122 in activity, however, involves a trade-off between decreasing risk of predation and obtaining resources because both risk (predation) and reward (resources) are positively related to activity. However, the relative costs of decreasing activity in response to predators are dependent on background resource level. In R. utricularia the presence of predators resulted in longer larval periods (i.e., delayed metamorphosis) regardless of food treatment. However, the effects on mass at metamorphosis were dependent on resource level. Tadpoles raised under the low food treatment metamorphosed at larger sizes when predators were present, whereas metamorphs on the high food treatments were larger when predators were absent. These results support the idea that population regulation is a multi-factor process. Behavioral responses made by individuals to changes in the environment affect risk of predation, resources acquisition, and therefore population level responses. Because the effects of predation can be direct or indirect, our understanding of these processes will be enhanced by considering behavior as an important factor in future experiments. Because long liydroperiod temporary wetlands oiler suitable breediuy habitat to a wide variety of anurans, the potential for behavioral differences among the species developing within these sites to affect assemblage dynamics may be great. There is much more to study.

PAGE 140

LITERATURE CITED Abrams, P. A. 1993. Why predation rates should not be proportional to predator density. Ecology. 94:726-733. Alford, R.A. 1986. Habitat use and positional behavior of anuran larvae in a northern Florida temporary pond. Copeia. 1986:408-423. Alford, R.A. and M.L, Crump. 1982. Habitat partitioning among size classes of larval southern leopard frogs, Rana utricularia Copeia. 1982:367-373. Alford, R.A. and R.N. Harris. 1988. Effects of larval growth history on anuran metamorphosis. Am. Nat. 131:91-106. Anholt. B.R. and E.E. Werner. 1995. Interaction between food availability and predation mortality mediated by adaptive behavior. Ecology. 7 6:2230-2234. Azevedo-Ramos, C, M. Van Sluys, J.-M. Hero, and W.E. Magnusson. 1992. Influence of tadpole movement on predation by odonate naiads. J. Herp. 26:335-338. Babbitt, K.J. and F. Jordan. 1996. Predation on Bufo terrestris tadpoles: effects of cover and predator identity. Copeia. 1996:485-488. Banks, B. and T.C. Beebee. 1988. Reproductive success of natterjack toads Bufo calamita in two contrasting habitats. J. Anim. Ecol. 57:475-492. Berven, K.A. 1990. Factors affecting population fluctuations in larval and adult stages of the wood frog ( Rana sylvatica ) Ecology. 71:1599-1608. Bowker, R. G. and M.H. Bowker. 1979. Abundance and distribution of anurans in a Kenyan pond. Copeia. 1979:278-285. Brodie, E.D. and D.R. Formanowicz. 1983. Prey size preferences of predators: differential vulnerability of larval anurans. Herpetologica 39:67-75. 123

PAGE 141

124 Caldwell, J. P., J.H. Thorpe, and T.O. Jervey. 1980. Predator-prey relations among larval dragonflies, salamander, and frogs. Oecologia. 46:285-289. Calef, G.W. 1973. Natural mortality of tadpoles in a population of Rana aurora Ecology. 54:741-758. Carpenter, S.R., J.F. Kitchell, J.R. Hodgson, P. A. Cochran, J.J. Elser, M.M. Elser, D.M. Lodge, D. Dretchmer, X. He, and C.N. von Ende 1987. Regulation of lake primary productivity by food web structure. Ecology. 68:18631876. Carr, A.F., Jr. 1940. A contribution to the herpetology of Florida. Uni Fla. Publ Biol. Sci. Ser. 3:1-118. Chick, J.H., F. Jordan, J. P. Smith, and C.C. Mclvor. 1992. A comparison of four enclosure traps and methods used to sample fishes in aquatic macrophytes. J. Freshwater Ecol. 7:353-361. Connell, J.H. 1978. Diversity in tropical rain forests and coral reefs. Science. 199:1302-1310. Corbet, P.S. 1980. Biology of Odonata. Ann. Rev. Entomol 25:189-217 Cronin, J.T. and J. Travis. 1986. Size-limited predation on larval Rana areolata (Anura : Ranidae) by two species of backswimmer (Insecta: Hemiptera: Notonectidae) Herpetologica. 42:171-174. Crowder, L.B. and W.E. Cooper. 1982. Habitat structural compler.ity and the interaction between bluegills and their prey. Ecology. 63:1802-1813. Crump, M.L. 1971. Quantitative analysis of the ecological distribution of a tropical herpetof auna Mus. Nat. Hist. Misc. Publ. Univ. Kans 61:1-68. Crump, M.L. 1982. Amphibian reproductive ecology on the community level, pp. 21-36. Tn: N.J. Scott (ed.). Herpetological Conmiunities U.S. Dept. interior, Fish and Wildlife Service, Wildlife Res. Rep. 13. 239pp. Crump, M.L. 1984. Ontogenetic changes in vulnerability to predation in tadpoles of Hyla pseudopuma Herpetologica. 40:265-271. Crump, M.L., and M. Vaira. 1991. Vulnerability of Pleuroderma borelli tadpoles to an avian predator: effect of body size and density. Herpetologica 47:316321

PAGE 142

125 Dalyrumple, G.H. 1988. The herpetofauna of Long Pine Key, Everglades National Park, in relation to vegetation and hydrology, pp. 72-86. In: R.C. Szaro, K.E. Severson, and D.R. Patton (eds.). Management of amphibians, reptiles, and small mammals in North America. U.S.D.A. For. Ser. Gen. Tech Rep. RM-166., Fort Collins, Colorado. 458pp. Day, R.W. and G.P Quinn. 1989. Comparisons of treatments after an analysis of variance in ecology. Ecol. Monog. 59:433-463. Diaz-Paniagua, C. 1987. Tadpole distribution in relation to vegetal heterogeneity in temporary ponds. Herpetological Journal. 1:167-169. Diaz-Paniagua, C. 1990. Temporary ponds as breeding sites of amphibians at a locality in southwestern Spain, Herpetological Journal. 1:447-453. Diehl, S. 1988. Foraging efficiency of three freshwater fishes: effects of structural compler.ity and light. Oikos. 53:207-214. Diehl, S. 1992. Fish predation and benthic comnmnity structure: the role of omnivory and habitat compler.ity. Ecology. 73:1646-1661. Diiion, J.R. and W.R. Heyer. 1968. Anuran succession in a temporary pond in Colima, Me;:ico. Bull. So. Cal. Acad. Sci. 67:129-137. Dodd, K.C., Jr. 1992. Biological diversity of a temporary pond herpetofauna in north Florida sandhills. Biodiversity and Conservation. 1:125-142. Dodd, K.C., Jr. 1995. The ecology of a sandhills population of the eastern narrow-mouth toad, Gastrophryne carolinensis during a drought. Bull. Fla. Mus Nat. Hist., Biol. Sci. 38:11-41. Duellman, W.E. and A. Schwartz. 1958. Amphibians and reptiles of southern Florida. Bull. Fl State Mus. 3: 181-323. Duellman, W.E. and L. Trueb. 1986. Biology of Amphibians. McGraw-Hill Book Co. New York. 670pp. Dupre, R.K. and J.W. Petranka. 1985. Ontogeny of temperature selection in larval amphibians. Copeia. 1985:462-467.

PAGE 143

126 Ebenman, B. and L. Persson (eds.). 1988. Size-structured populations. Springer-Verlag Berlin. 284pp. Enge, K.E. and W.R. Marion. 1986. Effects of clearcutting and site preparation on herpetofauna of a north Florida flatwoods. For. Ecol. and Manage. 14:177-192. Figiel, C.R., Jr. and R.D. Semlitsch. 1991. Effects of nonlethal injury and habitat complexity on predation in tadpole populations. Can. J. Zool. 69:830-834. Folsom, T.C. and N.C. Collins. 1984. The diet and foraging behavior of the larval dragonfly Ana;: Junius (Aeshnidae) with and assessment of the role of refuges and prey activity. Oikos. 42:105-113. Formanowicz, D.R., Jr. 1986. Anuran tadpole/aquatic insect predator-prey interactions: tadpole size and predator capture success. Herpetologica 42:367-373. Formanowicz, D.R., Jr. and E.D. Brodie. 1982. Relative palatabilities of members of a larval amphibian community. Copeia. 1982:91-97. Eraser, D.F. and J.F. Gilliam. 1987. Feeding under predation hazard: response of the guppy and Hart's rivulus from sites with contrasting predation hazard. Behav. Ecol. Sociobiol. 21:203-209. Freeman, B.J., H.S. Greening, and J.D. Oliver. 1984. Comparison of three methods for sampling fishes and macroinvertebrates in a vegetated freshwater wetland. J. Freshwater Ecol. 2:603-610. Gascon, C. 1991. Population and community-level analyses of species occurrences of central Amazonian rainforest tadpoles. Ecology. 72:1731-174 6. Gascon, C. 1992. Aquatic predators and tadpole prey in central Amazonia: field data and e;:perimental manipulations. Ecology. 73:971-980. Gilliam, J.F. 1982. Habitat use and competitive bottlenecks in size-structured fish populations. Dissertation. Michigan State University, Lansing, Michigan. Goater, CP. 1994. Growth and survival of postmetamorphic toads: interactions among larval history, density, and parasitism. Ecology. 75:2264-2274. Gosner, K.L. 1960. A simplified table for staging anuran embryos and larvae with notes on identification. Herpetologica. 16:183-190.

PAGE 144

127 Heck, K,L. and L.B. Crowder 1991. Habitat structure and predator-prey interactions in vegetated aquatic systems, pp. 281-299. In: S.S. Bell, E.D. McCoy, and H.R. Mushinsky (eds.). Habitat structure: the physical arrangement of objects in space. Chapman Hall. New York, NY. Hews, D.K. 1995. Overall predator feeding rates and relative susceptibility of large and small tadpoles to fish predation depend on microhabitat : a laboratory study. J. Herp. 29:142-145. Heyer, W.R. 1973. Ecological interactions of frog larvae at a seasonal tropical location in Thailand. J. Herp. 7: 337-361 Heyer, W.R. 1974. Niche measurements of frog larvae from a seasonal tropical location in Thailand. Ecology. 55: 651-656. Heyer, W.R. 1976. Studies of larval habitat partitioning. Smithsonian Contributions to Zoology. 242:1-27. Heyer, W.R., R.W. McDiarmid, and D.L. Weigmann. 1975. Tadpoles, predation, and pond habitats in the tropics. Biotropica. 7:100-111. Holomuzki, J.R. 1986a. Effects of microhabitat on fitness components of larval tiger salamanders, Ambystoma tigrinum nebulosum Oecologia. 71:142-148. Holomuzki, J.R. 1986b. Predator avoidance and diel patterns of microhabitat use by larval tiger salamanders. Ecology. 67:737-748. Holomuzki, J.R. 1989. Predation risk and macroalga use by the stream-dwelling salamander Ambystoma te::anum Copeia. 1989:22-28. Holt, R.D. 1987. Population dynamics and evolutionary processes: the manifold roles of habitat selection. Evol. Ecol. 1:331-347. Hopey, M.E. and J.W. Petranka 1994. Restriction of wood frogs to fish-free habitats: how important is adult choice? Copeia. 1994:1023-1025. Horat, P. and R.D. Semlitsch. 1994. Effects of predation risk and hunger on the behaviour of two species of tadpoles. Behav. Ecol. Sociobiol. 34:393-401.

PAGE 145

128 Kats, L.B., J.W. Petranka, and A. Sih. 1988. Antipredator defenses and the persistence of amphibian larva with fishes. Ecology. 69:1865-1870. Kohler, S.L. and M.A. McPeek. 1989. Predation risk and the foraging behavior of competing stream insects. Ecology. 70: 1811-1825. Kushlan, J. A. 1981. Sampling characteristics of enclosure traps. Trans. Amer. Fish. Soc. 110:557-562. Kutka, F.J. and M.D. Bachmann. 1990. Acid sensitivity and water chemistry correlates of amphibian breeding ponds in northern Wisconsin, USA. Hydrobiologia 208:153-160. LaClaire, L.V. and R. Franz. 1990. Importance of isolated wetlands in upland landscapes, pp. 9-15. In: M. Kelly (ed.) Florida Lake Management Society, Proc. 2nd Annual Meeting, Winter Haven, Florida. Lann, R. and B. Verboom. 1990. Effects of pool size and isolation on amphibian communities. Biol. Conserv. 54 :251-202. Lawler, S.P. 1989. Behavioural responses to predators and predation risk in four species of larval anurans. Anim. Behav. 38:1039-1047. Ligas, F.J. 1960. The Everglades bullfrog: life history and management. Fla. Game and Freshwater Fish Comm., Tallahassee. 129pp. Lima, S.L. and L.M. Dill. 1990. Behavioral decisions made under the risk of predation: a review and prospectus. Can. J. Zool. 68:619-640. Lizana, M., V. Perez-Mellado, and M.J. Cuidad. 1990. Analysis of the structure of an amphibian community in the central system of Spain. Herpetological Journal. 1 : 435-446. Loschenkohl, A. 1986. Niche partitioninq and competition in tadpoles, pp. 399-402. In: Z. Kocek (ed.). Studies in herpetology. Proc. 3rd. Annual Gen. Mtg. Societas European Herpetologica Charles Univ., Prague. Magnusson, W.E. and J.-M. Hero. 1991. Predation and the evolution of comple:: oviposition behaviour in Amazon rainforest frogs. Oecologia. 86:310-318. Mclvor, C.C. and W.E. Odum. 1988. Food, predation risk, and microhabitat selection in a marsh fish assemblage. Ecology. 69:1341-1351.

PAGE 146

129 McNamara, J.M. and A.I. Houston. 1987. Starvation and predation as factors limiting population size. Ecology. 68: 1515-1519. McNamara, J.M. and A.I. Houston. 1994. The effect of a change in foraging options on intake rate and predation rate. Am. Nat. 114:978-1000. McPeek, M.A. 1990. Determination of species composition in the Enallagma damselfly assemblages of permanent lakes. Ecology. 71:83-98. Moler, P.E. and R. Franz. 1987. Wildlife values of small isolated wetlands in the southeastern Coastal Plain, pp. 234-241. In: R.R. Odum, K.A. Riddleberger and J.C. Ozier (eds.). Proc. 3rd S.E. Nongame and Endangered Wildlife Symposium. Georgia Dept. of Nat. Res., Atlanta. 253pp. Morin, P.J. 1981. Predatory salamanders reverse the outcome of competition among three species of anuran tadpoles. Science. 212:1284-1286. Morin, P.J. 1983. Predation, competition, and the composition of larval anuran guilds. Ecol. Mono. 53:119-138. Morin, P.J. 1986. Interactions between intraspeci f ic competition and predation in an amphibian predator-prey system. Ecology. 67:713-720. Newman, R.A. 1987. Effects of density and predation on Scaphiopus couchii tadpoles in desert ponds. Oecologia. 71:301-307. Newman, R.A. 1989. Developmental plasticity of Scaphiopus couchii tadpoles in an unpredictable environment. Ecology. 70:1775-1787. O'Neill, L.D. 1995. Amphibian and reptile communities of temporary ponds in a managed pine flatwoods. M.S. Thesis, University of Florida, Gainesville. 106pp. Paine, R.T. 1969. The Pisaster Tegula interaction: prey patches, predator food preferences and intertidal community structure. Ecology. 50:950-961. Pechmann. J.H.K, D.E. Scott, J.W. Gibbons, and R.D. Semlitsch. 1989. Influence of wetland hydroperiod on diversity and abundance of metamorphosing juvenile amphibians. Wetlands Ecol. and Manage. 1:3-11.

PAGE 147

130 Peckarsky, B.L., C.A. Cowan, M.A. Penton, and C. Anderson, 1993. Sublethal consequences of stream-dwelling predatory stoneflies on mayfly growth and fecundity. Ecology. 74:1836-1846. Peterson, A.G., CM. Bull, and L.M. Wheeler. 1992. Habitat choice and predator avoidance in tadpoles. J. Herp. 26:142-146. Pfenning, D.W. 1990. The adaptive significance of an environmentally-cued developmental switch in an anuran tadpole. Oecologia. 85:101-107. Power, M.E. 1990. Effects of fish in river f oodwebs Science. 250:811-814. Pritchard, G. 1965. Prey capture by dragonfly larvae (Odonata; Anisoptera) Can. J. Zool. 43:271-289. Resetarits, W.J. and H.M. Wilbur. 1991. Calling site choice by Hyla chrysoscelis : effects of predators, competitors, an oviposition sites. Ecology. 72:778-786. Richards, S.J. and CM. Bull. 1990. Size-limited predation on tadpoles of three Australian frogs. Copeia. 1990: 1041-1046. Richter, K.O. and A.L. Azous. 1995. Amphibian occurrence and wetland characteristics in the Puget Sound basin. Wetlands. 15:305-312. Rowe, CL. and W.A. Dunson. 1993. Relationships among abiotic parameters and breeding effort by three amphibians in temporary wetlands in central Pennsylvania. Wetlands. 13:237-246. Rowe, CL. and W.A. Dunson. 1995. Impacts of hydroperiod on growth and survival of larval amphibians in temporary ponds of central Pennsylvania, USA. Oecologia. 102:397403. Rozas, L.P., and W.E. Odum. 1988. Occupation of submerged aquatic vegetation by fishes: testing the roles of food and refuge. Oecologia 7:101-106. Savino, J.F., andR.A. Stein. 1902. Predator-prey interactions between largemouth bass and bluegills as influenced by simulated, submersed vegetation. Trans. Amer. Fish. Soc. 111:255-266.

PAGE 148

131 Seale, D.B. and N. Beckvar. 1980. The comparative ability of anuran larvae (genera: Hyla Bufo, and Rana) to ingest suspended blue-green algae. Copeia. 1980:495503. Semlitsch, R.D. 1987. Relationship of pond drying to the reproductive success of the salamander Ambystoma talpoideum Copeia. 1987:61-69. Semlitsch, R.D. 1990. Effects of body size, sibship, and tail injury on the susceptibility of tadpoles to dragonfly predation. Can. J. Zool. 68:1027-1030. Semlitsch, R.D. and J.W. Gibbons. 1988. Fish predation in size-structured populations of treefrog tadpoles. Oecologia. 75:321-326. Semlitsch, R.D., D.E. Scott, and J.H.K. Pechmann. 1988. Time and size at metamorphosis related to adult fitness in Ambystoma talpoideum Ecology. 69:184-192. Semlitsch, R.D. and S. Gavasso. 1992. Behavioural responses of Bufo bufo and Bufo calamita tadpoles to chemical cues of vertebrate and invertebrate predators. Ethol. Ecol. & Evol. 4:165-173. Semlitsch, R.D. and H.U. Reyer. 1992. Modification of antipredator behaviour in tadpoles by environmental conditioning. J. Anim. Ecol. 61:353-360. Sih, A. 1987. Predators and prey lifestyles: an evolutionary and ecological overview, pp. 203-24. In: W.C. Kerfoot and A. Sih (eds.). Predation: direct and indirect impacts on aquatic communities. University Presses of New England, Hanover. Sih, A, P. Crowley, M. McPeek, J. Petranka and K. Strohmeier. 1985. Predation, competition, and prey communities: a review of field e::periments Ann. Rev. Ecol. Syst. 16:269-311. Skelly, D.K. 1992. Field evidence for a cost of behavioral antipredaLor behavioral response in a larval ampliibian. Ecology. 73:704-708. Skelly, D.K. 1994. Activity level and the susceptibility of anuran larvae to predation. Anim. Behav. 47:465-468. Skelly, D.K. 1995. A behavioral trade-off and its consequences for the distribution of Pseudacris treefrog larvae. Ecology. 76:150-164.

PAGE 149

132 Skelly, D.K. and W.E. Werner. 1990. Behavioral and lifehistorical responses of larval American toads to an odonate predator. Ecology. 71:2313-2322. Smith, D.C. 1983. Factors controlling tadpole populations of the chorus frog ( Pseudacris triseriata ) on Isle Royale, Michigan. Ecology. 64:501-510. Smith, D.C. 1987. Adult recruitment in chorus frogs: effects of size and date at metamorphosis. Ecology. 68:344-350. Snedecor, G.W. and W.G. Cochran. 1980. Statistical methods, 7th edition. Iowa State University Press, Ames. Sokol, R.R. and F.J. Rolf. 1981. Biometry. Second Edition. W.H. Freeman and Co. New York. Sredl, M.J., and J. P. Collins. 1992. The interaction of predation, competition, and habitat complexity in structuring an amphibian community. Copeia 1992:607614. Stauffer, H.-P. and R.D. Semlitsch. 1993. Effects of visual, chemical and tactile cues of fish on the behavioral responses of tadpoles. Anim. Behav. 46:355-364. Stein, R.A. and J.J. Magnuson. 1976. Behavioral responses of crayfish to a fish predator. Ecology. 57:751-761. Strauss, S.Y. 1991. Indirect effects in community ecology: their definition, study and importance. Trends in Ecol. and Evol. 6:206-209. Tejado, M. 1993. Size-dependent vulnerability and behavioral responses of tadpoles of two anuran species to beetle larvae predators. Herpetologica 49:287-294. Toft, C.A. 1985. Resource partitioning in amphibians and reptiles. Copeia. 1985:1-21. Travis, J. 1983. Variation in growth and survival of Hyla gratio sa larvae in er.perimentati on enclosures. Copeia. 1983:232-237 Travis, J., W.H. Keen, and J. Juilianna. 1985a. The role of relative body size in a predator-prey relationship between dragonfly naiads and larval anurans Oikos. 45: 59-65. Travis, J., W.H. Keen, and J. Juilianna. 1985b. The effects of multiple factors on viability selection in Hyla gratiosa tadpoles. Evolution. 39:1087-1099.

PAGE 150

133 Travis, J. and J.C. Trexler. 1986. Interactions among factors affecting growth, development and survival in experimental populations of Buf o terrestris (Anura: Bufonidae) Oecologia. 69:110-116. Vickers, C.R., L.D. Harris, and B.F. Swindel. 1985. Changes in herpetofauna resulting from ditching of cypress ponds in Coastal Plains flatwoods. Forest Ecol. and Manage. 11:17-29. Ward, J.V. 1992. Aquatic insect ecology. 1. Biology and habitat. John Wiley & Sons, Inc. New York. 438pp. Weist, J. A. 1981. Anuran succession at temporary ponds in post oak savanna region of Texas, pp. 39-47. In: N.J. Scott (ed.). Herpetological Communities. U.S. Dept. Interior, Fish and Wildlife Service, Wildlife Res. Rep. 13. 239pp. Weins, J. A. 1989. Spatial scaling in ecology. Functional Ecology. 3:385-397. Weins, J. A. 1992. The ecology of bird communities. Vol. 2 Processes and variations. Cambridge University Press, Cambridge Werner, E.E. 1986. Amphibian metamorphosis: growth rate, predation risk, and the optimal size at transformation. Am. Nat. 128:319-341. Werner, E.E. 1991. Nonlethal effects of a predator on competitive interactions between two anuran larvae. Ecology. 72:1709-1720. Werner, E.E. 1992a. Competitive interactions between wood^ frog and northern leopard frog larvae: the role of size and activity. Copeia. 1992:26-35. Werner, E.E. 1992b. Individual behavior and higher-order species interactions. Am. Nat. 140:S5-S32. Werner, E.E., J.F. Gilliam, D.J. Hall and G.G. Mittlebach. 1983. An experimental test of the effects of predation risk on habitat use in fish. Ecology. 64:1540-1548. Werner. E.E. and B.R. Anholt. 1993. Ecological consequences of the trade-off between growth and mortality rates mediated by foraging activity. Am. Nat. 142:242-272. Werner. E.E. and B.R. Anholt. 1996. Predator-induced behavioral indirect effects: consequences to competitive interactions in anuran larvae. Ecology. 77 : 157-169.

PAGE 151

134 Werner, E.E. and M.A. McPeek. 1994. Direct and indirect effects of predators on two anuran species along an environmental gradient. Ecology. 75:1368-1382. Wilbur, H.M. 1980. Complex life cycles. Ann. Rev. Ecol. Syst. 11:67-93. Wilbur, H.M. 1984. Complex life cycles and community organization in amphibians, pp. 195-224. In: P.W. Price, C.N. Slobodchikof f and W.S. Gaed (eds.). A new ecology: novel approaches to interactive systems. John Wiley and Sons, New York. Wilbur, H.M. 1987. Regulation of structure in complex systems: experimental temporary pond communities. Ecology. 68:1437-1452. Wilbur, H.M. 1988. Interactions between growing predators and growing prey. pp. 157-172. In: B. Ebenman and L. Persson (eds.). Size-structured populations. SpringerVerlag. Berlin. 284pp. Wilbur, H.M. and J. P. Collins. 1973. Ecological aspects of amphibian metamorphosis. Science. 182:1305-1314. Wilbur, H.M. and J.E. Fauth. 1990. Ex.perimental aquatic food webs: interactions between two predators and two prey. Am. Nat. 135:176-204. Woodward, B.D. 1982. Tadpole competition in a desert anuran community. Oecologia. 54:96-100 Woodward, B.D. 1983. Predator-prey interactions and breeding-pond use of temporary-pond species in a desert anuran community. Ecology. 64:1549-1555.

PAGE 152

BIOGRAPHICAL SKETCH Kimberly J. Babbitt was born on 15 February 1962 in Putnum, Connecticut. She graduated from John Jay Senior High in 1980. Kimberly received a Bachelor of Science degree in wildlife management from the University of New Hampshire in 1984. After coordinating a statewide lake monitoring program in New Hampshire, she attended Texas A&M University to study parent-offspring relationships in collared peccaries. She graduated in 1988 with a Master of Science degree in wildlife science. After almost three years of working dual jobs for the Natural Resources Department in Sarasota County, Florida, as a biologist for the 33,000 acre T. Mabry Carlton Memorial Reserve and as a researcher/administrator of a comprehensive study of the Myakka River, Kimberly entered a doctoral program at the University of Florida. She studied the spatial and temporal dynamics of tadpoles and aquatic insect predators for which she received a doctorate in August 1996. Kimberly has accepted a position as assistant professor in wildlife ecology in the Natural Resources Department at the University of New Hampshire where she will continue working on amphibians. 135

PAGE 153

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. George W. Tanner, Chair Associate Professor of Wildlife Ecology and Conservation I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor o^ Philosophy. Lyn C. Branch Associate Professor of Wildlife Ecology and Conservation I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Associate in Wildlife Ecology and Conservation I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Carmine A. Lanciani Professor of Zoology

PAGE 154

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Carole C. Mclvor Assistant Professor of Wildlife Ecology and Conservation This dissertation was submitted to the Graduate Faculty of the College of Agriculture and to the Graduate School and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy. August, 1996 Dean, College of Agriculture Dean, Graduate School