Citation
Testing the Influence of Cuban Treefrogs (Osteopilus septentrionalis) on Native Treefrog Detection and Abundance

Material Information

Title:
Testing the Influence of Cuban Treefrogs (Osteopilus septentrionalis) on Native Treefrog Detection and Abundance
Creator:
Hoffmann, Kristine E
Place of Publication:
[Gainesville, Fla.]
Florida
Publisher:
University of Florida
Publication Date:
Language:
english
Physical Description:
1 online resource (70 p.)

Thesis/Dissertation Information

Degree:
Master's ( M.S.)
Degree Grantor:
University of Florida
Degree Disciplines:
Wildlife Ecology and Conservation
Committee Chair:
Johnson, Steve A.
Committee Members:
Moulton, Michael P.
Krysko, Kenneth L.
Dorazio, Robert
Graduation Date:
12/14/2007

Subjects

Subjects / Keywords:
Amphibians ( jstor )
Ecology ( jstor )
Frogs ( jstor )
Introduced species ( jstor )
Null hypothesis ( jstor )
Predators ( jstor )
Refuge habitats ( jstor )
Species ( jstor )
Tadpoles ( jstor )
Wetlands ( jstor )
Wildlife Ecology and Conservation -- Dissertations, Academic -- UF
cinerea, femoralis, hyla, nonindigenous, osteopilus, pvc, refugia, septentrionalis, squirella, treefrog
Wekiwa Springs State Park ( local )
Genre:
bibliography ( marcgt )
theses ( marcgt )
government publication (state, provincial, terriorial, dependent) ( marcgt )
born-digital ( sobekcm )
Electronic Thesis or Dissertation
Wildlife Ecology and Conservation thesis, M.S.

Notes

Abstract:
The Cuban treefrog (Osteopilus septentrionalis) is a nonindigenous species in Florida that anecdotally replaces native treefrog species where it becomes established. However, few empirical data are available to verify this claim; thus, I sought to determine the effects of O. septentrionalis on native treefrog populations. I removed O. septentrionalis from natural areas and monitored the response of native treefrog populations using polyvinyl chloride (PVC) pipe refugia. In conjunction with my field study, I conducted laboratory experiments to determine the frequency at which treefrogs use PVC pipe refugia and to test the null hypothesis that O. septentrionalis does not affect behaviorally or chemically exclude native treefrogs from refugia. In my first laboratory study, I compared the variation by species in frequency of refugia used when treefrogs were presented with the choice of either a PVC pipe refugia or a plant. I found that the native squirrel treefrog (Hyla squirella) rested in the pipe twice as often as on the plant, and Osteopilus septentrionalis used the pipes infrequently. The other two natives tested, green treefrog (H. cinerea) and pinewoods treefrog (H. femoralis), did not strongly favor either the PVC pipe refugia or the plant. In my other laboratory studies, I tested if Osteopilus septentrionalis interferes with the use of PVC pipe refugia by native treefrogs. First, I paired a native treefrog (Hyla sp.) with an O. septentrionalis in an aquarium with only one pipe, and then observed which frog(s) rested inside the PVC pipe refugium 14 hours later. I found that O. septentrionalis did not behaviorally exclude native treefrogs from refugia. In another study, I presented treefrogs with a choice of a new PVC pipe refugium and one that had been recently used by O. septentrionalis to determine if treefrogs avoided refugia recently used by this predator. I found that Hyla cinerea and H. femoralis did not avoid refugia that were recently used by O. septentrionalis. These laboratory studies suggest that native treefrog detection in PVC pipe refugia is not altered by the presence of similar sized O. septentrionalis. In my field experiment, I established paired grids of PVC pipe refugia to attract treefrogs at three study sites at Wekiwa Springs State Park, Apopka, Florida, and at three study sites at Flatwoods Park, Hillsborough Wilderness Park, Tampa, Florida. One grid of each pair was used as a treatment grid where I removed Osteopilus septentrionalis, and the other served as a control. I monitored the community of native treefrogs within each grid, and hypothesized that if O. septentrionalis reduces the abundance of native treefrogs, then at sites where this species was removed native treefrogs should become more abundant. However, capture rates were low and few O. septentrionalis were removed to potentially decrease its abundance. Therefore, I was not able to determine if its removal positively impacts native treefrog abundance. However, my data did indicate a negative correlation between the abundance of O. septentrionalis and Hyla femoralis. More research is needed to determine the cause of this relationship. ( en )
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Bibliography:
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Thesis:
Thesis (M.S.)--University of Florida, 2007.
Local:
Adviser: Johnson, Steve A.
Statement of Responsibility:
by Kristine E Hoffmann.

Record Information

Source Institution:
UFRGP
Rights Management:
Copyright Hoffmann, Kristine E. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Classification:
LD1780 2007 ( lcc )

Downloads

This item has the following downloads:


Full Text





refugia should be aware of this sampling bias when interpreting their results, especially when

individuals are not marked.

The hypothesis that Osteopilus septentrionalis does not interfere with the detection of

native treefrogs using refugia was supported by results of my laboratory experiments (Chapter

3). I did not find evidence of behavioral or chemical exclusion of native treefrogs by O.

septentrionalis. This implies that native treefrogs sympatric with 0. septentrionalis can be

compared to those where 0. septentrionalis is absent; populations before and after establishment

of this nonindigenous species can be compared, as can populations before and after their

removal. However, although I determined that 0. septentrionalis does not behaviorally exclude

native treefrogs when they reach the PVC pipe refugia simultaneously, I did not test to see if they

displayed territorial behavior as residents when an invader frog attempts to use the refugia. Also,

I did not determine if native frogs that use the PVC pipe refugia in the field are more likely to be

preyed upon than those occurring away from artificial refugia. It is possible that the population

may be underestimated if frogs in PVC pipe refugia are consumed before they can be counted.

While native treefrogs do not avoid PVC pipe refugia that had recently been used by O.

septentrionalis, frogs of this nonindigenous species did avoid refugia that had been used by their

conspecifics. These frogs may be avoiding their batracophagic conspecifics, whereas native

treefrogs may be unaware of this potential predator, and may be more susceptible to predation.

These are two more questions that need to be addressed to be sure that 0. septentrionalis does

not interfere with the detection of native frogs.

Capture rates were low during the field study and an insufficient number of Osteopilus

septentrionalis was caught to examine the effects of its removal on native treefrogs. However,

the results of this study did reveal a negative correlation between the frequency of observations









septentrionalis in an aquarium with only one pipe, and then observed which frog(s) rested inside

the PVC pipe refugium 14 hours later. I found that 0. septentrionalis did not behaviorally

exclude native treefrogs from refugia. In another study, I presented treefrogs with a choice of a

new PVC pipe refugium and one that had been recently used by 0. septentrionalis to determine

if treefrogs avoided refugia recently used by this predator. I found that Hyla cinerea and H.

femoralis did not avoid refugia that were recently used by 0. septentrionalis. These laboratory

studies suggest that native treefrog detection in PVC pipe refugia is not altered by the presence

of similar sized 0. septentrionalis.

In my field experiment, I established paired grids of PVC pipe refugia to attract treefrogs at

three study sites at Wekiwa Springs State Park, Apopka, Florida, and at three study sites at

Flatwoods Park, Hillsborough Wilderness Park, Tampa, Florida. One grid of each pair was used

as a treatment grid where I removed Osteopilus septentrionalis, and the other served as a control.

I monitored the community of native treefrogs within each grid, and hypothesized that if O.

septentrionalis reduces the abundance of native treefrogs, then at sites where this species was

removed native treefrogs should become more abundant. However, capture rates were low and

few 0. septentrionalis were removed to potentially decrease its abundance. Therefore, I was not

able to determine if its removal positively impacts native treefrog abundance. However, my data

did indicate a negative correlation between the abundance of 0. septentrionalis and Hyla

femoralis. More research is needed to determine the cause of this relationship.









Table 3-1. Number of each species of native treefrog (Hyla squirella and H. cinerea) that chose
to rest in the PVC pipe refuge when paired with either Osteopilus septentrionalis or a
conspecific
Both Native only 0. septentrionalis Neither
(AB) (AO) or conspecific only (00)
(OB)
H. squirella with 0. septentrionalis 6 13 4 7
H. squirella with conspecific 12 2 9 7
H. cinerea with 0. septentrionalis 3 7 12 18
H. cinerea with conspecific 5 3 5 17


Table 3-2. Total number oftreefrogs resting in each location after 14 hours in an enclosure with
a choice of two PVC pipe refugia during experimental trials*
Experimental Control Neither


0. septentrionalis 3 11 16
H. squirella 11 9 10
H. cinerea 5 9 16
*Experimental pipes had several Osteopilus septentrionalis resting in them during the day of the
trial and were placed in the aquaria just before the start of the trial. Control pipes are new pipes,
and Neither indicates the frog was not in a pipe.









Heflick, S.K. 2001. Ecology of the exotic Cuban tree frog, Osteopilus septentrionalis, within
Brevard County, Florida. Unpublished thesis. Florida Institute of Technology,
Melbourne.

Hirai, T. 2006. Effects of bank compacting on frogs: Can the density of Hylajaponica be
recovered by installment of PVC pipe artificial refuges in the banks of rice fields?
Japanese Journal of Applied Entomology and Zoology 50:31-335. [in Japanese.]

Jaeger, R. G., E. D. Prosen, and D. C. Adams. 2002. Character displacement and aggression in
two species of terrestrial salamanders. Copeia, 2002:391-401.

Johnson, J. R. 2005. A novel arboreal pipe-trap designed to capture the gray treefrog (Hyla
versicolor). Herpetological Review 36:274-277.

Johnson, S. A. 2004. Osteopilus septentrionalis (Cuban treefrog) geographic distribution.
Herpetological Review 35:405.

Johnson, S. A. 2007. Osteopilus septentrionalis (Cuban treefrog) geographic distribution.
Herpetological Review 38:349.

Kiesecker, J. M. 2003. Invasive species as a global problem: toward understanding the
worldwide decline of amphibians. In R. D. Semlitsch (ed.), Amphibian Conservation, pp.
113-126. Smithsonian Books, Washington.

Kiesecker, J. M. and A. R. Blaustein. 1997. Population differences in responses of red-legged
frogs (Rana aurora) to introduced bullfrogs (Rana catesbeiana). Ecology, 78:1752-1760.

Kiesecker, J. M., D. P. Chivers, and A. R. Blaustein. 1996. The use of chemical cues in predator
recognition by western toad tadpoles. Animal Behavior, 52: 1237-1245.

Kupferberg, S. J. 1997. Bullfrog (Rana catesbeiana) invasion of a California river: the role of
larval competition. Ecology 78:1736-1751.

Krysko, K.L., K.M. Enge, J.H. Townsend, E.M. Langan, S.A. Johnson, and T.S. Campbell.
2005. New county records of amphibians and reptiles from Florida. Herpetological
Review. 3:85-87.

Lee, D. S. 1969. The Treefrogs of Florida. Florida Naturalist, 42:117-120.

Lockwood, J. L., P. Cassey, and T. Blackburn. 2005. The role of propagule pressure in
explaining species invasions. Trends in Ecology and Evolution, 20: 223-227.

Lutz, B. 1960. Fighting and an incipient notion of territory in male tree frogs. Copeia. 1960:61-
63.









BIOGRAPHICAL SKETCH

Kristine E. Hoffmann grew up in Sterling, Massachusetts, where she lived with her

parents, two brothers, and her twin sister, Carolyn. She graduated with honors from the

University of Massachusetts in 2005 with a B.S. in biology and a minor in wildlife and fisheries

conservation, and entered the University of Florida that fall. She hopes to teach biology and

conservation at a junior college, continue field and lab work on amphibians and reptiles, and to

help support her sister, who recently entered graduate school for costume design at New York

University.









CHAPTER 5
CONCLUSIONS

Summary

For decades it has been speculated that predation and competition by the nonindigenous

Cuban treefrog (Osteopilus septentrionalis) have caused declines in native treefrogs in Florida

(Allen and Neil 1953, Lee 1969). With the use of PVC pipe refugia, treefrog populations can

now be studied outside of the breeding season and these concerns can be addressed. However,

the use of PVC pipe refugia is a technique that is still developing, with new biases identified and

new designs reported in recent years (Zacharow et al. 2003, Johnson 2005). If we are to rely on

this method for the study of treefrogs then we must be aware of associated biases and

assumptions. I addressed two factors that contribute to the assumption that treefrogs sampled by

PVC pipe refugia accurately represent the population. With these assumptions in mind, I tried to

examine the effects of 0. septentrionalis on native treefrogs.

The results from the first laboratory study (Chapter 2) suggest that treefrog species vary

considerably in their frequency of occupying PVC pipe refugia. While the squirrel treefrog (Hyla

squirella) often chose PVC pipe refugia over a plant for cover, Osteopilus septentrionalis did so

rarely. The pinewoods treefrog (H. femoralis) and the green treefrog (H. cinerea) showed no

preference for the PVC pipe refugium or the plant. In some field studies, H. squirella was the

most common frog observed (Boughton et al. 2000, Zacharow et al. 2003, Borg et al. 2004). This

could be because either there are more H. squirella present in that area or due to its affinity for

PVC pipe refugia. The low frequency of pipe use by 0. septentrionalis is alarming as it is the

only frog found in pipes in some areas (pers. obs.). My results suggest that PVC pipe refugia are

poor indicators of the abundance of this frog compared to other species, and that there may be

many more in an area than observed using this sampling technique. Researchers using PVC









TABLE OF CONTENTS

page

A C K N O W L E D G M E N T S ..............................................................................................................4

L IST O F T A B L E S ...................................................................................................... . 7

LIST OF FIGURES .................................. .. ..... ..... ................. .8

ABSTRAC T .........................................................................................

CHAPTER

1 IN TR OD U CTION ............... ................. ............. ................. .............. 11

N onindigenou s Species ........................ .... ........................ ........ ......... 11
Nonindigenous Frogs ............. .................... .............. ............ .. 12
Three N active Treefrogs ........... .... ........ .................... ............................ 14
PVC Pipe Sampling ............. .................. ....................... ...... ..... ............ 16
O b j e ctiv e s .. ........ ................ ...........................................................................1 7

2 VARIATION IN PVC PIPE REFUGIA USE AMONG TREEFROGS ............................18

In tro d u ctio n ................. .......................................................... ................ 1 8
M methods ................. ............................................................ ............... 20
Collection and m maintenance of frogs........................................................ ............... 20
L laboratory T rials ................................................. 2 1
R e su lts .................22............................................
D iscu ssio n .................22............................................

3 TESTING FOR INTERFERENCE COMPETITION BETWEEN Osteopilus
septentrionalis AND NATIVE TREEFROGS AT PVC PIPE REFUGIA ...........................27

Introdu action ......... .............. ......................................... ............................27
M eth o d s ............................................................................................. 2 9
B ehavioral E xclusion........... ..................................................................... ........ .......... 29
C hem ical A voidance ..................... .......... .................................... .............. 30
Collection and M maintenance of Captive Frogs..................................... ............... 30
Statistical A analysis ...............................................................3 1
R results ......... .. ................ ............................ ......................... 32
B ehavioral E xclusion........... ..................................................................... ........ .......... 32
Chemical Avoidance .................................... .. .. ........ .. ............33
D iscu ssion ............... ......................................... .............................. 33









ACKNOWLEDGMENTS

I thank my committee, Steve A. Johnson, Kenneth L. Krysko, Michael Moulton, and

Robert Dorazio for their support, suggestions, help with statistics, and especially their

encouragement when the field study did not go as planned.

I thank Wekiwa Springs State Park biologists Paul Lammardo and Greg Walker, park

service specialist Jessica King, and environmental specialist Rick "Bubba" Owen for their aid in

selecting study sites, securing campsites, and for their support in general. Paul Elliot from

Southwest Florida Water Management District and the staff at Plant Adoption Landscapes

helped me secure sites to establish PVC pipe arrays to collect frogs for my laboratory studies.

Jack Rechcigl allowed me to use facilities at the IFAS Gulf Coast Research and Education

Center, and Jim Price facilitated use of growth rooms there. Mark Dykes and Monica McGarrity

housed me during my visits to Flatwoods Park. Jessie Calkins translated articles written in

Japanese. Elizabeth Roznik created maps of the field sites.

I am thankful to the many people who assisted with the fieldwork for this study.

Volunteers helped install pipe refugia, accompanied me in checking my PVC pipe refugia grids,

and aided in hand capturing frogs usually at night in the rain in ankle-deep water. For this, I

thank Travis Blunden, Erin Brown, Jill Comeau, Mark Dykes, Camille Esmel, Michelle Hawk,

Roger Herring, Carolyn Hoffmann, Steven Hoffmann, Yurii Kornilev, Monica McGarrity,

Elizabeth O'Leary, Elizabeth Roznik, Claire Sunquist, and Joshua Watson.









Table A-6. Total individual frogs captured in Wekiwa Grid F during each sampling session.
Date H. cinerea H. femoralis H. squirella 0. septentrionalis
21-Jul 0 8 0 0


Table A-7. Total individual frogs captured in Wekiwa Grid G during each sampling session.
Date H. cinerea H. femoralis H. squirella 0. septentrionalis
4-Sep 1 1 0 8
25-Sep 1 1 0 23
14-Oct 2 2 0 18
6-Nov 1 1 0 17
20-Nov 0 2 0 2


Table A-8. Total individual frogs captured in Wekiwa Grid H during each sampling session.
Date H. cinerea H. femoralis H. squirella 0. septentrionalis
4-Sep 0 1 0 1
25-Sep 0 3 0 20
14-Oct 1 1 0 22
6-Nov 0 3 0 4
20-Nov 0 0 0 8


Table A-9. Total individual frogs captured in Wekiwa Grid I during each sampling session.
Date H. cinerea H. femoralis H. squirella 0. septentrionalis
4-Sep 0 1 0 2
25-Sep 1 0 0 14
14-Oct 0 0 0 7
6-Nov 0 1 0 6
20-Nov 0 0 0 1


Table A-10. Total individual frogs captured in Wekiwa Grid J during each sampling session.
Date H. cinerea H. femoralis H. squirella 0. septentrionalis
4-Sep 1 1 0 3
25-Sep 1 3 0 5
14-Oct 1 2 0 8
6-Nov 0 4 0 2
20-Nov 0 4 0 4


Table A-11. Total individual frogs captured in Wekiwa Grid K during each sampling session.
Date H. cinerea H. femoralis H. squirella 0. septentrionalis
28-Jan 0 1 0 1
18-Feb 0 0 0 0









Wilcove, D. S., D. Rothstein, J. Dubow, A. Phillips, and E. Losos. 1998. Quantifying threats to
imperiled species in the United States: assessing the relative importance of habitat
destruction, alien species, pollution, overexploitation, and diseases. BioScience 48:607-
615

Williamson, M., and A. Fitter. 1996. The varying success of invaders. Ecology, 77:1161-1666.

Wright, A. H. 1932. Life-histories of the frogs of Okefinokee Swamp, Georgia. Macmillan
Company, New York.

Wright, A. H. and A. A. Wright. 1949. Handbook of frogs and toads of the United States and
Canada. Comstock Publishing Associates, Ithaca, New York.

Wyatt, J. L. and E. A. Forys. 2004. Conservation implications of predation by Cuban treefrogs
(Osteopilus septentrionalis) on native hylids in Florida. Southeastern Naturalist 3:695-
700.

Zacharow, M., W. J. Barichivich, and C. K. Dodd, Jr. 2003. Using ground-placed PVC pipes to
monitor hylid treefrogs: capture biases. Southeastern Naturalist 24:575-590.









frog for future identification using a unique combination of toe-clipping and the injection of

visual implant elastomer (VIE) in up to two toes. I released all native frogs and Osteopilus

septentrionalis caught within control grids, and euthanized all 0. septentrionalis captured within

the experimental grid after the initial two sampling periods to estimate initial population sizes.

Wekiwa Springs State Park

Wekiwa Springs State Park is located in Apopka, Florida, less than 20 km north of

Orlando. The park is managed by the Florida Park Service and consists of 2,800 ha surrounded

by urban development to the south, Rock Springs Run State Preserve to the northeast, and urban

development and agriculture to the west (Fig. 4-1).

During June 2006, I established my original grids at the park. I installed Grids A and B in

the southeastern corner of the park in a small mesic flatwoods opening surrounded by sandhill to

the south, hydric hammock to the north and west, and an urban/wildland interface to the east

(Fig. 4-1). Grids C, D, E, and F were in a large tract of mesic flatwoods within the interior of the

park. I checked these grids for treefrogs during 19-21 July 2006. Osteopilus septentrionalis were

captured in Grids A and B, and these grids remained in their original position throughout the

study. However, during this first sampling period it became apparent that the grids located in the

interior of the park (C, D, E, and F) did not contain 0. septentrionalis in detectable numbers, if at

all. Therefore, I removed these four grids on 8 August 2006, and established new Grids G, H, I,

and J in mesic flatwoods in the southwestern corner of the park. These new grids were

surrounded on two sides by an urban/wildland interface, and seemed likely to contain O.

septentrionalis. Indeed, during the next survey period, all grids contained 0. septentrionalis.

Removal of 0. septentrionalis from experimental grids (A, H, and J) began on 12 October 2006,

after the completion of two survey periods.









place, it is possible to use PVC pipe refugia to compare native treefrog populations prior to the

invasion of 0. septentrionalis to populations of treefrogs sampled after this nonindigenous

species has become established. However, researchers should keep in mind the possibility that O.

septentrionalis may be underrepresented.

My results are inconsistent with those of Buchanan (1988), who found that Hyla squirella

was territorial and excluded conspecific invaders from the PVC pipe refugia in the lab. I often

observed H. squirella sharing refugia with conspecifics or Osteopilus septentrionalis, although

this inconsistency may be due to the fact that I did not allow one frog to become established in

the aquarium before introducing a second frog. Stewart and Rand (1991) found that while

intruder Eleutherodactylus coqui were often evicted from refugia, resident frogs were rarely

displaced from their refugia by invaders. Thus, it may be important for a frog to establish claim

over a refugium before defending it, which might explain the difference between my results and

those of Buchanan (1988). Still, H. squirella has been observed in the wild sharing a PVC pipe

refugium with up to five adult conspecifics (Johnson, unpubl. data). Future research studies

should be conducted in which the native frog is allowed to reside in the aquarium for sufficient

time to establish residency prior to the introduction of 0. septentrionalis, and in which frogs in

the field are examined for territorial behavior.

The avoidance of PVC pipe refugia that had recently been used by Osteopilus

septentrionalis by conspecifics and not by native treefrogs suggests that these frogs may be

avoiding a potential predator that natives do not recognize. The ability to recognize and respond

appropriately to predators is extremely important to the fitness of an individual as those that fail

to do so are likely to be eaten (Chivers and Smith 1994, Kiesecker and Blaustein 1997, Griffin et

al. 2000). Although native frogs readily respond to native predators, they may not recognize and









beginning of removals. Shortly after, this species became scarce in both experimental and control

grids, and thus removal was infrequent.

Despite the lack of significant observable effects of Osteopilus septentrionalis removal, it

is noteworthy that Hylafemoralis was observed less frequently at grids where this species was

observed more frequently. While this correlation is not necessarily indicative of a cause and

effect relationship, it lends credibility to anecdotal evidence that 0. septentrionalis may be

replacing native treefrogs in areas where it has become established. However, further research is

needed to assess the potential effects of this invasive frog to actually cause declines of native

treefrogs.









these species in the field. I conducted a laboratory study in which frogs were given the choice of

resting in a PVC pipe refugium or on a plant. By using a known number of frogs in a laboratory

setting, I was able to examine not only how many frogs rested within PVC pipe refugia, but also

how many frogs chose to rest elsewhere. I tested the null hypothesis that treefrog species did not

vary in their frequency of refugia choice.

Methods

Collection and maintenance of frogs

I sampled four species of treefrogs (three natives and one nonindigenous) commonly

encountered in PVC pipe refugia studies in central Florida: green treefrog (Hyla cinerea),

squirrel treefrog (H. squirella), pinewoods treefrog (H. femoralis), and Cuban treefrog

(Osteopilus septentrionalis). I opportunistically collected treefrogs in Alachua, Hillsborough, and

Polk Counties, Florida. Collection methods included use of PVC pipe refugia and hand capture

during breeding events, rainstorms, and at night. I caught frogs at retention ponds, natural ponds,

swamps, roadside and agricultural ditches, suburban streets, county parks, and on buildings.

I transferred these frogs to growth rooms at the University of Florida Gulf Coast Research

and Education Center in Hillsborough County. The growth rooms were equipped with individual

heating, ventilation, and air conditioning (HVAC) systems, and had full spectrum fluorescent

lights suspended 0.66 m above 1-m high benches. These lights were set to mimic the natural

photoperiod during the study by automatically turning on and off at 0630 h and 2030 h,

respectively. Temperature in the growth rooms was set to 240C during the day and dropped to

210C during the night. Due to the limitations of the growth rooms, I was not able to raise the

temperature higher to more closely imitate the natural temperature regime. So that frogs would

not be foraging in absolute darkness, a high-intensity desk lamp with a 40-watt bulb was set to

shine into a corner of each growth room.









respond to nonindigenous predators (Kiesecker and Blaustein 1997, Pearl et al. 2003). I did not

investigate the possibility that native treefrogs attracted to the pipes might be more likely to be

preyed upon by Osteopilus septentrionalis, thus making the artificial refugia used in research

efforts into population sinks. If 0. septentrionalis residing inside a PVC pipe refugium consume

native treefrogs attracted to that refugium, native populations might be underrepresented in

surveys using this sampling technique. I recommend this topic for further research.





























I I I
0 1 2km


1 2 3

IG

HI I EL
SK D
NH AD B
[I] ED

FD

JD

0 100 200m

Figure 4-1. Location of each grid at Wekiwa Springs State Park (outlined in red). Above is an
image of the area, and maps 1-3 below details the location of each grid relative to the
other grids.









Left unmanaged, Osteopilus septentrionalis will continue to spread. As it invades new

areas it may cause native treefrogs to decline throughout Florida and other areas in the

southeastern United States.









1993, Kiesecker et al. 1996, Chivers, et al. 1999), and it has yet to be determined if secretion

residue left by 0. septentrionalis may be detected and avoided by native treefrogs. As detecting

chemical cues is important in amphibian ecology (Stauffer and Semlitsch 1993, Kiesecker, et al

1996, Chivers, et al. 1999), there is the potential that if native treefrogs do avoid this residue,

then 0. septentrionalis resting on a substrate might chemically interfere with the frogs that later

use that area.

I conducted two laboratory experiments to examine possible interference competition

between Osteopilus septentrionalis and native treefrogs for refugia sites. I tested the null

hypothesis that 0. septentrionalis does not affect the detection probability of native treefrogs in

PVC pipe refugia via behavioral exclusion and that native treefrogs do not avoid pipes

previously used by 0. septentrionalis (chemical exclusion).

Methods

Behavioral Exclusion

To determine if Osteopilus septentrionalis behaviorally excludes native treefrogs from

PVC pipe refugia, I conducted 30 trials for each of two native species: green treefrog (Hyla

cinerea) and squirrel treefrog (H. squirella). I paired a native frog with an 0. septentrionalis in

an aquarium with a single PVC pipe refugium; as a control, I paired frogs with conspecifics. To

prevent predation by 0. septentrionalis, I paired frogs of similar size such that frogs did not

differ more than 20% in snout-vent length (SVL). At 2000 h, I placed each pair into a 37.85-L

aquarium with a screen lid, moist sand substrate, and a 20-cm tall and 3.81-cm diameter PVC

pipe refugia planted vertically in the sand substrate in the center of the enclosure. At 1000 h the

following morning, I recorded the location of each frog as either in or out of the PVC pipe

refugium. Laboratory conditions, described below, were held constant and did not vary from

those of the growth room in which the frogs were housed.









LIST OF TABLES


Table page

3-1 Number ofHyla squirella and H. cinerea that chose to rest in PVC pipe refuge when
paired with either Osteopilus septentrionalis or a conspecific.......................................36

3-2 Total number of frogs that chose to rest in either a new PVC pipe refugia or one
used by Osteopilus septentrionalis. .............................................................................. 36

A-i Total frogs captured in Wekiwa Grid A. ........................................ ....................... 61

A-2 Total frogs captured in W ekiw a Grid B ............................................................ ........... 61

A-3 Total frogs captured in W ekiwa Grid C....................................... ......................... 61

A-4 Total frogs captured in Wekiwa Grid D. ........................................ ....................... 61

A-5 Total frogs captured in W ekiwa Grid E ........................................ ......................... 61

A-6 Total frogs captured in Wekiwa Grid F. ........................................ ........................ 62

A-7 Total frogs captured in Wekiwa Grid G. ........................................ ....................... 62

A-8 Total frogs captured in W ekiwa Grid H ........................................ ........................ 62

A-9 Total frogs captured in Wekiwa Grid I. ........................................ ......................... 62

A-10 Total frogs captured in Wekiwa Grid J....................................................... ........... 62

A-11 Total frogs captured in Wekiwa Grid K. ........................................ ....................... 62

A-12 Total frogs captured in W ekiwa Grid L ........................................ ......................... 63

A-13 Total frogs captured in W ekiwa Grid M .................................... ........................ .......... 63

A-14 Total frogs captured in Wekiwa Grid N. ........................................ ....................... 63









APPENDIX
TOTAL CAPTURES AT WEKIWA SPRINGS STATE PARK


Table A-1. Total individual frogs captured in Wekiwa Grid A during each sampling session.
Date H. cinerea H. femoralis H. squirella 0. septentrionalis
21-Jul 5 0 0 18
11-Aug 8 1 0 9
4-Sep 11 0 0 5
25-Sep 7 2 0 6
14-Oct 7 0 0 2
6-Nov 2 0 0 1
20-Nov 6 1 0 1
6-Jan 4 5 0 1
28-Jan 15 4 0 5
18-Feb 2 4 0 2


Table A-2. Total individual frogs captured in Wekiwa Grid B during each sampling session.
Date H. cinerea H. femoralis H. squirella 0. septentrionalis
21-Jul 4 0 1 16
11-Aug 7 0 0 4
4-Sep 6 0 0 10
25-Sep 5 0 0 12
14-Oct 3 1 0 7
6-Nov 2 0 0 6
20-Nov 4 0 0 7
6-Jan 3 5 0 15
28-Jan 10 0 0 3
18-Feb 6 0 0 1


Table A-3. Total individual frogs captured in Wekiwa Grid C during each sampling session.
Date H. cinerea H. femoralis H. squirella 0. septentrionalis
21-Jul 0 11 0 0


Table A-4. Total individual frogs captured in Wekiwa Grid D during each sampling session.
Date H. cinerea H. femoralis H. squirella 0. septentrionalis
21-Jul 0 9 0 0


Table A-5. Total individual frogs captured in Wekiwa Grid E during each sampling session.
Date H. cinerea H. femoralis H. squirella 0. septentrionalis
21-Jul 0 5 0 0









assist in the control of Osteopilus septentrionalis in both developed and natural settings. PVC

pipe refugia attract treefrogs that can then be captured for individual identification and marking

(Moulton et al. 1996), or removed and euthanized, as in the case of 0. septentrionalis. As this

technique becomes more widespread it is increasingly important to test the underlying basic

assumptions of this method. The assumption that PVC pipe refugia occupancy by O.

septentrionalis does not interfere with occupancy by native species through interference

competition, is a potential source of bias when sampling native treefrogs and has not yet been

evaluated.

Territorial behavior exhibited by a nonindigenous species may result in competition with

native species for refuge sites when this resource is limiting. Although most reported territorial

aggression in anurans is related to courtship (Duellman and Trueb 1986, Lutz 1960), several

treefrogs have been reported to exclude other frogs from refugia. Both male and female coquis

(Eleutherodactylus coqui) defend diurnal retreat sites, using aggressive calls and biting the heads

of intruders (Stewart and Rand 1991). Male Mexican leaf-frogs (Pachymedusa dacnicolor) use

prepollical spines on their wrists to defend burrows from conspecifics (Wiewandt 1971).

Resident male squirrel treefrogs (Hyla squirella) are known to force invaders out of short PVC

pipe refugia Buchanan (1988). Osteopilus septentrionalis is often found in aggregations

(Meshaka 1996), and territoriality has not yet been reported in this species.

In addition to the potential for behavioral interactions, there is also the potential for

chemically mediated effects between nonindigenous and native species. When threatened,

Osteopilus septentrionalis produces a milky toxin with a distinct odor (Meshaka 2001). This

toxin irritates mammalian mucous membranes (Meshaka 2001) and may also affect amphibians.

Tadpoles and adult frogs often use chemical cues to detect predators (Stauffer and Semlitsch














8-



S6
2o

a
.(
S4











Date

Figure 4-3. Total monthly precipitation at the Dover weather station in Hillsborough County
(adapted from Florida Automated Weather Network, 2007). This weather station is
located near Flatwoods Park.









On 18 November 2006, I discovered mowed firebreaks approximately 15 m wide cut

through Grids G and I, both of which were control grids. This disturbance had the potential to

alter the distribution of the frogs within each grid and introduce edge effects. Therefore, I

removed these grids on 3 January 2007. To replace the grids destroyed by mowing, I established

new Grids K and L in scrubby flatwoods just east of the previous grids and new Grids M and N

between Grids H and I. After sampling these new grids for two survey periods, and observing

less than one frog per grid, I decided I would be unable to obtain adequate sample sizes to

conduct this study. On 10 May 2007 all grids were removed from Wekiwa Springs State Park.

Flatwoods Park

Flatwoods Park, located northwest of Tampa, Florida, is part of the Hillsborough

Wilderness Park System, which contains over 2,200 ha (Fig. 4-2), and is managed by the

Southwest Florida Water Management District as the Morris Bridge Wellfield. The park is a

popular cycling location, and an 11.2-km paved bicycle loop encircles the center of the park. The

park consists predominantly of flatwoods habitat, and is surrounded by ongoing urban

development to the north and other parks in the Hillsborough Wilderness Park System to the east

and south.

On 11 July 2006, I installed six grids at Flatwoods Park (Fig. 4-2), all of which were

located north of the bicycle loop in mesic flatwoods habitat. Grids A and B were located in the

western side of the park, Grids C and D were central, and Grids E and F were near the eastern

corner of the park. The urban/wildland interface was visible from the majority of the six grids. I

began monitoring the grids on 8 August 2006 and began to remove Osteopilus septentrionalis

from experimental Grids A, D, and E on 30 September 2006. Monitoring and removal continued

through 29 June 2007, at which time the study was concluded and the grids were removed.









Statistical Analysis

I used SYSTAT, version 10.2, to perform a paired t-test to test if the total number of

Osteopilus septentrionalis varied by grid type, and thus determine if a significant number of

these invasive frogs were removed from experimental grids. I also visually inspected the number

of 0. septentrionalis captured at each grid over time to determine if fewer frogs of this species

were observed in experimental grids throughout the study than in control grids. Using program

MARK, version 4.3, I conducted a robust closed capture model with constant emigration rates to

estimate abundance of native treefrogs in each grid over time. I plotted the estimated abundances

of each grid, and visually inspected estimates for patterns. I expected that if removal of O.

septentrionalis caused native treefrog abundance to increase, then estimates of abundance of

native treefrogs in experimental grids would increase after removals began, while the abundance

in control grids would drop due to continued effects of 0. septentrionalis. As another indication

of fitness of native treefrogs within each grid, I used SYSTAT to perform a Pearson's chi-square

test to test the null hypothesis that equal numbers of gravid female H. femoralis were captured in

control and experimental grids from 7 May 2007 through 8 August 2007. Finally, as a final

attempt to determine if 0. septentrionalis affected native treefrog abundance, I used a Pearson's

Correlation Coefficient to determine if the number of 0. septentrionalis and native treefrogs in

each grid were negatively correlated.

Results

Flatwoods Park

Weather conditions were dry during the year leading up to the experiment (Fig 4-3; Florida

Automated Weather Network 2007), and capture rates of frogs at all sites throughout the

experiment were lower than anticipated. Hylafemoralis occurred regularly in PVC pipe refugia

and was observed on a total of 1,517 occasions. I captured fewer Osteopilus septentrionalis,









CHAPTER 3
TESTING FOR INTERFERENCE COMPETITION BETWEEN Osteopilus septentrionalis
AND NATIVE TREEFROGS AT PVC PIPE REFUGIA

Introduction

Competition within and among species is an important ecological interaction that can limit

the distribution and abundance of coexisting species (Gause 1934, Jaeger et al. 2002). In

exploitative competition, competition reduces the availability of a resource (Greenlees 2007). In

interference competition, the competitor reduces the access to resources (Greenlees 2007).

Nonindigenous species can harm native species through competition. When competing

taxa co-occur naturally, competition is often reduced through resource partitioning and character

displacement (Pacala and Roughgarden 1982, Jaeger et al 2002). However, when a species is

introduced into new area competition can be intense because the native fauna have not evolved

mechanisms to accommodate competition by nonindigenous species (Greenlees 2007). If the

introduced species is a superior competitor, it may cause declines, or even local extinction, in the

native species.

The Cuban treefrog (Osteopilus septentrionalis) is a nonindigenous species in Florida. It is

known to consume native treefrogs and compete with their tadpoles (Allen and Neil 1953,

Meshaka 2001, Smith 2005), and the establishment of this species in an area has been associated

with a decline in native treefrogs (Meshaka 2001). This species was first reported on Key West

in 1931 and is now established throughout most of peninsular Florida (Barbour 1931, Meshaka

2001, Krysko et al. 2005). As 0. septentrionalis continues to spread, an understanding of how it

impacts native communities, especially treefrogs, is essential to the eventual development of

management strategies for this nonindigenous frog.

The use of PVC pipe refugia is becoming a common method of sampling treefrogs

(Moulton et al. 1996, Boughton 1997, Hirai 2006). Additionally, this method has the potential to









grids with higher 0. septentrionalis abundances (Figure 4-10). A negative correlation was found

between total number of individual 0. septentrionalis and H. femoralis (log transformed data;

Pearson correlation = -0.820; Figure 4-11).

Wekiwa Springs State Park

At Wekiwa Springs State Park, removals of Osteopilus septentrionalis were only

conducted during two sessions at the western grids, and the eastern grids experienced extremely

low recapture rates (Appendix A). Thus my analysis focused on results from the Flatwoods Park

site.

Discussion

These results suggest that Osteopilus septentrionalis was not removed in sufficient

numbers to cause a decline in abundance and thereby exert an effect on Hylafemoralis. The

number of observations of 0. septentrionalis did not differ significantly between control and

experimental grids, so it is unclear if the abundance of H. femoralis would have increased had I

been able to remove more of this nonindigenous species. More research will need to be

conducted to determine if removal of 0. septentrionalis is an appropriate strategy to manage

native treefrogs.

The low capture rates experienced at the Flatwoods Park site during the study period may

have been caused by a combination of factors. First, although 2006 started off fairly wet, no

precipitation was recorded in March at the nearby Hillsborough River State Park, located within

1 km from Flatwoods Park, and the months of April and May were also somewhat dry (P. Potts,

Hillsborough River State Park, pers. comm.). In addition, Flatwoods Park is also the site of the

Morris Bridge Wellfield, an active, long-standing site of groundwater pumping by Tampa Bay

Water. This dry spell, combined with the potential effects of groundwater pumping, likely

affected treefrog communities at Flatwoods Park. Secondly, numbers of frogs at the grid













-- Grid A (removal)
O
G. -0- Grid B (control)

S40 -



C30
-o20
I1
3O









0






W ~Q~9

Date


Figure 4-7. Estimated abundance of Hylafemoralis (+ SE) at Grids A and B at Flatwoods Park.
Osteopilus septentrionalis were removed from Grid A starting on 9 September 2006.









LITERATURE CITED


Allen, E. R. and W. T. Neil. 1953. The tree frog, Hyla septentrionalis, in Florida. Copeia
1953:127-128.

Austin, S. 1975. Exotics. The Florida Naturalist, June 1975: 2-5.

Barbeau, T.R., and H. Lillywhite. 2005. Body wiping behaviors associated with cutaneous lipids
in hylid tree frogs of Florida. Journal of Experimental Biology 208:2147-2156.

Barbour, T. 1931. Another introduced frog in North America. Copeia 1931:140.

Bartareau, T M. 2004. PVC pipe diameter influences the species and sizes of treefrogs captured
in a Florida coastal oak scrub community. Herpetological Review 35:150-152.

Borg, C. K., S. K. Hoss, L. L. Smith, and L. M. Conner. 2004. From the field: a method for
preventing flying squirrel mortality in PVC pipe treefrog refugia. Wildlife Society
Bulletin 32:1313-1315.

Boughton, R G. 1997. The use of PVC pipe refugia as a trapping technique for hylid treefrogs.
Thesis, University of Florida, Gainesville, USA.

Boughton, R. G., J. Staiger, and R. Franz. 2000. Use of PVC pipe as a sampling technique for
hylid treefrogs. American Midland Naturalist 144:168-177.

Brown, R.L. 1974. Diets and habitat preferences of selected anurans in southeast Arkansas. The
American Midland Naturalist 91:468-473.

Buchanan, B. W. 1988. Territoriality in the squirrel treefrog, Hyla squirella: competition for
diurnal retreat sites. Unpublished Thesis, University of Southwestern Louisiana,
Lafayette, USA.

Carr, A. F. 1940. A contribution to the herpetology of Florida. University of Florida Biological
Publications Science Series, Vol 3, University of Florida Press, Gainesville FL.

Chivers, Douglas P., Joseph M. Kiescker, Erica L. Wildy, Lisa K. Blenden, Lee B. Kats, and
Andrew R. Blaustein. 1999. Avoidance response of post-metamorphic anurans on cues of
injured conspecifics and predators. Journal of Herpetology, 33: 472-476.

Chivers, Douglas P., and R. Jan F. Smith. 1994. Fathead minnows. Pimephalespromelas,
acquire predator recognition when alarm substance is associated with the sight of
unfamiliar fish. Animal Behavior, 48: 597-605.

Colautti, R. I., and H. J. MacIsaac. 2004. A neutral terminology to define "invasive" species.
Diversity and Distributions, 10: 135-141.









treefrogs. Before we can consider costly management plans to conserve native frogs, we need to

know if 0. septentrionalis truly does cause declines in native frogs. It is also necessary to

examine possible management strategies, such as the removal of 0. septentrionalis using PVC

pipe refugia.

I conducted a large-scale field study to determine if native treefrogs respond to the removal

of Osteopilus septentrionalis. I tested the hypothesis that if 0. septentrionalis reduces the

abundance of native treefrogs, then native treefrogs should be more abundant at sites where O.

septentrionalis was removed. I did not seek to identify the mechanism by which O.

septentrionalis could be affecting native treefrogs.

Methods

Sampling

At Wekiwa Springs State Park and Flatwoods Park in Central Florida, I selected three

study blocks based on accessibility, similarity of habitat, and proximity to development (and thus

probability of containing an established 0. septentrionalis population). In each block I

constructed a pair of 50- by 50-m grids of 1-m tall, 3.81-cm diameter PVC pipes installed

vertically at 5-m intervals. Every three weeks, I checked each grid once daily for three

consecutive days. I monitored each grid over two sampling periods to estimate initial treefrog

abundance. I then randomly selected one experimental grid of each pair at which I removed all

0. septentrionalis I encountered. Frogs were not removed from the other grids, which were used

as controls.

During each sampling period, I removed frogs from PVC pipe refugia by gently coaxing

them into a plastic bag using a "plunger" constructed from a sponge cut to the inside diameter of

the PVC pipe and attached to the end of a dowel. I identified each frog to species, measured

snout-vent length (SVL) to the nearest mm, and weighed to the nearest 0.1 g. I also marked each









LIST OF FIGURES


Figure p e

2-1 The frequency at which plant or PVC pipe was used as a refuge............... .................26

4-1 Location of each grid at Wekiwa Springs State Park. ...............................................46

4-2 Location of each grid at Flatwoods Park. ............................................... ............... 47

4-3 Total monthly precipitation for Flatwoods Park......... ..............................................48

4-4 Seasonal trend in Osteopilus septentrionalis at Grids A and B at Flatwoods Park..........49

4-5 Seasonal trend in Osteopilus septentrionalis at Grids C and D at Flatwoods Park...........50

4-6 Seasonal trend in Osteopilus septentrionalis at Grids E and F at Flatwoods Park..........51

4-7 Estimated abundance of Hylafemoralis at Grids A and B at Flatwoods Park ................52

4-8 Estimated abundance of Hylafemoralis at Grids C and D at Flatwoods Park..................53

4-9 Estimated abundance of Hylafemoralis at Grids E and F at Flatwoods Park .................54

4-10 Total Hylafemoralis and Osteopilus septentrionalis per grid .......................................55

4-11 Scatterplot matrix of Hylafemoralis and Osteopilus septentrionalis......................... 56









TESTING THE INFLUENCE OF CUBAN TREEFROGS (Osteopilus septentrionalis) ON
NATIVE TREEFROG DETECTION AND ABUNDANCE




















By

KRISTINE E. HOFFMANN


A THESIS PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE

UNIVERSITY OF FLORIDA

2007

































2007 Kristine E. Hoffmann









to conduct a Pearson chi-square test (a = 0.05) to test the null hypothesis that the frequency of

refugia use by treefrogs did not vary among species.

Results

Frequency of refugia use varied significantly by species (x2 = 21.61, df = 6, P = 0.001; Fig.

2-1); thus, I rejected the null hypothesis that these values did not differ among species.

Osteopilus septentrionalis used the plant eight times as often as the PVC pipe refugia. Hyla

squirella used the pipe refugia over twice as often as the plant. The frequency of pipe use did not

vary considerably from that of plant use for H. femoralis and H. cinerea. Frogs of all species

rarely chose to rest elsewhere within the aquarium, with three or fewer individuals per species

resting away from the PVC pipe refugia or the plant (Fig. 2-1).

Discussion

Refugia choice varied significantly among the four species of treefrogs I tested. Hyla

cinerea and H. femoralis were similar in their choice of refuges and appeared to show no

preference in their use of PVC pipe refugia or plants. In contrast, H. squirella and Osteopilus

septentrionalis showed a preference for one refuge type over the other. Hyla squirella preferred

the enclosed shelter of a PVC pipe refugium over resting on a plant, whereas 0. septentrionalis

rarely rested in a location other than on the plant. Frogs of all four species preferred a sheltered

hiding place rather than resting exposed on the wall of the aquarium or the sand substrate.

My results are similar to other observations of refugia use by the native species. Capture

and recapture rates of Hyla squirella in artificial refugia have been reported as higher than those

ofH. cinerea and H. femoralis (Goin 1958, Boughton et al. 2000, Zacharow et al. 2003). Goin

(1958) opportunistically observed H. squirella using artificial retreats more often than natural

retreats, and noted that retreats ofH. squirella were often more enclosed than the retreats of H.

cinerea. Boughton et al. (2000) captured more H. squirella than H. cinerea using PVC pipe













-0- Grid C (control)
c 25 O. 0 --0-- Grid D (removal)
O0
U)

S20 -
a
C..
15 0


e 10 -
CQ)

O5 5 -

0 0 0 .'
0





W--


Date


Figure 4-5. Seasonal trend in total individual Osteopilus septentrionalis observed per sampling
session at Grids C and D at Flatwoods Park.











4 REMOVAL OF Osteopilus septentrionalis FROM FLATWOODS HABITAT IN
C E N T R A L F L O R ID A ................................................................................... ....................37

In tro du ctio n ................... ...................3...................7..........
M e th o d s ............................................................................3 8
S am p lin g ..............................................................................................3 8
W ekiw a Spring s State P ark ........................................................................................ 39
F latw o o d s P ark .....................................................................................................4 0
Statistical A n aly sis ................................................................4 1
R e su lts ...............................................................................................4 1
F latw oods P ark .................................................................................................4 1
W ekiw a Spring s State P ark ........................................................................................ 43
D iscu ssio n ................... ...................4...................3..........

5 C O N C L U SIO N S ................................................................57

S u m m ary ................... ...................5...................7..........
R e co m m en d atio n s.............................................................................................................. 5 9

APPENDIX: TOTAL CAPTURES AT WEKIWA SPRINGS STATE PARK ............................61

L IT E R A T U R E C IT E D ............................................................................................................. 64

B IO G R A PH IC A L SK E T C H .................................................................................................... 70




























6










H. femoralis
m 0. septentrionalis


Grid E Grid B Grid A Grid F Grid D Grid C


Grid

Figure 4-10. Total individual Hylafemoralis and Osteopilus septentrionalis per grid over the
course of the study. Grids are arranged in order of increasing 0. septentrionalis.











30 Grid C (control)
-0- Grid D (removal)

o 25 -


S20










E
(0





W 0



-5
-5 I5 I




Date

Figure 4-8. Estimated abundance of Hylafemoralis (+ SE) at Grids C and D at Flatwoods Park.
Osteopilus septentrionalis were removed from Grid D starting on 9 September 2006.
Missing values were unrealistically estimated as over 500 individuals, due to low
capture and recapture rates, and are not shown on this figure.









suggested that predation on other treefrogs is infrequent (Meshaka 2001, S. A. Johnson unpubl.

data), another study (Heflick 2001) suggests that 0. septentrionalis preys heavily on H. cinerea

in some urban areas. Because of the stresses 0. septentrionalis exerts on native communities in

Florida, this frog might negatively impact native treefrog populations.

Three Native Treefrogs

Similar habitat affinities and behavior of other treefrogs in the family Hylidae make them

especially vulnerable to the effects of Osteopilus septentrionalis. These frogs are mostly arboreal

and have adhesive disks on the tip of each digit that allow them to climb (Dickerson 1908, Wells

2007). This adaptation allows treefrogs to utilize microhabitats not available to other frogs.

Treefrogs are nocturnal foragers, and are relatively small (Dickerson 1908). Native treefrogs can

easily be preyed on by 0. septentrionalis and have been found repeatedly in diet studies of 0.

septentrionalis (Meshaka 2001, Heflick 2001, S. A. Johnson unpubl. data). Thus, native

treefrogs are active in similar habitats at the same time as 0. septentrionalis, and may be more

likely than other frogs to be influenced by interactions with this nonindigenous species.

Hyla cinerea is a medium-sized treefrog with mature females measuring 41-63 mm and

males 37-59 mm (Wright and Wright 1949). This species occurs from Virginia to Florida, west

to Texas, and north along the Mississippi River to southern Illinois (Wright and Wright 1949,

Hardy 1972, Redmer and Brandon 2005). Hyla cinerea is common throughout central Florida,

and is often observed on buildings at night. This species consumes arthropods, including

Cicadellidae, Acrididae, Lepidoptera larvae, Chrysomelidae, and Arachnida (Brown 1974). Its

breeding season ranges from April to August (Wright and Wright 1949). Eggs are laid near the

surface of permanent, stagnant water bodies and hatch within 2-3 days into tadpoles that

metamorphose in about 60 days (Hardy 1972). Adult H. cinerea remain near wetlands on









CHAPTER 4
REMOVAL OF Osteopilus septentrionalis FROM FLATWOODS HABITAT IN CENTRAL
FLORIDA

Introduction

Some nonindigenous species may directly affect communities of native species by preying

on or competing with native species, or indirectly by introducing new diseases, altering nutrient

cycles, fire regimes, and other aspects of the environment that then create habitat less suitable for

native species (Simberloff 2000). Wilcove et al. (1998) ranked competition and predation by

nonindigenous species as the second greatest threat to imperiled plants and animals in the United

States, surpassed only by habitat loss caused by humans.

The potential for the nonindigenous Cuban treefrog (Osteopilus septentrionalis) to

negatively affect native treefrogs has been evident since its discovery on Key West, Florida

(Barbour 1931). In a letter to Barbour (1931), A. G. Elbon reported that this frog was common

on the island, had been there since before he moved to the island, and that it fed almost

exclusively on smaller frogs. Twenty years later, Schwartz (1951) documented 0. septentrionalis

on mainland Florida and the rapid development of its eggs. Shortly after, Allen and Neill (1952)

observed 0. septentrionalis feeding on breeding frogs, and warned that the presence of this frog

might affect native frogs. Lee (1969) repeated this warning, although he cautioned that there

were no data to substantiate this potential threat at that time.

Recently, studies have begun to confirm the adverse effects of Osteopilus septentrionalis

on native frogs. Osteopilus septentrionalis tadpoles have been found to consume and reduce the

growth rate of native tadpoles (Smith 2005a, Smith 2005b). Although 0. septentrionalis has

been observed consuming native frogs since the time of its introduction, adults appear to prey

mainly upon insects and less frequently on other frogs (Lee 1969, Wyatt and Forys 2004,

Meshaka 2001). These and other stresses may add up to a population-level effect on native
























LOGHYFE


LOGOSSE


Figure 4-11. Scatterplot matrix of the log of the total number of individual Hylafemoralis
(LOGHYFE) and Osteopilus septentrionalis (LOGOSSE). The Pearson correlation
coefficient of these two logs was -0.820.









of the nonindigenous frog and the native Hylafemoralis. This finding relates back to the original

question of the severity of predation and competition by 0. septentrionalis on the natives, and

suggests that this nonindigenous frog might cause declines in native treefrog populations. Further

research is needed to examine other possible causes of this correlation and to determine the

effects of removing this frog, and to determine if removal using PVC pipe is a suitable

management strategy for 0. septentrionalis.

The detection of native frogs in PVC pipe refugia does not appear to be affected by the

presence of Osteopilus septentrionalis, but populations of native treefrogs do appear to be

negatively affected. These findings support the anecdotes that 0. septentrionalis might be

replacing native treefrogs in central Florida.

Recommendations

As Osteopilus septentrionalis seems to negatively affect populations of native treefrogs, I

recommend that these nonindigenous frogs be removed when observed. This species appears

abundant in areas where they are established, and management of this species may require great

effort.

As I had low capture rates of this frog in flatwoods habitat, I recommend that PVC pipe

refugia be placed where these frogs are more likely to occur. These areas should include near

wetland and around buildings, but further research should be conducted to determine where PVC

pipe refugia would be most effective in attracting this nonindigenous frog.

Other methods, such as road cruising during rainstorms or hand capture at breeding sites,

should supplement removals using PVC pipe refugia. As I found Osteopilus septentrionalis to

use PVC pipe refugia less frequently than native treefrogs, these additional methods may be

necessary to remove individuals that do not use PVC pipe refugia.









Schwartz, A.. 1952. Hyla septentrionalis Dumeril and Bibron on the Florida mainland. Copeia,
1952:117-118.

Schwarzkopf, L. and R. A. Alford. 1996. Desiccation and shelter-site use in a tropical
amphibian: comparing toads with physical models. Functional Ecology 10:193-200.

Simberloff, D. 2000. Nonindigenous species: a global threat to biodiversity and stability. In P.H.
Raven (ed.), Nature and human society, pp. 325-334. National Academic Press,
Washington D.C.

Smith, K. G. 2004. Osteopilus septentrionalis (Cuban treefrog); Reproductive behavior.
Herpetological Review, 35:374-375.

Smith, K. G. 2005a. Effects of nonindigenous tadpoles on native tadpoles in Florida: evidence of
competition. Biological Conservation 123:443-441.

Smith, K.G. 2005b. An exploratory assessment of Cuban treefrog (Osteopilus septentrionalis)
tadpoles as predators of native and nonindigenous tadpoles in Florida. Amphibia-reptilia
26:571-575.

Stauffer, H. P. and R. D. Semlitsch. 1993. Effects of visual, chemical and tactile cues of fish on
the behavioral responses of tadpoles. Animal Behavior, 43: 355-364.

Stewart. M. M. and F. H. Pough. 1983. Population density of tropical forest frogs: relation to
retreat sites. Science 221:570-572.

Stewart, M. M. and A. S. Rand. 1991. Vocalization and the defense of retreat sites by male and
female frogs, Eleutherodactylus coqui. Copeia 1991:1013-1024.

Torchin, M. E., K. D. Lafferty, A. P. Dobson, V. J. McKenzie, and A. M. Kuris. 2003.
Introduced species and their missing parasites. Nature, 421: 628-630.

Walsh, P. T. and J. R. Downie. 2005. The effects of shelter availability and substrate quality on
behavior and post-metamorphic growth in three species of anurans: implications for
captive breeding. Herpetological Journal 15:245-255.

Walters, E.C. and J. M. Kneitel. 2004. Use of water-filled red-cockaded woodpecker cavities by
other organisms. In R. Costa and S. J. Daniels (eds). Red-cockaded woodpeckers: road to
recovery. Hancock House Publishers. Baline, Washington pp 492-497.

Welker, M.E. 2004. Geographic distribution: Osteopilus septentrionalis (Cuban treefrog).
Herpetological Review, 35:283.

Wiewandt, T. A. 1971. Breeding biology of the Mexican leaf frog. Fauna. 2:29-34.









Variation in design and placement of PVC pipe refugia can bias sampling and may alter

the probability of detection of some treefrog species or size classes. Differences in pipe design

such as pipe diameter, length, shape (straight or T-shaped), or if the pipe is capped at one end

can alter the species composition and size class of treefrogs sampled (Boughton 1997, Zacharow

et al. 2003, Bartareau 2004). Location of PVC pipe refugia in terms of distance to buildings,

height above ground, tree type (i.e., hardwood of softwood), arboreal or ground placement, and

distance to water can also affect the sample (McComb and Nobel 1981, Boughton 1997,

Zacharow et al. 2003). Since frogs change their behavior according to temperature, humidity,

and rain, detection also varies seasonally (Buchanan 1988, Boughton 1997, Zacharow et al.

2003). Regardless of pipe configuration or placement, some species are not effectively sampled

with this method. Although present at their study site, Boughton et al. (2000) did not observe

spring peepers (Pseudacris crucifer) and only rarely found barking treefrogs (Hyla gratiosa) in

their PVC pipe refugia. Despite our understanding about the effects of pipe design and

placement, variation in frequency of refugia use among species easily sampled with this method

has not been examined.

The propensity to use cavities may vary among treefrogs due to differences in their natural

history and physiology. Treefrogs that remain near wetlands or have evolved morphological

adaptations to prevent desiccation or predation may not frequently seek out refugia. Treefrogs

vary in the amount of cutaneous lipid secretions they produce, and in their repertoire of

associated body wiping behaviors that aid in water retention and thus inhibit desiccation

(Barbeau and Lillywhite 2005).

I examined variation in frequency of PVC pipe refugia use by four species of treefrogs in

order to better understand potential variation in detection probability, and thus occurrence, of









CHAPTER 1
INTRODUCTION

Nonindigenous Species

Nonindigenous species are those that have been transported to an area in which they do not

naturally occur (Austin 1975, Kiesecker and Blaustein 1997). Some nonindigenous species arrive

as stowaways on cargo, plants, and with human travelers; some are released or escape from

laboratories, aquaculture operations, or the global trade in live organisms; some have been

released deliberately by humans as crops, game, unwanted pets, bait, biological controls, or for

aesthetics; and others have moved in on their own aided by roads and canals (Elton 1958,

Simberloff 2000, Kiesecker 2003). Some nonindigenous species rapidly colonize new areas

(Austin 1975), becoming widespread (stage IVa), dominant (stage IVb), or both (stage V)

(Colautti and MacIssac 2004). Without natural controls nonindigenous species may alter the

ecosystem into which they have been introduced. They may directly affect natural communities

by preying on native species that do not have adequate defenses against invading species.

Nonindigenous species may also compete with native species for food and essential microhabitat

(Simberloff 2000, Meshaka 2001). They may also produce toxins and harm native predators that

try to consume them (Phillips et al. 2003). Nonindigenous species can harm natives indirectly by

introducing new diseases, or altering nutrient cycles, fire regimes, and other aspects of the

environment that then create less suitable habitat for native species (Simberloff 2000). The result

is an increase in abundance of the nonindigenous species and a decrease in abundance of natives,

and even the possible extinction of native species (Elton 1958, Smith 2005a). Wilcove et al.

(1998) ranked competition and predation by nonindigenous species as the second greatest threat

to imperiled plants and animals in the United States, surpassed only by habitat loss due to

humans.









locations may have been extremely low. Squirrel treefrogs (Hyla squirella) chose PVC pipes

over other refugia more often than other treefrogs and thus are most likely to be observed (see

Chapter 2); however few individuals of this species were observed in the grids. Green treefrogs

(H. cinerea) are associated with wetlands (Goin 1958) and although I have often observed this

species a distance from permanent water, it might not have occupied the grid areas in detectable

numbers. Third, frogs living at the grid sites might simply not be using the PVC pipe refugia.

Recaptured frogs were often encountered in or near refugia of initial capture after avoiding

detection for multiple days or sampling sessions. This suggests that either they returned to the

same area after absence from the grid or, perhaps more likely, that they made use of other

available refugia within the grid. Other researchers have observed higher capture rates in PVC

pipe refugia positioned > 2 m above the ground (Boughton et al. 2000), and it is possible that

frogs within the study area utilized taller structures as refugia and did not encounter the PVC

pipes. Finally, native frog populations in the grids may have already experienced substantial

reductions subsequent to the establishment of Osteopilus septentrionalis within the park.

Additionally, this nonindigenous species seems to prefer natural refugia over PVC pipe refugia,

and may have been underrepresented during sampling for this study (see Chapter 2).

Seasonal changes are known to occur in the activity of Osteopilus septentrionalis

(Meshaka 2001) and in the use of PVC refugia by some native treefrogs (Boughton 1997,

Zacharow et al. 2003); thus, it is not surprising that the use of PVC pipe refugia by O.

septentrionalis varied seasonally during the study period. The observed behavioral variation of

0. septentrionalis in conjunction with low capture rates served to confound potential effects of

removal, as seasonal variation in PVC pipe refugia use by this species was asynchronous to its

removal. Captures of this species peaked near the beginning of the study, before and during the



































For Ron
Rest in Peace









which I found a total of 532 times. H. squirella was only observed 14 times, and H. cinerea only

once. Therefore, my analysis focused on H. femoralis and 0. septentrionalis.

The total number of individual Osteopilus septentrionalis observed in control grids did not

differ significantly from that of the experimental grids (t = -0.77, df =2, P = 0.52). Due to low

capture rates of this species, I was unable to estimate their abundance using MARK. I plotted and

inspected counts of 0. septentrionalis, which did not vary visibly between experimental and

control grids, and it appears that my removal had no effect on the abundance of this species.

However, these counts did vary seasonally, with more frogs being observed in summer and early

fall than in winter and spring (Figures 4-4, 4-5, 4-6).

Based on visual observation of abundance estimates of Hylafemoralis over time, I did not

find compelling trends of increased abundance in experimental grids and decreased abundance in

control grids. It does not appear the removals affected the abundance of this native species.

Abundance of this frog in experimental Grid A was often, though not consistently, higher than in

its partner control grid (Figure 4-7). On occasions when the estimated abundance of Grid A

differed most from that of Grid B, standard errors were high. Populations of H. femoralis in

Grids C and D appeared to be small, and on many days no frogs of this species were observed.

The low capture rates in these grids created highly variable estimates with high standard errors

(Figure 4-8). Only in Grid E did abundance of H. femoralis appear to increase with the removal

of 0. septentrionalis (Figure 4-9). Other indicators of fitness did not vary by grid type. The ratio

of gravid females observed during the 2007 breeding season did not differ significantly

according to grid type (x2 = 12, df = 10, P = 0.285). All H. femoralis observed were adults at first

capture, and did not experience sufficient growth (snout-vent length and weight) to enable me to

rule out error as a source of variation in measurements. However, there were less H. femoralis in









Table A-12. Total individual frogs captured in Wekiwa Grid L during each sampling session.
Date H. cinerea H. femoralis H. squirella 0. septentrionalis
28-Jan 0 0 0 2
18-Feb 0 0 0 1


Table A-13. Total individual frogs captured in Wekiwa Grid M during each sampling session.
Date H. cinerea H. femoralis H. squirella 0. septentrionalis
28-Jan 0 1 0 0
18-Feb 0 0 0 2


Table A-14. Total individual frogs captured in Wekiwa Grid N during each sampling session.
Date H. cinerea H. femoralis H. squirella 0. septentrionalis
28-Jan 0 0 0 1
18-Feb 0 1 0 1









Chemical Avoidance

To determine if native treefrogs avoid PVC pipe refugia that have recently been used by

Osteopilus septentrionalis, I tested treefrogs by placing them in an aquarium with one pipe that

had been used recently by 0. septentrionalis and one new, clean pipe. In control trials, I placed

each treefrog in an aquarium with two clean pipes. To create the treatment pipes, I housed 30

large (> 45 mm SVL) 0. septentrionalis in a 113-L plastic enclosure with a moist sand substrate,

small dish of water, screen lid, potted vinca plant (Catharanthus roseus), and 20 PVC pipe

refugia. On nights before a trial, I replaced the refugia and the plant in this container with 15

new, clean refugia so that the 0. septentrionalis would rest in these pipes. At 2000 h the next

night, I removed 0. septentrionalis from these now used pipes, and placed these pipes opposite

new, clean pipes in 37-L aquaria with sand substrate and screen lids. One frog was then

introduced to each aquarium. At 1000 h the location of the frog was recorded. To control for

possible directional bias, the side of the enclosure on which each treatment pipe was placed was

alternated among tanks. To discourage smaller frogs from resting between the lid and lip of the

aquarium, I filled this space with a rope that was secured with non-toxic glue. A high-intensity

desk lamp with a 40-watt bulb was placed in the growth room where the tanks were housed to

provide light so that the frogs would not be in complete darkness at night. Laboratory conditions,

described below, were held constant and did not vary from those of the growth room in which

the frogs were housed.

Collection and Maintenance of Captive Frogs

I opportunistically collected treefrogs by use of PVC pipe refugia and by hand. Collection

occurred primarily at night, particularly during breeding events and rainstorms. Frogs were

collected from 10 May 2007 through 26 August 2007 in Alachua, Hernando, Hillsborough,

Manatee, Orange, Pinellas, Polk, Sarasota, and Seminole counties, Florida. I caught frogs at














0.


O ..... 0


-*- Grid E (removal)
.0 -- Grid F (control)


14 -


S12-
0
U)
U)
(a 10
()

8-


S6


Q 4-


2

0
I-- -
1-^


I I I I I I I I I I I I


q < < rV (V V NVX'

Date


Figure 4-6. Seasonal trend in total individual Osteopilus septentrionalis observed per sampling
session at Grids E and F at Flatwoods Park.


O ......









emergent vegetation (Wright and Wright 1949). Habitats include marshes, cypress swamps,

artificial wetlands, and edificarian areas (Hardy 1972, Delis et al. 1996).

Hyla squirella is a small frog; mature females measure 23-37 mm and males 23-36 mm

(Wright and Wright 1949). The range of Hyla squirella is similar to that of H. cinerea, but does

not extend as far north along the Mississippi River Valley (Mitchell and Lannoo 2005, Redmer

and Brandon 2005). Its diet of consists of small arthropods including Elateridae, Gryllidae,

Formicidae, and small crustaceans and spiders (Duellman and Schwartz 1958). During the

spring and summer breeding season adults migrate to ephemeral ponds (Wright 1932, Goin

1958, Mitchell and Lannoo 2005). Breeding sites include roadside ditches, wooded ponds, and

grassy pools (Wright 1932). Eggs hatch in about two days, and tadpoles metamorphose within

25-55 days (Wright 1932). Adult frogs are generally found away from wetlands, seeking shelter

under bark, in bromeliads, and along the axils of palmettos (Wright 1932, Goin 1958). This frog

occurs in urbanized areas and all major habitat types in Florida (Duellman and Schwartz 1958,

Delis et al. 1996).

The pinewoods treefrog (Hylafemoralis) is also a small frog; mature females measure 23-

40 mm and males 24-37 mm (Wright and Wright 1949). Hylafemoralis occurs along the coastal

plain southward from Virginia to Louisiana (Mitchell 2005). It feeds on a variety of arthropods,

including Tetrigidae, Gryllidae, Carabidae, Elateridae, Scarabaeidae, Formacidiae, Vespidae,

Attidae, Hymenoptera, and Lepidoptera (Duellman and Schwartz 1958, Carr 1940). Adults breed

in temporary wetlands from June to October (Duellman and Schwartz 1958), eggs hatch in three

days, and tadpoles metamorphose within 50-75 days (Wright and Wright 1949). As adults, these

frogs are generally associated with pine forests (Duellman and Scwartz 1958). Hylafemoralis is

sensitive to habitat degradation and is not often found in edificarian areas (Delis et al. 1996).




























0 1 2km


1 2 3



A ] C D F
O O E D


BW
C-1






0 100 200m

Figure 4-2. Location of each grid at Flatwoods Park (outlined in red). Above is an image of the
area, and maps 1-3 below details the location of each grid relative to the other grids.
















-0- Grid A (removal)
** 0- Grid B (control)


I I I I I I I I I I I I


c b C', N le,' V b' a' -'


Date


Figure 4-4. Seasonal trend in total individual Osteopilus septentrionalis observed per sampling
session at Grids A and B at Flatwoods Park.


--
O 20-
U)
()
U)

(D 15-



*O 10 -



) 5-


0
- 0









Not all species introduced into new areas become established, and not all that do become

established cause problems. Some are quickly preyed upon, out-competed, parasitized, or

infected with native diseases (Elton 1958). When individuals die without reproducing, or when

there is no recruitment, a recently introduced species cannot persist. Those that do establish

themselves tend to have similar characteristics. These include reduced parasite load, high

reproductive output, broad feeding niche, ability to thrive in disturbed ecosystems, tolerance of a

wide variety of conditions, introduction in high numbers, and introduction to areas where

predators are absent (Willamson and Fitter 1996, Crossland 2000, Torchin et al. 2003, Lockwood

et al. 2005).

Nonindigenous Frogs

The ability of nonindigenous frogs to become established outside their native range is

demonstrated by two well known examples, the cane toad (Bufo marinus) and the bull frog

(Rana catesbeiana). Bufo marinus, native to South America north to southern Texas, was

introduced into Australia in 1935 in an effort to control beetles that damaged sugar cane crops.

This toad did not control the beetles as planned, but became a nuisance when it began to

reproduce explosively and consume and poison native fauna. Bufo marinus produces strong

toxins (Bufotoxins) throughout its life cycle that can result in mortality of native predators at

such high frequency that the toad now threatens native populations (Phillips et al. 2003,

Crossland 2000). Rana catesbeiana, native to the eastern United States, was translocated to the

western United States and is now associated with the decline of native lowland frogs (McAuliffe

1978, Kiesecker 2003). This frog consumes a variety of prey, including tadpoles and adults of

other frogs. The yellow-legged frog (R. muscosa), a rare species impacted by R. catesbeiana, is

not naturally sympatric with this nonindigenous frog, and when prey upon by R. catesbeiana it

does not recognize this larger frog as a predator, and does not behave defensively in its presence









'J-J
-- grid E (removal)
S--0- grid F (control)





30 -

0- 40
S3o








(0




S20 4

Date


Figure 4-9. Estimated abundance of Hylafemoralis (+ SE) at Grids E and F at Flatwoods Park.
Osteopilus septentrionalis were removed from Grid E starting on 9 September 2006.









retention ponds, along the edges of natural wetlands, in roadside and agricultural ditches, on

suburban streets, in forested county parks, on walls of buildings, and in parking lots.

I transferred these frogs to growth rooms at the University of Florida's Gulf Coast

Research and Education Center in Hillsborough County. I maintained frogs in captivity in

accordance with University of Florida Institutional Animal Care and Use Committee (protocol

number E870). Growth rooms were equipped with individual heating, ventilation, and air

conditioning (HVAC) systems set to 240C. The rooms also had full-spectrum fluorescent lights

set to automatically turn on at 0630 h and off at 2030 h to mimic the natural photoperiod. A high-

intensity desk lamp with a 40-watt bulb shining into a corner provided light so that the frogs

would not be in complete darkness at night. I housed frogs individually in 0.75-L Sterilite

containers with 3-mm air holes drilled in the lids during the entire length of their captivity. I fed

them crickets, misted their containers daily, and replaced their paper towel substrate as needed.

I conducted experiments from 15 May 2007 through 31 August 2007. After completion of

each trial I marked native treefrogs with Visual Implant Elastomer (VIE) and released them at

their initial capture site. All Osteopilus septentrionalis were humanely euthanized, as this is a

nonindigenous species in Florida.

Statistical Analysis

To test null the hypothesis that Osteopilus septentrionalis does not behaviorally exclude

native frogs from PVC pipe refugia, I compared the probability of refuge use by native frogs in

the presence of 0. septentrionalis with the probability of refuge use when paired with a

conspecific. Probabilities were calculated using a maximum likelihood model base on frequency

of refugia use by pairs of frogs. To calculate probabilities when two native frogs were paired,

each was randomly assigned post hoc identification as frog A or frog B. Four combinations of

outcomes were possible: (1) both frogs were in the pipe (AB), (2) only native frog A was in the









refugia, but also caught very few H. femoralis. Hyla squirella was the only species of treefrog

found by Walters and Kneitel (2004) to use red-cockaded woodpecker (Picoides borealis)

cavities. Similar numbers of individual H. cinerea and H. squirella were caught in PVC pipe

refugia by Zacharow et al (2003), but H. squirella was more frequently recaptured. Hyla

squirella occurred more often than H. cinerea in PVC pipe refugia during a study by Bartareau

(2004). However, McComb and Nobel (1981) found H. cinerea to use natural cavities and

wooden nest boxes more frequently than H. squirella. These observations are similar to my

results as H. squirella used the PVC pipe refugia more often than H. cinerea and H. femoralis,

but differed in that H. femoralis did not often occur in refugia during these field studies.

Variation in the use of refugia among treefrogs may be due to differences in their natural

history and physiology. Hyla cinerea exhibits relatively high evaporative water loss, and engages

in vigorous head wiping to stimulate secretion of cutaneous lipids that limit desiccation (Barbeau

and Lillywhite 2005). Frogs of this species breed in permanent wetlands, and are often found in

vegetation within and around wetlands (Wright and Wright 1949, Redmer and Brandon 2005).

As they typically remain near permanent water, these frogs may rely on wetlands for moisture

and may be less likely to seek out refugia primarily for water retention (Goin 1958). In contrast,

Hyla squirella breeds in temporary ponds, and Goin (1958) hypothesized that when this frog

disperses to upland habitats it seeks refugia that are more sheltered in order to prevent water loss.

In addition, Barbeau and Lillywhite (2005) reported that H. squirella exhibits a higher number of

body wipes before entering a water conservation posture than other Florida treefrogs. Hyla

femoralis also breeds in temporary wetlands and appears to spend much of its time in tree

canopies, though specific adaptations to avoid desiccation are unknown (Wright and Wright

1949, Mitchell 2005). The nonindigenous Osteopilus septentrionalis typically breeds in









Information regarding detection probability of frogs using this method and the effects of

species interactions are lacking. Although trends in species composition are apparent in the

literature, variation in the frequency of PVC pipe refugia use, and thus detection probability,

have not been compared across species encountered using this method. It is also not known how

behavioral interactions among treefrogs affect use of the refugia, or if these interactions vary

depending on gender, size, and species of frogs. These factors also need to be addressed to

identify potential biases in sampling with PVC pipe refugia.

Objectives

I examined the use of PVC pipe refugia as a method for sampling treefrogs, especially

when Osteopilus septentrionalis is present, and to examine the potential influence of 0.

septentrionalis on native treefrog populations. Using laboratory studies, I determined the

frequency of use of PVC pipe refugia when vegetative refugia were present for four species of

treefrogs (Hyla cinerea, H. squirella, H. femoralis, and 0. septentrionalis), and tested the null

hypothesis that this frequency did not vary among species. I tested the hypothesis that 0.

septentrionalis affects the probability of pipe use by two native treefrog species (H. cinerea and

H. squirella). I tested the null hypotheses that 0. septentrionalis does not affect the use of PVC

pipe refugia of native treefrogs by behavioral exclusion and that native treefrogs do not avoid

refugia previously used by 0. septentrionalis. I also designed a field study to examine potential

ecological effects of 0. septentrionalis on native treefrogs (H. cinerea, H. squirella, and H.

femoralis). I tested the hypothesis that if 0. septentrionalis does reduce the abundance of native

treefrogs, then native treefrogs should be more abundant at sites where 0. septentrionalis was

removed.









Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Master of Science

TESTING THE INFLUENCE OF CUBAN TREEFROGS (Osteopilus septentrionalis) ON
NATIVE TREEFROG DETECTION AND ABUNDANCE

By

Kristine E. Hoffmann

December 2007

Chair: Steve A. Johnson
Major: Wildlife Ecology and Conservation

The Cuban treefrog (Osteopilus septentrionalis) is a nonindigenous species in Florida that

anecdotally replaces native treefrog species where it becomes established. However, few

empirical data are available to verify this claim; thus, I sought to determine the effects of O.

septentrionalis on native treefrog populations. I removed 0. septentrionalis from natural areas

and monitored the response of native treefrog populations using polyvinyl chloride (PVC) pipe

refugia. In conjunction with my field study, I conducted laboratory experiments to determine the

frequency at which treefrogs use PVC pipe refugia and to test the null hypothesis that O.

septentrionalis does not affect behaviorally or chemically exclude native treefrogs from refugia.

In my first laboratory study, I compared the variation by species in frequency of refugia

used when treefrogs were presented with the choice of either a PVC pipe refugia or a plant. I

found that the native squirrel treefrog (Hyla squirella) rested in the pipe twice as often as on the

plant, and Osteopilus septentrionalis used the pipes infrequently. The other two natives tested,

green treefrog (H. cinerea) and pinewoods treefrog (H. femoralis), did not strongly favor either

the PVC pipe refugia or the plant.

In my other laboratory studies, I tested if Osteopilus septentrionalis interferes with the use

of PVC pipe refugia by native treefrogs. First, I paired a native treefrog (Hyla sp.) with an O.









CHAPTER 2
VARIATION IN PVC PIPE REFUGIA USE AMONG TREEFROGS

Introduction

The use of cavities provides many benefits to frogs, such as predator avoidance,

thermoregulation, breeding sites, and osmoregulation (Stewart and Rand 1991, Schwarzkopf and

Alford 1996, Walsh and Downie 2005). For example, in the southeastern United States, treefrogs

are known to use tree cavities as diurnal shelters presumably to avoid desiccation (McComb and

Noble 1981, Ritke and Gabb 1991, Walters and Kneitel 2004). Researchers have taken advantage

of this behavior to study the ecology of treefrogs by deploying artificial refugia that mimic

natural cavities. As early as the 1950s, Goin (1958) used aluminum cans inverted on wooden

stakes to compare the non-breeding habits of two treefrog species. More recently, researchers

have used polyvinyl chloride (PVC) pipes to attract treefrogs for study.

The use of PVC pipe refugia is becoming commonplace as a method to monitor and

sample treefrogs (Moulton et al. 1996, Boughton et al. 2000, Zacharow et al. 2001). This

method is superior to call surveys for studies focused specifically on treefrogs because PVC pipe

refugia can be used to sample both genders during both the breeding and non-breeding seasons.

PVC pipe refugia allow for capture and identification of individual frogs as well. Refugia are

also better suited than drift fence arrays to the study of treefrogs, as treefrogs may traverse these

barriers (Dodd 1991). Furthermore, refugia passively attract (rather than trap) treefrogs (Moulton

et al. 1996). As treefrogs are not restrained and can enter and leave at will, use of this method

eliminates trap mortality and allows researchers to check PVC pipe refugia infrequently. Because

of this, there is no need to remove PVC pipe refugia between sampling periods, which reduces

the amount of labor needed for repeat sampling or long-term monitoring.









temporary wetlands and has skin co-ossified to the dermal bones in its skull, an adaptation that

reduces the amount of vascular tissue in the skin on its head and thus the rate of water loss from

this area (Duellman and Treub 1986, Meshaka 2001). Osteopilus septentrionalis is relatively

large and secretes a slimy toxin when disturbed (Meshaka 2001), so it may be less reliant on

refugia for protection against predators than native treefrogs that lack this protection.

Researchers using PVC pipe refugia to sample treefrog populations should take this

variation in refugia use, and thus detection probability, into account when using this method to

study frog communities. A dominance ofHyla squirella in a sample may not necessarily indicate

that this species is more abundant than other treefrogs in an area, and a lack of Osteopilus

septentrionalis in a sample may not indicate absence of this species.

The low detectability of Osteopilus septentrionalis obfuscates interpretation of the

occurrence of this frog when PVC pipe refugia are used in monitoring programs. As this species

does not frequently use the refugia, managers cannot be certain if 0. septentrionalis is absent

when it is not observed. When this species is observed, there may be several times more

individuals present than sampled using this method. However, as this species uses PVC pipe

refugia considerably less often than native treefrogs, managers should be concerned when it is

the dominant frog found in the PVC pipe refugia (see Chapter 3). Researchers need to consider

using mark-recapture methods rather than counts when comparing abundance across species.

Management of this nonindigenous species is complicated by the strong aversion of

Osteopilus septentrionalis to PVC pipe refugia. These frogs should be removed when observed;

however, my results indicate that only a small portion of the frogs can be captured using this

method. To attract a larger percentage of the population of 0. septentrionalis, researchers need to

investigate the optimal placement and configuration of PVC pipe refugia. Until a better









management strategy is available, managers should not rely on PVC pipe refugia alone to catch

these frogs. Other techniques, such as dip netting larvae and hand capturing breeding adults,

must be used to supplement removal efforts. Otherwise, frogs that do not use the pipes will not

be captured.









(Kiesecker and Blaustein 1997, Kupferberg 1997). Also, R. catesbeiana has larger tadpoles that

out-compete other tadpoles (Kiesecker 2003).

Numerous characteristics, several of which are shared between Rana catesbeiana and Bufo

marinus, are suspected to contribute to their successful invasion of new areas. These include

large body size, broad feeding niche, high fecundity, ability to thrive in developed areas, and

toxicity (e.g. B. marinus). The Cuban treefrog (Osteopilus septentrionalis) displays all of these

traits (Schwartz 1952, Lee 1969, Meshaka 2001). Perhaps then it is no coincidence that this frog

may also have negative impacts on native fauna where it has been introduced.

Native to Cuba, Isla de Juventud, the Cayman Islands, and the Bahamas (Meshaka 2001),

Osteopilus septentrionalis was first observed in the United States on Key West, Florida, in 1931

(Barbour 1931), and on the Florida mainland in 1951 (Schwartz 1952). Since then, this frog has

expanded its range northward throughout most of peninsular Florida and isolated locations in the

Coastal Plain (Welker 2004, Krysko et al. 2005, Johnson 2004, 2007). Osteopilus septentrionalis

thrives in disturbed areas, but also invades natural habitats (Meshaka 2001, Meshaka and Babbitt

2005). Although 0. septentrionalis consumes mostly invertebrates (Heflick 2001, Meshaka

2001), it is large enough that it also preys upon and may compete with native treefrogs (Barbour

1931, Lee 1969, Meshaka 2001). Additionally, Smith (2005a) found that southern toad (Bufo

terrestris) and green treefrog (Hyla cinerea) tadpoles metamorphosed later and at a smaller size

in the presence of introduced 0. septentrionalis tadpoles. Breeding 0. septentrionalis can also

interfere with the breeding of native frogs, as males have been observed amplexing with other

species (Smith 2004). Anecdotal accounts of native treefrogs being replaced by O.

septentrionalis are common, yet the severity of the effects of competition and predation on the

community of native treefrogs remains to be demonstrated. Whereas some studies of gut contents









I housed frogs individually in 0.75-L Sterilite containers with 3-mm air holes drilled in the

lids. I held frogs separately during the entire length of their captivity. I fed the frogs crickets,

misted their containers daily, and replaced their paper towel substrate as needed.

Laboratory Trials

Frogs were maintained in captivity and tested following protocols approved by the

Institutional Animal Care and Use Committee of the University of Florida (protocol number

E870). Frogs were collected from 2 June 2007 until 23 July 2007, and trials were conducted from

18 June 2007 until 24 July 2007. No frog remained in captivity longer than 18 days. After

completion of each trial I marked the native frogs with Visual Implant Elastomer (VIE) and

released these treefrogs at their site of initial capture. All Osteopilus septentrionalis were

euthanized humanely, as this is a nonindigenous species in Florida and cannot be legally

released.

For each species, I conducted 30 overnight trials, whereby individual frogs were offered a

choice between a plant and a PVC pipe refugium as a diurnal refuge. During each trial, a frog

was individually placed into a 37.8-L aquarium with moist sand substrate, screen lid, 3.8 cm

diameter PVC pipe that stood 20 cm tall near one end of the tank, and an approximately 20-cm

tall vinca (Catharanthus roseus) planted at the other end of the tank. To ensure the tank would

remain humid throughout the night, I saturated the substrate until standing water was observed.

Trials began at 2000 h and ended the next day at 1000 h, when frogs in natural habitats would no

longer be active and would have sought diurnal shelter. At the completion of each trial I recorded

the location of the frog as being inside the pipe, on the plant (i.e., on leaves, stems, or at the base

of the plant), or elsewhere (on the side of the tank or on the sand substrate). I measured snout-

vent length (SVL) to the nearest mm and weighed each frog to the nearest 0.1 g. Each individual

was used for only one trial, and lab conditions were the same for all trials. I used SYSTAT 10.2









PVC Pipe Sampling

Treefrogs are not easily captured outside their breeding season using standard amphibian

trapping techniques. Because of their arboreal nature, they do not normally utilize cover boards

and cannot be found by raking through forest debris. Additionally, their toe pads allow them to

climb over typical drift fence arrays (Dodd 1991), and therefore fences must be modified with

plastic barriers (Murphy 1993). As an alternative to these trapping techniques, some researchers

have been developing refugia to attract treefrogs.

The use of refugia has progressed over the last fifty years. Goin (1958) used tin cans

resting upside down on stakes to attract treefrogs. While these cans attracted numerous treefrogs,

a single pipe found nearby on the ground attracted nearly twice as many frogs as all six cans

combined (Goin 1958). McComb and Noble (1981) used wooden nest boxes of varying sizes to

sample reptiles and amphibians, and found treefrogs to use boxes more frequently than natural

tree cavities. Stewart and Pough (1983) used bamboo refugia to study the coqui

(Eleutherodactylus coqui), and Moulton et al. (1996) introduced the use of polyvinyl chloride

(PVC) pipes to capture treefrogs.

With the use of PVC pipes established, researchers turned their attention to identifying and

improving on sources of bias. To design a PVC pipe refugium that would capture all species and

size classes of treefrogs present in central Florida, Boughton et al. (2000) examined the effects of

pipe diameter and length, the use of caps to retain water, placement on hardwoods versus pines,

and height at which pipes were placed; Zacharow et al. (2003) studied the effects of PVC pipe

diameter of ground-placed pipes; and Bartareau (2004) examined the effects of pipe diameter in

oak scrub habitat. Borg et al. (2004) modified pipe refugia to curtail entrapment and mortality of

non-target flying squirrels (Glaucomys spp.), a source of bias that reduced refugia use by hylids.









McAuliffe, J. R. 1978. Biological survey and management of sport-hunted bullfrog populations
in Nebraska. Nebraska Game and Parks Commission, Lincoln, Nebraska.

McComb, W. C., and R. E. Noble. 1981. Herpetofaunal use of natural tree cavities and nest
boxes. The Wildlife Society Bulletin. 9:261-267.

Meshaka, W. E. 1996. Retreat use by the Cuban treefrog (Osteopilus septentrionalis);
implications for successful colonization in Florida. Journal of Herpetology. 30:443-445.

Meshaka, W. E. 2001. The Cuban treefrog in Florida: life history of a successful colonizing
species. University Press of Florida, Gainesville, Florida.

Meshaka, W. E. and K. J. Babbitt. 2005. Amphibians and reptiles: status and conservation in
Florida. Krieger Publishing Company, Malabar, Florida.

Mitchell, J. C. 2005. Hylafemoralis Bosc, 1800, pine woods treefrog. Pages 454-455 in M.
Lannoo, editor. Amphibian Declines: Conservation Status of the United States Species.
University of California Press, Berkley, USA.

Mitchell, J. C. and M. J. Lannoo. 2005. Hyla squirella Bosc,1800, squirrel treefrog. Pages 456-
458 in M. Lannoo, editor. Amphibian Declines: Conservation Status of the United States
Species. University of California Press, Berkley, USA.

Murphy, C. G. 1993. A modified drift fence for capturing treefrogs. Herpetological Review,
24:143-145.

Moulton, C. A., W. J. Flemming, and B. R. Nerney. 1996. The use of PVC pipes to capture hylid
frogs. Herpetological Review 27:86-187.

Pacala, S. W., and J. Roughgarden. 1982. Resource partitioning and interspecific competition in
two two-species insular Anolis lizard communities. Science 217:444-446.

Pearl, Christopher A., Michael J. Adams, Gerald S. Schuytema, and Alan V. Nebeker. 2003.
Behavioral responses of anuran larvae to chemical cues of native and introduced
predators in the pacific northwestern United States. Journal ofHerpetology, 37: 572-576.

Phillips, B. L., G. P. Brown, and R. Shine. 2003. Assessing the potential impact of cane toads on
Australian snakes. Conservation Biology 17:1738-1747.

Redmer, M. and R. A. Brandon. 2005. Hyla cinerea (Schneider, 1799), green treefrog. Pages
452-454 in M. Lannoo, editor. Amphibian Declines: Conservation Status of the United
States Species. University of California Press, Berkley, USA.

Ritke, M. E. and J. G. Babb. 1991. Behavior of the gray treefrog (Hyla chrysoscelis) during the
non-breeding season. Herpetological Review 22:5-8.









pipe (AO), (3) only Osteopilus septentrionalis or the native frog B was in the pipe (OB), and (4)

neither frog was in the pipe (00). The probabilities of these outcomes are as follows:

PAB = PA x PB
PAO = PA (1- PB)
POB= (- PA) X PB
Po = (1- PA) (1- PB)

I used Program R, version 2.5.1, to calculate the maximum likelihood estimates of the

probability that native treefrog (frog A when paired with a conspecific) was in the pipe (PA), and

that the 0. septentrionalis or frog B was in the pipe (PB). I used the program to calculate the

likelihood ratio -2 (log Lreduced model Lfull model) for each species to compare the probability of the

native frog being in the pipe when paired with an 0. septentrionalis to the probability of native

frog A being in the pipe when paired with a conspecific (native frog B). I used the likelihood

ratios in a chi-square analysis to test the null hypotheses that the native treefrogs are equally

likely to occupy the pipe in the presence of 0. septentrionalis as with a conspecific.

To analyze the outcome of my chemical avoidance experiment, I tested the null hypothesis

that there was no difference in occupancy of clean PVC pipe refugia and those that had recently

housed Osteopilus septentrionalis. To do this I used Microsoft Excel to conduct a chi-square test

for each species to determine if the frogs were more likely to use one type of pipe (used by O.

septentrionalis vs. clean) over the other. I used the results of the control trials as expected values

for this analysis.

Results

Behavioral Exclusion

The probability ofHyla squirella occupying PVC pipe refugia in the presence of

Osteopilus septentrionalis was not significantly different from the probability ofH. squirella

occupying the refugia in the presence of a conspecific (LR = 1.69, P =0.193). These probabilities









Crossland, M. R. 2000. Direct and indirect effects of the introduced Cane Toad Bufo marinus
(Anura: Bufonidae) on populations of native anuran larvae in Australia. Ecography, 23:
283-290.

Delfino, G., R. Brizzi, D. Nosi, and A. Terreni. 2002. Serous cutaneous glands in New World
hylid frogs: an ultrastructural study on skin poisons confirms phylogenetic relationships
between Osteopilus septentrionalis and Phrynohyas venulosa. Journal of Morphology
253:176-186.

Delis, P.R., H.R. Mushinsky, and E.D. McCoy. 1996. Decline of some west-central Florida
anuran populations in response to habitat degradation. Biodiversity and Conservation
5:1579-1595.

Dodd, C. K. Jr. 1991. Drift fence-associated sampling bias of amphibians at a Florida sandhills
temporary pond. Journal of Herpetology 25:296-301.

Dodd, C. K. Jr. 1996. Use of terrestrial habitats by amphibians in the sandhill upland on north-
central Florida. Alytes 14:42-52.

Duellman, W. E. and Schwartz, A. 1958. Amphibians and reptiles of southern Florida. Bulletin
of the Florida State Museum 3:181-324.

Duellman, W. E., and L. Trueb. 1986. Biology of Amphibians. McGraw-Hill Publishing
Company, New York, USA.

Elton, C. S. 1958. The ecology of invasions by animals and plants. University of Chicago Press,
Chicago and London.

Florida Automated Weather Network. 2007. Archived Weather Data. IFAS. University of
Florida. . Accessed 18 October 2007.

Gause, G. F. 1934. The struggle for existence. Williams and Wilkins, Baltimore.

Goin, O. B. 1958. A comparison of the nonbreeding habits of two treefrogs, Hyla squirella and
Hyla cinerea. Quarterly Journal of the Florida Academy of Sciences 21:49-60.

Greenlees, M. J. G. P. Brown, J. K. Webb, B. L. Philips, and R. Shine. 2007. Do invasive cane
toads (Chaunua marinus) compete with Australian frogs (Cyclorana australis)? Austral
Ecology 32: in press.

Griffin, Andrea S., Daniel T. Blumstein, and Christopher S. Evans. 2000. Training captive-bred
or translocated animals to avoid predators. Conservation Biology, 14: 1317-1326.

Hardy, J. 1972. Tentative outline for inventory of amphibians: Hyla cinerea (green tree frog).
Chesapeake Science, Vol 13; Supplement: Biota of the Chesapeake Bay, pp. S180-S190.









were also not significant for H. cinerea (LR = 0.130, P = 0.718; Table 3-1). Therefore, I failed to

reject the null hypothesis that native treefrogs were equally as likely to rest in a PVC pipe

refugium with an 0. septentrionalis as with a conspecific.

Chemical Avoidance

Native treefrogs did not demonstrate significant avoidance of PVC pipe refugia that had

been recently used by Osteopilus septentrionalis (x2 =0.86, df = 2, P = 0.65 for Hyla squirella,

and x2 =1.54, df = 2, P = 0.46 for H. cinerea; Table 3-2). Although 0. septentrionalis used to

treat the pipes were often found sharing refugia in their holding enclosure, with 30 frogs

congregated in approximately half of the 15 available clean pipes, 0. septentrionalis in

experimental trials rested in used PVC pipe refugia significantly less frequently than in control

pipes or other locations (on glass or sand substrate; x2 =7.11, df = 2, P = 0.03). Therefore, I failed

to reject the null hypothesis that native treefrogs were equally likely to use a PVC pipe refugium

that had recently been used by 0. septentrionalis as new, clean refugia, but did reject the null

than 0. septentrionalis would use these pipes equally.

Discussion

Neither experiment indicated that Osteopilus septentrionalis interferes with native treefrog

refugia use and detection during studies using PVC pipe refugia. The frequency of pipe use by

native treefrogs was not altered by the presence of 0. septentrionalis or by any residue that

might have been left on pipes recently used by this nonindigenous frog. However, O.

septentrionalis avoided refugia that had recently been used by its batracophagic conspecifics.

These results support the underlying assumption that the presence of 0. septentrionalis does not

reduce the detection probability of native treefrogs, and thus competition for refugia is unlikely.

However, the results indicate that as it avoids refugia recently used by conspecifics, 0.

septentrionalis may be underrepresented in samples using this method. With this assumption in
















20

4--
0
15
-0
E

z 10



5



0I
Osteopilus H. squirella H. cinerea H. femoralis

Species


Figure 2-1. The frequency of refugia type used by individuals of four treefrog species after 14
hours in an aquarium with a plant and a PVC pipe. Other resting sites included the
side of the tank, the corner where the lid met the tank, and the sand substrate.




Full Text

PAGE 1

1 TESTING THE INFLUENCE OF CUBAN TREEFROGS ( Osteopilus septentrionalis ) ON NATIVE TREEFROG DETECTI ON AND ABUNDANCE By KRISTINE E. HOFFMANN A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2007

PAGE 2

2 2007 Kristine E. Hoffmann

PAGE 3

3 For Ron Rest in Peace

PAGE 4

4 ACKNOWLEDGMENTS I thank m y committee, Steve A. Johnson, Kenneth L. Krysko, Michael Moulton, and Robert Dorazio for their support, suggestions, help with statistics, and especially their encouragement when the field study did not go as planned. I thank Wekiwa Springs State Park biologists Paul Lammardo and Greg Walker, park service specialist Jessica King, and environmenta l specialist Rick Bubba Owen for their aid in selecting study sites, securing campsites, and for their support in general. Paul Elliot from Southwest Florida Water Management District and the staff at Pl ant Adoption Landscapes helped me secure sites to establish PVC pipe ar rays to collect frogs fo r my laboratory studies. Jack Rechcigl allowed me to use facilities at the IFAS Gulf Coast Research and Education Center, and Jim Price facilitated use of growth rooms there. Ma rk Dykes and Monica McGarrity housed me during my visits to Flatwoods Park. Jessie Calkins translated articles written in Japanese. Elizabeth Roznik created maps of the field sites. I am thankful to the many people who a ssisted with the fieldwork for this study. Volunteers helped install pipe refugia, accompanie d me in checking my PVC pipe refugia grids, and aided in hand capturing frogs usually at night in the rain in ankle-deep water. For this, I thank Travis Blunden, Erin Brown, Jill Comeau, Ma rk Dykes, Camille Esmel, Michelle Hawk, Roger Herring, Carolyn Hoffmann, Steven Hoffmann, Yurii Kornilev, Monica McGarrity, Elizabeth OLeary, Elizabeth Roznik, Cl aire Sunquist, and Joshua Watson.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4 LIST OF TABLES................................................................................................................. ..........7 LIST OF FIGURES.........................................................................................................................8 ABSTRACT.....................................................................................................................................9 CHAP TER 1 INTRODUCTION..................................................................................................................11 Nonindigenous Species...........................................................................................................11 Nonindigenous Frogs..............................................................................................................12 Three Native Treefrogs......................................................................................................... ..14 PVC Pipe Sampling................................................................................................................16 Objectives...............................................................................................................................17 2 VARIATION IN PVC PIPE REFUGIA USE AMONG TREEFROGS................................ 18 Introduction................................................................................................................... ..........18 Methods..................................................................................................................................20 Collection and maintenance of frogs............................................................................... 20 Laboratory Trials............................................................................................................. 21 Results.....................................................................................................................................22 Discussion...............................................................................................................................22 3 TESTING FOR INTERFERENCE COMPETITION BETWEEN Osteopilus septen trionalis AND NATIVE TREEFROGS AT PVC PIPE REFUGIA............................ 27 Introduction................................................................................................................... ..........27 Methods..................................................................................................................................29 Behavioral Exclusion....................................................................................................... 29 Chemical Avoidance....................................................................................................... 30 Collection and Maintenance of Captive Frogs................................................................30 Statistical Analysis.......................................................................................................... 31 Results.....................................................................................................................................32 Behavioral Exclusion....................................................................................................... 32 Chemical Avoidance....................................................................................................... 33 Discussion...............................................................................................................................33

PAGE 6

6 4 REMOVAL OF Osteopilus septentr ionalis FROM FLATWOODS HABITAT IN CENTRAL FLORIDA............................................................................................................ 37 Introduction................................................................................................................... ..........37 Methods..................................................................................................................................38 Sampling..........................................................................................................................38 Wekiwa Springs State Park............................................................................................. 39 Flatwoods Park................................................................................................................ 40 Statistical Analysis.......................................................................................................... 41 Results.....................................................................................................................................41 Flatwoods Park................................................................................................................ 41 Wekiwa Springs State Park............................................................................................. 43 Discussion...............................................................................................................................43 5 CONCLUSIONS.................................................................................................................... 57 Summary.................................................................................................................................57 Recommendations................................................................................................................ ...59 APPENDIX: TOTAL CAPTURES AT WEKI WA SPRINGS STATE PARK ............................ 61 LITERATURE CITED..................................................................................................................64 BIOGRAPHICAL SKETCH.........................................................................................................70

PAGE 7

7 LIST OF TABLES Table page 3-1 Number of Hyla squirella and H. cinerea that chose to rest in PVC pipe refuge when paired with eithe r Osteopilus septentrionalis or a conspecific..........................................36 3-2 Total number of frogs that chose to rest in either a new PVC pipe refugia or one used by Osteopilus septentrionalis ....................................................................................36 A-1 Total frogs captured in Wekiwa Grid A. ........................................................................... 61 A-2 Total frogs captured in Wekiwa Grid B. ............................................................................ 61 A-3 Total frogs captured in Wekiwa Grid C ............................................................................. 61 A-4 Total frogs captured in Wekiwa Grid D. ........................................................................... 61 A-5 Total frogs captured in Wekiwa Grid E. ............................................................................ 61 A-6 Total frogs captured in Wekiwa Grid F. ............................................................................ 62 A-7 Total frogs captured in Wekiwa Grid G. ........................................................................... 62 A-8 Total frogs captured in Wekiwa Grid H ............................................................................ 62 A-9 Total frogs captured in Wekiwa Grid I. ............................................................................. 62 A-10 Total frogs captured in Wekiwa Grid J. ............................................................................. 62 A-11 Total frogs captured in Wekiwa Grid K. ........................................................................... 62 A-12 Total frogs captured in Wekiwa Grid L. ............................................................................ 63 A-13 Total frogs captured in Wekiwa Grid M. ........................................................................... 63 A-14 Total frogs captured in Wekiwa Grid N. ........................................................................... 63

PAGE 8

8 LIST OF FIGURES Figure page 2-1 The frequency at which plant or PVC pipe was used as a refuge. .....................................26 4-1 Location of each grid at Wekiwa Springs State P ark........................................................ 46 4-2 Location of each grid at Flatwoods Park........................................................................... 47 4-3 Total monthly precipita tion for Flatwoods Park.. ..............................................................48 4-4 Seasonal trend in Osteop ilus septentrionalis at Grids A and B at Flatwoods Park........... 49 4-5 Seasonal trend in Osteop ilus septentrionalis at Grids C and D at Flatwoods Park........... 50 4-6 Seasonal trend in Osteop ilus septentrionalis at Grids E and F at Flatwoods Park............ 51 4-7 Estimated abundance of Hyla femor alis at Grids A and B at Flatwoods Park.................. 52 4-8 Estimated abundance of Hyla femor alis at Grids C and D at Flatwoods Park.................. 53 4-9 Estimated abundance of Hyla femor alis at Grids E and F at Flatwoods Park................... 54 4-10 Total Hyla femoralis and Osteopilus septentrionalis per grid .......................................... 55 4-11 Scatterplot matrix of Hyla femoralis and Osteopilus septentr ionalis ...............................56

PAGE 9

9 Abstract of Thesis Presen ted to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science TESTING THE INFLUENCE OF CUBAN TREEFROGS ( Osteopilus septentrionalis ) ON NATIVE TREEFROG DETECTI ON AND ABUNDANCE By Kristine E. Hoffmann December 2007 Chair: Steve A. Johnson Major: Wildlife Ecology and Conservation The Cuban treefrog (Osteopilus septentrionalis ) is a nonindigenous spec ies in Florida that anecdotally replaces native treefrog species where it becomes established. However, few empirical data are available to verify this claim; thus, I sough t to determine the effects of O. septentrionalis on native treefrog populations. I removed O. septentrionalis from natural areas and monitored the response of native treefrog po pulations using polyvinyl chloride (PVC) pipe refugia. In conjunction with my field study, I co nducted laboratory experiments to determine the frequency at which treefrogs use PVC pipe re fugia and to test th e null hypothesis that O. septentrionalis does not affect behaviorally or chemically exclude native treefrogs from refugia. In my first laboratory study, I compared the va riation by species in frequency of refugia used when treefrogs were presented with the choi ce of either a PVC pipe refugia or a plant. I found that the native squirrel treefrog ( Hyla squirella) rested in the pipe twice as often as on the plant, and Osteopilus septentrionalis used the pipes infrequently. The other two natives tested, green treefrog ( H. cinerea) and pinewoods treefrog ( H. femoralis ), did not strongly favor either the PVC pipe refugia or the plant. In my other laboratory studies, I tested if Osteopilus septentrionalis interferes with the use of PVC pipe refugia by native treefrogs. First, I paired a native treefrog ( Hyla sp.) with an O.

PAGE 10

10 septentrionalis in an aquarium with only one pipe, and th en observed which frog(s) rested inside the PVC pipe refugium 14 hours later. I found that O. septentrionalis did not behaviorally exclude native treefrogs from refugia. In anot her study, I presented treefro gs with a choice of a new PVC pipe refugium and one that had been recently used by O. septentrionalis to determine if treefrogs avoided refugia recently used by this predator. I found that Hyla cinerea and H. femoralis did not avoid refugia that were recently used by O. septentrionalis These laboratory studies suggest that nativ e treefrog detection in PVC pipe refugia is not altered by the presence of similar sized O. septentrionalis. In my field experiment, I established paired grid s of PVC pipe refugia to attract treefrogs at three study sites at Wekiwa Springs State Park, Apopka, Florida, and at three study sites at Flatwoods Park, Hillsborough Wilderness Park, Tampa, Florida. One grid of each pair was used as a treatment grid where I removed Osteopilus septentrionalis and the other served as a control. I monitored the community of native treefrogs within each grid, and hypothesized that if O. septentrionalis reduces the abundance of native treefrogs, then at sites where this species was removed native treefrogs should become more a bundant. However, capture rates were low and few O. septentrionalis were removed to potentially decrease its abundance. Therefore, I was not able to determine if its removal positively imp acts native treefrog abundance. However, my data did indicate a negative correla tion between the abundance of O. septentrionalis and Hyla femoralis More research is needed to determin e the cause of this relationship.

PAGE 11

11 CHAPTER 1 INTRODUCTION Nonindigenous Species Nonindigenous species are those that have been transported to an ar ea in which they do not naturally occur (Austin 1975, Kiesecker and Blau stein 1997). Som e nonindi genous species arrive as stowaways on cargo, plants, and with human trav elers; some are released or escape from laboratories, aquaculture operati ons, or the global trade in live organisms; some have been released deliberately by humans as crops, game, unwanted pets, bait, biol ogical controls, or for aesthetics; and others have moved in on th eir own aided by roads and canals (Elton 1958, Simberloff 2000, Kiesecker 2003). Some nonindige nous species rapidly colonize new areas (Austin 1975), becoming widespread (stage IVa) dominant (stage IVb), or both (stage V) (Colautti and MacIssac 2004). Without natural controls nonindigenous species may alter the ecosystem into which they have been introduced. They may directly affect natural communities by preying on native species that do not have adequate defenses against invading species. Nonindigenous species may also compete with na tive species for food and essential microhabitat (Simberloff 2000, Meshaka 2001). They may also pr oduce toxins and harm native predators that try to consume them (Phillips et al. 2003). Noni ndigenous species can harm natives indirectly by introducing new diseases, or altering nutrient cycles, fire regi mes, and other aspects of the environment that then create less suitable habitat for native species (Simberloff 2000). The result is an increase in abundance of the nonindigenous species and a d ecrease in abundance of natives, and even the possible extinction of native sp ecies (Elton 1958, Smith 2005a). Wilcove et al. (1998) ranked competition and predation by nonindi genous species as the second greatest threat to imperiled plants and animals in the Unite d States, surpassed only by habitat loss due to humans.

PAGE 12

12 Not all species introduced into new areas become established, and not all that do become established cause problems. Some are quickly preyed upon, out-competed, parasitized, or infected with native diseases (Elton 1958). When individuals die without reproducing, or when there is no recruitment, a recen tly introduced species cannot pers ist. Those that do establish themselves tend to have similar characteristic s. These include reduced parasite load, high reproductive output, broad feeding niche, ability to thrive in disturbed ecosystems, tolerance of a wide variety of conditions, introduction in high numbers, and introduction to areas where predators are absent (Willamson and Fitter 1996, Crossland 2000, Torchin et al. 2003, Lockwood et al. 2005). Nonindigenous Frogs The ability of nonindigenous frogs to becom e established outside their native range is demonstrated by two well known examples, the cane toad ( Bufo marinus ) and the bull frog ( Rana catesbeiana ). Bufo marinus, native to South America north to southern Texas, was introduced into Australia in 1935 in an effort to control beetles that damaged sugar cane crops. This toad did not control the beetles as pl anned, but became a nuisance when it began to reproduce explosively and consume and poison native fauna. Bufo marinus produces strong toxins (Bufotoxins) throughout its life cycle that can re sult in mortality of native predators at such high frequency that the toad now threat ens native populations (Phillips et al. 2003, Crossland 2000). Rana catesbeiana native to the eastern United States, was translocated to the western United States and is now associated with the decline of native lo wland frogs (McAuliffe 1978, Kiesecker 2003). This frog consumes a variety of prey, including ta dpoles and adults of other frogs. The yellow-legged frog ( R. muscosa), a rare species impacted by R. catesbeiana, is not naturally sympatric with this no nindigenous frog, and when prey upon by R. catesbeiana it does not recognize this larger frog as a predator, and does not behave defens ively in its presence

PAGE 13

13 (Kiesecker and Blaustein 1997, Kupferberg 1997). Also, R. catesbeiana has larger tadpoles that out-compete other tadpoles (Kiesecker 2003). Numerous characteristics, severa l of which are shared between Rana catesbeiana and Bufo marinus are suspected to contribute to their successful invasion of new areas. These include large body size, broad feeding niche, high fecundit y, ability to thrive in developed areas, and toxicity (e.g. B. marinus ). The Cuban treefrog ( Osteopilus septentrionalis ) displays all of these traits (Schwartz 1952, Lee 1969, Meshaka 2001). Perhap s then it is no coincidence that this frog may also have negative impacts on nativ e fauna where it has been introduced. Native to Cuba, Isla de Juventud, the Cayman Islands, and the Bahamas (Meshaka 2001), Osteopilus septentrionalis was first observed in the United States on Key West, Florida, in 1931 (Barbour 1931), and on the Florida mainland in 1951 (Schwartz 1952). Since then, this frog has expanded its range northward throughout most of pe ninsular Florida and isol ated locations in the Coastal Plain (Welker 2004, Krysko et al. 2005, Johnson 2004, 2007). Osteopilus septentrionalis thrives in disturbed areas, but also invades natural habitats (Meshaka 2001, Meshaka and Babbitt 2005). Although O. septentrionalis consumes mostly invert ebrates (Heflick 2001, Meshaka 2001), it is large enough that it also preys upon and may compete with native treefrogs (Barbour 1931, Lee 1969, Meshaka 2001). Additionally, Smith (2005a) found that southern toad (Bufo terrestris ) and green treefrog ( Hyla cinerea ) tadpoles metamorphosed later and at a smaller size in the presence of introduced O. septentrionalis tadpoles. Breeding O. septentrionalis can also interfere with the breeding of native frogs, as males have been observed amplexing with other species (Smith 2004). Anecdotal accounts of native treefrogs being replaced by O. septentrionalis are common, yet the severity of the effects of competition and predation on the community of native treefrogs remains to be demonstrated. Whereas some studies of gut contents

PAGE 14

14 suggested that predation on ot her treefrogs is infrequent (Meshaka 2001, S. A. Johnson unpubl. data), another study (Hef lick 2001) suggests that O. septentrionalis preys heavily on H. cinerea in some urban areas. Because of the stresses O. septentrionalis exerts on native communities in Florida, this frog might negatively impact native treef rog populations. Three Native Treefrogs Sim ilar habitat affinities and behavior of other treefrogs in the family Hylidae make them especially vulnerable to the effects of Osteopilus septentrionalis These frogs are mostly arboreal and have adhesive disks on the tip of each digit that allow them to climb (Dickerson 1908, Wells 2007). This adaptation allows treef rogs to utilize microhabitats not available to other frogs. Treefrogs are nocturnal foragers and are relatively small (Dickerson 1908). Native treefrogs can easily be preyed on by O. septentrionalis and have been found repeatedly in diet studies of O. septentrionalis (Meshaka 2001, Heflick 2001, S. A. Johnson unpubl. data). Thus, native treefrogs are active in similar habitats at the same time as O. septentrionalis, and may be more likely than other frogs to be influenced by interactions with th is nonindigenous species Hyla cinerea is a medium-sized treefrog with mature females measuring 41-63 mm and males 37-59 mm (Wright and Wright 1949). This speci es occurs from Virginia to Florida, west to Texas, and north along the Mississippi Rive r to southern Illinois (Wright and Wright 1949, Hardy 1972, Redmer and Brandon 2005). Hyla cinerea is common throughout central Florida, and is often observed on buildi ngs at night. This species c onsumes arthropods, including Cicadellidae, Acrididae, Lepidopt era larvae, Chrysomelidae, and Arachnida (Brown 1974). Its breeding season ranges from April to August (Wri ght and Wright 1949). Eggs are laid near the surface of permanent, stagnant water bodies an d hatch within 2 days into tadpoles that metamorphose in about 60 days (Hardy 1972). Adult H. cinerea remain near wetlands on

PAGE 15

15 emergent vegetation (Wright and Wright 1949). Habitats include marshes, cypress swamps, artificial wetlands, and edificarian areas (Hardy 1972, Delis et al. 1996). Hyla squirella is a small frog; mature female s measure 23-37 mm and males 23-36 mm (Wright and Wright 1949). The range of Hyla squirella is similar to that of H. cinerea but does not extend as far north along the Mississippi River Valley (Mitchell and Lannoo 2005, Redmer and Brandon 2005). Its diet of consists of sm all arthropods including Elateridae, Gryllidae, Formicidae, and small crustaceans and spiders (Duellman and Schwartz 1958). During the spring and summer breeding season adults mi grate to ephemeral ponds (Wright 1932, Goin 1958, Mitchell and Lannoo 2005). Breeding sites in clude roadside ditches, wooded ponds, and grassy pools (Wright 1932). Eggs hatch in abou t two days, and tadpoles metamorphose within 25-55 days (Wright 1932). Adult frogs are generally found away from wetlands, seeking shelter under bark, in bromeliads, and along the axils of palmettos (Wright 1932, Goin 1958). This frog occurs in urbanized areas and all major habita t types in Florida (Due llman and Schwartz 1958, Delis et al. 1996). The pinewoods treefrog (Hyla femoralis ) is also a small frog; mature females measure 2340 mm and males 24-37 mm (Wright and Wright 1949). Hyla femoralis occurs along the coastal plain southward from Virginia to Louisiana (Mitchell 2005). It f eeds on a variety of arthropods, including Tetrigidae, Gryllidae, Carabidae, Elateridae, Scarabaeidae, Formacidiae, Vespidae, Attidae, Hymenoptera, and Lepidoptera (Duellman and Schwartz 1958, Carr 1940). Adults breed in temporary wetlands from June to October (Due llman and Schwartz 1958), eggs hatch in three days, and tadpoles metamorphose within 50-75 da ys (Wright and Wright 1949). As adults, these frogs are generally associated with pi ne forests (Duellman and Scwartz 1958). Hyla femoralis is sensitive to habitat degradation and is not often found in edificarian areas (Delis et al. 1996).

PAGE 16

16 PVC Pipe Sampling Treefrogs are not eas ily capture d outside their breeding season using standard amphibian trapping techniques. Because of their arboreal na ture, they do not normally utilize cover boards and cannot be found by raking through forest debris Additionally, their to e pads allow them to climb over typical drift fence arrays (Dodd 1991), a nd therefore fences must be modified with plastic barriers (Murphy 1993). As an alternative to these trapping techniques, some researchers have been developing ref ugia to attract treefrogs. The use of refugia has progressed over the la st fifty years. Goin (1958) used tin cans resting upside down on stakes to attract treefrogs While these cans attrac ted numerous treefrogs, a single pipe found nearby on the ground attracted nearly twice as many frogs as all six cans combined (Goin 1958). McComb and Noble (1981) used wooden nest boxes of varying sizes to sample reptiles and amphibians, and found treefrogs to use boxes more frequently than natural tree cavities. Stewart and Pough (1983) used bamboo refugia to study the coqui ( Eleutherodactylus coqui), and Moulton et al. (1996) intr oduced the use of polyvinyl chloride (PVC) pipes to capture treefrogs. With the use of PVC pipes established, research ers turned their attention to identifying and improving on sources of bias. To design a PVC pipe refugium that would capture all species and size classes of treefrogs present in central Florida, Boughton et al (2000) examined the effects of pipe diameter and length, the use of caps to reta in water, placement on hardwoods versus pines, and height at which pipes were placed; Zacharow et al. (2003) st udied the effects of PVC pipe diameter of ground-placed pipes; an d Bartareau (2004) examined the effects of pipe diameter in oak scrub habitat. Borg et al. ( 2004) modified pipe refugia to curt ail entrapment and mortality of non-target flying squirrels ( Glaucomys spp.), a source of bias that re duced refugia use by hylids.

PAGE 17

17 Information regarding detection probability of frogs using th is method and the effects of species interactions are lacking. Although trends in species composition are apparent in the literature, variation in the frequency of PVC pipe refugia use, and thus detection probability, have not been compared across species encountered using this method. It is also not known how behavioral interactions among treefrogs affect use of the refugia, or if these interactions vary depending on gender, size, and species of frogs. These factors also need to be addressed to identify potential biases in sampling with PVC pipe refugia. Objectives I exam ined the use of PVC pipe refugia as a method for sampling treefrogs, especially when Osteopilus septentrionalis is present, and to examine the potential influence of O. septentrionalis on native treefrog populations. Using la boratory studies, I determined the frequency of use of PVC pipe refugia when vegetative refugia were presen t for four species of treefrogs ( Hyla cinerea, H. squirella, H. femoralis, and O. septentrionalis ), and tested the null hypothesis that this frequency did not vary among species. I tested the hypothesis that O. septentrionalis affects the probability of pipe use by two native treefrog species (H. cinerea and H. squirella ). I tested the null hypotheses that O. septentrionalis does not affect the use of PVC pipe refugia of native treefrogs by behavioral exclusion and that native treefrogs do not avoid refugia previously used by O. septentrionalis. I also designed a field study to examine potential ecological effects of O. septentrionalis on native treefrogs ( H. cinerea, H. squirella, and H. femoralis ). I tested the hypothesis that if O. septentrionalis does reduce the abundance of native treefrogs, then native tr eefrogs should be more abundant at sites where O. septentrionalis was removed.

PAGE 18

18 CHAPTER 2 VARIATION IN PVC PIPE REFUGIA USE AMONG TREEFROGS Introduction The use of cavities provides m any benefits to frogs, such as predator avoidance, thermoregulation, breeding sites, and osmoregul ation (Stewart and Rand 1991, Schwarzkopf and Alford 1996, Walsh and Downie 2005). For example, in the southeastern United States, treefrogs are known to use tree cavities as diurnal shelte rs presumably to avoid desiccation (McComb and Noble 1981, Ritke and Gabb 1991, Walters and Kneite l 2004). Researchers ha ve taken advantage of this behavior to study the ecology of treefrogs by deploying artificial refugia that mimic natural cavities. As ear ly as the 1950s, Goin (1958) used aluminum cans inverted on wooden stakes to compare the non-breeding habits of two treefrog species. More recently, researchers have used polyvinyl chloride (PVC) pipes to attract treefrogs for study. The use of PVC pipe refugia is becoming commonplace as a method to monitor and sample treefrogs (Moulton et al. 1996, Bought on et al. 2000, Zacharow et al. 2001). This method is superior to call surveys for studies focused specifically on treefrogs because PVC pipe refugia can be used to sample both genders during both the breeding and non-breeding seasons. PVC pipe refugia allow for capture and identifi cation of individual frogs as well. Refugia are also better suited than drift fence arrays to the st udy of treefrogs, as tree frogs may traverse these barriers (Dodd 1991). Furthermore, re fugia passively attract (rather than trap) treefrogs (Moulton et al. 1996). As treefrogs are not restrained and can enter and l eave at will, use of this method eliminates trap mortality and allows researchers to check PVC pipe refugia infrequently. Because of this, there is no need to remove PVC pipe refugia between sampling periods, which reduces the amount of labor needed for repeat sampling or long-term monitoring.

PAGE 19

19 Variation in design and placement of PVC pipe refugia can bias sampling and may alter the probability of detection of some treefrog spec ies or size classes. Diffe rences in pipe design such as pipe diameter, length, shap e (straight or T-shaped), or if the pipe is capped at one end can alter the species composition and size cla ss of treefrogs sampled (Boughton 1997, Zacharow et al. 2003, Bartareau 2004). Location of PVC pipe refugia in terms of distance to buildings, height above ground, tree type (i.e., hardwood of softwood), arboreal or ground placement, and distance to water can also affect th e sample (McComb and Nobel 1981, Boughton 1997, Zacharow et al. 2003). Since frogs change their behavior according to te mperature, humidity, and rain, detection also va ries seasonally (Buchanan 1988, Boughton 1997, Zacharow et al. 2003). Regardless of pipe configuration or placement, some speci es are not effectively sampled with this method. Although presen t at their study site, Boughton et al. (2000) did not observe spring peepers ( Pseudacris crucifer ) and only rarely found barking treefrogs ( Hyla gratiosa ) in their PVC pipe refugia. Despite our understa nding about the effects of pipe design and placement, variation in frequency of refugia us e among species easily sampled with this method has not been examined. The propensity to use cavities may vary among tr eefrogs due to differences in their natural history and physiology. Treefrogs that remain near wetlands or have evolved morphological adaptations to prevent desiccation or predation may not frequently seek out refugia. Treefrogs vary in the amount of cutaneous lipid secretions they produce, and in their repertoire of associated body wiping behaviors that aid in water retention and thus inhibit desiccation (Barbeau and Lillywhite 2005). I examined variation in frequency of PVC pipe refugia use by four sp ecies of treefrogs in order to better understand potential variation in detection probab ility, and thus occurrence, of

PAGE 20

20 these species in the field. I conducted a laboratory study in which frogs were given the choice of resting in a PVC pipe refugium or on a plant. By using a known number of frogs in a laboratory setting, I was able to examine not only how many frogs rested within PVC pipe refugia, but also how many frogs chose to rest elsewhere. I tested the null hypothesis that treefrog species did not vary in their frequency of refugia choice. Methods Collection and maintenance of frogs I sam pled four species of treefrogs (t hree natives and one nonindigenous) commonly encountered in PVC pipe refugia studie s in central Florida: green treefrog ( Hyla cinerea ) squirrel treefrog ( H. squirella), pinewoods treefrog ( H. femoralis ), and Cuban treefrog ( Osteopilus septentrionalis ). I opportunistically collected tr eefrogs in Alachua, Hillsborough, and Polk Counties, Florida. Collection methods incl uded use of PVC pipe refugia and hand capture during breeding events, rainstorms, and at night. I caught frogs at retention ponds, natural ponds, swamps, roadside and agricultural ditches, s uburban streets, county pa rks, and on buildings. I transferred these frogs to grow th rooms at the University of Florida Gulf Coast Research and Education Center in Hillsborough County. The growth rooms were equipped with individual heating, ventilation, and air c onditioning (HVAC) systems, and had full spectrum fluorescent lights suspended 0.66 m above 1-m high benches. Th ese lights were set to mimic the natural photoperiod during the study by automatically turning on and off at 0630 h and 2030 h, respectively. Temperature in the growth rooms was set to 24C during the day and dropped to 21C during the night. Due to the limitations of the growth rooms, I was not able to raise the temperature higher to more closel y imitate the natural temperature regime. So that frogs would not be foraging in absolute darkness, a high-inte nsity desk lamp with a 40-watt bulb was set to shine into a corner of each growth room.

PAGE 21

21 I housed frogs individually in 0.75-L Sterilite containers with 3-mm ai r holes drilled in the lids. I held frogs separately during the entire le ngth of their captivity. I fed the frogs crickets, misted their containers daily, and replaced their paper towel substrate as needed. Laboratory Trials Frogs were m aintained in captivity and te sted following protocols approved by the Institutional Animal Care and Use Committee of the University of Florida (protocol number E870). Frogs were collected from 2 June 2007 un til 23 July 2007, and trials were conducted from 18 June 2007 until 24 July 2007. No frog remained in captivity longer than 18 days. After completion of each trial I marked the native frogs with Visual Implant Elastomer (VIE) and released these treefrogs at their site of initial capture. All Osteopilus septentrionalis were euthanized humanely, as this is a nonindigenous species in Florida and cannot be legally released. For each species, I conducted 30 overnight tria ls, whereby individual frogs were offered a choice between a plant and a PVC pipe refugium as a diurnal refuge. During each trial, a frog was individually placed into a 37.8-L aquarium with moist sand substrate, screen lid, 3.8 cm diameter PVC pipe that stood 20 cm tall near one end of the tank, and an approximately 20-cm tall vinca (Catharanthus roseus ) planted at the other end of th e tank. To ensure the tank would remain humid throughout the night, I saturated the substrate until standing water was observed. Trials began at 2000 h and ended the next day at 1000 h, when frogs in natural habitats would no longer be active and would have so ught diurnal shelter. At the co mpletion of each trial I recorded the location of the frog as being inside the pipe, on the plant (i.e., on leaves, stems, or at the base of the plant), or elsewhere (on the side of the tank or on the sand substrate). I measured snoutvent length (SVL) to the nearest mm and wei ghed each frog to the nearest 0.1 g. Each individual was used for only one trial, and lab conditions were the same for all trials. I used SYSTAT 10.2

PAGE 22

22 to conduct a Pearson chi-square test ( = 0.05) to test the null hypothe sis that the frequency of refugia use by treefrogs di d not vary among species. Results Frequency of refugia use vari ed s ignificantly by species ( = 21.61, df = 6, P = 0.001; Fig. 2-1); thus, I rejected the null hypothesis that these values did not differ among species. Osteopilus septentrionalis used the plant eight times as often as the PVC pipe refugia. Hyla squirella used the pipe refugia over twice as often as the plant. The frequency of pipe use did not vary considerably from that of plant use for H. femoralis and H. cinerea. Frogs of all species rarely chose to rest elsewhere within the aquari um, with three or fewer individuals per species resting away from the PVC pipe re fugia or the plant (Fig. 2-1). Discussion Refugia choice varied significantly among th e four species of treefrogs I tested. Hyla cinerea and H. femoralis were sim ilar in thei r choice of refuges and appeared to show no preference in their use of PVC pipe refugia or plants. In contrast, H. squirella and Osteopilus septentrionalis showed a preference for one refuge type over the other. Hyla squirella preferred the enclosed shelter of a PVC pipe refugium over resting on a plant, whereas O. septentrionalis rarely rested in a location other than on the plant. Frogs of all four species preferred a sheltered hiding place rather than resti ng exposed on the wall of the aquarium or the sand substrate. My results are similar to other observations of refugia use by the native species. Capture and recapture rates of Hyla squirella in artificial refugi a have been reported as higher than those of H. cinerea and H. femoralis (Goin 1958, Boughton et al. 2000, Z acharow et al. 2003). Goin (1958) opportunistically observed H. squirella using artificial retreats more often than natural retreats, and noted that retreats of H. squirella were often more enclosed than the retreats of H. cinerea. Boughton et al. (2000) captured more H. squirella than H. cinerea using PVC pipe

PAGE 23

23 refugia, but also caught very few H. femoralis. Hyla squirella was the only species of treefrog found by Walters and Kne itel (2004) to use redcockaded woodpecker ( Picoides borealis ) cavities. Similar numbers of individual H. cinerea and H. squirella were caught in PVC pipe refugia by Zacharow et al (2003), but H. squirella was more frequently recaptured. Hyla squirella occurred more often than H. cinerea in PVC pipe refugia dur ing a study by Bartareau (2004). However, McComb and Nobel (1981) found H. cinerea to use natural cavities and wooden nest boxes more frequently than H. squirella. These observations are similar to my results as H. squirella used the PVC pipe refugia more often than H. cinerea and H. femoralis but differed in that H. femoralis did not often occur in ref ugia during these field studies. Variation in the use of refugi a among treefrogs may be due to differences in their natural history and physiology. Hyla cinerea exhibits relatively high eva porative water loss, and engages in vigorous head wiping to stimul ate secretion of cutaneous lipids that limit desiccation (Barbeau and Lillywhite 2005). Frogs of this species breed in permanent wetlands, and are often found in vegetation within and around wetlands (Wright and Wright 1949, Redmer and Brandon 2005). As they typically remain near permanent water, these frogs may rely on wetlands for moisture and may be less likely to seek out refugia primarily for water re tention (Goin 1958). In contrast, Hyla squirella breeds in temporary ponds, and Goin ( 1958) hypothesized that when this frog disperses to upland habitats it seek s refugia that are more sheltered in order to prevent water loss. In addition, Barbeau and Lillywhite (2005) reported that H. squirella exhibits a higher number of body wipes before entering a water conservati on posture than othe r Florida treefrogs. Hyla femoralis also breeds in temporary wetlands and app ears to spend much of its time in tree canopies, though specific adaptations to avoid desiccation are unknown (Wright and Wright 1949, Mitchell 2005). The nonindigenous Osteopilus septentrionalis typically breeds in

PAGE 24

24 temporary wetlands and has skin co -ossified to the dermal bones in its skull, an adaptation that reduces the amount of vascular tissu e in the skin on its head and t hus the rate of water loss from this area (Duellman and Treub 1986, Meshaka 2001). Osteopilus septentrionalis is relatively large and secretes a slimy toxin when disturbe d (Meshaka 2001), so it may be less reliant on refugia for protection against predators than na tive treefrogs that l ack this protection. Researchers using PVC pipe refugia to sample treefrog popul ations should take this variation in refugia use, and t hus detection probability, into a ccount when using this method to study frog communities. A dominance of Hyla squirella in a sample may not necessarily indicate that this species is more abundant than ot her treefrogs in an area, and a lack of Osteopilus septentrionalis in a sample may not indicate absence of this species. The low detectability of Osteopilus septentrionalis obfuscates interpretation of the occurrence of this frog when PVC pipe refugia ar e used in monitoring programs. As this species does not frequently use the refugia, managers cannot be certain if O. septentrionalis is absent when it is not observed. When this species is observed, there may be several times more individuals present than sampled using this me thod. However, as this species uses PVC pipe refugia considerably less often than native treefrogs, manage rs should be concerned when it is the dominant frog found in the PVC pipe refugia (s ee Chapter 3). Researchers need to consider using mark-recapture methods rather than counts when comparing abundance across species. Management of this nonindigenous species is complicated by the strong aversion of Osteopilus septentrionalis to PVC pipe refugia. These frogs should be removed when observed; however, my results indicate that only a small portion of the frogs can be captured using this method. To attract a larger percentage of the population of O. septentrionalis researchers need to investigate the optimal placement and configur ation of PVC pipe refugia. Until a better

PAGE 25

25 management strategy is available, managers should not rely on PVC pipe refugia alone to catch these frogs. Other techniques, such as dip ne tting larvae and hand capturing breeding adults, must be used to supplement removal efforts. Otherwise, frogs that do not use the pipes will not be captured.

PAGE 26

26 OsteopilusH. squirellaH. cinereaH. femoralisNumber of trials 0 5 10 15 20 25 plant tube other Species Figure 2-1. The frequency of refugia type used by individuals of four treefrog species after 14 hours in an aquarium with a plant and a PVC pipe. Other resting sites included the side of the tank, the corner where the lid met the tank, and the sand substrate.

PAGE 27

27 CHAPTER 3 TESTING FOR INTERFERENCE COMPETITION BETWEEN Osteopilus sep tentrionalis AND NATIVE TREEFROGS AT PVC PIPE REFUGIA Introduction Com petition within and among species is an im portant ecological interaction that can limit the distribution and abundance of coexisting sp ecies (Gause 1934, Jaeg er et al. 2002). In exploitative competition, competition reduces the av ailability of a resource (Greenlees 2007). In interference competition, the competitor reduces the access to resources (Greenlees 2007). Nonindigenous species can harm native sp ecies through competition. When competing taxa co-occur naturally, competition is often redu ced through resource partitioning and character displacement (Pacala and Roughgard en 1982, Jaeger et al 2002). However, when a species is introduced into new area competition can be intense because the native fauna have not evolved mechanisms to accommodate competition by non indigenous species (Greenlees 2007). If the introduced species is a superior co mpetitor, it may cause declines, or even local extinction, in the native species. The Cuban treefrog ( Osteopilus septentrionalis ) is a nonindigenous specie s in Florida. It is known to consume native treefrogs and compet e with their tadpoles (Allen and Neil 1953, Meshaka 2001, Smith 2005), and the es tablishment of this species in an area has been associated with a decline in native treefrogs (Meshaka 2001). This specie s was first reported on Key West in 1931 and is now established throughout most of peninsular Florida (Barbour 1931, Meshaka 2001, Krysko et al. 2005). As O. septentrionalis continues to spread, an understanding of how it impacts native communities, especially treefrogs, is essential to the eventual development of management strategies for this nonindigenous frog. The use of PVC pipe refugia is becoming a common method of sampling treefrogs (Moulton et al. 1996, Boughton 1997, Hirai 2006). Additionally, this method has the potential to

PAGE 28

28 assist in the control of Osteopilus septentrionalis in both developed and natural settings. PVC pipe refugia attract treefrogs th at can then be captured for indi vidual identification and marking (Moulton et al. 1996), or removed and euthanized, as in the case of O. septentrionalis As this technique becomes more widespr ead it is increasingly important to test the underlying basic assumptions of this method. The assumpti on that PVC pipe refugia occupancy by O. septentrionalis does not interfere with occupancy by native species through interference competition, is a potential source of bias when sampling native treefrogs and has not yet been evaluated. Territorial behavior exhibited by a nonindigenous species may result in competition with native species for refuge sites when this reso urce is limiting. Although most reported territorial aggression in anurans is related to courts hip (Duellman and True b 1986, Lutz 1960), several treefrogs have been reported to exclude other fr ogs from refugia. Both male and female coquis ( Eleutherodactylus coqui) defend diurnal retreat sites, using aggressive calls and biting the heads of intruders (Stewart and Rand 1991). Male Mexican leaf-frogs (Pachymedusa dacnicolor ) use prepollical spines on their wrists to defend burrows from conspecifics (Wiewandt 1971). Resident male squirrel treefrogs ( Hyla squirella) are known to force inva ders out of short PVC pipe refugia Buchanan (1988). Osteopilus septentrionalis is often found in aggregations (Meshaka 1996), and territoriality has not yet been reported in this species. In addition to the potential for behavioral interactions, there is also the potential for chemically mediated effects between nonindige nous and native species. When threatened, Osteopilus septentrionalis produces a milky toxin with a di stinct odor (Meshaka 2001). This toxin irritates mammalian mucous membranes (Mes haka 2001) and may also affect amphibians. Tadpoles and adult frogs often use chemical cues to detect predators (Stauffer and Semlitsch

PAGE 29

29 1993, Kiesecker et al. 1996, Chivers, et al. 1999), and it has yet to be determined if secretion residue left by O. septentrionalis may be detected and avoided by native treefrogs. As detecting chemical cues is important in amphibian ecol ogy (Stauffer and Semlitsch 1993, Kiesecker, et al 1996, Chivers, et al. 1999), there is the potential that if native treefrogs do avoid this residue, then O. septentrionalis resting on a substrate might chemically interfere with the frogs that later use that area. I conducted two laboratory experiments to examine possible interference competition between Osteopilus septentrionalis and native treefrogs for refugia sites. I tested the null hypothesis that O. septentrionalis does not affect the detection pr obability of native treefrogs in PVC pipe refugia via behavioral exclusion a nd that native treefr ogs do not avoid pipes previously used by O. septentrionalis (chemical exclusion). Methods Behavioral Exclusion To determ ine if Osteopilus septentrionalis behaviorally excludes native treefrogs from PVC pipe refugia, I conducted 30 trials fo r each of two native species: green treefrog ( Hyla cinerea ) and squirrel treefrog ( H. squirella). I paired a native frog with an O. septentrionalis in an aquarium with a single PVC pipe refugium; as a control, I paired frogs with conspecifics. To prevent predation by O. septentrionalis I paired frogs of similar size such that frogs did not differ more than 20% in snout-vent length (SVL ). At 2000 h, I placed each pair into a 37.85-L aquarium with a screen lid, moist sand substrat e, and a 20-cm tall and 3.81-cm diameter PVC pipe refugia planted vertically in the sand substrate in the center of the enclosure. At 1000 h the following morning, I recorded the location of each frog as either in or out of the PVC pipe refugium. Laboratory conditions, described below, were held constant and did not vary from those of the growth room in which the frogs were housed.

PAGE 30

30 Chemical Avoidance To determ ine if native treefrogs avoid PVC pipe refugia that have recently been used by Osteopilus septentrionalis I tested treefrogs by placing them in an aquarium with one pipe that had been used recently by O. septentrionalis and one new, clean pipe. In control trials, I placed each treefrog in an aquarium with two clean pipe s. To create the treatment pipes, I housed 30 large (> 45 mm SVL) O. septentrionalis in a 113-L plastic enclosure with a moist sand substrate, small dish of water, scre en lid, potted vinca plant ( Catharanthus roseus ), and 20 PVC pipe refugia. On nights before a tria l, I replaced the refugi a and the plant in this container with 15 new, clean refugia so that the O. septentrionalis would rest in these pi pes. At 2000 h the next night, I removed O. septentrionalis from these now used pipes, and placed these pipes opposite new, clean pipes in 37-L aqua ria with sand substrate and sc reen lids. One frog was then introduced to each aquarium. At 1000 h the locati on of the frog was recorded. To control for possible directional bias, the side of the enclos ure on which each treatment pipe was placed was alternated among tanks. To discourage smaller frogs from resting between the lid and lip of the aquarium, I filled this space with a rope that wa s secured with non-toxic glue. A high-intensity desk lamp with a 40-watt bulb was placed in the growth room where the tanks were housed to provide light so that the frogs would not be in complete darkness at night. Laboratory conditions, described below, were held cons tant and did not vary from those of the growth room in which the frogs were housed. Collection and Maintenance of Captive Frogs I opportunistically collected treefrogs by use of PVC pipe refugia and by hand. Collection occurred prim arily at night, particularly dur ing breeding events and rainstorms. Frogs were collected from 10 May 2007 through 26 A ugust 2007 in Alachua, Hernando, Hillsborough, Manatee, Orange, Pinellas, Polk, Sarasota, and Seminole counties, Florida. I caught frogs at

PAGE 31

31 retention ponds, along the edges of natural wetlands, in roadside and agricultural ditches, on suburban streets, in forested county parks, on walls of buildings, and in parking lots. I transferred these frogs to growth rooms at the University of Floridas Gulf Coast Research and Education Center in Hillsborough County. I maintained frogs in captivity in accordance with University of Florida Institutional Animal Care and Use Committee (protocol number E870). Growth rooms were equipped with individual heati ng, ventilation, and air conditioning (HVAC) systems set to 24C. The ro oms also had full-spectr um fluorescent lights set to automatically turn on at 0630 h and off at 2030 h to mimic the natural photoperiod. A highintensity desk lamp with a 40-watt bulb shining into a corner provided lig ht so that the frogs would not be in complete darkness at night. I housed frogs individually in 0.75-L Sterilite containers with 3-mm air holes drilled in the lid s during the entire length of their captivity. I fed them crickets, misted their containers daily, a nd replaced their paper towe l substrate as needed. I conducted experiments from 15 May 2007 through 31 August 2007. After completion of each trial I marked native treefrogs with Visual Implant Elastomer (VIE) and released them at their initial capture site. All Osteopilus septentrionalis were humanely euthanized, as this is a nonindigenous species in Florida. Statistical Analysis To test null the hypothesis that Osteopilus septentrionalis does not behaviorally exclude native frogs from PVC pipe refugia, I compared th e probability of refuge use by native frogs in the presence of O. septentrionalis with the probability of ref uge use when paired with a conspecific. Probabilities were calculated using a maximum likelihood model base on frequency of refugia use by pairs of frogs. To calculate pr obabilities when two native frogs were paired, each was randomly assigned post hoc identification as frog A or frog B. Four combinations of outcomes were possible: (1) both frogs were in th e pipe (AB), (2) only native frog A was in the

PAGE 32

32 pipe (A0), (3) only Osteopilus septentrionalis or the native frog B was in the pipe (0B), and (4) neither frog was in the pipe (00). The proba bilities of these outcomes are as follows: PAB = PA PB PA0 = PA (1PB) P0B = (1PA) PB P00 = (1PA) (1PB) I used Program R, version 2.5.1, to calculat e the maximum likelihood estimates of the probability that native treefrog (frog A when pa ired with a conspecific) was in the pipe (PA), and that the O. septentrionalis or frog B was in the pipe (PB). I used the program to calculate the likelihood ratio -2 (log Lreduced model Lfull model) for each species to compare the probability of the native frog being in the pipe when paired with an O. septentrionalis to the probability of native frog A being in the pipe when paired with a conspecific (native frog B). I used the likelihood ratios in a chi-square analysis to test the null hypotheses that the native treefrogs are equally likely to occupy the pipe in the presence of O. septentrionalis as with a conspecific. To analyze the outcome of my chemical avoi dance experiment, I tested the null hypothesis that there was no difference in occupancy of clea n PVC pipe refugia and those that had recently housed Osteopilus septentrionalis. To do this I used Microsoft Ex cel to conduct a chi-square test for each species to determine if the frogs were more likely to use one type of pipe (used by O. septentrionalis vs. clean) over the other. I used the results of the control trials as expected values for this analysis. Results Behavioral Exclusion The probability of Hyla squirella occupying PVC pipe refugia in the presence of Osteopilus septentrionalis was not significantly different from the probability of H. squirella occupying the refugia in the presence of a conspecific (LR = 1.69, P =0.193). These probabilities

PAGE 33

33 were also not significant for H. cinerea (LR = 0.130, P = 0.718; Table 3-1). Therefore, I failed to reject the null hypothesis that na tive treefrogs were equally as likely to rest in a PVC pipe refugium with an O. septentrionalis as with a conspecific. Chemical Avoidance Native treefrogs did not dem onstrate significant avoidance of PVC pipe refugia that had been recently used by Osteopilus septentrionalis ( =0.86, df = 2, P = 0.65 for Hyla squirella, and =1.54, df = 2, P = 0.46 for H. cinerea ; Table 3-2). Although O. septentrionalis used to treat the pipes were often found sharing refugi a in their holding enclosure, with 30 frogs congregated in approximately half of the 15 available clean pipes, O. septentrionalis in experimental trials rested in used PVC pipe ref ugia significantly less frequently than in control pipes or other locations ( on glass or sand substrate; =7.11, df = 2, P = 0.03). Therefore, I failed to reject the null hypothesis that native treefrogs were equally likel y to use a PVC pipe refugium that had recently been used by O. septentrionalis as new, clean refugia, but did reject the null than O. septentrionalis would use these pipes equally. Discussion Neither exp eriment indicated that Osteopilus septentrionalis interferes with native treefrog refugia use and detection during studies using PV C pipe refugia. The frequency of pipe use by native treefrogs was not a ltered by the presence of O. septentrionalis or by any residue that might have been left on pipes recently used by this nonindigenous frog. However, O. septentrionalis avoided refugia that had recently been us ed by its batracophagic conspecifics. These results support the underlying assumption that the presence of O. septentrionalis does not reduce the detection probability of native treefr ogs, and thus competition for refugia is unlikely. However, the results indicate that as it a voids refugia recently used by conspecifics, O. septentrionalis may be underrepresented in samples usi ng this method. With this assumption in

PAGE 34

34 place, it is possible to use PVC pipe refugia to compare native treefrog populations prior to the invasion of O. septentrionalis to populations of treefrogs sampled after this nonindigenous species has become established. However, resear chers should keep in mind the possibility that O. septentrionalis may be underrepresented. My results are inconsistent with th ose of Buchanan (1988), who found that Hyla squirella was territorial and excluded conspecific invaders from the PVC pipe refugia in the lab. I often observed H. squirella sharing refugia with conspecifics or Osteopilus septentrionalis although this inconsistency may be due to the fact that I did not allow one frog to become established in the aquarium before introducing a second fr og. Stewart and Rand (1991) found that while intruder Eleutherodactylus coqui were often evicted from refugi a, resident frogs were rarely displaced from their refugia by invaders. Thus, it may be important for a frog to establish claim over a refugium before defending it, which might explain the difference between my results and those of Buchanan (1988). Still, H. squirella has been observed in the wild sharing a PVC pipe refugium with up to five adult conspecifics (Johnson, unpubl. data). Future research studies should be conducted in which the native frog is allo wed to reside in the aquarium for sufficient time to establish residency pr ior to the introduction of O. septentrionalis, and in which frogs in the field are examined fo r territorial behavior. The avoidance of PVC pipe refugia that had recently been used by Osteopilus septentrionalis by conspecifics and not by native treef rogs suggests that these frogs may be avoiding a potential predator that natives do not recognize. The ability to recognize and respond appropriately to predators is extremely important to the fitness of an indivi dual as those that fail to do so are likely to be eaten (Chivers and Smith 1994, Kiesecker and Bl austein 1997, Griffin et al. 2000). Although native frogs readily respond to native predators, they may not recognize and

PAGE 35

35 respond to nonindigenous predators (Kiesecker a nd Blaustein 1997, Pearl et al. 2003). I did not investigate the possibility that na tive treefrogs attracted to the pi pes might be more likely to be preyed upon by Osteopilus septentrionalis thus making the artificial refugia used in research efforts into population sinks. If O. septentrionalis residing inside a PVC pipe refugium consume native treefrogs attracted to th at refugium, native populations might be underrepresented in surveys using this sampling technique. I recommend this topic for further research.

PAGE 36

36 Table 3-1. Number of each species of native treefrog ( Hyla squirella and H. cinerea) that chose to rest in the PVC pipe ref uge when paired with either Osteopilus septentrionalis or a conspecific Both (AB) Native only (A0) O. septentrionalis or conspecific only (0B) Neither (00) H. squirella with O. septentrionalis 6 13 4 7 H. squirella with conspecific 12 2 9 7 H. cinerea with O. septentrionalis 3 7 12 18 H. cinerea with conspecific 5 3 5 17 Table 3-2. Total number of treefrogs resting in e ach location after 14 hours in an enclosure with a choice of two PVC pipe refugia during experimental trials* Experimental Control Neither O. septentrionalis 3 11 16 H. squirella 11 9 10 H. cinerea 5 9 16 *Experimental pipes had several Osteopilus septentrionalis resting in them during the day of the trial and were placed in the aquaria just before th e start of the trial. Cont rol pipes are new pipes, and Neither indicates the frog was not in a pipe.

PAGE 37

37 CHAPTER 4 REMOVAL OF Osteopilus septentr ionalis FROM FLATWOODS HABITAT IN CENTRAL FLORIDA Introduction Som e nonindigenous species may directly affect communities of native species by preying on or competing with native species, or indire ctly by introducing new dise ases, altering nutrient cycles, fire regimes, and other aspects of the e nvironment that then create habitat less suitable for native species (Simberloff 2000). Wilcove et al. (1998) ranked competition and predation by nonindigenous species as the second greatest threat to imperiled pl ants and animals in the United States, surpassed only by habita t loss caused by humans. The potential for the noni ndigenous Cuban treefrog ( Osteopilus septentrionalis ) to negatively affect native treefrogs has been evident since its discovery on Key West, Florida (Barbour 1931). In a letter to Barbour (1931), A. G. Elbon reported that this frog was common on the island, had been there si nce before he moved to the island, and that it fed almost exclusively on smaller frogs. Twenty y ears later, Schwartz (1951) documented O. septentrionalis on mainland Florida and the rapid development of its eggs. Shortly after, Allen and Neill (1952) observed O. septentrionalis feeding on breeding frogs, and warned that the presence of this frog might affect native frogs. Lee ( 1969) repeated this warning, a lthough he cautioned that there were no data to substantiate this potential threat at that time. Recently, studies have begun to confirm the adverse effects of Osteopilus septentrionalis on native frogs. Osteopilus septentrionalis tadpoles have been found to consume and reduce the growth rate of native tadpoles (Smith 2005a, Smith 2005b). Although O. septentrionalis has been observed consuming native frogs since the time of its introduction, adults appear to prey mainly upon insects and less frequently on other frogs (Lee 1969, Wyatt and Forys 2004, Meshaka 2001). These and other stresses may a dd up to a population-le vel effect on native

PAGE 38

38 treefrogs. Before we can consider costly manageme nt plans to conserve native frogs, we need to know if O. septentrionalis truly does cause declines in native frogs. It is also necessary to examine possible management strategies, such as the removal of O. septentrionalis using PVC pipe refugia. I conducted a large-scale field st udy to determine if native tree frogs respond to the removal of Osteopilus septentrionalis I tested the hypothesis that if O. septentrionalis reduces the abundance of native treefrogs, then native treefr ogs should be more a bundant at sites where O. septentrionalis was removed. I did not seek to identify the mechanism by which O. septentrionalis could be affecting native treefrogs. Methods Sampling At Wekiwa Springs State Park and Flatwoods Park in Central Florida, I selected three study blocks based on accessibility, similarity of habitat, and proximity to development (and thus probability of containing an established O. septentrionalis population). In each block I constructed a pair of 50by 50-m grids of 1m tall, 3.81-cm diameter PVC pipes installed vertically at 5-m intervals. Every three w eeks, I checked each grid once daily for three consecutive days. I monitored each grid over two sampling periods to estimate initial treefrog abundance. I then randomly selected one experiment al grid of each pair at which I removed all O. septentrionalis I encountered. Frogs were not removed fr om the other grids, which were used as controls. During each sampling period, I removed frogs fr om PVC pipe refugia by gently coaxing them into a plastic bag using a plunger construc ted from a sponge cut to the inside diameter of the PVC pipe and attached to the end of a dow el. I identified each frog to species, measured snout-vent length (SVL) to the nearest mm, and wei ghed to the nearest 0.1 g. I also marked each

PAGE 39

39 frog for future identification us ing a unique combination of toeclipping and the injection of visual implant elastomer (VIE) in up to tw o toes. I released all native frogs and Osteopilus septentrionalis caught within control grids, and euthanized all O. septentrionalis captured within the experimental grid after the initial two sampling periods to es timate initial population sizes. Wekiwa Springs State Park Wekiwa Springs State P ark is located in Apopka, Florida, less than 20 km north of Orlando. The park is managed by the Florida Pa rk Service and consists of 2,800 ha surrounded by urban development to the south, Rock Springs Run State Preserve to the northeast, and urban development and agriculture to the west (Fig. 4-1). During June 2006, I established my original grid s at the park. I installed Grids A and B in the southeastern corner of th e park in a small mesic flatw oods opening surrounded by sandhill to the south, hydric hammock to the north and we st, and an urban/wildland interface to the east (Fig. 4-1). Grids C, D, E, and F were in a large tr act of mesic flatwoods within the interior of the park. I checked these grids for treefrogs during 19-21 July 2006. Osteopilus septentrionalis were captured in Grids A and B, and these grids rema ined in their original position throughout the study. However, during this first sa mpling period it became apparent that the grids located in the interior of the park (C, D, E, and F) did not contain O. septentrionalis in detectable numbers, if at all. Therefore, I removed these four grids on 8 August 2006, and established new Grids G, H, I, and J in mesic flatwoods in the southwestern corner of the park. These new grids were surrounded on two sides by an urban/wildland interface, and seemed likely to contain O. septentrionalis Indeed, during the next surv ey period, all grids contained O. septentrionalis Removal of O. septentrionalis from experimental grids (A, H, and J) began on 12 October 2006, after the completion of two survey periods.

PAGE 40

40 On 18 November 2006, I discovered mowed fi rebreaks approximately 15 m wide cut through Grids G and I, both of which were control grids. This disturbanc e had the potential to alter the distribution of the frogs within each grid and introduce edge effects. Therefore, I removed these grids on 3 January 2007. To replac e the grids destroyed by mowing, I established new Grids K and L in scrubby flatwoods just east of the previous grids and new Grids M and N between Grids H and I. After sampling these ne w grids for two survey periods, and observing less than one frog per grid, I decided I would be unable to obtain adequate sample sizes to conduct this study. On 10 May 2007 all grids were removed from Wekiwa Springs State Park. Flatwoods Park Flatwoods Park, located northwest of Tamp a, Florida, is part of the Hillsborough Wilderness Park System, which contains over 2,200 ha (Fig. 4-2), and is managed by the Southwest Florida Water Management District as the Morris Bridge Wellfield. The park is a popular cycling location, and an 11.2-km paved bicycle loop encircles the cen ter of the park. The park consists predominantly of flatwoods habitat, and is surrounded by ongoing urban development to the north and other parks in the Hillsborough Wilderness Park System to the east and south. On 11 July 2006, I installed six grids at Flat woods Park (Fig. 4-2), all of which were located north of the bicycle loop in mesic flatwoods habitat. Grid s A and B were located in the western side of the park, Grids C and D were central, and Grids E and F were near the eastern corner of the park. The urban/wildland interface wa s visible from the major ity of the six grids. I began monitoring the grids on 8 August 2006 and began to remove Osteopilus septentrionalis from experimental Grids A, D, and E on 30 Se ptember 2006. Monitoring and removal continued through 29 June 2007, at which time the study was concluded and the grids were removed.

PAGE 41

41 Statistical Analysis I used SYSTAT, version 10.2, to perform a paired t-test to test if the total number of Osteopilus septentrionalis varied by grid type, and thus determine if a significant number of these invasive frogs were removed from experimental grids. I also visually inspected the number of O. septentrionalis captured at each grid over time to de termine if fewer frogs of this species were observed in experimental grids throughout th e study than in control grids. Using program MARK, version 4.3, I conducted a robus t closed capture model with constant emigration rates to estimate abundance of native treefrogs in each grid over time. I plotted the estimated abundances of each grid, and visually inspected estimates for patterns. I expected that if removal of O. septentrionalis caused native treefrog abundance to increase, then estimates of abundance of native treefrogs in experimental grids would in crease after removals be gan, while the abundance in control grids would drop due to continued effects of O. septentrionalis. As another indication of fitness of native treefrogs with in each grid, I used SYSTAT to perform a Pearsons chi-square test to test the null hypo thesis that equal numbers of gravid female H. femoralis were captured in control and experimental grids from 7 Ma y 2007 through 8 August 2007. Finally, as a final attempt to determine if O. septentrionalis affected native treefrog a bundance, I used a Pearsons Correlation Coefficient to determine if the number of O. septentrionalis and native treefrogs in each grid were negatively correlated. Results Flatwoods Park Weather conditions were dry duri ng the year leading up to the e xperim ent (Fig 4-3; Florida Automated Weather Network 2007), and capture rates of frogs at all sites throughout the experiment were lower than anticipated. Hyla femoralis occurred regularly in PVC pipe refugia and was observed on a total of 1,517 occasions. I captured fewer Osteopilus septentrionalis

PAGE 42

42 which I found a total of 532 times. H. squirella was only observed 14 times, and H. cinerea only once. Therefore, my analysis focused on H. femoralis and O. septentrionalis. The total number of individual Osteopilus septentrionalis observed in control grids did not differ significantly from that of the experimental grids ( t = -0.77, df =2, P = 0.52). Due to low capture rates of this species, I was unable to estimate their abundance using MARK. I plotted and inspected counts of O. septentrionalis, which did not vary visibl y between experimental and control grids, and it appears that my removal had no effect on the abundance of this species. However, these counts did vary seasonally, with more frogs being observed in summer and early fall than in winter and spring (Figures 4-4, 4-5, 4-6). Based on visual observation of abundance estimates of Hyla femoralis over time, I did not find compelling trends of increased abundance in experimental grids and decreased abundance in control grids. It does no t appear the removals affected th e abundance of this native species. Abundance of this frog in experimental Grid A wa s often, though not consiste ntly, higher than in its partner control grid (Figur e 4-7). On occasions when the estimated abundance of Grid A differed most from that of Grid B, st andard errors were high. Populations of H. femoralis in Grids C and D appeared to be small, and on many days no frogs of this species were observed. The low capture rates in these grids created highl y variable estimates with high standard errors (Figure 4-8). Only in Grid E did abundance of H. femoralis appear to increase with the removal of O. septentrionalis (Figure 4-9). Other indicators of f itness did not vary by grid type. The ratio of gravid females observed during the 2007 bree ding season did not differ significantly according to grid type ( = 12, df = 10, P = 0.285). All H. femoralis observed were adults at first capture, and did not experience sufficient growth (snout-vent length and weight) to enable me to rule out error as a source of variation in measurements. However, there were less H. femoralis in

PAGE 43

43 grids with higher O. septentrionalis abundances (Figure 4-10). A negative correlation was found between total numbe r of individual O. septentrionalis and H. femoralis (log transformed data; Pearson correlation = -0.820; Figure 4-11). Wekiwa Springs State Park At W ekiwa Springs State Park, removals of Osteopilus septentrionalis were only conducted during two sessions at the western grids, and the eastern grids experienced extremely low recapture rates (Appendix A). Thus my analysis focused on results from the Flatwoods Park site. Discussion These results suggest that Osteopilus septen trionalis was not removed in sufficient numbers to cause a decline in abunda nce and thereby exert an effect on Hyla femoralis The number of observations of O. septentrionalis did not differ significantly between control and experimental grids, so it is unclear if the abundance of H. femoralis would have increased had I been able to remove more of this nonindige nous species. More research will need to be conducted to determine if removal of O. septentrionalis is an appropriate strategy to manage native treefrogs. The low capture rates experien ced at the Flatwoods Park si te during the study period may have been caused by a combination of factors. First, although 2006 star ted off fairly wet, no precipitation was recorded in March at the near by Hillsborough River State Park, located within 1 km from Flatwoods Park, and th e months of April and May were also somewhat dry (P. Potts, Hillsborough River State Park, pers. comm.). In addi tion, Flatwoods Park is also the site of the Morris Bridge Wellfie ld, an active, long-standing site of groundwater pumping by Tampa Bay Water. This dry spell, combined with the potential effects of gr oundwater pumping, likely affected treefrog communities at Flatwoods Pa rk. Secondly, numbers of frogs at the grid

PAGE 44

44 locations may have been extremely low. Squirrel treefrogs ( Hyla squirella ) chose PVC pipes over other refugia more often than other treefrogs and thus are most likely to be observed (see Chapter 2); however few individuals of this specie s were observed in the grids. Green treefrogs ( H. cinerea ) are associated with wetlands (Goin 1958) and although I have often observed this species a distance from permanent water, it might not have occupied the grid areas in detectable numbers. Third, frogs living at th e grid sites might simply not be using the PVC pipe refugia. Recaptured frogs were often encountered in or near refugia of initia l capture after avoiding detection for multiple days or sampling sessions. Th is suggests that either they returned to the same area after absence from the grid or, perh aps more likely, that they made use of other available refugia within the gri d. Other researchers have observed higher capture rates in PVC pipe refugia positioned > 2 m above the ground (Boughton et al 2000), and it is possible that frogs within the study area utili zed taller structures as refugi a and did not encounter the PVC pipes. Finally, native frog populations in the grids may have already experienced substantial reductions subsequent to the establishment of Osteopilus septentrionalis within the park. Additionally, this nonindigenous spec ies seems to prefer natural re fugia over PVC pipe refugia, and may have been underrepresented during sampling for this study (see Chapter 2). Seasonal changes are known to occur in the activity of Osteopilus septentrionalis (Meshaka 2001) and in the use of PVC refugia by some native treefrogs (Boughton 1997, Zacharow et al. 2003); thus, it is not surpri sing that the use of PVC pipe refugia by O. septentrionalis varied seasonally during the study period. The observed behavioral variation of O. septentrionalis in conjunction with low cap ture rates served to c onfound potential effects of removal, as seasonal variation in PVC pipe ref ugia use by this species was asynchronous to its removal. Captures of this species peaked near the beginning of the study, before and during the

PAGE 45

45 beginning of removals. Shortly after, this specie s became scarce in both experimental and control grids, and thus removal was infrequent. Despite the lack of signifi cant observable effects of Osteopilus septentrionalis removal, it is noteworthy that Hyla femoralis was observed less frequently at grids where this species was observed more frequently. While this correlation is not necessarily indi cative of a cause and effect relationship, it lends credib ility to anecdotal evidence that O. septentrionalis may be replacing native treefrogs in areas where it has b ecome established. However, further research is needed to assess the potential effects of this invasive frog to actually cause declines of native treefrogs.

PAGE 46

46 Figure 4-1. Location of each grid at Wekiwa Springs State Park (outlined in red). Above is an image of the area, and maps 1-3 below detail s the location of each grid relative to the other grids.

PAGE 47

47 Figure 4-2. Location of each grid at Flatwoods Park (outlined in red). Above is an image of the area, and maps 1-3 below details the location of each grid relative to the other grids.

PAGE 48

48 Date J an-05 Fe b -05 Mar-0 5 Apr-05 M a y0 5 Jun-05 Jul-05 Aug-05 Sep-05 Oct-05 Nov-05 D e c-05 Jan-06 F eb -06 Ma r -0 6 Ap r0 6 M a y0 6 Jun-0 6 Jul-06 Aug-06 Sep-06 Oct 0 6 Nov-0 6 D e c-06 Jan-0 7 Feb -07 Ma r0 7 A p r-07 M a y0 7 Jun-0 7 J ul-0 7Precipitation 0 2 4 6 8 10 Figure 4-3. Total monthly precipitation at th e Dover weather stati on in Hillsborough County (adapted from Florida Automated Weathe r Network, 2007). This weather station is located near Flatwoods Park.

PAGE 49

49 20-Aug-06 9-Sep-06 3 0-Se p-0 6 21-Oc t -0 6 10-Nov-06 4D ec-06 1 J a n 0 7 21-J a n-0 7 2M ar-07 2 3 -Ma r0 7 14-Apr-07 16-May0 7Total O. septentrionalis per session 0 5 10 15 20 25 Grid A (removal) Grid B (control) Date Figure 4-4. Seasonal tre nd in total individual Osteopilus septentrionalis observed per sampling session at Grids A and B at Flatwoods Park.

PAGE 50

50 20-Aug-06 9-Se p -06 30 -Se p-0 6 2 1 -O c t-06 10 -Nov -0 6 4 -D ec-0 6 1-Ja n -07 21-Jan-07 2 -M ar-07 23-Mar-07 14-Apr-07 16-May-07Total O. septentrionalis per session 0 5 10 15 20 25 30 Grid C (control) Grid D (removal) Date Figure 4-5. Seasonal tre nd in total individual Osteopilus septentrionalis observed per sampling session at Grids C and D at Flatwoods Park.

PAGE 51

51 20 -Au g-0 6 9-Sep-06 30-Sep-0 6 21-Oct-06 10-Nov-06 4 -Dec -0 6 1 -J a n-0 7 2 1 -J a n-0 7 2 -M ar -07 2 3 -Ma r-0 7 1 4 -Apr -0 7 1 6 -Ma y-0 7Total O. septentrionalis per session 0 2 4 6 8 10 12 14 Grid E (removal) Grid F (control) Date Figure 4-6. Seasonal tre nd in total individual Osteopilus septentrionalis observed per sampling session at Grids E and F at Flatwoods Park.

PAGE 52

52 Figure 4-7. Estimated abundance of Hyla femoralis ( SE) at Grids A and B at Flatwoods Park. Osteopilus septentrionalis were removed from Grid A starting on 9 September 2006.

PAGE 53

53 Figure 4-8. Estimated abundance of Hyla femoralis ( SE) at Grids C and D at Flatwoods Park. Osteopilus septentrionalis were removed from Grid D starting on 9 September 2006. Missing values were unrealistically estimat ed as over 500 individuals, due to low capture and recapture rates, and are not shown on this figure.

PAGE 54

54 Figure 4-9. Estimated abundance of Hyla femoralis ( SE) at Grids E and F at Flatwoods Park. Osteopilus septentrionalis were removed from Grid E starting on 9 September 2006.

PAGE 55

55 Grid Grid EGrid BGrid AGrid FGrid DGrid CNumber of individuals 0 20 40 60 80 100 H. femoralis O. septentrionalis Figure 4-10. Total individual Hyla femoralis and Osteopilus septentrionalis per grid over the course of the study. Grids are a rranged in orde r of increasing O. septentrionalis

PAGE 56

56 L O G H Y F E LOGHYFEL O G O S S E LOGOSSE Figure 4-11. Scatterplot ma trix of the log of the total number of individual Hyla femoralis (LOGHYFE) and Osteopilus septentrionalis (LOGOSSE). The Pearson correlation coefficient of these two logs was -0.820.

PAGE 57

57 CHAPTER 5 CONCLUSIONS Summary For decades it has b een speculated that predation and competition by the nonindigenous Cuban treefrog ( Osteopilus septentrionalis ) have caused declines in native treefrogs in Florida (Allen and Neil 1953, Lee 1969). With the use of PVC pipe refugia, treefrog populations can now be studied outside of the breeding season an d these concerns can be addressed. However, the use of PVC pipe refugia is a technique that is still developing, with new biases identified and new designs reported in recent years (Zacharow et al. 2003, Johnson 2005). If we are to rely on this method for the study of treefrogs then we must be aware of associated biases and assumptions. I addressed two factor s that contribute to the assu mption that treefrogs sampled by PVC pipe refugia accurately represent the populati on. With these assumptions in mind, I tried to examine the effects of O. septentrionalis on native treefrogs. The results from the first laboratory study (Cha pter 2) suggest that treefrog species vary considerably in their frequency of occupying PV C pipe refugia. While the squirrel treefrog ( Hyla squirella ) often chose PVC pipe refugia over a plant for cover, Osteopilus septentrionalis did so rarely. The pinewoods treefrog ( H. femoralis ) and the green treefrog ( H. cinerea) showed no preference for the PVC pipe refugium or the plant. In some field studies, H. squirella was the most common frog observed (Boughton et al. 2000, Zacharow et al. 2003, Borg et al. 2004). This could be because either there are more H. squirella present in that area or due to its affinity for PVC pipe refugia. The low frequency of pipe use by O. septentrionalis is alarming as it is the only frog found in pipes in some areas (pers. obs.) My results suggest that PVC pipe refugia are poor indicators of the abundance of this frog compar ed to other species, and that there may be many more in an area than observed using th is sampling technique. Researchers using PVC

PAGE 58

58 refugia should be aware of this sampling bias when interpreting their results, especially when individuals are not marked. The hypothesis that Osteopilus septentrionalis does not interfere with the detection of native treefrogs using refugia was supported by re sults of my laboratory experiments (Chapter 3). I did not find evidence of behavioral or chemical exclusion of native treefrogs by O. septentrionalis This implies that native treefrogs sympatric with O. septentrionalis can be compared to those where O. septentrionalis is absent; populations befo re and after establishment of this nonindigenous species can be compare d, as can populations before and after their removal. However, although I determined that O. septentrionalis does not behaviorally exclude native treefrogs when they reach the PVC pipe ref ugia simultaneously, I did not test to see if they displayed territorial behavior as residents when an invader frog attempts to use the refugia. Also, I did not determine if native frogs that use the PVC pipe refugia in the field are more likely to be preyed upon than those occurring away from artificial refugia. It is possible that the population may be underestimated if frogs in PVC pipe ref ugia are consumed before they can be counted. While native treefrogs do not avoid PVC pipe refugia that had recently been used by O. septentrionalis frogs of this nonindigenous species did a void refugia that had been used by their conspecifics. These frogs may be avoiding th eir batracophagic conspecifics, whereas native treefrogs may be unaware of this potential predat or, and may be more susceptible to predation. These are two more questions that need to be addressed to be sure that O. septentrionalis does not interfere with the detection of native frogs. Capture rates were low during the field study and an insufficient number of Osteopilus septentrionalis was caught to examine the effects of its removal on native treefrogs. However, the results of this study did reve al a negative correla tion between the frequency of observations

PAGE 59

59 of the nonindigenous frog and the native Hyla femoralis This finding relates back to the original question of the severity of predation and competition by O. septentrionalis on the natives, and suggests that this nonindigenous frog might cause declines in native tr eefrog populations. Further research is needed to examine other possible causes of this correlation and to determine the effects of removing this frog, and to determin e if removal using PVC pipe is a suitable management strategy for O. septentrionalis The detection of native frogs in PVC pipe re fugia does not appear to be affected by the presence of Osteopilus septentrionalis but populations of native treefrogs do appear to be negatively affected. These findi ngs support the anecdotes that O. septentrionalis might be replacing native treefrogs in central Florida. Recommendations As Osteopilus septentrionalis seems to negatively affect popul ations of native treefrogs, I recomm end that these nonindigenous frogs be removed when observed. This species appears abundant in areas where they are established, and management of this species may require great effort. As I had low capture rates of this frog in fl atwoods habitat, I reco mmend that PVC pipe refugia be placed where these frogs are more lik ely to occur. These areas should include near wetland and around buildings, but fu rther research should be conduc ted to determine where PVC pipe refugia would be most effectiv e in attracting this nonindigenous frog. Other methods, such as road cruising during ra instorms or hand capture at breeding sites, should supplement removals using PVC pipe refugia. As I found Osteopilus septentrionalis to use PVC pipe refugia less frequently than na tive treefrogs, these a dditional methods may be necessary to remove individuals th at do not use PVC pipe refugia.

PAGE 60

60 Left unmanaged, Osteopilus septentrionalis will continue to sp read. As it invades new areas it may cause native treefrogs to declin e throughout Florida and other areas in the southeastern United States.

PAGE 61

61 APPENDIX TOTAL CAPTURES AT WEKI WA SPRINGS STATE PARK Table A-1. Total individual frogs captured in Wekiwa Grid A during each sam pling session. Date H. cinerea H. femoralis H. squirella O. septentrionalis 21-Jul 5 0 0 18 11-Aug 8 1 0 9 4-Sep 11 0 0 5 25-Sep 7 2 0 6 14-Oct 7 0 0 2 6-Nov 2 0 0 1 20-Nov 6 1 0 1 6-Jan 4 5 0 1 28-Jan 15 4 0 5 18-Feb 2 4 0 2 Table A-2. Total individual frogs captured in Wekiwa Grid B during each sampling session. Date H. cinerea H. femoralis H. squirella O. septentrionalis 21-Jul 4 0 1 16 11-Aug 7 0 0 4 4-Sep 6 0 0 10 25-Sep 5 0 0 12 14-Oct 3 1 0 7 6-Nov 2 0 0 6 20-Nov 4 0 0 7 6-Jan 3 5 0 15 28-Jan 10 0 0 3 18-Feb 6 0 0 1 Table A-3. Total individual fr ogs captured in Wekiwa Grid C during each sampling session. Date H. cinerea H. femoralis H. squirella O. septentrionalis 21-Jul 0 11 0 0 Table A-4. Total individual frogs captured in Wekiwa Grid D during each sampling session. Date H. cinerea H. femoralis H. squirella O. septentrionalis 21-Jul 0 9 0 0 Table A-5. Total individual fr ogs captured in Wekiwa Grid E during each sampling session. Date H. cinerea H. femoralis H. squirella O. septentrionalis 21-Jul 0 5 0 0

PAGE 62

62 Table A-6. Total individual fr ogs captured in Wekiwa Grid F during each sampling session. Date H. cinerea H. femoralis H. squirella O. septentrionalis 21-Jul 0 8 0 0 Table A-7. Total individual fr ogs captured in Wekiwa Grid G during each sampling session. Date H. cinerea H. femoralis H. squirella O. septentrionalis 4-Sep 1 1 0 8 25-Sep 1 1 0 23 14-Oct 2 2 0 18 6-Nov 1 1 0 17 20-Nov 0 2 0 2 Table A-8. Total individual fr ogs captured in Wekiwa Grid H during each sampling session. Date H. cinerea H. femoralis H. squirella O. septentrionalis 4-Sep 0 1 0 1 25-Sep 0 3 0 20 14-Oct 1 1 0 22 6-Nov 0 3 0 4 20-Nov 0 0 0 8 Table A-9. Total individual frogs captured in Wekiwa Grid I during each sampling session. Date H. cinerea H. femoralis H. squirella O. septentrionalis 4-Sep 0 1 0 2 25-Sep 1 0 0 14 14-Oct 0 0 0 7 6-Nov 0 1 0 6 20-Nov 0 0 0 1 Table A-10. Total individual fr ogs captured in Wekiwa Grid J during each sampling session. Date H. cinerea H. femoralis H. squirella O. septentrionalis 4-Sep 1 1 0 3 25-Sep 1 3 0 5 14-Oct 1 2 0 8 6-Nov 0 4 0 2 20-Nov 0 4 0 4 Table A-11. Total individual fr ogs captured in Wekiwa Grid K during each sampling session. Date H. cinerea H. femoralis H. squirella O. septentrionalis 28-Jan 0 1 0 1 18-Feb 0 0 0 0

PAGE 63

63 Table A-12. Total individual fr ogs captured in Wekiwa Grid L during each sampling session. Date H. cinerea H. femoralis H. squirella O. septentrionalis 28-Jan 0 0 0 2 18-Feb 0 0 0 1 Table A-13. Total individual frogs captured in Wekiwa Grid M during each sampling session. Date H. cinerea H. femoralis H. squirella O. septentrionalis 28-Jan 0 1 0 0 18-Feb 0 0 0 2 Table A-14. Total individual fr ogs captured in Wekiwa Grid N during each sampling session. Date H. cinerea H. femoralis H. squirella O. septentrionalis 28-Jan 0 0 0 1 18-Feb 0 1 0 1

PAGE 64

64 LITERATURE CITED Allen, E. R. and W T. Neil. 1953. The tree frog, Hyla septentrionalis in Florida. Copeia 1953:127-128. Austin, S. 1975. Exotics. The Flor ida Naturalist, June 1975: 2-5. Barbeau, T.R., and H. Lillywhite. 2005. Body wiping behaviors associated with cutaneous lipids in hylid tree frogs of Florida. Jour nal of Experimental Biology 208:2147-2156. Barbour, T. 1931. Another introduced frog in North America. Copeia 1931:140. Bartareau, T M. 2004. PVC pipe diameter influences the species and sizes of treefrogs captured in a Florida coastal oak scrub commun ity. Herpetological Review 35:150-152. Borg, C. K., S. K. Hoss, L. L. Smith, and L. M. Conner. 2004. From the field: a method for preventing flying squirrel mortality in PV C pipe treefrog refugia. Wildlife Society Bulletin 32:1313-1315. Boughton, R G. 1997. The use of PVC pipe refugia as a trapping technique for hylid treefrogs. Thesis, University of Florida, Gainesville, USA. Boughton, R. G., J. Staiger, and R. Franz. 2000. Use of PVC pipe as a sampling technique for hylid treefrogs. American Midland Naturalist 144:168-177. Brown, R.L. 1974. Diets and habitat preferences of selected anuran s in southeast Arkansas. The American Midland Naturalist 91:468-473. Buchanan, B. W. 1988. Territorial ity in the squi rrel treefrog, Hyla squirella : competition for diurnal retreat sites. Unpub lished Thesis, University of Southwestern Louisiana, Lafayette, USA. Carr, A. F. 1940. A contribution to the herpetology of Florida. University of Florida Biological Publications Science Series, Vol 3, Univers ity of Florida Press, Gainesville FL. Chivers, Douglas P., Joseph M. Kiescker, Erica L. Wildy, Lisa K. Blenden, Lee B. Kats, and Andrew R. Blaustein. 1999. Avoidance respons e of post-metamorphic anurans on cues of injured conspecifics and predator s. Journal of Herpetology, 33: 472-476. Chivers, Douglas P., and R. Jan F. Smith. 1994. Fathead minnows. Pimephales promelas acquire predator recognition when alarm subs tance is associated with the sight of unfamiliar fish. Animal Behavior, 48: 597-605. Colautti, R. I., and H. J. MacIsaac. 2004. A neutral terminology to define invasive species. Diversity and Distributions, 10: 135-141.

PAGE 65

65 Crossland, M. R. 2000. Direct and indirect effects of the introduced Cane Toad Bufo marinus (Anura: Bufonidae) on populations of native anuran larvae in Australia. Ecography, 23: 283-290. Delfino, G., R. Brizzi, D. Nosi, and A. Terreni. 2002. Serous cutaneous glands in New World hylid frogs: an ultrastructural study on skin poisons confirms phylogenetic relationships between Osteopilus septentrionalis and Phrynohyas venulosa Journal of Morphology 253:176-186. Delis, P.R., H.R. Mushinsky, and E.D. McCoy. 1996. Decline of some west-central Florida anuran populations in respons e to habitat degradation. Bi odiversity and Conservation 5:1579-1595. Dodd, C. K. Jr. 1991. Drift fence-associated samp ling bias of amphibians at a Florida sandhills temporary pond. Journal of Herpetology 25:296-301. Dodd, C. K. Jr. 1996. Use of terrestrial habitats by amphibians in the sandhill upland on northcentral Florida. Alytes 14:42-52. Duellman, W. E. and Schwartz, A. 1958. Amphibians and reptiles of southern Florida. Bulletin of the Florida State Museum 3:181-324. Duellman, W. E., and L. Trueb. 1986. Biol ogy of Amphibians. McGraw-Hill Publishing Company, New York, USA. Elton, C. S. 1958. The ecology of invasions by animal s and plants. University of Chicago Press, Chicago and London. Florida Automated Weather Network. 2007. Archiv ed Weather Data. IFAS. University of Florida. . Accessed 18 October 2007. Gause, G. F. 1934. The struggle for existen ce. Williams and Wilkins, Baltimore. Goin, O. B. 1958. A comparison of the nonbreeding habits of two treefrogs, Hyla squirella and Hyla cinerea Quarterly Journal of the Florid a Academy of Sciences 21:49-60. Greenlees, M. J. G. P. Brown, J. K. Webb, B. L. Philips, and R. Shine. 2007. Do invasive cane toads ( Chaunua marinus ) compete with Australian frogs ( Cyclorana australis )? Austral Ecology 32: in press. Griffin, Andrea S., Daniel T. Blumstein, and Ch ristopher S. Evans. 2000. Training captive-bred or translocated animals to avoid predators. Conservation Biology, 14: 1317-1326. Hardy, J. 1972. Tentative outline for inventory of amphibians: Hyla cinerea (green tree frog). Chesapeake Science, Vol 13; Supplement: Biota of the Chesapeake Bay, pp. S180-S190.

PAGE 66

66 Heflick, S.K. 2001. Ecology of the exotic Cuban tree frog, Osteopilus septentrionalis within Brevard County, Florida. Unpublished thesis. Florida Inst itute of Technology, Melbourne. Hirai, T. 2006. Effects of bank comp acting on frogs: Can the density of Hyla japonica be recovered by installment of PVC pipe artificial refuges in the banks of rice fields? Japanese Journal of Applied Entomol ogy and Zoology 50:31-335. [in Japanese.] Jaeger, R. G., E. D. Prosen, and D. C. Adam s. 2002. Character displacement and aggression in two species of terrestrial sa lamanders. Copeia, 2002:391-401. Johnson, J. R. 2005. A novel arboreal pipe-trap designed to capture the gray treefrog (Hyla versicolor ). Herpetological Review 36:274-277. Johnson, S. A. 2004. Osteopilus septentrionali s (Cuban treefrog) geographic distribution. Herpetological Review 35:405. Johnson, S. A. 2007. Osteopilus septentrionali s (Cuban treefrog) geographic distribution. Herpetological Review 38:349. Kiesecker, J. M. 2003. Invasive species as a global problem: toward understanding the worldwide decline of amphibians. In R. D. Semlitsch (ed.), Amphibian Conservation, pp. 113-126. Smithsonian Books, Washington. Kiesecker, J. M. and A. R. Blaustein. 1997. Popul ation differences in responses of red-legged frogs ( Rana aurora) to introduced bullfrogs ( Rana catesbeiana). Ecology, 78:1752-1760. Kiesecker, J. M., D. P. Chivers, and A. R. Blau stein. 1996. The use of chemical cues in predator recognition by western toad tadpoles Animal Behavior, 52: 1237-1245. Kupferberg, S. J. 1997. Bullfrog ( Rana catesbeiana) invasion of a California river: the role of larval competition. Ecology 78:1736-1751. Krysko, K.L., K.M. Enge, J.H. Townsend, E.M. Langan, S.A. Johnson, and T.S. Campbell. 2005. New county records of amphibians and reptiles from Florida. Herpetological Review. 3:85-87. Lee, D. S. 1969. The Treefrogs of Florida. Florida Naturalist, 42:117-120. Lockwood, J. L., P. Cassey, and T. Blackbur n. 2005. The role of propagule pressure in explaining species invasi ons. Trends in Ecology and Evolution, 20: 223-227. Lutz, B. 1960. Fighting and an incipient notion of territory in male tree frogs. Copeia. 1960:6163.

PAGE 67

67 McAuliffe, J. R. 1978. Biological survey and management of sport-hunte d bullfrog populations in Nebraska. Nebraska Game and Park s Commission, Lincol n, Nebraska. McComb, W. C., and R. E. Noble. 1981. Herpetof aunal use of natural tree cavities and nest boxes. The Wildlife Society Bulletin. 9:261-267. Meshaka, W. E. 1996. Retreat use by the Cuban treefrog ( Osteopilus septentrionalis ); implications for successful colonization in Florida. Journal of Herpetology. 30:443-445. Meshaka, W. E. 2001. The Cuban treefrog in Flor ida: life history of a successful colonizing species. University Press of Fl orida, Gainesville, Florida. Meshaka, W. E. and K. J. Babbitt. 2005. Amphibians and reptiles: status and conservation in Florida. Krieger Publishing Co mpany, Malabar, Florida. Mitchell, J. C. 2005. Hyla femoralis Bosc, 1800, pine woods treefrog. Pages 454-455 in M. Lannoo, editor. Amphibian Declines: Conservation Status of the United States Species. University of California Press, Berkley, USA. Mitchell, J. C. and M. J. Lannoo. 2005. Hyla squirella Bosc,1800, squirrel treefrog. Pages 456458 in M. Lannoo, editor. Amphibian Declines: Conservation Stat us of the United States Species. University of Calif ornia Press, Berkley, USA. Murphy, C. G. 1993. A modified drift fence for capturing treefrogs. Herpetological Review, 24:143-145. Moulton, C. A., W. J. Flemming, and B. R. Nerney. 1996. The use of PVC pipes to capture hylid frogs. Herpetological Review 27:86-187. Pacala, S. W., and J. Roughgarden. 1982. Resource partitioning and interspecific competition in two two-species insular Anolis lizard communities. Science 217:444. Pearl, Christopher A., Michael J. Adams, Gera ld S. Schuytema, and Alan V. Nebeker. 2003. Behavioral responses of anuran larvae to chemical cues of native and introduced predators in the pacific northwestern Unite d States. Journal of Herpetology, 37: 572-576. Phillips, B. L., G. P. Brown, and R. Shine. 2003. Assessing the potential impact of cane toads on Australian snakes. Conservation Biology 17:1738-1747. Redmer, M. and R. A. Brandon. 2005. Hyla cinerea (Schneider, 1799), gr een treefrog. Pages 452-454 in M. Lannoo, editor. Amphibian Declin es: Conservation Status of the United States Species. University of California Press, Berkley, USA. Ritke, M. E. and J. G. Babb. 1991. Behavior of the gray treefrog ( Hyla chrysoscelis ) during the non-breeding season. Herpetol ogical Review 22:5-8.

PAGE 68

68 Schwartz, A.. 1952. Hyla septentrionalis Dumeril and Bibron on the Florida mainland. Copeia, 1952:117-118. Schwarzkopf, L. and R. A. Alford. 1996. Desi ccation and sh elter-site use in a tropical amphibian: comparing toads with physic al models. Functional Ecology 10:193-200. Simberloff, D. 2000. Nonindigenous sp ecies: a global threat to biodive rsity and stability. In P.H. Raven (ed.), Nature and human societ y, pp. 325-334. National Academic Press, Washington D.C. Smith, K. G. 2004. Osteopilus septentrionalis (Cuban treefrog); Reproductive behavior. Herpetological Review, 35:374-375. Smith, K. G. 2005a. Effects of nonindigenous tadpoles on native tadpoles in Florida: evidence of competition. Biological Conservation 123:443-441. Smith, K.G. 2005b. An exploratory assessment of Cuban treefrog ( Osteopilus septentrionalis ) tadpoles as predators of nativ e and nonindigenous ta dpoles in Florida. Amphibia-reptilia 26:571-575. Stauffer, H. P. and R. D. Semlitsch. 1993. Effects of visual, chemical and tactile cues of fish on the behavioral responses of tadpoles. Animal Behavior, 43: 355-364. Stewart. M. M. and F. H. Pough. 1983. Population density of tropical forest frogs: relation to retreat sites. Science 221:570-572. Stewart, M. M. and A. S. Rand. 1991. Vocalization and the defense of retreat sites by male and female frogs, Eleutherodactylus coqui Copeia 1991:1013-1024. Torchin, M. E., K. D. Lafferty, A. P. Dobson, V. J. McKenzie, and A. M. Kuris. 2003. Introduced species and their missing parasites. Nature, 421: 628-630. Walsh, P. T. and J. R. Downie. 2005. The effects of shelter availability and substrate quality on behavior and post-metamorphic growth in th ree species of anuran s: implications for captive breeding. Herpetol ogical Journal 15:245-255. Walters, E.C. and J. M. Kneitel. 2004. Use of wa ter-filled red-cockaded woodpecker cavities by other organisms. In R. Costa and S. J. Dani els (eds). Red-cockaded woodpeckers: road to recovery. Hancock House Publishers. Baline, Washington pp 492-497. Welker, M.E. 2004. Geographic distribution: Osteopilus septentrionalis (Cuban treefrog). Herpetological Review, 35:283. Wiewandt, T. A. 1971. Breeding biology of the Mexican leaf frog. Fauna. 2:29-34.

PAGE 69

69 Wilcove, D. S., D. Rothstein, J. Dubow, A. Phillips, and E. Losos. 1998. Quantifying threats to imperiled species in the United States: a ssessing the relative im portance of habitat destruction, alien species, pollution, overe xploitation, and diseas es. BioScience 48:607615 Williamson M., and A. Fitter. 1996. The varying success of invaders. Ecology, 77:1161-1666. Wright, A. H. 1932. Life-histories of the fr ogs of Okefinokee Swamp, Georgia. Macmillan Company, New York. Wright, A. H. and A. A. Wr ight. 1949. Handbook of frogs and toads of the United States and Canada. Comstock Publishing Associates, Ithaca, New York. Wyatt, J. L. and E. A. Forys. 2004. Conservati on implications of pred ation by Cuban treefrogs ( Osteopilus septentrionalis ) on native hylids in Florida. Southeastern Naturalist 3:695700. Zacharow, M., W. J. Barichivich, and C. K. Dodd, Jr. 2003. Using groundplaced PVC pipes to monitor hylid treefrogs: capture bias es. Southeastern Naturalist 24:575-590.

PAGE 70

70 BIOGRAPHICAL SKETCH Kristine E. Hoffm ann grew up in Sterling, Massachusetts, where she lived with her parents, two brothers, and her twin sister, Carolyn. She graduated with honors from the University of Massachusetts in 2005 with a B.S. in biology and a minor in wildlife and fisheries conservation, and entered the University of Fl orida that fall. She hopes to teach biology and conservation at a junior college, continue field and lab work on amphibians and reptiles, and to help support her sister, who recently entered gr aduate school for costume design at New York University.