Citation
Microhabitat preference of the introduced gecko Hemidactylus turcicus in an urban environment

Material Information

Title:
Microhabitat preference of the introduced gecko Hemidactylus turcicus in an urban environment
Creator:
Zlatar, Patricia A. Gomez ( Author, Primary )
Publication Date:
Copyright Date:
2003

Subjects

Subjects / Keywords:
Aluminum ( jstor )
Bricks ( jstor )
Cements ( jstor )
Construction materials ( jstor )
Ecology ( jstor )
Logistic regression ( jstor )
Microhabitats ( jstor )
Modeling ( jstor )
Vegetation ( jstor )
Wall temperature ( jstor )
City of Gainesville ( local )

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright Zlatar, Patricia A. Gomez. 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.
Embargo Date:
9/9/1999
Resource Identifier:
53208108 ( OCLC )

Downloads

This item has the following downloads:

PDF ( 2 MBs ) ( .pdf )

gomez_p_Page_081.txt

gomez_p_Page_073.txt

gomez_p_Page_082.txt

gomez_p_Page_058.txt

gomez_p_Page_027.txt

gomez_p_Page_074.txt

gomez_p_Page_096.txt

gomez_p_Page_028.txt

gomez_p_Page_038.txt

gomez_p_Page_052.txt

gomez_p_Page_034.txt

gomez_p_Page_101.txt

gomez_p_Page_060.txt

gomez_p_Page_076.txt

gomez_p_Page_091.txt

gomez_p_Page_047.txt

gomez_p_Page_049.txt

gomez_p_Page_092.txt

gomez_p_Page_005.txt

gomez_p_Page_099.txt

gomez_p_Page_057.txt

gomez_p_Page_032.txt

gomez_p_Page_022.txt

gomez_p_Page_030.txt

gomez_p_Page_011.txt

gomez_p_Page_059.txt

gomez_p_Page_088.txt

gomez_p_Page_066.txt

gomez_p_Page_093.txt

gomez_p_Page_029.txt

gomez_p_Page_012.txt

gomez_p_Page_085.txt

gomez_p_Page_041.txt

gomez_p_Page_035.txt

gomez_p_Page_051.txt

gomez_p_Page_024.txt

gomez_p_Page_070.txt

gomez_p_Page_018.txt

gomez_p_Page_033.txt

gomez_p_Page_020.txt

gomez_p_Page_105.txt

gomez_p_Page_015.txt

gomez_p_Page_003.txt

gomez_p_pdf.txt

gomez_p_Page_079.txt

gomez_p_Page_054.txt

gomez_p_Page_025.txt

gomez_p_Page_077.txt

gomez_p_Page_102.txt

gomez_p_Page_036.txt

gomez_p_Page_001.txt

gomez_p_Page_084.txt

gomez_p_Page_056.txt

gomez_p_Page_026.txt

gomez_p_Page_067.txt

gomez_p_Page_017.txt

gomez_p_Page_064.txt

gomez_p_Page_100.txt

gomez_p_Page_069.txt

gomez_p_Page_098.txt

gomez_p_Page_083.txt

gomez_p_Page_071.txt

gomez_p_Page_002.txt

gomez_p_Page_044.txt

gomez_p_Page_043.txt

gomez_p_Page_007.txt

gomez_p_Page_046.txt

gomez_p_Page_021.txt

gomez_p_Page_065.txt

gomez_p_Page_087.txt

gomez_p_Page_050.txt

gomez_p_Page_013.txt

gomez_p_Page_075.txt

gomez_p_Page_019.txt

gomez_p_Page_094.txt

gomez_p_Page_039.txt

gomez_p_Page_031.txt

gomez_p_Page_061.txt

gomez_p_Page_104.txt

gomez_p_Page_008.txt

gomez_p_Page_078.txt

gomez_p_Page_037.txt

gomez_p_Page_086.txt

gomez_p_Page_103.txt

gomez_p_Page_095.txt

gomez_p_Page_062.txt

gomez_p_Page_068.txt

gomez_p_Page_023.txt

gomez_p_Page_040.txt

gomez_p_Page_006.txt

gomez_p_Page_080.txt

gomez_p_Page_072.txt

gomez_p_Page_014.txt

gomez_p_Page_045.txt

gomez_p_Page_048.txt

gomez_p_Page_004.txt

gomez_p_Page_009.txt

gomez_p_Page_097.txt

gomez_p_Page_053.txt

gomez_p_Page_042.txt

gomez_p_Page_090.txt

gomez_p_Page_063.txt

gomez_p_Page_089.txt

gomez_p_Page_010.txt

gomez_p_Page_016.txt

gomez_p_Page_055.txt


Full Text












MICROHABITAT PREFERENCE OF THE INTRODUCED GECKO Hemidactylus turcicus
IN AN URBAN ENVIRONMENT















By

PATRICIA A.GOMEZ ZLATAR


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


2003

































Copyright 2003

by

Patricia A. Gomez Zlatar















ACKNOWLEDGMENTS

I would first like to thank my committee chair, Mike Moulton, for his constant

support, mentoring, and enthusiasm. His friendship has been instrumental in the

completion of this degree. I owe a great deal of thanks to committee member Ken Portier

for the endless hours of guidance he provided. I am also grateful to committee member

Dick Franz, for providing useful comments and ideas throughout the study. It has been

both a privilege and pleasure to interact with them all.

Many people made my fieldwork possible. I owe thanks to both the UF Physical

Plant Division and to J. Darcy White for providing me with detailed maps and a wealth of

information. I would like to thank the UF Campus Police Force and the VA Hospital

Police Force for keeping my helpers and me safe during our nighttime surveys. I also

want to give a big thanks to Chuck Knapp, Ester Langan, Elza Kephart, Alex Martin, and

Robin Sternberg for providing endless hours of field assistance.

Finally, I want to give immense gratitude to those closest to me. I give thanks to

my family, especially my parents. Also, I want to thank my close friends (they know

they are). Last, but not least, I want to thank Alex Martin for his patience, friendship, and

love. I could not have done it without him.
















TABLE OF CONTENTS

page

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

LIST OF TABLES .................................................................... .. .... ..... vi

LIST OF FIGURES ................................. .............. ............. .......... viii

ABSTRACT ........ .............. ............. ..... .......... .......... ix

CHAPTER

1 IN TR O D U C T IO N ............................................................. .. ......... ...... .....

U rban Ecology ................................................... ................ .............
G general O bj active ................................................................. ............. 2

2 STUDY SPECIES AND STUDY SITES ............. ............................ ...............3

Stu dy Sp ecies ................................................ 3
Study Sites ............................ .................................... 6
The Study Species in the Study Sites ........................................ ........................ 8

3 PRELIM INARY SURVEY ............................................................. .................9

Introduction..................................... ........................... .... ..... ........ 9
M e th o d s ............................................................. ................ 9
R results ....................................................... ...................... ............... 11
D iscu ssio n ...................................... ................................................. 13

4 ADDITIONAL NATURAL HISTORY NOTES .......................................................21

Intro du action ...................................... ................................................ 2 1
M eth o d s ..............................................................................2 3
R e su lts .................. ........................................................ ................ 2 4
Su m m er 2 00 1 .................................................................. ................. 24
Fall/W inter 2001 .................................... .......................... ........... 25
S p rin g 2 0 0 2 .................................................................................................... 2 6
D isc u ssio n .............................................................................................................. 2 6









5 MICROHABITAT PREFERENCE IN THE INTRODUCED GECKO,
H EM ID A CTYL US TURCICU S .......... ................. .................................................39

Introduction............ ........... ....................................................... ..... 39
M methods and Results ........... .... .... .. ........................ ........ ....... ..... 41
Sam pling M methods .............. ..... .... ........ .. .. .................... ...............41
D ata A n aly sis 1.............................................................4 4
R esu lts 1 ................................................................... 4 6
D ata A n aly sis 2 .............................................................4 8
R esu lts 2 .................................................................. 4 9
D ata A n aly sis 3 .............................................................50
R esu lts 3 .................................................................. 50
D ata A n aly sis 4 .............................................................5 1
R esu lts 4 .................................................................. 52
D ata A n aly sis 5 ..............................................................52
R esu lts 5 .................................................................. 52
D discussion ............... .. ..... .....................................................................54

6 MANAGEMENT AND CONSERVATION IN AN URBAN ENVIRONMENT ....75

APPENDIX

A PRELIMINARY SURVEY DATA ....................................................................79

B NATURAL HISTORY DATA..................... ...............81

C MICROHABITAT PREFERENCE DATA ........................................ .....84

R E F E R E N C E S ................................................................9 1

B IO G R A PH IC A L SK E T C H ........................................................................................ 96
















LIST OF TABLES


Table pge

3-1 Criteria of wall characterization variables used in preliminary survey .................. 16

4-1 Criteria of natural history variables................................... .................. ... ......... 31

4-2 Temperature measurements of wall surface for adult and subadult H. turcicus;
fall/winter 2001 .................................... ............................... ........31

4-3 Temperature measurements of wall surface for adult H. turcicus; spring 2002.......31

5-1 Description of wall characterization variables for microhabitat study...................57

5-2 Chi-square and Fisher Exact p-values for wall characterization variables .............58

5-3 The dependency of age, color, length, and texture on material.............................59

5-4 Logistic regression models resulting from the backward elimination method
using the variables: material; cardinal orientation, vegetation, light .....................60

5-5 Summary statistics of the significant predictor variables of the
M *CO*V+M *CO*L m odel .............................................................................. 62

5-6 Possible wall combinations involving the variables: material, cardinal
orientation, vegetation, and light .................................... ................. ........ ....... 63

5-7 Number of observations per wall combination involving the variables: material,
vegetation (2 levels), and light ...................................................................... ..... 64

5-8 Logistic regression models resulting from the backward elimination method
using the variables: material, vegetation (2 levels), and light..................................65

5-9 Number of observations per wall combination involving the variables: material,
and light-vegetation (3 levels)........................................................ ............... 67

5-10 Logistic regression models resulting from the backward elimination method
using the variables: material, and light-vegetation (3 levels)................................67

5-11 Chi-square test and Fisher's Exact test for light-vegetation and gecko presence,
controlling for m aterial.......... ......................................................... .. .... ...... 68









5-12 Chi-square test and Fisher's Exact test for material and gecko presence,
controlling for light-vegetation ........................................... ......................... 69

5-13 Chi-square test for light-vegetation and gecko presence ......................................70

5-14 Chi-square test for material and gecko presence................................ ..............71

5-15 Three-way contingency table of light-vegetation, controlling for material, with
associated percentages and marginal associations .........................................72

5-16 Three-way contingency table of material, controlling for light-vegetation, with
associated percentages and marginal associations .........................................73

A-1 Temperature readings of walls with respect to material and cardinal orientation....79

B-l Temperature readings (C) for individual adult H. turcicus recorded during the
fall/w inter 200 1 survey .............................................................................. 8 1

B-2 Temperature readings (C) for individual sub-adult H. turcicus recorded during
the fall/w inter 2001 survey ........................................................ ............. 82

B-3 Temperature readings (C) for individual adult H. turcicus recorded during the
spring 2002 survey .................................... .... ... ...... ...............83
















LIST OF FIGURES


Figure page

3-1 Depiction of the 45 angle leeway employed in the determination of wall
cardinal orientation ......... ................................................................ ...... .... ....17

3-2 Average wall temperature of different construction material ................................. 17

3-3 Average wall temperature of different cardinal locations ........... ...............18

3-4 Average wall temperature of different construction materials at different
cardin al lo cation s............................................................................ ............... 18

3-5 Average number of H. turcicus per building of different construction material......19

3-6 Number of H. turcicus on walls of different cardinal locations............................19

3-7 Number of H. turcicus on walls of different vegetation levels .............................20

3-8 Number of H. turcicus on walls of different light intensities................................20

4-1 Perch height preference of adult and subadult H. turcicus (summer 2001) ............32

4-2 Perch height preference of adult and subadult H. turcicus (fall/winter 2001) .........33

4-3 Social preference of adult and subadult H. turcicus (fall/winter 2001) ...................34

4-4 Exposure preference of adult and subadult H. turcicus (fall/winter 2001)..............35

4-5 Perch height preference of adult H. turcicus (spring 2002) ...................................36

4-6 Social preference of adult H. turcicus (spring 2002) ............................................37

4-7 Exposure preference of adult H. turcicus (spring 2002) .......................................38















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

MICROHABITAT PREFERENCE OF THE INTRODUCED GECKO Hemidactylus turcicus
IN AN URBAN ENVIRONMENT


By

Patricia A. Gomez Zlatar

August 2003

Chair: Michael Moulton
Major Department: Wildlife Ecology and Conservation

I investigated microhabitat preference in the introduced gecko H. turcicus in

Gainesville, Florida from summer 2001 through spring 2002. After collecting extensive

natural history data in 2001, I then attempted to construct a model for microhabitat

preference during 2002, using logistic regression. I characterized 160 walls by

construction material, vegetative cover, artificial light intensity, cardinal orientation, age

of building, length of wall, surface color, and surface texture. I sampled each wall twice

for the presence or absence ofH. turcicus.

Chi-square analyses indicated that the age of building, length of wall, surface color,

and surface texture were dependent on construction material (p< 0.05). I then fit a

logistic regression model with the variables: construction material, vegetative cover,

artificial light intensity, and cardinal orientation. I was unable to obtain a functional

logistic regression model, probably owing to a small sample size. Thus, I condensed the

dataset by eliminating the variable cardinal orientation and combining the variables









vegetative cover and artificial light intensity. I was unable to obtain a significant logistic

regression model (p<0.05) with this reduced dataset; and I therefore performed chi-square

analyses instead. Results revealed no significance (p<0.05) between the presence ofH.

turcicus and the three wall variables examined. Hence, the presence ofH. turcicus on a

wall appears to be independent of construction material, vegetative cover, and artificial

light intensity. This conclusion indicates that H. turcicus does not demonstrate a

preference among walls of different material type, vegetative cover levels, and artificial

light intensities. These results could reflect the generalist tendency of this gecko; and

thus explain its overall resiliency as an introduced species. Alternatively, the inability to

detect significance could reflect failure to gather a sufficient sample size; or failure to

properly select variables relevant to microhabitat preference in this species.














CHAPTER 1
INTRODUCTION

Urban Ecology

Urbanization is a significant global phenomenon (Grimm et al., 2000). Roughly

half of the world's population currently resides in cities, as the tendency toward

urbanization is characteristic of both developed and developing nations (Lord et al.,

2003). This trend is projected to increase over the next few decades, whereby the number

and sizes of cities are expected to grow extensively (Pickett et al., 2001). The inevitable

growth of urban areas, and their subsequent ecological impacts, makes the field of urban

ecology timely and ultimately essential (Grimm et al., 2000).

Urban ecology is a fairly young discipline that has increased in prominence over

the last two decades (Rebele, 1994). Before this, ecological studies in urban

environments were rare as ecologists generally considered urban areas ecologically

inferior to natural ones; and thus demonstrated little interest and overall disregard for

cities (Botkin and Beveridge, 1997; Gilbert, 1989). However, this anti-urban attitude

began to progressively vanish when ecologists started to recognize and become

concerned with the influence of humans on ecosystems (Niemela, 1999). Urban ecology

first emerged as a discipline that mainly dealt with the ecology of habitats and organisms

within cities (Pickett et al., 2001). It eventually expanded when it embraced and then

advocated the notion that cities were ecosystems in themselves, with humans occupying

the position of keystone species (Rees, 1997). With this new perspective, urban

ecologists have recognized that the urban setting cannot be adequately understood and









that findings are inapplicable without accounting for human influence (Grimm et al.,

2000). As a consequence, urban ecology is in the process of developing into an

integrated discipline that incorporates the social, behavioral, economic, physical, and

ecological sciences (Niemela, 1999).

Despite the surge in popularity of urban ecology, few studies have strictly dealt

with urban species and/or been conducted in urban settings (McIntyre et al., 2000). In a

review of leading ecology journals between 1993 and 1997, Collins at al. (2000)

concluded that a mere 0.4% of papers surveyed (25 of 6157) were restricted specifically

to urban habitat and/or urban wildlife. Although reasons for this paucity have not been

specified, urban areas offer the distinct challenge of being controlled by strong and

diverse human actions (Dow, 2000). An ongoing history of intense and varied

micromanagement has made metropolitan landscapes into highly heterogeneous areas,

their uniqueness wrought with logistical constraints (Dow, 2000; McIntyre et al., 2000).

Of the few wildlife ecological studies performed in urban environments, most

feature birds, mammals, and terrestrial invertebrates. Significantly less popular subjects

are aquatic fauna, amphibians, and reptiles (Luniak and Pisarki, 1994). Many of these

studies, in turn, are anecdotal in nature; thus making comparisons between locations and

the formation of a general body of knowledge unfeasible (McIntyre et al., 2000).

General Objective

In light of the current shortcomings in urban ecology, the main objective of this

study was to conduct a repeatable, quantitative ecological study on the microhabitat

preferences of an introduced reptile in an urban environment.














CHAPTER 2
STUDY SPECIES AND STUDY SITES

Study Species

The Mediterranean gecko Hemidactylus turcicus is an old world geckkonid lizard

that has successfully extended its range into India and North America through human-

assisted introductions (Conant and Collins, 1998). This species occurs naturally in the

Middle East and Mediterranean regions. Hemidactylus turcicus is thought to have

reached North America after being initially introduced through human agencies into the

Antilles and Gulf-coastal Mexico. It first appeared in the United States in 1915 in Key

West, Florida (Stejneger, 1922). Since then, H. turcicus has expanded onto the mainland

where it has established itself in a number of localities throughout the southeastern and

south-central states; specifically Alabama, Arizona, Arkansas, California, Florida,

Georgia, Louisiana, Mississippi, and Texas (Barbour, 1936; Etheridge, 1952; Conant,

1955; Conant and Collins, 1998).

The Mediterranean gecko possesses a life history that, with human assistance,

favors successful colonization of new areas. Both Davis (1974) and Meshaka Jr. (1995)

have reported that the dispersal ofH. turcicus in Texas and Florida parallels that of major

highways; and that produce trucks were the most likely source of transportation for this

gecko. The calcareous shelled eggs ofH. turcicus are fairly resistant to desiccation; and

egg survivorship appears to be extremely high. With an incubation period of 40 to 45

days, these eggs are ideal for surviving lengthy truck rides (Selcer, 1986). In 1986,









Selcer confirmed this high egg survivorship when he obtained a 100% hatching success

rate from 100 eggs he had collected in the field.

The secretive nesting behavior of the Mediterranean gecko also favors it as a

colonizing species. Nests in this species are usually constructed in hidden locations such

as attics, storage rooms, under eaves of houses, closets, and rock crevices; and on a wide

array of surfaces, including cardboard boxes, wood planks, and old clothing (Davis,

1974; Punzo, 2001a; Selcer, 1986; Trauth, 1985). Furthermore, eggs have reduced

visibility as they are often covered with debris, including dirt, paper, eggshell and shed

skin (Punzo, 2001a; Rose and Barbour, 1968; Selcer, 1986). Many of the nesting sites

are in prime positions to be moved or transported in vehicles. Nests in H. turcicus range

from solitary to communal, with some communal nests containing as many as 20 eggs

(Selcer, 1986). Therefore, the possibility exists of unknowingly transporting a mini

colony to a new locale. Mediterranean gecko hatchlings have also shown remarkable

survivorship in dry environments, requiring no food or water for up to a month (Rose and

Barbour, 1968). Thus, H. turcicus eggs/hatchling can survive dry and nutrient-poor

environments for extended periods of time (over two months) making this gecko a

resilient and ultimately successful stowaway on vehicles.

Once at a new site, H. turcicus often occurs at extremely high densities: as many as

544 to 2210 geckos/hectare in Texas (Selcer, 1986); and 497 to 1463 geckos/hectare in

Florida (Punzo, 2001a). A high population density coupled with a consistently

encountered 1:1 sex ratio allows for the potential of a large, annual reproductive output

by a population. Females are reproductively active between the months of April and

September; and are believed to have two to three clutches a season, each clutch









consisting of two eggs (Rose & Barbour, 1968; Selcer, 1986; Meshaka Jr., 1995; Punzo,

2001a). Hemidactylus turcicus is also characterized by being an early maturing species

with a long lifespan; on average, juveniles require between eight and nine months to

mature, and routinely live at least three years (Selcer, 1986).

Hemidactylus turcicus is highly pre-adapted for life in urbanized areas; and this

further aids in dispersal. The presence of scansors (adhesive pads on toes) allows the

Mediterranean gecko to perch on vertical walls of buildings. In fact, the Mediterranean

gecko is a familiar resident in many cities and towns around the world; and according to

Luiselli and Capizzi (1999) is more often found in human-disturbed areas than in natural

environments. In addition, H. turcicus is considered to be a generalized predator, as a

number of studies have reported a wide array of mostly arthropod prey in this gecko's

diet (Carey, 1988; Punzo, 2001a; Saenz, 1996). Although Punzo (2001a) failed to detect

age- or sex-related differences in diets, Saenz (1996) observed food partitioning between

both juveniles and adults and males and females, although the latter remained

inconclusive due to small sample size. Thus, under some circumstances, food

partitioning could further contribute to the success of H turcicus, since it would reduce

intraspecific competition, and increase the feeding efficiency of a given population

(Saenz, 1996).

The Mediterranean gecko further increases its chances for establishment with a

number of predator escape and avoidance tactics. As a nocturnal, arboreal, and

cryptically colored lizard, H. turcicus has few known predators (Selcer, 1986). In a study

conducted in Tampa Bay, Florida, Punzo (2001a) listed bats, Cuban tree frogs, large

heteropodid crab and wolf spiders, giant tail scorpions, and feral domestic cats as









possible predators. As expected, the most vulnerable period for this gecko is shortly after

hatching; both Selcer (1986) and Punzo (2001a) showed that mortality was significantly

greater for small juveniles when compared to large juveniles, adult females, and adult

males. Accordingly, Selcer (1986) regularly observed an avoidance mechanism in

juveniles termed "tail wagging"; when disturbed, H. turcicus juveniles draw attention

away from their bodies by wagging their conspicuously banded tails, which can easily be

autotomized for a rapid escape. Adult H. turcicus are adept at fleeing danger too, as

noted by Selcer (1986) they use routine escape routes when harassed. Adults also have

easily autotomized tails, which can startle and/or distract a predator while the gecko

retreats to safety (Selcer, 1986).

Most of the areas invaded by H. turcicus have little or no competitive pressure.

Noted exceptions include competition with the introduced gecko Cyrtopodion scabrum in

Texas, and the nonindigenous geckos Hemidactylus garnotii and Hemidactylus mabouia

in south Florida (Klawinski et al., 1994; Punzo, 2001b). In both previously mentioned

cases, H. turcicus appears to be competitively excluded and replaced in many locations.

In Texas, the competitive failure ofH. turcicus has been linked to the ability of

C. scabrum to monopolize prey and force H. turcicus to undertake a dietary shift

(Klawinski et al., 1994). In Florida, increased digestive and assimilation efficiencies, and

continuous reproduction have been suggested as factors giving H. garnotii and

H. mabouia a competitive edge over H. turcicus (Meshaka Jr., 1995; Punzo, 2001b).

Study Sites

My primary study area was the University of Florida campus located in the town of

Gainesville in north central Florida. Since its official inception in 1906, the University of

Florida has continuously expanded in size to become the fourth largest university in the









United States. Following an extended construction boom starting in 1950 and ending in

1999, the 2,000-acre University of Florida campus has roughly 1,251 buildings that

provide approximately 18, 670, 086 gross square feet of area. The University of Florida

has just nearly 60,000 full and part-time students and employees, resulting in a vast

amount of pedestrian and vehicular traffic. (University of Florida, 2002)

My secondary study site was the Gainesville VA Medical Hospital Center within

the boundaries of the University of Florida. Construction of the VA Medical Center

commenced in 1964 and was completed in 1967. Presently it consists of 38 buildings

(US Department of Veterans Affairs, 2003).

The landscapes of both the University of Florida campus and the Gainesville VA

Medical Center are highly heterogeneous, as both contain a wide assortment of facilities,

each with its distinct type of architecture, ornamental vegetation and/or decorative natural

scenery, walkway(s), road(s), and other associated features. This diversity is emphasized

by the amount of variation present on each individual building, to the extent that it is a

rare occurrence to encounter two identical walls. Walls are the principle habitat of

Mediterranean geckos in this environment. Complexity depends on human factors such

as stylistic trends, budgetary schemes, and logistics. This complexity was apparent

during a preliminary survey that I undertook between the months of July 2001 and

November 2001. A wide array of human activity continuously affected my sampling

locations; these activities ranged from complete modification of existing vegetation, to

the failure to repair damaged building structures. Thus, in an attempt to control for the

fluctuating quality of campus buildings, I used categorical variables, as opposed to

continuous variables, to quantify sampling locations.






8


The Study Species in the Study Sites

Records of H. turcicus for the University of Florida campus in Gainesville first

began in 1956 (King, 1959). Although no official records ofH. turcicus exist from the

Gainesville VA Medical Center, the gecko's presence on the premises was confirmed

during a preliminary survey. Evidence of large populations of other nonindigenous

hemidactyline geckos have not yet been recorded at either location, thus making these

localities ideal for the investigation of the ecology of H. turcicus in the absence of

potential behavioral changes caused by interference of directly competing species.














CHAPTER 3
PRELIMINARY SURVEY

Introduction

Despite its widespread distribution and abundance on the campus of the University

of Florida, the Mediterranean gecko has not been studied intensively. In 1956, King

collected specimens ofH. turcicus from a wood frame building on the campus. The

following year, King and a colleague collected 49 H. turcicus individuals over three

nights, also on frame buildings. A year later, Riemer collected an additional 44

specimens from the same buildings during one night of sampling (King, 1959). In an

attempt to elucidate additional natural history patterns of H. turcicus, particularly with

respect to habitat preference, and to obtain some general insight on their environment, I

conducted a survey on the UF campus and VA Hospital, between July and August of

2001.

Methods

I randomly sampled 48 buildings over seven nights. I sampled each building by

directing a flashlight systematically over all accessible walls from top to bottom, from

right to left. I further characterized each wall by its type of construction material (one of

four types: aluminum, brick, cement, and wood) and cardinal orientation. I arbitrarily

used 50% of the surface area of a given wall for my material classification. Thus, I

categorized a wall constructed with > 50% wood, as wood. The majority of the buildings

were constructed from only one material. Some of the buildings consisted of more than









one material. However, of these, I limited my survey to only those buildings that

featured a predominate material.

I determined the cardinal orientation of walls using official maps of the University

of Florida (UF Physical Plant Division, 2000), and of the VA Hospital Engineering

Department. For the cardinal orientation categories I used north, south, west, and east.

Nearly all of the walls were clearly oriented in one of these directions. For those walls

that were not clear-cut, I allowed a 450 angle of leeway on each side of the cardinal

direction (Figure 3-1).

For each gecko, I also recorded one of three vegetation levels. The three vegetation

levels were all based on the height of the vegetation, rather than diversity. To facilitate

measurement in the field, I based these heights using a simple system. Thus, the low

level referred to flora no higher than my knees (< 0.51 m), the medium level to floral

height between my knees and shoulder (> 0.51 m and < 1.37 m), and the high level

comprised of flora reaching higher than my shoulders ( >1.37 m ). I limited my

vegetation classification to an imaginary half sphere, coming out of the wall, with a one-

meter radius and with the gecko as its center.

I also determined light intensity with respect to each gecko, along the same lines as

my vegetation classification. I used two categories to describe light intensity, again in an

imaginary half sphere with a one-meter radius around each gecko. I assigned an

area to the low light category if it had at most one dim light, whereas I classified an area

as high light if there was either at least one bright light or at least two dim lights. The

difference between dim and bright was fairly arbitrary; however, as a general rule of

thumb, if I had to use a flashlight in the presence of a light to detect a gecko, I classified









the area low light. A list of the criteria of these variables, and their associated levels is

summarized in Table 3-1.

A final variable I recorded for each gecko was the temperature of the center of each

wall. I recorded the temperature using a Raytek Raynger ST model temperature gun.

I also noted the number of geckos on each wall. Appendix A contains the complete

dataset. I calculated averages, standard deviations and percentages where relevant.

Results

With respect to construction material, the average wall temperature was highest for

brick (26.89 + 2.06 C; n = 16), followed by cement (26.14 + 1.42 C; n = 52), wood

(25.13 + 1.38 C; n = 38), and aluminum (23.84 + 0.99 C; n = 78) (Figure 3-2). The

average wall temperature for the four cardinal orientations is displayed in Figure 3-3.

These average temperatures spanned a smaller range with north (25.17 + 1.89 C, n = 46)

having the highest value, proceeded by east (25.03 + 1.74 C; n = 46), south (24.75 +

1.46 C, n = 47), and then west (24.5 + 1.80 C, n = 45). As summarized in Figure 3-4,

brick walls of any cardinal location had a higher average temperature than any other

construction material, while aluminum walls consistently had the lowest average

temperature. Cement walls followed in second place for north, south and west facing

walls, only to drop to third place for east facing walls. Wood walls, in turn, occupied

third place for northern, southern and western walls, rising to second place for eastern

walls. Brick walls reached their highest temperature on north-facing walls (27.16 + 3.1

C, n = 4), followed by west-facing walls (27 + 2.0 C; n = 4), east-facing walls (26.9 +

1.86 C; n = 4), and south-facing walls (26.42 + 1.92 C; n = 4). Average cement wall

temperatures peaked on western walls (26.37 + 1.77 C; n = 13), and then continually

decreased on northern walls (26.16 + 1.45 C, n = 13), western walls (25.82 + 0.89 C, n









= 13), and eastern walls (24.02 + 1.55 C, n = 13). For wood walls, average wall

temperature was highest on northern walls (25.41 + 1.66 C, n = 10), and then

progressively decreased on eastern (25.17 + 1.29 C, n = 9), western (25.04 + 1.40 C, n

= 9), and southern walls (24.89 + 1.31 C, n = 10) respectively. Aluminum walls attained

their highest average temperature on west-facing walls (23.95 + 1.190C, n = 19),

followed by, in decreasing order, west-facing walls (23.92 + 0.91 C, n = 19), east-facing

walls (23.82 + 1.10 C, n = 20), and south-facing walls (23.65 + 0.77 C, n = 20).

Of the 48 buildings I surveyed, five were brick, ten were wood, 13 were cement,

and 20 were aluminum. All the brick buildings and cement buildings had geckos with

totals of 27 and 40 individuals, respectively. Geckos were absent on one of the wood

buildings, whereas the remaining nine contained 14 individuals. Aluminum buildings

contained 30 individuals, although eight of these buildings had no geckos. I found the

highest average number of H. turcicus on brick buildings with 5.4 geckos/building (27/5).

Cement buildings had the second highest average with 3.08 geckos/building (40/13),

whereas aluminum averaged 1.5 geckos/building (30/20), slightly ahead of wood, which

averaged 1.4 geckos/building (14/10) (Figure 3-5).

I sampled a total of 48 south walls, 47 north walls, 47 east walls, and 46 west walls.

I found 86 geckos on north walls, resulting in an average of 1.83 geckos/wall (86/47).

On south walls I tallied 76 geckos for an average of 1.58 geckos/wall (76/48), whereas

walls oriented toward the east had a total of 49 geckos and averaged 1.04 geckos/wall

(49/47), and west oriented walls had 47 geckos with an average of 1.02 geckos/wall

(47/46) (Figure 3-6).









I did not record the number of walls and/or areas within each type of vegetation

level. However, of the 203 geckos that I recorded, 126 of them were located in areas that

possessed a medium level of vegetation (62%). Meanwhile, areas bordered by low and

high vegetation had considerably fewer geckos with 44 (22%) and 33 (16%), respectively

(Figure 3-7).

I also did not record the number of walls and/or areas of each type of light intensity

level. I recorded a total of 235 geckos for this portion of the study. Of these 235 geckos,

I observed 119 (51%) in areas with high light level and 116 (49%) in areas with a low

light level (Figure 3-8).

Discussion

The materials brick, cement and wood all have similar average wall temperatures.

The small difference in average wall temperature they display becomes irrelevant when

their highly overlapping standard deviations are included. Aluminum possesses a lower

average wall temperature, even when its standard deviation is considered. A wall's

cardinal orientation does not appear to have an effect on average wall temperature as all

four directions have comparable temperatures, especially when assessed with their

standard deviations. Each material has its own pattern of average wall temperature with

respect to cardinal orientation. These results, however, should not be taken at face value,

as the sample sizes are small and the standard deviations overlap each other.

The thermal properties of walls are extremely complex and are only briefly

mentioned here as they are beyond the scope of this study. Heat flow involves a variety

of thermal parameters specific to the surface in question, such as its conductivity,

convection capacity, and radiation constant. These parameters, in turn, are highly

influenced by both climate conditions and the thermal property of the proximate









environment (Nave, 2000). Additional work in describing the thermal habitat of H.

turcicus is essential, as activity in ectothermic animals greatly depends on ambient

temperature (Bartholomew, 1959). Furthermore, Frankenberg (1979) suggested that

nocturnal animals are especially dependent on environmental temperature for activity

since they cannot directly use the sun for thermoregulation.

Brick and cement appear to be popular wall material types, as they both possess

more H. turcicus per building than either aluminum or wood. However, the validity of

these results rests on the assumption that I adequately sampled the buildings. I did not

document the size of the buildings. Also, occurrence on a building constructed of a

particular material type might be a result of other factors such as dispersal constraints

rather than preference. A similar fate befalls the results dealing with the average number

of H. turcicus on walls of different cardinal orientations. Although all four cardinal

orientations share comparable averages, the areas of the walls were never recorded and

thus, the results could be an artifact of this lack in rigor.

Mediterranean geckos appear to prefer medium vegetation levels. Perhaps this

indicates a balance between having sufficient vegetation for cover, but not too much

foliage that it raises the vulnerability to predators. As for light levels, H. turcicus does

not appear to have a preference for a particular light intensity. This was a surprising

result as many researchers have reported that H. turicus has an affinity for lights because

it aids in the capture of insect prey (Capula and Luiselli, 1994; Conant and Collins, 1998;

Davis, 1974; Punzo, 2001a). This finding could imply that these geckos are merely

easier to spot near lights, and that their attraction to light is a conclusion stemming from

unintentional bias of the investigator. However, the reliability of my results is fairly









limited, as I did not record the number of walls and/or areas of each vegetation and light

level. Therefore, the pattern in my data might simply reflect a higher incidence of one

vegetation and/or light level rather than a difference in gecko preference.

Although the majority of these results have limitations, they served the function of

illuminating some potentially interesting patterns, some of which merited further

investigation. In addition, this preliminary survey allowed me to develop a reasonable

and rigorous sampling regime for future work. First, it familiarized me with the cryptic

coloration and secretive nature of the study species. Secondly, it revealed the numerous

logistical considerations of the study site such as safety, accessibility, and the high degree

of variation. Lastly, it encouraged me to make some important improvements on my

sampling methods. Among these changes are the inclusion of wall size measurements,

and the sampling of walls rather than buildings so as to avoid the possibility of counting

the same gecko more than once. Also, to deal with the immense variability resulting

from constant human manipulation, vegetation and light need to have broad and

quantifiable levels. These levels, in turn, must pertain to the overall habitat and not the

immediate vicinity of the gecko, as this vicinity will change with movement. Ultimately,

this initial survey would prove to be key in shaping all aspects of my eventual

microhabitat study.











Table 3-1. Criteria of wall characterization variables used in preliminary survey

Variables Levels Criteria

North *Location of wall on official UF map
Cardil O n South *Location of wall on official UF map
Cardinal Orientation
West *Location of wall on official UF map
East *Location of wall on official UF map
Presence of one dull light** or no
Low light source within a Im radius
of the gecko
Light Presence of at least two dull light**
sources or at least one bright
light*** source within a 1m
radius of the gecko
Al m Physical observation; >50% of
Aluminum
building surface area
Brick Physical observation; >50% of
Mateckal building surface area
Cement Physical observation; >50% of
building surface area
Material
W d Physical observation; >50% of
Cement
building surface area
Physical observation; >50% of
Wood
building surface area
Vertical measurement < 0.51 m
Low
within a Im radius of the gecko
0.51 m Vegetation Medium 1.37 m within a Im radius of the
gecko
Hih Vertical measurement > 1.37 m
Highwithin a Im radius of the gecko
*Maps taken from 2000 Building Information List for the University of Florida, prepared by the UF Physical Plant
Division, and official VA Hospital Engineering maps
** A dull light source is one where a flashlight is still needed to locate a gecko
***A bright light source is one where a flashlight is not needed to locate a gecko













45


Figure 3-1. Depiction of the 45 angle leeway employed in the determination of wall
cardinal orientation


26.89


26.14


25.13


23.84


l-l- -------------------------------
Aluminum Brick Cement Wood
Construction Material

Figure 3-2. Average wall temperature of different construction material


7


W










28




26




E 24
I-



E 22


24.5


West


East


Cardinal Location

Figure 3-3. Average wall temperature of different cardinal locations


West


-- Aluminum
--- Brick
Cement
Wood


East


Figure 3-4. Average wall temperature of different construction materials at different
cardinal locations


25.03


25.17


24.75


20 --


North


South


27.16 26.42 27 26.9
26. 16 26.37
25.82251
25.17
25.41 24.89 25.04 24.02

23.95 *23.92 23.84
23.65


North


South


Cardinal Locations


t---


t---











6



5
U)
0

4
4-
0

n3
E
z

S2



1


0 -




Figure 3-5.


3



2.5



t 2
0
4-
5 1.5


E
z 1


0.5


3.08


Cement Brick Aluminum Wood

Construction Material

Average number of H. turcicus per building of different construction material


1.58


1.02


1.04


0 I -- I I-
North South West East

Cardinal Location

Figure 3-6. Number of H. turcicus on walls of different cardinal locations


A I=











140
126

120


100
0
S80

0
60
E
44
S 40 33


20


0
Low Medium High

Vegetation Level

Figure 3-7. Number ofH. turcicus on walls of different vegetation levels


150 i


0 100


lO
0

o
--
.0
E
z 50


0!
Low High

Light Intensity

Figure 3-8. Number ofH. turcicus on walls of different light intensities














CHAPTER 4
ADDITIONAL NATURAL HISTORY NOTES

Introduction

A large portion of the information regarding H. turcicus pertains to the

reproductive cycle and associated activities. Few studies have documented natural

history variables such as perch height, degree of sociality and exposure. And not much is

known about any preferred temperature regime.

Perch height in H. turcicus has only rarely been investigated. The only

comprehensive study was conducted on a university campus in Texas, where Saenz

(unpublished) found substantial dietary diversity among geckos at different perch heights.

The diet ofH. turcicus encountered below 1.52 m in height overlapped only 22.71%

(Schoener's percent overlap; Schoener, 1970) with conspecifics occupying a perch over

3.05 m in height. Specifically, the geckos with lower perches ingested mostly ground-

dwelling prey, whereas those at higher perches fed predominantly on flying insect taxa.

In general, females tended to use perches of lower height than males, as 55.8% of females

were recorded below 1.52 m in comparison to 30.2% of males. Meanwhile, 41.5% of

males were recorded above 3.05 m as opposed to only 11.6% of females. Geckos

captured at different perch heights also demonstrated a difference in the number of empty

stomachs, with 13.04% for low geckos and 25% for high geckos (Saenz, unpublished).

Other perch height studies include one by Capula and Luiselli (1994) in Rome,

Italy. These authors concluded a close-to-the-ground existence for H. turcicus, as they

showed that its diet consisted mainly of ground-dwelling prey, with 55.2% of the gecko's









diet being made up of ants and flightless insects. In addition, Klawinski (unpublished)

observed a large number of Mediterranean geckos positioned close to the ground

(40.95%), a result he felt stemmed from the need for shade from surrounding lights. In

New Orleans, Rose and Barbour (1968) noticed several geckos on the third-floor level of

a building, and also on the roof of another.

Both Selcer (1986) and Klawinski (unpublished) recorded low average home range

areas of 0.93 m2 and 4.073 m2, respectively, with very little home-range overlap. These

results suggest that this species is territorial. Furthermore, observations have indicated

that male H. turcicus emerge from winter retreats earlier then females, perhaps to

establish territories before the breeding season (Klawinski, unpublished). In addition,

Rose and Barbour (1968), Frankenberg (1982), Marcellini (1977), Klawinski

(unpublished) have all witnessed aggressive displays ranging from tail waving to neck

biting. A study of the vocal activity ofH. turcicus revealed that only the dominant male

in a group produces a multiple click call in response to an intruder of either sex

(Frankenberg, 1978). In this same study, Frankenberg (1978) found that most of the

vocalization in H. turcicus occurred during the day, a time when this gecko is grouped

together in retreat-sites. This result indicates that social activity is perhaps separate from

this species' nocturnal foraging. In turn, this division between sociality and foraging

could explain the peaceful interaction between two male H. turcicus behind a drainpipe

witnessed by Rose and Barbour (1968).

The degree of exposure once H. turcicus has emerged from its daytime retreat has

not been formally studied. Rose and Barbour (1968) observed H. turcicus behind vertical









storm drains. Through personal observations, I have noted geckos behind electrical

boxes, pipes, and signs on walls.

Few studies have focused on H. turcicus with respect to temperature. An exception

is the study by Angilletta et al. (1999) where the body temperature for eight

Mediterranean geckos was measured to be 27.8 C in the morning, and 29.1 C in the

evening.

Hoping to shed further light on the natural history of the Mediterranean gecko, I

explored perch height, sociality, exposure, and selected surface temperature in three

distinct field surveys.

Methods

The first survey took place during the months of July and August 2001 whereby I

randomly sampled walls on 48 one-story buildings on the University of Florida campus

and the VA Hospital. I systematically sampled each wall by scanning it with a flashlight

from top to bottom, left to right. I classified each gecko I encountered as either an adult

(greater than 40 mm) or subadult (less than 30 mm). Note that I omitted any geckos that

I could not accurately size from any analysis. Lastly, I recorded the perch height of each

gecko using two categories; low if the gecko was at most one meter above the ground,

and high otherwise. Appendix B contains the complete dataset.

In the second survey, I sampled 50 one-story walls located on the University of

Florida campus and the VA Hospital. Between September 2001 and January 2002, I

sampled each of these 50 walls on a weekly basis. Upon each visit, I again examined a

wall by methodically examining it with a flashlight from top to bottom, left to right.

Geckos were sized according to the method mentioned previously. In addition to

documenting perch height (as above), I also quantified sociality, exposure, and selected









surface temperature. I used three categories to describe sociality: a gecko that had no

other individual within a 50 cm radius was termed alone, while two geckos within the

same radius were designated a pair, and three or more geckos were considered a group.

Furthermore, I labeled a gecko as exposed if it was in plain sight, and not exposed if the

snout-vent portion of its body was hidden behind a wall fixture. Using a Raytek Raynger

ST model temperature gun, I determined the selected surface temperature by measuring

the temperature of a spot adjacent to the gecko. Appendix B contains the complete

dataset.

For the third survey, also on the University of Florida campus and the VA Hospital,

I sampled 160 one-story walls between the months of March and June 2002. I visited

each wall twice throughout the study. The sampling regime and equipment I used were

identical to those previously mentioned in the second study. Appendix B contains the

complete dataset. A list of the criteria of these variables, and their associated levels is

summarized in Table 4-1.

I calculated percentages, averages, standard deviations, maximum values and

minimum values where appropriate.

Results

Summer 2001

During this study, I recorded a total of 187 gecko observations. Of these, 125

(67%) occurred at a high height, whereas 62 (33%) were at a low height. Upon

separating these observations into adults and subadults, two distinct patterns emerged. Of

the 131 adults I sampled, 97 (74%) were located high on walls and 34 (26%) were

observed at a low height. The opposite was true for subadults, as 38 (84%) out of the 45









recorded were found within one meter of the ground whereas the remaining 7 (16%) were

located at a high height (Figure 4-1).

Fall/Winter 2001

I used 576 gecko observations to test for perch height preference; 412 (71.5%)

occupied high positions, whereas 164 (28.5%) occupied low positions. Of these 576

gecko observations, 355 were adults of which 296 (83.4%) were at a high position and 59

(16.6%) were at low positions. I recorded 221 subadult observations, where 116 (52.5%)

were at a high position, and 105 (47.5%) were at a low portion of a wall (Figure 4-2).

I used 577 gecko observations to investigate sociality. Of these, 506 (87.7%)

geckos were alone, 56 (14.9%) were part of a pair, and 15 (4.2%) were part of a group. I

counted 355 adult gecko observations where 287 (80.9%) were alone, 53 (14.9%)

belonged to a pair, and 15 (4.2%) belonged to a group. Following a comparable trend,

the 222 subadult observations resulted in 219 (98.7%) alone counts and 3 (1.3%) pair

counts. No (0%) subadults were seen in any groups (Figure 4-3).

Of 574 exposure observations, I tallied 397 (69.2%) as being exposed and 177

(30.8%) as not exposed. Further breakdown of these results revealed that 199 (56.7%) of

the 351 adult observations were exposed, whereas 152 (43.3%) were not exposed. Of the

223 subadult observations I recorded, 198 (88.8%) fell in the exposed group, which

contrasts with the 25 (11.2%) observations that I placed in the not exposed group

(Figure 4-4).

As summarized by Table 4-2, the average substrate temperature for adults was

23.22 C (+ 3.05), with a maximum value of 36.3 C and a minimum value of 12.7 C.

Likewise, the average substrate temperature for subadults was 22.67 C (+ 2.80), with the

maximum and minimum values being 30.8 C and 13.9 C, respectively.









Spring 2002

I observed few subadults in this study, so I included only adult observations. With

respect to perch height, I obtained 237 observations, with 170 (71.7%) at high positions

and 67 (28.3%) at low positions (Figure 4-5).

For sociality, I had 236 observations, of which 223 (94.5%) were alone, 13 (5.5%)

belonged to a pair, and none (0%) were part of a group (Figure 4-6).

I collected 237 exposure observations, of which 178 (75%) fell into the exposed

category, and 59 (25%) into the not exposed category (Figure 4-7).

The average substrate temperature for adults was 24.89 C (+ 2.94). The maximum

temperature value I measured was 31.5 C, whereas the minimum value was 16.4 C

(Table 4-3).

Discussion

Adult H. turcicus consistently occupied wall habitats that were greater than one

meter above the ground. This trend was observed regardless of the time of year.

Although not quantified, the majority of the "high" sightings were located in the upper

portion of the walls, in close proximity to the roof awning. This high perch height could

be beneficial for escaping predators, as I repeatedly witnessed startled geckos escape into

crevices in the roof awning. However, it is important to mention that this result could be

an artifact of the difference in surface area between the two height categories; the low

category is restricted to a significantly smaller area than the high category. Thus, the

greater number of adults located in high wall positions may be directly related to the

greater area available.

Interestingly, despite this disparity in area, subadults were recorded on wall habitats

that were a maximum of one meter above the ground. This result was more pronounced









during the summer survey, although it was still evident during the fall/winter sampling

period.

This tendency for subadults to occupy a low habitat could stem from a variety of

reasons. First, the portion of a wall near the ground might not be an optimal habitat as it

could increase a gecko's vulnerability to predators. Subadults might be obliged to use

this less desirable habitat as a result of being out-competed by adults for the high optimal

ones. This scenario would be congruent with the significantly higher mortality rate found

in geckos measuring less than 30 mm (Selcer, 1986).

Secondly, the subadult age period could be the dispersal stage of H. turcicus. If

subadults were the dispersers, they would frequently be on the lower portion of a wall as

they would be continuously on the move. This idea might be supported by some

circumstantial evidence that I have witnessed during the course of this study; on several

occasions I have observed subadults on the ground some distance away from any

building/wall. In fact, one particular individual was recorded in the middle of an

expansive cement parking lot. Also, Rose and Barbour (1968) showed that hatchlings

could survive without food or water for up to one month. This resilient quality of

subadults would be ideal for the uncertainties of dispersing. Furthermore, this idea

would be compatible with Selcer's findings (1986), as dispersal would be expected to

make subadults more vulnerable to predation and thus increase their mortality rate.

Thirdly, subadult preference for low wall habitats might be a consequence of diet.

Perhaps the preferred prey of subadults and/or prey size suitable for small mouths is more

abundant on low dwellings. This is highly possible, as Saenz (unpublished), upon

conducting a detailed dietary study on H. turcicus, concluded that the gecko's height on a









wall greatly influenced a gecko's diet. Finally, the actual cause behind subadults

occupying a low habitat could be any one of the three mentioned hypotheses, or a

combination of these, or even still, none of these.

During the summer survey, a greater proportion of subadults in my sample used

perches at a low height than during the fall/winter survey. This discrepancy could be

attributed to a number of factors such as seasonal changes in prey consumption, a

decrease in adult competition due to adult turnover and/or the decrease of breeding

activities, an increase in the establishment of subadults on walls due to a decrease in

dispersal, an increase in subadult population, a combination of these, or none of the

above.

Sociality in the Mediterranean gecko during nocturnal foraging appears to be quite

minimal. Both adults and subadults preferred being alone. This result is in agreement

with the belief, held by many investigators, that H. turcicus is largely territorial. The

majority of the pair and group observations involved only adults. Group sizes rarely

exceeded three. I observed no aggressive display in any of these observations. Although

no copulation was witnessed, perhaps the gecko pairings were associated in some way

with breeding. Gecko groupings could also be linked to other stages of reproduction, as

communal nesting has been documented in this species (Selcer, 1986). Since geckos

were not sexed in this study, there is no way of telling if the pairs and groups consisted of

same or different sex individuals. The tendency for H. turcicus to be solitary does not

necessarily imply that this gecko species is not social. It has been shown that other

nocturnal gecko species (Nephrurus milii and Christinus marmoratus) form large, non-

random aggregations within retreat-sites (Kearney et al., 2001). The possibility that H.









turicus displays this behavior has been alluded to by Frankenberg (1978), who showed

that most vocalization in this species occurs in daytime retreats. Thus perhaps, H. turicus

socializes during the daytime within retreat-sites, and then forages in solitude during the

night, occasionally interacting with others for reproductive purposes and/or for prey

exploitation.

In general, the Mediterranean gecko remains exposed once it emerges from its

retreat-site at night. Since the number of possible hiding places is extremely difficult to

quantify and/or locate, this result could reflect my inability to find all hidden geckos and

thus be skewed. In addition, this result could also simply be a product of availability of

hiding places rather than preference. Subadults showed a greater propensity for

remaining exposed than adults. This difference might stem from adult competition for

hiding spaces, which could be intense if these spaces offer a significant increase in

predator protection. This subadult trend might also be an artifact of a lower incidence of

hiding places situated in the bottom portion of walls.

The field temperatures of the wall surface selected by both adult and subadult

H. turcicus were comparable. Although these temperatures are lower than previously

measured field body temperatures (29.1 C, n = 8), it is not entirely surprising as some

species of nocturnal geckos have been known to thermoregulate and achieve their

preferred body temperature during the day in their retreats rather than at night (Angilletta

et al., 1999).

An important point when considering these results is the high possibility of

pseudoreplication in the data. For the fall/winter 2001 survey, I visited the same 50 walls

every week, whereas in the spring 2002 survey I visited the same 160 walls twice. These






30


two sampling methods did not allow me to distinguish among individuals, and it is likely

that I counted the same individual many times. Thus, conclusions should be made with

caution. However, despite the fact that these results are limited in scope, they

nevertheless provide some insight for future investigators.











Table 4-1. Criteria of natural history variables
Variable Levels Criteria
Gecko < Im above the
Low
Height ground
H Gecko > Im above the
High ground
ground
Alone 1 gecko within a 50cm
Alone
radius
2 geckos within a 50cm
Sociality Pair
radius
3 or more geckos within
Group
Group a 50cm radius
Exp d Snout-vent portion of
Exposed
gecko in plain sight
Exposure Snout-vent portion of
posd gecko not in plain
Exposed sight
sight


Table 4-2. Temperature measurements of wall
fall/winter 2001


surface for adult and subadult H. turcicus;


Temperature Adult Subadult
Measurements

Average 23.220C 22.670C

Standard Deviation 3.050C 2.800C

Maximum Value 36.300C 30.800C

Minimum Value 12.70C 13.900C


Table 4-3. Temperature measurements of wall surface for adult H. turcicus; spring 2002
Temperature Adult
Measurements

Average 24.890C


Standard Deviation 2.940C


Maximum Value 31.50C


Minimum Value 16.40C












100%

90%

80%

e 70%

60%
O
Ch Low
._ 50%
SHigh
CD 40%

30%

20%

10%

0%
Adult Subadult
Perch Height

Figure 4-1. Perch height preference of adult and subadult H. turcicus (summer 2001)













100%-

90%

80%

70%

60%

50%

40%

30%

20%

10%

0%


Adult


SHigh
SLow


Subadult


Gecko Age

Figure 4-2. Perch height preference of adult and subadult H. turcicus (fall/winter 2001)












100% 4.2% 1.3%

90%

80%

70%

60%
O 6E Group
o 50% I Pair
gAlone
0. 40%
30%

20%

10%

0%
Adult Subadult

Gecko Age

Figure 4-3. Social preference of adult and subadult H. turcicus (fall/winter 2001)














100%

90%

80%

70%

60%

50%

40%

30%

20%

10%

0%


Subadult


Gecko Age

Figure 4-4. Exposure preference of adult and subadult H. turcicus (fall/winter 2001)


Adult


I


* Exposed
* Not Exposed










100%

90%

80%72%

r 70%

60%
0
U)
o 50%

40%
S28%
o 30%

.20%

10%

0%
High Low
Perch Height


Figure 4-5. Perch height preference of adult H. turcicus (spring 2002)











100%
94.5%
90%

80%

70%

60%

50%
UC


o 40%

30%

20%
5.5%
10%
0% 0.0%

Alone Pair Group
Sociality


Figure 4-6. Social preference of adult H. turcicus (spring 2002)













100%

90%

80%75%

70%

.o 60%
0
0
5 50%

S40%
25%
30%

20%

10%

0%
Yes No

Exposure

Figure 4-7. Exposure preference of adult H. turcicus (spring 2002)














CHAPTER 5
MICROHABITAT PREFERENCE IN THE INTRODUCED GECKO,
HEMIDACTYLUS TURCICUS

Introduction

Although frequently seen on buildings in both its native and nonnative range, little

is known of the microhabitat preferences ofH. turcicus. Habitat studies on this species

are largely nonexistent, and the little information available has mostly originated as

incidental observations made during other studies.

The only direct study, conducted in Rome, Italy, was a comparative study between

H. turcicus and the sympatric gecko species Tarentola mauritanica. In this study,

Luiselli and Cappizzi (1999) found that H. turcicus was more abundant on recently

constructed buildings than debilitated ancient buildings dating back to the Roman

Empire. Although this study provided some quantitative information on microhabitat

preference in H. turcicus, the confounding effects of competition influenced the

conclusions. Evidence of this is demonstrated by a survey Capula and Luiselli (1994)

conducted in Rome years earlier, which concluded that H. turcicus was particularly

common in Roman age archeological sites.

Additional information consists primarily of incidental observations, and is

anecdotal in nature. Although the natural habitat ofH. turcicus may have been rocky

cliffs, the primary habitat today appears to be structures associated with human habitation

(Arnold, 1984). Investigators have reported seeing H. turcicus on a number of human-

made constructions such as rock walls, burial vaults, and buildings of varying material









including granite, cement, wood, metal, and stucco (Davis, 1974; Klawinski,

unpublished; Meshaka Jr., 1995; Punzo, 2001a; Rose and Barbour, 1968; Saenz,

unpublished; Selcer, 1986). In Texas, Selcer (1986) found that H. turcicus occurred at

higher densities on brick versus metal structures, whereas in Florida Punzo (2001a) found

higher densities of H. turcicus on wood as opposed to metal buildings.

Investigators have always linked artificial light to the presence ofH. turcicus on

buildings, as lights presumably facilitate the capture of their insect prey (Capula and

Luiselli, 1994; Conant and Collins, 1998; Davis, 1974; Punzo, 2001a). In Texas, Davis

(1974) reported that H. turcicus preferred buildings that were lit by mercury-vapor lights.

However, a number of studies have collected H. turcicus from buildings with varying

light intensities, including complete darkness (Klawinski, unpublished; Meshaka Jr.,

1995).

Vegetation is another factor that has often been associated with Mediterranean

gecko habitat. Throughout its native range, H. turcicus has been found commonly on

trees (Loveridge, 1947). In its introduced range, buildings inhabited by H. turcicus have

possessed grass, shrubs and/or trees in close proximity to its walls (Klawinski,

unpublished, Saenz, 1996). Vegetative cover has been hypothesized by Saenz (1996) to

provide H. turcicus with retreats. In a separate study, Klawinski (unpublished) found a

weak association between the occurrence ofH. turcicus on walls with both high light

intensity and high vegetative cover.

The majority of these findings are qualitative observations. The few results

supported by quantitative data often lack a rigorous framework, making conclusions

difficult to formulate. Thus, I embarked on a systematic study of microhabitat









preference, in an attempt to contribute quantitative baseline data on the Mediterranean

gecko. Specifically, I investigated the previously studied variables: construction

material; vegetative cover; and light intensity. Furthermore, I examined cardinal

location, building age, surface texture, surface color, and wall length, as I considered

these characteristics to be potentially important to the habitat preference of this species.

Methods and Results

Sampling Methods

Between the months of March and June of 2002, I conducted a survey detailing the

microhabitat preference of H. turcicus on the University of Florida campus and VA

Hospital in Gainesville. I selected 160 buildings according to their accessibility, and, for

the sake of accurately detecting geckos, being one-story in height. I sampled only one

randomly selected wall per building. I used a die to make my selection; the numbers 1, 2,

5, and 6 indicated north, south, west and east, respectively, whereas 3 and 4 denoted

rolling the dice over again. Walls were considered a good representation of microhabitat

use in this species, as H. turcicus has been shown to possess a small home range; Rose

and Barbour (1968) reported an average recapture distance from initial capture site of

5.7 m, while Selcer (1986) estimated the mean range movement to be 0.93 m.

Furthermore, Trout and Schwaner (1994) reported that H. turcicus maintains itself in

discrete subpopulations in which differences in allele frequency have been found between

populations only 100 m apart. Each selected wall was characterized by building material,

presence or absence of light sources, surface color, surface texture, age of building,

vegetation level, cardinal orientation, and length.

Construction material was confined to four types: aluminum, brick, cement, and

wood. Identical to my preliminary survey, I arbitrarily used 50% of the surface area of a









given wall for my material classification. I restricted my survey to walls that featured a

predominate material.

I used two categories of light intensity, high and low. To account for gecko

movement, light intensity was measured with respect to the wall as opposed to a small

area around a gecko. Specifically, I classified walls as high light if they possessed at

least one light source (any brightness), whereas walls containing no light source I

assigned to the low light category.

I classified wall color into two levels using a 3-inch by 5-inch white index card. I

categorized a wall as dark if, upon fastening the index card to the wall, I could distinguish

it at a perpendicular distance of 10 feet (3.048 m) in daylight. The opposite was true of

walls appointed to the light level; the index card could not be perceived at a

perpendicular distance of 10 feet. All the walls I surveyed were uniformly colored across

their entire surface.

I also quantified wall texture using two levels. I considered a wall to be smooth if I

was able to draw a straight line, roughly10 cm in length, on white printer paper propped

on three distinct points on a wall. These three positions were subjectively selected as the

left edge, right edge, and middle point of the wall, approximately mid-wall in height. If I

was unable to draw a straight line at all three points on a wall, I classified the wall as

rough. Note that all lines were drawn with a relaxed handgrip.

I determined building age information from literature provided by the University of

Florida (UF Physical Plant Division, 2000), and an unofficial list created specifically for

this study by the Veterans Administration Hospital Engineering Department. I arbitrarily

assigned three age groups, all based on these sources. I categorized walls built between









1900 and 1969 as early, between 1970 and 1989 as modern, and between 1990 and 2002

as contemporary. I ignored any possible renovations. The age of a building was used as

an approximation for the number of daytime retreats (cracks and/or crevices) since

Luiselli and Capizzi (1999) found that the age of a building and the condition of its walls

were highly correlated. The necessity of using this approximation arose when, during my

preliminary survey, it became apparent that estimating the number of retreats with the

naked eye was highly unreliable.

I classified walls into one of three vegetation levels. The three vegetation levels

were based on the cement/vegetation ratio bordering the wall, rather than the diversity or

height of the vegetation. I quantified the vegetation in this fashion because of the

unpredictable management techniques encountered during my preliminary survey;

primarily, workers constantly altered vegetation variety and height. Thus, the cement

level referred to a wall where at least 60% of the length was bordered by cement, the mix

level to a wall whose length was bordered more than 40% but less than 60% by either

cement or vegetation, and the vegetation level to a wall whose length was bordered at

least 60% by vegetation.

I determined the cardinal orientation of walls using the same methods discussed in

my preliminary survey. Using the 2000 Building Information List for the University of

Florida and official maps of the VA Hospital Engineering Department, I classified walls

as north, south, west, or east. For walls that were not clearly oriented in one of these

directions, I allowed a 45 angle of leeway on each side of the cardinal direction. Figure

3-1 in chapter 3 has further details.









I described length by means of three randomly created categories. I classified walls

as small if their greatest length measured less than 20 m, as medium if their length was

between 20 m and 40 m, and as large if they were 40 or more meters in length. Length

was used as a general measure for size, since all walls had roughly the same height (one-

story). A list of the criteria of these variables, and their associated levels is summarized

in Table 5-1.

My sampling regime consisted of visiting 10 walls per night, on two nights per

week. Each visit occurred approximately two hours after sunset, which according to

King (1959) is a period of high activity for H. turcicus in Gainesville. The number of

walls I inspected per night was limited to 10 to keep sampling duration under two hours.

This was done to homogenize weather conditions among walls. The four months I

selected for the survey period coincided with part of the reproductive season of H.

turcicus and further ensured gecko activity (Selcer, 1986). Thus, I examined each of the

160 walls twice for completeness, once during March/April and once during May/June. I

sampled each wall by passing a flashlight systematically across the entire surface, going

from right to left, top to bottom. I recorded the presence or absence ofH. turcicus for

each wall. I pooled data from the two visits, and considered a wall to have a gecko if at

least 1 gecko was present during at least one of the two visits. The complete dataset is

presented in Appendix C.

Data Analysis 1

The data were initially analyzed using either a chi-square test or a Fisher's exact

test in order to detect dependency among the eight variables of interest. The Fisher's

exact test is based on the hypergeometric distribution, rather than a chi-square

distribution; when testing for independence, the p-value is obtained by adding the









probabilities of outcomes as favorable to the alternative hypothesis (dependence) as the

observed outcome (Agresti, 1996). I then used logistic regression in an effort to model

the wall variables to the presence/absence ofH. turcicus on a wall. I selected logistic

regression because it functions on binary data, which was the format of the data I had

collected. Only those variables deemed independent from the chi-square tests were used

as predictors for the logistic regression in an attempt to avoid multicollinearity in the

model.

I determined the logistic regression model by using the backward elimination

method (Agresti, 1996). In this method, one essentially begins with the full model,

containing all possible variables and the interactions between them, and then

systematically removes one term at a time, starting with the highest-order term. With the

removal of each term, the deviation (G2 test of goodness of fit) of the new model is

compared to that of the full model. Removals continue until the difference in deviation

between the two models either reaches a specified value determined by the investigator

and/or the difference reaches a large enough value that it becomes significant. Once

significance is attained, further term removal would result in losing the integrity of the

information provided by the dataset. In other words, this method is a balance between

simplifying the model, and preserving a sufficient amount of the dataset information.

Therefore, the number of retained terms is directly related to the objective of the

investigator (Agresti, 1996). In this case, I decided to choose integrity of information

over simplification, and thus I elected to use the simplest model that possessed the

smallest difference in deviation from the full model.









Once I had chosen an appropriate model, I calculated the odds ratio and a 95%

confidence interval of the odds ratio for each significant predictor. The odds ratio (0), as

defined by Agresti (1996) is the ratio of the odds of two events (0 = odds,/ odds2), where

the odds of each event is defined as the odds of success for that event. Thus for example,

if the probability of success for event 1 is 0.75, then odds= probability of

success/probability of failure = 0.75/1-0.75 = 0.75/025 = 3. This signifies that success in

event 1 is three times as likely as failure. Finally, if odds= 3 and odds2 = 4, then the

odds ratio 0 = 3/4 = 0.75, which indicates that the odds of success for event 1 is 0.75 the

odds of event 2. To put it into perspective, if the odds of finding a gecko on aluminum

walls is 3, while the odds of finding a gecko on wood walls is 4, then the odds ratio of

aluminum walls to wood walls would be 0.75; this signifies that the odds of finding a

gecko on aluminum walls is 0.75 that of finding a gecko on wood walls. The odds ratio

for each predictor was calculated by taking the exponent of the predictor's estimate (eB,

where B is the predictor estimate). I then calculated a 95% confidence interval for each

odds ratio by finding the lower and upper bounds, and then taking their exponents; the

bounds were found using the equation Bi +/- 1.96 (ASE), where Bi denotes the estimate

of the predictor in question and ASE stands for the estimate's asymptotic standard error.

Thus, the 95% confidence intervals take the following form: (eB-196 (), eBi+.96 (ASE)

(Agresti, 1996). In all of my statistical analyses, the significance value was set at the

0.05 level. All statistical analyses were performed using SAS, version 8.2.

Results 1

The chi-square and the Fisher's exact analyses revealed a number of associations

among the variables. Specifically, material was highly dependent on building age, color,









length, and texture. Furthermore, light was dependent on both length and vegetation,

while color was dependent on both age and length. Table 5-2 contains additional

information, including p-values. To avoid multicollinearity in the eventual model, I

omitted the variables age, color, length, and texture from further analysis since they could

be accounted for by the variable material. Table 5-3 describes the relationship between

material and age, color, length, and texture.

I fitted a logistic regression model using the variables cardinal orientation (CO),

light (L), material (M), and vegetation (V). The backward elimination method supported

the use of the interaction variables M*CO*V + M*CO*L model instead of the full model

M*CO*V*L. The M*CO*V + M*CO*L model explained the same amount of variability

as the full model (the difference in deviation was zero), but with fewer terms, thus

making it easier to interpret. Table 5-4 presents additional details on the backward

elimination method.

The expanded, symbolic version of the logistic regression model M*CO*V +

M*CO*L is the Equation 5-1.

(Eq. 5-1) Y =M*CO*V + M*CO*L + M*CO + M*V + M*L + CO*V + CO*L +
V*L +M+CO + V+L

The M*CO*V + M*CO*L model is modeling the probability that y = 1, i.e. that H.

turcicus is present on a wall. Table 5-5 contains the estimate, asymptotic standard error,

p-value, odds ratio, and 95% confidence interval of the odds ratio for all of the significant

predictors. The numerical representation of the model is the Equation 5-2.

(Eq. 5-2) Y= 68.3825 68.3825m1 91.7478m2 46.1655m3 + 18.3578v2 +
109.586coi 45.0170co2 67.283911 + 131.7139 mlcoi + 44.3238 mico2
17.0373m2co3 18.3581v2 c02
Where


mi = 1 = Aluminum, and 0 otherwise










m2 = 1 = Brick, and 0 otherwise
m3 = 1 = Cement, and 0 otherwise
m = m2 = m3 = 0 = Wood

vi = 1 = Cement, and 0 otherwise
v2= 1 = Mix, and 0 otherwise
vl = v2 = 0 = Vegetation

col = 1 = East, and 0 otherwise
co2 = 1 = North, and 0 otherwise
co3 = 1 = South, and 0 otherwise
co = co2 = 3 = 0 = West

11 = 1 = No light, and 0 otherwise
11 = 0 = Light


However, parameter estimates of this model were unusually high or unusually low,

resulting in equally extreme odd ratios. This suggests that despite this model being the

"best" one my input could generate, it was inadequate as the estimates were too

unrealistic. Although there is no written rule on estimate magnitude, the magnitudes of

my estimates were so large as to render them unpredictable and useless (Ken Portier,

personal communication). In general, I have noticed odds ratios take values between

zero and five, occasionally larger but never greater than 10. Thus, I recommend using a

value of 10 as a cut-off point for odds ratios. Accordingly, anything above 10 should be

used with caution, and definitely examined further. With respect to a low boundary for

odds ratio values, the mathematical minimum is zero. However, values close to zero that

take the form of a low order decimal should also be considered with caution, and

investigated further.

Data Analysis 2

In an attempt to determine if the difficulty encountered in my first analysis was a

result of the model selected, I fitted all of the possible models, even those that would

explain less overall variability. Specifically, these were










M*CO*V+ M*CO*L+ M*V*L + CO*V*L
M*CO*V+ M*CO*L+ M*V*L
M*CO*V+ M*CO*L + CO*V*L
M*CO*V+ M*V*L + CO*V*L
M*CO*L + M*V*L + CO*V*L
M*CO*V+ M*V*L
M*CO*L + M*V*L
M*CO*L + CO*V*L
M*CO*V
M*CO*L
These models correspond to #2, #3, #4, #5, #6, #8, #9, #12, #13, and, #14 in Table 5.4.

In addition, I constructed a table of all the wall combinations I encountered during

my survey in order to verify that my dataset was appropriate for logistic regression; too

few observations per wall combination could cause the model to produce unrealistic

estimates.

Results 2

All 10 additional fitted models produced similar estimate and odds ratio values,

either extremely high or extremely low in magnitude. In other words, these 10 models

were inadequate too. Thus, this indicated that model selection was not the source of the

problem.

A possible cause of the problem could have been the large number of zeros and low

among the various wall combinations (Table 5-6). The wall combination table revealed

that of 96 possible wall combinations, 32 were never encountered during this study, and

an additional 32 combinations occurred just once. Indeed, only 10 combinations of the

96 possible were represented by more than five occurrences, although none over 10.

These results strongly point toward the conclusion that this dataset is not sufficiently

large enough to accommodate logistic regression (Agresti, 1996; Ken Portier, personal

communication).









Data Analysis 3

Following a common statistical practice, I collapsed the wall combination table into

a smaller, more concise table with fewer variables and/or levels. This can be

accomplished most simply by removing and/or combining. Therefore, I removed

cardinal orientation, as it was the least repeatable of all the variables; compass readings

are linked to the magnetic poles, which are constantly changing locations (Natural

Resources Canada, 2003). Furthermore, I decided to combine the cement and mix

vegetation levels to increase cell numbers and eliminate zeros within the table. Thus,

vegetation was described by two levels; the cement level referred to walls whose length

were bordered more than 40% by cement, whereas the vegetation level referred to walls

whose lengths were bordered at least 60% by vegetation of any species and height.

Although there existed a number of distinct ways I could have collapsed my wall

combinations table, I believe that the arrangement I selected was ecologically, the most

parsimonious one. I then fitted a logistic regression model with the backward elimination

method. I used the variables material (M), light (L), and the newly described vegetation

(V). Estimates and odds ratios for parameters were calculated where pertinent.

Results 3

The wall combination table resulting from the removal of cardinal orientation and

the merging of two vegetation levels was significantly condensed; instead of 96 cells, this

new table totaled only 16 (Table 5-7). This decrease in the number of cells eliminated

zeros from the table, although six of the remaining 16 (37.5%; 6/16) values were less

than five.

The backward elimination method resulted in 16 logistic regression models,

ranging from the highly complex three-factor interaction model to the simple no-factor









model (Table 5-8). Theoretically, all 16 models were found to be functional, each having

its own balance between simplicity and the amount of variability explained. Despite all

models being usable, only the model M*V*L contained significant parameters.

However, these parameter estimates and their resulting odds ratios were again found to be

unusually high or unusually low. Although the remaining models had typical parameter

estimates and odds ratios, none of the parameters were significant; in other words, these

models would contain zero parameters, since only significant parameters are included in

a model, thus these models were impractical. These results show that logistic regression

models continue to crash with these data.

Data Analysis 4

In a final attempt to obtain a successful logistic regression model, I combined the

light and vegetation variables into one variable with three levels. Specifically, I pooled

both vegetation levels under the high light level, and left the two vegetation levels under

the low light level unchanged. Thus, I described the new light-vegetation variable

(LVEG) as HCV (high/cement-vegetation) if a wall was characterized by high light and

any kind of vegetation, as LC (low/cement) if a wall was classified as low light and

cement, and as LV (low/vegetation) if a wall was classified as low light and vegetation.

From an ecological perspective, this merging assumes that in the presence of a light

source, vegetation is irrelevant to choice of walls by a gecko. Conversely, in the absence

of light, vegetation becomes an important consideration in choice of walls by a gecko.

For example, the overall significance could be that light sources always attract insect prey

and thus render vegetation level irrelevant; whereas in the absence of light, vegetation

might dictate the type and/or amount of insect prey and thus become a key component in

microhabitat choice.









I then fitted a logistic regression model using the variables material (M) and the

newly formed light-vegetation (LVEG). I used the backward elimination method for

model fitting.

Results 4

Further variable condensation produced a wall combination table with no zeros or

ones, with only two values smaller than five (Table 5-9).

The backward elimination method generated five functional logistic regression

models (Table 5-10). However, none of the models contained significant parameters.

Consequently, these models were unusable.

Data Analysis 5

A final analysis was performed on these data. I computed chi-square tests for

material and gecko presence, light-vegetation and gecko presence, material and gecko

presence while controlling for the light-vegetation variable, and light-vegetation and

gecko presence while controlling for the material variable. I used the Fisher's Exact test

instead of the chi-square test for small samples. Lastly, I calculated relevant percentages

for these data. Significance was set at the 5% level.

Results 5

Using the Fisher's exact test, light-vegetation was found to be independent of

gecko presence at the 5% significance level when controlling for material (Table 5-11).

Significance values for light-vegetation and gecko presence were 0.4098 for aluminum,

1.0000 for brick, 0.2416 for cement, and 0.6431 for wood. The cell chi-square values

revealed that fewer cement walls of high light and any vegetation type had geckos than

expected. Although this result is not significant, it points to a potential trend.









When light-vegetation was controlled, the Fisher's exact test concluded that

material and gecko presence were independent at the 5% significance level (Table 5-12).

Specifically, significance values were 0.8594 for high light/cement-vegetation, 0.6111 for

low light/cement, and 0.2642 for low light/vegetation. A possible trend was also

uncovered via the cell chi-square values; there was a greater number of aluminum, low

light/vegetation walls that contained geckos than would be expected by chance.

Two-way chi-square tests concluded that both light-vegetation (Table 5-13) and

material (5-14) were independent of gecko presence at the 5% significance level. The

significance value for light vegetation was 0.1402, whereas that of material was 0.2281.

General, yet non-significant, trends included fewer high light/cement-vegetation walls

inhabited by geckos, and a greater number of gecko-populated aluminum walls than

predicted by randomness.

These analyses revealed that 34% (13/38) of high light/cement-vegetation walls,

48% (40/83) of low light/vegetation walls and 56% (22/39) of low light/cement walls

contained geckos (Table 5-15). With respect to aluminum walls, 42% (5/12) of high

light/cement-vegetation walls, 63% (17/27) of low light/vegetation walls and 71% (5/7)

of low light/cement walls recorded gecko presence. Following a similar pattern of gecko

occurrence, cement and wood walls achieved the low values of 22% (2/9) and 25% (1/4)

respectively when described as high light/cement-vegetation, followed by 44% (12/27)

and 33% (5/15) when characterized as low light/vegetation, and peaked at 57% (13/23)

and 67% (2/3) when categorized as low light/cement. When considering brick walls, in

turn, 33% (2/6) of high light/cement-vegetation walls, 39% (5/13) of low light/cement

walls and 43% (6/14) of low light/vegetation walls recorded geckos.









With respect to material, 59% (27/46) of aluminum walls, 46% (27/59) of cement

walls, 39% (13/33) of brick walls, and 36% (8/22) of wood walls had geckos

(Table5-16). When controlling for high light/cement-vegetation, geckos occurred on

42% (5/12) of aluminum walls, 39% (5/13) of brick walls, 25% (1/4) of wood walls, and

22% (2/9) of cement walls. Walls described as low light/cement contained geckos 71%

(5/7) of the time if they were made of aluminum, 67% (2/3) of the time if they were

constructed of wood, 57% (13/23) of the time if they were cement, and 33% (2/6) of the

time if they were build out of brick. Regarding low light/vegetation walls, 63% (17/27)

of aluminum walls, 44% (12/27) of cement walls, 43% (6/14) of brick walls, and 33%

(5/15) of wood walls were populated by geckos.

Discussion

Sample size proved to be a defining component in the outcome of my analyses.

The realization that the number of sampled walls was too small to accommodate a

logistic regression was unexpected, as my sample size (160 walls) greatly exceeded the

general rule of 10 observations per variable (Agresti, 1996). Indeed, my sample size of

160 was twice that required for my original eight variables, and four times that needed for

the four variables eventually used for the logistic regression model. Further

investigation, however, revealed that overall sample size was not responsible for the

model inadequacy. Instead, it was the mostly small, sometimes zero, sample size of

individual wall combinations that contributed to the collapse of the model. Thus, even if

the sample size for this study was substantially larger, a logistic regression model would

continue to fail if there were wall combinations with a small number of samples.

This could be a common dilemma in urban ecology studies, where investigators

must work within rigid landscapes that offer few opportunities for manipulation. Thus,









the possibility that study sites might not contain specific combinations of variables is

often true, and highly unpredictable due to the human dimension involved. These

statistical considerations should be an integral step in the planning of any urban ecology

study.

Once the data were sufficiently collapsed to compensate for the small sample size

of certain wall combinations, both the logistic regression model and chi-square tests

showed no significance at the 5% level. This lack of significance suggests a variety of

scenarios.

First, the microhabitat preference ofH. turcicus with respect to material, light, and

vegetation might be subtle and thus require a larger sample size to be exposed.

Second, perhaps the microhabitat variables that were selected for this study are

irrelevant to the microhabitat choice ofH. turcicus, and the statistical conclusions merely

report this. Theoretically, habitat selection in reptiles is believed to be most effective

when controlled by reliable environmental cues that are independent of daily and/or

seasonal fluctuations, and are evident in all situations (Heatwole, 1982). Although this

was the case for material, light and vegetation were less consistent due to management

regimes, and thus might not be used as a stimulus because they fail to accurately

represent a given habitat. Other factors such as, but not limited to, microclimate quality,

behavioral aspects and structural attributes may also play an important role in

microhabitat selection and need to be considered in future studies (Heatwole, 1982).

Lastly, it is possible that these results reflect the robust character of this proficient

colonizing gecko species. Perhaps the extensive non-native range ofH. turcicus stems









from this gecko's habitat flexibility, which allows it to thrive on any type of building/wall

environment.

Although no significant results were obtained in this study, some general trends

were uncovered, and may pave the way for future research. In view of material, more

aluminum walls than expected contained geckos, particularly those possessing low light

and vegetation. Aluminum walls were also described as having a smooth texture, light

color, small length and being modem in age. With respect to light-vegetation, fewer high

light/cement-vegetation walls than expected possessed geckos, especially those

constructed of cement. Further work is required to officially establish these trends, and to

tease out the mechanisms behind them.

As with all scientific studies, these data and its conclusions have some limitations.

For instance, it is impossible to determine if the absence of H. turcicus on a wall is due to

preference or if it is an artifact of this species dispersal ability and/or of extraneous

circumstances. Also, shortcomings in sampling technique, such as not observing a wall

throughout the entire night and frightening geckos as I approached a wall, could have

resulted in an underestimation of walls containing geckos.

Additional work, both in the field and in the lab, is needed to shed light on the

microhabitat preferences ofH. turcicus in urban environments. Key to future studies is

the establishment of strict protocols that all investigators can follow anywhere in the

world. This, in turn, would allow for meaningful comparisons between different sites,

and when all studies are pooled, for meta-analyses on general patterns. Moreover, future

work should explore and develop sampling and statistical methods that will enhance

ecological studies in urban environments.











Table 5-1. Description of wall characterization variables for microhabitat study


Levels
Early
Modem
Contemporary
North
South
West
East


Criteria
*Built between 1900 to 1969
*Built between 1970 to 1989
*Built between 1990 to 2002
**Location of wall on official maps
**Location of wall on official maps
**Location of wall on official maps
**Location of wall on official maps


*** The inability to perceive a 3"x 5" white index
Light card at a perpendicular distance of 10ft
(3.048m) from the wall; during the day
***The ability to perceive a 3"x 5" white index
Dark card at a perpendicular distance of 10ft
(3.048m) from the wall; during the day
Small Measurement < 20m

Length Medium 20m < measurement < 40m

Large Measurement > 40m

High Presence of at least one light source on the wall
Light
Low No light source present on the wall
Aluminum Physical observation; > 50% of wall surface
Brick Physical observation; > 50% of wall surface
Material
Cement Physical observation; > 50% of wall surface
Wood Physical observation; > 50% of wall surface
Sm h Ability to draw a straight line -10cm in length on
Smooth
Texture white printer paper propped on the wall
Roh Inability to draw a straight line -10cm in length on
white printer paper propped on the wall
Cement > 60% of wall length bordered by cement

Veget n Mx > 40% to <60% of wall length bordered by
Vegetation Mix
cement or vegetation
> 60% of wall length bordered by any type or
Vegetation
height of vegetation
*Sources used: 2000 Building Information List for the University of Florida prepared by the UF Physical Plant Division,and
unofficial list prepared by the VA Hospital Engineering Department
**Sources used: 2000 Building Information List for the University of Florida prepared by the UF Physical Plant Division,and
official VA Hospital Engineering maps
*** Index cards were provided by AMPAQ, Dallas, TX, 75252


Variables


Cardinal
Orientation

















Table 5-2. Chi-square and Fisher Exact p-values for wall characterization variables
Cardinal
Variables Age Ort Color Length Light Material Texture Vegetation
Orientation
Age -- 0.6695 0.0297 *0.0845 0.7228 <0.0001 0.9564 0.1976

Cardinal
Cardinal 0.6043 *0.8805 0.1941 0.8859 0.7580 0.3484
Orientation

Color ------- ------- ------ *0.0481 *0.0922 <0.0001 *<0.0001 0.3426

Length ------- ------- ------- ------- 0.0005 *0.0041 0.0573 *0.2682

Light ------- ------- ------- ------- ------- 0.0622 *0.0846 0.0464

Material ------- ------- ------- ------- ------- ------ <0.0001 0.0691

Texture ------- ------- ------- ------- ------- ------- ------- 0.7696

Vegetation --- --- --- --- --- --- --- ---

Indicates the use of the Fisher's Exact test
Numbers in bold indicate significance







59


Table 5-3. The dependency of age, color, length, and texture on material

Material Age Color Length Texture

Aluminum Modem Light Small Smooth

Brick Early Dark Small/Moder Smooth

Cement Early Light Small Rough
Modern/
Wood Cotem Dark Small Rough
Contemporary
















Table 5-4. Logistic regression models resulting from the backward elimination method using the variables: material; cardinal
orientation, vegetation, light
Models
Model Predictors Deviance DF s Difference P-Value
Compared
1 M*CO*V*L 141.5044 95

M*CO*V+ M*CO*L+
2 M*C M* + 141.5044 95 2-1 0 >0.999
M*V*L+ CO*V*L
M*CO*V+M*CO*L+
3 *V *141.5044 98 3-1 0 >0.999
M*V*L
M*CO*V+M*CO*L+
4 141.5044 96 4-1 0 >0.999
CO*V*L
M*CO*V+M*V*L+
5 + 152.4738 100 5-1 10.9694 -0.100
CO*V*L
M*CO*L+M*V*L+
6 M*C+ 152.9663 103 6-1 11.4619 > 0.250
CO*V*L

7 M*CO*V+M*CO*L 141.5044 101 7-1 0 >0.999

8 M*CO*V+M*V*L 152.4738 104 8-1 10.9694 >0.100

9 M*CO*L+M*V*L 155.9364 108 9-1 14.4320 >0.100

10 M*CO*V+CO*V*L 156.7944 103 10-1 15.2900 <0.050

11 M*V*L+CO*V*L 171.1805 110 11-1 29.6761 <0.010

12 M*CO*L+CO*V*L 162.5828 109 12-1 21.0784 -0.100














Table 5-4. Continued.
Models
Model Predictors Deviance DF s Difference P-Value
Compared

13 M*CO*V 158.7768 108 13-1 17.2724 -0.100

14 M*CO*L 164.7604 114 14-1 23.2560 >0.100

15 M*V*L 176.4308 115 15-1 34.9264 <0.050

16 CO*V*L 178.6783 116 16-1 37.1739 <0.050

M*CO+M*V+M*L+
17 CO+*+*L185.0743 121 17-1 43.5699 <0.050
CO*V+CO*L+V*L
Numbers in bold indicates significance
Model 7, underlined and in bold, is the chosen model














Table 5-5. Summary statistics of the significant predictor variables of the M*CO*V+M*CO*L model

Predictor Estimate ASE P-value Odds Ratio 95%C.I.of Odds Ratio


Aluminum

Brick

Cement

Wood

Mix

Vegetation

East

North

West

No Light

Light

Aluminum / East

Aluminum / North

Aluminum / West

Brick / South

Brick/ West

North / Mix Veg

North / Veg


-68.3825

-91.7478

-46.1655

0.0000

18.3578

0.0000

109.5860

-45.0170

0.0000

-67.2839

0.0000

131.7139

44.3238

0.0000

- 17.0373

0.0000

-18.3581

0.0000


1.4434

1.6833

1.3663

0.0000

1.5916

0.0000

1.0954

1.5275

0.0000

1.6583

0.0000

1.5652

2.4152

0.0000

2.0897

0.0000

2.0083

0.0000


<0.0001

<0.0001

<0.0001


< 0.0001


< 0.0001

< 0.0001


< 0.0001


< 0.0001

< 0.0001


<0.0001


<0.0001


2.0038 x 10'30 (1.18 x 1031, 3.39 x 1030)

1.4270 x 10-40 (5.27 x 10-42, 3.87 x 1039)

8.9244 x 10-21 (6.13 x 10-22, 1.30 x 1019)


9.3906 x 10 (4.15 x 106,2.13 x 10)
9.3906 x 10' (4.15 x 106 2.13 x 109


3.9138 x 1047

2.8143 x 10-20


6.0114 x 10-30


1.5945 x 1057

1.7766 x 1019


3.9884 x 10-8


1.0646 x 10-


(4.57 x 1046, 3.35 x 1048)

(1.41 x 10-21, 5.62 x 1019)


(2.33 x 1031, 1.55 x 10-28)


(7.42 x 1055, 3.43 x 105 )

(1.56 x 101, 2.02 x 1021)


(6.64 x 10-10, 2.40 x 10-6)


(2.08 x 10-0, 5.45 x 10 )















Table 5-6. Possible wall combinations involving the variables: material, cardinal orientation, vegetation, and light
Cardinal Light Level = High Light Level = Low
Orientation Material Vegetation Level Vegetation Level
Cement Mix Vegetation Cement Mix Vegetation
Aluminum 0 0 3 1 3 5
N h Brick 2 2 2 0 0 0
North
Cement 1 1 1 2 5 6
Wood 1 1 1 0 0 4
Aluminum 1 1 3 1 0 7
Brick 2 1 1 1 0 5
South
Cement 0 1 1 3 5 7
Wood 0 0 0 1 1 7
Aluminum 0 0 2 0 1 6
Brick 0 0 1 2 2 3
West
Cement 1 0 1 1 2 8
Wood 0 0 0 0 0 4
Aluminum 0 1 1 1 1 8
East Brick 2 0 0 1 1 3
East
Cement 1 0 1 3 2 6
Wood 1 0 0 1 0 0










Table 5-7.


Number of observations per wall combination involving the variables:
material vegetation (2 levels) and lig t


Light Level = High Light Level = Low
Material Vegetation Level Vegetation Level

Cement Vegetation Cement Vegetation

Aluminum 3 9 8 26

Brick 9 4 7 13

Cement 5 4 23 27

Wood 3 1 3 15














Table 5-8. Logistic regression models resulting from the backward elimination method using the variables: material, vegetation
(2 levels), and light
Models
Model Predictors Deviance DF odes Difference P-Value
Compared
1 M*V*L 207.9758 144

2 M*L+M*V+V*L 208.5680 147 2-1 0.5922 > 0.250

3 M*L+M*V 208.8361 148 3-1 0.8603 > 0.250

4 M*L+V*L 211.0082 150 4 -1 3.0324 > 0.250

5 M*V+V*L 210.8852 150 5-1 2.9094 > 0.250

6 M*L 211.6953 151 6-1 3.7195 > 0.250

7 M*V 211.1258 151 7-1 3.15 > 0.250

8 V*L 212.4750 153 8-1 4.4992 > 0.250

9 M+V+L 212.8583 154 9-1 4.8825 > 0.250

10 M+L 213.4205 155 10-1 5.4447 > 0.250

11 M+V 216.6519 155 11-1 8.6761 > 0.250

12 V+L 217.7235 157 12-1 9.7477 > 0.250














Table 5-8. Continued.
Models
Model Predictors Deviance DF s Difference P-Value
Compared

13 M 216.8310 156 13-1 8.8552 > 0.250

14 V 221.1641 158 14-1 13.1883 > 0.250

15 L 217.9192 158 15-1 9.9434 > 0.250

16 NONE 221.1817 159 16-1 13.2059 > 0.250

Number in bold indicates model that contain significant parameters










Table 5-9. Number of observations per wall combination involving the variables:
material, and light-vegetation (3 levels)

Light-Vegetation Levels
Material
High Light /
ateial High Ligt Low Light/Cement Low Light/Vegetation
Cement-Vegetation
Aluminum 12 8 26

Brick 13 7 13

Cement 9 23 27

Wood 4 3 15



Table 5-10. Logistic regression models resulting from the backward elimination method
using the variables: material, and light-vegetation (3 levels)
Model Predictors Deviance DF Models Compared Difference P-Value

1 M*LVEG 209.8897 148

2 M+LVEG 212.3961 154 2-1 2.5064 > 0.250

3 M 216.8310 156 3-1 6.9413 > 0.250

4 LVEG 217.2007 157 4-1 7.311 > 0.250

5 NONE 221.1817 159 5-1 11.292 > 0.250

Note that there is no model with significant parameters














Table 5-11. Chi-square test and Fisher's Exact test for light-vegetation and gecko presence, controlling for material


High Light /
Cement -
Vegetation


Material Aluminum Brick Cement Wood


Gecko Presence

Frequency


Yes No


Yes No


Expected 7.0435 4.9565 5.1212 7.8788 4.1186 4.8814 1.4545 2.5455
Frequency


Cell Chi-Square


0.5929


0.8425


0.0029


0.0019


1.0898


0.9196


0.1420


0.0812


Frequency 5 2 2 4 13 10 2 1

Low Light / Expected
Cement Frequenc 4.1087 2.8913 2.3636 3.6364 10.525 12.475 1.0909 1.9091
Cement Frequency

Cell Chi-Square 0.1934 0.2748 0.0559 0.0364 0.5818 0.4909 0.7576 0.4329

Frequency 17 10 6 8 12 15 5 10


Low Light /
Vegetation


Expected 15.8480 11.1520 5.5152 8.4848 12.3560 14.644 5.4545 9.5455
Frequency


Cell Chi-Square


0.0838


0.1190


0.0426


0.0277


0.0103


0.0087


0.0379


0.0216


- r 4


Total Chi-Square


2.1063


0.1674


3.1010


1.4732


P-Value for
Chi-Square/ 0.3488 / 0.4098* 0.9197 / 1.0000* 0.2121 /0.2416* 0.4787/ 0.6431*
Fisher's Exact*














Table 5-12. Chi-square test and Fisher's Exact test for material and gecko presence, controlling for light-vegetation


Aluminum


High Light /
Light-Vegetation Low Light / Cement Low Light / Vegetation
LCement Vegetation


Gecko Presence


No


No


Yes


Frequency 5 7 5 2 17 10

Expected 4.1053 7.8947 3.9487 3.0513 13.012 13.9880
Frequency


Cell Chi-Square


0.195


0.1014


0.2799


0.3622


1.2222


1.1370


Frequency 5 8 2 4 6 8

Brick Expected 4.4474 8.5526 3.3846 2.6154 6.7470 7.2530
Frequency
Cell Chi-Square 0.0687 0.0357 0.5664 0.7330 0.0827 0.0769

Frequency 2 7 13 10 12 15

Cement Expected 3.0789 5.9211 12.9740 10.0260 13.0120 13.9880
Frequency
Cell Chi-Square 0.3781 0.1966 5.07x 10-5 0.0001 0.0787 0.0732

Frequency 1 3 2 1 5 10


Expected
Frequency
Cell Chi-Square


1.3684

0.0992


2.6316

0.0516


1.1263


1.6923

0.0559


1.3077

0.0724


2.0700


7.2289

0.6872


7.7711

0.6393


3.9973


P-Value for
Chi-Square/ 0.7707 / 0.8594* 0.5580 / 0.6111* 0.2618 / 0.2642*
Fisher's Exact*


Wood


-I 4 1


Total Chi-Square







70


Table 5-13. Chi-square test for light-vegetation and gecko presence


High Light /
Cement -
Vegetation


Gecko Presence


Frequency 13 25

Expected 17.8130 20.1880
Frequency


Cell Chi-Square


1.3002


1.1473


Frequency 22 17

Low Light / Expected 20.7190
Cement Frequency

Cell Chi-Square 0.7565 0.6675

Frequency 40 43


Low Light /
Vegetation


Expected 38.9060 44.0940
Frequency


Cell Chi-Square


Total Chi-Square


P-Value for
Chi-Square


0.0307


0.0271


3.9293


4


0.1402


- ~ 4.







71


Table 5-14. Chi-square test for material and gecko presence


Gecko Presence


Frequency 27 19


Aluminum


Expected 21.5630 24.4380
Frequency


Cell Chi-Square


1.3712


1.2099


Frequency 13 20

Brick Expected 15.4690 17.5310
Frequency

Cell Chi-Square 0.3940 0.3476

Frequency 27 32

Cement Expected 27.6560 31.3440
Frequency

Cell Chi-Square 0.0156 0.0137

Frequency 8 14


Wood


Expected 10.3130 11.6880
Frequency


Cell Chi-Square


0.5186


- 4.


Total Chi-Square


P-Value for
Chi-Square


0.4576


4.3282


0.2281










Table 5-15. Three-way contingency table of light-vegetation, controlling for material,
with associated percentages and marginal associations
Gecko Presence Percent
Material Light-Vegetation Gecko
Yes No Present

High Light / 7 42%
Cement -Vegetation
Aluminum Low Light / Cement 5 2 71%

Low Light / Vegetation 17 10 63%

High Light / 5 39
Cement -Vegetation
Brick Low Light / Cement 2 4 33%

Low Light / Vegetation 6 8 43%

High Light / 7 22%
Cement -Vegetation
Cement Low Light / Cement 13 10 57%

Low Light / Vegetation 12 15 44%

High Light / 1 3 25
Cement -Vegetation
Wood Low Light / Cement 2 1 67%

Low Light / Vegetation 5 10 33%

High Light / 25 34
Cement -Vegetation

All Materials Low Light / Cement 22 17 56%

Low Light / Vegetation 40 43 48%










Table 5-16. Three-way contingency table of material, controlling for light-vegetation,
with associated percentages and marginal associations

Light-Vegetation Material Gecko Presence Percent Gecko
Present
Yes No

Aluminum 5 7 42%

High Light/ Brick 5 8 39%
Cement Vegetation Cement 2 7 22%

Wood 1 3 25%

Aluminum 5 2 71%
Brick 2 4 33%
Low Light /Cement
Cement 13 10 57%
Wood 2 1 67%

Aluminum 17 10 63%

Low Light/ Brick 6 8 43%
Vegetation Cement 12 15 44%

Wood 5 10 33%
Aluminum 27 19 59%
All Brick 13 20 39%
Light /Vegetation
Levels Cement 27 32 46%
Wood 8 14 36%














CHAPTER 6
MANAGEMENT AND CONSERVATION IN AN URBAN ENVIRONMENT

Wildlife management and conservation in urban areas have a number of key goals.

These include traditional objectives such as the promotion and/or maintenance of species

composition, and the control of species abundance by either directly increasing or

decreasing numbers (Nilon and Pais, 1997). More importantly, urban wildlife

management and conservation programs also provide the public with opportunities to

interact with wildlife, and disseminate information to the general public and appropriate

professionals (Anderson, 2002). The latter two goals are essential to formulate

ecologically advantageous policy, as the urban public's voting strength in legislatures is

on the rise (Bolen and Robinson, 2003).

The management of urban wildlife follows a holistic style, which differs

dramatically from the conventional agriculture and hunting orientated approach (Bolen

and Robinson, 2003). This difference is directly related to the non-consumptive,

recreational attitude that many urban residents maintain toward wildlife. A study

undertaken in metropolitan areas of New York State demonstrated this idea, as urbanites

were found to prefer butterflies and songbirds to species sought for hunting such as

waterfowl and pheasants (Brown et al., 1979). With the exception of nuisance animals

and/or pests (cockroaches, rats, pigeons, sometimes raccoons, etc.) the urban public

generally supports the notion of having wildlife in their surrounding environment (Bolen

and Robinson, 2003).









Despite the fact that the desire for wildlife is present in cities, it is a challenge to

implement management and/or conservation plans in the myriad of interests and human

beliefs usually present in urban areas (Lyons, 1997). Although public opinion is an

integral part of most management and conservation initiatives, it is especially the case in

urban environments (Anderson, 2002). 7

The intricacies that govern public perception are many, subtle, and highly

susceptible to change. It has been shown that public perception varies enormously within

a city, particularly between communities of different income, education, and race (Nilon

and Pais, 1997). Knox (1991) alluded to this concept by illustrating that land use was

directly related to social, economic, and demographic factors. This variation in public

perception, in turn, tends to lead to equally variable public preferences (Schauman et al.,

1986). Whitney and Adams (1980) showed this concept when they found that the types

of plants in gardens were linked to fashion, taste, species availability, property value, and

age of the house in question.

Developing urban wildlife management and conservation strategies involves a

multitude of participants and considerations, making each case unique (Lyons, 1997).

The heterogeneity of public opinion, and the limitations imposed by both the intensity of

urbanization and the land-use history of a given area require a great deal of creativity and

cross-disciplinary thinking from an urban wildlife manger (Bolen and Robinson, 2003;

Loeb, 1998; Lyons, 1997). Additional complexity results from the fact that few urban

spaces are dedicated solely to wildlife, which requires most management plans to

consider and, sometimes favor, other land uses (Bolen and Robinson, 2003). This

multiple-use management approach requires urban wildlife managers to interact with a









wide range of professionals (Lyons, 1997). A particularly important professional group

is the urban landscape planners, as they execute the majority of urban wildlife initiatives

(Nilon and Pais, 1997). Lastly, public approval and participation is a necessity: the

public is directly affected by any management scheme, as the latter invariably becomes

part of their everyday life. Thus, the public is a preeminent force behind any change

(Lyons, 1997).

In general, urban wildlife managers assume the role of solution facilitators rather

than active problem solvers, as they provide insight, tools, and ideas to a number of

different interest groups in an attempt to coalesce their sensibilities into common, feasible

goals (Lyons, 1997). For example, if the creation of a neighborhood park is being

considered, an urban wildlife manager might suggest programs that unite features

attractive to wildlife with other objectives such as safety, specific recreational purposes,

following city ordinance, remaining within a certain budget, and others. An urban

wildlife manager will not, however, physically carry out the selected program. Instead,

the logistics are left to specific groups such as the police force for safety, landscape

planners for recreational structures, and city officials for monitoring compliance with city

regulations. An equally important function of urban wildlife managers is as educators to

the public via the media, pamphlets, and/or seminars (Anderson, 2002

In the case ofH. turcicus in Gainesville, management strategies would be specific

to the locality due to the gecko's nonnative status. Although H. turcicus is an introduced

species, it is has managed to occupy a vacant niche in Gainesville, and thus is not

believed to possess a threat to any native species. Consequently, the eradication of this

gecko, which would probably be costly due to its prolific colonization abilities and its









potentially generalist microhabitat habits, would not be a necessity. In fact, the presence

ofH. turcicus could be beneficial; being easily observable on walls, this gecko could be

wildlife that many in the public could interact with on a daily basis, hence familiarizing a

number of citizens (and voters) with reptiles, a group often seen in a negative light. An

integral part of the appreciation of H. turcicus would be the use of a number of formats to

educate both adults and children on the interesting facts of not just this species, but of

geckos and reptiles in general. Benefits of this species, such as their consumption of

insects, would especially have to be emphasized since they have been labeled as "pests"

by some homeowners in Gainesville, as a result of the mess they sometimes leave when

they nest (shredded paper, etc.) (Franz, personal communication). Ultimately, I believe

that a primary goal of urban ecologists, regardless of whether they hold a management

position, should be to convey the intrinsic value of nature to all sectors of the urban

population.
















APPENDIX A
PRELIMINARY SURVEY DATA



Table A-1. Temperature readings of walls with respect to material and cardinal
orientation
.l Cardinal Orientation
Material
North South West East
Cement 25.00 25.43 24.75 24.98
Cement 25.28 24.68 24.48 24.75
Cement 24.48 25.03 24.78 24.58
Cement 25.53 26.30 25.88 25.38
Cement 25.33 24.83 25.03 25.10
Cement 26.73 25.93 27.25 27.18
Cement 26.40 26.53 27.15 27.03
Cement 25.45 25.38 25.33 24.78
Cement 25.58 25.53 25.30 26.10
Cement 25.68 24.98 25.55 25.35
Cement 28.60 26.98 29.10 29.08
Cement 26.50 26.58 29.35 27.98
Cement 29.55 27.50 28.88 28.40
Aluminum 23.10 23.08 Not Recorded 23.58
Aluminum 23.75 23.98 24.15 23.93
Aluminum 24.18 24.03 23.93 23.80
Aluminum 24.63 24.05 23.98 24.13
Aluminum 23.40 22.48 23.05 23.33
Aluminum 22.90 22.48 22.90 22.60
Aluminum 22.68 23.08 22.53 22.65
Aluminum Not Recorded 23.28 22.90 22.55
Aluminum 21.75 22.23 23.38 21.85
Aluminum 23.88 24.45 24.75 24.65
Aluminum 25.03 23.15 24.00 24.08
Aluminum 24.10 23.85 24.33 24.25
Aluminum 22.85 23.73 23.28 23.40
Aluminum 23.50 24.75 24.10 24.13
Aluminum 24.33 24.80 24.95 24.93
Aluminum 26.23 23.70 26.10 26.73
Aluminum 24.98 23.88 23.83 23.98
Aluminum 25.88 24.08 24.38 24.25
Aluminum 22.63 23.10 22.88 22.78
Aluminum 25.33 24.87 25.07 25.17












Table A-1. Continued.
Matel Cardinal Orientation
Material
North South West East
Brick 25.13 24.60 24.75 25.00
Brick 25.03 26.48 27.08 26.53
Brick 26.85 25.53 26.55 26.63
Brick 31.63 29.05 29.60 29.45
Wood 24.15 24.85 24.55 24.33
Wood 25.20 25.45 Not Recorded Not Recorded
Wood 25.38 24.13 23.83 24.48
Wood 25.03 25.13 25.30 26.15
Wood 23.05 22.68 23.45 23.80
Wood 26.60 24.75 25.83 25.45
Wood 25.58 25.53 24.80 24.75
Wood 25.75 25.68 25.53 25.50
Wood 24.15 23.33 23.95 24.10
Wood 29.20 27.35 28.10 27.95

















APPENDIX B
NATURAL HISTORY DATA

Table B-1. Temperature readings ( C) for individual adult H. turcicus recorded during
the fall/winter 2001 survey
12.7 19.7 21.1 21.9 22.4 23.1 23.7 24.6 25.8 26.7
14.0 19.7 21.1 21.9 22.5 23.1 23.7 24.6 25.8 26.7
14.9 19.8 21.1 21.9 22.5 23.2 23.8 24.6 25.8 26.8
15.6 19.9 21.2 21.9 22.6 23.2 23.8 24.7 25.9 26.8
15.7 19.9 21.2 21.9 22.6 23.3 23.8 24.7 25.9 26.9
16.1 19.9 21.2 21.9 22.6 23.3 23.8 24.7 25.9 26.9
16.3 19.9 21.3 21.9 22.6 23.3 23.9 24.8 25.9 26.9
16.3 19.9 21.3 21.9 22.6 23.3 23.9 24.8 26.1 26.9
16.6 20.2 21.3 21.9 22.6 23.3 23.9 24.8 26.1 26.9
17.3 20.2 21.4 21.9 22.6 23.3 23.9 24.8 26.1 27.1
17.4 20.2 21.4 21.9 22.7 23.3 23.9 24.9 26.2 27.3
17.6 20.2 21.4 21.9 22.7 23.3 24.0 24.9 26.2 27.4
18.2 20.3 21.4 21.9 22.7 23.3 24.1 24.9 26.2 27.4
18.3 20.3 21.4 22.0 22.7 23.3 24.1 24.9 26.2 27.4
18.3 20.4 21.4 22.0 22.7 23.4 24.1 24.9 26.2 27.6
18.3 20.5 21.4 22.1 22.7 23.4 24.1 25.0 26.3 27.6
18.4 20.5 21.4 22.1 22.7 23.4 24.1 25.1 26.3 27.8
18.7 20.5 21.4 22.1 22.8 23.4 24.1 25.1 26.3 27.9
18.7 20.6 21.4 22.1 22.8 23.4 24.1 25.2 26.4 28.1
18.7 20.7 21.5 22.1 22.8 23.4 24.1 25.2 26.4 28.2
18.8 20.7 21.5 22.1 22.8 23.4 24.1 25.2 26.4 28.3
18.8 20.7 21.5 22.2 22.9 23.5 24.1 25.2 26.4 28.3
18.8 20.8 21.5 22.2 22.9 23.5 24.3 25.3 26.4 28.4
18.9 20.8 21.5 22.2 22.9 23.5 24.3 25.3 26.5 28.6
18.9 20.8 21.6 22.2 22.9 23.5 24.4 25.3 26.5 28.7
19.0 20.8 21.7 22.2 22.9 23.5 24.4 25.3 26.5 28.7
19.1 20.9 21.7 22.2 22.9 23.6 24.4 25.4 26.5 30.2
19.2 20.9 21.7 22.2 22.9 23.6 24.4 25.5 26.6 30.5
19.2 20.9 21.7 22.3 23.0 23.6 24.4 25.5 26.6 30.6
19.2 20.9 21.7 22.3 23.0 23.6 24.5 25.5 26.6 30.9
19.3 21.0 21.8 22.3 23.1 23.7 24.6 25.6 26.7 31.2
19.5 21.0 21.8 22.3 23.1 23.7 24.6 25.7 26.7 31.2
19.6 21.0 21.8 22.3 23.1 23.7 24.6 25.7 26.7 31.8
19.6 21.1 21.8 22.4 23.1 23.7 24.6 25.7 26.7 34.3
19.6 21.1 21.8 22.4 23.1 23.7 24.6 25.8 26.7 36.3
19.6 21.1











Table B-2. Temperature readings (C) for individual sub-adult H. turcicus recorded
during the fall/winter 2001 survey.



21.1 13.9 18.8 20.1 21.9 22.6 23.3 23.7 24.2 25.1 26.4
21.1 16.3 18.9 20.2 21.9 22.7 23.3 23.7 24.3 25.1 26.5
21.2 16.6 18.9 20.3 22.1 22.7 23.3 23.7 24.3 25.2 26.6
21.2 16.8 19.2 20.3 22.1 22.7 23.3 23.7 24.4 25.2 26.7
21.3 17.1 19.3 20.3 22.1 22.7 23.3 23.7 24.4 25.2 26.7
21.3 17.4 19.5 20.4 22.1 22.7 23.3 23.7 24.4 25.2 26.8
21.3 17.5 19.5 20.4 22.2 22.7 23.3 23.8 24.4 25.2 27.1
21.4 17.5 19.5 20.5 22.2 22.8 23.4 23.8 24.6 25.2 27.2
21.6 17.7 19.6 20.6 22.2 22.8 23.4 23.8 24.6 25.3 27.3
21.6 17.7 19.7 20.6 22.3 22.8 23.4 23.9 24.6 25.3 27.3
21.6 17.8 19.7 20.6 22.3 22.8 23.4 23.9 24.6 25.3 27.6
21.7 18.0 19.7 20.7 22.4 22.9 23.4 23.9 24.6 25.6 27.8
21.7 18.3 19.7 20.7 22.4 22.9 23.4 24.0 24.7 25.6 27.8
21.7 18.4 19.7 20.7 22.4 22.9 23.5 24.0 24.8 25.7 28.3
21.7 18.4 19.8 20.8 22.6 23.0 23.5 24.1 24.8 25.7 28.4
21.7 18.4 19.8 20.8 22.6 23.1 23.5 24.1 24.9 25.9 28.9
21.7 18.7 19.8 20.8 22.6 23.1 23.5 24.1 24.9 26.1 29.3
21.8 18.7 19.8 20.9 22.6 23.1 23.6 24.2 24.9 26.2 30.3
21.8 18.7 19.9 20.9 22.6 23.1 23.6 24.2 25.0 26.2 30.5
21.8 18.7 20.0 21.0 22.6 23.2 23.6 24.2 25.0 26.4 30.8
21.8 18.7 20.1 21.1











Table B-3. Temperature readings (C) for individual adult H. turcicus recorded during
the spring 2002 survey


16.4 22.6 24.1 25.1 26.2 27.3
16.6 22.7 24.1 25.1 26.2 27.3
16.8 22.7 24.2 25.2 26.3 27.4
17.0 22.7 24.2 25.2 26.3 27.4
17.2 22.8 24.2 25.2 26.3 27.6
17.7 22.8 24.2 25.2 26.4 27.6
17.7 22.8 24.2 25.2 26.4 27.6
18.6 22.8 24.3 25.2 26.5 27.7
18.6 22.9 24.3 25.2 26.6 27.9
18.7 22.9 24.3 25.3 26.6 27.9
18.9 22.9 24.3 25.3 26.6 28.1
19.0 23.0 24.3 25.5 26.6 28.1
19.2 23.0 24.3 25.5 26.7 28.3
19.3 23.0 24.3 25.5 26.7 28.3
19.3 23.2 24.5 25.5 26.7 28.4
19.4 23.2 24.5 25.6 26.7 28.6
19.4 23.3 24.6 25.6 26.7 28.7
19.9 23.3 24.6 25.6 26.7 28.8
19.9 23.3 24.6 25.6 26.7 28.8
20.7 23.4 24.7 25.6 26.7 28.8
20.7 23.4 24.7 25.6 26.8 28.9
20.8 23.4 24.7 25.7 26.8 29.2
20.9 23.4 24.7 25.7 26.8 29.4
21.2 23.5 24.7 25.7 26.8 29.4
21.2 23.5 24.8 25.7 26.8 29.7
21.3 23.6 24.8 25.7 26.8 29.7
21.4 23.6 24.8 25.7 26.8 29.8
21.7 23.7 24.8 25.8 26.9 29.9
21.8 23.7 24.8 25.8 26.9 30.1
21.9 23.8 24.9 25.8 26.9 30.1
22.1 23.8 24.9 25.8 26.9 30.2
22.1 23.8 24.9 25.8 27.1 30.2
22.1 23.9 24.9 25.9 27.1 30.3
22.1 23.9 25.0 25.9 27.1 30.4
22.3 24.0 25.0 25.9 27.1 30.6
22.3 24.1 25.1 26.1 27.1 30.8
22.3 24.1 25.1 26.1 27.2 30.8
22.4 24.1 25.1 26.2 27.2 30.9
22.4 24.1 25.1 26.2 27.2 31.5
22.5 24.1 25.1















APPENDIX C
MICROHABITAT PREFERENCE DATA












Cardinal Presence Presence
Material Vegetation Texture Color Age Length
Orientation ______of Light of H.turcicus
Wood North Cement Rough Light Contemporary Small Yes 1
Brick South Cement Smooth Dark Early Medium Yes 1
Aluminum East Vegetation Smooth Dark Contemporary Small No 0
Wood West Vegetation Rough Dark Contemporary Small No 1
Cement West Vegetation Rough Light Early Small No 1
Brick West Mix Smooth Dark Early Medium No 1
Brick North Mix Rough Light Early Medium Yes 0
Aluminum East Vegetation Rough Light Modern Medium No 0
Aluminum South Vegetation Rough Light Modern Small No 1
Wood North Vegetation Smooth Light Early Medium No 1
Wood South Vegetation Smooth Light Contemporary Small No 0
Cement North Mix Rough Light Early Large No 0
Brick West Cement Smooth Dark Modern Small No 0
Brick West Vegetation Smooth Dark Early Medium No 1
Brick North Vegetation Smooth Dark Early Medium Yes 1
Brick East Cement Smooth Dark Early Large Yes 1
Brick South Cement Smooth Dark Early Medium Yes 1
Brick East Vegetation Smooth Dark Modern Small No 1
Brick North Vegetation Smooth Dark Early Small No 0
Aluminum East Mix Rough Light Modern Small Yes 0
Brick South Vegetation Smooth Dark Early Medium No 1
Brick West Vegetation Smooth Dark Contemporary Small No 0
Cement South Mix Rough Light Early Small No 1
Cement North Mix Rough Light Early Small No 1
Brick South Vegetation Smooth Dark Contemporary Small No 0
Aluminum West Vegetation Rough Light Modern Small No 1












Aluminum East Vegetation Rough Light Modern Medium No 1
Wood South Vegetation Rough Dark Modern Medium No 1
Aluminum East Vegetation Smooth Dark Modern Small Yes 1
Wood West Vegetation Rough Dark Modern Small No 0
Wood North Vegetation Rough Dark Contemporary Small No 0
Wood South Vegetation Rough Dark Modern Large No 1
Brick South Vegetation Smooth Dark Early Medium Yes 0
Cement West Vegetation Rough Light Early Small No 1
Cement South Vegetation Rough Light Early Small No 0
Cement South Cement Rough Light Early Small No 0
Cement South Cement Rough Light Early Small No 0
Cement East Cement Rough Light Early Small No 0
Aluminum South Mix Smooth Dark Early Medium Yes 1
Wood South Cement Smooth Light Early Small No 0
Aluminum South Cement Rough Light Modem Small No 0
Brick North Vegetation Smooth Dark Early Large Yes 1
Brick West Cement Smooth Dark Early Medium No 1
Aluminum West Vegetation Smooth Light Early Medium Yes 0
Cement West Mix Rough Light Early Small No 0
Cement South Mix Rough Light Early Small No 1
Wood West Vegetation Rough Light Early Large No 1
Cement North Cement Rough Light Modern Small Yes 0
Cement South Cement Rough Light Modem Medium No 1
Cement East Cement Rough Light Modern Small No 0
Cement East Vegetation Rough Light Modern Small No 0
Aluminum East Mix Smooth Light Modern Small No 0
Cement North Mix Rough Light Modern Small No 1
Cement West Cement Rough Light Modern Medium Yes 1












Cement East Mix Rough Light Early Small No 0
Cement South Mix Rough Light Early Small No 0
Aluminum West Vegetation Rough Dark Modern Small Yes 1
Aluminum West Vegetation Smooth Dark Modern Small No 0
Cement North Vegetation Rough Light Modern Small No 0
Aluminum North Vegetation Smooth Light Modern Small No 0
Cement North Vegetation Rough Light Modern Small No 0
Cement West Vegetation Rough Light Early Small No 1
Cement North Mix Rough Light Modern Medium Yes 1
Cement East Vegetation Rough Light Modern Small No 0
Aluminum North Mix Smooth Light Modern Small No 0
Cement West Vegetation Rough Light Modern Small No 0
Brick East Vegetation Smooth Dark Modern Small No 1
Cement South Vegetation Smooth Light Early Small Yes 1
Aluminum South Vegetation Smooth Light Early Small No 0
Brick South Vegetation Smooth Dark Contemporary Medium No 0
Aluminum West Vegetation Smooth Light Modern Medium No 0
Aluminum East Vegetation Smooth Light Modern Small No 0
Wood South Vegetation Rough Dark Modern Small No 0
Aluminum South Vegetation Rough Light Contemporary Small No 0
Cement West Vegetation Rough Light Early Small No 0
Wood North Vegetation Rough Light Early Small Yes 1
Cement South Vegetation Rough Light Early Small No 1
Brick West Mix Smooth Dark Early Medium No 1
Wood East Cement Rough Light Early Small Yes 0
Aluminum North Vegetation Smooth Dark Early Small No 0
Aluminum East Vegetation Smooth Dark Modern Small No 0
Aluminum North Mix Smooth Light Modern Small No 0












Cement North Vegetation Rough Light Early Small No 1
Aluminum North Mix Smooth Light Modern Small No 1
Cement North Vegetation Rough Light Early Small No 1
Cement South Vegetation Rough Light Modern Small No 1
Cement East Mix Rough Light Early Small No 0
Aluminum North Vegetation Rough Light Early Small No 0
Aluminum South Vegetation Smooth Light Modern Medium Yes 1
Brick North Cement Smooth Dark Contemporary Small Yes 0
Aluminum South Vegetation Smooth Dark Modern Small No 1
Wood North Vegetation Rough Dark Contemporary Small No 1
Wood East Cement Rough Dark Contemporary Small No 1
Wood South Vegetation Rough Dark Contemporary Small No 1
Cement South Vegetation Rough Light Contemporary Medium No 0
Cement North Mix Rough Light Contemporary Small No 0
Aluminum North Vegetation Rough Dark Contemporary Small Yes 1
Aluminum South Vegetation Smooth Light Modern Small No 0
Brick North Cement Smooth Dark Contemporary Medium Yes 1
Brick East Cement Smooth Dark Contemporary Small No 0
Cement West Vegetation Rough Light Early Large No 0
Aluminum North Vegetation Rough Light Modern Small No 0
Cement North Vegetation Rough Light Early Small Yes 0
Cement South Mix Rough Light Early Small No 1
Cement North Cement Rough Light Early Small No 0
Cement East Vegetation Rough Light Early Small No 0
Cement North Cement Rough Light Earl Small No 0
Cement North Mix Rough Light Early Small No 0
Brick East Vegetation Smooth Dark Early Large No 1
Cement East Vegetation Smooth Light Early Medium No 1












Brick South Mix Smooth Dark Modern Medium Yes 1
Brick North Mix Smooth Dark Early Medium Yes 1
Wood South Vegetation Rough Dark Modern Small No 1
Aluminum South Vegetation Smooth Dark Early Small No 1
Cement North Vegetation Rough Light Modern Small No 0
Cement East Vegetation Rough Light Modern Medium No 1
Aluminum West Vegetation Smooth Light N/A Small No 0
Aluminum South Cement Smooth Light Modern Small Yes 1
Aluminum West Mix Rough Light Modern Small No 0
Aluminum South Vegetation Rough Light Contemporary Small No 1
Cement South Vegetation Rough Light Early Small No 0
Cement South Mix Rough Light Modern Medium No 1
Wood South Mix Rough Dark Modern Medium No 0
Aluminum East Vegetation Rough Light Modern Small No 1
Aluminum North Cement Rough Light Contemporary Small No 0
Aluminum West Vegetation Rough Light Modern Small No 0
Aluminum East Vegetation Smooth Dark Modern Small No 1
Brick East Mix Smooth Dark Early Small No 1
Aluminum East Vegetation Smooth Light Modern Small No 1
Cement East Cement Rough Light Early Medium No 1
Wood West Vegetation Rough Dark Contemporary Small No 1
Wood North Vegetation Rough Dark Modern Small No 0
Brick East Cement Rough Dark Modern Small Yes 0
Wood South Vegetation Rough Dark Modern Small No 1
Wood North Mix Rough Dark Modern Medium Yes 1
Aluminum South Vegetation Rough Light Modern Small Yes 0
Aluminum West Vegetation Rough Light Contemporary Small No 1
Aluminum North Vegetation Smooth Light Contemporary Small Yes 0












Brick South Cement Rough Dark Contemporary Small No 1
Aluminum East Cement Smooth Dark Modern Small No 1
Brick West Vegetation Rough Dark Early Small No 1
Aluminum North Vegetation Smooth Light Modern Small No 0
Cement North Vegetation Rough Dark Early Medium No 1
Cement West Vegetation Rough Dark Modern Small No 1
Cement South Mix Rough Dark Early Large Yes 1
Cement West Mix Rough Light Contemporary Medium No 1
Aluminum South Vegetation Rough Light Modern Medium Yes 1
Cement East Vegetation Rough Dark Early Small Yes 1
Cement East Vegetation Rough Dark Early Small No 1
Cement East Cement Rough Dark Early Medium Yes 1
Cement South Vegetation Rough Light Early Small No 1
Cement West Cement Rough Light Modern Small No 1
Cement West Vegetation Rough Light Early Small Yes 1
Cement South Vegetation Rough Light Modern Small No 1
Cement West Vegetation Rough Light Modern Small No 1
Aluminum North Vegetation Smooth Light Modern Small Yes 0
Brick South Vegetation Smooth Dark Early Small No 0
Brick South Vegetation Smooth Dark Modern Small No 1
Brick West Vegetation Smooth Dark Contemporary Medium Yes 0
Brick North Vegetation Rough Dark Contemporary Small No 0
















REFERENCES


Agresti, A. 1996. An introduction to categorical data. John Wiley & Sons Inc., New
York, New York.

Anderson, S. H. 2002. Managing our wildlife resources. Prentice Hall, Upper Saddle
River, New Jersey.

Angilletta, M. L., L. G. Montgomery, and Y. L. Werner. 1999. Temperature preference
in geckos: diel variation in juveniles and adults. Herpetologica 55 (2): 212-222.

Arnold, E. N. 1984. Ecology of lowland lizards in the eastern United Arab Emirates.
Journal of Zoology: Preceedings of the Zoological Society of London 204: 329-
354.

Barbour, T. 1936. Two introduced lizards in Miami, Florida. Copeia 1936 (2): 113.

Bartholomew, G. A. 1959. Photoperiodism in reptiles. In: Photoperiodism and related
phenomena in plants and animals, edited by R. B. Withrow, American
Association for the Advancement of Science, Washington D.C.

Bolen, E. G. and W. L. Robinson. 2003. Wildlife ecology and management. Prentice
Hall, Upper Saddle River, New Jersey.

Botkin, D. B. and C. E. Beveridge. 1997. Cities as environments. Urban Ecosystems 1:
3-19.

Brown T. L., C. P. Dawson, and R. L. Miller. 1979. Interests and attitudes of
metropolitan New York residents about wildlife. Transactions of the North
American Wildlife and Natural Resource Conference 44: 289-297.

Capula, M. and L. Luiselli. 1994. Trophic niche overlap in sympatric Tarentola
mauritanica and Hemidactylus turcicus: a preliminary study. Herpetological
Journal 4: 24-25.

Carey, S. D. 1988. Food habits of gulf coast Mediterranean geckos (Hemidactylus
turcicus). Journal of the Alabama Academy of Sciences 59 (3): 103.

Collins, J. P., A. Kinzig, N. B. Grimm, W. F. Fagan, D. Hope, J. G. Wu, and E. T. Borer.
2000. A new urban ecology. American Scientist 88: 416-425.




Full Text

PAGE 1

MICROHABITAT PREFERENCE OF THE INTRODUCED GECKO Hemidactylus turcicus IN AN URBAN ENVIRONMENT By PATRICIA A.GOMEZ ZLATAR 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 2003

PAGE 2

Copyright 2003 by Patricia A. Gomez Zlatar

PAGE 3

ACKNOWLEDGMENTS I would first like to thank my committee chair, Mike Moulton, for his constant support, mentoring, and enthusiasm. His friendship has been instrumental in the completion of this degree. I owe a great deal of thanks to committee member Ken Portier for the endless hours of guidance he provided. I am also grateful to committee member Dick Franz, for providing useful comments and ideas throughout the study. It has been both a privilege and pleasure to interact with them all. Many people made my fieldwork possible. I owe thanks to both the UF Physical Plant Division and to J. Darcy White for providing me with detailed maps and a wealth of information. I would like to thank the UF Campus Police Force and the VA Hospital Police Force for keeping my helpers and me safe during our nighttime surveys. I also want to give a big thanks to Chuck Knapp, Ester Langan, Elza Kephart, Alex Martin, and Robin Sternberg for providing endless hours of field assistance. Finally, I want to give immense gratitude to those closest to me. I give thanks to my family, especially my parents. Also, I want to thank my close friends (they know they are). Last, but not least, I want to thank Alex Martin for his patience, friendship, and love. I could not have done it without him. iii

PAGE 4

TABLE OF CONTENTS page ACKNOWLEDGMENTS.................................................................................................iii LIST OF TABLES.............................................................................................................vi LIST OF FIGURES.........................................................................................................viii ABSTRACT.......................................................................................................................ix CHAPTER 1 INTRODUCTION........................................................................................................1 Urban Ecology..............................................................................................................1 General Objective.........................................................................................................2 2 STUDY SPECIES AND STUDY SITES.....................................................................3 Study Species................................................................................................................3 Study Sites....................................................................................................................6 The Study Species in the Study Sites...........................................................................8 3 PRELIMINARY SURVEY..........................................................................................9 Introduction...................................................................................................................9 Methods........................................................................................................................9 Results.........................................................................................................................11 Discussion...................................................................................................................13 4 ADDITIONAL NATURAL HISTORY NOTES.......................................................21 Introduction.................................................................................................................21 Methods......................................................................................................................23 Results.........................................................................................................................24 Summer 2001.......................................................................................................24 Fall/Winter 2001..................................................................................................25 Spring 2002.........................................................................................................26 Discussion...................................................................................................................26 iv

PAGE 5

5 MICROHABITAT PREFERENCE IN THE INTRODUCED GECKO, HEMIDACTYLUS TURCICUS...................................................................................39 Introduction.................................................................................................................39 Methods and Results...................................................................................................41 Sampling Methods...............................................................................................41 Data Analysis 1....................................................................................................44 Results 1..............................................................................................................46 Data Analysis 2....................................................................................................48 Results 2..............................................................................................................49 Data Analysis 3....................................................................................................50 Results 3..............................................................................................................50 Data Analysis 4....................................................................................................51 Results 4..............................................................................................................52 Data Analysis 5....................................................................................................52 Results 5..............................................................................................................52 Discussion...................................................................................................................54 6 MANAGEMENT AND CONSERVATION IN AN URBAN ENVIRONMENT....75 APPENDIX A PRELIMINARY SURVEY DATA............................................................................79 B NATURAL HISTORY DATA...................................................................................81 C MICROHABITAT PREFERENCE DATA...............................................................84 REFERENCES..................................................................................................................91 BIOGRAPHICAL SKETCH.............................................................................................96 v

PAGE 6

LIST OF TABLES Table page 3-1 Criteria of wall characterization variables used in preliminary survey....................16 4-1 Criteria of natural history variables..........................................................................31 4-2 Temperature measurements of wall surface for adult and subadult H. turcicus; fall/winter 2001........................................................................................................31 4-3 Temperature measurements of wall surface for adult H. turcicus; spring 2002.......31 5-1 Description of wall characterization variables for microhabitat study.....................57 5-2 Chi-square and Fisher Exact p-values for wall characterization variables..............58 5-3 The dependency of age, color, length, and texture on material................................59 5-4 Logistic regression models resulting from the backward elimination method using the variables: material; cardinal orientation, vegetation, light.......................60 5-5 Summary statistics of the significant predictor variables of the M*CO*V+M*CO*L model.....................................................................................62 5-6 Possible wall combinations involving the variables: material, cardinal orientation, vegetation, and light..............................................................................63 5-7 Number of observations per wall combination involving the variables: material, vegetation (2 levels), and light.................................................................................64 5-8 Logistic regression models resulting from the backward elimination method using the variables: material, vegetation (2 levels), and light..................................65 5-9 Number of observations per wall combination involving the variables: material, and light-vegetation (3 levels)..................................................................................67 5-10 Logistic regression models resulting from the backward elimination method using the variables: material, and light-vegetation (3 levels)...................................67 5-11 Chi-square test and Fishers Exact test for light-vegetation and gecko presence, controlling for material.............................................................................................68 vi

PAGE 7

5-12 Chi-square test and Fishers Exact test for material and gecko presence, controlling for light-vegetation................................................................................69 5-13 Chi-square test for light-vegetation and gecko presence.........................................70 5-14 Chi-square test for material and gecko presence......................................................71 5-15 Three-way contingency table of light-vegetation, controlling for material, with associated percentages and marginal associations...................................................72 5-16 Three-way contingency table of material, controlling for light-vegetation, with associated percentages and marginal associations...................................................73 A-1 Temperature readings of walls with respect to material and cardinal orientation....79 B-1 Temperature readings (C) for individual adult H. turcicus recorded during the fall/winter 2001 survey.............................................................................................81 B-2 Temperature readings (C) for individual sub-adult H. turcicus recorded during the fall/winter 2001 survey.......................................................................................82 B-3 Temperature readings (C) for individual adult H. turcicus recorded during the spring 2002 survey...................................................................................................83 vii

PAGE 8

LIST OF FIGURES Figure page 3-1 Depiction of the 45 angle leeway employed in the determination of wall cardinal orientation...................................................................................................17 3-2 Average wall temperature of different construction material..................................17 3-3 Average wall temperature of different cardinal locations........................................18 3-4 Average wall temperature of different construction materials at different cardinal locations......................................................................................................18 3-5 Average number of H. turcicus per building of different construction material......19 3-6 Number of H. turcicus on walls of different cardinal locations...............................19 3-7 Number of H. turcicus on walls of different vegetation levels................................20 3-8 Number of H. turcicus on walls of different light intensities...................................20 4-1 Perch height preference of adult and subadult H. turcicus (summer 2001).............32 4-2 Perch height preference of adult and subadult H. turcicus (fall/winter 2001).........33 4-3 Social preference of adult and subadult H. turcicus (fall/winter 2001)...................34 4-4 Exposure preference of adult and subadult H. turcicus (fall/winter 2001)..............35 4-5 Perch height preference of adult H. turcicus (spring 2002).....................................36 4-6 Social preference of adult H. turcicus (spring 2002)...............................................37 4-7 Exposure preference of adult H. turcicus (spring 2002)..........................................38 viii

PAGE 9

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 MICROHABITAT PREFERENCE OF THE INTRODUCED GECKO Hemidactylus turcicus IN AN URBAN ENVIRONMENT By Patricia A. Gomez Zlatar August 2003 Chair: Michael Moulton Major Department: Wildlife Ecology and Conservation I investigated microhabitat preference in the introduced gecko H. turcicus in Gainesville, Florida from summer 2001 through spring 2002. After collecting extensive natural history data in 2001, I then attempted to construct a model for microhabitat preference during 2002, using logistic regression. I characterized 160 walls by construction material, vegetative cover, artificial light intensity, cardinal orientation, age of building, length of wall, surface color, and surface texture. I sampled each wall twice for the presence or absence of H. turcicus. Chi-square analyses indicated that the age of building, length of wall, surface color, and surface texture were dependent on construction material (p< 0.05). I then fit a logistic regression model with the variables: construction material, vegetative cover, artificial light intensity, and cardinal orientation. I was unable to obtain a functional logistic regression model, probably owing to a small sample size. Thus, I condensed the dataset by eliminating the variable cardinal orientation and combining the variables ix

PAGE 10

vegetative cover and artificial light intensity. I was unable to obtain a significant logistic regression model (p<0.05) with this reduced dataset; and I therefore performed chi-square analyses instead. Results revealed no significance (p<0.05) between the presence of H. turcicus and the three wall variables examined. Hence, the presence of H. turcicus on a wall appears to be independent of construction material, vegetative cover, and artificial light intensity. This conclusion indicates that H. turcicus does not demonstrate a preference among walls of different material type, vegetative cover levels, and artificial light intensities. These results could reflect the generalist tendency of this gecko; and thus explain its overall resiliency as an introduced species. Alternatively, the inability to detect significance could reflect failure to gather a sufficient sample size; or failure to properly select variables relevant to microhabitat preference in this species. x

PAGE 11

CHAPTER 1 INTRODUCTION Urban Ecology Urbanization is a significant global phenomenon (Grimm et al., 2000). Roughly half of the worlds population currently resides in cities, as the tendency toward urbanization is characteristic of both developed and developing nations (Lord et al., 2003). This trend is projected to increase over the next few decades, whereby the number and sizes of cities are expected to grow extensively (Pickett et al., 2001). The inevitable growth of urban areas, and their subsequent ecological impacts, makes the field of urban ecology timely and ultimately essential (Grimm et al., 2000). Urban ecology is a fairly young discipline that has increased in prominence over the last two decades (Rebele, 1994). Before this, ecological studies in urban environments were rare as ecologists generally considered urban areas ecologically inferior to natural ones; and thus demonstrated little interest and overall disregard for cities (Botkin and Beveridge, 1997; Gilbert, 1989). However, this anti-urban attitude began to progressively vanish when ecologists started to recognize and become concerned with the influence of humans on ecosystems (Niemela, 1999). Urban ecology first emerged as a discipline that mainly dealt with the ecology of habitats and organisms within cities (Pickett et al., 2001). It eventually expanded when it embraced and then advocated the notion that cities were ecosystems in themselves, with humans occupying the position of keystone species (Rees, 1997). With this new perspective, urban ecologists have recognized that the urban setting cannot be adequately understood and 1

PAGE 12

2 that findings are inapplicable without accounting for human influence (Grimm et al., 2000). As a consequence, urban ecology is in the process of developing into an integrated discipline that incorporates the social, behavioral, economic, physical, and ecological sciences (Niemela, 1999). Despite the surge in popularity of urban ecology, few studies have strictly dealt with urban species and/or been conducted in urban settings (McIntyre et al., 2000). In a review of leading ecology journals between 1993 and 1997, Collins at al. (2000) concluded that a mere 0.4% of papers surveyed (25 of 6157) were restricted specifically to urban habitat and/or urban wildlife. Although reasons for this paucity have not been specified, urban areas offer the distinct challenge of being controlled by strong and diverse human actions (Dow, 2000). An ongoing history of intense and varied micromanagement has made metropolitan landscapes into highly heterogeneous areas, their uniqueness wrought with logistical constraints (Dow, 2000; McIntyre et al., 2000). Of the few wildlife ecological studies performed in urban environments, most feature birds, mammals, and terrestrial invertebrates. Significantly less popular subjects are aquatic fauna, amphibians, and reptiles (Luniak and Pisarki, 1994). Many of these studies, in turn, are anecdotal in nature; thus making comparisons between locations and the formation of a general body of knowledge unfeasible (McIntyre et al., 2000). General Objective In light of the current shortcomings in urban ecology, the main objective of this study was to conduct a repeatable, quantitative ecological study on the microhabitat preferences of an introduced reptile in an urban environment.

PAGE 13

CHAPTER 2 STUDY SPECIES AND STUDY SITES Study Species The Mediterranean gecko Hemidactylus turcicus is an old world geckkonid lizard that has successfully extended its range into India and North America through human-assisted introductions (Conant and Collins, 1998). This species occurs naturally in the Middle East and Mediterranean regions. Hemidactylus turcicus is thought to have reached North America after being initially introduced through human agencies into the Antilles and Gulf-coastal Mexico. It first appeared in the United States in 1915 in Key West, Florida (Stejneger, 1922). Since then, H. turcicus has expanded onto the mainland where it has established itself in a number of localities throughout the southeastern and south-central states; specifically Alabama, Arizona, Arkansas, California, Florida, Georgia, Louisiana, Mississippi, and Texas (Barbour, 1936; Etheridge, 1952; Conant, 1955; Conant and Collins, 1998). The Mediterranean gecko possesses a life history that, with human assistance, favors successful colonization of new areas. Both Davis (1974) and Meshaka Jr. (1995) have reported that the dispersal of H. turcicus in Texas and Florida parallels that of major highways; and that produce trucks were the most likely source of transportation for this gecko. The calcareous shelled eggs of H. turcicus are fairly resistant to desiccation; and egg survivorship appears to be extremely high. With an incubation period of 40 to 45 days, these eggs are ideal for surviving lengthy truck rides (Selcer, 1986). In 1986, 3

PAGE 14

4 Selcer confirmed this high egg survivorship when he obtained a 100% hatching success rate from 100 eggs he had collected in the field. The secretive nesting behavior of the Mediterranean gecko also favors it as a colonizing species. Nests in this species are usually constructed in hidden locations such as attics, storage rooms, under eaves of houses, closets, and rock crevices; and on a wide array of surfaces, including cardboard boxes, wood planks, and old clothing (Davis, 1974; Punzo, 2001a; Selcer, 1986; Trauth, 1985). Furthermore, eggs have reduced visibility as they are often covered with debris, including dirt, paper, eggshell and shed skin (Punzo, 2001a; Rose and Barbour, 1968; Selcer, 1986). Many of the nesting sites are in prime positions to be moved or transported in vehicles. Nests in H. turcicus range from solitary to communal, with some communal nests containing as many as 20 eggs (Selcer, 1986). Therefore, the possibility exists of unknowingly transporting a mini colony to a new locale. Mediterranean gecko hatchlings have also shown remarkable survivorship in dry environments, requiring no food or water for up to a month (Rose and Barbour, 1968). Thus, H. turcicus eggs/hatchling can survive dry and nutrient-poor environments for extended periods of time (over two months) making this gecko a resilient and ultimately successful stowaway on vehicles. Once at a new site, H. turcicus often occurs at extremely high densities: as many as 544 to 2210 geckos/hectare in Texas (Selcer, 1986); and 497 to 1463 geckos/hectare in Florida (Punzo, 2001a). A high population density coupled with a consistently encountered 1:1 sex ratio allows for the potential of a large, annual reproductive output by a population. Females are reproductively active between the months of April and September; and are believed to have two to three clutches a season, each clutch

PAGE 15

5 consisting of two eggs (Rose & Barbour, 1968; Selcer, 1986; Meshaka Jr., 1995; Punzo, 2001a). Hemidactylus turcicus is also characterized by being an early maturing species with a long lifespan; on average, juveniles require between eight and nine months to mature, and routinely live at least three years (Selcer, 1986). Hemidactylus turcicus is highly pre-adapted for life in urbanized areas; and this further aids in dispersal. The presence of scansors (adhesive pads on toes) allows the Mediterranean gecko to perch on vertical walls of buildings. In fact, the Mediterranean gecko is a familiar resident in many cities and towns around the world; and according to Luiselli and Capizzi (1999) is more often found in human-disturbed areas than in natural environments. In addition, H. turcicus is considered to be a generalized predator, as a number of studies have reported a wide array of mostly arthropod prey in this geckos diet (Carey, 1988; Punzo, 2001a; Saenz, 1996). Although Punzo (2001a) failed to detect ageor sex-related differences in diets, Saenz (1996) observed food partitioning between both juveniles and adults and males and females, although the latter remained inconclusive due to small sample size. Thus, under some circumstances, food partitioning could further contribute to the success of H. turcicus, since it would reduce intraspecific competition, and increase the feeding efficiency of a given population (Saenz, 1996). The Mediterranean gecko further increases its chances for establishment with a number of predator escape and avoidance tactics. As a nocturnal, arboreal, and cryptically colored lizard, H. turcicus has few known predators (Selcer, 1986). In a study conducted in Tampa Bay, Florida, Punzo (2001a) listed bats, Cuban tree frogs, large heteropodid crab and wolf spiders, giant tail scorpions, and feral domestic cats as

PAGE 16

6 possible predators. As expected, the most vulnerable period for this gecko is shortly after hatching; both Selcer (1986) and Punzo (2001a) showed that mortality was significantly greater for small juveniles when compared to large juveniles, adult females, and adult males. Accordingly, Selcer (1986) regularly observed an avoidance mechanism in juveniles termed "tail wagging"; when disturbed, H. turcicus juveniles draw attention away from their bodies by wagging their conspicuously banded tails, which can easily be autotomized for a rapid escape. Adult H. turcicus are adept at fleeing danger too, as noted by Selcer (1986) they use routine escape routes when harassed. Adults also have easily autotomized tails, which can startle and/or distract a predator while the gecko retreats to safety (Selcer, 1986). Most of the areas invaded by H. turcicus have little or no competitive pressure. Noted exceptions include competition with the introduced gecko Cyrtopodion scabrum in Texas, and the nonindigenous geckos Hemidactylus garnotii and Hemidactylus mabouia in south Florida (Klawinski et al., 1994; Punzo, 2001b). In both previously mentioned cases, H. turcicus appears to be competitively excluded and replaced in many locations. In Texas, the competitive failure of H. turcicus has been linked to the ability of C. scabrum to monopolize prey and force H. turcicus to undertake a dietary shift (Klawinski et al., 1994). In Florida, increased digestive and assimilation efficiencies, and continuous reproduction have been suggested as factors giving H. garnotii and H. mabouia a competitive edge over H. turcicus (Meshaka Jr., 1995; Punzo, 2001b). Study Sites My primary study area was the University of Florida campus located in the town of Gainesville in north central Florida. Since its official inception in 1906, the University of Florida has continuously expanded in size to become the fourth largest university in the

PAGE 17

7 United States. Following an extended construction boom starting in 1950 and ending in 1999, the 2,000-acre University of Florida campus has roughly 1,251 buildings that provide approximately 18, 670, 086 gross square feet of area. The University of Florida has just nearly 60,000 full and part-time students and employees, resulting in a vast amount of pedestrian and vehicular traffic. (University of Florida, 2002) My secondary study site was the Gainesville VA Medical Hospital Center within the boundaries of the University of Florida. Construction of the VA Medical Center commenced in 1964 and was completed in 1967. Presently it consists of 38 buildings (US Department of Veterans Affairs, 2003). The landscapes of both the University of Florida campus and the Gainesville VA Medical Center are highly heterogeneous, as both contain a wide assortment of facilities, each with its distinct type of architecture, ornamental vegetation and/or decorative natural scenery, walkway(s), road(s), and other associated features. This diversity is emphasized by the amount of variation present on each individual building, to the extent that it is a rare occurrence to encounter two identical walls. Walls are the principle habitat of Mediterranean geckos in this environment. Complexity depends on human factors such as stylistic trends, budgetary schemes, and logistics. This complexity was apparent during a preliminary survey that I undertook between the months of July 2001 and November 2001. A wide array of human activity continuously affected my sampling locations; these activities ranged from complete modification of existing vegetation, to the failure to repair damaged building structures. Thus, in an attempt to control for the fluctuating quality of campus buildings, I used categorical variables, as opposed to continuous variables, to quantify sampling locations.

PAGE 18

8 The Study Species in the Study Sites Records of H. turcicus for the University of Florida campus in Gainesville first began in 1956 (King, 1959). Although no official records of H. turcicus exist from the Gainesville VA Medical Center, the geckos presence on the premises was confirmed during a preliminary survey. Evidence of large populations of other nonindigenous hemidactyline geckos have not yet been recorded at either location, thus making these localities ideal for the investigation of the ecology of H. turcicus in the absence of potential behavioral changes caused by interference of directly competing species.

PAGE 19

CHAPTER 3 PRELIMINARY SURVEY Introduction Despite its widespread distribution and abundance on the campus of the University of Florida, the Mediterranean gecko has not been studied intensively. In 1956, King collected specimens of H. turcicus from a wood frame building on the campus. The following year, King and a colleague collected 49 H. turcicus individuals over three nights, also on frame buildings. A year later, Riemer collected an additional 44 specimens from the same buildings during one night of sampling (King, 1959). In an attempt to elucidate additional natural history patterns of H. turcicus, particularly with respect to habitat preference, and to obtain some general insight on their environment, I conducted a survey on the UF campus and VA Hospital, between July and August of 2001. Methods I randomly sampled 48 buildings over seven nights. I sampled each building by directing a flashlight systematically over all accessible walls from top to bottom, from right to left. I further characterized each wall by its type of construction material (one of four types: aluminum, brick, cement, and wood) and cardinal orientation. I arbitrarily used 50% of the surface area of a given wall for my material classification. Thus, I categorized a wall constructed with > 50% wood, as wood. The majority of the buildings were constructed from only one material. Some of the buildings consisted of more than 9

PAGE 20

10 one material. However, of these, I limited my survey to only those buildings that featured a predominate material. I determined the cardinal orientation of walls using official maps of the University of Florida (UF Physical Plant Division, 2000), and of the VA Hospital Engineering Department. For the cardinal orientation categories I used north, south, west, and east. Nearly all of the walls were clearly oriented in one of these directions. For those walls that were not clear-cut, I allowed a 45 angle of leeway on each side of the cardinal direction (Figure 3-1). For each gecko, I also recorded one of three vegetation levels. The three vegetation levels were all based on the height of the vegetation, rather than diversity. To facilitate measurement in the field, I based these heights using a simple system. Thus, the low level referred to flora no higher than my knees ( < 0.51 m), the medium level to floral height between my knees and shoulder ( > 0.51 m and < 1.37 m), and the high level comprised of flora reaching higher than my shoulders ( >1.37 m ). I limited my vegetation classification to an imaginary half sphere, coming out of the wall, with a onemeter radius and with the gecko as its center. I also determined light intensity with respect to each gecko, along the same lines as my vegetation classification. I used two categories to describe light intensity, again in an imaginary half sphere with a one-meter radius around each gecko. I assigned an area to the low light category if it had at most one dim light, whereas I classified an area as high light if there was either at least one bright light or at least two dim lights. The difference between dim and bright was fairly arbitrary; however, as a general rule of thumb, if I had to use a flashlight in the presence of a light to detect a gecko, I classified

PAGE 21

11 the area low light. A list of the criteria of these variables, and their associated levels is summarized in Table 3-1. A final variable I recorded for each gecko was the temperature of the center of each wall. I recorded the temperature using a Raytek Raynger ST model temperature gun. I also noted the number of geckos on each wall. Appendix A contains the complete dataset. I calculated averages, standard deviations and percentages where relevant. Results With respect to construction material, the average wall temperature was highest for brick (26.89 + 2.06 C; n = 16), followed by cement (26.14 + 1.42 C; n = 52), wood (25.13 + 1.38 C; n = 38), and aluminum (23.84 + 0.99 C; n = 78) (Figure 3-2). The average wall temperature for the four cardinal orientations is displayed in Figure 3-3. These average temperatures spanned a smaller range with north (25.17 + 1.89 C, n = 46) having the highest value, proceeded by east (25.03 + 1.74 C; n = 46), south (24.75 + 1.46 C, n = 47), and then west (24.5 + 1.80 C, n = 45). As summarized in Figure 3-4, brick walls of any cardinal location had a higher average temperature than any other construction material, while aluminum walls consistently had the lowest average temperature. Cement walls followed in second place for north, south and west facing walls, only to drop to third place for east facing walls. Wood walls, in turn, occupied third place for northern, southern and western walls, rising to second place for eastern walls. Brick walls reached their highest temperature on north-facing walls (27.16 + 3.1 C, n = 4), followed by west-facing walls (27 + 2.0 C; n = 4), east-facing walls (26.9 + 1.86 C; n = 4), and south-facing walls (26.42 + 1.92 C; n = 4). Average cement wall temperatures peaked on western walls (26.37 + 1.77 C; n = 13), and then continually decreased on northern walls (26.16 + 1.45 C, n = 13), western walls (25.82 + 0.89 C, n

PAGE 22

12 = 13), and eastern walls (24.02 + 1.55 C, n = 13). For wood walls, average wall temperature was highest on northern walls (25.41 + 1.66 C, n = 10), and then progressively decreased on eastern (25.17 + 1.29 C, n = 9), western (25.04 + 1.40 C, n = 9), and southern walls (24.89 + 1.31 C, n = 10) respectively. Aluminum walls attained their highest average temperature on west-facing walls (23.95 + 1.19C, n = 19), followed by, in decreasing order, west-facing walls (23.92 + 0.91 C, n = 19), east-facing walls (23.82 + 1.10 C, n = 20), and south-facing walls (23.65 + 0.77 C, n = 20). Of the 48 buildings I surveyed, five were brick, ten were wood, 13 were cement, and 20 were aluminum. All the brick buildings and cement buildings had geckos with totals of 27 and 40 individuals, respectively. Geckos were absent on one of the wood buildings, whereas the remaining nine contained 14 individuals. Aluminum buildings contained 30 individuals, although eight of these buildings had no geckos. I found the highest average number of H. turcicus on brick buildings with 5.4 geckos/building (27/5). Cement buildings had the second highest average with 3.08 geckos/building (40/13), whereas aluminum averaged 1.5 geckos/building (30/20), slightly ahead of wood, which averaged 1.4 geckos/building (14/10) (Figure 3-5). I sampled a total of 48 south walls, 47 north walls, 47 east walls, and 46 west walls. I found 86 geckos on north walls, resulting in an average of 1.83 geckos/wall (86/47). On south walls I tallied 76 geckos for an average of 1.58 geckos/wall (76/48), whereas walls oriented toward the east had a total of 49 geckos and averaged 1.04 geckos/wall (49/47), and west oriented walls had 47 geckos with an average of 1.02 geckos/wall (47/46) (Figure 3-6).

PAGE 23

13 I did not record the number of walls and/or areas within each type of vegetation level. However, of the 203 geckos that I recorded, 126 of them were located in areas that possessed a medium level of vegetation (62%). Meanwhile, areas bordered by low and high vegetation had considerably fewer geckos with 44 (22%) and 33 (16%), respectively (Figure 3-7). I also did not record the number of walls and/or areas of each type of light intensity level. I recorded a total of 235 geckos for this portion of the study. Of these 235 geckos, I observed 119 (51%) in areas with high light level and 116 (49%) in areas with a low light level (Figure 3-8). Discussion The materials brick, cement and wood all have similar average wall temperatures. The small difference in average wall temperature they display becomes irrelevant when their highly overlapping standard deviations are included. Aluminum possesses a lower average wall temperature, even when its standard deviation is considered. A walls cardinal orientation does not appear to have an effect on average wall temperature as all four directions have comparable temperatures, especially when assessed with their standard deviations. Each material has its own pattern of average wall temperature with respect to cardinal orientation. These results, however, should not be taken at face value, as the sample sizes are small and the standard deviations overlap each other. The thermal properties of walls are extremely complex and are only briefly mentioned here as they are beyond the scope of this study. Heat flow involves a variety of thermal parameters specific to the surface in question, such as its conductivity, convection capacity, and radiation constant. These parameters, in turn, are highly influenced by both climate conditions and the thermal property of the proximate

PAGE 24

14 environment (Nave, 2000). Additional work in describing the thermal habitat of H. turcicus is essential, as activity in ectothermic animals greatly depends on ambient temperature (Bartholomew, 1959). Furthermore, Frankenberg (1979) suggested that nocturnal animals are especially dependent on environmental temperature for activity since they cannot directly use the sun for thermoregulation. Brick and cement appear to be popular wall material types, as they both possess more H. turcicus per building than either aluminum or wood. However, the validity of these results rests on the assumption that I adequately sampled the buildings. I did not document the size of the buildings. Also, occurrence on a building constructed of a particular material type might be a result of other factors such as dispersal constraints rather than preference. A similar fate befalls the results dealing with the average number of H. turcicus on walls of different cardinal orientations. Although all four cardinal orientations share comparable averages, the areas of the walls were never recorded and thus, the results could be an artifact of this lack in rigor. Mediterranean geckos appear to prefer medium vegetation levels. Perhaps this indicates a balance between having sufficient vegetation for cover, but not too much foliage that it raises the vulnerability to predators. As for light levels, H. turcicus does not appear to have a preference for a particular light intensity. This was a surprising result as many researchers have reported that H. turicus has an affinity for lights because it aids in the capture of insect prey (Capula and Luiselli, 1994; Conant and Collins, 1998; Davis, 1974; Punzo, 2001a). This finding could imply that these geckos are merely easier to spot near lights, and that their attraction to light is a conclusion stemming from unintentional bias of the investigator. However, the reliability of my results is fairly

PAGE 25

15 limited, as I did not record the number of walls and/or areas of each vegetation and light level. Therefore, the pattern in my data might simply reflect a higher incidence of one vegetation and/or light level rather than a difference in gecko preference. Although the majority of these results have limitations, they served the function of illuminating some potentially interesting patterns, some of which merited further investigation. In addition, this preliminary survey allowed me to develop a reasonable and rigorous sampling regime for future work. First, it familiarized me with the cryptic coloration and secretive nature of the study species. Secondly, it revealed the numerous logistical considerations of the study site such as safety, accessibility, and the high degree of variation. Lastly, it encouraged me to make some important improvements on my sampling methods. Among these changes are the inclusion of wall size measurements, and the sampling of walls rather than buildings so as to avoid the possibility of counting the same gecko more than once. Also, to deal with the immense variability resulting from constant human manipulation, vegetation and light need to have broad and quantifiable levels. These levels, in turn, must pertain to the overall habitat and not the immediate vicinity of the gecko, as this vicinity will change with movement. Ultimately, this initial survey would prove to be key in shaping all aspects of my eventual microhabitat study.

PAGE 26

16 Table 3-1. Criteria of wall characterization variables used in preliminary survey Variables Levels Criteria North *Location of wall on official UF map South *Location of wall on official UF map West *Location of wall on official UF map Cardinal Orientation East *Location of wall on official UF map Low Presence of one dull light** or no light source within a 1m radius of the gecko Light High Presence of at least two dull light** sources or at least one bright light*** source within a 1m radius of the gecko Aluminum Physical observation; >50% of building surface area Brick Physical observation; >50% of building surface area Cement Physical observation; >50% of building surface area Material Wood Physical observation; >50% of building surface area Low Vertical measurement < 0.51 m within a 1m radius of the gecko Medium 0.51 m 1.37 m within a 1m radius of the gecko *Maps taken from 2000 Building Information List for the University of Florida, prepared by the UF Physical Plant Division, and official VA Hospital Engineering maps ** A dull light source is one where a flashlight is still needed to locate a gecko ***A bright light source is one where a flashlight is not needed to locate a gecko

PAGE 27

17 N 45 45 W E S Figure 3-1. Depiction of the 45 angle leeway employed in the determination of wall cardinal orientation 23.8426.8926.1425.1322232425262728AluminumBrick CementWoodConstruction MaterialAverage Wall Temperature (0C ) Figure 3-2. Average wall temperature of different construction material

PAGE 28

18 25.1724.7524.525.032022242628North SouthWest EastCardinal LocationAverage Wall Temperature (0C ) Figure 3-3. Average wall temperature of different cardinal locations 26.3723.9523.6523.9223.8427.1626.92726.4224.0225.8226.1625.0425.1724.8925.412021222324252627282930North SouthWest EastCardinal LocationsAverage Wall Temperature (0C ) Aluminum Brick Cement Wood Figure 3-4. Average wall temperature of different construction materials at different cardinal locations

PAGE 29

19 3.085.41.41.50123456CementBrickAluminumWoodConstruction MaterialAverage Number of Geckos Figure 3-5. Average number of H. turcicus per building of different construction material 1.581.021.041.8300.511.522.53NorthSouthWest EastCardinal LocationNumber of Geckos Figure 3-6. Number of H. turcicus on walls of different cardinal locations

PAGE 30

20 4412633020406080100120140LowMediumHighVegetation LevelNumber of Geckos Figure 3-7. Number of H. turcicus on walls of different vegetation levels 116119050100150LowHighLight IntensityNumber of Geckos Figure 3-8. Number of H. turcicus on walls of different light intensities

PAGE 31

CHAPTER 4 ADDITIONAL NATURAL HISTORY NOTES Introduction A large portion of the information regarding H. turcicus pertains to the reproductive cycle and associated activities. Few studies have documented natural history variables such as perch height, degree of sociality and exposure. And not much is known about any preferred temperature regime. Perch height in H. turcicus has only rarely been investigated. The only comprehensive study was conducted on a university campus in Texas, where Saenz (unpublished) found substantial dietary diversity among geckos at different perch heights. The diet of H. turcicus encountered below 1.52 m in height overlapped only 22.71% (Schoeners percent overlap; Schoener, 1970) with conspecifics occupying a perch over 3.05 m in height. Specifically, the geckos with lower perches ingested mostly ground-dwelling prey, whereas those at higher perches fed predominantly on flying insect taxa. In general, females tended to use perches of lower height than males, as 55.8% of females were recorded below 1.52 m in comparison to 30.2% of males. Meanwhile, 41.5% of males were recorded above 3.05 m as opposed to only 11.6% of females. Geckos captured at different perch heights also demonstrated a difference in the number of empty stomachs, with 13.04% for low geckos and 25% for high geckos (Saenz, unpublished). Other perch height studies include one by Capula and Luiselli (1994) in Rome, Italy. These authors concluded a close-to-the-ground existence for H. turcicus, as they showed that its diet consisted mainly of ground-dwelling prey, with 55.2% of the geckos 21

PAGE 32

22 diet being made up of ants and flightless insects. In addition, Klawinski (unpublished) observed a large number of Mediterranean geckos positioned close to the ground (40.95%), a result he felt stemmed from the need for shade from surrounding lights. In New Orleans, Rose and Barbour (1968) noticed several geckos on the third-floor level of a building, and also on the roof of another. Both Selcer (1986) and Klawinski (unpublished) recorded low average home range areas of 0.93 m2 and 4.073 m2, respectively, with very little home-range overlap. These results suggest that this species is territorial. Furthermore, observations have indicated that male H. turcicus emerge from winter retreats earlier then females, perhaps to establish territories before the breeding season (Klawinski, unpublished). In addition, Rose and Barbour (1968), Frankenberg (1982), Marcellini (1977), Klawinski (unpublished) have all witnessed aggressive displays ranging from tail waving to neck biting. A study of the vocal activity of H. turcicus revealed that only the dominant male in a group produces a multiple click call in response to an intruder of either sex (Frankenberg, 1978). In this same study, Frankenberg (1978) found that most of the vocalization in H. turcicus occurred during the day, a time when this gecko is grouped together in retreat-sites. This result indicates that social activity is perhaps separate from this species nocturnal foraging. In turn, this division between sociality and foraging could explain the peaceful interaction between two male H. turcicus behind a drainpipe witnessed by Rose and Barbour (1968). The degree of exposure once H. turcicus has emerged from its daytime retreat has not been formally studied. Rose and Barbour (1968) observed H. turcicus behind vertical

PAGE 33

23 storm drains. Through personal observations, I have noted geckos behind electrical boxes, pipes, and signs on walls. Few studies have focused on H. turcicus with respect to temperature. An exception is the study by Angilletta et al. (1999) where the body temperature for eight Mediterranean geckos was measured to be 27.8 C in the morning, and 29.1 C in the evening. Hoping to shed further light on the natural history of the Mediterranean gecko, I explored perch height, sociality, exposure, and selected surface temperature in three distinct field surveys. Methods The first survey took place during the months of July and August 2001 whereby I randomly sampled walls on 48 one-story buildings on the University of Florida campus and the VA Hospital. I systematically sampled each wall by scanning it with a flashlight from top to bottom, left to right. I classified each gecko I encountered as either an adult (greater than 40 mm) or subadult (less than 30 mm). Note that I omitted any geckos that I could not accurately size from any analysis. Lastly, I recorded the perch height of each gecko using two categories; low if the gecko was at most one meter above the ground, and high otherwise. Appendix B contains the complete dataset. In the second survey, I sampled 50 one-story walls located on the University of Florida campus and the VA Hospital. Between September 2001 and January 2002, I sampled each of these 50 walls on a weekly basis. Upon each visit, I again examined a wall by methodically examining it with a flashlight from top to bottom, left to right. Geckos were sized according to the method mentioned previously. In addition to documenting perch height (as above), I also quantified sociality, exposure, and selected

PAGE 34

24 surface temperature. I used three categories to describe sociality: a gecko that had no other individual within a 50 cm radius was termed alone, while two geckos within the same radius were designated a pair, and three or more geckos were considered a group. Furthermore, I labeled a gecko as exposed if it was in plain sight, and not exposed if the snout-vent portion of its body was hidden behind a wall fixture. Using a Raytek Raynger ST model temperature gun, I determined the selected surface temperature by measuring the temperature of a spot adjacent to the gecko. Appendix B contains the complete dataset. For the third survey, also on the University of Florida campus and the VA Hospital, I sampled 160 one-story walls between the months of March and June 2002. I visited each wall twice throughout the study. The sampling regime and equipment I used were identical to those previously mentioned in the second study. Appendix B contains the complete dataset. A list of the criteria of these variables, and their associated levels is summarized in Table 4-1. I calculated percentages, averages, standard deviations, maximum values and minimum values where appropriate. Results Summer 2001 During this study, I recorded a total of 187 gecko observations. Of these, 125 (67%) occurred at a high height, whereas 62 (33%) were at a low height. Upon separating these observations into adults and subadults, two distinct patterns emerged. Of the 131 adults I sampled, 97 (74%) were located high on walls and 34 (26%) were observed at a low height. The opposite was true for subadults, as 38 (84%) out of the 45

PAGE 35

25 recorded were found within one meter of the ground whereas the remaining 7 (16%) were located at a high height (Figure 4-1). Fall/Winter 2001 I used 576 gecko observations to test for perch height preference; 412 (71.5%) occupied high positions, whereas 164 (28.5%) occupied low positions. Of these 576 gecko observations, 355 were adults of which 296 (83.4%) were at a high position and 59 (16.6%) were at low positions. I recorded 221 subadult observations, where 116 (52.5%) were at a high position, and 105 (47.5%) were at a low portion of a wall (Figure 4-2). I used 577 gecko observations to investigate sociality. Of these, 506 (87.7%) geckos were alone, 56 (14.9%) were part of a pair, and 15 (4.2%) were part of a group. I counted 355 adult gecko observations where 287 (80.9%) were alone, 53 (14.9%) belonged to a pair, and 15 (4.2%) belonged to a group. Following a comparable trend, the 222 subadult observations resulted in 219 (98.7%) alone counts and 3 (1.3%) pair counts. No (0%) subadults were seen in any groups (Figure 4-3). Of 574 exposure observations, I tallied 397 (69.2%) as being exposed and 177 (30.8%) as not exposed. Further breakdown of these results revealed that 199 (56.7%) of the 351 adult observations were exposed, whereas 152 (43.3%) were not exposed. Of the 223 subadult observations I recorded, 198 (88.8%) fell in the exposed group, which contrasts with the 25 (11.2%) observations that I placed in the not exposed group (Figure 4-4). As summarized by Table 4-2, the average substrate temperature for adults was 23.22 C ( + 3.05), with a maximum value of 36.3 C and a minimum value of 12.7 C. Likewise, the average substrate temperature for subadults was 22.67 C ( + 2.80), with the maximum and minimum values being 30.8 C and 13.9 C, respectively.

PAGE 36

26 Spring 2002 I observed few subadults in this study, so I included only adult observations. With respect to perch height, I obtained 237 observations, with 170 (71.7%) at high positions and 67 (28.3%) at low positions (Figure 4-5). For sociality, I had 236 observations, of which 223 (94.5%) were alone, 13 (5.5%) belonged to a pair, and none (0%) were part of a group (Figure 4-6). I collected 237 exposure observations, of which 178 (75%) fell into the exposed category, and 59 (25%) into the not exposed category (Figure 4-7). The average substrate temperature for adults was 24.89 C ( + 2.94). The maximum temperature value I measured was 31.5 C, whereas the minimum value was 16.4 C (Table 4-3). Discussion Adult H. turcicus consistently occupied wall habitats that were greater than one meter above the ground. This trend was observed regardless of the time of year. Although not quantified, the majority of the high sightings were located in the upper portion of the walls, in close proximity to the roof awning. This high perch height could be beneficial for escaping predators, as I repeatedly witnessed startled geckos escape into crevices in the roof awning. However, it is important to mention that this result could be an artifact of the difference in surface area between the two height categories; the low category is restricted to a significantly smaller area than the high category. Thus, the greater number of adults located in high wall positions may be directly related to the greater area available. Interestingly, despite this disparity in area, subadults were recorded on wall habitats that were a maximum of one meter above the ground. This result was more pronounced

PAGE 37

27 during the summer survey, although it was still evident during the fall/winter sampling period. This tendency for subadults to occupy a low habitat could stem from a variety of reasons. First, the portion of a wall near the ground might not be an optimal habitat as it could increase a geckos vulnerability to predators. Subadults might be obliged to use this less desirable habitat as a result of being out-competed by adults for the high optimal ones. This scenario would be congruent with the significantly higher mortality rate found in geckos measuring less than 30 mm (Selcer, 1986). Secondly, the subadult age period could be the dispersal stage of H. turcicus. If subadults were the dispersers, they would frequently be on the lower portion of a wall as they would be continuously on the move. This idea might be supported by some circumstantial evidence that I have witnessed during the course of this study; on several occasions I have observed subadults on the ground some distance away from any building/wall. In fact, one particular individual was recorded in the middle of an expansive cement parking lot. Also, Rose and Barbour (1968) showed that hatchlings could survive without food or water for up to one month. This resilient quality of subadults would be ideal for the uncertainties of dispersing. Furthermore, this idea would be compatible with Selcers findings (1986), as dispersal would be expected to make subadults more vulnerable to predation and thus increase their mortality rate. Thirdly, subadult preference for low wall habitats might be a consequence of diet. Perhaps the preferred prey of subadults and/or prey size suitable for small mouths is more abundant on low dwellings. This is highly possible, as Saenz (unpublished), upon conducting a detailed dietary study on H. turcicus, concluded that the geckos height on a

PAGE 38

28 wall greatly influenced a geckos diet. Finally, the actual cause behind subadults occupying a low habitat could be any one of the three mentioned hypotheses, or a combination of these, or even still, none of these. During the summer survey, a greater proportion of subadults in my sample used perches at a low height than during the fall/winter survey. This discrepancy could be attributed to a number of factors such as seasonal changes in prey consumption, a decrease in adult competition due to adult turnover and/or the decrease of breeding activities, an increase in the establishment of subadults on walls due to a decrease in dispersal, an increase in subadult population, a combination of these, or none of the above. Sociality in the Mediterranean gecko during nocturnal foraging appears to be quite minimal. Both adults and subadults preferred being alone. This result is in agreement with the belief, held by many investigators, that H. turcicus is largely territorial. The majority of the pair and group observations involved only adults. Group sizes rarely exceeded three. I observed no aggressive display in any of these observations. Although no copulation was witnessed, perhaps the gecko pairings were associated in some way with breeding. Gecko groupings could also be linked to other stages of reproduction, as communal nesting has been documented in this species (Selcer, 1986). Since geckos were not sexed in this study, there is no way of telling if the pairs and groups consisted of same or different sex individuals. The tendency for H. turcicus to be solitary does not necessarily imply that this gecko species is not social. It has been shown that other nocturnal gecko species (Nephrurus milii and Christinus marmoratus) form large, non-random aggregations within retreat-sites (Kearney et al., 2001). The possibility that H.

PAGE 39

29 turicus displays this behavior has been alluded to by Frankenberg (1978), who showed that most vocalization in this species occurs in daytime retreats. Thus perhaps, H. turicus socializes during the daytime within retreat-sites, and then forages in solitude during the night, occasionally interacting with others for reproductive purposes and/or for prey exploitation. In general, the Mediterranean gecko remains exposed once it emerges from its retreat-site at night. Since the number of possible hiding places is extremely difficult to quantify and/or locate, this result could reflect my inability to find all hidden geckos and thus be skewed. In addition, this result could also simply be a product of availability of hiding places rather than preference. Subadults showed a greater propensity for remaining exposed than adults. This difference might stem from adult competition for hiding spaces, which could be intense if these spaces offer a significant increase in predator protection. This subadult trend might also be an artifact of a lower incidence of hiding places situated in the bottom portion of walls. The field temperatures of the wall surface selected by both adult and subadult H. turcicus were comparable. Although these temperatures are lower than previously measured field body temperatures (29.1 C, n = 8), it is not entirely surprising as some species of nocturnal geckos have been known to thermoregulate and achieve their preferred body temperature during the day in their retreats rather than at night (Angilletta et al., 1999). An important point when considering these results is the high possibility of pseudoreplication in the data. For the fall/winter 2001 survey, I visited the same 50 walls every week, whereas in the spring 2002 survey I visited the same 160 walls twice. These

PAGE 40

30 two sampling methods did not allow me to distinguish among individuals, and it is likely that I counted the same individual many times. Thus, conclusions should be made with caution. However, despite the fact that these results are limited in scope, they nevertheless provide some insight for future investigators.

PAGE 41

31 Table 4-1. Criteria of natural history variables Variable Levels Criteria Low Gecko < 1m above the ground Height High Gecko > 1m above the ground Alone 1 gecko within a 50cm radius Pair 2 geckos within a 50cm radius Sociality Group 3 or more geckos within a 50cm radius Exposed Snout-vent portion of gecko in plain sight Exposure Not Exposed Snout-vent portion of gecko not in plain sight Table 4-2. Temperature measurements of wall surface for adult and subadult H. turcicus; fall/winter 2001 Temperature Measurements Adult Subadult Average 23.220C 22.670C Standard Deviation 3.050C 2.800C Maximum Value 36.300C 30.80 0 C Minimum Value 12.70C 13.90 0 C Table 4-3. Temperature measurements of wall surface for adult H. turcicus; spring 2002 Temperature Measurements Adult Average 24.890C Standard Deviation 2.940C Maximum Value 31.50C Minimum Value 16.40C

PAGE 42

32 74%16%26%84%0%10%20%30%40%50%60%70%80%90%100%AdultSubadultPerch HeightPercent Geckos Observed Low High Figure 4-1. Perch height preference of adult and subadult H. turcicus (summer 2001)

PAGE 43

33 16.6%47.5%83.4%52.5%0%10%20%30%40%50%60%70%80%90%100%AdultSubadultGecko AgePercent Gecko Observed High Low Figure 4-2. Perch height preference of adult and subadult H. turcicus (fall/winter 2001)

PAGE 44

34 80.9%98.7%14.9%1.3%4.2%0%10%20%30%40%50%60%70%80%90%100%AdultSubadultGecko AgePercent Geckos Observed Group Pair Alone Figure 4-3. Social preference of adult and subadult H. turcicus (fall/winter 2001)

PAGE 45

35 43.3%11.2%56.7%88.8%0%10%20%30%40%50%60%70%80%90%100%AdultSubadultGecko AgePercent Geckos Observed Exposed Not Exposed Figure 4-4. Exposure preference of adult and subadult H. turcicus (fall/winter 2001)

PAGE 46

36 72%28%0%10%20%30%40%50%60%70%80%90%100%HighLowPerch HeightPercent Geckos Observed Figure 4-5. Perch height preference of adult H. turcicus (spring 2002)

PAGE 47

37 0.0%5.5%94.5%0%10%20%30%40%50%60%70%80%90%100%AlonePair GroupSocialityPercent Geckos Observed Figure 4-6. Social preference of adult H. turcicus (spring 2002)

PAGE 48

38 25%75%0%10%20%30%40%50%60%70%80%90%100%YesNoExposurePercent Gecko Observed Figure 4-7. Exposure preference of adult H. turcicus (spring 2002)

PAGE 49

CHAPTER 5 MICROHABITAT PREFERENCE IN THE INTRODUCED GECKO, HEMIDACTYLUS TURCICUS Introduction Although frequently seen on buildings in both its native and nonnative range, little is known of the microhabitat preferences of H. turcicus. Habitat studies on this species are largely nonexistent, and the little information available has mostly originated as incidental observations made during other studies. The only direct study, conducted in Rome, Italy, was a comparative study between H. turcicus and the sympatric gecko species Tarentola mauritanica. In this study, Luiselli and Cappizzi (1999) found that H. turcicus was more abundant on recently constructed buildings than debilitated ancient buildings dating back to the Roman Empire. Although this study provided some quantitative information on microhabitat preference in H. turcicus, the confounding effects of competition influenced the conclusions. Evidence of this is demonstrated by a survey Capula and Luiselli (1994) conducted in Rome years earlier, which concluded that H. turcicus was particularly common in Roman age archeological sites. Additional information consists primarily of incidental observations, and is anecdotal in nature. Although the natural habitat of H. turcicus may have been rocky cliffs, the primary habitat today appears to be structures associated with human habitation (Arnold, 1984). Investigators have reported seeing H. turcicus on a number of human-made constructions such as rock walls, burial vaults, and buildings of varying material 39

PAGE 50

40 including granite, cement, wood, metal, and stucco (Davis, 1974; Klawinski, unpublished; Meshaka Jr., 1995; Punzo, 2001a; Rose and Barbour, 1968; Saenz, unpublished; Selcer, 1986). In Texas, Selcer (1986) found that H. turcicus occurred at higher densities on brick versus metal structures, whereas in Florida Punzo (2001a) found higher densities of H. turcicus on wood as opposed to metal buildings. Investigators have always linked artificial light to the presence of H. turcicus on buildings, as lights presumably facilitate the capture of their insect prey (Capula and Luiselli, 1994; Conant and Collins, 1998; Davis, 1974; Punzo, 2001a). In Texas, Davis (1974) reported that H. turcicus preferred buildings that were lit by mercury-vapor lights. However, a number of studies have collected H. turcicus from buildings with varying light intensities, including complete darkness (Klawinski, unpublished; Meshaka Jr., 1995). Vegetation is another factor that has often been associated with Mediterranean gecko habitat. Throughout its native range, H. turcicus has been found commonly on trees (Loveridge, 1947). In its introduced range, buildings inhabited by H. turcicus have possessed grass, shrubs and/or trees in close proximity to its walls (Klawinski, unpublished, Saenz, 1996). Vegetative cover has been hypothesized by Saenz (1996) to provide H. turcicus with retreats. In a separate study, Klawinski (unpublished) found a weak association between the occurrence of H. turcicus on walls with both high light intensity and high vegetative cover. The majority of these findings are qualitative observations. The few results supported by quantitative data often lack a rigorous framework, making conclusions difficult to formulate. Thus, I embarked on a systematic study of microhabitat

PAGE 51

41 preference, in an attempt to contribute quantitative baseline data on the Mediterranean gecko. Specifically, I investigated the previously studied variables: construction material; vegetative cover; and light intensity. Furthermore, I examined cardinal location, building age, surface texture, surface color, and wall length, as I considered these characteristics to be potentially important to the habitat preference of this species. Methods and Results Sampling Methods Between the months of March and June of 2002, I conducted a survey detailing the microhabitat preference of H. turcicus on the University of Florida campus and VA Hospital in Gainesville. I selected 160 buildings according to their accessibility, and, for the sake of accurately detecting geckos, being one-story in height. I sampled only one randomly selected wall per building. I used a die to make my selection; the numbers 1, 2, 5, and 6 indicated north, south, west and east, respectively, whereas 3 and 4 denoted rolling the dice over again. Walls were considered a good representation of microhabitat use in this species, as H. turcicus has been shown to possess a small home range; Rose and Barbour (1968) reported an average recapture distance from initial capture site of 5.7 m, while Selcer (1986) estimated the mean range movement to be 0.93 m. Furthermore, Trout and Schwaner (1994) reported that H. turcicus maintains itself in discrete subpopulations in which differences in allele frequency have been found between populations only 100 m apart. Each selected wall was characterized by building material, presence or absence of light sources, surface color, surface texture, age of building, vegetation level, cardinal orientation, and length. Construction material was confined to four types: aluminum, brick, cement, and wood. Identical to my preliminary survey, I arbitrarily used 50% of the surface area of a

PAGE 52

42 given wall for my material classification. I restricted my survey to walls that featured a predominate material. I used two categories of light intensity, high and low. To account for gecko movement, light intensity was measured with respect to the wall as opposed to a small area around a gecko. Specifically, I classified walls as high light if they possessed at least one light source (any brightness), whereas walls containing no light source I assigned to the low light category. I classified wall color into two levels using a 3-inch by 5-inch white index card. I categorized a wall as dark if, upon fastening the index card to the wall, I could distinguish it at a perpendicular distance of 10 feet (3.048 m) in daylight. The opposite was true of walls appointed to the light level; the index card could not be perceived at a perpendicular distance of 10 feet. All the walls I surveyed were uniformly colored across their entire surface. I also quantified wall texture using two levels. I considered a wall to be smooth if I was able to draw a straight line, roughly10 cm in length, on white printer paper propped on three distinct points on a wall. These three positions were subjectively selected as the left edge, right edge, and middle point of the wall, approximately mid-wall in height. If I was unable to draw a straight line at all three points on a wall, I classified the wall as rough. Note that all lines were drawn with a relaxed handgrip. I determined building age information from literature provided by the University of Florida (UF Physical Plant Division, 2000), and an unofficial list created specifically for this study by the Veterans Administration Hospital Engineering Department. I arbitrarily assigned three age groups, all based on these sources. I categorized walls built between

PAGE 53

43 1900 and 1969 as early, between 1970 and 1989 as modern, and between 1990 and 2002 as contemporary. I ignored any possible renovations. The age of a building was used as an approximation for the number of daytime retreats (cracks and/or crevices) since Luiselli and Capizzi (1999) found that the age of a building and the condition of its walls were highly correlated. The necessity of using this approximation arose when, during my preliminary survey, it became apparent that estimating the number of retreats with the naked eye was highly unreliable. I classified walls into one of three vegetation levels. The three vegetation levels were based on the cement/vegetation ratio bordering the wall, rather than the diversity or height of the vegetation. I quantified the vegetation in this fashion because of the unpredictable management techniques encountered during my preliminary survey; primarily, workers constantly altered vegetation variety and height. Thus, the cement level referred to a wall where at least 60% of the length was bordered by cement, the mix level to a wall whose length was bordered more than 40% but less than 60% by either cement or vegetation, and the vegetation level to a wall whose length was bordered at least 60% by vegetation. I determined the cardinal orientation of walls using the same methods discussed in my preliminary survey. Using the 2000 Building Information List for the University of Florida and official maps of the VA Hospital Engineering Department, I classified walls as north, south, west, or east. For walls that were not clearly oriented in one of these directions, I allowed a 45 angle of leeway on each side of the cardinal direction. Figure 3-1 in chapter 3 has further details.

PAGE 54

44 I described length by means of three randomly created categories. I classified walls as small if their greatest length measured less than 20 m, as medium if their length was between 20 m and 40 m, and as large if they were 40 or more meters in length. Length was used as a general measure for size, since all walls had roughly the same height (one-story). A list of the criteria of these variables, and their associated levels is summarized in Table 5-1. My sampling regime consisted of visiting 10 walls per night, on two nights per week. Each visit occurred approximately two hours after sunset, which according to King (1959) is a period of high activity for H. turcicus in Gainesville. The number of walls I inspected per night was limited to 10 to keep sampling duration under two hours. This was done to homogenize weather conditions among walls. The four months I selected for the survey period coincided with part of the reproductive season of H. turcicus and further ensured gecko activity (Selcer, 1986). Thus, I examined each of the 160 walls twice for completeness, once during March/April and once during May/June. I sampled each wall by passing a flashlight systematically across the entire surface, going from right to left, top to bottom. I recorded the presence or absence of H. turcicus for each wall. I pooled data from the two visits, and considered a wall to have a gecko if at least 1 gecko was present during at least one of the two visits. The complete dataset is presented in Appendix C. Data Analysis 1 The data were initially analyzed using either a chi-square test or a Fisher's exact test in order to detect dependency among the eight variables of interest. The Fishers exact test is based on the hypergeometric distribution, rather than a chi-square distribution; when testing for independence, the p-value is obtained by adding the

PAGE 55

45 probabilities of outcomes as favorable to the alternative hypothesis (dependence) as the observed outcome (Agresti, 1996). I then used logistic regression in an effort to model the wall variables to the presence/absence of H. turcicus on a wall. I selected logistic regression because it functions on binary data, which was the format of the data I had collected. Only those variables deemed independent from the chi-square tests were used as predictors for the logistic regression in an attempt to avoid multicollinearity in the model. I determined the logistic regression model by using the backward elimination method (Agresti, 1996). In this method, one essentially begins with the full model, containing all possible variables and the interactions between them, and then systematically removes one term at a time, starting with the highest-order term. With the removal of each term, the deviation (G2 test of goodness of fit) of the new model is compared to that of the full model. Removals continue until the difference in deviation between the two models either reaches a specified value determined by the investigator and/or the difference reaches a large enough value that it becomes significant. Once significance is attained, further term removal would result in losing the integrity of the information provided by the dataset. In other words, this method is a balance between simplifying the model, and preserving a sufficient amount of the dataset information. Therefore, the number of retained terms is directly related to the objective of the investigator (Agresti, 1996). In this case, I decided to choose integrity of information over simplification, and thus I elected to use the simplest model that possessed the smallest difference in deviation from the full model.

PAGE 56

46 Once I had chosen an appropriate model, I calculated the odds ratio and a 95% confidence interval of the odds ratio for each significant predictor. The odds ratio (), as defined by Agresti (1996) is the ratio of the odds of two events ( = odds1/ odds2), where the odds of each event is defined as the odds of success for that event. Thus for example, if the probability of success for event 1 is 0.75, then odds1= probability of success/probability of failure = 0.75/1-0.75 = 0.75/025 = 3. This signifies that success in event 1 is three times as likely as failure. Finally, if odds1= 3 and odds2 = 4, then the odds ratio = 3/4 = 0.75, which indicates that the odds of success for event 1 is 0.75 the odds of event 2. To put it into perspective, if the odds of finding a gecko on aluminum walls is 3, while the odds of finding a gecko on wood walls is 4, then the odds ratio of aluminum walls to wood walls would be 0.75; this signifies that the odds of finding a gecko on aluminum walls is 0.75 that of finding a gecko on wood walls. The odds ratio for each predictor was calculated by taking the exponent of the predictors estimate (eB, where B is the predictor estimate). I then calculated a 95% confidence interval for each odds ratio by finding the lower and upper bounds, and then taking their exponents; the bounds were found using the equation Bi +/1.96 (ASE), where Bi denotes the estimate of the predictor in question and ASE stands for the estimates asymptotic standard error. Thus, the 95% confidence intervals take the following form: (eBi-1.96 (ASE), eBi+1.96 (ASE)) (Agresti, 1996). In all of my statistical analyses, the significance value was set at the 0.05 level. All statistical analyses were performed using SAS, version 8.2. Results 1 The chi-square and the Fishers exact analyses revealed a number of associations among the variables. Specifically, material was highly dependent on building age, color,

PAGE 57

47 length, and texture. Furthermore, light was dependent on both length and vegetation, while color was dependent on both age and length. Table 5-2 contains additional information, including p-values. To avoid multicollinearity in the eventual model, I omitted the variables age, color, length, and texture from further analysis since they could be accounted for by the variable material. Table 5-3 describes the relationship between material and age, color, length, and texture. I fitted a logistic regression model using the variables cardinal orientation (CO), light (L), material (M), and vegetation (V). The backward elimination method supported the use of the interaction variables M*CO*V + M*CO*L model instead of the full model M*CO*V*L. The M*CO*V + M*CO*L model explained the same amount of variability as the full model (the difference in deviation was zero), but with fewer terms, thus making it easier to interpret. Table 5-4 presents additional details on the backward elimination method. The expanded, symbolic version of the logistic regression model M*CO*V + M*CO*L is the Equation 5-1. (Eq. 5-1) Y =M*CO*V + M*CO*L + M*CO + M*V + M*L + CO*V + CO*L + V*L + M + CO + V + L The M*CO*V + M*CO*L model is modeling the probability that y = 1, i.e. that H. turcicus is present on a wall. Table 5-5 contains the estimate, asymptotic standard error, p-value, odds ratio, and 95% confidence interval of the odds ratio for all of the significant predictors. The numerical representation of the model is the Equation 5-2. (Eq. 5-2) Y= 68.3825 68.3825m1 91.7478m2 46.1655m3 + 18.3578v2 + 109.586co1 45.0170co2 67.2839l1 + 131.7139 m1co1 + 44.3238 m1co2 17.0373m2co3 18.3581v2 co2 Where m1 = 1 = Aluminum, and 0 otherwise

PAGE 58

48 m2 = 1 = Brick, and 0 otherwise m3 = 1 = Cement, and 0 otherwise m1 = m2 = m3 = 0 = Wood v1 = 1 = Cement, and 0 otherwise v2 = 1 = Mix, and 0 otherwise v1 = v2 = 0 = Vegetation co1 = 1 = East, and 0 otherwise co2 = 1 = North, and 0 otherwise co3 = 1 = South, and 0 otherwise co1 = co2 = co3 = 0 = West l1 = 1 = No light, and 0 otherwise l1 = 0 = Light However, parameter estimates of this model were unusually high or unusually low, resulting in equally extreme odd ratios. This suggests that despite this model being the best one my input could generate, it was inadequate as the estimates were too unrealistic. Although there is no written rule on estimate magnitude, the magnitudes of my estimates were so large as to render them unpredictable and useless (Ken Portier, personal communication). In general, I have noticed odds ratios take values between zero and five, occasionally larger but never greater than 10. Thus, I recommend using a value of 10 as a cut-off point for odds ratios. Accordingly, anything above 10 should be used with caution, and definitely examined further. With respect to a low boundary for odds ratio values, the mathematical minimum is zero. However, values close to zero that take the form of a low order decimal should also be considered with caution, and investigated further. Data Analysis 2 In an attempt to determine if the difficulty encountered in my first analysis was a result of the model selected, I fitted all of the possible models, even those that would explain less overall variability. Specifically, these were

PAGE 59

49 M*CO*V+ M*CO*L+ M*V*L + CO*V*L M*CO*V+ M*CO*L+ M*V*L M*CO*V+ M*CO*L + CO*V*L M*CO*V+ M*V*L + CO*V*L M*CO*L + M*V*L + CO*V*L M*CO*V+ M*V*L M*CO*L + M*V*L M*CO*L + CO*V*L M*CO*V M*CO*L These models correspond to #2, #3, #4, #5, #6, #8, #9, #12, #13, and, #14 in Table 5.4. In addition, I constructed a table of all the wall combinations I encountered during my survey in order to verify that my dataset was appropriate for logistic regression; too few observations per wall combination could cause the model to produce unrealistic estimates. Results 2 All 10 additional fitted models produced similar estimate and odds ratio values, either extremely high or extremely low in magnitude. In other words, these 10 models were inadequate too. Thus, this indicated that model selection was not the source of the problem. A possible cause of the problem could have been the large number of zeros and low among the various wall combinations (Table 5-6). The wall combination table revealed that of 96 possible wall combinations, 32 were never encountered during this study, and an additional 32 combinations occurred just once. Indeed, only 10 combinations of the 96 possible were represented by more than five occurrences, although none over 10. These results strongly point toward the conclusion that this dataset is not sufficiently large enough to accommodate logistic regression (Agresti, 1996; Ken Portier, personal communication).

PAGE 60

50 Data Analysis 3 Following a common statistical practice, I collapsed the wall combination table into a smaller, more concise table with fewer variables and/or levels. This can be accomplished most simply by removing and/or combining. Therefore, I removed cardinal orientation, as it was the least repeatable of all the variables; compass readings are linked to the magnetic poles, which are constantly changing locations (Natural Resources Canada, 2003). Furthermore, I decided to combine the cement and mix vegetation levels to increase cell numbers and eliminate zeros within the table. Thus, vegetation was described by two levels; the cement level referred to walls whose length were bordered more than 40% by cement, whereas the vegetation level referred to walls whose lengths were bordered at least 60% by vegetation of any species and height. Although there existed a number of distinct ways I could have collapsed my wall combinations table, I believe that the arrangement I selected was ecologically, the most parsimonious one. I then fitted a logistic regression model with the backward elimination method. I used the variables material (M), light (L), and the newly described vegetation (V). Estimates and odds ratios for parameters were calculated where pertinent. Results 3 The wall combination table resulting from the removal of cardinal orientation and the merging of two vegetation levels was significantly condensed; instead of 96 cells, this new table totaled only 16 (Table 5-7). This decrease in the number of cells eliminated zeros from the table, although six of the remaining 16 (37.5%; 6/16) values were less than five. The backward elimination method resulted in 16 logistic regression models, ranging from the highly complex three-factor interaction model to the simple no-factor

PAGE 61

51 model (Table 5-8). Theoretically, all 16 models were found to be functional, each having its own balance between simplicity and the amount of variability explained. Despite all models being usable, only the model M*V*L contained significant parameters. However, these parameter estimates and their resulting odds ratios were again found to be unusually high or unusually low. Although the remaining models had typical parameter estimates and odds ratios, none of the parameters were significant; in other words, these models would contain zero parameters, since only significant parameters are included in a model, thus these models were impractical. These results show that logistic regression models continue to crash with these data. Data Analysis 4 In a final attempt to obtain a successful logistic regression model, I combined the light and vegetation variables into one variable with three levels. Specifically, I pooled both vegetation levels under the high light level, and left the two vegetation levels under the low light level unchanged. Thus, I described the new light-vegetation variable (LVEG) as HCV (high/cement-vegetation) if a wall was characterized by high light and any kind of vegetation, as LC (low/cement) if a wall was classified as low light and cement, and as LV (low/vegetation) if a wall was classified as low light and vegetation. From an ecological perspective, this merging assumes that in the presence of a light source, vegetation is irrelevant to choice of walls by a gecko. Conversely, in the absence of light, vegetation becomes an important consideration in choice of walls by a gecko. For example, the overall significance could be that light sources always attract insect prey and thus render vegetation level irrelevant; whereas in the absence of light, vegetation might dictate the type and/or amount of insect prey and thus become a key component in microhabitat choice.

PAGE 62

52 I then fitted a logistic regression model using the variables material (M) and the newly formed light-vegetation (LVEG). I used the backward elimination method for model fitting. Results 4 Further variable condensation produced a wall combination table with no zeros or ones, with only two values smaller than five (Table 5-9). The backward elimination method generated five functional logistic regression models (Table 5-10). However, none of the models contained significant parameters. Consequently, these models were unusable. Data Analysis 5 A final analysis was performed on these data. I computed chi-square tests for material and gecko presence, light-vegetation and gecko presence, material and gecko presence while controlling for the light-vegetation variable, and light-vegetation and gecko presence while controlling for the material variable. I used the Fishers Exact test instead of the chi-square test for small samples. Lastly, I calculated relevant percentages for these data. Significance was set at the 5% level. Results 5 Using the Fishers exact test, light-vegetation was found to be independent of gecko presence at the 5% significance level when controlling for material (Table 5-11). Significance values for light-vegetation and gecko presence were 0.4098 for aluminum, 1.0000 for brick, 0.2416 for cement, and 0.6431 for wood. The cell chi-square values revealed that fewer cement walls of high light and any vegetation type had geckos than expected. Although this result is not significant, it points to a potential trend.

PAGE 63

53 When light-vegetation was controlled, the Fishers exact test concluded that material and gecko presence were independent at the 5% significance level (Table 5-12). Specifically, significance values were 0.8594 for high light/cement-vegetation, 0.6111 for low light/cement, and 0.2642 for low light/vegetation. A possible trend was also uncovered via the cell chi-square values; there was a greater number of aluminum, low light/vegetation walls that contained geckos than would be expected by chance. Two-way chi-square tests concluded that both light-vegetation (Table 5-13) and material (5-14) were independent of gecko presence at the 5% significance level. The significance value for light vegetation was 0.1402, whereas that of material was 0.2281. General, yet non-significant, trends included fewer high light/cement-vegetation walls inhabited by geckos, and a greater number of gecko-populated aluminum walls than predicted by randomness. These analyses revealed that 34% (13/38) of high light/cement-vegetation walls, 48% (40/83) of low light/vegetation walls and 56% (22/39) of low light/cement walls contained geckos (Table 5-15). With respect to aluminum walls, 42% (5/12) of high light/cement-vegetation walls, 63% (17/27) of low light/vegetation walls and 71% (5/7) of low light/cement walls recorded gecko presence. Following a similar pattern of gecko occurrence, cement and wood walls achieved the low values of 22% (2/9) and 25% (1/4) respectively when described as high light/cement-vegetation, followed by 44% (12/27) and 33% (5/15) when characterized as low light/vegetation, and peaked at 57% (13/23) and 67% (2/3) when categorized as low light/cement. When considering brick walls, in turn, 33% (2/6) of high light/cement-vegetation walls, 39% (5/13) of low light/cement walls and 43% (6/14) of low light/vegetation walls recorded geckos.

PAGE 64

54 With respect to material, 59% (27/46) of aluminum walls, 46% (27/59) of cement walls, 39% (13/33) of brick walls, and 36% (8/22) of wood walls had geckos (Table5-16). When controlling for high light/cement-vegetation, geckos occurred on 42% (5/12) of aluminum walls, 39% (5/13) of brick walls, 25% (1/4) of wood walls, and 22% (2/9) of cement walls. Walls described as low light/cement contained geckos 71% (5/7) of the time if they were made of aluminum, 67% (2/3) of the time if they were constructed of wood, 57% (13/23) of the time if they were cement, and 33% (2/6) of the time if they were build out of brick. Regarding low light/vegetation walls, 63% (17/27) of aluminum walls, 44% (12/27) of cement walls, 43% (6/14) of brick walls, and 33% (5/15) of wood walls were populated by geckos. Discussion Sample size proved to be a defining component in the outcome of my analyses. The realization that the number of sampled walls was too small to accommodate a logistic regression was unexpected, as my sample size (160 walls) greatly exceeded the general rule of 10 observations per variable (Agresti, 1996). Indeed, my sample size of 160 was twice that required for my original eight variables, and four times that needed for the four variables eventually used for the logistic regression model. Further investigation, however, revealed that overall sample size was not responsible for the model inadequacy. Instead, it was the mostly small, sometimes zero, sample size of individual wall combinations that contributed to the collapse of the model. Thus, even if the sample size for this study was substantially larger, a logistic regression model would continue to fail if there were wall combinations with a small number of samples. This could be a common dilemma in urban ecology studies, where investigators must work within rigid landscapes that offer few opportunities for manipulation. Thus,

PAGE 65

55 the possibility that study sites might not contain specific combinations of variables is often true, and highly unpredictable due to the human dimension involved. These statistical considerations should be an integral step in the planning of any urban ecology study. Once the data were sufficiently collapsed to compensate for the small sample size of certain wall combinations, both the logistic regression model and chi-square tests showed no significance at the 5% level. This lack of significance suggests a variety of scenarios. First, the microhabitat preference of H. turcicus with respect to material, light, and vegetation might be subtle and thus require a larger sample size to be exposed. Second, perhaps the microhabitat variables that were selected for this study are irrelevant to the microhabitat choice of H. turcicus, and the statistical conclusions merely report this. Theoretically, habitat selection in reptiles is believed to be most effective when controlled by reliable environmental cues that are independent of daily and/or seasonal fluctuations, and are evident in all situations (Heatwole, 1982). Although this was the case for material, light and vegetation were less consistent due to management regimes, and thus might not be used as a stimulus because they fail to accurately represent a given habitat. Other factors such as, but not limited to, microclimate quality, behavioral aspects and structural attributes may also play an important role in microhabitat selection and need to be considered in future studies (Heatwole, 1982). Lastly, it is possible that these results reflect the robust character of this proficient colonizing gecko species. Perhaps the extensive non-native range of H. turcicus stems

PAGE 66

56 from this geckos habitat flexibility, which allows it to thrive on any type of building/wall environment. Although no significant results were obtained in this study, some general trends were uncovered, and may pave the way for future research. In view of material, more aluminum walls than expected contained geckos, particularly those possessing low light and vegetation. Aluminum walls were also described as having a smooth texture, light color, small length and being modern in age. With respect to light-vegetation, fewer high light/cement-vegetation walls than expected possessed geckos, especially those constructed of cement. Further work is required to officially establish these trends, and to tease out the mechanisms behind them. As with all scientific studies, these data and its conclusions have some limitations. For instance, it is impossible to determine if the absence of H. turcicus on a wall is due to preference or if it is an artifact of this species dispersal ability and/or of extraneous circumstances. Also, shortcomings in sampling technique, such as not observing a wall throughout the entire night and frightening geckos as I approached a wall, could have resulted in an underestimation of walls containing geckos. Additional work, both in the field and in the lab, is needed to shed light on the microhabitat preferences of H. turcicus in urban environments. Key to future studies is the establishment of strict protocols that all investigators can follow anywhere in the world. This, in turn, would allow for meaningful comparisons between different sites, and when all studies are pooled, for meta-analyses on general patterns. Moreover, future work should explore and develop sampling and statistical methods that will enhance ecological studies in urban environments.

PAGE 67

57 Table 5-1. Description of wall characterization variables for microhabitat study Variables Levels Criteria Early *Built between 1900 to 1969 Modern *Built between 1970 to 1989 Age Contemporary *Built between 1990 to 2002 North **Location of wall on official maps South **Location of wall on official maps West **Location of wall on official maps Cardinal Orientation East **Location of wall on official maps Light *** The inability to perceive a 3x 5 white index card at a perpendicular distance of 10ft (3.048m) from the wall; during the day Color Dark ***The ability to perceive a 3x 5 white index card at a perpendicular distance of 10ft (3.048m) from the wall; during the day Small Measurement < 20m Medium 20m < measurement < 40m Length Large Measurement > 40m High Presence of at least one light source on the wall Light Low No light source present on the wall Aluminum Physical observation; > 50% of wall surface Brick Physical observation; > 50% of wall surface Cement Physical observation; > 50% of wall surface Material Wood Physical observation; > 50% of wall surface Smooth Ability to draw a straight line ~10cm in length on white printer paper propped on the wall Texture Rough Inability to draw a straight line ~10cm in length on white printer paper propped on the wall Cement > 60% of wall length bordered by cement Mix > 40% to <60% of wall length bordered by cement or vegetation Vegetation Vegetation > 60% of wall length bordered by any type or height of vegetation *Sources used: 2000 Building Information List for the University of Florida prepared by the UF Physical Plant Division,and unofficial list prepared by the VA Hospital Engineering Department **Sources used: 2000 Building Information List for the University of Florida prepared by the UF Physical Plant Division,and official VA Hospital Engineering maps *** Index cards were provided by AMPAQ, Dallas, TX, 75252

PAGE 68

Table 5-2. Chi-square and Fisher Exact p-values for wall characterization variables 58 Variables Age Cardinal Orientation Color Length Light Material Texture Vegetation Age -----0.6695 0.0297 *0.0845 0.7228 <0.0001 0.9564 0.1976 Cardinal Orientation ------------0.6043 *0.8805 0.1941 0.8859 0.7580 0.3484 Color -----------------*0.0481 *0.0922 <0.0001 *<0.0001 0.3426 Length ------------------------0.0005 *0.0041 0.0573 *0.2682 Light ------------------------------0.0622 *0.0846 0.0464 Material -----------------------------------<0.0001 0.0691 Texture ------------------------------------------0.7696 Vegetation ------------------------------------------------* Indicates the use of the Fishers Exact test Numbers in bold indicate significance

PAGE 69

59 Table 5-3. The dependency of age, color, length, and texture on material Material Age Color Length Texture Aluminum Modern Light Small Smooth Brick Early Dark Small/Modern Smooth Cement Early Light Small Rough Wood Modern/ Contemporary Dark Small Rough

PAGE 70

Table 5-4. Logistic regression models resulting from the backward elimination method using the variables: material; cardinal orientation, vegetation, light 60 Model Predictors Deviance DF Models Compared Difference P-Value 1 M*CO*V*L 141.5044 95 2 M*CO*V+ M*CO*L+ M*V*L+ CO*V*L 141.5044 95 2-1 0 >0.999 3 M*CO*V+M*CO*L+ M*V*L 141.5044 98 3-1 0 >0.999 4 M*CO*V+M*CO*L+ CO*V*L 141.5044 96 4-1 0 >0.999 5 M*CO*V+M*V*L+ CO*V*L 152.4738 100 5-1 10.9694 ~ 0.100 6 M*CO*L+M*V*L+ CO*V*L 152.9663 103 6-1 11.4619 > 0.250 7 M*CO*V+M*CO*L 141.5044 101 7-1 >0.999 8 M*CO*V+M*V*L 152.4738 104 8-1 10.9694 >0.100 9 M*CO*L+M*V*L 155.9364 108 9-1 14.4320 >0.100 10 M*CO*V+CO*V*L 156.7944 103 10-1 15.2900 <0.050 11 M*V*L+CO*V*L 171.1805 110 11-1 29.6761 <0.010 12 M*CO*L+CO*V*L 162.5828 109 12-1 21.0784 ~0.100 0

PAGE 71

Table 5-4. Continued. 61 Model Predictors Deviance DF Models Compared Difference P-Value 13 M*CO*V 158.7768 108 13-1 17.2724 ~0.100 14 M*CO*L 164.7604 114 14-1 23.2560 >0.100 15 M*V*L 176.4308 115 15-1 34.9264 <0.050 16 CO*V*L 178.6783 116 16-1 37.1739 <0.050 17 M*CO+M*V+M*L+ CO*V+CO*L+V*L 185.0743 121 17-1 43.5699 <0.050 Numbers in bold indicates significance Model 7, underlined and in bold, is the chosen model

PAGE 72

Table 5-5. Summary statistics of the significant predictor variables of the M*CO*V+M*CO*L model Predictor Estimate ASE P-value Odds Ratio 95%C.I.of Odds Ratio Aluminum 68.3825 1.4434 < 0.0001 2.0038 x 10-30 (1.18 x 10-31, 3.39 x 10-30 ) Brick 91.7478 1.6833 < 0.0001 1.4270 x 10-40 (5.27 x 10-42, 3.87 x 10-39 ) Cement 46.1655 1.3663 < 0.0001 8.9244 x 10-21 (6.13 x 10-22, 1.30 x 10-19 ) Wood 0.0000 0.0000 ---------------Mix 18.3578 1.5916 < 0.0001 9.3906 x 107 (4.15 x 106, 2.13 x 109 ) Vegetation 0.0000 0.0000 ---------------East 109.5860 1.0954 < 0.0001 3.9138 x 1047 (4.57 x 1046, 3.35 x 1048 ) North 45.0170 1.5275 < 0.0001 2.8143 x 10-20 (1.41 x 10-21, 5.62 x 10-19 ) West 0.0000 0.0000 ---------------No Light 67.2839 1.6583 < 0.0001 6.0114 x 10-30 (2.33 x 10-31, 1.55 x 10-28 ) Light 0.0000 0.0000 ---------------Aluminum / East 131.7139 1.5652 < 0.0001 1.5945 x 1057 (7.42 x 1055, 3.43 x 1058 ) Aluminum / North 44.3238 2.4152 < 0.0001 1.7766 x 1019 (1.56 x 1017, 2.02 x 1021 ) Aluminum / West 0.0000 0.0000 ---------------Brick / South 17.0373 2.0897 < 0.0001 3.9884 x 10-8 (6.64 x 10-10, 2.40 x 10-6 ) Brick / West 0.0000 0.0000 ---------------North / Mix Veg 18.3581 2.0083 < 0.0001 1.0646 x 10-8 (2.08 x 10-10, 5.45 x 10-7 ) North / Veg 0.0000 0.0000 ---------------62

PAGE 73

Table 5-6. Possible wall combinations involving the variables: material, cardinal orientation, vegetation, and light 63 Light Level = High Light Level = Low Vegetation Level Vegetation Level Cardinal Orientation Material Cement Mix Vegetation Cement Mix Vegetation Aluminum 0 0 3 1 3 5 Brick 2 2 2 0 0 0 Cement 1 1 1 2 5 6 North Wood 1 1 1 0 0 4 Aluminum 1 1 3 1 0 7 Brick 2 1 1 1 0 5 Cement 0 1 1 3 5 7 South Wood 0 0 0 1 1 7 Aluminum 0 0 2 0 1 6 Brick 0 0 1 2 2 3 Cement 1 0 1 1 2 8 West Wood 0 0 0 0 0 4 Aluminum 0 1 1 1 1 8 Brick 2 0 0 1 1 3 Cement 1 0 1 3 2 6 East Wood 1 0 0 1 0 0

PAGE 74

64 Table 5-7. Number of observations per wall combination involving the variables: material, vegetation (2 levels), and light Light Level = High Light Level = Low Vegetation Level Vegetation Level Material Cement Vegetation Cement Vegetation Aluminum 3 9 8 26 Brick 9 4 7 13 Cement 5 4 23 27 Wood 3 1 3 15

PAGE 75

Table 5-8. Logistic regression models resulting from the backward elimination method using the variables: material, vegetation (2 levels), and light 65 Model Predictors Deviance DF Models Compared Difference P-Value 1 M*V*L 207.9758 144 2 M*L+M*V+V*L 208.5680 147 2-1 0.5922 > 0.250 3 M*L+M*V 208.8361 148 3-1 0.8603 > 0.250 4 M*L+V*L 211.0082 150 4 -1 3.0324 > 0.250 5 M*V+V*L 210.8852 150 5-1 2.9094 > 0.250 6 M*L 211.6953 151 6-1 3.7195 >0.250 7 M*V 211.1258 151 7-1 3.15 >0.250 8 V*L 212.4750 153 8-1 4.4992 >0.250 9 M+V+L 212.8583 154 9-1 4.8825 > 0.250 10 M+L 213.4205 155 10-1 5.4447 > 0.250 11 M+V 216.6519 155 11-1 8.6761 > 0.250 12 V+L 217.7235 157 12-1 9.7477 >0.250

PAGE 76

Table 5-8. Continued. 66 Model Predictors Deviance DF Models Compared Difference P-Value 13 M 216.8310 156 13-1 8.8552 >0.250 14 V 221.1641 158 14-1 13.1883 >0.250 15 L 217.9192 158 15-1 9.9434 >0.250 16 NONE 221.1817 159 16-1 13.2059 > 0.250 Number in bold indicates model that contain significant parameters

PAGE 77

67 Table 5-9. Number of observations per wall combination involving the variables: material, and light-vegetation (3 levels) Light-Vegetation Levels Material High Light / Cement-Vegetation Low Light/Cement Low Light/Vegetation Aluminum 12 8 26 Brick 13 7 13 Cement 9 23 27 Wood 4 3 15 Table 5-10. Logistic regression models resulting from the backward elimination method using the variables: material, and light-vegetation (3 levels) Model Predictors Deviance DF Models Compared Difference P-Value 1 M*LVEG 209.8897 148 2 M+LVEG 212.3961 154 2-1 2.5064 > 0.250 3 M 216.8310 156 3-1 6.9413 > 0.250 4 LVEG 217.2007 157 4 -1 7.311 > 0.250 5 NONE 221.1817 159 5-1 11.292 > 0.250 Note that there is no model with significant parameters

PAGE 78

Table 5-11. Chi-square test and Fishers Exact test for light-vegetation and gecko presence, controlling for material 68 Material Aluminum Brick Cement Wood Gecko Presence Yes No Yes No Yes No Yes No Frequency 5 7 5 8 2 7 1 3 Expected Frequency 7.0435 4.9565 5.1212 7.8788 4.1186 4.8814 1.4545 2.5455 High Light / Cement Vegetation Cell Chi-Square 0.5929 0.8425 0.0029 0.0019 1.0898 0.9196 0.1420 0.0812 Frequency 5 2 2 4 13 10 2 1 Expected Frequency 4.1087 2.8913 2.3636 3.6364 10.525 12.475 1.0909 1.9091 Low Light / Cement Cell Chi-Square 0.1934 0.2748 0.0559 0.0364 0.5818 0.4909 0.7576 0.4329 Frequency 17 10 6 8 12 15 5 10 Expected Frequency 15.8480 11.1520 5.5152 8.4848 12.3560 14.644 5.4545 9.5455 Light Vegetation Low Light / Vegetation Cell Chi-Square 0.0838 0.1190 0.0426 0.0277 0.0103 0.0087 0.0379 0.0216 Total Chi-Square 2.1063 0.1674 3.1010 1.4732 P-Value for Chi-Square/ Fishers Exact* 0.3488 / 0.4098* 0.9197 / 1.0000* 0.2121 / 0.2416* 0.4787 / 0.6431*

PAGE 79

Table 5-12. Chi-square test and Fishers Exact test for material and gecko presence, controlling for light-vegetation Light-Vegetation High Light / Cement Vegetation Low Light / Cement Low Light / Vegetation Gecko Presence Yes No Yes No Yes No Frequency 5 7 5 2 17 10 Expected Frequency 4.1053 7.8947 3.9487 3.0513 13.012 13.9880 Aluminum Cell Chi-Square 0.195 0.1014 0.2799 0.3622 1.2222 1.1370 Frequency 5 8 2 4 6 8 Expected Frequency 4.4474 8.5526 3.3846 2.6154 6.7470 7.2530 Brick Cell Chi-Square 0.0687 0.0357 0.5664 0.7330 0.0827 0.0769 Frequency 2 7 13 10 12 15 Expected Frequency 3.0789 5.9211 12.9740 10.0260 13.0120 13.9880 Cement Cell Chi-Square 0.3781 0.1966 5.07x 10-5 0.0001 0.0787 0.0732 Frequency 1 3 2 1 5 10 Expected Frequency 1.3684 2.6316 1.6923 1.3077 7.2289 7.7711 Material Wood Cell Chi-Square 0.0992 0.0516 0.0559 0.0724 0.6872 0.6393 Total Chi-Square 1.1263 2.0700 3.9973 P-Value for Chi-Square/ Fishers Exact* 0.7707 / 0.8594* 0.5580 / 0.6111* 0.2618 / 0.2642* 69

PAGE 80

70 Table 5-13. Chi-square test for light-vegetation and gecko presence Gecko Presence Yes No Frequency 13 25 Expected Frequency 17.8130 20.1880 High Light / Cement Vegetation Cell Chi-Square 1.3002 1.1473 Frequency 22 17 Expected Frequency 18.2810 20.7190 Low Light / Cement Cell Chi-Square 0.7565 0.6675 Frequency 40 43 Expected Frequency 38.9060 44.0940 Light Vegetation Low Light / Vegetation Cell Chi-Square 0.0307 0.0271 Total Chi-Square 3.9293 P-Value for Chi-Square 0.1402

PAGE 81

71 Table 5-14. Chi-square test for material and gecko presence Gecko Presence Yes No Frequency 27 19 Expected Frequency 21.5630 24.4380 Aluminum Cell Chi-Square 1.3712 1.2099 Frequency 13 20 Expected Frequency 15.4690 17.5310 Brick Cell Chi-Square 0.3940 0.3476 Frequency 27 32 Expected Frequency 27.6560 31.3440 Cement Cell Chi-Square 0.0156 0.0137 Frequency 8 14 Expected Frequency 10.3130 11.6880 Material Wood Cell Chi-Square 0.5186 0.4576 Total Chi-Square 4.3282 P-Value for Chi-Square 0.2281

PAGE 82

72 Table 5-15. Three-way contingency table of light-vegetation, controlling for material, with associated percentages and marginal associations Gecko Presence Material Light-Vegetation Yes No Percent Gecko Present High Light / Cement -Vegetation 5 7 42% Low Light / Cement 5 2 71% Aluminum Low Light / Vegetation 17 10 63% High Light / Cement -Vegetation 5 8 39% Low Light / Cement 2 4 33% Brick Low Light / Vegetation 6 8 43% High Light / Cement -Vegetation 2 7 22% Low Light / Cement 13 10 57% Cement Low Light / Vegetation 12 15 44% High Light / Cement -Vegetation 1 3 25% Low Light / Cement 2 1 67% Wood Low Light / Vegetation 5 10 33% High Light / Cement -Vegetation 13 25 34% Low Light / Cement 22 17 56% All Materials Low Light / Vegetation 40 43 48%

PAGE 83

73 Table 5-16. Three-way contingency table of material, controlling for light-vegetation, with associated percentages and marginal associations Gecko Presence Light-Vegetation Material Yes No Percent Gecko Present Aluminum 5 7 42% Brick 5 8 39% Cement 2 7 22% High Light / Cement Vegetation Wood 1 3 25% Aluminum 5 2 71% Brick 2 4 33% Cement 13 10 57% Low Light /Cement Wood 2 1 67% Aluminum 17 10 63% Brick 6 8 43% Cement 12 15 44% Low Light / Vegetation Wood 5 10 33% Aluminum 27 19 59% Brick 13 20 39% Cement 27 32 46% All Light /Vegetation Levels Wood 8 14 36%

PAGE 84

CHAPTER 6 MANAGEMENT AND CONSERVATION IN AN URBAN ENVIRONMENT Wildlife management and conservation in urban areas have a number of key goals. These include traditional objectives such as the promotion and/or maintenance of species composition, and the control of species abundance by either directly increasing or decreasing numbers (Nilon and Pais, 1997). More importantly, urban wildlife management and conservation programs also provide the public with opportunities to interact with wildlife, and disseminate information to the general public and appropriate professionals (Anderson, 2002). The latter two goals are essential to formulate ecologically advantageous policy, as the urban public's voting strength in legislatures is on the rise (Bolen and Robinson, 2003). The management of urban wildlife follows a holistic style, which differs dramatically from the conventional agriculture and hunting orientated approach (Bolen and Robinson, 2003). This difference is directly related to the non-consumptive, recreational attitude that many urban residents maintain toward wildlife. A study undertaken in metropolitan areas of New York State demonstrated this idea, as urbanites were found to prefer butterflies and songbirds to species sought for hunting such as waterfowl and pheasants (Brown et al., 1979). With the exception of nuisance animals and/or pests (cockroaches, rats, pigeons, sometimes raccoons, etc.) the urban public generally supports the notion of having wildlife in their surrounding environment (Bolen and Robinson, 2003). 75

PAGE 85

76 Despite the fact that the desire for wildlife is present in cities, it is a challenge to implement management and/or conservation plans in the myriad of interests and human beliefs usually present in urban areas (Lyons, 1997). Although public opinion is an integral part of most management and conservation initiatives, it is especially the case in urban environments (Anderson, 2002). 7 The intricacies that govern public perception are many, subtle, and highly susceptible to change. It has been shown that public perception varies enormously within a city, particularly between communities of different income, education, and race (Nilon and Pais, 1997). Knox (1991) alluded to this concept by illustrating that land use was directly related to social, economic, and demographic factors. This variation in public perception, in turn, tends to lead to equally variable public preferences (Schauman et al., 1986). Whitney and Adams (1980) showed this concept when they found that the types of plants in gardens were linked to fashion, taste, species availability, property value, and age of the house in question. Developing urban wildlife management and conservation strategies involves a multitude of participants and considerations, making each case unique (Lyons, 1997). The heterogeneity of public opinion, and the limitations imposed by both the intensity of urbanization and the land-use history of a given area require a great deal of creativity and cross-disciplinary thinking from an urban wildlife manger (Bolen and Robinson, 2003; Loeb, 1998; Lyons, 1997). Additional complexity results from the fact that few urban spaces are dedicated solely to wildlife, which requires most management plans to consider and, sometimes favor, other land uses (Bolen and Robinson, 2003). This multiple-use management approach requires urban wildlife managers to interact with a

PAGE 86

77 wide range of professionals (Lyons, 1997). A particularly important professional group is the urban landscape planners, as they execute the majority of urban wildlife initiatives (Nilon and Pais, 1997). Lastly, public approval and participation is a necessity: the public is directly affected by any management scheme, as the latter invariably becomes part of their everyday life. Thus, the public is a preeminent force behind any change (Lyons, 1997). In general, urban wildlife managers assume the role of solution facilitators rather than active problem solvers, as they provide insight, tools, and ideas to a number of different interest groups in an attempt to coalesce their sensibilities into common, feasible goals (Lyons, 1997). For example, if the creation of a neighborhood park is being considered, an urban wildlife manager might suggest programs that unite features attractive to wildlife with other objectives such as safety, specific recreational purposes, following city ordinance, remaining within a certain budget, and others. An urban wildlife manager will not, however, physically carry out the selected program. Instead, the logistics are left to specific groups such as the police force for safety, landscape planners for recreational structures, and city officials for monitoring compliance with city regulations. An equally important function of urban wildlife managers is as educators to the public via the media, pamphlets, and/or seminars (Anderson, 2002 In the case of H. turcicus in Gainesville, management strategies would be specific to the locality due to the geckos nonnative status. Although H. turcicus is an introduced species, it is has managed to occupy a vacant niche in Gainesville, and thus is not believed to possess a threat to any native species. Consequently, the eradication of this gecko, which would probably be costly due to its prolific colonization abilities and its

PAGE 87

78 potentially generalist microhabitat habits, would not be a necessity. In fact, the presence of H. turcicus could be beneficial; being easily observable on walls, this gecko could be wildlife that many in the public could interact with on a daily basis, hence familiarizing a number of citizens (and voters) with reptiles, a group often seen in a negative light. An integral part of the appreciation of H. turcicus would be the use of a number of formats to educate both adults and children on the interesting facts of not just this species, but of geckos and reptiles in general. Benefits of this species, such as their consumption of insects, would especially have to be emphasized since they have been labeled as pests by some homeowners in Gainesville, as a result of the mess they sometimes leave when they nest (shredded paper, etc.) (Franz, personal communication). Ultimately, I believe that a primary goal of urban ecologists, regardless of whether they hold a management position, should be to convey the intrinsic value of nature to all sectors of the urban population.

PAGE 88

APPENDIX A PRELIMINARY SURVEY DATA Table A-1. Temperature readings of walls with respect to material and cardinal orientation Cardinal Orientation Material North South West East Cement 25.00 25.43 24.75 24.98 Cement 25.28 24.68 24.48 24.75 Cement 24.48 25.03 24.78 24.58 Cement 25.53 26.30 25.88 25.38 Cement 25.33 24.83 25.03 25.10 Cement 26.73 25.93 27.25 27.18 Cement 26.40 26.53 27.15 27.03 Cement 25.45 25.38 25.33 24.78 Cement 25.58 25.53 25.30 26.10 Cement 25.68 24.98 25.55 25.35 Cement 28.60 26.98 29.10 29.08 Cement 26.50 26.58 29.35 27.98 Cement 29.55 27.50 28.88 28.40 Aluminum 23.10 23.08 Not Recorded 23.58 Aluminum 23.75 23.98 24.15 23.93 Aluminum 24.18 24.03 23.93 23.80 Aluminum 24.63 24.05 23.98 24.13 Aluminum 23.40 22.48 23.05 23.33 Aluminum 22.90 22.48 22.90 22.60 Aluminum 22.68 23.08 22.53 22.65 Aluminum Not Recorded 23.28 22.90 22.55 Aluminum 21.75 22.23 23.38 21.85 Aluminum 23.88 24.45 24.75 24.65 Aluminum 25.03 23.15 24.00 24.08 Aluminum 24.10 23.85 24.33 24.25 Aluminum 22.85 23.73 23.28 23.40 Aluminum 23.50 24.75 24.10 24.13 Aluminum 24.33 24.80 24.95 24.93 Aluminum 26.23 23.70 26.10 26.73 Aluminum 24.98 23.88 23.83 23.98 Aluminum 25.88 24.08 24.38 24.25 Aluminum 22.63 23.10 22.88 22.78 Aluminum 25.33 24.87 25.07 25.17 79

PAGE 89

80 Table A-1. Continued. Cardinal Orientation Material North South West East Brick 25.13 24.60 24.75 25.00 Brick 25.03 26.48 27.08 26.53 Brick 26.85 25.53 26.55 26.63 Brick 31.63 29.05 29.60 29.45 Wood 24.15 24.85 24.55 24.33 Wood 25.20 25.45 Not Recorded Not Recorded Wood 25.38 24.13 23.83 24.48 Wood 25.03 25.13 25.30 26.15 Wood 23.05 22.68 23.45 23.80 Wood 26.60 24.75 25.83 25.45 Wood 25.58 25.53 24.80 24.75 Wood 25.75 25.68 25.53 25.50 Wood 24.15 23.33 23.95 24.10 Wood 29.20 27.35 28.10 27.95

PAGE 90

APPENDIX B NATURAL HISTORY DATA Table B-1. Temperature readings (C) for individual adult H. turcicus recorded during the fall/winter 2001 survey 12.7 19.7 21.1 21.9 22.4 23.1 23.7 24.6 25.8 26.7 14.0 19.7 21.1 21.9 22.5 23.1 23.7 24.6 25.8 26.7 14.9 19.8 21.1 21.9 22.5 23.2 23.8 24.6 25.8 26.8 15.6 19.9 21.2 21.9 22.6 23.2 23.8 24.7 25.9 26.8 15.7 19.9 21.2 21.9 22.6 23.3 23.8 24.7 25.9 26.9 16.1 19.9 21.2 21.9 22.6 23.3 23.8 24.7 25.9 26.9 16.3 19.9 21.3 21.9 22.6 23.3 23.9 24.8 25.9 26.9 16.3 19.9 21.3 21.9 22.6 23.3 23.9 24.8 26.1 26.9 16.6 20.2 21.3 21.9 22.6 23.3 23.9 24.8 26.1 26.9 17.3 20.2 21.4 21.9 22.6 23.3 23.9 24.8 26.1 27.1 17.4 20.2 21.4 21.9 22.7 23.3 23.9 24.9 26.2 27.3 17.6 20.2 21.4 21.9 22.7 23.3 24.0 24.9 26.2 27.4 18.2 20.3 21.4 21.9 22.7 23.3 24.1 24.9 26.2 27.4 18.3 20.3 21.4 22.0 22.7 23.3 24.1 24.9 26.2 27.4 18.3 20.4 21.4 22.0 22.7 23.4 24.1 24.9 26.2 27.6 18.3 20.5 21.4 22.1 22.7 23.4 24.1 25.0 26.3 27.6 18.4 20.5 21.4 22.1 22.7 23.4 24.1 25.1 26.3 27.8 18.7 20.5 21.4 22.1 22.8 23.4 24.1 25.1 26.3 27.9 18.7 20.6 21.4 22.1 22.8 23.4 24.1 25.2 26.4 28.1 18.7 20.7 21.5 22.1 22.8 23.4 24.1 25.2 26.4 28.2 18.8 20.7 21.5 22.1 22.8 23.4 24.1 25.2 26.4 28.3 18.8 20.7 21.5 22.2 22.9 23.5 24.1 25.2 26.4 28.3 18.8 20.8 21.5 22.2 22.9 23.5 24.3 25.3 26.4 28.4 18.9 20.8 21.5 22.2 22.9 23.5 24.3 25.3 26.5 28.6 18.9 20.8 21.6 22.2 22.9 23.5 24.4 25.3 26.5 28.7 19.0 20.8 21.7 22.2 22.9 23.5 24.4 25.3 26.5 28.7 19.1 20.9 21.7 22.2 22.9 23.6 24.4 25.4 26.5 30.2 19.2 20.9 21.7 22.2 22.9 23.6 24.4 25.5 26.6 30.5 19.2 20.9 21.7 22.3 23.0 23.6 24.4 25.5 26.6 30.6 19.2 20.9 21.7 22.3 23.0 23.6 24.5 25.5 26.6 30.9 19.3 21.0 21.8 22.3 23.1 23.7 24.6 25.6 26.7 31.2 19.5 21.0 21.8 22.3 23.1 23.7 24.6 25.7 26.7 31.2 19.6 21.0 21.8 22.3 23.1 23.7 24.6 25.7 26.7 31.8 19.6 21.1 21.8 22.4 23.1 23.7 24.6 25.7 26.7 34.3 19.6 21.1 21.8 22.4 23.1 23.7 24.6 25.8 26.7 36.3 19.6 21.1 81

PAGE 91

82 Table B-2. Temperature readings (C) for individual sub-adult H. turcicus recorded during the fall/winter 2001 survey. 21.1 13.9 18.8 20.1 21.9 22.6 23.3 23.7 24.2 25.1 26.4 21.1 16.3 18.9 20.2 21.9 22.7 23.3 23.7 24.3 25.1 26.5 21.2 16.6 18.9 20.3 22.1 22.7 23.3 23.7 24.3 25.2 26.6 21.2 16.8 19.2 20.3 22.1 22.7 23.3 23.7 24.4 25.2 26.7 21.3 17.1 19.3 20.3 22.1 22.7 23.3 23.7 24.4 25.2 26.7 21.3 17.4 19.5 20.4 22.1 22.7 23.3 23.7 24.4 25.2 26.8 21.3 17.5 19.5 20.4 22.2 22.7 23.3 23.8 24.4 25.2 27.1 21.4 17.5 19.5 20.5 22.2 22.8 23.4 23.8 24.6 25.2 27.2 21.6 17.7 19.6 20.6 22.2 22.8 23.4 23.8 24.6 25.3 27.3 21.6 17.7 19.7 20.6 22.3 22.8 23.4 23.9 24.6 25.3 27.3 21.6 17.8 19.7 20.6 22.3 22.8 23.4 23.9 24.6 25.3 27.6 21.7 18.0 19.7 20.7 22.4 22.9 23.4 23.9 24.6 25.6 27.8 21.7 18.3 19.7 20.7 22.4 22.9 23.4 24.0 24.7 25.6 27.8 21.7 18.4 19.7 20.7 22.4 22.9 23.5 24.0 24.8 25.7 28.3 21.7 18.4 19.8 20.8 22.6 23.0 23.5 24.1 24.8 25.7 28.4 21.7 18.4 19.8 20.8 22.6 23.1 23.5 24.1 24.9 25.9 28.9 21.7 18.7 19.8 20.8 22.6 23.1 23.5 24.1 24.9 26.1 29.3 21.8 18.7 19.8 20.9 22.6 23.1 23.6 24.2 24.9 26.2 30.3 21.8 18.7 19.9 20.9 22.6 23.1 23.6 24.2 25.0 26.2 30.5 21.8 18.7 20.0 21.0 22.6 23.2 23.6 24.2 25.0 26.4 30.8 21.8 18.7 20.1 21.1

PAGE 92

83 Table B-3. Temperature readings (C) for individual adult H. turcicus recorded during the spring 2002 survey 16.4 22.6 24.1 25.1 26.2 27.3 16.6 22.7 24.1 25.1 26.2 27.3 16.8 22.7 24.2 25.2 26.3 27.4 17.0 22.7 24.2 25.2 26.3 27.4 17.2 22.8 24.2 25.2 26.3 27.6 17.7 22.8 24.2 25.2 26.4 27.6 17.7 22.8 24.2 25.2 26.4 27.6 18.6 22.8 24.3 25.2 26.5 27.7 18.6 22.9 24.3 25.2 26.6 27.9 18.7 22.9 24.3 25.3 26.6 27.9 18.9 22.9 24.3 25.3 26.6 28.1 19.0 23.0 24.3 25.5 26.6 28.1 19.2 23.0 24.3 25.5 26.7 28.3 19.3 23.0 24.3 25.5 26.7 28.3 19.3 23.2 24.5 25.5 26.7 28.4 19.4 23.2 24.5 25.6 26.7 28.6 19.4 23.3 24.6 25.6 26.7 28.7 19.9 23.3 24.6 25.6 26.7 28.8 19.9 23.3 24.6 25.6 26.7 28.8 20.7 23.4 24.7 25.6 26.7 28.8 20.7 23.4 24.7 25.6 26.8 28.9 20.8 23.4 24.7 25.7 26.8 29.2 20.9 23.4 24.7 25.7 26.8 29.4 21.2 23.5 24.7 25.7 26.8 29.4 21.2 23.5 24.8 25.7 26.8 29.7 21.3 23.6 24.8 25.7 26.8 29.7 21.4 23.6 24.8 25.7 26.8 29.8 21.7 23.7 24.8 25.8 26.9 29.9 21.8 23.7 24.8 25.8 26.9 30.1 21.9 23.8 24.9 25.8 26.9 30.1 22.1 23.8 24.9 25.8 26.9 30.2 22.1 23.8 24.9 25.8 27.1 30.2 22.1 23.9 24.9 25.9 27.1 30.3 22.1 23.9 25.0 25.9 27.1 30.4 22.3 24.0 25.0 25.9 27.1 30.6 22.3 24.1 25.1 26.1 27.1 30.8 22.3 24.1 25.1 26.1 27.2 30.8 22.4 24.1 25.1 26.2 27.2 30.9 22.4 24.1 25.1 26.2 27.2 31.5 22.5 24.1 25.1

PAGE 93

APPENDIX C MICROHABITAT PREFERENCE DATA

PAGE 94

85 Cardinal Presence Presence Material Orientation Vegetation Texture Color Age Length of Light of H.turcicus Wood North Cement Rough Light Contemporary Small Yes 1 Brick South Cement Smooth Dark Early Medium Yes 1 Aluminum East Vegetation Smooth Dark Contemporary Small No 0 Wood West Vegetation Rough Dark Contemporary Small No 1 Cement West Vegetation Rough Light Early Small No 1 Brick West Mix Smooth Dark Early Medium No 1 Brick North Mix Rough Light Early Medium Yes 0 Aluminum East Vegetation Rough Light Modern Medium No 0 Aluminum South Vegetation Rough Light Modern Small No 1 Wood North Vegetation Smooth Light Early Medium No 1 Wood South Vegetation Smooth Light Contemporary Small No 0 Cement North Mix Rough Light Early Large No 0 Brick West Cement Smooth Dark Modern Small No 0 Brick West Vegetation Smooth Dark Early Medium No 1 Brick North Vegetation Smooth Dark Early Medium Yes 1 Brick East Cement Smooth Dark Early Large Yes 1 Brick South Cement Smooth Dark Early Medium Yes 1 Brick East Vegetation Smooth Dark Modern Small No 1 Brick North Vegetation Smooth Dark Early Small No 0 Aluminum East Mix Rough Light Modern Small Yes 0 Brick South Vegetation Smooth Dark Early Medium No 1 Brick West Vegetation Smooth Dark Contemporary Small No 0 Cement South Mix Rough Light Early Small No 1 Cement North Mix Rough Light Early Small No 1 Brick South Vegetation Smooth Dark Contemporary Small No 0 Aluminum West Vegetation Rough Light Modern Small No 1

PAGE 95

86 Aluminum East Vegetation Rough Light Modern Medium No 1 Wood South Vegetation Rough Dark Modern Medium No 1 Aluminum East Vegetation Smooth Dark Modern Small Yes 1 Wood West Vegetation Rough Dark Modern Small No 0 Wood North Vegetation Rough Dark Contemporary Small No 0 Wood South Vegetation Rough Dark Modern Large No 1 Brick South Vegetation Smooth Dark Early Medium Yes 0 Cement West Vegetation Rough Light Early Small No 1 Cement South Vegetation Rough Light Early Small No 0 Cement South Cement Rough Light Early Small No 0 Cement South Cement Rough Light Early Small No 0 Cement East Cement Rough Light Early Small No 0 Aluminum South Mix Smooth Dark Early Medium Yes 1 Wood South Cement Smooth Light Early Small No 0 Aluminum South Cement Rough Light Modern Small No 0 Brick North Vegetation Smooth Dark Early Large Yes 1 Brick West Cement Smooth Dark Early Medium No 1 Aluminum West Vegetation Smooth Light Early Medium Yes 0 Cement West Mix Rough Light Early Small No 0 Cement South Mix Rough Light Early Small No 1 Wood West Vegetation Rough Light Early Large No 1 Cement North Cement Rough Light Modern Small Yes 0 Cement South Cement Rough Light Modern Medium No 1 Cement East Cement Rough Light Modern Small No 0 Cement East Vegetation Rough Light Modern Small No 0 Aluminum East Mix Smooth Light Modern Small No 0 Cement North Mix Rough Light Modern Small No 1 Cement West Cement Rough Light Modern Medium Yes 1

PAGE 96

87 Cement East Mix Rough Light Early Small No 0 Cement South Mix Rough Light Early Small No 0 Aluminum West Vegetation Rough Dark Modern Small Yes 1 Aluminum West Vegetation Smooth Dark Modern Small No 0 Cement North Vegetation Rough Light Modern Small No 0 Aluminum North Vegetation Smooth Light Modern Small No 0 Cement North Vegetation Rough Light Modern Small No 0 Cement West Vegetation Rough Light Early Small No 1 Cement North Mix Rough Light Modern Medium Yes 1 Cement East Vegetation Rough Light Modern Small No 0 Aluminum North Mix Smooth Light Modern Small No 0 Cement West Vegetation Rough Light Modern Small No 0 Brick East Vegetation Smooth Dark Modern Small No 1 Cement South Vegetation Smooth Light Early Small Yes 1 Aluminum South Vegetation Smooth Light Early Small No 0 Brick South Vegetation Smooth Dark Contemporary Medium No 0 Aluminum West Vegetation Smooth Light Modern Medium No 0 Aluminum East Vegetation Smooth Light Modern Small No 0 Wood South Vegetation Rough Dark Modern Small No 0 Aluminum South Vegetation Rough Light Contemporary Small No 0 Cement West Vegetation Rough Light Early Small No 0 Wood North Vegetation Rough Light Early Small Yes 1 Cement South Vegetation Rough Light Early Small No 1 Brick West Mix Smooth Dark Early Medium No 1 Wood East Cement Rough Light Early Small Yes 0 Aluminum North Vegetation Smooth Dark Early Small No 0 Aluminum East Vegetation Smooth Dark Modern Small No 0 Aluminum North Mix Smooth Light Modern Small No 0

PAGE 97

88 Cement North Vegetation Rough Light Early Small No 1 Aluminum North Mix Smooth Light Modern Small No 1 Cement North Vegetation Rough Light Early Small No 1 Cement South Vegetation Rough Light Modern Small No 1 Cement East Mix Rough Light Early Small No 0 Aluminum North Vegetation Rough Light Early Small No 0 Aluminum South Vegetation Smooth Light Modern Medium Yes 1 Brick North Cement Smooth Dark Contemporary Small Yes 0 South Vegetation Sm Modern Sm Aluminum W ooth Dark all No 1 ood North Vegetation Rough Dark ContemporarySmall No 1 Wood East Cement Rough Dark ContemporarySmall No 1 Wood South Vegetation Rough Dark ContemporarySmall No 1 Cement South Vegetation Rough Light ContemporaryMedium No 0 Cement North Mix Rough Light ContemporarySmall No 0 Aluminum North Vegetation Rough Dark ContemporarySmall Yes 1 Aluminum South Vegetation Smooth Light Modern Small No 0 Brick North Cement Smooth Dark ContemporaryMedium Yes 1 Brick East Cement Smooth Dark ContemporarySmall No 0 Cement West Vegetation Rough Light Early Large No 0 Aluminum North Vegetation Rough Light Modern Small No 0 Cement North Vegetation Rough Light Early Small Yes 0 Cement South Mix Rough Light Early Small No 1 Cement North Cement Rough Light Early Small No 0 Cement East Vegetation Rough Light Early Small No 0 Cement North Cement Rough Light Early Small No 0 Cement North Mix Rough Light Early Small No 0 Brick East Vegetation Smooth Dark Early Large No 1 Cement East Vegetation Smooth Light Early Medium No 1

PAGE 98

Aluminum South Vegetation Rough Light Modern Small Yes 0 Aluminum West Vegetation Rough Light ContemporarySmall No 1 Aluminum North Vegetation Smooth Light ContemporarySmall Yes 0 89Brick South Mix Smooth Dark Modern Medium Yes 1 Brick North Mix Smooth Dark Early Medium Yes 1 Wood South Vegetation Rough Dark Modern Small No 1 Aluminum South Vegetation Smooth Dark Early Small No 1 Cement North Vegetation Rough Light Modern Small No 0 Cement East Vegetation Rough Light Modern Medium No 1 Aluminum West Vegetation Smooth Light N/A Small No 0 Aluminum South Cement Smooth Light Modern Small Yes 1 Aluminum West Mix Rough Light Modern Small No 0 Aluminum South Vegetation Rough Light ContemporarySmall No 1 Cement South Vegetation Rough Light Early Small No 0 Cement South Mix Rough Light Modern Medium No 1 Wood South Mix Rough Dark Modern Medium No 0 Aluminum East Vegetation Rough Light Modern Small No 1 Aluminum North Cement Rough Light ContemporarySmall No 0 Aluminum West Vegetation Rough Light Modern Small No 0 Aluminum East Vegetation Smooth Dark Modern Small No 1 Brick East Mix Smooth Dark Early Small No 1 Aluminum East Vegetation Smooth Light Modern Small No 1 Cement East Cement Rough Light Early Medium No 1 Wood West Vegetation Rough Dark ContemporarySmall No 1 Wood North Vegetation Rough Dark Modern Small No 0 Brick East Cement Rough Dark Modern Small Yes 0 Wood South Vegetation Rough Dark Modern Small No 1 Wood North Mix Rough Dark Modern Medium Yes 1

PAGE 99

90Brick South Cement Rough Dark ContemporarySmall No 1 Aluminum East Cement Smooth Dark Modern Small No 1 Brick West Vegetation Rough Dark Early Small No 1 Aluminum North Vegetation Smooth Light Modern Small No 0 Cement North Vegetation Rough Dark Early Medium No 1 Cement West Vegetation Rough Dark Modern Small No 1 Cement South Mix Rough Dark Early Large Yes 1 Cement West Mix Rough Light ContemporaryMedium No 1 Aluminum South Vegetation Rough Light Modern Medium Yes 1 Cement East Vegetation Rough Dark Early Small Yes 1 Cement East Vegetation Rough Dark Early Small No 1 Cement East Cement Rough Dark Early Medium Yes 1 Cement South Vegetation Rough Light Early Small No 1 Cement West Cement Rough Light Modern Small No 1 Cement West Vegetation Rough Light Early Small Yes 1 Cement South Vegetation Rough Light Modern Small No 1 Cement West Vegetation Rough Light Modern Small No 1 Aluminum North Vegetation Smooth Light Modern Small Yes 0 Brick South Vegetation Smooth Dark Early Small No 0 Brick South Vegetation Smooth Dark Modern Small No 1 Brick West Vegetation Smooth Dark Contemporary Medium Yes 0 Brick North Vegetation Rough Dark Contemporary Small No 0

PAGE 100

91 2000. A new urban ecology. American Scientist 88: 416-425. REFERENCES Agresti, A. 1996. An introduction to categorical data. John Wiley & Sons Inc., New York, New York. Anderson, S. H. 2002. Managing our wildlife resources. Prentice Hall, Upper Saddle River, New Jersey. Angilletta, M. L., L. G. Montgomery, and Y. L. Werner. 1999. Temperature preference in geckos: diel variation in juveniles and adults. Herpetologica 55 (2): 212-222. Arnold, E. N. 1984. Ecology of lowland lizards in the eastern United Arab Emirates. Journal of Zoology: Preceedings of the Zoological Society of London 204: 329-354. Barbour, T. 1936. Two introduced lizards in Miami, Florida. Copeia 1936 (2): 113. Bartholomew, G. A. 1959. Photoperiodism in reptiles. In: Photoperiodism and related phenomena in plants and animals, edited by R. B. Withrow, American Association for the Advancement of Science, Washington D.C. Bolen, E. G. and W. L. Robinson. 2003. Wildlife ecology and management. Prentice Hall, Upper Saddle River, New Jersey. Botkin, D. B. and C. E. Beveridge. 1997. Cities as environments. Urban Ecosystems 1: 3-19. Brown T. L., C. P. Dawson, and R. L. Miller. 1979. Interests and attitudes of metropolitan New York residents about wildlife. Transactions of the North American Wildlife and Natural Resource Conference 44: 289-297. Capula, M. and L. Luiselli. 1994. Trophic niche overlap in sympatric Tarentola mauritanica and Hemidactylus turcicus: a preliminary study. Herpetological Journal 4: 24-25. Carey, S. D. 1988. Food habits of gulf coast Mediterranean geckos (Hemidactylus turcicus). Journal of the Alabama Academy of Sciences 59 (3): 103. Collins, J. P., A. Kinzig, N. B. Grimm, W. F. Fagan, D. Hope, J. G. Wu, and E. T. Borer.

PAGE 101

92 Conant, R. 1955. Notes on three Texas reptiles, including an addition to the fauna of the state. American Museum Novitates 1726: 1-6. Conant, R. and J. T. Collins. 1998. A field guide to reptiles and amphibians of eastern State University, Nacogdoches, Texas. and central North America. Houghton Mifflin Co., New York, New York. Davis, W. K. 1974. The Mediterranean gecko, Hemidactylus turcicus in Texas. Journal of Herpetology 8 (1): 77-80. Dow, K. 2000. Social dimensions of gradients in urban ecosystems. Urban Ecosystems 4: 255-275. Etheridge, R. E. 1952. The warty gecko, Hemidactylus turcicus turcicus (Linnaeus), in New Orleans, Louisiana. Copeia 1952 (1): 47-48. Frankenberg, E. 1978. Intraspecific and seasonal variation of daily activity times in gekkonid lizards (Reptilia, Lacertilia). Journal of Herpetology 12 (4): 505-519. ---. 1979. Influence of light and temperature on daily patterns of 3 Israeli forms of Ptyodactylus (Reptilia: Gekkoninae). Journal of the Zoological Society of London 189: 21-30. ---. 1982. Vocal behavior of the Mediterranean house gecko, Hemidactylus turcicus." Copeia 1982 (4): 770-775. Gilbert, O. L. 1989. The ecology of urban habitats. Chapman & Hall, New York, New York. Grimm, N. B., J. M. Grove, S. T. A. Pickett, and C. L. Redman. 2000. Integrated approaches to long-term studies of urban ecological systems. Bioscience 50 (7): 571-584. Heatwole, H. 1977. Habitat selection in reptiles. In: Biology of the Reptilia, Vol. 7 (Ecology and Behavior A), edited by C. Gans, Academic Press Inc., New York, New York. Kearney, M., R. Shine, S. Comber, and D. Pearson. 2001. Why do geckos group? An analysis of "social" aggregations in two species of Australian lizards. Herpetologica 57 (4): 411422. King, W. 1958. Observations on the ecology of a new population of Mediterranean gecko, Hemidactylus turcicus, in Florida. Quarterly Journal of the Florida Academy of Science 21 (4): 317-318. Klawinski, P. D. 1991. Home range, activity, and spatial distribution of the Mediterranean gecko, Hemidactylus turcicus. Master's Thesis, Stephen F. Austin

PAGE 102

93 Klawinski, P. D., R. K. Vaughan, D. saenz, and W. Godwin. 1994. Comparison of dietary overlap between allopatric and sympatric geckos. Journal of Herpetology 28 (2): 225-230. astr.gsu.edu/hbase/hframe.html, Accessed: April 19, 2003. Knox, P. L. 1991. The restless urban landscape: economic and sociocultural change and the transformation of metropolitan Washington, DC. Annals of the Association of American Geographers 81: 181-209. Loeb, R. E. 1998. Urban ecosystem management and change during the past millennium: a case study from New York City. Urban Ecosystems 2: 17-26. Lord, C. P., E. Strauss, A. Toffler. 2003. Natural cities: urban ecology and the restoration of urban ecosystems. Virginia Environmental Law Journal 21: 317-381. Loveridge, A. 1947. Revision of the African lizards of the family Gekkonidae. Bulletin of the Museum of Comparative Zoology 98 (1): 1-469. Luiselli, L. and D. Capizzi. 1999. Ecological distribution of the geckos Tarentola mauritanica and Hemidactylus turcicus in the urban area of Rome in relation to age of buildings and condition of walls. Journal of Herpetology 33 (2): 316-319. Luniak, M. and B. Pisarski. 1994. State of research into the fauna of Warsaw (up to 1990). Memorabilia Zoology 49: 155-65. Lyons, J. R. 1997. Urban ecosystem management: bringing science and policy together. Urban Ecosystems 1: 77-83. Marcellini, D. L. 1977. Acoustic and visual display behavior of gekkonid lizards. American Zoology 17: 251-260. McIntyre, N. E., K. Knowles-Yanez, D. Hope. 2000. Urban ecology as an interdisciplinary field: differences in the use of "urban" between the social and natural sciences. Urban Ecosystems 4: 5-24. Meshaka, W. E., Jr. 1995. Reproductice cycle and colonization ability of the Mediterranean gecko (Hemidactylus turcicus) in south-central Florida. Florida Scientist 58 (1): 10-15. Natural Resources Canada. 2003. Geomagnatism, Geological Survey of Canada, (Online). Available at: http://www.geolab.nrcan.gc. ca/geomag/, Accessed: April 12, 2003. Nave R. 2000. Hyperphysics, Department of Physics and Astronomy, Georgia State University, Atlanta, Georgia, (Online). Available at: http://230nsc1.phy-

PAGE 103

94 Niemela, J. 1999. Is there a need for a theory of urban ecology? Urban Ecosystems 3: 57-65. Nilon, C. H. and R. C. Pais. 1997. Terrestrial vertebrates in urban ecosystems: developing hypotheses for the Gwynns Falls Watershed in Baltimore, Maryland. Urban Ecosystems 1: 247-257. Pickett, S. T. A., M. L. Cadenasso, J. M. Grove, C. H. Nilon, R. V. Pouyat, W. C. Zipperer, and R. Constanza. 2001. Urban ecological systems: linking terrestrial ecological, physical, and socioeconomic components of metropolitan areas. Annual Review of Ecological and Systematics 32: 127-157. Punzo, F. 2001a. The Mediterranean gecko Hemidactylus turcicus: life in an urban landscape. Florida Scientist 64 (1): 56-66. Punzo, F. 2001b. A comparative study of digestive function in three species of hemidactyline geckos from Florida. Florida Scientist 64 (2): 124-130. Rebele, F. 1994. Urban ecology and special features of urban ecosystems. Global Ecology and Biogeography Letters 4: 173-187. Rees, W. E. 1997. Urban ecosystems: the human dimension. Urban Ecosystems 1: 63-75. Rose, F. L. and C. D. Barbour. 1968. Ecology and reproductive cycles of the introduced gecko, Hemidactylus turcicus, in the southern Unites States. The American Midland Naturalist 79 (1): 159-168. Saenz, D. 1992. Dietary analysis of Hemidactylus turcicus, the Mediterranean gecko, a population in Nagodoches, Texas. Master's Thesis, Stepehen F. Austin State University, Nacogdoches, Texas. Saenz, D. 1996. Dietary overview of Hemidactylus turcicus with possible implications of food partitioning. Journal of Herpetology 30 (4): 461-466. Schauman S., S. Penland, M. Freeman. 1987. Public knowledge of and preferences for wildlife habitats in urban open space. In: Integrating Man and Nature in the Metropolitan Environment, edited by L. W. Adams and D. L. Leedy, National Institute for Urban Wildlife, Columbia, Maryland. Schoener, T. W. 1970. Nonsynchronous spatial overlap of lizards in patchy habitats. Ecology 51 (3): 408-418. Selcer, K. W. 1986. Life history of a successful colonizer: the Mediterranean gecko, Hemidactylus turcicus, in southern Texas. Copeia 1986 (4): 956-962.

PAGE 104

95 Stejneger, L. 1922. Two geckos new to the fauna of the United States. Copeia 1922 (108): 56. Trauth, S. E. 1985. Nest, eggs, and hatchlings of the Mediterranean gecko, Hemidactylus turcicus (Sauria: Gekkonidae). The Southwestern Naturalist 30 (2): 309-310. Trout, L. and T. D. Schwaner. 1994. Allozyme evidence for insularity in exotic populations of the Mediterranean gecko (Hemidactylus turcicus). Journal of Herpetology 28 (3): 391-393. University of Florida. 2002. Physical Facilities and Their Utilization, Office of Institutional Research (Online). Available at: http://www.ir.ufl.edu/factbook/facilities.htm, Accessed: March 25, 2003. University of Florida Physical Plant Division, Architecture/ Engineering Department. 2000. 2000 Building information list for the University of Florida. Gainesville, Florida. United States Department of Veterans Affairs. 2003. Gainesville Division, North Florida/ South Georgia Veterans Health System (Online). Available at: http://www.va.gov/sta/guide/facility.asp?ID=54, Accessed: March 25, 2003. Whitney G. G. and S. D. Adams. 1980. Man as a maker of new plant communities. Journal of Applied Ecology 17: 431-448.

PAGE 105

BIOGRAPHICAL SKETCH Patricia Gomez Zlatar was born in the small town of Punta Arenas, located at the southern tip of Chile. She eventually found her way to Montreal, Canada, where she received a bachelors degree in biology in 1999 from Concordia University. After a brief respite from academic life, she entered the University of Florida Department of Wildlife Ecology and Conservation in Gainesville during the fall of 2000. She received her Master of Science degree, with a minor in statistics, in 2003. 96