A STUDY OF THE TIMBER RATTLESNAKE (Crotalus horridus) IN A FRAGMENTED AGRICULTURAL LANDSCAPE A STUDY OF THE TIMBER RATTLESNAKE (Crotalus horridus) IN A FRAGMENTED AGRICULTURAL LANDSCAPE
A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Biology
By
Rodney Dale Wittenberg Avila College Bachelor of Science in Biology, 1999 University of Texas at Tyler Master of Science in Biology, 2001
December 2009 University of Arkansas ABSTRACT
Few ecological studies have focused on understanding how reptiles use
anthropogenically disturbed habitats. Consequently, this research bias has constrained
efforts to conserve the ever-increasing number of reptiles threatened by human-induced
landscape changes. One such reptile, the Timber Rattlesnake (Crotalus horridus), is a
well-studied North American snake species considered to be imperiled by the loss and
fragmentation of closed-canopy forest. Despite the threat posed to the Timber
Rattlesnake by anthropogenic changes in habitat structure, most research on this species has been carried-out in large tracts of mature forest. Given the important conservation
implications of understanding how this species uses anthropogenic habitat, I studied the
ecology (spatial, foraging, and thermal) and life history (individual growth rates) of a
Timber Rattlesnake population that uses a fragmented agricultural landscape in west-
central Missouri.
Despite large differences in habitat structure, the spatial data and movement patterns of individuals in this study monitored with radiotelemetry were similar to those previously reported for individuals in closed-canopy forests. However, 9 of the 27 individuals implanted with radiotransmitters (33.3 %) exited the study when they were found dead upon relocation. Sources of mortality were known for 6 of the dead individuals (66.7%), all of which were human-induced. Snakes readily foraged in secondary woodland tracts and corridors, agricultural fields, and habitat edges. Dietary analysis found that snakes fed exclusively on mammals including shrews (Soricidae), mice in the genus Peromyscus, Prairie Voles {Microtus ochrogaster), Cotton Rats
(Sigmodon hispidus), Eastern Gray Squirrels (Sciurus caro linens is), and Eastern Cottontails (Sylvilagus floridanus). Although small mammal trapping indicated that fields contained fewer numbers of prey than woodlands and habitat edges, field dwelling
Prairie Voles were the most frequently consumed prey item.
Data obtained using temperature sensitive radio-transmitters indicated that median snake body temperature (7b) did not differ among woodland (median = 28.0 °C, range =
16.4 - 34.2 °C, n = 63), edge (median = 28.6 °C, range = 16.7 - 33.9 °C, n = 56), or field
(median = 28.2 °C, range = 20.2 - 34.4 °C, n = 61) habitats. Additionally, operative temperature models were used to quantify thermal constraints on activity in both a woodland and a field. During each two week interval of the sampling period (02 June to
26 August 2006), the overall percentage of thermally available habitat was high in both woodland (range 84.2 - 99.8 %) and field (range 78.1 - 94.9 %) habitats. However, sharp decreases in habitat availability occurred in both the woodland and field during midday. The lowest hourly percentages of thermally available habitat were recorded at
1400 hours in the field (19.7%) and at 1500 hours in the woodland (36.7%) during the two-week interval from 29 July to 11 August. Body temperatures of gravid females gestating within a man-made quarry (median = 32.5 °C, range = 25.3 - 34.7 °C, n = 113) were significantly warmer and less variable than 7b's of males and non-gravid females using woodlands, edges, and fields (median = 28.3 °C, range = 16.4 - 34.4 °C, n = 180).
Finally, mark-recapture data and rattle morphology were used to compare the birth size and early growth rates of timber rattlesnakes in a fragmented habitat of west central Missouri (MO) to those in a closed-canopy forest of northwest Arkansas (AR).
Missouri snakes increased in length more rapidly than their AR counterparts through their first eight ecdyses. Furthermore, males and females from MO diverge in size between the fifth and sixth ecdysis event, while growth trajectories of AR males and females remain indistinguishable through eight ecdyses. Despite climatic data suggesting that AR snakes may potentially have an average of 8.5% more time each season to acquire and assimilate prey, MO snakes still exhibited superior growth and early maturation.
Results of this study suggest that Timber Rattlesnakes may not require large tracts of closed-canopy forest as long as their thermal and dietary needs are met. Thus, employing habitat management techniques that enhance small mammal densities may be a more effective strategy for conserving Timber Rattlesnake populations than strictly attempting to preserve closed-canopy forest. Future studies should use habitat manipulation as a tool to better understand how forest-dwelling Timber Rattlesnakes respond to changes in habitat structure. This dissertation is approved for Recommendation to the Graduate Council
Dissertation Director:
Dr. Steven J. Beaupre
Dissertation Committee:
Dr. Ines Pinto
Dr. Kimberly G. Smith
Dr. Edward E. Gbur Jr. DISSERTATION DUPLICATION RELEASE
I hereby authorize the University of Arkansas Libraries to duplicate this dissertation when needed for research and/or scholarship.
Agreed Rodney D. Wittenberg
Refused Rodney D. Wittenberg ACKNOWLEDGEMENTS
Although few have the opportunity to learn from the best in their respective field,
I consider myself "one of the few". I will always be indebted to my friend and advisor,
Dr. Steven Beaupre, for generously sharing his time, knowledge, and resources to ensure
that I fulfilled a life long dream. I intend to do things in life, personally and
professionally, that will make you smile (with pride of course)!
As members of my graduate committee, Dr. Edward Gbur Jr., Dr. Ines Pinto, and
Dr. Kimberly G. Smith provided guidance and helpful comments on my dissertation. I
sincerely appreciate their service.
I benefited greatly from interactions with labmates past (Nick Haertl, Dr. Jacques
Hill, Dr. Marshal McCue, Carolina Monteiro, Dr. Chad Montgomery, Dr. Melissa
Pilgrim, and Dr. Fred Zaidan) and present (Joseph Agugliaro, Lara Douglas, Jason
Ortega, Matthew Smith, and James Van Dyke). In one way or another, each of you
contributed to my success. For that I thank you.
My best friend, Patrick Koontz, provided both assistance and companionship at my study site. Together we flirted with heat stroke, huddled in caves to avoid electrical
storms, proved we were immune to insect borne diseases, destroyed several pairs of boots, and had some amazing encounters with the local wildlife. I will always cherish our adventures.
I thank my family for their love, support, and encouragement. Extreme examples of their love and support are in order: My father, Robert Wittenberg, allowed me to process my study animals in his spare bedroom. My mother, Peggy Wittenberg, allowed
vii me to keep roadkill in the crisper of her refrigerator until it could be preserved or transported elsewhere. Amazing! My grandparents, Virgil and Betty Kempf, have always made me feel special. I thank my Uncle Don and Aunt Diana Wittenberg for always rooting for me. Even before I married my wife, Sara, the Ress family treated me as one of their own. Tom, Roberta, and Michael.. .you have been wonderful.
Aside from allowing my study to be conducted on their property, several landowners took a keen interest in me, my project, and the Timber Rattlesnake. For this,
I will always be indebted to the Bishop, Debrick, Franklin, Sutton, Thomas, and York families. I am honored to be your friend.
I thank the Beckers (Bill and Gwen), the Brewsters (Ambre, Brad, Owen, and
Addyson), Adam Crane, Jen Dorr, Matt Dekar, the Ewings (Scott, Linda, Desmond, and
Jackson), Andrea Green, Dr. Ron Gutberlet, Ken Hester, Chad Leslie, Jason Luscier, Dr.
Glenn Manning, Dr. Nancy McCartney, Alex Muensch, Dr. Douglas James, Dr. Robert
Powell, Dana Savorelli, Nora Schubert, the Slays (Mike and Christy), Dr. Sarah Spurrier, and Dr. James Walker. I trust that each of you know why I am saying, "thank you".
Finally, this research was approved by the University of Arkansas Institutional
Animal Care and Use Committee (protocol # 05001) and the Missouri Department of
Conservation (collecting permits # 12005, 12367, 12715, 13101). Partial funding was provided by a University of Arkansas Causey Grant-in-Aid Award, a Harry Steinman
Memorial Grant from the St. Louis Herpetological Society, and a grant from the
Arkansas Audubon Society Trust.
vm DEDICATION
I dedicate this dissertation to my wife, Sara (Ress) Wittenberg. Although I have given you an entire page, I am struggling to find the words to fill it. You have enriched my life in every way imaginable, and have done so much to ensure my success. I love you dearly and look forward to the adventures that lie ahead.
IX TABLE OF CONTENTS
CHAPTER 1. HABITAT FRAGMENTATION AND THE TIMBER RATTLESNAKE (CROTALUS HORRID US): AN INTRODUCTION
Abstract 1 Habitat Fragmentation Defined 1 Taxonomic Biases in Habitat Fragmentation Studies 3 The Timber Rattlesnake as a Model for Forest Fragmentation Studies 5 The Timber Rattlesnake in Fragmented Habitats: A Tabular Review 6 Regional Biases 8 Biases Related to Habitat 8 Timber Rattlesnakes in Logged Forests 10 Timber Rattlesnakes in Agricultural Fields 10 Use of Anthropogenic Structures by Timber Rattlesnakes 11 Directives for Future Research 11 A Study of the Timber Rattlesnake in a Fragmented Landscape 13 Study Site 13 Study Objectives 15 Dissertation Overview 16 Acknowledgments 18 Literature Cited 18 Tables 33 TABLE 1. Literature Survey of Timber Rattlesnake (Crotalus horridus) Studies and Observations 33
CHAPTER 2. SPATIAL ECOLOGY OF THE TIMBER RATTLESNAKE {Crotalus horridus) IN A FRAGMENTED AGRICULTURAL LANDSCAPE
Abstract 42 Introduction 43 Materials and Methods 45 Study Site 45 Radiotelemetry 45 Spatial Analyses 46 Results 49 Home Range and Core Areas 49 Dispersal Distances 50 Fidelity to Rookery Habitat and Individual Hibernacula 52 Human-Induced Mortality 53 Discussion 53 Home Range and Core Areas 53 Dispersal Distances and Movement Patterns 56 Fidelity to Rookery Habitat and Individual Hibernacula 58 Human-Induced Mortality 59 Conclusions 60
x Acknowledgments 60 Literature Cited 61 Tables 66 TABLE 1. Estimates of home range area and area(s) of core use for 9 adult Timber Rattlesnakes 66 TABLE 2. Published home range estimates of adult Timber Rattlesnakes using the MCP method 68 TABLE 3. Published home range estimates of adult Timber Rattlesnakes derived from two kernel based approaches 71 TABLE 4. Published core area estimates of adult Timber Rattlesnakes derived from two kernel based approaches 73 TABLE 5. Published maximum dispersal distances of adult Timber Rattlesnakes from hibernacula 75 Figures 77 FIGURE 1. Mean monthly dispersal distances of male Timber Rattlesnakes 77 FIGURE 2. Mean monthly dispersal distances of non-gravid female Timber Rattlesnakes 79 FIGURE 3. Mean monthly dispersal distances of female Timber Rattlesnakes during a season in which they were gravid 81 Appendix 1 83 MAP IMAGE 1. Home range, core areas, and radiotelemetry relocations of adult male 150 84 MAP IMAGE 2. Home range, core area, and radiotelemetry relocations of adult male 193 86 MAP IMAGE 3. Home range, core areas, and radiotelemetry relocations of adult male 099 88 MAP IMAGE 4. Home range, core area, and radiotelemetry relocations of adult male 210 90 MAP IMAGE 5. Home range, core areas, and radiotelemetry relocations of adult male 802 92 MAP IMAGE 6. Home range, core area, and radiotelemetry relocations of adult male260B 94 MAP IMAGE 7. Home range, core area, and radiotelemetry relocations of adult female 060 96 MAP IMAGE 8. Home range, core areas, and radiotelemetry relocations of adult female 821 98 MAP IMAGE 9. Home range, core area, and radiotelemetry relocations of adult female 760B 100 MAP IMAGE 10. Composite map of the MCP home ranges of nine individual Timber Rattlesnakes 102 MAP IMAGE 11. Exclusive use of an inactive rock quarry by radiotelemetry- monitored Timber Rattlesnakes for both hibernation and gestation 104
CHAPTER 3. FORAGING ECOLOGY OF THE TIMBER RATTLESNAKE (Crotalus horridus) IN A FRAGMENTED AGRICULTURAL LANDSCAPE
XI Abstract 106 Introduction 107 Materials and Methods 108 Study Site 108 Study Animals 109 Observations of Foraging Behavior 109 Dietary Analysis 110 Mammal Trapping Ill Results 113 Observations of Foraging Behavior 113 Dietary Analysis 114 Mammal Trapping 115 Discussion 116 Observations of Foraging Behavior 116 Dietary Analysis 119 Mammal Trapping 122 Conclusions and Management Implications 123 Acknowledgments 124 Literature Cited 125 Tables 131 TABLE 1. Records of habitat use and foraging behavior within each habitat for twenty-six radio-tagged Timber Rattlesnakes 131 TABLE 2. Results of contingency table analyses testing the null hypothesis that the number of small mammals captured in fields and edges would not differ significantly from those captured in woodlands 134 TABLE 3. Number of individuals from each of seven different small mammal species captured in woodland, field, and edge habitats 136 Figures 138 FIGURE 1. Total number of each type of mammal prey recovered from Timber Rattlesnakes in the study population 138 FIGURE 2. Relationship between snake size (SVL) and type of mammal prey consumed 140 Appendix I. Tabular listing of mammal species that 1) as juveniles and/or adults could possibly be prey for Timber Rattlesnakes and 2) have geographic distributions that overlap or come very near the study site 142 Appendix II. List of dietary material examined 145
CHAPTER 4. THERMAL ECOLOGY OF THE TIMBER RATTLESNAKE (Crotalus horridus) IN A FRAGMENTED AGRICULTURAL LANDSCAPE
Abstract 148 Introduction 149 Methods 152 Study Species 152 Study Site 152 Characterization of Habitat Types and Population Segments 153
xn Body Temperature Measurements 154 Operative Temperature Measurements 157 Results 161 Body Temperatures in Woodlands, Edges, and Fields 161 Thermal Constraints 161 Body Temperatures of Females Gestating Within an Inactive Quarry 162 Discussion 163 Habitat-Specific Thermal Biology 163 Thermal Constraints on Activity 164 Thermal Biology of Gestating Females 168 Use of Rock Quarry as a Rookery 172 Conclusions 173 Acknowledgments 174 References 174 Figures 181 FIGURE 1. Comparison of Jj, distributions for Timber Rattlesnakes occupying three distinct habitat types 181 FIGURE 2. Temporal distribution of Timber Rattlesnake Ji, records during well- sampled hours of the day (ca. 1400 - 2030 hours) 183 FIGURE 3. Overall percentage of habitat thermally available for Timber Rattlesnakes in both a woodland and field between 02 June and 26 August 2006 (presented in ca. two week intervals) 185 FIGURE 4. Hourly percentages of habitat thermally available for Timber Rattlesnakes in both a woodland and field during the approximate two week intervals of 02 June- 16 June and 17 June-30 June 187 FIGURE 5. Hourly percentages of habitat thermally available for Timber Rattlesnakes in both a woodland and field during the approximate two week intervals of 01 July-14 July and 15 July-28 July 189 FIGURE 6. Hourly percentages of habitat thermally available for Timber Rattlesnakes in both a woodland and field during the approximate two week intervals of 29 July - 11 August and 12 August - 26 August 191 FIGURE 7. Comparison of T\, distributions from two behaviorally distinct segments of the study population 193
CHAPTER 5. THE TIMBER RATTLESNAKE {Crotalus horridus) EXHIBITS RAPID GROWTH IN A FRAGMENTED HABITAT
Abstract 195 Introduction 196 Materials and Methods 198 Field Sites 198 Data Collection 199 Potential Effects of Climate on Resource Acquisition 200 Rattle Morphology as an Indicator of Growth 200 Common Garden Study 202 Data Analysis 203
xin Results 204 Potential Effects of Climate on Resource Acquisition 204 Size at Birth 205 Basal Segment Width and SVL Correlation 205 Comparison of Individual Growth Rates 206 Common Garden Analysis 206 Discussion 207 Acknowledgments 211 References 212 Tables 218 TABLE 1. Results of a repeated measures ANOVA testing the effects of population and sex on rattle segment width 218 Figures 220 FIGURE 1. Basal segment width correlates positively with snout-vent length (SVL), and the relationship is consistent between MO and AR timber rattlesnake populations 220 FIGURE 2. Significant rattle segment x population x sex interaction indicates that MO timber rattlesnakes grow more rapidly and mature earlier than snakes from the AR population 222
CONCLUSION 224
xiv CHAPTER 1: HABITAT FRAGMENTATION AND THE TIMBER RATTLESNAKE
(CROTALUSHORRIDUS): AN INTRODUCTION
Abstract
As a group, reptiles have been largely underrepresented in studies aimed at
understanding how vertebrates use fragmented habitats. One such reptile, the timber
rattlesnake {Crotalus horridus), is an imperiled species considered to be greatly
threatened by the loss and fragmentation of closed-canopy forest. Nevertheless, a review
of 137 documents revealed that research on timber rattlesnakes in fragmented habitats has
been slow to emerge. The single greatest contribution of documents (39.4%) stemmed
from the United States Fish and Wildlife Service's Northeast Region, where the species
has been studied primarily in heavily forested mountains. More documents (48.9%)
reported on timber rattlesnakes in closed-canopy forests than any other habitat type,
regardless of geographic region. In fact, habitats that could be described as fragmented
were represented in only 12.4% of the reviewed documents. Interestingly, reports of
timber rattlesnakes using logged areas, agricultural fields, and anthropogenic structures
such as abandoned quarries and mines do exist in the literature. However, little is known
regarding the ecology of timber rattlesnakes in these disturbed, fragmented habitats. To better understand the threat posed to timber rattlesnakes by habitat fragmentation, future research must be conducted on populations residing in fragmented landscapes.
Habitat fragmentation defined
1 Habitat fragmentation is a process that breaks large continuous tracts of a species'
habitat into smaller patches that are separated by areas of unsuitable habitat (Franklin et
al. 2002). Occasionally these habitat patches remain loosely connected via corridors of
suitable habitat (Fahrig and Merriam 1985). Habitat fragmentation may occur naturally
as a result of long-term geological processes (Watson 2003), or as a short-term
consequence of natural events such as fire and windfall (Andren 1994). However, the
most ecologically important and widespread form of habitat fragmentation is caused by
human activities (Andren 1994). Most studies focus on the latter form of habitat
fragmentation and are conducted in landscapes where the fragmentation process has
occurred in the past 100 years (Watson 2002). Such emphasis on anthropogenic habitat
fragmentation may be born out of necessity as human activities continue to threaten
biodiversity through habitat loss, species extinction, disrupted communities, and damaged
ecosystems (Templeton et al. 2001). Land use practices linked to agriculture and
development are considered primary agents of anthropogenic habitat fragmentation
(Saunders et al. 1991).
Many studies have examined the adverse effects of habitat fragmentation on local
wildlife, whether it be a single species or an entire community (Andren 1992; Burke and
Nol 1998; Zanette et al. 2000; Sumner 2005). Suffering from a critical loss or reduction
of habitat (Templeton et al. 2001), terrestrial vertebrates may adopt altered movement patterns that increase mortality (Quinn and Hastings 1988), experience density reductions that may threaten population survival (Fahrig and Merriam 1985, 1994), face incursion threats from non-resident species through competition or predation (Cantrell et al. 2001), or suffer from a loss of genetic diversity (Templeton et al. 2001; Ujvari et al. 2002).
2 Despite an emphasis on negative impacts of habitat fragmentation in ecological literature,
some species are positively impacted through the creation of habitat edges (Alverson et
al. 1988). At the interface of contrasting habitats, organisms may encounter structural,
thermal, or dietary resources that were unavailable prior to fragmentation (Carfagno et al.
2006; Carfagno and Weatherhead 2006).
Taxonomic biases in habitat fragmentation studies
Habitat loss and fragmentation have been identified as important factors in the
global decline of reptile populations (Gibbons et al. 2000). Nevertheless, Mac Nally and
Brown (2001) report a significant taxonomic bias in the scientific literature concerning
habitat fragmentation and vertebrates. Their search of two different bibliographical
databases revealed that habitat fragmentation studies conducted on birds (n=327) and
mammals (n=294) far outnumbered those on amphibians (n=34) and reptiles (n=22). In
eastern Australia, Heard et al. (2004) report that although the movement and habitat use
patterns of birds and mammals have been well-studied in fragmented forests, there is a
paucity of comparable data for many reptilian species residing in the same habitat.
According to Bury (2006), our knowledge of herpetofaunal natural history stems mostly
from past studies conducted in pristine habitats, largely undermining our efforts to
conserve these organisms. In order to successfully manage reptile and amphibian populations in human-impacted landscapes, Bury indicates future studies must be
conducted in areas where human perturbations occur.
It is disconcerting that reptiles are the least represented vertebrate group in studies
of habitat fragmentation considering they possess two attributes that make them
3 especially vulnerable to changes in structural habitat that are not shared by their
endothermic counterparts. First, most reptile species exhibit site fidelity and lack the
short-term vagility needed to migrate from disturbance and reestablish themselves in
more favorable habitat elsewhere (Gibbons et al. 2000; Diaz et al. 2000). Second,
changes in structural habitat (i.e. canopy cover) affect the physical properties of the
earth's surface, thereby affecting microclimates greatly (Robinson, 1966; Porter and
Gates, 1969; Newmark, 2001; Schlaepfer and Gavin, 2001; Pringle et al., 2003). Because
reptiles must obtain heat from the surrounding physical environment, the effects of
habitat fragmentation on the behavior and physiology of these organisms may be
pervasive (Huey, 1991). The relationship between organism and thermal environment
directly affects physiological processes such as locomotory performance, digestion, reproduction, and immune function (Huey 1982; Adolph and Porter 1993). Indirect
effects arise through thermal constraints on activity (spatial and/or temporal) that may
limit the individual's ability to extract resources from the environment (Grant and
Dunham 1988; Beaupre 1995). Resource-limited individuals may spend a
disproportional amount of their activity budget foraging and have less available energy to
allocate to the competing functions of growth, maintenance, storage, and reproduction
(Dunham et al. 1989). The result maybe decreased growth, delayed maturation, fewer opportunities to mate, and reduced fecundity (Dunham et al. 1989). Therefore, both direct and indirect effects of the thermal environment on individual life history and fitness can greatly influence reptilian population dynamics (Beaupre et al. 1993; Beaupre
1995; Beaupre 2002).
4 The timber rattlesnake as a model for forest fragmentation studies
The timber rattlesnake (Crotalus horridus) is an ambush-foraging pitviper that associates with the deciduous forest biome of eastern North America (Brown 1993). The species is currently recognized as vulnerable, imperiled, or critically imperiled in 20 of the 30 U.S. states where populations still remain (Waldron et al. 2006). Despite being afforded a degree of legal protection in many of these states, the timber rattlesnake has vanished throughout parts of its former range (Brown 1993). In fact, the timber rattlesnake has been extirpated from Rhode Island, Maine, and possibly the Canadian province of Ontario (Brown 1993). Most consider this rattlesnake species to rely on a closed-canopy ecosystem (Collins and Knight 1980; Reinert 1984a; Reinert and
Zappalorti 1988a; Martin 1992a; Rudolph et al. 1998), and fragmentation of this woodland habitat has been cited as a primary threat to the survival of the species (Martin
1992a; Brown 1993; Clark et al. 2003; Furman 2007).
Timber rattlesnake behaviors that incorporate trees and logs may support the notion that this species is a woodland specialist. There have been numerous reports of arboreal behavior (Saenz et al. 1996; Coupe 2001; Fogell et al. 2002b; Bartz and Sajdak
2004; Rudolph et al. 2004; Sajdak and Bartz 2004), including observations of arboreal courtship (Bartz and Sajdak 2004) and the striking and holding of avian prey (Sajdak and
Bartz 2004). A telemetry study in east Texas detected a high frequency of arboreal activity among 12 subadult snakes, as 35 of 218 relocations occurred in trees at heights ranging between 0.8 and 14.5 meters (Rudolph et al. 2004).
Reinert et al. (1984) describe a unique foraging strategy where tightly coiled snakes lay beside fallen logs and place their chin on the log itself, perpendicular to the
5 log's long axis. Reinert and colleagues propose 1) chemoreception is initially used to select logs that serve as active small mammal runways and 2) even a sleeping snake would be alerted to the presence of a traveling rodent through ground born vibrations transmitted from the log to the snake's lower jaw. Brown and Greenberg (1992) observed an adult male timber rattlesnake that repeatedly positioned itself at the base of trees with its head and loosely coiled neck extended vertically up the tree's trunk. During one such postural episode, the authors moved a freshly-killed chipmunk (Tamias striatus) down the tree trunk towards the snake. The snake immediately delivered a predatory strike from its position. The same vertical-tree ambush posture has been observed in south-central Indiana, where adult timber rattlesnakes select microhabitats based on the presence of tree trunks and log cover (Walker 2000). In South Carolina, 31% of female foraging took place near logs, while 23% of male foraging occurred at the base of trees
(Waldron et al. 2006). Specialized foraging tactics that incorporate woodland structure may suggest that timber rattlesnakes are not habitat generalists. Therefore, the species may not be able to successfully expand its activities beyond shrinking woodland tracts and into more open habitats.
The timber rattlesnake in fragmented habitats: a tabular review
Worldwide, there are approximately 2,700 snake species currently recognized
(Greene 1997). Of these, the timber rattlesnake is one of the most intensely studied.
Given that habitat fragmentation is a threat to the species' survival, I was interested in determining how much emphasis has been placed on studying timber rattlesnakes in disturbed areas. Therefore, I surveyed the literature to determine the following:
6 1) Are studies and reports on the timber rattlesnake biased towards the
Northeast geographic region, where remote areas of heavily forested,
mountainous terrain remain intact?
2) Are studies and reports biased towards timber rattlesnakes that reside in
closed- canopy forests?
3) What is known about timber rattlesnakes in logged forests, agricultural fields,
or other anthropogenically modified habitats?
The result of this effort was a tabulation of 137 documents on the timber rattlesnake,
consisting primarily of field and laboratory studies as well as natural history observations
(Table 1). An effort was made to obtain ecologically relevant literature such as that
pertaining to movement, habitat selection or use, reproduction, diet, behavior, thermal
biology, historical distribution, population genetics, morphology, systematics, and causes
of decline. Studies of timber rattlesnake venom and the clinical aspects of their bite were
considered to be beyond the scope of this review. I summarized published material from journals, unpublished theses and dissertations, and herpetological society newsletters and
bulletins. Two publications from the Society for the Study of Amphibian and Reptiles, a
loose-leaf catalogue account for the species (Collins and Knight 1980) and a guide for
conservation (Brown 1993), have been included because of their importance to
researchers. Unpublished government reports, regional field guides, popular books, magazine articles, meeting abstracts, and distributional notes that lacked any meaningful habitat data were purposefully omitted. Literature surveyed spans a period of 80 years
(1929-2008) and includes data gathered on timber rattlesnakes from 30 different states. Regional biases
For each timber rattlesnake study or report reviewed, an attempt was made to
determine the state(s) where the subject animals were from. Documents were then
assigned to a geographic region, using the regional delineation of states adopted by the
United States Fish and Wildlife Service. The largest contribution of documents from any
single region (39.4%) were from the Northeast Region, where dwindling populations
have made timber rattlesnake conservation a priority. Long-term, pioneering studies
have been conducted by William S. Brown (New York), William H. Martin (Maryland,
Pennsylvania, Virginia, West Virginia), Howard K. Reinert (New Jersey, Pennsylvania)
and their colleagues. The Southeast Region accounted for 22.6% of documents,
including many from Steven J. Beaupre's well-studied population in northwest Arkansas.
Surprisingly, only 13.1%) of documents were from the Midwest Region, given that timber
rattlesnakes occur in every state but Michigan. The species' range extends into the
Mountain-Prairie Region in the eastern portions of Kansas and Nebraska. Here, research
conducted by Henry S. Fitch and colleagues in Kansas, and Dan Fogell in Nebraska, has
contributed largely to the 8.8% of documents from this region. Finally, 5.8% of
documents report on timber rattlesnakes in the Southwest Region, where the species can be found in eastern portions of Texas and Oklahoma. Due to a paucity of information, I
was unable to assign 12.4% of documents to a geographic region.
Biases related to habitat
Nearly half (48.9%) of all reviewed documents on the timber rattlesnake were based on individuals or populations from closed-canopy forests. Considering the range
8 of the species, and their presumed preference for this habitat, such a bias in the literature was not unexpected. Much of this woodland habitat is described as "upland",
"deciduous", "hardwood", "mixed hardwood", or "oak-hickory". Additionally, the species has been studied in bottomland forest, pine forest, mixed forests containing both pine and hardwood, and the Pine Barrens of the Atlantic Coastal Plain.
Habitat descriptions were "na" (not available / not applicable) for 35.0% of the documents reviewed. Many laboratory studies did not provide a description of the habitat from which the study subjects originated, as it lacked relevance to the research being conducted. Examples of studies that did not address habitat structure included those pertaining to morphology, phylogeography, fossil history, physiology, recognition of chemical cues, and impacts of human harvest. Unfortunately, observations of timber rattlesnakes in the wild were occasionally published with little or no description of the habitat in which the observation occurred. Another 3.6% of documents provided habitat descriptions that could not be assigned to either closed-canopy or fragmented types.
Examples of such problematic descriptions include "talus slope" or "road near marsh".
Only 12.4%o of the documents reported on timber rattlesnakes residing in fragmented habitats. Of these 17 documents, 8 were from eastern Kansas where the species inhabits mixed habitat consisting of woodlands and grassy fields (Fitch and Shirer
1971; Fitch 1999; Fitch and Pisani 2002, 2005, 2006; Fitch et al. 2004; Pisani and Fitch
2006; Walker et al. 2008). A timber rattlesnake population in southeast Nebraska that frequents crop fields (Fogell 2000; Fogell et al. 2002a) was the subject of 3 documents.
South Carolina populations using pine plantations interspersed with open fields were also the subject of 3 documents (Gibbons 1972; Piatt et al. 2001; Waldron et al. 2006).
9 Timber rattlesnakes in logged forests
In West Virginia, Adams (2005) studied a population of timber rattlesnakes in an
Allegheny Mountain production forest. Telemetry revealed that although the snakes preferred stands of mature forest, they would also use edges and clearcuts in recently logged areas. In South Carolina, the species can be found in managed areas of planted pine (Gibbons 1972; Piatt et al. 2001; Waldron et al. 2006). A bobwhite quail being monitored with telemetry was found consumed by a timber rattlesnake in a recently cut and burned site of a loblolly pine plantation (Piatt et al. 2001). According to Stechert
(1982), timber rattlesnakes in New York initially benefited when mountains "too heavily timbered" were cleared, however, the escalation of logging and quarrying practices in the
1900s led to the eventual decline of these populations.
Timber rattlesnakes in agricultural fields
In Vermont, timber rattlesnakes reportedly forage in freshly mowed hayfields that border hardwood forests (Furman 2007). In Virginia during the first half of the 20th century, Martin (1979) describes how rattlesnake populations responded when abandoned farms fell under government control during the establishment of Shenandoah National
Park and the George Washington and Jefferson National Forests. The author remarks,
"Old fields with their rock piles and stone walls became overgrown. Rodents multiplied and rattlers followed, in artificially restored favorable environments." In South Carolina, timber rattlesnakes use both old fields (Gibbons 1972) and cultivated fields (Waldron et al. 2006). Populations use a mixture of woodland, prairie, and field habitats in eastern
Kansas (Fitch and Shirer 1971, Fitch 1999; Fitch and Pisani 2002, 2005, 2006; Fitch et al.
10 2004; Pisani and Fitch 2006). Fields of row crops have been used by timber rattlesnakes
in southeastern Nebraska (Fogell 2000; Fogell et al. 2002a).
Use of anthropogenic structures by timber rattlesnakes
Interestingly, I found several records of timber rattlesnakes using structures such
as old quarries, abandoned mines, and concrete dams. In a wooded region of Illinois
containing dolomite and sandstone bluffs, Walley (1963) describes finding specimens
near an abandoned mine. In eastern Missouri, Anderson (2006) found that the remnants
of an old limestone quarry contained two of the largest hibernacula located within his
study site. A rock quarry was also used by a gestating female in south-central Indiana
(Gibson 2003). Lardie (1999) documented a specimen taken from a concrete lake dam in
Oklahoma, and a study population in eastern Kansas hibernates within a string of rock
outcroppings located along a recently constructed reservoir (Fitch et al. 2004; Fitch and
Pisani 2005, 2006). At a pig farm in northern Missouri, rocks along a pond dam are used
extensively by timber rattlesnakes (Robert Powell, pers. comm.). Furthermore, broken
slabs of concrete spilling down an embankment provided hibernacula for both timber
rattlesnakes and copperheads in a suburb of Kansas City, Kansas (pers. obs.). It appears
timber rattlesnakes may not differentiate between anthropogenic and natural sources of
exposed rock within their habitat boundaries.
Directives for future research
The timber rattlesnake was one of the first snake species to be studied using telemetry (Fitch and Shirer 1971). Researchers have used telemetry to gain knowledge of
11 timber rattlesnake movement patterns (Galligan and Dunson 1979; Brown et al. 1982;
Reinert and Zappalorti 1988a; Reinert and Rupert 1999; Coupe 2002, Fitch et al. 2004;
Waldron et al. 2006), habitat selection (Reinert 1984a,b; Reinert and Zappalorti 1988a;
Waldron et al. 2006), thermal biology (Brown 1982; Brown et al. 1982; Wills and
Beaupre 2000), and foraging behavior (Reinert et al. 1984; Brown and Greenberg 1992;
Cundall and Beaupre 2001; Clark 2006a,b). Modern molecular techniques have resulted
in studies of population genetics, addressing questions relating to genetic variation and
gene transfer between hibernacula (Villarreal et al. 1996; Bushar et al. 1998; Anderson
2006; Clark et al. 2008). Even videography and remotely triggered cameras have been
employed to study timber rattlesnake behavior in the field (Sadighi et al. 1995; Cundall
and Beaupre 2001; Clark 2006a,b). Long-term monitoring and research has been
conducted on wild populations in many states (Brown 1993), owing to the imperiled
status of the species. Nevertheless, despite threats posed by habitat loss and
fragmentation, timber rattlesnake populations in relatively pristine, forested habitats
remain the focus of most studies.
Evidence does suggest that timber rattlesnakes can be successful in habitats
devoid of large, continuous tracts of closed-canopy forest (Fitch and Shirer 1971; Fitch
1999; Fogell 2000; Fitch and Pisani 2002, 2005, 2006; Fogell et al. 2002a; Fitch et al.
2004; Pisani and Fitch 2006). Rapid growth, a favorable life history attribute, has even been reported for a handful of specimens using a woodland-grassland habitat in eastern
Kansas (Pisani and Fitch 2006). Mechanism-driven studies aimed at learning how timber rattlesnakes use fragmented habitats are clearly needed. Information gained from these
12 studies may lead to the development of alternative conservation strategies for forest- dwelling populations when habitat preservation is no longer an option.
A study of the timber rattlesnake in a fragmented landscape
Study site-
To investigate the ecology of timber rattlesnakes in fragmented habitats, I studied the species in a diverse, agricultural landscape of west-central Missouri. The study took place between the springs of 2003 and 2007 in a highland area of what is considered the prairie geographic region (Schwartz and Schwartz 1959). Here, a string of hardwood hills extend across a flat landscape of what was once tall grass prairie (Schwartz and
Schwartz 1959; Sims 1988). Presently, the areas of relatively flat topography contain a mixture of structural habitat created and maintained by human land use practices. Small farms and rural residences are spread across an agricultural matrix interconnected by corridors comprised of wooded creek bottoms and fencerows. Field type is variable and consists of active cattle pasture, old growth field, frequently cut hay field, or cultivated row crops (primarily soybean and corn). Additionally, some areas of the highlands have become denuded, as trees have been cleared and replaced by homes, lawns, fields, and the occasional manmade pond or lake. Exposed surface rock is not uncommon, and glades can be found in both woodland tracts and open pasture.
Schwartz and Schwartz (1959) commented on the mammal diversity of
Missouri's prairie geographic region, noting that species typical of both forested and open habitat could be found. During my fieldwork, I noted that this trend could be extended to all terrestrial vertebrate taxa. For example, species that prefer
13 prairie/grassland habitat such as the prairie vole {Microtus ochrogaster), dickcissel (Spiza americana), eastern meadowlark (Sturnella magna), slender glass lizard (Ophisaurus attenuatus), bullsnake (Pituophis melanoleucus), prairie kingsnake {Lampropeltis calligaster) and ornate box turtle (Terrapene ornata) were all present. Likewise, in addition to the study organism, primarily forest dwelling species occurring at the study site included the raccoon (Procyon lotor), bobcat {Lynx rufus), Cooper's hawk (Accipiter cooperii), gray treefrog (Hyla versicolor/chrysoscelis), western worm snake (Carphophis vermis), and eastern box turtle {Terrapene Carolina). Schwartz and Schwartz (1959) attributed the region's mammalian diversity to both natural and anthropogenic factors.
The authors believed that the historical intermingling of two major vegetation types
(grassland and forest), as well as the habitat heterogeneity resulting from human land use practices, maintain a landscape capable of supporting both forest and grassland species.
Timber rattlesnakes in my study over-wintered in an approximately 32 hectare rock quarry located along a partially denuded fringe of the highlands. By interviewing landowners I determined that the quarry had been inactive for approximately 25 years prior to when my study began in 2003. According to these individuals, timber rattlesnakes were consistently found in the quarry even while active mining was underway. Although disuse of the quarry has allowed some vegetative re-growth to occur, canopy cover remains sparse. The mining created a wealth of outcroppings that serve as hibernacula and extensive boulder piles that function as rookeries for gestating females. In fact, gravid females spend most of the active season in the quarry gestating beneath one or more boulders. Males and non-gravid females, however, disperse from the quarry in all directions to forage in old fields, wooded fencerows, small woodland
14 tracts, roadside ditches, and occasionally crop fields. Such movements would often force individuals to cross roads or take them into residential settings.
Study objectives-
My research efforts were focused at answering one primary question: "If the fragmentation and loss of forested habitat is so detrimental to the timber rattlesnake, a specialized woodland predator, then why does the study population appear to be so viable?" I was interested in determining how the ecology of timber rattlesnakes in fragmented habitat may differ from those in closed-canopy forests. I hypothesized that habitat structure may affect individual movement patterns, foraging strategy, thermal biology, and life history traits. Therefore, I developed the following questions to be investigated using field and laboratory methods.
1) Is there evidence that habitat fragmentation affects the spatial ecology of timber rattlesnakes? Do individuals in the study population differ from those in closed-canopy forests with respect to:
a. home range and core area size?
b. maximum dispersal distances from the hibernaculum?
c. movement patterns of males, females, and gravid females?
2) Is there evidence that habitat fragmentation affects the foraging ecology of timber rattlesnakes?
a. In what types of habitat (woodland, field, or edge) does the study population
forage?
15 b. Do individuals in the study population consume different types of prey than
those in closed-canopy forests?
c. How do prey abundances compare among the different types of foraging
habitat (woodland, field, and edge)?
3) Is there evidence that habitat fragmentation affects the thermal biology of timber
rattlesnakes?
a. Are there significant differences in body temperature between snakes using
woodland, field, or edge habitat?
b. Do snakes that forage in open-canopy habitats such as old fields experience
thermal constraints on activity?
c. Does the anthropogenic habitat (boulder piles within quarry) used by
gestating females allow them to maintain warmer and less variable body
temperatures than males and non gravid females?
4) If active season lengths are comparable between populations, do timber rattlesnakes
from a fragmented habitat in Missouri exhibit growth rates similar to individuals from a
closed-canopy forest in Arkansas?
a. Are early growth rates of snakes in the Missouri population greater than those
of individuals from a closed-canopy forest population located approximately 260
km due south in Arkansas?
Dissertation overview-
Chapter two examines the spatial ecology of the study population. For 9 adult individuals with telemetry data spanning one or more complete active seasons, I report
16 home range and core area size as well as the maximum distance these snakes dispersed
from their hibernacula. In order to investigate general patterns of movement, I plot the
mean dispersal distance for males, females, and reproductive females through each month
of the active season. Lastly, I use previously published studies to compare the home
range size, core area size, maximum dispersal distance, and movement patterns of study
subjects to those in closed-canopy forests (addressing questions la, lb, and lc).
Foraging ecology is the focus of chapter three. Telemetry data is used to
investigate where foraging occurs, thereby addressing question 2a. Dietary data obtained through examination of fecal material and the dissection of roadkilled specimens address
question 2b. Small mammal trapping was conducted to determine relative prey
abundance in woodland, field, and edge habitats (question 2c). Additionally, I discuss
differences in foraging behavior between the study population and timber rattlesnakes in closed-canopy forests.
Chapter four examines the potential thermal consequences of a closed-canopy species using open-canopy habitat. I compare body temperatures (acquired via temperature sensitive telemetry) of snakes located in woodland, field, and edge habitat to answer question 3a. Thermal constraints on activity in both an old field and an adjacent woodland tract were investigated by using copper snake models containing small, dedicated temperature loggers to compare hourly operative temperatures in both habitats throughout midsummer (question 3b). To address question 3c, I compare body temperatures of gravid females gestating within a quarry to those of males and non- gravid females located in summer foraging habitat.
17 Finally, the fifth chapter is a collaborative effort with Dr. Steven J. Beaupre in
which rattle morphology is used to compare the early individual growth rates of snakes
from our respective study populations (question 4a). We establish 1) that the width of the
basal rattle segment correlates with a particular snout-vent length (SVL), and 2) that the
relationship between basal segment width and SVL does not differ between our study
populations. Because a new rattle segment is added at each ecdysis event from the time a
snake is born, snakes with complete (unbroken) rattle strings carry with them a per shed
history of their growth since birth. A repeated measures statistical analysis treated each
segment of an individual's complete rattle string as a repeated measurement of that
individual's SVL through time. This approach allowed us to detect variation in
individual growth rate between populations. A common garden study supports the hypothesis that differences in individual growth can be attributed to population-level
differences in the resource environment.
Acknowledgments
I thank Steven J. Beaupre, Kimberly G. Smith, Ines Pinto, and Edward E. Gbur Jr. for comments on this manuscript. Joseph Agugliaro and George R. Pisani assisted me in obtaining obscure and/or hard to acquire documents on the timber rattlesnake. I am indebted to Sara Wittenberg for her diligent proofreading and words of encouragement while this work was being completed.
Literature cited
18 Adams, J.P. 2005. Home range and behavior of the timber rattlesnake (Crotalus horridus). M.Sc. Thesis, Marshall University, Huntington, West Virginia, USA. 104 p.
Adolph, S.C, and W.P. Porter. 1993. Temperature, activity, and lizard life histories. The American Naturalist 142:273-295.
Agugliaro, J., and H.K. Reinert. 2005. Comparative skin permeability of neonatal and adult timber rattlesnakes {Crotalus horridus). Comparative Biochemistry and Physiology-Part A: Molecular and Integrative Physiology 141:70-75.
Aldridge, R.D., and W.S. Brown. 1995. Male reproductive cycle, age at maturity, and cost of reproduction in the timber rattlesnake {Crotalus horridus). Journal of Herpetology 29:399-407.
Allen, E.R. 1949. Range of cane-brake rattlesnake in Florida. Copeia 1949:73-74.
Allsteadt, J., A.H. Savitzky, C.E. Petersen, and D.N. Naik. 2006. Geographic variation in the morphology of Crotalus horridus (Serpentes: Viperidae). Herpetological Monographs 20:1-63.
Alverson, W.S., D.M. Waller, S.L. Solheim. 1988. Forests too deer: edge effects in northern Wisconsin. Conservation Biology 2:348-358.
Anderson, CD. 2006. Landscape genetics and landscape ecology of a local timber rattlesnake population. Ph.D. Dissertation, Washington University, St. Louis, Missouri, USA. 134 p.
Anderson, CD., and W.J. Drda. 2005. Crotalus horridus (timber rattlesnake). Behavior. Herpetological Review 36:456.
Andren, H. 1992. Corvid density and nest predation in relation to forest fragmentation: a landscape perspective. Ecology 73:794-804.
Andren, H. 1994. Effects of habitat fragmentation on birds and mammals in landscapes with different proportions of suitable habitat: a review. Oikos 71:355-366.
Atwood, W.H. 1955. The timber rattlesnake in Michigan. The Chicago Academy of Sciences, Natural History Miscellanea 139:1-2.
Babcock, H.L. 1929. Food habits of the timber rattlesnake. Bulletin of the Boston Society of Natural History 59:13-14.
Barton, A.J. 1950. Replacement fangs in newborn timber rattlesnakes. Copeia 1950:235-236.
19 Bartz, A.D., and R.A. Sajdak. 2004. Crotalus horridus (timber rattlesnake). Arboreality, courtship. Herpetological Review 35:61.
Beaupre, S.J. 1995. Effects of geographically variable thermal environment on bioenergetics of mottled rock rattlesnakes. Ecology 76:1655-1665.
Beaupre, S.J. 2002. Modeling time-energy allocation in vipers: individual responses to environmental variation and implications for populations. Pages 463-481 in Biology of the Vipers. G.W. Schuett, M. Hoggren, M.E. Douglas, and H.W. Greene. (Eds.). Eagle Mountain Publishing, Eagle Mountain, Utah, USA.
Beaupre, S.J., A.E. Dunham, and K.L. Overall. 1993. The effects of consumption rate and temperature on apparent digestibility coefficient, urate production, metabolizable energy coefficient and passage time in canyon lizards {Sceloporus merriami) from two populations. Functional Ecology 7:273-280.
Beaupre, S.J., and F. Zaidan III. 2001. Scaling of CO2 production in the timber rattlesnake {Crotalus horridus), with comments on cost of growth in neonates and comparative patterns. Physiological and Biochemical Zoology 74:757-768.
Briggler, J.T., and J.W. Prather. 2002. Crotalus horridus (timber rattlesnake). Cave use. Herpetological Review 33:139.
Brown, C.W., and C.H. Ernst. 1986. A study of variation of eastern timber rattlesnakes, Crotalus horridus Linnaeus (Serpentes: Viperidae). Brimleyana 12:57-74.
Brown, W.S. 1982. Overwintering body temperatures of timber rattlesnakes {Crotalus horridus) in northeastern New York. Journal of Herpetology 16:145-150.
Brown, W.S. 1984. Background information for the protection of the timber rattlesnake in New York State. Bulleting of the Chicago Herpetological Society 19:94-97.
Brown, W.S. 1991. Female reproductive ecology in a northern population of the timber rattlesnake, Crotalus horridus. Herpetologica 47:101-115.
Brown, W.S. 1992. Emergence, ingress, and seasonal captures at dens of northern timber rattlesnakes, Crotalus horridus. Pages 251-258 in Biology of the Pitvipers. J.A. Campbell, and E.D. Brodie, Jr. (Eds.). Selva, Tyler, Texas, USA.
Brown, W.S. 1993. Biology, status, and management of the timber rattlesnake {Crotalus horridus): a guide for conservation. SSAR Herpetol Circ No. 22:1-78. Society for the Study of Amphibians and Reptiles. Univ. of Kansas, Lawrence
Brown, W.S. 1995. Heterosexual groups and the mating season in a northern population of timber rattlesnakes, Crotalus horridus. Herpetological Natural History 3:127- 133.
20 Brown, W.S., and D.B. Greenberg. 1992. Vertical-tree ambush posture in Crotalus horridus. Herpetological Review 23:67.
Brown, W.S., L. Jones, and R. Stechert. 1994. A case in herpetological conservation: notorious poacher convicted of illegal trafficking in timber rattlesnakes. Bulletin of the Chicago Herpetological Society 29:74-79.
Brown, W.S., M. Kery, and J.E. Hines. 2007. Survival of timber rattlesnakes {Crotalus horridus) estimated by capture-recapture models in relation to age, sex, color morph, time, and birthplace. Copeia 2007:656-671.
Brown, W.S., and F.M. MacLean. 1983. Conspecific scent-trailing by newborn timber rattlesnakes, Crotalus horridus. Herpetologica 39:430-436.
Brown, W.S., D.W. Pyle, K.R. Greene, and J.B. Friedlaender. 1982. Movements and temperature relationships of timber rattlesnakes {Crotalus horridus) in northeastern New York. Journal of Herpetology 16:151-161.
Browning, D.M., S.J. Beaupre, and L. Duncan. 2005. Using partitioned mahalanobis D2{k) to formulate a GIS-based model of timber rattlesnake hibernacula. Journal of Wildlife Management 69:33-44.
Burger, J.W. 1934. The hibernation habits of the rattlesnake of the New Jersey Pine Barrens. Copeia 1934:142.
Burke, D.M., and E. Nol. 1998. Influence of food abundance, nest-site habitat, and forest fragmentation on breeding Ovenbirds. The Auk 115:96-104.
Burnett, CD. 2001. Comparative diets of syntopic pit vipers: co-existence through trophic niche separation. Unpublished honors thesis, The University of Arkansas, Fayetteville, Arkansas, USA.
Bury, R.B. 2006. Natural history, field ecology, conservation biology and wildlife management: time to connect the dots. Herpetological Conservation and Biology 1:56-61.
Bushar, L.M., M. Maliga, and H.K. Reinert. 2001. Cross-species amplification of Crotalus horridus microsatellites and their application in phylogenetic analysis. Journal of Herpetology 35:532-537.
Bushar, L.M., H.K. Reinert, and L. Gelbert. 1998. Genetic variation and gene flow within and between local populations of the timber rattlesnake, Crotalus horridus. Copeia 1998:411-422.
21 Cale, W.G., Jr., and J.W. Gibbons. 1972. Relationships between body size, size of the fat bodies, and total lipid content in the canebrake rattlesnake {Crotalus horridus) and the black racer {Coluber constrictor). Herpetologica 28:51-53.
Cantrell, R.S., C. Cosner, and W.F. Fagan. 2001. How predator incursions affect critical patch size: the role of the functional response. The American Naturalist 158:368- 375.
Carfagno, G.L.F., E.J. Heske, and P.J. Weatherhead. 2006. Does mammalian prey abundance explain forest-edge use by snakes? Ecoscience 13:293-297.
Carfagno, G.L.F., and P.J. Weatherhead. 2006. Intraspecific and interspecific variation in use of forest-edge habitat by snakes. Canadian Journal of Zoology 84:1440- 1452.
Carle, H.D. 1948. Rattle snakes on Wantastiquet Mountain. Proceedings of the New Hampshire Academy of Science 1:23-24.
Carmichael, E.B., and P.W. Petcher. 1945. Constituents of the blood of the hibernating and normal rattlesnake, Crotalus horridus. Journal of Biological Chemistry 161:693-696.
Cavanaugh, C.J. 1994. Crotalus horridus (timber rattlesnake). Longevity. Herpetological Review 25:70.
Chiszar, D., H.M. Smith, CM. Bogert, and J. Vidaurri. 1991. A chemical sense of self in timber and prairie rattlesnakes. Bulletin of the Psychonomic Society 29:153- 154.
Christoffel, R.A. 2007. Using human dimensions insights to improve conservation efforts for the eastern massasauga rattlesnake (Sistrurus catenatus catenatus) in Michigan and the timber rattlesnake {Crotalus horridus horridus) in Minnesota. Ph.D. Dissertation, Michigan State University, East Lansing, Michigan, USA. 286 p.
Clark, A.M., P.E. Moler, E.E. Possardt, A.H. Savitzky, W.S. Brown, and B.W. Bowen. 2003. Phylogeography of the timber rattlesnake {Crotalus horridus) based on mtDNA sequences. Journal of Herpetology 37:145-154.
Clark, R.W. 2002. Diet of the timber rattlesnake, Crotalus horridus. Journal of Herpetology 36:494-499.
Clark, R.W. 2004a. Feeding experience modifies the assessment of ambush sites by the timber rattlesnake, a sit-and-wait predator. Ethology 110:471-483.
22 Clark, R.W. 2004b. Kin recognition in rattlesnakes. Proceedings of the Royal Society of London Series B Biology Letters 27LS243-S245.
Clark, R.W. 2004c. Timber rattlesnakes {Crotalus horridus) use chemical cues to select ambush sites. Journal of Chemical Ecology 30:607-617.
Clark, R.W. 2005. Pursuit-deterrent communication between prey animals and timber rattlesnakes {Crotalus horridus): the response of snakes to harassment displays. Behavioral Ecology and Sociobiology 59:258-261.
Clark, R.W. 2006a. Post-strike behavior of timber rattlesnakes {Crotalus horridus) during natural predation events. Ethology 112:1089-1094.
Clark, R.W. 2006b. Fixed videography to study predation behavior of an ambush foraging snake, Crotalus horridus. Copeia 2006:181-187.
Clark, R.W. 2007. Public information for solitary foragers: timber rattlesnakes use conspecific chemical cues to select ambush sites. Behavioral Ecology 18:487- 490.
Clark, R.W., W.S. Brown, R. Stechert, and K.R. Zamudio. 2008. Integrating individual behaviour and landscape genetics: the population structure of timber rattlesnake hibernacula. Molecular Ecology 17:719-730.
Cobb, V.A., J.J. Green, T. Worrall, J. Pruett, and B. Glorioso. 2005. Initial den location behavior in a litter of neonate Crotalus horridus (timber rattlesnakes). Southeastern Naturalist 4:723-730.
Collins, J.T., and J.L. Knight. 1980. Crotalus horridus. Catalogue of American Amphibians and Reptiles 253:1-2.
Conner, R.N., D.C. Rudolf, D. Saenz, R.R. Schaefer, and S.J. Burgdorf. 2003. Growth rates and post-release survival of captive neonate timber rattlesnakes, Crotalus horridus. Herpetological Review 34:314-317.
Cooper, J.E., and F. Groves. 1959. The rattlesnake, Crotalus horridus, in the Maryland Piedmont. Herpetologica 15:33-34.
Coupe, B.H. 1997. Factors affecting movement of radio-tracked timber rattlesnakes {Crotalus horridus) in southern Ohio. M.Sc. Thesis. Ohio State University, Columbus, Ohio, USA. 52 p.
Coupe, B. 2001. Arboreal behavior in timber rattlesnakes {Crotalus horridus). Herpetological Review 32:83-85.
23 Coupe, B. 2002. Pheromones, search patterns, and old haunts: How do male timber rattlesnakes (Crotalus horridus) locate mates? Pages 139-148 in Biology of the Vipers. G.W. Schuett, M. Hoggren, M.E. Douglas, and H.W. Greene. (Eds.). Eagle Mountain Publishing, Eagle Mountain, Utah, USA.
Cundall, D., and S.J. Beaupre. 2001. Field records of predatory strike kinematics in timber rattlesnakes, Crotalus horridus. Amphibia-Reptilia 22:492-498.
Diaz, J.A., R. Carbonell, E. Virgos, T. Santos, and J.L. Telleria. 2000. Effects of forest fragmentation on the distribution of the lizard Psammodromus algirus. Animal Conservation 3:235-240.
Diemer Berish, J.E. 1998. Characterization of rattlesnake harvest in Florida. Journal of Herpetology 32:551-557.
Dundee, H.A. 1994a. Crotalus horridus (timber rattlesnake). Coloration. Herpetological Review 25:28.
Dundee, H.A. 1994b. Crotalus horridus (timber rattlesnake). Herpetological Review 25:33-34.
Dunham, A.E., B.W. Grant, and K.L. Overall. 1989. Interfaces between biophysical and physiological ecology and the population ecology of terrestrial vertebrate ectotherms. Physiological Zoology 62:335-355.
Edgren, R.A., Jr. 1948. Notes on a litter of young timber rattlesnakes. Copeia 1948:132.
Fahrig, L., and G. Merriam. 1985. Habitat patch connectivity and population survival. Ecology 66:1762-1768.
Fahrig, L., and G. Merriam. 1994. Conservation of fragmented populations. Conservation Biology 8:50-59.
Fitch, H.S. 1985. Observations on rattle size and demography of prairie rattlesnakes {Crotalus viridis) and timber rattlesnakes (Crotalus horridus) in Kansas. Occasional Papers of the Museum of Natural History, The University of Kansas, Lawrence, Kansas 118:1-11.
Fitch, H.S. 1999. A Kansas snake community: composition and changes over 50 years. Krieger Publishing Company, Malabar, Florida, USA.
Fitch, H.S., and G.R. Pisani. 2002. Longtime recapture of a timber rattlesnake (Crotalus horridus) in Kansas. Journal of Kansas Herpetology 3:15-16.
Fitch, H.S., and G.R. Pisani. 2005. Disappearance of radio-monitored timber rattlesnakes. Journal of Kansas Herpetology 14:14-15.
24 Fitch, H.S., and G.R. Pisani. 2006. The timber rattlesnake in northeastern Kansas. Journal of Kansas Herpetology 19:11-15.
Fitch, H.S., G.R. Pisani, H.W. Greene, A.F. Echelle, and M. Zerwekh. 2004. A field study of the timber rattlesnake in Leavenworth County, Kansas. Journal of Kansas Herpetology 11:18-24.
Fitch, H.S., and H.W. Shirer. 1971. A radiotelemetric study of spatial relationships in some common snakes. Copeia 1971:118-128.
Fogell, D.D. 2000. Seasonal activity, habitat preferences, and natural history of the timber rattlesnake (Crotalus horridus) in southeastern Nebraska. M.Sc. Thesis, University of Nebraska at Omaha, Omaha, Nebraska, USA. 91 p.
Fogell, D.D., T.J. Leonard, and J.D. Fawcett. 2002a. Crotalus horridus horridus (timber rattlesnake). Habitat. Herpetological Review 33:211-212.
Fogell, D.D., T.J. Leonard, and J.D. Fawcett. 2002b. Crotalus horridus horridus (timber rattlesnake). Climbing. Herpetological Review 33:212.
Fox, J., and B. Hamilton. 2007. Crotalus horridus (timber rattlesnake). Behavior. Herpetological Review 38:86.
Franklin, A.B, B.R. Noon, and T.L. George. 2002. What is habitat fragmentation? Studies in Avian Biology 25:20-29.
Furman, J. 2007. Timber Rattlesnakes in Vermont and New York: biology, history, and the fate of an endangered species. University Press of New England, Lebanon, New Hampshire, USA.
Galligan J.H., and W.A. Dunson. 1979. Biology and status of timber rattlesnake {Crotalus horridus) populations in Pennsylvania. Biological Conservation 15:13- 58.
Gibbons, J.W. 1972. Reproduction, growth, and sexual dimorphism in the canebrake rattlesnake (Crotalus horridus atricaudatus). Copeia 1972:222-226.
Gibbons, J.W., D. E. Scott, T.J. Ryan, K.A. Buhlmann, T.D. Tuberville, B.S. Metts, J.L. Greene, T. Mills, Y. Leiden, S. Poppy, and C.T. Winne. 2000. The global decline of reptiles, deja vu amphibians. BioScience 50:653-666.
Gibson, S.E. 2003. Movements and habitat preference of the timber rattlesnake (Crotalus horridus) in southern Indiana. M. Sc. Thesis, Purdue University, Fort Wayne, Indiana, USA. 72 p.
Gloyd, H.K. 1935. The cane-brake rattlesnake. Copeia 1935:175-178.
25 Grant, B.W., and A.E. Dunham. 1988. Thermally imposed time constraints on the activity of the desert lizard Sceloporus merriami. Ecology 69:167-176.
Greene, H.W. 1997. Snakes: the Evolution of Mystery in Nature. University of California Press, Berkeley, California, USA.
Heard, G.W., D. Black, and P. Robertson. 2004. Habitat use by the inland carpet python (Morelia spilota metcalfei: Pythonidae): seasonal relationships with habitat structure and prey distribution in a rural landscape. Austral Ecology 29:446-460.
Huey, R.B. 1991. Physiological consequences of habitat selection. American Naturalist 137:S91-S115.
Huey, R.B. 1982. Temperature, physiology, and the ecology of reptiles. Pages 25-91 in Biology of the Reptilia. C. Gans, and F.H. Pough (Eds.). Academic Press, New York, New York, USA.
Jensen, J.B., B.W. Mansell, and P.E. Moler. 1994. Crotalus horridus (timber rattlesnake). Herpetological Review 25:166.
Keenlyne, K.D. 1972. Sexual differences in feeding habits of Crtalus horridus horridus. Journal of Herpetology 6:234-237.
Keenlyne, K.D. 1978. Reproductive cycles in two species of rattlesnakes. The American Midland Naturalist 100:368-375.
Klenzendorf, S.A., D.J. Lee, M.R. Vaughan, and R.B. Duncan, Jr. 2004. Crotalus horridus (timber rattlesnake). Defense and black bear death. Herpetological Review 35:61.
Lardie, R.L. 1999. Crotalus horridus (timber rattlesnake). Herpetological Review 30:112.
Lasher, D.N. 1980. A bicephalic Crotalus horridus from Alabama. Herpetological Review 11:89.
Mac Nally, R., and G.W. Brown. 2001. Reptiles and habitat fragmentation in the box- ironbark forests of central Victoria, Australia: predictions, compositional change and faunal nestedness. Oecologia 128:116-125.
Martin, W.H. 1979. The timber rattlesnake in Virginia: its distribution and present status. Virginia Herpetological Society Bulletin 89:1-4.
Martin, W.H. 1982. The timber rattlesnake in the northeast; its range, past and present. HERP Bulletin of the New York Herpetological Society 17:15-20.
26 Martin, W.H. 1988. Life history of the timber rattlesnake. Catesbeiana 8:9-12.
Martin, W.H. 1992a. The timber rattlesnake: its distribution and natural history. Pages 13-22 in Conservation of the timber rattlesnake in the northeast. T.F. Tyning (Ed.). Massachusetts Audubon Society, Lincoln, MA, USA.
Martin, W.H. 1992b. Phenology of the timber rattlesnake (Crotalus horridus) in an unglaciated section of the Appalachian Mountains. Pages 259-277 in Biology of the Pitvipers. J.A. Campbell, and E.D. Brodie, Jr. (Eds.). Selva, Tyler, Texas, USA.
Martin, W.H. 1993. Reproduction of the timber rattlesnake (Crotalus horridus) in the Appalachian Mountains. Journal of Herpetology 27:133-143.
Martin, W.H. 1996. Crotalus horridus (timber rattlesnake). Reproductive phenology. Herpetological Review 27:144-145.
Martin, W.H. 2002. Life history constraints on the timber rattlesnake (Crotalus horridus) at its climatic limits. Pages 285-306 in Biology of the Vipers. G.W. Schuett, M. Hoggren, M.E. Douglas, and H.W. Greene. Eagle Mountain Publishing, Eagle Mountain, Utah, USA.
McAllister, C.T., S.R. Goldberg, H.J. Holshuh, and S.E. Trauth. 1993. Disseminated mycotic dermatitis in a wild-caught timber rattlesnake, Crotalus horridus (Serpentes: Viperidae), from Arkansas. Texas Journal of Science 45:279-281.
McGowan, E.M. 2003. Reproductive strategies of the timber rattlesnake, Crotalus horridus. Ph.D. Dissertation, State University of New York at Binghamton, Binghamton, New York, USA. 241 p.
Merrow, J.S., and T. Aubertin. 2005. Crotalus horridus (timber rattlesnake). Reproduction. Herpetological Review 36:192.
Minton, S.A., Jr. 1969. The feeding strike of the timber rattlesnake. Journal of Herpetology 3:121-124.
Monteiro, C.S. 2006. Effect of food availability on energy allocation between growth and reproduction in timber rattlesnakes, Crotalus horridus (Serpentes, Viperidae). M.Sc. Thesis, The University of Arkansas, Fayetteville, Arkansas, USA. 41 p.
Montgomery, C.E. 2005. Bioenergetic and ecological correlates of foraging mode in copperheads, Agkistrodon contortrix, and timber rattlesnakes, Crotalus horridus, in northwest Arkansas. Ph.D. Dissertation, The University of Arkansas, Fayetteville, Arkansas, USA. 208 p.
27 Newmark, W.D. 2001. Tanzanian forest edge microclimatic gradients: dynamic patterns. Biotropica 33:2-11.
Parmley, D., and A.M. Parmley. 2001. Food habits of the canebrake rattlesnake {Crotalus horridus atricaudatus) in central Georgia. Georgia Journal of Science 59:172-178.
Petri, P. 2002. Timber rattlesnakes and changing land use: a management plan for the Catskill Park, New York. M.Sc. Thesis, Bard College, Annandale-On-Hudson, New York, USA. 77 p.
Pisani, G.R., J.T. Collins, and S.R. Edwards. 1973. A re-evaluation of the subspecies of Crotalus horridus. Transactions of the Kansas Academy of Science 75:255-263.
Pisani, G.R., and H.S. Fitch. 2006. Rapid early growth in northeastern Kansas timber rattlesnakes. Journal of Kansas Herpetology 20:19-20.
Piatt, S.G., A.W. Hawkes, and T.R. Rainwater. 2001. Diet of the canebrake rattlesnake {Crotalus horridus atricaudatus): an additional record and review. Texas Journal of Science 53:115-120.
Porter, W.P., and D.M. Gates. 1969. Thermodynamic equilibria of animals with environment. Ecological Monographs 39:227-244.
Pringle, R.M., J.K. Webb, and R. Shine. 2003. Canopy structure, microclimate, and habitat selection by a nocturnal snake, Hoplocephalus bungaroides. Ecology 84:2668-2679.
Quinn, J.F., and A. Hastings. 1988. Extinction in subdivided habitats: reply to Gilpin. Conservation Biology 2:293-296.
Reinert, H.K. 1984a. Habitat separation between sympatric snake populations. Ecology 65:478-486.
Reinert, H.K. 1984b. Habitat variation within sympatric snake populations. Ecology 65:1673-1682.
Reinert, H.K. 1990. A profile and impact assessment of organized rattlesnake hunts in Pennsylvania. Journal of the Pennsylvania Academy of Science 64:136-144.
Reinert, H.K., D. Cundall, and L.M. Bushar. 1984. Foraging behavior of the timber rattlesnake, Crotalus horridus. Copeia 1984:976-981.
Reinert, H.K., and R.R. Rupert Jr. 1999. Impacts of translocation on behavior and survival of timber rattlesnakes, Crotalus horridus. Journal of Herpetology 33:45- 61.
28 Reinert, H.K., and R.T. Zappalorti. 1988a. Timber rattlesnakes {Crotalus horridus) of the pine barrens: their movement patterns and habitat preference. Copeia 1988:964-978.
Reinert, H.K., and R.T. Zappalorti. 1988b. Field observation of the association of adult and neonatal timber rattlesnakes, Crotalus horridus, with possible evidence for conspecific trailing. Copeia 1988:1057-1059.
Richards, R.L. 1990. Quaternary distribution of the timber rattlesnake {Crotalus horridus) in southern Indiana. Proceedings of the Indiana Academy of Science 99:113-122.
Robinson, T.S. 1966. Effect of canopy density and slope exposure on the subcanopy microenvironment of a northern hardwood forest. The American Midland Naturalist 75:339-346.
Rudolph, D.C., and S.J. Burgdorf. 1997. Timber rattlesnakes and Louisiana pine snakes of the West Gulf Coastal Plain: hypotheses of decline. Texas Journal of Science 49:111-122.
Rudolph, D.C., S.J. Burgdorf, R.N. Conner, and J.G. Dickson. 1998. The impact of roads on the timber rattlesnake, {Crotalus horridus), in eastern Texas. Pages 236- 240 in G. L. Evink, P. Garrett, D. Zeigler and J. Berry, editors. Proceedings of the International Conference on Wildlife Ecology and Transportation. FL-Er-69- 98, Florida Department of Transportation, Tallahassee, Florida.
Rudolph, D.C., R.R. Schaefer, D. Saenz, and R.N. Conner. 2004. Arboreal behavior in the timber rattlesnake, Crotalus horridus, in eastern Texas. Texas Journal of Science 56:395-404.
Sadighi, K., R.M. DeGraff, and W.R. Danielson. 1995. Experimental use of remotely- triggered cameras to monitor occurrence of timber rattlesnakes {Crotalus horridus). Herpetological Review 26:189-190.
Saenz, D., S.J. Burgdorf, D.C. Rudolph, and CM. Duran. 1996. Crotalus horridus (timber rattlesnake). Climbing. Herpetological Review 27:145.
Sajdak, R.A., and A.W. Bartz. 2004. Crotalus horridus (timber rattlesnake). Arboreality, diet. Herpetological Review 35:60.
Sasaki, K., and D. Duvall. 2006. Field notes: negative impacts of herping on timber rattlesnakes {Crotalus horridus). Bulletin of the Herpetological Society of Japan 2006:45-47.
Saunders, D.A., R.J. Hobbs, and C.R. Margules. 1991. Biological consequences of ecosystem fragmentation: a review. Conservation Biology 5:18-32.
29 Savage, T. 1967. The diet of rattlesnakes and copperheads in the Great Smoky Mountains National Park. Copeia 1967:226-227.
Schaefer, G.C. 1969. Sex independent ground color in the timber rattlesnake, Crotalus horridus horridus. Herpetologica 25:65-66.
Schlaepfer, M.A., and T.A. Gavin. 2001. Edge effects on lizards and frogs in tropical forest fragments. Conservation Biology 15:1079-1090.
Schwartz, C.W., and E.R. Schwartz. 1959. The Wild Mammals of Missouri. University of Missouri Press, Columbia, Missouri, USA.
Sealy, J.B. 1996. Crotalus horridus (timber rattlesnake). Mating. Herpetological Review 27:23-24.
Sealy, J.B. 1997. Short-distance translocations of timber rattlesnakes in a North Carolina state park: a successful conservation and management program. Sonoran Herpetologist 10:94-99.
Sealy, J.B. 2002. Ecology and behavior of the timber rattlesnake {Crotalus horridus) in the upper Piedmont of North Carolina: identified threats and conservation recommendations. Pages 561-578 in Biology of the Vipers. G.W. Schuett, M. Hoggren, M.E. Douglas, and H.W. Greene. (Eds.). Eagle Mountain Publishing, Eagle Mountain, Utah, USA.
Sims, P.L. 1988. Grasslands. Pages 265-286 in North American Terrestrial Vegetation. M.G. Barbour and W.D. Billings. (Eds.). Cambridge University Press, New York, New York, USA.
Sinclair, T.A., and B.L. Tipton. 2005. Crotalus horridus (timber rattlesnake). Herpetological Review 36:81.
Smyth, T. 1949. Notes on the timber rattlesnake at Mountain Lake, Virginia. Copeia 1949:78.
Stechert, R. 1980. Observations on northern snake dens. HERP Bulletin of the New York Herpetological Society 15:7-14.
Stechert, R. 1982. Historical depletion of timber rattlesnake colonies in New York state. HERP Bulletin of the New York Herpetological Society 17:23-24.
Steen, D.A., L.L. Smith, L.M. Conner, J.C. Brock, and S.K. Hoss. 2007. Habitat use of sympatric rattlesnake species within the Gulf Coastal Plain. Journal of Wildlife Management 71:759-764. Stewart, M.M., G.E. Larson, and T.H. Matthews. 1960. Morphological variation in a litter of timber rattlesnakes. Copeia 1960:366-367.
30 Storment, D. 1990. Field observations of sexual dimorphism in head pattern (markings) in timber rattlesnakes (Crotalus horridus). Bulletin of the Chicago Herpetological Society 25:160-162.
Stroupe, D.A., and M.E. Dorcas. 2001. The apparent persistence of Crotalus horridus in the western Piedmont of North Carolina. Herpetological Review 32:287-288.
Sumner, J. 2005. Decreased relatedness between male prickly forest skinks (Gnypetoscincus queenslandiae) in habitat fragments. Conservation Genetics 6:333-340.
Sutherland, I.D.W. 1958. The "combat dance" of the timber rattlesnake. Herpetologica 14:23-24.
Templeton, A.R., R.J. Robertson, J. Brisson, and J. Strasburg. 2001. Disrupting evolutionary processes: the effect of habitat fragmentation on collared lizards in the Missouri Ozarks. Proceedings of the National Academy of Sciences 98:5426- 5432.
Trapido, H. 1939. Parturition in the timber rattlesnake, Crotalus horridus horridus Linne. Copeia 1939:230.
Ujvari, B., T. Madsen, T. Kotenko, M. Olsson, R. Shine, and H. Wittzell. 2002. Low genetic diversity threatens imminent extinction for the Hungarian meadow viper (Vipera ursinii rakosiensis). Biological Conservation 105:127-130.
Villarreal, X., J. Bricker, H.K. Reinert, L. Gelbert, and L.M. Bushar. 1996. Isolation and characterization of microsatellite loci for use in population genetic analysis in the timber rattlesnake, Crotalus horridus. The Journal of Heredity 87:152-155.
Waldron, J.L., J.D. Lanham, and S.H. Bennett. 2006. Using behaviorally-based seasons to investigate canebrake rattlesnake {Crotalus horridus) movement patterns and habitat selection. Herpetologica 62:389-398.
Walker, M.L., J.A. Dorr, and G.R. Pisani. 2008. Observation of aberrant growth in a Timber rattlesnake {Crotalus horridus). Transactions of the Kansas Academy of Science 111:156-158.
Walker, Z.J. 2000. The spatial ecology of the timber rattlesnake {Crotalus horridus) in south central Indiana. M.Sc. Thesis, Purdue University, Fort Wayne, Indiana, USA. 45 p.
Walley, H.D. 1963. The rattlesnake, Crotalus horridus horridus, in north-central Illinois. Herpetologica 19:216.
31 Watson, D.M. 2002. A conceptual framework for studying species composition in fragments, islands and other patchy ecosystems. Journal of Biogeography 29:823-834.
Watson, D.M. 2003. Long-term consequences of habitat fragmentation-highland birds in Oaxaca, Mexico. Biological Conservation 111:283-303.
Wills, C.A., and S.J. Beaupre. 2000. An application of randomization for detecting evidence of thermoregulation in timber rattlesnakes (Crotalus horridus) from northwest Arkansas. Physiological and Biochemical Zoology 73:325-334.
Zaidan III, F. 1996. Behavioral thermoregulation in Crotalus horridus and Agkistrodon piscivorus: the importance of radiated and conducted heat sources. M.Sc. Thesis, The University of Mississippi, Oxford, Mississippi, USA. 62 p.
Zaidan III, F., and S.J. Beaupre. 2003. Effects of body mass, meal size, fast length, and temperature on specific dynamic action in the timber rattlesnake (Crotalus horridus). Physiological and Biochemical Zoology 76:447-458.
Zanette, L., P. Doyle, and S.M. Tremont. 2000. Food shortage in small fragments: evidence from an area-sensitive passerine. Ecology 81:1654-1666.
32 ID OS
CO o ON ON NO in t- OS o m ON u ON 1/N ON "cd SO ND NO ON "O ON ON ON ON ON I I ON ON —< § t> w ^^ 73 •J d '—' O t c •a to c cd >> 60 _u O u o o J3 o i "L^ *-*60 £ * 3 as *a Ste w Sav a Coope r m CQ 5 u O W CD Wal l Min t
i-i >N u § 60 •J3 co 2JO S ; "3 O 2 c -C T3 cd g fe m c hol o erna i *-• 60 £ •a c -a "cd o =4- o. c & u rS leonat e .S X o o. cd o 1 o C E _c o o tes t § g T3 'P .S •a _c C cd
"3 collec t ^ cd
CO tte r o f n e Bar r / afte r h ulu m ."ti XI u £
.spu n "c S 5 2 60 — a CJ o "« '5 .c u esc r
1 ncapa c snake s co n du r ;urren i C •^
icai ibernac i •a nt h M sin g <£ CO u • ^- u
60 ie s " x c CD ••? X5 ^i
T3 c u c "5) on s 5 60 o ^c 3 Z. _o a \cd c C "3 co 'co > •2 O X) Tl o XI o 3 o cd -a •c^d co 4-1 cd cd aria t ho w ber n bse r u o co m x: o 1 var i Id iso n data , ecor d -a -a ofrep 1 O § ca l
o f h i o o f partur i istr y ca l v
a l d v:
"cUO l d ak e s ^ cd 13 o. c c c c p I" c cd c a '5b g sO
o og i o io n CO o o io n o £ o +-i 1 1 CJ "3 -•-?* 3 o 3 o .s- 60 X> J3 X) b & > cd o •S cd J= script i signat i rura l h •ra y o f ao d c h arpho l stribut i str i CO or p
bse i id to u CD Cd X •o -a 3 C E T3 o T3 E -a -a
to
« c o o m (oN 73 00 (N § o x' CO cd &H 6u0 O ~53
,o en s c cd c« E p. 'E C ar r cd cd o _o X3 m 73 p. Id Id i- u w t« u Ido l o c 60 S 73 z =a for e servoi r c ode c ain s c hru b cd -a w a> o o cd WOO ' au , r Park , c SU 1 ar d 15 Id 'g y Mou n J3 o S 55 73 o U 3 O 'c §• O O >^ tyiE 3 o cd 3: E 60 idon e 73 Id u CJ cu csd 73 (S 'o
•J x i-i < z; P* S 2: < z 0- > OH
33 opsurt i ^^ x 60 !-. <""- XH X ON iS u CO ^ ON ~ c -a oo —• ON ON ON •-H ON C8 ca •o ^j ^*-^ c u ^—' cd C s a er t c C O X X £ o ? IS
X O s art i u = e =o o o o U do pa PQ U 5 ^ a s
60 _o x >N O •c % X 60 o _o c & X o "o -a .2 X u •S o £ x ^ o 5 •I x E c "E E
VI c a & o o no l V- O j^ a a 'Si 3 IS •a o o o O o IS J4 s, 5 & oQ an d o T3 , , of •m S3 60 =9 S D. O X i- _o
fiel & "OT ge o
cd od e
_o CO 1 o ixe d o
"• es c 'Eo C 3 8" S g _o CO c« X> > X "g u « .SP 3 >CO> a GJ t> 60 •a u {-• 'C c« C o '3 gra s es , h '3 o •o O u o ter n 3 3 o c ttH c« O o o o es co •a u 3 ur e & c X -a * T3 u^ J3 t 5&? 5 E "E. o u X 3 XI c c« CO O u < > GO > co z 34 t o o 15 ^-« OS ft ca OS & ^ Zap p N 990 ) W 990 ) o ^^ ON OS -a OS C- w OS la OS § Cfl an d cBa o> -o B 1 B t •K e U c ca u -—E ' u u ca N g g S o B E JB » CW '5 o O S c o o 1 '5 '53 '53 ^ 15 ft a1 CQ 'S oi Bi 55 CQ U ft1 s x> S b •a c3 c J3 E ca 1> <*- O o m *C t/1 E .£> ta £ P. 60 t B "C ft u o c o S o UD E E o >> O u >-, O o -o <*> x> a L* .B ca <" g ft -a u M ca -3 J3 ^E u 3 g o '5 s u G O oB u ca -^ ecifi ' '3 tern s ds u CJ ~G r} ^ ca ca •- o o "rrt ^ o 111 ft o se n o -ti E o u u J2 -£> & B o tw ca ft U h ft^ o to 3 "O ft t« ca u ? -E O B ,u E £ ft-E B t+H u 60 60 o c 5? u ••c c -a u cfl •; « ft a o o "3.2 § 'So ^^ 60 ita t pr e 3 a a. f mov e Delation s :e d hu n ^ £ .-s o g ca a> O -O c« o ca -.5 ° 60 rs o >* C3 -T-J >> a g bo o & tH T3 B -o -B c o a) O O 5 o ta "o > 5 • _ "o S 3 E cfl stud y ibeh a lat e a s organ i 60 O w iC u 3 v> rs -S -E ca u OT >> b a cl e ^ ~ £ H o -= « « to n J= © 3 E- ^5 3 u ft a M c« s(U ^.'o 13 .2 de s O -a t/n e g £ _S net r aj _>. Cl. o 60 o ^£ 'cfol u u •g o ft i, ft o3 pac t o f 2 g O s s 3 ca E H c c/3 a> E 3 •u iS -a g E ,1- O a .s I -§1 -a o o is -a M O * •g T3 -o M o O (D O o O u CJ X t/} ^ca dwo o T3 cta: cfca ca CD *d CL) m sn o 60 «J u -o B c 3 d> d/o r O g ft ft E 3 3 3 J* a E ca 3, -a ita t de c •o •s -B J3 -E 1 O o "3 O U ;- ;0-0 3 U 3 3 O O tj -a -a •a 3 3 E B c 2 o o o -a o u > < in >- < < ft < ft OH > 2 ft 35 ON ON 00 m i~ o> u o> c X) Tt c —. o CD o ^—-' ON CD , , 15 i~ cd X^> - m CN (N CO m ON o ON ON ON ON i) •o ON ON ON ON i- _cd' .— XH —=8 =cd c CD 03 1 *—• c« - u a 00 C '—c ' >—C ' —C ' C C « -a 5 rt i t. rt i £ o » < > o cd a o cd O o 3 P 2 cd 03 m s S m s s m O CQ 3 -o o I o o ucti > CO 3 -8.2 -o o IE o 11 o a o QJ cd w S-g V- s- l- 2 CD 3 2 -o zn O cd c •o c cd ft T3 .P C "cd en cd CO c O co spe r P mb u e, P cd ~3 •o _u .o O XI "3) Id EH .S "3 Cn OH cd O C4H & O QJ S 3 3 "S3 > H-. J^ c CO £ f P cd 3 P U _ o « 2 cd U U O CO C 'pi u Ml en +-» cod u o > c £ cd cd -C •<"C= u CO o *+^ 3 o OH 3 U 00 o JO -a o •c — u •c o cd O CD P en OH l U "Td e T3 •a -a -a O o o o O o o o s -o M O -a f -a o u o. S5 CD CD oo X o 'c^ XI >> t y pr e eRiv i J= y mi x y mi x O. .c 8 CO u cd O Scd ivers i 1/5 -p 13) M) CO cfl~ <£ PH o. c _3 3 1) o HD S3 3 3 I en V} CL) "cd -2^ en nea t E •« •3 CD M M Ia n w is 3 cl> CL) «3 M g P ri d ri d ire s ire s H B S CH-H •o o o eg cd 4H ^ cd cd <+s c/J CO o O O p P 3 3 T3 cd H- 3 p O O •a T3 CL> c 3 3 u U tj o 3 p -10 3 a •a •a -a "Og -a 2 ^H U cd ec i cd cd cd cd o o u -o -a C wo o P wo o P P CJ XI T3 "O 13 > 1 < | ^ ^ < •J >- > •z, Z 1 > 1 •5 5! •J UH Z 2 >• 36 OS OS OS OS <+- OS SO i- OS 0 T3 OS OS os OS t! so bo OS .& er t D. "0 T3 3 T3 •§ -V3H c '5 O 3 3 E a CD N O C* CQ Pi •J Di u .8 1 •5 60 ^ PH C o C -a 1-2 II if c - 60 3 O E 60 O (_> 1) O 5 o .2 o 3 U ^ O 3 c £ ~o £ a o 2 a -p # OS « S a cu c u -e E tu ^ f3 « 60 a u s c II o •gi -O •£. g2 o •-„ E=° fee" u S a1 U 'M C" ^0 T^ 3 o .2 £> N O fa o- c a "SI S o II o o tu 3 -* •S-e a 8 .5 '% •§>" o a o ° .&! •2 o a — "O C • c 5 "' u 5 tn 3 ° •S ° £ o S 3 a J o ^ -S -S? en .3 „- a '§ 'E ™ ft « e .-s o u 2 £> CU 2 5 '-3 „ c I a C .0 i§ 3 2 •£ h C tu I c a bo "S S "S « C a c V. o o -^ CL _^- o o 11 •— CU h o c o -c E - 60 •c 650 tnS ga« E .«* tu r^ -0 0 • 3 a cu a e J -0 c > HG 0 H O en J E^ § 'opu l 1 a a o •5 -a x> '_ 0. e3n a I&0 P X E "3 60 o o LuH o dwoo d s d can o -a X! H I* 1 a D. a _o -*-> 03 X! H- -aJ3 0 ,3 •a -o bo s c o T3 tn -a lu s O 0 a en < > 5 J ^£ X O CO < GO H O PH E O :z s M" s S OH 37 o o CN - £ o o ^ o 20 0 o •"-H—' (20 C cs c t« CJ tS —ca S^ -a o o .g o. o u ^ '3 CN u ca O O "S § N ^ , v PH o O P^H -o o ca CN ,^—^ ffl c "3 T3 (N O C3 CoN C O o -a t3 ~3 ca • an d O (200 1 1 an d *—' L- c y an d (N V: fa U ca u u P. M U — u ft -a "p 3 60 4^ c 3 c O 60 3 M 15 3 3 O 13 •c la r O 3 i CD CO m u U PH p- &0 CQ U u (U u fc- (&3 1c3 4= a> 5 o 45 pe r o o c > -o g c co p -o "S o f+- ifc if o o o -a plifi e E owt h : s an d •£ XS o> u fg r lak e ylogen y .g o o> en c (L) M 43 o 43 3 o > •o T3 1) ar e ca CJ U 00 ts u o ^j la X .3 a ca E E 13 ri n 8 13 & C3 > _U3 uc« o ca u o .E C C c a ca 13 "r/i ca u > c ,«3 e comp a behav i 1 Crot a o n an d ° trS imbe r Dsatell i 60 £ — ca •8 parti a oduct i o f die t 43 ctiv e seaso : . stud y a n of t e mie n borea l i l an d cott o ca intogeneti c i f male/fer n these s o f m dul t mal e i u O S ca o o ca (D T3 & t»-c || o £ -p Ida t > £ o o o C & § o O t- -a > ca w ca •3s 4 3 lesn ; ulat i u nparis i pee r c B o O O o o WD c 1 1 c" 60 <4H 5z o &£ O S - S S t I 60 c Crotalus thermor e observat i literatur e distributi i interacti o scalin g o dietar y co i timbe r a radio-tag ; observat i geograp h telemetr y recaptur e ID OH •n -a Q? o '5 u u O & G I 3 •" 60 U -a c -a -o 4 $ o o ® .S3 ca O o o to S — % -a •p S3 43 l-s 43 O -a u -^ 3 O I l£l •a *• T3 1 CDo I p. 60 >< i T3 •S. B O 3 o « a is O o c ed >-i c -a P.T3 3 C C O S3 g ca 43 43 ™^ ^ -a 4^ -a — ca P* > Pi K & < u O Pi ffi < O t/3 < < a ^ < o 38 ,—, •st- •* o o o o CN ca -D a. (N 60 60 ar k bs o len z on m art z lar k a lar k lar k o O .1 Mar t U U 5 N CQ U u U £ ^ % -H •oc c 60 U o ts E > 2 S D C3 CD 3 -T3 -a c3 & -5 U o X g .2 o td 60 £ 9 '§ C i = o •a E C •s i- c o CD T3 Q bo IB g -1 '•C ° VJ O o OH 1 •* u 60 £ >; N c 60 t/1 if i •rt o o .» CD o 60 -5 'B "O CD b & B C o "i >-"% «?• 2 E & o g s CD ^ o 'is C o O O C •-= o -a E u o o o u >; „ ca E 'E o a >> y 60 J3 60 O O O S> 60 3 •§ 60 ft o ^ a c T3 u ~W VI U rse i 60 >> & -° 60 eari n i d fo r T3 C 60 ca •o O 1 s^ u ca o 60 60 i t interspe ; Sa ^2 ^ ^ ca o CD" o" uo u & (/) V} se d ro c and : O idge s O >- o e d deci O J3 - d, pastu r d , pastu r ; separate d -C u o Zft & Ia n ge s Ia n •a mi x aOJ O , D. o atur e 72 ard w fore s o ravine : J= atches , -a cr o an d woodlan d islan E <4H a 13 CH-O -a Vl 3« o o 3 c a o 3 Ms O CD ca lop e u CD •a O O 3 x CD tH J- u 3 3 2 3 .C T3 om l E? ^ > 3 CD O ci d C CD S T3 O u «o O > ca C E 'g woo d -o J3 o ID > 3 D w w S >H u >- < < < OH 55 £ z > 2 z, s P. P- 39 o o o u-i o ^^m , 60 1.3 XI O •8 cd 3 o u XI CJ r- 60 o E *- ft E O 60 M o o XI § 2 X X CJ o o N o u *3 cd -G c cd o cd Ct-1 -5 , ft *J = S >> XuI x: c o -* .9 u o 60 ON " c o XI 60 — m x) ft ii c« >. o u o — O c3 60 5 'C o o ^ eco l cd u X olo g J3 O >> 4- o tura l eg > 5 M o •<= c -? E -C CJ cd C rs c 60 S-i ft cd -2 c •a c -C O •ft o c -t-J TO S. W Q 60 0> X) ^S eo - x2 o c g se a c -° 60 60 XI 'C "a .5 C 3 J3 o >^ 3 > S G, 60 c s j> M ft g c a XI O N 2 -ft — Ji "" ft o cd ,- ft x> •c E Cd CO XI •2 -C cd •£ o c o cd w o x> 60 _£" Cad u .a O U "B o C >> 60 CJ -a "> o cd -O XI CO ™ 1,H cd (D cd "o c ft Ji '> ft u o ft 13 1) J3 60 xt ft !? C O CJ E ft £2 § 1) H 60 X •§ o u O a IS 60 w 60 cj CJ 73 c t- J-H c « '5b rese r 60 o £ 60 vT 'S60i 60 J3 -a o c 60 o o ft > Q ov e o ar g cj" ft In 3 3 & O L» CD o -o O XI 2c d 1) J-i •a #C 3 re c tu r 3 o J- o -60 -^J o o X XI o =5 o ; fore s in d m i '3 t2 V) ita t sa. 1 3 XI 'c CO -T? .3 M xc>d 3 "a o O o JS cbd o 3 s XI X! ^4-,o1 -o oas t XI '5 C s O C -a =9 2 CJ tur a c o XI 13 cj •3 S c •o^ o 1o cj -G 60 T3 om l tru e T3 c -o O •^ s s t/3 e O N o O •3 ^S ^3 X x o > ft > o >- Z 40 CO O J£ ca in € T3 3 o o. VO •a •!1i -w o O VO VO o O o o CN CN o o >t ov IN tN o 3—1 ' o o o t-~ r- oo o u II o o -D O CoN o (N = 06 ) o o ca o •^ 2 c ca T3 ca C CL) O en •c II o PH •H3 O o Jt I o c/> c ta "i3 . _ si c o s (V) 1> o \£ * £.£ ca H3 3 .£ ca o ^ HHJ "3 3 3 ^3 a i» u .: uct i rag i > 1 ca c E o S — -S -o P- b •5 « u > im p o -a •- nse r * 3 cfl o C-5rg60 - " ca t. o c rep r rin g -a an d r e T3 o 3 3 .E°l 3 C E so -3 v> •3 .2 impa c £ S o H-J C« o*s- 3 c O n t an d rnatio n ' ffecti n .3 V5 nts , d e 1 locatio n rl y gr o •z & CJ u ca •S CJ o cj ca I! 3 (« O cj o 8 u ca xi fl o E P ,o E ->. > 0>0. la g 3C ta a IS o 60 ov e ier g ov e ctor s pid . -a 'o 3 CJ O 60 gen e E 2' E ca 11 rattl e mal e u c 1 E CJ s I [/I -a 2 ca cj -a oi r cU1 B j X o - E 'E o o o er v is II t/> o ^ Vj T3 ton e ID OwH o i* 3 ->- ca O CL) S i* 2 u cj 2 •si 00 J3 cj Cu 3 *3 ^H w" if"oi C o "S 3 c -a _o 1) c "OT E g ^J 60 o CJ « w O t/3 O o II 3 O LH O E ~ ta 3 fores t vate d o as o p al e n fie l an d -bore a T3 H ca ^Ht9H -O WD ^ cd co C •-uH a. u CO .- M" •a £ ^ •o st u oi l Ia n ras s a 1 d gras s na , cu l te d pi n ciduo u t an d o 3 S T3 nefo i s ted , c -a 00 id 60 o ount a -u ^ u I-U0 U u 6 E E O u *5b fie st u 3 'i» c CJ co o CJ fores i kH Adiron d uplan d l i interfac e pin e na n bottom s woodla n C mixtur e W -S CJ -3 1 PH °- —CJ '' C "" is CJ C+^ CJ 3s 1/3 on ^ < en ca PH o 3 j5 o 1 O £ > U [ W 'S H-^ I! ca CO P, c3 ^ cB f 41 CHAPTER 2: Spatial Ecology of the Timber Rattlesnake (Crotalus horridus) in a Fragmented Agricultural Landscape ABSTRACT Although forest fragmentation is considered a threat to the imperiled Timber Rattlesnake {Crotalus horridus), few studies of this North American pitviper have taken place in fragmented habitats. Therefore, I investigated the spatial ecology of a Timber Rattlesnake population in an agricultural region of west-central Missouri. Home range estimates for 6 male Timber Rattlesnakes ranged from 28.2 to 226.5 ha using the MCP method and from 38.6 to 280.9 ha using the 95% fixed kernel analysis. Two females tracked during gravid and non-gravid seasons had MCP home ranges of 4.7 and 30.5 ha and 95% kernel home ranges of 9.9 and 49.4 ha. A single female tracked during one complete gravid season had a MCP home range of 19.5 ha and a 95% kernel home range of 17.0 ha. Hibernacula, transient habitat, gestation sites, and locations of intense foraging were identified as core areas. Maximum dispersal distances from hibernacula documented for males (2.9 km) and non-gravid females (2.8 km) approached 3.0 km. Ten gravid females gestated exclusively within an inactive rock quarry at an average distance of less than 0.2 km from their respective hibernacula. Snakes exhibited a high degree of fidelity to a specific hibernaculum, 81.8% of which were south-facing. Additionally, 33.3% of radiotelemetry subjects incurred mortality during the monitoring period. Despite large differences in habitat structure, the spatial data and movement 42 patterns of individuals in this study were similar to those previously reported for individuals in closed-canopy forests. INTRODUCTION Agriculture, development, and other anthropogenic land use practices have resulted in the loss and fragmentation of wildlife habitat on a global scale (Saunders et al., 1991). The challenge of finding suitable habitat in which to forage (Zanette et al., 2000), search for mates (Sumner, 2005), reproduce (Burke and Nol, 1998), and evade predators (Andren, 1992) maybe substantial for organisms residing in disturbed, fragmented landscapes. Consequently, the difficult task of utilizing fragmented habitat may be reflected in the movements of individuals (Moore and Russell, 2004; Hinam and St. Clair, 2008). Perhaps the most effective way for researchers to determine the influence of structural habitat change on movement patterns would be to compare the spatial ecology of study subjects prior to and after a specific disturbance event. Because many habitat altering processes occur unexpectedly or gradually take place over long periods of time, obtaining such data may not always be feasible. However, valuable information regarding a species' response to habitat fragmentation may be obtained by comparing the movement patterns of individuals from both fragmented and undisturbed habitats. Perhaps the greatest threat posed to snakes as a group is habitat destruction (Dodd, 1987). In North America, loss of snake habitat occurs primarily from residential and agricultural development, logging and forestry practices, and the impoundment of 43 streams and rivers (Dodd, 1987). Habitat structure, and the resource environment that it influences, is perhaps the most important extrinsic factor influencing snake movement patterns and home range size (Gregory et al, 1987). Considering that geographically widespread snake species do indeed exhibit population level variability in their movements (Macartney et al., 1988), emergent trends that may be related to habitat structure are worthy of investigation. The Timber Rattlesnake (Crotalus horridus) is a well-studied pitviper that primarily inhabits forested regions of eastern North America (Brown, 1993). Most consider this species to favor mature woodlands with a high degree of canopy closure (Collins and Knight, 1980; Reinert, 1984a; Reinert and Zappalorti, 1988; Martin, 1992a; Rudolph et al., 1998), and in fact, most ecological studies of Timber Rattlesnakes have been conducted in mature forests lacking large-scale habitat heterogeneity (reviewed in Wittenberg, chapter one). Although the loss and fragmentation of forested habitat is considered a threat to the Timber Rattlesnake (Martin, 1992a; Brown, 1993; Clark et al., 2003; Furman, 2007), studies of the species' movements in more open, disturbed habitats have been slow to emerge. The purpose of this study was to examine the spatial ecology of a Timber Rattlesnake population located in a fragmented, agricultural region of west-central Missouri. For 9 adult individuals with radiotelemetry data spanning one or more complete active seasons, I report home range and core area size as well as the maximum distance these snakes dispersed from their hibernacula. General patterns of movement are investigated by plotting the mean dispersal distance for males, females, and gravid females through each month of the active season. I describe patterns of fidelity to 44 rookery habitat and individual hibernacula within an inactive rock quarry. Additionally, I report and identify sources of mortality among radiotelemetry subjects. Finally, I discuss similarities in home range size, core area size, and movement patterns between individuals utilizing this fragmented habitat and those reported for closed-canopy woodland populations in the literature. MATERIALS AND METHODS Study Site.— The study took place in a highland region of west-central Missouri in what is considered the prairie geographic region (Schwartz and Schwartz, 1959). The highlands consist of hills containing hardwood forests that extend across a flat landscape of what was once tall grass prairie (Schwartz and Schwartz, 1959; Sims, 1988). This historical intermingling of closed and open canopy habitats has been further diversified by human land use practices. Anthropogenic habitat fragmentation has created a mosaic of woodland, agricultural, and residential habitats. An approximately 32 hectare inactive rock quarry provides hibernacula for all Timber Rattlesnakes in the study population. Remnant boulder piles strewn throughout the quarry are used as rookery habitat for gestating females during summer months. After leaving the quarry, males and non- gravid females spend the summer foraging in woodland, edge, and field habitats. Radiotelemetry.— Telemetry data were obtained from 26 adult Timber Rattlesnakes (13 males, 13 females) that were part of an ecological study conducted between April 2003 and May 2007. Transmitters (Holohil Systems models SI-2, SI-2T, 45 and SB-2T, Carp, Ontario) were surgically implanted into the coelomic cavities of these individuals following the protocol of Reinert and Cundall (1982). Snout-vent lengths (SVL) of implanted males were between 78.1 and 101.2 cm (mean 88.3 cm, SD = 7.68) while those of females were between 73.4 and 95.9 cm (mean 83.0 cm, SD = 6.32). During the study, 10 of the 13 females were gravid for one season. Palpation and/or ultrasound were used to determine the reproductive stage of captured females. As a precautionary measure to prevent follicular resorption, surgery was delayed for females possessing ovarian follicles until ovulation was complete (Graves and Duvall, 1993). All other subjects were typically released 48 hours following surgery at the point of capture. During the active season, snakes were radio-tracked approximately once per week. Surface activity was typically abandoned between 01 November and 31 March; therefore, snakes were only radio-tracked 2-3 times during this period. Telemetry relocations were made using a Wildlife Materials (Carbondale, Illinois) TRX-1000S receiver and Yagi 3- element directional antenna. Upon relocation, the UTM (Universal Transverse Mercator) coordinates of each snake were recorded using a WAAS (Wide Area Augmentation System) enabled GPS 72 hand held GPS receiver (Garmin International, Inc., Olathe, Kansas). On average, this unit displayed an estimated accuracy of 5.3 m (SD = 0.9). Spatial Analyses.— Home range estimates were calculated for 9 adult Timber Rattlesnakes for which approximately 30 or more radiotelemetry relocations were obtained. Movement data for these snakes encompassed between one and two complete active seasons (379 to 741 days). The sample included 6 adult males, 2 adult females that were tracked during gravid and non-gravid seasons, and 1 adult female tracked through a 46 single, gravid season. Individual snake coordinates were imported into Arc View GIS version 3.2 (Environmental Systems Research Institute, Inc., Redlands, California) and plotted on a black and white, digital orthophoto quarter quadrangle (DOQQ) map of the study site (Appendix 1). Home range and core area estimates were calculated using the Animal Movement Analyst Extension (AMAE) version 2.0 (Hooge et al., 1999) for Arc View, and these area estimates were graphically outlined on each snake's DOQQ map (Appendix 1). Home range areas were calculated using two methods, the 100% minimum convex polygon (MCP) (Mohr, 1947) and 95% fixed kernel (Worton, 1989), both of which are reported in other studies of Timber Rattlesnake spatial ecology. The MCP method calculates the area within the smallest polygon formed by connecting peripheral locations of the animal that contain all locations with each internal angle not exceeding 180 degrees (Worton, 1987). Kernel methods estimate a utilization distribution (UD) by placing probability density functions, referred to as kernels, over each location of that animal (Worton, 1989; Seaman and Powell, 1996). The UD measures the probability of distribution (i.e. area use) throughout the home range, and home range areas can be estimated for any probability contour (Worton, 1989; Gitzen et al., 2006). Superimposing a grid on the data allows density estimates to be made at each grid intersection, essentially estimating the average density of every overlapping kernel at that given location (Seaman and Powell, 1996). Thus, a major difference between the methods is that the MCP simply depicts the extent of the animal's home range whereas the fixed kernel actually reflects the intensity of areas used within the range (Worton, 1987). A 50%o fixed kernel, representing a more conservative estimate of the animal's 47 utilization distribution, was used to identify core areas within the 95% fixed kernel home range. Least-squares cross-validation (LSCV), described by Silverman (1986), was used to select the smoothing parameter (h) in all fixed kernel analyses. The LSCV method accurately selects the value of h that gives the lowest mean integrated squared error for the density estimate and performs well in kernel home range analysis (Worton, 1989; Seaman and Powell, 1996; Hooge et al., 1999; Seaman et al., 1999). Additionally, Seaman et al. (1999) recommend a minimum sample size of 30 relocations for kernel estimates of home range. The linear dispersal distance from the hibernaculum was calculated for each snake at every point of relocation. Distances were obtained by taking the square root of the following: ((ECH - ECR)2 + (NCH - NCR)2) where: ECH = Easting coordinate of the individual's hibernaculum ECR = Easting coordinate of the individual upon relocation NCH = Northing coordinate of the individual's hibernaculum NCR = Northing coordinate of the individual upon relocation Linear dispersal data were used to examine two aspects of Timber Rattlesnake spatial ecology. First, maximum dispersal distances are reported for the 9 individuals used in deriving home range estimates. Maximum dispersal distances have been reported in studies of Timber Rattlesnakes in both woodland (Brown et al., 1982; Brown, 1993; Sealy, 2002; Adams, 2005) and agricultural habitat (Fogell, 2000). Second, mean dispersal distance for males, females, and gravid females are plotted through each month 48 of the active season to determine if foraging individuals exhibit the progressive dispersal patterns described in studies of woodland populations (Reinert and Zappalorti, 1988; Reinert and Rupert, 1999; Adams, 2005). Radiotelemetry subjects with complete and partial seasonal movement data were used in calculating mean monthly dispersal distances. For each individual, dispersal distances within a given month (regardless of year) were averaged to obtain a mean value. In cases where only a single relocation was recorded for a snake during a particular month, this value took the place of a monthly mean. Mean monthly dispersal distances of individual males, females, and gravid females were then averaged to obtain mean values for these segments of the adult population. Individual females occasionally contributed dispersal data to both the "female" and "gravid female" category when they were tracked across two active seasons. RESULTS Home Range and Core Areas.— Home range estimates for the 6 male Timber Rattlesnakes ranged from 28.2 to 226.5 ha (mean = 96.3, SD = 73.9) using the MCP method and from 38.6 to 280.9 ha (mean = 122.8, SD = 82.4) using the 95% fixed kernel (Table 1). Both females that were tracked during gravid and non-gravid seasons had MCP home ranges of 4.7 and 30.5 ha and 95% kernel home ranges of 9.9 and 49.4 ha (Table 1). The female tracked during a single gravid season had an MCP home range of 19.5 ha and a 95% kernel home range of 17.0 ha (Table 1). Kernel estimates exceeded those obtained via the MCP method for 7 of the 9 snakes by an average of 28.3 ha (range 49 = 5.2 - 57.3 ha, SD = 20.4). However, the MCP estimates for the gravid season-only female and one adult male exceeded those obtained via the kernel method by 2.5 and 14.5 ha, respectively. Male core areas ranged from 3.3 to 35.3 ha (mean = 16.2, SD = 10.8) (Table 1). Core areas for the gravid/ non-gravid season females were 2.0 and 9.4 ha (Table 1). The gravid season-only female had a core area of 4.1 ha (Table 1). A single core area was identified for 3 males and 2 females. For the males, this single area encompassed the hibernaculum and transient habitat. For the females, the core area included the hibernaculum, transient habitat, and the boulders used during gestation. Two core areas were identified for the remaining 3 males. One included the hibernaculum and transient habitat while a second encompassed an area intensively used while foraging. Interestingly, 3 small core areas were identified for the remaining female that were partitioned into 1) the den and transient habitat, 2) the boulders used during gestation, and 3) an area intensively used while foraging. Dispersal Distances.— Maximum dispersal distance from the hibernacula ranged from 0.8 to 2.7 km (mean = 1.6, SD = 0.7) for the 6 males (Table 1). The gravid/ non- gravid season females dispersed maximum distances of 0.4 and 1.3 km during their non- gravid seasons (Table 1). The gravid season-only female, leaving the quarry to forage prior to gestation and again following parturition, traveled a maximum distance of 1.1 km from the hibernaculum (Table 1). Despite lacking sufficient relocations to estimate their home range, two radiotelemetry subjects actually dispersed farther than individuals for which home range estimates were calculated. A non-gravid female traveled 2.8 km 50 northeast from her hibernaculum to forage near the edges of soybean fields and wooded creek bottoms. During the late summer mating season, at the peak of her dispersal, an unmarked adult male was found associating with this snake. However, it is unclear if this male overwintered in the quarry or elsewhere. Another male subject traveled 2.9 km southeast from his hibernaculum, crossing a state highway and several large agricultural fields in doing so. After relocating this snake on 27 July of 2005 at a dispersal distance of 2.4 km, radio contact was lost and never reestablished. Mean monthly dispersal data (Figs 1, 2 and 3) indicate the annual movement patterns of both male and non-gravid females take the general form of a "loop", as described in woodland studies of Timber Rattlesnakes (Reinert and Zappalorti, 1988; Reinert and Rupert, 1999; Adams, 2005). After emerging from the hibernaculum in April, these individuals typically spent days, occasionally weeks, in transient habitat (Ditmars, 1942) before leaving the quarry to forage. Brown (1992) describes transient habitat used by Timber Rattlesnakes as rocky terrain located within 200 m of the hibernacula, supporting more open woodland with exposed clearings and shelter rocks. Rocks and boulder piles near the access roads and heavily mined areas of the quarry are consistent with Brown's description and serve as transient habitat for the study population. Foraging males traveled away from the quarry reaching mean dispersal distances of approximately 1.3 km during June and July. Mean dispersal distance decreased to 0.5 km in August as males began to reverse course and return to the quarry, again visiting transient habitat before entering their hibernaculum in September or October. Although the movements of non-gravid females resembled the looping pattern 51 of their male counterparts, the mean dispersal distance for these individuals actually peaked during August at 1.3 km. Fidelity to Rookery Habitat and Individual Hibernacula .— All 10 of the gravid females in the study gestated among boulders located in the quarry (Appendix 1, map image 11). On average, gestation occurred less than 0.2 km from the female's hibernaculum. Because 9 of the females were implanted after ovulation, the early season movements of these individuals are unknown. Extended movements were not observed for most females following parturition. Of the 5 confirmed and 2 suspected (snake was not visible beneath field vegetation) records of post-partum foraging, all but one occurred within the quarry or near its edges. One female provided movement data for an entire gravid season because she was implanted in September of the previous year. Like the others, this female gestated in boulder-laden habitat within 0.2 km of her hibernaculum. However, on two occasions this female foraged beyond the quarry, dispersing 0.8 km prior to gestation and 1.1 km following parturition. Twenty-two adult radiotelemetry subjects used 11 distinct hibernacula within the quarry (Appendix 1, map image 11). Twelve snakes (5 males, 7 females) were tracked back to their hibernacula in successive years. Only a single male switched hibernacula between years. When relocated on 02 October 2006, this individual had entered the previous year's hibernaculum. However, when relocated on 20 November 2006, the snake had moved 98.8 m and was overwintering in a south-facing rock outcropping that had served as a hibernaculum for two other radiotelemetry subjects. The majority of snakes (n = 18) overwintered in hibernacula that were primarily south-facing. The 52 remaining individuals used hibernacula that were either west-facing (n = 2), north-facing (n = 1), or located on a flat ridge (n = 1). Human-Induced Mortality.— Acquisition of spatial data was impeded by the disappearance and mortality of radiotelemetry subjects. Nine of the 27 individuals implanted with radiotransmitters (33.3 %) exited the study when they were found dead upon relocation. Home ranges could not be estimated for 8 of these snakes as they died before the requisite number of relocations needed for such calculations was obtained. Sources of mortality were known for 6 of the dead individuals (66.7%), all of which were human-induced. Four were killed by farm implements while foraging in fields (3 females, 1 male). One male was destroyed by a landowner when it was found in his driveway, while another male was killed by a vehicle on a state highway. Additionally, each season non-radio-tagged Timber Rattlesnakes were found dead on gravel roads surrounding the quarry. DISCUSSION Home Range and Core Areas.— Comparisons of MCP and kernel-based home range estimates between Timber Rattlesnakes in the present study and those in woodland habitats did not reveal pronounced differences in home range size. Focusing on the best sampled and widest ranging members of the study population, the males, I found that the few differences between my population and those in the literature were often driven by a single individual. For example, after removing a male a with an MCP home range 53 estimate of 381.6 ha from the dataset of Walker (2000), the range of MCP estimates for my males (28.2 - 226.5 ha) actually exceeded the total range of male MCP estimates derived from all published studies (15.5-212.6 ha) (Table 2). By contrast, after removing a male with a 280.9 ha 95% fixed kernel home range from my sample, the range of estimates for the remaining males in the study population (38.6 - 115.1 ha) does not approach the total range of kernel-based estimates for forest dwelling males in the literature (23.7 - 256.7 ha) (Table 3). Interestingly, the range of 50% fixed kernel core area estimates for males in the study population (3.3-35.3 ha) is comparable to the total range of published, kernel-based, core area estimates for males in woodland habitats (4.5 - 30.5 ha) (Table 4). Home range (MCP = 4.7, 30.5 ha; 95% fixed kernel = 9.9, 49.4 ha) and core area (50% fixed kernel = 2.0, 9.4 ha) estimates for the two females that were tracked during gravid and non-gravid seasons fell in the lower range of all published home range (MCP = 7.4 - 180.8 ha; 95% kernel-based = 15.3 - 127.9 ha) (Tables 2 and 3) and core area (50% kernel-based = 1.1 - 25.16 ha) (Table 4) estimates for woodland females. According to Seaman and Powell (1996), combining spatial data from gravid and non- gravid seasons may lead to smaller kernel-based estimates of home range when LSCV is used to select the amount of smoothing. The authors indicate that the inclusion of tightly clumped spatial data (i.e. data obtained on nesting animals) causes LSCV to reduce the smoothing parameter, thereby resulting in an overall lower estimate of home range area. Because the MCP home range estimates are also somewhat low for the two females in this study, inadequate sampling may better explain differences in area use between these individuals and females in woodland habitats. 54 The female tracked through one complete gravid season traveled 0.8 km from her hibernaculum to forage before returning to the quarry to gestate. Following parturition, this female was relocated at a distance of 1.1 km from her hibernaculum as she engaged in a second bout of foraging. The MCP home range estimate for this individual (19.5 ha) was near the maximum of the total range of MCP estimates reported for gravid females in woodland habitats (1.6 - 20.4 ha) (Table 2), and was inflated substantially by these foraging movements. Although gestating Timber Rattlesnakes were relatively sedentary, observations of pre-gestation and post-partum foraging underscore the importance of using datasets spanning the entire active season when deriving home range estimates for these individuals. Published home range estimates using kernel-based methods for gravid females in woodland habitats range from 2.6 to 28.1 ha (Table 3). The 95% kernel home range estimate for the gravid season-only female (17.0 ha) fell in the middle of this range. The 50% fixed kernel core area estimate for this female (4.1 ha) was approximately twice that of the largest, kernel-based, core area estimate published for gravid females in woodland habitats (range = 0.3-1.95) (Table 4). According to Gregory et al. (1987), cross-study comparisons of snake movements and home range are often plagued by methodological differences and variable quality of the data. Indeed, differences in the sample size of observations (Worton, 1987; Seaman and Powell, 1996; Seaman et al., 1999), estimation method (Worton, 1987), and computational discrepancies among software programs (Home and Garton, 2006) influence home range estimates. Furthermore, extensive variation in the movements of individuals makes it difficult to generalize about snake movement patterns within a population (Reinert and Kodrich, 1982; Reinert and Zappalorti, 1988). Nevertheless, 55 home range and core area comparisons between my study population and those in woodland habitats served as an initial attempt at determining if habitat fragmentation may affect the spatial ecology of Timber Rattlesnakes. I only drew comparisons between my data and those obtained from studies using similar methodology. Dispersal Distances and Movement Patterns.— Maximum dispersal distances for both males and non-gravid females in this study approached 3.0 km. The total range of published, maximum dispersal distances for Timber Rattlesnakes in woodland habitat was 1.1 to 3.6 km for males, but only 0.2 to 1.1 km for females (Table 5). In a West Virginia forest managed for timber production, Adams (2005) reports maximum dispersal distances of 0.3 and 0.8 km for two gravid females. As previously mentioned, the female in this study tracked for a complete gravid season traveled 1.1 km from the hibernaculum to forage following parturition. For Timber Rattlesnakes in northeastern New York, Brown (2003) reports dispersal distances for males (mean = 4.1 km, max. = 7.2 km) and females (mean 2.1 km, max. = 3.7 km). Unfortunately, Brown does not indicate if these previously unpublished data were obtained from radiotelemetry studies. For instance, Martin (1992b) reports park service personnel in the Appalachian Mountains finding Timber Rattlesnakes between 5.5 and 6.0 km from the "nearest den". These records are of unknown precision as they are not based on radiotelemetry data. In southeastern Nebraska, Fogell (2000) reports a maximum dispersal distance of 2.6 km for Timber Rattlesnakes in agricultural habitat. The dispersal distances reported by Fogell (2000) constitute the only quantitative spatial data from an agricultural population that could be 56 used for comparative purposes in this study (but see Fitch and Shirer, 1971; Fitch et al., 2004; Waldron et al, 2006). The loop-like seasonal movements of foraging snakes do not always consist solely of a single movement away from the quarry in spring and another movement back in late summer. Aside from prey acquisition, activities such as ecdysis and male mate searching appeared to influence snake movement patterns. Although many individuals remained in their foraging habitat while undergoing ecdysis, apparently some individuals sought the warmth of exposed rocks prior to shedding. One female, for instance, returned 0.6 km to the quarry from distant foraging habitat in early August and entered a pile of boulders, remaining there for at least one week. After emerging from the boulder pile, where she was believed to have shed, this individual traveled 0.4 km back to her foraging habitat where she continued to hunt through mid September. Although speculative, mate-searching behavior may explain why the mean monthly dispersal distance for males decreased in August (Fig. 1), while dispersal distances for non-gravid females did not (Fig. 2). Four observations of mating activity, consisting of 3 male-female associations and one copulation event, took place in summer foraging habitat outside of the quarry. However, male-female associations also occurred within the quarry, where males were often captured in late summer as they began appearing in core areas containing hibernacula and birthing rookeries. In the Appalachian Mountains, Martin (1992) found that observations of males at birthing rookeries peaked during August and early September. Martin also reports that most male-female associations were observed in early August. Observations of male rattlesnakes visiting birthing rookeries and associating with pregnant and/or post-partum 57 females have been reported for other Crotalus species including molossus (Dunkle and Smith, 1937), viridis, and willardi (Holycross and Fawcett, 2002). Thus, it is possible that many males returned to the quarry during late summer in order to mate with post partum females. Although foraging snakes dispersed from the quarry in all directions (Appendix 1, map image 10), all 7 individuals (5 males, 2 females) tracked into foraging habitats in successive years dispersed in the same direction as the previous year. After leaving the quarry, these snakes also displayed a tendency to hunt in the same general areas between years. Activity range overlap has also been documented in woodland studies of Timber Rattlesnakes where adult individuals were monitored for more than one active season (Reinert and Zappalorti, 1988; Reinert and Rupert, 1999; Walker, 2000; Gibson, 2003; Adams, 2005). On occasion in this study, individuals were observed using specific structures (i.e. boulders and logs) when dispersing to and from the den in the same season, or during dispersal in successive seasons. Similarly, in the deciduous forests of northeastern New York, Brown (1992) has observed individuals repeatedly using specific shelter rocks located in transient habitat year after year. Fidelity to Rookery Habitat and Individual Hibernacula .— Gravid females monitored with radiotelemetry gestated exclusively among exposed boulders in the quarry and in close proximity to their respective hibernacula (Appendix 1, map image 11). Interestingly, these anthropogenic gestation sites appear to be analogous to the birthing rookeries used by females in closed-canopy forests. Birthing rookeries consist of rocky, sparsely wooded sites that provide thermally favorable habitat for gestation and 58 eventually parturition (Keenlyne, 1972; Reinert, 1984b; Martin, 1993), and are often located near the hibernacula (Martin, 1993). Furthermore, radiotelemetry subjects in the study population hibernated exclusively within the quarry (Appendix 1, map image 11), typically used the same hibernaculum in successive years, and favored dens with south- facing exposures. Individual fidelity to a given hibernaculum and preference for hibernacula with south-facing exposures have also been documented in woodland populations of Timber Rattlesnakes (Brown, 1992; Bushar et al., 1998; Adams, 2005). Human-Induced Mortality.— Nine of 27 radiotelemetry subjects were found dead while being monitored (33.3%), and 6 were confirmed to have died as a consequence of their interactions with man (22.2%). Of the few studies in woodland habitat that describe sources of mortality among radiotelemetry subjects, none report human-induced mortality exceeding 22.2%. Although Adams (2005) reported a mortality of 76.5% for the 17 telemetry subjects in her study, only 3 individuals (17.6%) were killed by vehicles. Other known sources of mortality reported by Adams were predation (23.5%), surgical complications associated with transmitter implantation (23.5%), and a single female died from a post-partum condition (5.9%). Reinert and Rupert (1999) report mortality rates for 11 translocated (54.5%) and 18 residential (11.1%) Timber Rattlesnakes monitored with radiotelemetry. However, causes of mortality reported by Reinert and Rupert were not human related. Although Sealy (2002) documented road-killed Timber Rattlesnakes at his North Carolina study site (including 17 during a 3-year period), the author did not report instances of mortality among his 10 radiotelemetry subjects. 59 Conclusions.— Comparisons between the study population and Timber Rattlesnakes in woodland habitats revealed more similarities than differences with respect to area use and general movement patterns. Regardless of habitat structure, those attempting to protect previously unstudied Timber Rattlesnake populations would be well-advised to locate the hibernacula and focus management efforts on all surrounding habitat within 3.0 km (at minimum) of these structures. As the study population routinely foraged in open fields and habitat edges (see Appendix 1, map images 1-9), future studies of the species in agricultural habitats should investigate patch selection and the foraging (Wittenberg chapter 3), thermal (Wittenberg chapter 4), and life history (Wittenberg chapter 5) implications of anthropogenic habitat use. Furthermore, the suggestion that Timber Rattlesnakes evolving in historically grassland regions may be pre-adapted to using agricultural habitat appears valid (Fogell, 2000). Therefore, although informative, an understanding of the factors that allow this species to persist in agricultural areas of the Midwest may not allow researchers to accurately predict the response of forest- dwelling populations to habitat fragmentation. Mechanisms driving the causal relationship between anthropogenic changes in structural habitat and the decline of woodland Timber Rattlesnake populations may be best understood through habitat- manipulation studies conducted within closed-canopy forests. ACKNOWLEDGMENTS This research was conducted with the approval of the University of Arkansas Institutional Animal Care and Use Committee (protocol # 05001) and the Missouri 60 Department of Conservation (collecting permits # 12005, 12367, 12715, 13101). I thank Steven J. Beaupre, Kimberly G. Smith, Ines Pinto, and Edward E. Gbur Jr. for comments on this manuscript. I am indebted to Sara Wittenberg for her diligent proofreading and words of encouragement while this work was being completed. Joseph Agugliaro and Adam L. Crane kindly introduced me to the animal movement software used in this study. I thank the numerous private landowners in Missouri who allowed this work to be conducted on their property. This research was supported in part by a University of Arkansas Causey Grant-in-Aid Award, a Harry Steinman Memorial Grant from the St. Louis Herpetological Society, and a grant from the Arkansas Audubon Society Trust. LITERATURE CITED ADAMS, J.P. 2005. Home range and behavior of the timber rattlesnake (Crotalus horridus). Unpubl. M.Sc. thesis. Marshall Univ., Huntington. ANDREN, H. 1992. Corvid density and nest predation in relation to forest fragmentation: a landscape perspective. Ecology 73:794-804. BROWN, W.S. 1992. Emergence, ingress, and seasonal captures at dens of northern timber rattlesnakes, Crotalus horridus). In J.A. Campbell and E.D. Brodie, Jr. (eds.), Biology of the Pitvipers, pp. 251-258. Selva, Tyler, TX. BROWN, W.S. 1993. Biology, status, and management of the timber rattlesnake {Crotalus horridus): a guide for conservation. SSAR Herpetol Circ No. 22:1-78. Society for the Study of Amphibians and Reptiles. Univ. of Kansas, Lawrence BROWN, W.S., D.W. PYLE, K.R. GREENE, AND J.B. FRIEDLAENDER. 1982. Movements and temperature relationships of timber rattlesnakes {Crotalus horridus) in northeastern New York. Journal of Herpetology 16:151-161. BURKE, D.M., AND E. NOL. 1998. Influence of food abundance, nest-site habitat, and forest fragmentation on breeding Ovenbirds. The Auk 115:96-104. 61 BUSHAR, L.M., H.K. REINERT, AND L. GELBERT. 1998. Genetic variation and gene flow within and between local populations of the timber rattlesnake, Crotalus horridus. Copeia 1998:411-422. CLARK, A.M., P.E. MOLER, E.E. POSSARDT, A.H. SAVITZKY, W.S. BROWN, AND B.W. BOWEN. 2003. Phylogeography of the timber rattlesnake {Crotalus horridus) based on mtDNA sequences. Journal of Herpetology 37:145-154. COLLINS, J.T., AND J.L. KNIGHT. 1980. Crotalus horridus. Catalogue of American Amphibians and Reptiles 253: 1-2. DITMARS, R.L. 1942. Snakes of the world. Macmillan, New York. DODD, C.K., JR. 1987. Status, conservation, and management. In R.A. Seigel, J.T. Collins, and S.S. Novak (eds.), Snakes: Ecology and Evolutionary Biology, pp. 478-513. McGraw-Hill, New York, NY. DUNKLE, D.H., AND H.M. SMITH. 1937. Notes on some Mexican ophidians. Occasional Papers of the Museum of Zoology, University of Michigan 363:1-15. FITCH, H.S., G.R. PISANI, H.W. GREENE, A.F. ECHELLE, AND M. ZERWEKH. 2004. A field study of the timber rattlesnake in Leavenworth County, Kansas. Journal of Kansas Herpetology 11:18-24. FITCH, H.S., AND H.W. SHIRER. 1971. A radiotelemetric study of spatial relationships in some common snakes. Copeia 1971:118-128. FOGELL, D.D. 2000. Seasonal activity, habitat preferences, and natural history of the timber rattlesnake {Crotalus horridus) in southeastern Nebraska. Unpubl. M.Sc. thesis, Univ. of Nebraska at Omaha, Omaha. FURMAN, J. 2007. Timber Rattlesnakes in Vermont and New York: biology, history, and the fate of an endangered species. University Press of New England, Lebanon, New Hampshire. GIBSON, S.E. 2003. Movements and habitat preference of the timber rattlesnake {Crotalus horridus) in southern Indiana. Unpubl. M.Sc. thesis, Purdue University, Fort Wayne, Indiana. GITZEN, R.A., J.J. MILLSPAUGH, AND B.J. KERNOHAN. 2006. Bandwidth selection for fixed-kernel analysis of animal utilization distributions. Journal of Wildlife Management 70:1334-1344. GRAVES, B.M., AND D. DUVALL. 1993. Reproduction, rookery use, and thermoregulation in free-ranging, pregnant Crotalus v. viridis. Journal of Herpetology 27:33-41. 62 GREGORY, P.T., J.M. MACARTNEY, AND K.W. LARSEN. 1987. Spatial patterns and movements. In R.A. Seigel, J.T. Collins, and S.S. Novak (eds.), Snakes: Ecology and Evolutionary Biology, pp. 366-395. McGraw-Hill, New York, NY. HINAM, H.L., AND C.C. ST.CLAIR. 2008. High levels of habitat loss and fragmentation limit reproductive success by reducing home range size and provisioning rates of Northern Saw-whet Owls. Biological Conservation 141:524-535. HOLYCROSS, A.T., AND J.D. FAWCETT. 2002. Observations on neonatal aggregations and associated behaviors in the prairie rattlesnake, Crotalus viridis viridis. American Midland Naturalist 148:181-184. HOOGE, P.N., W. ElCHENLAUB, AND E. SOLOMON. 1999. The Animal Movement Program version 2.0 Beta. Anchorage, Alaska Biological Science Center, U.S. Geological Survey. HORNE, J.S., AND E.O. GARTON. 2006. Likelihood cross-validation versus least squares cross-validation for choosing the smoothing parameter in kernel home-range analysis. Journal of Wildlife Management 70:641-648. KEENLYNE, K.D. 1972. Sexual differences in feeding habits of Crotalus horridus horridus. Journal of Herpetology 6:234-237. MACARTNEY, J.M., P.T. GREGORY, AND K.W. LARSEN. 1988. A tabular survey of data on movements and home ranges of snakes. Journal of Herpetology 22:61-73. MARTIN, W.H. 1992a. The timber rattlesnake: its distribution and natural history. In T.F. Tyning (ed.), Conservation of the Timber Rattlesnake in the Northeast, pp. 13-22. Massachusetts Audubon Society, Lincoln, MA. MARTIN, W.H. 1992b. Phenology of the timber rattlesnake {Crotalus horridus) in an unglaciated section of the Appalachian Mountains. In J. A. Campbell and E.D. Brodie, Jr. (eds.), Biology of the Pitvipers, pp. 259-277. Selva, Tyler, TX. MARTIN, W.H. 1993. Reproduction of the timber rattlesnake {Crotalus horridus) in the Appalachian Mountains. Journal of Herpetology 27:133-143. MOHR, CO. 1947. Table of equivalent populations of North American small mammals. American Midland Naturalist 37:223-249. MOORE, J.E., AND R.E. RUSSELL. 2004. Empirical considerations for modeling animal movements in human-dominated landscapes. In R.K. Swihart and J.E. Moore (eds.), Conserving Biodiversity in Agricultural Landscapes: Model-Based Planning Tools, pp. 165-180. Purdue University Press, West Lafayette, IN. 63 REINERT, H.K. 1984a. Habitat separation between sympatric snake populations. Ecology 65:478-486. REINERT, H.K. 1984b. Habitat variation within sympatric snake populations. Ecology 65: 1673-1682. REINERT, H.K., AND D. CUNDALL., 1982. An improved surgical implantation technique for radio tracking snakes. Copeia 1982:702-704. REINERT, H.K., AND W.R. KODRICH. 1982. Movements and habitat utilization by the Massasauga, Sistrurus catenatus catenatus. Journal of Herpetology 16:162-171. REINERT, H.K., AND R.R. RUPERT Jr. 1999. Impacts of translocation on behavior and survival of timber rattlesnakes, Crotalus horridus. Journal of Herpetology 33:45- 61. REINERT, H.K., AND R.T. ZAPPALORTI. 1988. Timber rattlesnakes (Crotalus horridus) of the pine barrens: their movement patterns and habitat preference. Copeia 1988:964-978. RUDOLPH, D.C., AND S.J. BURGDORF. 1997. Timber rattlesnakes and Louisiana pine snakes of the West Gulf Coastal Plain: hypotheses of decline. Texas Journal of Science 49:111-122. RUDOLPH, D.C., S.J. BURGDORF, R.N. CONNER, AND J.G. DICKSON. 1998. The impact of roads on the timber rattlesnake, (Crotalus horridus), in eastern Texas. In G.L. Evink, P. Garrett, D. Zeigler, and J. Berry, (eds.), Proceedings of the International Conference on Wildlife Ecology and Transportation, FL-Er-69-98, pp. 236-240. Florida Department of Transportation, Tallahassee, FL. SAUNDERS, D.A., R.J. HOBBS, AND C.R. MARGULES. 1991. Biological consequences of ecosystem fragmentation: a review. Conservation Biology 5:18-32. SCHWARTZ, C.W., AND E.R. SCHWARTZ. 1959. The Wild Mammals of Missouri. University of Missouri Press, Columbia, Missouri, USA. SEALY, J.B. 2002. Ecology and behavior of the timber rattlesnake (Crotalus horridus) in the upper Piedmont of North Carolina: identified threats and conservation recommendations. In G.W. Schuett, M. Hoggren, M.E. Douglas, and H.W. Greene, (eds.), Biology of the Vipers, pp. 561-578. Eagle Mountain Publishing, Eagle Mountain, UT. SEAMAN, D.E., J.J. MILLSPAUGH, B.J. KERNOHAN, G.C. BRUNDIGE, K.J. RAEDEKE, AND R.A. GITZEN. 1999. Effects of sample size on kernel home range estimates. Journal of Wildlife Management 63:739-747. 64 SEAMAN, D.E., AND R.A. POWELL. 1996. An evaluation of the accuracy of kernel density estimators for home range analysis. Ecology 77:2075-2085. SILVERMAN, B.W. 1986. Density estimation for statistics and data analysis. Chapman and Hall, London, UK. SIMS, P.L. 1988. Grasslands. In: M.G. Barbour and W.D. Billings (eds.), North American Terrestrial Vegetation, pp. 265-286. Cambridge University Press, New York, NY. SUMNER, J. 2005. Decreased relatedness between male prickly forest skinks {Gnypetoscincus queenslandiae) in habitat fragments. Conservation Genetics 6:333-340. WALDRON, J.L., J.D. LANHAM, AND S.H. BENNETT. 2006. Using behaviorally-based seasons to investigate canebrake rattlesnake {Crotalus horridus) movement patterns and habitat selection. Herpetologica 62:389-398. WALKER, Z.J. 2000. The spatial ecology of the timber rattlesnake {Crotalus horridus) in south central Indiana. Unpubl. M.Sc. thesis, Purdue University, Fort Wayne, Indiana. WORTON, B.J. 1987. A review of models of home range for animal movement. Ecological Modelling 38:277-298. WORTON, B.J. 1989. Kernel methods for estimating the utilization distribution in home- range studies. Ecology 70:164-168. WORTON, B.J. 1995. Using monte carlo simulation to evaluate kernel-based home range estimators. Journal of Wildlife Management 59:794-800. ZANETTE, L., P. DOYLE, and S.M. TREMONT. 2000. Food shortage in small fragments: evidence from an area-sensitive passerine. Ecology 81:1654-1666. 65 Table 1. Estimates of home range area (100% MCP, 95% fixed kernel) and area(s) of core use (50% fixed kernel) for 9 adult Timber Rattlesnakes. Also provided for each snake is the sex and reproductive condition, snout-vent length (SVL), number of telemetry observations (Obs.), sampling period, duration of study, smoothing parameter for kernel analyses (selected using least-squares cross-validation), and the maximum dispersal distance from the hibernaculum. 66 a a a *O 33 S u g § s a r- r- ON C-- ^^ OO ^H o O •* 00 NO ^H f- ON 8 -2 u "a" o ^oH J2 " •° 9 r-~ oo ^ CN t-; CN co S ^ S O -H o •-< °)>-< —^ CN .-H o »—1 NO >n O ON NO o o ON r- p ON o co o ON OO ON p NO l> ON r-; NO CN CN O CO oo NO ON CN r-- "3" O ON >n ON m r- oo NC o CN ON CN CN O O co ON o r- NO ON o r- •* CN ON ON r- •* o IT) O oo 9. 6 oo OO 00 9. 4 6. 9 a e co ON 00 CN •* ON -—< CN co oo >n 00 CN in "3- CN oo ON co NO in cn oo CN CN o ON CN CN co o3 •* NO oo ON ON t- r-- in ON I—1 ^* ON >n oo •* r- m NO NO ^f lO NO NO m NO 04 - 7 o 03 - 3 o o 1/1 in •a CM >n o o o TJ- in CN CN CN O o o _o o 1H f-^. o ° -3- o o o r~ co JQ CoN j5 CN CN CN CN 4 ^ o CN ° o XI X> x> ^ ^H o o © u< o O-(^l S "K o a X> © ° X o (N o CN p ON in ON o O OO NO CO NO in SV L (cm ) r^ '"' oo OO ON oo r- r~ ON •o T3 0>3 '>cd VH 1- 3 S 00 bfl O ft +J 3 3 o3 3 3 3 3 U s , v T3 T3 3T3 T3 T3 T3 ,—* tz> C3 c3 03 3 3 an d r e vid/no : vid/no ; ditio n "3 "ca o CZo1 « PQ o CO ON O CN o O o m ON ON r-H O NO NO NO ^^ --* © CN oo CN o 67 Table 2. Published home range estimates of adult Timber Rattlesnakes using the MCP method. Both the range and mean of area estimates are given in hectares (ha). Data are partitioned by snake sex and reproductive condition (male, female, gravid female) as well as by published data source (citation). The habitat and/or geographic region where the study took place and the sample size of telemetry subjects are also provided. 68 CN in ON . rn 0\ r^ •—! ^r CN *—I o NO m ON ^ 7 ON ^ m CN r- oo NO in o CN o °° NO m o o CN en ™ CN I I I a t> ^H in NO in 7 ON in o >n in O O ON ID (^ _H 60 d a I.S re re d Pi Pi re Pi 1 e- re a 73 re re re a d d d re re re re re •3 '3 •3 73 '3 U d 1 d d re re a X a re re ID re ij > > d d H It) re H 2; re ID D ID w O O d O re .g IU •3 ID "o -g 00 d oo r- oo oo ON oo ON ON ON ON ON I—i T-H ,^-~ON , ON *~s ON ON .-r-t ^—<+H ' ON ON •e ^H O O 73 _o 73 alo i 60 •C "re 60 OH i-T ID a IH 0) £H 3 OH d a, C? PQ 3 & PQ d o in m N Pi N 73 o> o >n -o o CN o O -a o 73 0 73 o O d C CI O o d re o re IN CN CN CN (20 0 O re and ] (20 0 !002 ) -d I.N CZ) _&, IU d •e ti 53 d io n er t a er t a _>N o ID 03 CN OO CN s o 43 PH u 0s n3 45 ON u £ en ON r- „ o NO he r 00 H o 3 PH .g s U •a S u o 45 ft I* loc a 4j o (D o a e o 3 •S ,o > Pw a o to» '^=3 'So o 3 -a o o o 3 > 73 o [» •a o O o 03 J-i 45 a re m a 03 ft o 03 o o 43 T3 t _o3 a> O o3 60 OH 13 03 al s 3 a0) m 03 •8 a Hi X) a 03 c O u O T3 O ^ CN o oo o ke r ON s ON 6 13 ON VN ON £ T3 >N U 42 ft T3 ft ft 3 3 c? m,— , m O N P< o o o c According to Worton (1995), an inherent problem with home range estimates based upon the harmonic mean method is that central data points have an excessive influence on outer areas. Both the range and mean of area estimates are given in hectares (ha). Data are partitioned by snake sex and reproductive condition (male, female, gravid female) as well as by published data source (citation). The habitat and/or geographic region where the study took place and the sample size of telemetry subjects are also provided. 71 03 ;3 u (N oo w oo in t-l 13 o © •*' O CN o CN ,_; ON ON ON «a3 CN ON o >n CN 207 . CJ 3 t» 03 mea i •3- oa ^—' ON o CN r- '3 p -t—* ON o 03 ON CN '8o .—< CN OO CN CN oo CN g CN S CN J 1 o3 a I I I J3 •t^» I 1 03 I OS t/3 03 03 03 a 3 a 2 S •3 •3 •3 e C3 a 03 3 X X 'S a H o D u a .3 03 o -t-» .3 cj •p" «3 Ba 3 3 '5b| T3 o .a -a 3 O T3 3 o o O 'B. o o o o ft o O o -a -a o 3 a s 3 a oj 133 03 •a S3 u 03 03 3 U o CO o .3 73 cd cd cd r- l . so •_£ r- m .-t 6 co (N O Tf CO ID i-H o CI o § cd «3 nj D u s c cd a cd a cd cd SO O in in a 1 cd o in O CN m (N § § I I 8 I in r- C3 o OS cd o o cd o 1) so ao so o T3 3 a c a cd CO cd pi Pi 3 75 in o CN O o CN 00 CN oo ON in CN CN ON ^pi valu e d lea n CN O CN CN o oo oo m OO CN O o d ca 13 © I I I I I I > ON oo o oo > <4-l OO o r- r- CN oo •a o © r- CJ m O 00 a> oo a CJ a. c d a ca ca Pi Pi > ca u a o T3 cj "O 3 S3 c3 C/5 les m J3 ca a Ratt l ca Pi C3 id ca 3 3 .3 'a O '3 o 'oo • a JO 'ob 1 > CJ > ca "ca to -§ a B > a ca >, ca a C3 ca o T3 ca CJ 3 ca -a U 3 cd .3 U 3 .3 o c/5 +J 0) •4—' 3 J3 U t» 3 3 o % O € O o jo -o u CN CN oo OO a ON o o in o o >n o o o CN O o o O o o O O o o ca o CN CN CN O Values appearing above error bars represent the number of telemetry individuals that contributed data to the mean monthly dispersal distance calculation. All spatial data collected between the springs of 2003 and 2007 were pooled to illustrate these general dispersal patterns. 77 8 2.5 - 7 2.0 - 1.5 - 9 V 8 1.0 - 0.5 - \. 8 6 / 4 0.0 - i i 1 Apr May Jun Jul Aug Sep Oct Month 78 Figure 2. Mean monthly dispersal distances of non-gravid female Timber Rattlesnakes. Dispersal patterns of individuals were quite variable, as indicated by error bars representing minimum and maximum mean dispersal distances of individuals for a given month. Values appearing above error bars represent the number of telemetry individuals that contributed data to the mean monthly dispersal distance calculation. All spatial data collected between the springs of 2003 and 2007 were pooled to illustrate these general dispersal patterns. 79 3.5 3.0 4 5 2.5 2.0 4 6 1.5 3 1.0 6 . 0.5 \1 0.0 i i 1 1 i Apr May Jun Jul Aug Sep Oct Month 80 Figure 3. Mean monthly dispersal distances of female Timber Rattlesnakes during a season in which they were gravid. Dispersal patterns of individuals were quite variable, as indicated by error bars representing minimum and maximum mean dispersal distances of individuals for a given month. Values appearing above error bars represent the number of telemetry individuals that contributed data to the mean monthly dispersal distance calculation. All spatial data collected between the springs of 2003 and 2007 were pooled to illustrate these general dispersal patterns. 81 Apr May Jun Jul Aug Sep Oct Month 82 Appendix I Map images 1-9: Home ranges (MCP and 95% fixed kernel), core areas (50% fixed kernel), and relocations points of individual Timber Rattlesnakes. Map image 10: Composite map of the MCP home ranges of nine individual Timber Rattlesnakes. Map image 11: Gestation sites and confirmed hibernacula within an approximately 32 hectare inactive rock quarry. 83 Map Image 1. Adult male 150 was a 100.0 cm SVL individual that was relocated 43 times between 03 October 2004 and 28 August 2006. Individual dots represent one or more relocations of this snake. The polygon represents a 100% MCP home range area of 85.434 ha. Two outermost contour lines comprise a 95% fixed kernel home range area of 114.270 ha, while the innermost contour lines represent two 50% fixed kernel core areas within the 95% home range. The core area in the southwest portion of the map encompasses the hibernaculum (rock outcropping with south and west-facing edges) and rocky, transient habitat within the quarry. In the northeast portion of the map, a subset of this snake's foraging habitat is identified as another area of core-use. Collectively, the 50%o fixed kernel core areas represent 12.397 ha. In both 2005 and 2006, snake 150 emerged from the same hibernaculum and traveled northeast to forage in habitat edges, fields, and wooded creek-bottoms. The maximum distance from which this snake was relocated from its hibernaculum was 1.707 km. 84 £ #* 85 Map Image 2. Adult male 193 was a 101.2 cm SVL individual that was relocated 36 times between 20 September 2003 and 06 August 2005. Individual dots represent one or more relocations of this snake. The polygon represents a 100% MCP home range area of 28.221 ha. Two outermost contour lines comprise a 95% fixed kernel home range area of 38.576 ha, while a single, innermost contour line represents a 50% fixed kernel core area of 3.285 ha. The core area encompasses the hibernaculum (west-facing rock outcropping) and rocky, transient habitat within the quarry. In both 2004 and 2005, snake 193 emerged from the same hibernaculum and traveled northwest to forage primarily in a tract of secondary woodland habitat containing glades. The tract was bisected by a gravel road that the snake would occasionally cross and even forage alongside. After relocating this individual on 06 August 2005, radio contact was permanently lost. The maximum distance from which this snake was relocated from its hibernaculum was 0.847 km. 86 *« -.;•>* Mr •** Ft s V 87 Map Image 3. Adult male 099 was an 88.2 cm SVL individual that was relocated 31 times between 13 June 2004 and 05 August 2005. Individual dots represent one or more relocations of this snake. The polygon represents a 100% MCP home range area of 75.843 ha. Two outermost contour lines comprise a 95% fixed kernel home range area of 98.511 ha, while the innermost contour lines represent two 50% fixed kernel core areas within the 95% home range. The core area in the northwest portion of the map encompasses the hibernaculum (rock outcropping at the crest of a brushy, south-facing slope) and nearby foraging habitat, including both a narrow strip of unmowed vegetation at the southern edge of the quarry and a grassy field above the hibernaculum. In the southeast portion of the map, a subset of this snake's foraging habitat is identified as another area of core-use. Collectively, the 50% fixed kernel core areas represent 10.447 ha. In both 2004 and 2005, snake 099 traveled southeast of the quarry to forage. This individual foraged primarily in habitat edges, but was relocated in old field and woodland habitats as well. When relocated on 05 August 2005, snake 099 was found dead on a state highway while in the process of extending its range to the south (see southernmost relocation on map). The maximum distance from which this snake was relocated from its hibernaculum was 1.189 km. 88 3 Ml -\ •MP M 89 Map Image 4. Adult male 210 was an 86.0 cm SVL individual that was relocated 33 times between 07 September 2004 and 18 September 2006. Individual dots represent one or more relocations of this snake. The polygon represents a 100% MCP home range area of 129.584 ha. Two outermost contour lines comprise a 95% fixed kernel home range area of 115.116 ha, while a single, innermost contour line represents a 50% fixed kernel core area of 16.694 ha. The core area encompasses the hibernaculum (rock outcropping with south and west-facing edges) and transient habitat within the quarry (including a massive pile of waste tires). In both 2005 and 2006, snake 210 emerged from the same hibernaculum and traveled west/southwest foraging primarily in habitat edges. The maximum distance from which this snake was relocated from its hibernaculum was 1.967 km. 90 &Mtf tt~ •VAT.'- *£ 91 Map Image 5. Adult male 802 was a 94.4 cm SVL individual that was relocated 26 times between 20 September 2003 and 03 October 2004. Individual dots represent one or more relocations of this snake. The polygon represents a 100% MCP home range area of 32.341 ha. Two outermost contour lines comprise a 95% fixed kernel home range area of 89.649 ha, while the innermost contour lines represent two 50% fixed kernel core areas within the 95% home range. The core area in the south-central portion of the map encompasses the hibernaculum (rock outcropping with south and west-facing edges) and rocky, transient habitat within the quarry. In the north-central portion of the map, a subset of this snake's foraging habitat is identified as another area of core-use. Collectively, the 50% fixed kernel core areas represent an area of 19.070 ha. In 2004, snake 802 emerged from its hibernaculum and traveled north to forage in both secondary woodland habitat containing glades and habitat edges, including old-growth vegetation approximately 10.0 m from a gravel road. After entering its hibernaculum in the fall of 2004,1 lost radio contact with this snake (presumably due to battery failure). The maximum distance from which this snake was relocated from its hibernaculum was 1.251 km. 92 r mm wm 93 Map Image 6. Adult male 260B was an 83.9 cm SVL individual that was relocated 30 times between 03 October 2005 and 25 May 2007. Individual dots represent one or more relocations of this snake. The polygon represents a 100% MCP home range area of 226.452 ha. Two outermost contour lines comprise a 95% fixed kernel home range area of 280.904 ha, while a single, innermost contour line represents a 50% fixed kernel core area of 35.257 ha. The core area encompasses two hibernacula separated by 98.838 m (snake 260B was the only telemetry subject observed to switch hibernacula between years) and rocky, transient habitat within the quarry. During 2005/2006, the snake overwintered in the southeast-facing slope of a rocky bowl. When relocated on 02 October 2006, snake 260B had entered the previous year's hibernaculum. However, when relocated on 20 November 2006, the snake had moved to a south-facing rock outcropping used by two other telemetry subjects during their time in the study. Snake 260B traveled north of the quarry to forage in habitat edges, fields, and woodlands. The maximum distance from which this snake was relocated from its hibernaculum was 2.718 km. 94 v.-..-'/ • \l MS 95 Map Image 7. Adult female 060 was a 73.4 cm SVL individual that was relocated 36 times between 19 July 2004 and 07 May 2006. Individual dots represent one or more relocations of this snake. The polygon represents a 100% MCP home range area of 30.548 ha. Two outermost contour lines comprise a 95% fixed kernel home range area of 49.407 ha, while a single, innermost contour line represents a 50% fixed kernel core area of 9.418 ha. The core area encompasses a significant portion of the quarry including 1) the hibernaculum located on a south-facing, boulder-laden slope, 2) rocky, transient habitat, and 3) boulder piles used for gestation during the summer of 2004. In 2005, a non-gravid year, snake 060 traveled west/northwest to forage primarily in a narrow, wooded fencerow separating a field of soybeans from a cut, grassy field. The maximum distance from which this snake was relocated from its hibernaculum was 1.271 km. 96 Jfe*. H (, & . ,> 1 ?^(. .?•:••• &:J L , -> >- -v'i"!• ; •''> -#* ? •s f.f- •••- 97 Map Image 8. Adult female 821 was a 76.5 cm SVL individual that was relocated 43 times between 22 July 2003 and 09 May 2005. Individual dots represent one or more relocations of this snake. The polygon represents a 100% MCP home range area of 4.733 ha. Two outermost contour lines comprise a 95% fixed kernel home range area of 9.879 ha, while the innermost contour lines represent three 50% fixed kernel core areas within the 95%o home range. The northernmost core area represents the portion of the quarry where the snake gestated during 2003. Directly to the south of this rookery habitat, a second area of core-use primarily encompasses the hibernaculum (north-facing rock outcropping). Farther to the south, the third and smallest area of core use represents the most intensively used portion of the old field in which this female foraged during 2004. Collectively, the 50% fixed kernel core areas represent an area of 2.027 ha. Interestingly, a rock wall adjacent to the road was used by the snake, possibly for ecdysis. The maximum distance from which this snake was relocated from its hibernaculum was 0.390 km. 98 w W-'* -V »V" S*i 1 • r„ ^ . m iH £wfa ***** j"*--. •3$ ••••a a* t *', 99 Map Image 9. Adult female 760B was a 95.9 cm SVL individual that was relocated 28 times between 17 October 2005 and 15 April 2007. Movements made during this time period represent a single, gravid season. Individual dots represent one or more relocations of this snake. The polygon represents a 100% MCP home range area of 19.517 ha. Three outermost contour lines comprise a 95% fixed kernel home range area of 16.954 ha, while a single, innermost contour line represents a 50% fixed kernel core area of 4.094 ha. The core area encompasses the hibernaculum (rock outcropping with south and west- facing edges) and quarry habitat containing boulders used for gestation. Although females in the study population stayed within the quarry and tended to display restricted movement patterns during gestation, snake 760B left the quarry to forage prior to gestating and again following parturition (see relocations to the east and southeast of quarry). The foraging pattern of this snake underscores the importance of using data from the entire active season when investigating the home range of gravid individuals, regardless of how sedentary they may seem. The maximum distance from which this snake was relocated from its hibernaculum was 1.100 km. 100 '&*•;,: -!*•-*- * m 101 Map Image 10. Composite map of the MCP home ranges of nine individual Timber Rattlesnakes. This map illustrates how individuals in the study population dispersed from the quarry in essentially every direction to use the surrounding habitat. 102 •n % 103 Map Image 11. Exclusive use of an inactive rock quarry by radiotelemetry-monitored Timber Rattlesnakes for both hibernation and gestation. Dots represent 11 distinct hibernacula used by 22 adult telemetry subjects. Any location where a snake overwintered, regardless of physical structure, was considered to be a hibernaculum. Symbols in the form of an "X" represent one or more relocations of a gestating female. Pooling the telemetry data of 10 gestating females (including females 060, 821, and 760B) yielded 124 relocations within the quarry. 104 HI vM V •£ $tt 105 CHAPTER 3: Foraging Ecology of the Timber Rattlesnake {Crotalus horridus) in a Fragmented Agricultural Landscape ABSTRACT Habitat loss and fragmentation pose perhaps the greatest threats to snake populations. However, as with other terrestrial ectotherms, snakes are poorly represented in vertebrate studies focusing on the effects of habitat fragmentation. Despite habitat fragmentation being cited as a primary threat to the Timber Rattlesnake {Crotalus horridus), our knowledge of this species' foraging ecology is largely based on studies of populations in mature forests. I present data obtained from a 4-year study of the Timber Rattlesnake in west-central Missouri where snakes readily foraged in secondary woodland tracts and corridors, agricultural fields, and habitat edges. Dietary analysis found that snakes fed exclusively on mammals including shrews (Soricidae), mice in the genus Peromyscus, Prairie Voles {Microtus ochrogaster), Cotton Rats (Sigmodon hispidus), Eastern Gray Squirrels {Sciurus caro linens is), and Eastern Cottontails {Sylvilagus floridanus). Although small mammal trapping indicated that fields contained fewer numbers of prey than woodlands and habitat edges, field dwelling Prairie Voles were the most frequently consumed prey item. Arboreal foraging and ambush strategies utilizing logs and tree trunks were not prevalent in the study population. Although behavioral responses to habitat fragmentation may be population-specific, the present study indicates that large tracts of mature forest are not required for Timber Rattlesnakes to forage effectively. Therefore, prey availability may be more important to this species than the physical structure of the habitat in which it forages. 106 INTRODUCTION Studies of terrestrial vertebrates in fragmented habitats have primarily focused on birds and mammals, while in these same habitats, reptiles and amphibians remain largely unstudied (Mac Nally and Brown, 2001; Heard et al., 2004). Conservation efforts targeting reptiles and amphibians have been significantly undermined because little is known regarding their natural history in human-impacted landscapes (Bury, 2006). In order to successfully manage herpetofauna, researchers cannot strictly rely on past studies conducted in pristine habitats (Bury, 2006). This may be especially true for snakes because, according to Dodd (1987), the greatest threat posed to ophidians as a group is habitat destruction. In North America, Dodd (1987) claims that loss of snake habitat has occurred primarily from residential and agricultural development, logging and forestry practices, and the impoundment of streams and rivers. The Timber Rattlesnake (Crotalus horridus) is a North American snake species threatened by loss and fragmentation of closed-canopy forests (Martin, 1992; Brown, 1993; Clark et al., 2003; Furman, 2007). This well-studied pitviper primarily inhabits forested regions of eastern North America (Brown, 1993), where it is considered to favor mature woodlands with a high degree of canopy closure (Collins and Knight, 1980; Reinert, 1984; Reinert and Zappalorti, 1988; Martin, 1992; Rudolph et al., 1998). In fact, most ecological studies of Timber Rattlesnakes have been conducted in mature forests lacking large-scale habitat heterogeneity (reviewed in Wittenberg, chapter one). Thus, despite the threat posed by habitat fragmentation to the Timber Rattlesnake, our understanding of this species' foraging ecology stems primarily from research on forest dwelling populations (Reinert et al., 1984; Brown and Greenberg, 1992; Clark 2006a,b). 107 Efforts to conserve declining Timber Rattlesnake populations in fragmented forests may be enhanced by studying how individuals acquire prey in more open, disturbed habitats. From April 2003 through May 2007,1 studied the ecology of a Timber Rattlesnake population in west-central Missouri that inhabited a fragmented, agricultural landscape. Data obtained through dietary analysis, radiotelemetry, and small mammal trapping are used to answer three primary questions regarding the foraging ecology of snakes in this population. In what types of habitat (i.e. woodland, field, or edge) do individuals forage? What prey species are being consumed? How do prey abundances compare among different foraging habitats? Finally, I discuss how the foraging behavior and diet of the study population compares to that of previously studied Timber Rattlesnake populations in closed-canopy forests. MATERIALS AND METHODS Study Site.— The study took place in a highland region of west-central Missouri in what is considered the prairie geographic region (Schwartz and Schwartz, 1959). The highlands consist of hills containing hardwood forests that extend across a flat landscape of what was once tall grass prairie (Schwartz and Schwartz, 1959; Sims, 1988). This historical intermingling of closed and open canopy habitats has been further diversified by human land use practices. Anthropogenic habitat fragmentation has created a mosaic of woodland, agricultural, and residential habitats. An approximately 32 hectare inactive rock quarry provides hibernacula for all timber rattlesnakes in the study population. Remnant boulder piles strewn throughout the quarry are used as rookery habitat for gestating females during summer months. Males and non-gravid females spend the 108 summer using an array of habitat types including small woodlots, narrow wooded fencerows, old fields, row crops, and areas where these habitats converge (edge habitat). Study Animals.— During the course of the study, radio transmitters (Holohil Systems models SI-2, SI-2T, and SB-2T, Carp, Ontario) were surgically implanted into the coelomic cavities of 26 timber rattlesnakes (13 males, 13 females) following the protocol of Reinert and Cundall (1982). Snout-vent lengths (SVL) of implanted snakes were measured in a squeeze box (Quinn and Jones, 1974) and ranged from 68.0 to 101.2 cm. Snake masses ranged from 245 to 1126.5 g prior to implantation and transmitter weight (range 5.2 - 14.2 g) never exceeded 5% of the snake's total body mass. During the study, 10 of the 13 females were gravid for one season. Observations of Foraging Behavior.— Telemetry relocations provided the basis for observational data on the foraging habits of snakes in the study population. Snakes were determined to be foraging when observed in a typical ambush posture (Beaupre, 2008). In nearly every instance, ambushing snakes were found on the surface in a tight, concentric coil. The neck extended from the middle of this coil in an S-shaped fashion, allowing the head to rest upon an outer coil of the body. From this "cocked" position, snakes can readily strike passing prey (Cundall and Beaupre, 2001). Snakes were not considered to be foraging if they were in ecdysis, digesting prey (could only be determined if the body was visibly distended), on the move, associating with an individual of the opposite sex, or gestating among boulders. Lush undergrowth associated with edge habitats and the dense tussocks of vegetation in fields made 109 observing snakes in these habitats difficult. Often, all or part of an individual could only be seen by slightly moving the vegetation concealing the snake. Surrounding vegetation was not manipulated if disturbing the snake was likely, thus preventing snakes from being directly observed in 17.2% of edge and 22.7% of field relocations. By contrast, snakes were at least partially visible during 95.6% of woodland relocations. Because snake behavior could not be determined if the individual was concealed by surface vegetation, instances of foraging in edges and fields were likely underestimated. Dietary Analysis.— Thirty fecal samples were opportunistically collected from Timber Rattlesnakes while in captivity for mark-recapture processing or transmitter implantation/removal. The identities of individual snakes that deposited feces were recorded so that dietary data could be analyzed with respect to snake size and sex. Occasionally, snakes were transported or housed together which made assigning feces to an individual snake impossible. Upon dissection, five of twelve Timber Rattlesnakes killed while crossing roads at my study site contained dietary material. I explored the entire digestive tract of each roadkilled specimen to recover intact prey items from the stomach as well as ingesta distributed throughout the intestines. Finally, dietary material was recovered from the scattered remains of two telemetry subjects killed by farm implements, bringing the total number of dietary samples examined in this study to thirty- seven. With the aid of a dissecting microscope, feces and the ingesta recovered from salvaged specimens were dissected in a Petri dish containing 70% ethanol. Consumed prey items were readily identified to vertebrate class owing largely to the presence of 110 undigested hair. Samples of the hair were cleaned and mounted on microscope slides for further examination using a compound light microscope. Characteristics such as hair length, width, color, shape, cuticular scale pattern, cortex width, and medullary configuration have been used to identify mammals to the species level (Mayer, 1952; Tumlison, 1983). Using Burt and Grossenheider (1980) and Schwartz and Schwartz (1959), I compiled a comprehensive list of mammals that as juveniles or adults might serve as prey for Timber Rattlesnakes and with ranges that overlapped or approached my study site (Appendix 1). Guided by this list, a reference set of microscope slides was constructed using mammal skins from the University of Arkansas Museum Collections. Prey identification was accomplished by comparing slides of hair recovered from feces to the reference slides of hair obtained from mammals of known identity. Additionally, 51% of dietary samples (including feces) contained undigested mammal skeletal fragments and/or teeth. Cranial components and teeth greatly facilitated prey identification. Mammal Trapping.— In order to determine if prey abundance could explain the use of disturbed habitat by Timber Rattlesnakes, small mammals were trapped in woodland, field, and edge habitats during periods of the active season when snakes were typically foraging. During July and August of 2005,1 conducted four trapping sessions to compare mammal abundances in woodland habitats to those in fields and edges. Trapping sessions one and three directly compared woodland tracts to uncut grassy fields. Trapping sessions two and four directly compared woodland tracts to wooded fencerows (edge habitat). A trapping session consisted of simultaneously placing 60 traps in each of 111 the two habitats for three nights. Two additional trapping sessions were carried-out in June of 2006, in which 40 traps were placed in all three habitats simultaneously. To better understand small mammal diversity at the study site, specific locations were never trapped more than once throughout the study. Therefore, six woodland tracts, four uncut grassy fields, and four wooded fencerows were sampled during the six trapping sessions. Trapping grid configuration was rectangular, and although the shape of the rectangle varied to accommodate the habitat, the distance between adjacent traps was always 10 m. For instance, because wooded fencerows were approximately 10m wide, traps were placed in two parallel rows along opposing edges of this habitat. Consequently, grid configurations in edge habitat were long and narrow compared to those in woodland tracts and fields. Non-folding Sherman live traps were opened in the evening just before sunset and closed in the morning after being checked. Closing traps during the day prevented trapping mammals when they could be subjected to lethal temperatures. Traps were kept baited with a mixture of oats and peanut butter. Captured mammals were identified to species, weighed, and given a batch-mark before being released. The activity patterns of small mammals can vary on a temporal scale and are affected by weather and moonlight (Stokes et al., 2001). Therefore, because all three habitats were not trapped simultaneously during trapping sessions one through four, capture data were not pooled across all six trapping sessions by habitat type. Instead, habitat-specific capture data were first compared independently within each of the six trapping sessions. A total of 232 capture events were recorded during the study, 85 (36.6%) of which were recaptures of previously marked individuals. Only data 112 pertaining to first-captures are reported and analyzed herein. Contingency table analyses (a = 0.05) tested the null hypothesis that the number of small mammals captured in fields and edges would not differ from those captured in woodlands. Because only two habitats were trapped simultaneously during 2005 (trapping sessions 1 - 4), capture data were analyzed using 2x2 contingency tables and the Fisher exact test (Zar, 1999). Three habitats were trapped simultaneously during 2006 (trapping sessions 5, 6) and capture data were therefore analyzed using 3 x 2 contingency tables and the Chi-square test (Zar, 1999). Contingency table analyses were performed using JMP version 8.0.1 (SAS Institute). RESULTS Observations of Foraging Behavior.— A total of 627 telemetry relocations of Timber Rattlesnakes occurred in habitats classified as either woodland (n = 69), field (n = 79), edge (n = 65), or other (n = 414). Relocations of snakes inside the quarry (n = 401) account for the large number of relocations classified as occurring in "other" habitat. Within the quarry, snakes were often relocated in boulder piles (transient habitat and birthing rookeries), rocky slopes and outcroppings (hibernacula), and to a lesser degree, anthropogenic material including old tires and scrap metal. In woodland, field, edge, and other habitats, foraging was observed in 53.6, 48.1, 66.2, and 1.4 percent of relocations, respectively. Although snakes were given sufficient time to move into a different habitat between relocations (typically days), these data do represent repeated observations of twenty-six individual telemetry subjects. In fact, snakes did exhibit individual tendencies with respect to where they foraged. Nevertheless, among fifteen individuals that were 113 observed foraging more than once, thirteen foraged in more than one habitat type (Table 1). Despite a repeated-measures bias, the telemetry data clearly show that snakes in the study population forage in woodland, field, and edge habitat. Dietary Analysis.— Forty-one mammalian prey items were identified from 37 dietary samples (Figure 1). Prairie Voles (Microtus ochrogaster) comprised 41.5% of the identified prey. The remaining prey consisted of 24.4% mice in the genus Peromyscus, 14.6% Cotton Rats {Sigmodon hispidus), 9.8% shrews (Soricidae), 4.9% Eastern Cottontail Rabbits (Sylvilagus floridanus), 2.4% Eastern Gray Squirrels (Sciurus caro linens is), and 2.4% unidentified mammal (owing to a small and degraded hair sample). As would be expected, larger prey such as rats, rabbits, and squirrels, were only consumed by large snakes. However, smaller prey such as shrews, mice, and voles were not excluded from the diet of larger snakes (Figure 2). Sexual differences in types of prey consumed were not apparent. Although shrews were only recovered from females, this is most likely an artifact of small sample size (n = 4). Four samples contained two prey items. A single roadkilled specimen contained a relatively intact Prairie Vole in its stomach and Cotton Rat hair in its large intestine. Three other snakes had each consumed two Prairie Voles, as was evident by the presence of two distinct skulls. Other items recovered from feces and ingesta included straw-like vegetation, exoskeletons of small insects, and shed fangs. The vegetation was likely inadvertently swallowed during intraoral prey transport. Secondary consumption would explain the presence of insect exoskeletons, as these were recovered with Peromyscus hair. Mice in the genus Peromyscus are known to occasionally prey upon insects (Burt and Grossenheider, 1980). 114 No fewer than nine dietary samples contained shed fangs (24.3%), and as many as six shed fangs were recovered from a single sample. Mammal Trapping.— In 2005, two trapping sessions (1 and 3) compared small mammal captures in woodlands to those in fields (Table 2). Woodland habitat produced significantly more captures than field habitat during session three (P = 0.0355) but not during session one (P = 0.8187). Similarly, two trapping sessions (2 and 4) comparing captures in woodland and edge habitats were also split with respect to their statistical significance (Table 2). Edge habitat produced significantly more captures than woodland habitat during session four (P < 0.0001) but not during session two (P = 0.6385). In 2006, two trapping sessions (5 and 6) compared woodland, field, and edge habitats simultaneously (Table 2). Significant differences in captures among the three habitats were detected in both sessions five (P = 0.0029) and six (P = 0.0006). During both sessions the number of captures in woodland and edge habitats was similar, however, woodlands produced markedly more captures than fields. Viewed collectively, the contingency table analyses indicate that woodland and edge habitats did not differ in three of the four sessions where both habitats were trapped simultaneously (Table 2). By contrast, in three of the four sessions where both woodland and field habitats were trapped simultaneously, woodlands produced more captures (Table 2). On the basis of trapping effort, edge habitat was the most productive (75 captures, 600 trap nights, capture rate = 0.1250), followed by woodlands (57 captures, 960 trap nights, capture rate = 0.0594) and fields (15 captures, 600 trap nights, capture rate = 0.0250) (Table 3). However, habitat-specific capture rates obtained by pooling 115 data across trapping sessions (i.e. time periods) should be interpreted cautiously, given that field and edge habitats were not trapped simultaneously in sessions 1 through 4. Mice in the genus Peromyscus (P. leucopus and P. maniculatus) accounted for the majority of both woodland (96.5%) and edge (61.3%) captures (Table 3). A second mammal, the Cotton Rat, was also prevalent in habitat edges (32.0%) (Table 3). Despite a paucity of captures, fields produced the greatest number of species (n = 6) (Table 3). Prairie Voles were the most frequently captured mammal in fields (40.0%), the only habitat in which they were detected (Table 3). DISCUSSION Observations of Foraging Behavior.— Studies conducted in mature woodlands suggest that logs and trees are important to foraging Timber Rattlesnakes. In Pennsylvania, Reinert et al. (1984) described snakes coiling tightly beside logs on the forest floor. From this position, snakes placed their chins on the lateral surface of the log in order to ambush woodland mammals using the log as a runway. The authors indicated that this was the "typical" ambush posture for snakes in the population and suggest that "fallen logs play a functional role in the predatory behavior of this species." On several occasions in northeastern New York, Brown and Greenberg (1992) documented an adult male rattlesnake adopting an upward-facing ambush posture at the base of trees. Using tongs to slide a freshly killed Eastern Chipmunk (Tamias striatus) down the tree towards the snake, the authors watched as it quickly struck the prey from this position. The same vertical-tree ambush posture has been observed in south-central Indiana, where adult timber rattlesnakes select microhabitats based on the presence of tree trunks and log 116 cover (Walker, 2000). In South Carolina, 31% of female foraging took place near logs, while 23% of male foraging occurred at the base of trees (Waldron et al., 2006). The ambush posture described by Reinert et al. (1984) was not prevalent among snakes in my study population. On a single occasion, I observed a large adult male adopt this posture on a man-made pile of woody debris. Secondary woodlands and wooded fencerows typically lacked sizeable deadfall, which may provide a simple explanation as to why snakes were not observed lying in ambush beside logs. Although mature forest containing significant deadfall was present at the study site, few telemetry relocations occurred in this habitat. Interestingly, this mature forest was the only woodland habitat that failed to produce a single capture when trapped for small mammals (trial 4, Table 2). Furthermore, snakes were never seen foraging at the base of trees as described by Brown and Greenberg (1992). As perhaps the most arboreal members of the genus Crotalus, the tendency for Timber Rattlesnakes to climb trees is well documented (Klauber, 1972; Saenz et al., 1996; Coupe, 2001; Fogell et al., 2002b; Sealy, 2002; Bartz and Sajdak, 2004; Rudolph et al., 2004; Sajdak and Bartz, 2004). Although subjects are known to engage in basking, shedding, and even courtship while in trees (Fogell et al., 2002b; Bartz and Sajdak, 2004), Rudolph et al. (2004) suggest foraging may provide the most general explanation for arboreal behavior. While both arboreal mammals and birds are consumed by this species, the only natural observation of arboreal predation involved the striking and holding of avian prey (Sajdak and Bartz, 2004). On a single occasion, I observed an adult female (77.4 cm SVL, 364 g) approximately 2 m above the ground lying elongate in an understory tree. It was not clear if this animal was ambushing prey. Such a low 117 frequency of arboreal activity among snakes in the study population may explain the lack of birds in their diet. Foraging in fields exposed snakes to the risk of being killed by farm implements. Three telemetry subjects were killed by mowers as they hunted in grassy and old fields, while the plowing of a fallow crop field resulted in the death of a fourth individual. In some instances, foraging snakes would spend much of the active season in a single field where they moved very little between successive relocations. Intuitively, snakes making intermittent forays into fields were less likely to encounter a tractor than those largely dedicated to a single field throughout the summer. Although Furman (2007) reports an anecdotal account of Timber Rattlesnakes being killed during the haying of a Vermont dairy farm, I am unaware of other studies where telemetry subjects have been killed by farm implements. However, a telemetry study in northwestern Missouri reported that mowing resulted in a 43% mortality rate for grassland dwelling Massasauga Rattlesnakes {Sistrurus catenatus) (Durbian, 2006). Clark (2006a) used fixed videography to study the predatory behavior of ambushing Timber Rattlesnakes in the forests of New York. The author found that 81% of feeding events occurred at night and estimated that snakes fed successfully 12 to 15 times per year. During the present study, snakes were never observed consuming prey. Telemetry subjects were found visibly distended at midbody on five different occasions, indicating they had recently ingested prey. As most telemetry relocations occurred during the day, the lack of observed prey handling supports Clark's findings that prey capture typically occurs at night. If Clark's estimates of feeding frequency are accurate, it might seem that snakes with visible body distensions should have been encountered 118 more often. However, small prey may not be visually detected when consumed by relatively large snakes. Furthermore, concealing surface vegetation often made diagnosing snake behavior difficult. In Wisconsin, Keenlyne (1972) found that during the period of time between ovulation and parturition, female Timber Rattlesnakes remained near large rocks and probably did not feed. Similarly, gravid females in this study gestated exclusively within the rock quarry while other members of the population dispersed to forage in surrounding habitats. During 123 relocations of gestating females (n = 10), predatory behavior was never definitively observed. Gravid females remained beneath rocks or within boulder piles much of the time. Although snakes were sometimes found coiled tightly on the surface near their chosen maternal rock, in most instances it seemed unlikely that they were ambushing prey. At least some females in the study population forage during seasons in which they reproduce, prior to gestation and/or following parturition (Wittenberg, chapter 2). Successful spring foraging prior to gestation, coupled with a long fecal retention time characteristic of vipers (Lillywhite et al., 2002), might explain why hair-containing feces were collected from seven gravid females (Appendix 2). Although Keenlyne (1972) dissected 23 gravid females and reported finding no prey items, it is unclear if the author examined the lower digestive tract for feces. Dietary Analysis.— Recovering only mammalian prey in the feces and digestive tracts of snakes from west-central Missouri is consistent with other dietary studies of the Timber Rattlesnake. Regional studies in Pennsylvania (Surface, 1906; Reinert et al., 119 1984), Wisconsin (Keenlyne, 1972),Virginia (Uhler et al., 1939; Smyth, 1949), Georgia (Hamilton and Pollack, 1955; Parmley and Parmley, 2001), Tennessee (Savage, 1967), and Arkansas (Burnett, 2001; Montgomery, 2005) found that dietary samples contained between 73.1 and 100% mammalian prey. Clark (2002) examined the diet of the Timber Rattlesnake throughout its range by reviewing 400 published dietary records, including records from the aforementioned studies, and dissecting 1108 museum specimens (resulting in the recovery of 189 prey items from the stomachs of 178 individuals). Clark (2002) found that mammals comprised 91.1% of the total diet, followed by birds (7.2%), reptiles (1.2%), and amphibians (0.3%). Although mammals belonging to 22 different genera were consumed, Clark (2002) only considered those that comprised more than 1% of total prey records to be of dietary importance. These 11 genera consisted of Peromyscus (33.3%), Microtus (10.9%), Tamias (10.6%), Sylvilagus (10.4%), Sigmodon (5.3%), Sciurus (4.2%), Clethrionomys (3.4%), Napaeozapus (2.6%), Sorex (2.2%), Mus (2.2%), and Blarina (1.2%). Birds were the second most commonly recovered prey in nearly every study. In the George Washington National Forest of Virginia, birds comprised as much as 13% of the Timber Rattlesnake's diet (Uhler et al., 1939). Interestingly, specimens examined from Fort Benning, Georgia contained 73.1% mammals, 23.1% lizards, and 7.7% birds (Hamilton and Pollack, 1955). Piatt et al. (2001) reviewed published dietary data for the species in the south where it was once considered a distinct subspecies of the Timber Rattlesnake known as the Canebrake Rattlesnake, Crotalus horridus atricaudatus. The authors found that the diet of southern horridus consisted of 80.4% mammals, 13.0% birds, and 6.5% reptiles. However, it should be noted that the 6.5% reptilian contribution 120 to the Timber Rattlesnake's diet reported by Piatt et al. (2001) consists solely of the lizard consumption data previously reported in Hamilton and Pollack (1955). As snakes increase in size, they often add larger prey to their diet (Mushinsky, 1987). With this upward shift in prey size, some species exhibit a concomitant refusal to target smaller prey (Reynolds and Scott, 1982; Arnold, 1993; Shine and Sun, 2003). In this study, the three larger prey types (rats, rabbits, and squirrels) were in fact taken by snakes greater than 70.0 cm SVL. However, snakes greater than 80.0 cm SVL continued to take shrews, mice, and voles (Figure 2). Clark (2002) also confirmed this trend regarding the size of Timber Rattlesnakes and the prey they consume. The willingness of large snakes to consume both large and small prey may be an important factor contributing to this species ability to utilize anthropogenic habitat. Many New World pitviper species prey on ectotherms as juveniles and later shift to endothermic prey as adults (Campbell and Lamar, 2004). I examined dietary material from 7 young Timber Rattlesnakes ranging between 31.0 and 66.3 cm SVL and recovered 1 shrew, 3 mice, 2 voles, and 1 unidentifiable mammal. Both voles were taken by the largest two individuals of this group (60.0 and 66.3 cm SVL). Clark (2002) also found the diet of young Timber Rattlesnakes to consist entirely of small mammalian prey. Consistent with the findings of this study, scant dietary records from other Timber Rattlesnake populations using disturbed, agricultural habitats indicate field-dwelling voles are important prey for such populations. Snakes in northeastern Kansas have been shown to forage in secondary woodlands, grasslands, and old fields (Fitch, 1982, 1999; Fitch and Pisani, 2006; Fitch et al., 2004; Pisani and Fitch, 2006). Of seventeen prey items recovered from snakes in northeastern Kansas, Prairie Voles (n = 6) comprised the 121 largest dietary contribution of any single prey species (Fitch and Pisani, 2006). Pisani and Fitch (2006) suggested that high Prairie Vole densities contributed to the rapid growth rates of some snakes in the population. In Southeastern Nebraska, Timber Rattlesnakes have been shown to leave wooded areas to forage in fields, including those planted in row-crops (Fogell, 2000; Fogell et al., 2002a). Fogell (2000) reported that thirteen identified prey items from this Nebraska population included one Prairie Vole and three Meadow Voles (Microtus pennsylvanicus). Mammal Trapping.— Small mammals were captured with greater success in edge and woodland habitats than in fields (Table 3). Given the degree of field use by foraging snakes in the population, as well as the importance of Prairie Voles in their diet, low capture rates in fields were surprising. Because Timber Rattlesnakes typically consume small mammals in proportions similar to the local abundance of those species (Reinert et al., 1984; Clark, 2002), it is unlikely that Prairie Voles were rare at the study site. In nearby northeastern Kansas, Fitch (1957) found that Prairie Voles were the most abundant mammal in grassland areas. Martin (1956) stated that Prairie Voles were perhaps the most important species of mammal in the grasslands of Kansas and neighboring states. Nevertheless, possible explanations as to why Prairie Voles were not captured in large numbers may be related to the natural history of the species. First, the dedicated use of well-established runway systems by Prairie Voles may limit trap encounters (Stokes et al., 2001). Second, in addition to surface runways, the semifossorial Prairie Vole may spend much of its time in an elaborate system of tunnels, some of which are entirely below ground (Jameson, 1947). 122 Conclusions and Management Implications.— Other studies have effectively combined telemetry, dietary, and prey sampling data in order to validate (Heard et al., 2004) or potentially refute (Carfagno et al., 2006) foraging-related hypotheses that explain the use of disturbed habitats by snakes. The present study underscores the importance of using all three methods, because if viewed alone, the mammal trapping data would indicate that fields are poor resource environments for snakes at the study site. However, both telemetry observations of snakes foraging in fields and the frequent occurrence of Prairie Voles in the population's diet suggest otherwise. Snakes clearly acquired small mammal prey in a diversity of habitats that included secondary woodlands, overgrown fields, and habitat edges. Arboreal foraging and ambush postures that utilize logs and tree trunks were not prevalent in the population, and individuals often foraged in areas with little or no canopy closure from overstory trees. Therefore, it would seem that prey availability may be more important to this species than the physical structure of the habitat in which it forages. However, snakes are subjected to the risk of being killed by farm implements when they remain static in fields for long periods while ambushing prey. Furthermore, snake activity within fields may be constrained by warm midday temperatures (Wittenberg, chapter 4). Near the western edge of their geographic range, Timber Rattlesnakes appear to forage readily in anthropogenically fragmented landscapes (Fitch, 1999; Fitch and Pisani, 2006; Fitch et al., 2004; Pisani and Fitch, 2006; Fogell, 2000; Fogell et al., 2002a; this study). However, as suggested by Fogell (2000), populations in geographic regions historically containing grassland may be pre-adapted to using agricultural habitats. Therefore, can the findings from this study be applied to forest-dwelling populations 123 threatened by the loss and fragmentation of woodland habitat? Is there any evidence that Timber Rattlesnakes in closed-canopy forests will venture into more open, disturbed areas created by anthropogenic processes? In Virginia during the first half of the 20 century, Timber Rattlesnakes exploited high rodent densities on abandoned farms that had fallen under government control during the establishment of Shenandoah National Park and the George Washington and Jefferson National Forests (Martin, 1979). In West Virginia, Adams (2005) reported that telemetry subjects used anthropogenic habitat edges as well as clearcuts. However, a telemetry study conducted in the Piedmont region of North Carolina indicated that individuals tended to avoid open canopy habitats (Sealy, 2002). In South Carolina, one of the few detailed studies of the Timber Rattlesnake in Coastal Plain habitat revealed that both males and nongravid females prefer to forage in forests but primarily use fields during the mating season (Waldron et al., 2006). Thus, although the current study shows that large tracts of mature forest are not required for Timber Rattlesnakes to forage effectively, behavioral responses to habitat fragmentation may be population-specific. Nevertheless, given that clearcutting has been shown to increase small mammal densities within mature forests (Kirkland, 1977, 1990), silviculture practices that promote high prey densities may be an important tool for those attempting to preserve the species in woodland habitats (Beaupre and Douglas, 2009). ACKNOWLEDGMENTS This research was conducted with the approval of the University of Arkansas Institutional Animal Care and Use Committee (protocol # 05001) and the Missouri 124 Department of Conservation (collecting permits # 12005, 12367, 12715, 13101). I thank Steven J. Beaupre, Kimberly G. Smith, Ines Pinto, and Edward E. Gbur Jr. for comments on this manuscript. I am indebted to Sara Wittenberg for her diligent proofreading and words of encouragement while this work was being completed. Douglas A. James graciously allowed me to use his mammal traps. I thank the numerous private landowners in Missouri who allowed this work to be conducted on their property. This research was supported in part by a University of Arkansas Causey Grant-in-Aid Award, a Harry Steinman Memorial Grant from the St. Louis Herpetological Society, and a grant from the Arkansas Audubon Society Trust. LITERATURE CITED ADAMS, J.P. 2005. Home range and behavior of the timber rattlesnake {Crotalus horridus). Unpubl. M.Sc. thesis. Marshall Univ., Huntington. ARNOLD, S.J. 1993. Foraging theory and prey-size-predator-size relations in snakes. In R.A. Seigel and J.T. Collins (eds.), Snakes: Ecology and Behavior, pp. 87-115. McGraw-Hill, Inc., New York, NY. BARTZ, A.D., AND R.A. SAJDAK. 2004. Crotalus horridus (timber rattlesnake). Arboreality, courtship. Herpetological Review 35:61. BEAUPRE, S.J., 2008. Annual variation in time-energy allocation by timber rattlesnakes {Crotalus horridus) in relation to food acquisition. In W.K. Hayes, K.R. Beaman, M.D. Cardwell, and S.P. Bush (eds.), Biology of the Rattlesnakes, pp. 111-122. Loma Linda Univ. Press, Loma Linda, CA. BEAUPRE, S.J., AND L.A. DOUGLAS. 2009. Snakes as indicators and monitors of ecosystem properties. In S.J. Mullin and R.A. Seigel (eds.), Snakes: Ecology and Conservation, pp. 244-261. Cornell Univ. Press, Ithaca, NY. BROWN, W.S. 1993. Biology, status, and management of the timber rattlesnake {Crotalus horridus): a guide for conservation. SSAR Herpetol Circ No. 22:1-78. Society for the Study of Amphibians and Reptiles. Univ. Kansas, Lawrence. 125 BROWN, W.S., AND D.B. GREENBERG. 1992. Vertical-tree ambush posture in Crotalus horridus. Herpetological Review 23:67. BURNETT, CD. 2001. Comparative diets of syntopic pit vipers: co-existence through trophic niche separation. Unpubl. honors thesis, Univ. Arkansas, Fayetteville. BURT, W.H., AND R.P. GROSSENHEIDER. 1980. A Field Guide to the Mammals. Houghton Mifflin Co., NY. BURY, R.B. 2006. Natural history, field ecology, conservation biology and wildlife management: time to connect the dots. Herpetological Conservation and Biology 1:56-61. CAMPBELL, J. A., AND W.W. LAMAR. 2004. The Venomous Reptiles of the Western Hemisphere. Cornell University Press, Ithaca, NY. CARFAGNO, G.L.F., E.J. HESKE, AND P.J. WEATHERHEAD. 2006. Does mammalian prey abundance explain forest-edge use by snakes? Ecoscience 13:293-297. CLARK, A.M., P.E. MOLER, E.E. POSSARDT, A.H. SAVITZKY, W.S. BROWN, AND B.W. BOWEN. 2003. Phylogeography of the timber rattlesnake {Crotalus horridus) based on mtDNA sequences. Journal of Herpetology 37:145-154. CLARK, R.W. 2002. Diet of the timber rattlesnake, Crotalus horridus. Journal of Herpetology 36:494-499. . 2006a. Fixed videography to study predation behavior of an ambush foraging snake, Crotalus horridus. Copeia2006: 181-187. . 2006b. Post-strike behavior of timber rattlesnakes {Crotalus horridus) during natural predation events. Ethology 112: 1089-1094. COLLINS, J.T., AND J.L. KNIGHT. 1980. Crotalus horridus. Catalogue of American Amphibians and Reptiles 253: 1-2. COUPE, B. 2001. Arboreal behavior in timber rattlesnakes {Crotalus horridus). Herpetological Review 32:83-85. CUNDALL, D., AND S.J. BEAUPRE. 2001. Field records of predatory strike kinematics in timber rattlesnakes, Crotalus horridus. Amphibia-Reptilia 22:492-498. DODD, C.K., JR. 1987. Status, conservation, and management. In R.A. Seigel, J.T. Collins, and S.S. Novak (eds.), Snakes: Ecology and Evolutionary Biology, pp. 478-513. McGraw-Hill, New York, NY. 126 DURBIAN, F. E. 2006. Effects of mowing and summer burning on the massasauga (Sistrurus catenatus). American Midland Naturalist 155:329-334. FITCH, H.S. 1957. Aspects of reproduction and development in the prairie vole {Microtus ochrogaster). Univ. of Kansas Publications, Museum of Natural History 10:129-161. . 1982. Resources of a snake community in prairie-woodland habitat of northeastern Kansas. In N.J. Scott, Jr. (ed.), Herpetological Communities, U.S. Fish and Wildlife Service Wildlife Research Report 13:83-97. . 1999. A Kansas snake community: composition and changes over 50 years. Krieger Publishing Company, Malabar, FL. FITCH, H.S., AND G.R. PISANI. 2006. The timber rattlesnake in northeastern Kansas. Journal of Kansas Herpetology 19:11-15. FITCH, H.S., G.R. PISANI, H.W. GREENE, A.F. ECHELLE, AND M. ZERWEKH. 2004. A field study of the timber rattlesnake in Leavenworth County, Kansas. Journal of Kansas Herpetology 11: 18-24. FOGELL, D.D. 2000. Seasonal activity, habitat preferences, and natural history of the timber rattlesnake (Crotalus horridus) in southeastern Nebraska. Unpubl. M.Sc. thesis, Univ. Nebraska at Omaha, Omaha. FOGELL, D.D., T.J. LEONARD, AND J.D. FAWCETT. 2002a. Crotalus horridus horridus (timber rattlesnake). Habitat. Herpetological Review 33:211-212. . 2002b. Crotalus horridus horridus (timber rattlesnake). Climbing. Herpetological Review 33:212. FURMAN, J. 2007. Timber Rattlesnakes in Vermont and New York: biology, history, and the fate of an endangered species. Univ. Press of New England, Lebanon, NH. HAMILTON, W.J., JR., AND J.A. POLLACK. 1955. The food of some crotalid snakes from Fort Benning, Georgia. The Chicago Academy of Sciences, Natural History Miscellanea 140:1-4. HEARD, G.W., D. BLACK, AND P. ROBERTSON. 2004. Habitat use by the inland carpet python (Morelia spilota metcalfei: Pythonidae): seasonal relationships with habitat structure and prey distribution in a rural landscape. Austral Ecology 29:446-460. JAMESON, E.W., JR. 1947. Natural history of the prairie vole (mammalian genus Microtus). Univ. of Kansas Publications, Museum ofNatural History 1:125-151. 127 KEENLYNE, K.D. 1972. Sexual differences in feeding habits of Crotalus horridus horridus. Journal of Herpetology 6:234-237. KlRKLAND, G.L., JR. 1977. Responses of small mammals to the clearcutting of northern Appalachian forests. Journal of Mammalogy 58:600-609. KlRKLAND, G.L., JR. 1990. Patterns of initial small mammal community change after clearcutting of temperate North American forests. Oikos 59:313-320. KLAUBER, L.M. 1972. Rattlesnakes: Their Habits, Life Histories, and Influence on Mankind. Univ. of California Press, Berkeley. LILLYWHITE, H.B., P. DE DELVA, AND B.P. NOONAN. 2002. Patterns of gut passage time and the chronic retention of fecal mass in viperid snakes. In G.W. Schuett, M. Hoggren, M.E. Douglas, and H.W. Greene (eds.), Biology of the Vipers, pp. 497- 506. Eagle Mountain Publishing, Eagle Mountain, UT. MAC NALLY, R., AND G.W. BROWN. 2001. Reptiles and habitat fragmentation in the box-ironbark forests of central Victoria, Australia: predictions, compositional change and faunal nestedness. Oecologia 128:116-125. MARTIN, E.P. 1956. A population study of the prairie vole {Microtus ochrogaster) in northeastern Kansas. Univ. of Kansas Publications, Museum of Natural History 8:361-416. MARTIN, W.H. 1979. The timber rattlesnake in Virginia: its distribution and present status. Virginia Herpetological Society Bulletin 89:1-4. MARTIN, W.H. 1992. The timber rattlesnake: its distribution and natural history. In T.F. Tyning (ed.), Conservation of the Timber Rattlesnake in the Northeast, pp. 13-22. Massachusetts Audubon Society, Lincoln, MA. MAYER, W. V. 1952. The hair of California mammals with keys to the dorsal guard hairs of California mammals. The American Midland Naturalist 48:480-512. MONTGOMERY, C.E. 2005. Bioenergetic and ecological correlates of foraging mode in copperheads, Agkistrodon contortrix, and timber rattlesnakes, Crotalus horridus, in northwest Arkansas. Unpubl. Ph.D. Diss., Univ. Arkansas, Fayetteville. MUSHINSKY, H.R. 1987. Foraging ecology. In R.A. Seigel, J.T. Collins, and S.S. Novak (eds.), Snakes: Ecology and Evolutionary Biology, pp. 302-334. McGraw-Hill, New York, NY. PARMLEY, D., AND A.M. PARMLEY. 2001. Food habits of the canebrake rattlesnake {Crotalus horridus atricaudatus) in central Georgia. Georgia Journal of Science 59:172-178. 128 PISANI, G.R., AND H.S. FITCH. 2006. Rapid early growth in northeastern Kansas timber rattlesnakes. Journal of Kansas Herpetology 20:19-20. PLATT, S.G., A.W. HAWKES, AND T.R. RAINWATER. 2001. Dietofthecanebrake rattlesnake {Crotalus horridus atricaudatus): an additional record and review. Texas Journal of Science 53:115-120. QUINN, H., AND J.P. JONES. 1974. Squeeze box technique for measuring snakes. Herpetological Review 5:35. REINERT, H.K. 1984. Habitat separation between sympatric snake populations. Ecology 65:478-486. REINERT, H.K., AND D. CUNDALL. 1982. An improved surgical implantation technique for radio tracking snakes. Copeia 1982:702-704. REINERT, H.K., D. CUNDALL, AND L.M. BUSHAR. 1984. Foraging behavior of the timber rattlesnake, Crotalus horridus. Copeia 1984:976-981. REINERT, H.K., AND R.T. ZAPPALORTI. 1988. Timber rattlesnakes {Crotalus horridus)^ the pine barrens: their movement patterns and habitat preference. Copeia 1988:964-978. REYNOLDS, R.P., AND N.J. SCOTT JR. 1982. Use of a mammalian resource by a Chihuahuan desert snake community. In N.J. Scott, Jr. (ed.), Herpetological Communities, U.S. Fish and Wildlife Service Wildlife Research Report 13:99- 118. RUDOLPH, D.C., S.J. BURGDORF, R.N. CONNER, AND J.G. DICKSON. 1998. The impact of roads on the timber rattlesnake, {Crotalus horridus), in eastern Texas. In G.L. Evink, P. Garrett, D. Zeigler, and J. Berry, (eds.), Proceedings of the International Conference on Wildlife Ecology and Transportation, FL-Er-69-98, pp. 236-240. Florida Department of Transportation, Tallahassee. RUDOLPH, D.C., R.R. SCHAEFER, D. SAENZ, AND R.N. CONNER. 2004. Arboreal behavior in the timber rattlesnake, Crotalus horridus, in eastern Texas. Texas Journal of Science 56:395-404. SAENZ, D., S.J. BURGDORF, D.C. RUDOLPH, AND CM. DURAN. 1996. Crotalus horridus (timber rattlesnake). Climbing. Herpetological Review 27:145. SAJDAK, R.A., AND A.W. BARTZ. 2004. Crotalus horridus (timber rattlesnake). Arboreality, diet. Herpetological Review 35:60. SAVAGE, T. 1967. The diet of rattlesnakes and copperheads in the Great Smoky Mountains National Park. Copeia 1967:226-227. 129 SCHWARTZ, C.W., AND E.R. SCHWARTZ. 1959. The Wild Mammals of Missouri. Univ. of Missouri Press, Columbia. SEALY, J.B. 2002. Ecology and behavior of the timber rattlesnake {Crotalus horridus) in the upper Piedmont of North Carolina: identified threats and conservation recommendations. In G.W. Schuett, M. Hoggren, M.E. Douglas, and H.W. Greene, (eds.), Biology of the Vipers, pp. 561-578. Eagle Mountain Publishing, Eagle Mountain, UT. SHINE, R., AND SUN, L.-X. 2003. Attack strategy of an ambush predator: which attributes of the prey trigger a pit-viper's strike? Functional Ecology 17:340-348. SIMS, P.L. 1988. Grasslands. In: M.G. Barbour and W.D. Billings (eds.), North American Terrestrial Vegetation, pp. 265-286. Cambridge Univ. Press, New York, NY. SMYTH, T. 1949. Notes on the timber rattlesnake at Mountain Lake, Virginia. Copeia 1949:78. STOKES, M.K., N.A. SLADE, AND S.M. BLAIR. 2001. Influences of weather and moonlight on activity patterns of small mammals: a biogeographical perspective. Canadian Journal of Zoology 79:966-972. SURFACE, H.A. 1906. Snakes and turtles of Pennsylvania. Bulletin of the Pennsylvania State Department of Agriculture 1906-1908. TUMLISON, R. 1983. An annotated key to the dorsal guard hairs of Arkansas game mammals and furbearers. The Southwestern Naturalist 28:315-323. UHLER, F.M., C. COTTAM, AND T.E. CLARKE. 1939. Food of snakes of the George Washington National Forest, Virginia. Transactions of the Fourth North American Wildlife Conference 605-622. WALDRON, J.L., J.D. LANHAM, AND S.H. BENNETT. 2006. Using behaviorally-based seasons to investigate canebrake rattlesnake {Crotalus horridus) movement patterns and habitat selection. Herpetologica 62:389-398. WALKER, Z.J. 2000. The spatial ecology of the timber rattlesnake {Crotalus horridus) in south central Indiana. Unpubl. M.Sc. thesis. Purdue Univ., Fort Wayne. ZAR, J.H. 1999. Biostatistical Analysis. 4th ed. Prentice Hall, Upper Saddle River, NJ. 130 Table 1. Records of habitat use and foraging behavior within each habitat for twenty-six radio-tagged Timber Rattlesnakes (13 males, 13 females). Values represent the number of telemetry relocations occurring in each habitat type, values in parentheses represent the number of telemetry relocations where foraging was observed. "Other" relocations describe those occurring in habitats that could not be characterized as woodland, field, or edge. 131 so t-- ©, O —i — C" ° o o o o o OS i/-> so o r- — m so — SO m o o Os so — CM CO -^ ^ r- o ,—^ ? ,—i >— r- so oo OS -H ro an c 'C.J O O M^ ^ Irf ^/ ^ Srf •** *1^ ^^ **^ ^ 'W *4^ j EEEEEEEEEEEEEJJ CQ <• < OQ < CQ o o O o O M o o o O O o ^ w o o ON ON ON OO r- CN CN CO CN 41 4 w « CO ON ON 1 NO u E C A < m is ca -^ — — ON o ^ O ••=• HmcicNvocN'—ivO° 133 Table 2. Results of contingency table analyses testing the null hypothesis that the number of small mammals captured in fields and edges would not differ significantly from those captured in woodlands. Because only two habitats were trapped simultaneously during 2005 (trapping sessions 1 - 4), capture data were analyzed using 2 x 2 contingency tables and the Fisher exact test. Three habitats were trapped simultaneously during 2006 (trapping sessions 5, 6) and capture data were therefore analyzed using 3x2 contingency tables and the Chi-square test. 134 * * 3 in © * *00 0 ON cO © NO © .638 5 03 5 > .818 7 00 2 © © a, o © o © V II ll o o o o cO cO CO CO w GO S- ^ n-c a >. o aneous l taneous nr a CO GO simu l — © (N OO in © 3 ^ ON OO -1 CO CO T} Tl TJ Tj ha b Tl Tj hab i ft G fl fl C d c o +-> CO T> CO a> CO T) CO (i) CO Tl « CO Tl o Tj £ Ed g Fie l Ed g Ed g Fie l Fie l Fie l ood l ood l ood l ood l ood ] ood l Habita t £ £ £ £ £ £ Trappin g (t w Trappin g 200 5 P. cO 200 6 GO •;? o © © © © © f^-S NO NO NO NO Ti ra eg •* rVH ^ H -a o -*-» -*-^ J? ^ ne , ne , 3 NO 3 NO ^ © ^ © H-, o >-> © Augu s Augu s 00 5 ON o 00 5 r-oHo (No) CN (N ON in 1 ^T Three- d 24 - 6- 0 appin g p VH © <—1 °fl - fl S • " o 135 Table 3. Number of individuals from each of 7 different small mammal species captured in woodland, field, and edge habitats. Values in parentheses represent the percent contribution made by that species to the total number of first-captures recorded in that habitat type. The number of trap nights in each habitat represent trapping effort, and were calculated by multiplying the number of traps deployed by the number of nights trapped. Capture rates were calculated by dividing the number of first-captures by the number of trap nights. 136 •* o o ON i/"> m m CN OJ u o o — o o © o 3 H so - t/> O O cu.« o .2Ml NO O O H,»- '5 ON NO VC So P o _: o •S S CO >! O Is o s 1 a, S O NO NO >/1 ~ 3 — "3- o CQ BO •o PH w 137 Figure 1. Total number of each type of mammal prey recovered from Timber Rattlesnakes in the study population. Identified prey included shrews (Soricidae), mice in the genus Peromyscus, Prairie Voles (Microtus ochrogaster), Cotton Rats (Sigmodon hispidus), an Eastern Gray Squirrel (Sciurus carolinensis), and Eastern Cottontails {Sylvilagus floridanus). 138 0> Q. H Qj 1- so|duiB§ ui paj3A033^ Joqmn\[ 139 Figure 2. Relationship between snake size (SVL) and type of mammal prey consumed, based on 33 dietary samples collected from individuals of known identity. Prey are ordered on the Y-axis by ascending average adult body mass. 140 o CM o O o O o o 0 o 0 o o o o £ 8 00 u 8 o o s—' JS 8 o •4-i WD S3 V - +* o o CD •Ye n -*j 3 0 O a en o snu: &3 S v M^V nsis ts »3 "§ s> "a .e Z5/7 Z rid 5< ^^ 'CW 5 o o ^ c/jr t 3 <*> 1 3 u -§ o o »3 o »0 w £ g bfl O a. •^^ _s !v ^^^» u to O c^ co pauinsuo^ jfojj 141 APPENDIX 1 - Tabular listing of mammal species that 1) as juveniles and/or adults could possibly be prey for Timber Rattlesnakes and 2) have geographic distributions that overlap or come very near the study site. Species verified by means other than small mammal trapping are denoted in the middle column. The right-hand column denotes which species were captured using Sherman live traps and the habitat(s) in which they were caught. 142 u u o 60 60 60 •a •o T3 Id ) u t> U CD 43 •a -a -a Uo u u u o 60 w W SB •o lan d •a" -a" •o" -a c c o ra§ ra cua >^> X o -a -a T3 o o O o o o X X £ s £ XX X l£\ a X X X « II51 o > '« tu b 3 CT" CO -a c a 3 o ^ E 2 irre l dg r t -3 u p o g^ .s <*• £? 60 -a le ) re w £? -C § s C •£ o 13 © 3 E M O '3 « • 5- E a- c w C 60 $ Cfi x « C 5 o shrew ) ^ tus (thir t horttai l s Easter n m tai l weas > nkli n gr o ,o -5 i "3 F, w. f= ° 'S o 3 ca --s S 52- CO Co ca c .2 J3 W a S3 c w-5 •—- to I S r£ <$ Co 2 .00 B a ^ tridecemlinea (squirre l fam i (shre w famil y brevicauda ( s 3 O frenata (lon g franklini (Fra : esh-eaters ) I is parva (leas t (mol e family ) a aquaticus ( E : vison (mink ) iwin g mam m .3 fopw da e i elida e (wease l fa m =§•1 5 g ra( i 1 co Co h ida e T3 >5 *5 B Citellus Citellus 1V0 1 Se a vo r ori c cm r 3 O 5 ^ ll| Mus t +* GO tH s enti a to • 1 • • uCAeu •o a^- o a U « 143 t/3 P. cs s- X o X> a cd u J3 CO 60 C t/3 3 •a u B»- &cd o cfl U O u C&O . •o Io- 3 a. cd o o c 3 -D T3 o IC •oc C>O O o a) a< CO -2 CW •>a% 3 c« ca c XA> cd "a? to c 3 ^ u O cd o g C G 60 C/l co c +J o u r , ±2 o >.'& , 15 o o o E o o. 'E t« 3 s^ d r a fam i (ju m *O.3 W^* s 0> 3 ^ Cd a cs -&«3 d -o di d 3 eu-c 1 o > ou no , & s< Xto N 3 O M M < -«J 144 APPENDIX 2 - List of dietary material examined including sample ID, sex of snake from which the sample was collected, reproductive condition of snake, snout-vent length of snake in centimeters, mass of snake in grams, approximate date when the dietary material was collected, and the prey identified in the sample. 145 ^ K ^ s^ § -CU -22 •82 CD e2 c S ^3 "CO en 'in e w B 55 B 3 a :§ tX) 60 tw -idst e ^ si C d d d d d d d s? d d £! w ja en -sg: -sg: en en en -s en cn -sg: en en o tn =o O U raysc lagusft OT^SC otus otus mjtt c my.s e my.s e mys c O mj'i c :ida e mj'i c otus . mysc myj c lagusft :ida e o S O o o o o o c^ o o o 2 ej ricida e ^cj ^Cj v i< s^ ^cj 'C ^cj i. o ^0) ^1) cu t^o CD ^ o e=u< c»a. *o^ & § ^ ^ Sigm a. ^ oo 00 a. s§ £ 00 1 NO NO •t en o O o o o c^l O o o NO CN o^t- O CN CN CN o^r o NO o in NO NO m b r^ o •* -^- •* U cu o oCN o Xi O o CN O o o CN o Xi Xl X) o CN o O p o en o o o en iber : CN 3 CN CN CN 3 CN CN CN 53 3 CN ud y i ft.o d u 8- o m in Jun e Jul y Jun e Ma y Au g Jul y Oct c Jun e Aug i Jun e Aug i GO Ma y Ma y t/3 oo o ON •* ON o ON NO ^- NO m oo •t o m r~- m oo .^ o r~ o NO CN © CN CN o CN O o o CN CN CN (N o CN OS CN cn co r^ O O in in in in in o oo CN co o •* m o O © - CN •*' ON NO r^ O ON o rn •* CN •*' ON •* C3^-N •<* oo NO O CN NO •* r- j NO CN in NO NO CN o NO in m ON m •* r- cn ^r 1 TT m CN t- NO CN O C— •* O M- NO O o m o m r^ (-; —' oo CN CN © ON K o NO OO ON ON ON 1 rn OO O | o ON ,—t r- r- ON oo in t~- r- 1 oo r- NO i oo r-~ in „ 3 CU 1) *3 -a -o T3 3 -o -a -o o ni l ni l •a ni l tu CL> ca -caa i> > 3 3 3 3 C>3 -O > 3 > 3 3 > L-ca< ca X> £ .>3 -o *o -a -a .>3 3 -a 1 v-. T3 1 3 1 -a 5i« 3> b;-o. CO ca ca ca !0~0 en 0B ca 1 bo ca 1 w 1 ca bo •.—i ON CO NO t- ON (N O-) •& "J") NO r- oo o CN •* m oo o o o O o o O O o o o CN "5. ca ca cC ca ca ca C3 ca ca ca ca ca ca ca ca ca ca ca ca ca E o o o y o o c_> o o o o o o o o o CJ o o o 1 C^H c+^ c+-< e" C4^ e^ e^ J " C+H £ c+^ c+- c+^ c+* 146 ^ 'o' o •S * i. s< s< Co CO CO -o •SJ c^o co ^ b 5J CJ t? t? co co CO CO CO to CO 3 .3 .3 3 g co IS a "a^ s; a 73 a S>0 « bo 73 to bo "a bo /3 >3 bO & •S p p g g §• g §• g .C^O -sC -a .oa .'1o - ,co -s: s.. cj cj -o« cj u rol -*: -5 •3 •S p -5 •S o o o s; a o C s; s; Co CO ^ s cj 52 S3 co o o o p o jS jS co 73 73 3 73 ^a o Cs -«o ^o P p O o -§• .1 o Sawri i Sigm Sigm Mien Sigm Mien Sigm Mien Mien Mien ^ :§ i§ i§ .Co : -* SO o © © o © O en CN 7f CN CN © © © 7f CO o co in CD •* SO CJ r~ •* CJ o CN o o jO o CN © J2 o o © JO O o o P o -t_. © o o © CN .(_CO, CN CN CN CN 1 CN CN CN 3 CJ 3 U 1 CJ CJ U U bo >s •tern c p >-§ . p 3 c^a cd c>a^ c SJ ^3 3 CJ 3 SO 3 CJ Aug i t/3 2- 8- t/3 s < s © s ON in OS r~ CO SO OS •* OS OO © CN CN r~ © o CN © CN CN CN CN ^-< © OO so r- OO t-; SO OS OS © in in © © in in CN so OS © CN K in TT c~- CO o•* 1 o 1 OS in so SO 1 O—O in 1 SO p SD o p m •—' © in CO r^ *cr od M: Os © so GO K I o ^H CO ^OO oo so OO r-~ OO r- 1 Os 73 C O •a 3 -a -o •I-. ^ <-. p cd > 3 -O 3 3 3 > 3 3 3 3 -3 3 ^ i- 73 3 -o •a •a u -a -a -a bcao a CO ca cd ca 6ca0 ca cd cd T3 -O 73 S rrt m rri .r±. u O CJ o CJ CJ ca CD ca ca cd cd ca ca CO ca ca cd cd cd £ cj o o a o o a o o o cj CJ cj cj CJ cj CO CJ CJ .CJ 45 {Crotalus horridus) IN A FRAGMENTED AGRICULTURAL LANDSCAPE Abstract Changes in structural habitat (i.e. canopy cover) affect the physical properties of the earth's surface and affect microclimates greatly. Because reptiles derive their body temperature (Tb) from the surrounding physical environment, the effects of habitat fragmentation on their behavior and physiology may be pervasive. Nevertheless, reptiles remain underrepresented in studies of habitat fragmentation. The purpose of this study was to investigate potential effects of forest fragmentation on the thermal ecology of timber rattlesnakes (Crotalus horridus) in an agricultural region of west-central Missouri. Median snake T\, did not differ among woodland (median = 28.0 °C, range = 16.4 - 34.2 °C, n = 63), edge (median = 28.6 °C, range = 16.7 - 33.9 °C, n = 56), or field (median = 28.2 °C, range = 20.2 - 34.4 °C, n = 61) habitats. During each two week interval of the sampling period (02 June to 26 August 2006), the overall percentage of thermally available habitat was high in both woodland (range 84.2 - 99.8 %) and field (range 78.1 - 94.9 %) habitats. However, sharp decreases in habitat availability occurred in both the woodland and field during midday. The lowest hourly percentages of thermally available habitat were recorded at 1400 hours in the field (19.7%) and at 1500 hours in the woodland (36.7%) during the two-week interval from 29 July to 11 August. Body temperatures of gravid females gestating within a man-made quarry (median = 32.5 °C, range = 25.3 - 34.7 °C, n = 113) were significantly warmer and less variable than Jb's of males and non-gravid females using woodlands, edges, and fields (median = 28.3 °C, 148 range = 16.4 - 34.4 °C, n = 180). Although direct effects of forest fragmentation on the thermal biology of timber rattlesnakes were not detected, results indicate that warm midday temperatures may place considerable constraints on snakes foraging in field habitat. Introduction Habitat loss and fragmentation have been identified as important factors in the global decline of reptile populations (Gibbons et al., 2000). Nevertheless, studies of habitat fragmentation and its impact on wildlife tend to target birds and mammals (Mac Nally and Brown, 2001; Heard et al., 2004). Such biases are disconcerting, given that reptiles possess two attributes that make them especially vulnerable to changes in structural habitat that are not shared by their endothermic counterparts. First, lacking the short-term vagility of birds and the larger mammals, many reptilian species cannot easily migrate from disturbance and reestablish themselves in more favorable habitat elsewhere (Gibbons et al., 2000). Second, changes in structural habitat (i.e. canopy cover) affect the physical properties of the earth's surface, thereby affecting microclimates greatly (Robinson, 1966; Porter and Gates, 1969; Newmark, 2001; Schlaepfer and Gavin, 2001; Pringle et al., 2003). Because reptiles must obtain heat from the surrounding physical environment, the effects of habitat fragmentation on the behavior and physiology of these organisms may be pervasive (Huey, 1991). Many reptiles primarily thermoregulate and thus select a narrow range of body temperatures (TVs) from the surrounding thermal environment (Brattstrom, 1965), thereby enhancing one or more physiological tasks such as locomotory performance, 149 digestion, reproduction, and immune function (Huey, 1982; Adolph and Porter, 1993). Other species, including some burrowing and aquatic forms, primarily thermoconform (Brattstrom, 1965) and effectively carry-out these same tasks by allowing their Tj, to fluctuate with the thermal environment around them (Huey, 1982). When thermoregulation requires elevated levels of activity (i.e. lizards shuttling between shaded and sunlit patches), individuals may experience a concomitant increase in both energy expenditure (Christian et al., 1997) and risk of predation (Rose, 1981; Adolph and Porter, 1993). Theory and field observations suggest that when the costs and potential risks associated with behavioral thermoregulation are high, lizards (Huey, 1974; Huey and Slatkin, 1976) and snakes (Brown and Weatherhead, 2000; Webb and Whiting, 2005) will thermoregulate less precisely (but see Blouin-Demers and Nadeau, 2005). Measuring the degree to which an individual regulates its T\, with respect to the environment remains a fundamental aspect of reptilian thermal studies (Hertz et al., 1993; Seebacher and Shine, 2004). All woodland reptile species, regardless of how precisely they maintain their 7b, have the potential to suffer adverse, physiological consequences (both direct and indirect) from the thermal effects of anthropogenic forest fragmentation. First, elevated TVs may directly result from individuals utilizing warmer, open-canopy habitats. Metabolism and water loss may increase with 7b, resulting in physiological stress. In extreme circumstances, individuals unable to maintain their 71, within critical physiological limits will face immediate and lethal consequences. Second, thermal constraints on activity arise when portions of an individual's habitat become thermally unsuitable and cannot be occupied for activities such as foraging, territory defense, mate acquisition, and predator 150 avoidance (Grant and Dunham, 1988; Beaupre, 1995). Under such spatial and/or temporal constraints, extracting resources from the environment may prove difficult. Resource limited individuals will typically spend a disproportional amount of their activity budget foraging (Beaupre, 2008) and have less available energy to allocate to the competing functions of growth, maintenance, storage, and reproduction (Dunham et al., 1989). The result will be individuals suffering from decreased growth, delayed maturation, less opportunity to mate, and reduced fecundity (Dunham et al., 1989). Environmentally imposed thermal constraints have been shown to markedly impact reptilian life histories (Beaupre, 1995; Martin, 2002). The purpose of this study was to explore potential effects of forest fragmentation on the thermal ecology of the timber rattlesnake (Crotalus horridus). I compare the Tb's of snakes using woodland, edge, and field habitats in an attempt to detect differences in thermal ecology possibly related to forest fragmentation. To determine if timber rattlesnakes experience thermal constraints on activity when foraging in disturbed areas, I deployed operative temperature (Te) models in both a woodland and a field during June, July, and August of 2006. I was therefore able to directly compare the percentage of thermally available habitat between the two habitat types. Assuming clear conditions, a lack of canopy cover will allow more solar radiation to strike ground-level vegetation in a field than in a woodland habitat per unit time. I hypothesized that such differences in insolation may result in fields having considerably warmer microclimates than woodlands throughout much of the timber rattlesnake's active season. Finally, I compare the Tb's of gestating females to those of males and non-gravid females. Because gravid females gestated exclusively among the boulders of an inactive rock quarry (Wittenberg, chapter 151 2), I discuss the apparent thermal suitability of this anthropogenic habitat as a rookery site. Methods Study species The timber rattlesnake (Crotalus horridus) is an ambush-foraging pitviper that associates with the deciduous forest biome of eastern North America (Brown, 1993). The species is currently recognized as vulnerable, imperiled, or critically imperiled in 20 of the 30 U.S. states where populations still remain (Waldron et al., 2006). Despite being afforded a degree of legal protection in many of these states, the timber rattlesnake has vanished throughout parts of its former range (Brown, 1993). In fact, the timber rattlesnake has been extirpated from Rhode Island, Maine, and possibly the Canadian province of Ontario (Brown, 1993). Most consider this rattlesnake species to rely on a closed-canopy ecosystem (Collins and Knight, 1980; Reinert, 1984a; Reinert and Zappalorti, 1988; Martin, 1992; Rudolph et al., 1998; but see Fitch, 1999), and fragmentation of this woodland habitat has been cited as a primary threat to the survival of the species (Martin, 1992; Brown, 1993; Clark et al., 2003; Furman, 2007). Study site The study took place in a highland region of west-central Missouri in what is considered the prairie geographic region (Schwartz and Schwartz, 1959). The highlands consist of hills containing hardwood forests that extend across a flat landscape of what was once tall grass prairie (Schwartz and Schwartz, 1959; Sims, 1988). This historical 152 intermingling of closed and open canopy habitats has been further diversified by human land use practices. Anthropogenic habitat fragmentation has created a mosaic of woodland, agricultural, and residential habitats. An approximately 32 hectare inactive rock quarry provides hibernacula for all timber rattlesnakes in the study population. Remnant boulder piles strewn throughout the quarry are used as rookery habitat for gestating females during summer months. After leaving the quarry, males and non- gravid females spend the summer foraging in woodland, edge, and field habitats. Characterization of habitat types and population segments Upon relocation outside of the quarry, males and non-gravid females could typically be described as using woodland, edge, or field habitat. Woodland habitat consisted primarily of secondary forest that lacked both canopy stratification as well as abundant deadfall in the form of large logs. Edge habitat was characterized by the dense, leafy, transitional vegetation found at the interface of woodland and field habitats or along major breaks within woodlands. With respect to fields, snakes occupied uncut grassy fields, old fields, crop fields (primarily soybean), fallow crop fields, and pasture. Individual records of habitat use and observations of foraging behavior within each habitat type are summarized in Wittenberg (chapter 3). While gestating, gravid females associated exclusively with boulders inside of the rock quarry (Wittenberg, chapter 2). Noting a similar divergence in habitat use between gestating females and males/non- gravid females, Reinert (1984b) considered these two groups to comprise behaviorally distinct segments within the timber rattlesnake population. 153 Body temperature measurements Body temperature data were obtained via temperature-sensitive radio-transmitters (Holohil Systems models SI-2T and SB-2T, Carp, Ontario) surgically implanted in the coelomic cavities of twenty-two free-ranging timber rattlesnakes (10 males, 12 females). During the study, nine of the twelve females were gravid for one season. Prior to implantation, transmitters were calibrated in a series of temperature controlled water baths that ranged at minimum from 0° to 40°C in approximately 5° increments. The range of physiologically feasible snake 7b's fell within this selected range of calibration temperatures. To calibrate the temperature-sensitive transmitters, regression equations describing the relationship between transmitter temperature and interpulse period were calculated using JMP version 7.0.1 (SAS Institute, Inc.). Regression analyses indicated that interpulse period explains from 98.2% to 100% of the variance in transmitter temperature. Average weight of the transmitter package varied by model (SI-2T = 13.7g, SB-2T = 5.4g), but always constituted less than 5% of the snake's total body mass. Snout-vent length (SVL) of implanted snakes ranged from 68.0 to 101.2 cm (mean = 85.05, SD = 7.84) while mass ranged from 245.0 to 1126.5 g (mean = 571.56, SD = 225.34). Transmitters were implanted following the surgical protocol of Reinert and Cundall (1982). Afterwards, snakes were given a minimum recovery period of 48 hours before release at the initial point of capture. Sampling of snake 7Vs occurred during routine relocations of the telemetry subjects using a Wildlife Materials TRX 1000S tracking receiver and a Yagi three- element directional antenna. A stopwatch was used to measure interpulse period in the field. To reduce human error associated with using a stopwatch, interpulse period was 154 obtained by measuring the total time needed for ten consecutive interpulse periods to occur and then dividing this value by ten. Stopwatch measurements were repeated four times and then averaged to obtain a single interpulse period (i.e. lb) per relocation event. Relocations typically took place during mid to late afternoon. With the exception of unfavorable weather conditions, it is assumed that individuals sampled during this portion of the day would have had the opportunity to acquire and maintain some of the warmest Ib's within their activity temperature range. Occasionally sampling would occur in the morning or after sunset in the evening. A series of relocations occurring between sunset and sunrise during the summer of 2006 provided a limited sample of nocturnal 7b's. The number of T\> records obtained from each individual varied (from 1 to 33) depending on how long a given snake remained in the study. Body temperature distributions of snakes within each habitat type (woodland, edge, field) and within each population segment (gestating females, males and non-gravid females) were evaluated for normality with Shapiro-Wilk tests and by imposing normality curves on histograms of the 7i, data. Body temperature distributions within each group were negatively skewed and failed to meet the normality criterion for parametric analyses. Therefore, a non-parametric Kruskal-Wallis test was used to determine if snake TVs differed among habitat types. Because a Levene's test detected variance heterogeneity in Tj, distributions between population segments, non-parametric analyses could not be used to determine if the Ib's of gravid females were significantly different than those of males and non-gravid females. Instead, I conducted a randomization analysis (Manly, 1997) to test the null hypothesis that any difference in mean 7t> between population segments was purely due to chance. Randomization 155 analyses do not assume random sampling, equal variances among groups, or that the response variable is distributed normally (Manly, 1997). However, randomization analyses only indicate whether or not patterns in the data are likely to have arisen by chance and the results cannot be extrapolated to the population from which the sample came (Manly, 1997). Previous studies of snake thermal biology have employed randomization procedures to analyze temperature data that failed to meet the assumptions of parametric and/or non-parametric tests (Wills and Beaupre, 2000). The two-sample randomization analysis used in this study is described in detail by Manly (1997). First, a D\ value was obtained by finding the difference in mean 7b's between gestating females and the population segment consisting of males and non- gravid females. Next, each of the individual Jb's were randomly reallocated between these two groups while maintaining original group sample sizes. Following the reallocation step, a £>2 value was obtained by finding the new difference in mean 7b's between population segments. This reallocation step and calculation of Di was carried- out 9,999 times. Combining the values of £>2 with D\ yields a randomization distribution consisting of 10,000 values. (The inclusion of D\ in the randomization distribution is warranted because if the null hypothesis is true then D\ is simply another value in the lower 95% of the distribution.) Finally, the total number of D values recovered in the randomization distribution greater than or equal to D\ (including D\) is divided by 10,000 to determine the significance of D\. Thus, if D\ resides in the upper 5% of the randomization distribution the mean TVs of the two population segments can be considered significantly different at the 5% level. 156 Additionally, I examined the degree of 7b dispersion exhibited by snakes at different times of the day by plotting 7b's during the well-sampled period from ca. 1400 - 2030 hours. Active season data from four years (2003-2007) were pooled in all comparisons to increase sample size. Although repeated measurements of T\, were collected from individual telemetry subjects, sufficient time passed between relocations (typically several days) for snakes to seek-out new microclimates or relocate to different habitat types altogether. A type I error of 0.05 was applied for all hypothesis tests. Normality testing and non-parametric analyses were performed using SAS (v. 9.3.1, SAS Institute). The program PopTools (v. 3.0.6) for Microsoft Excel (Hood, 2008) was used to perform the two-sample randomization analysis. Operative temperature measurements The percentage of thermally available habitat within two habitat types (woodland, field) was quantified by using Te modeling. Operative temperature is defined by Bakken and Gates (1975) as the temperature of an inanimate object of zero heat capacity with the same size, shape, and radiative properties as the animal and exposed to the same microclimate. Environmental TVs recorded by such inanimate objects (Te models) estimate actual equilibrium 7b's that the study animal would have experienced in that particular microhabitat. Of the three types of foraging habitat, woodlands and fields represented the extremes with respect to overstory canopy cover. Therefore, a representative woodland tract and uncut grassy field were selected for this investigation. Following the methods outlined by Peterson et al. (1993) and adapted by Wills and Beaupre (2000) for the study of timber rattlesnakes, I constructed 60 Te models to 157 simulate the thermal properties of snakes in my study population. For constructing Te models of snakes, Peterson et al. (1993) suggest that 1) models be constructed of copper because of its high conductivity and tendency to minimize temperature gradients within the model, 2) model length should span the approximate width of a coiled study specimen, 3) model diameter should approximate mean study specimen thickness, and 4) models should be painted to match the spectral reflectivity of the study species. Therefore, I selected rigid copper tubing with an inside diameter of 2.54 cm (wall thickness 0.159 cm) and cut models 15.24 cm in length. The dorsal background coloration of snakes in the study population ranged from a yellowish-tan to a silvery- gray. Data published by Peterson et al. (1993) indicate that gray models match the reflectivity of several snake species with a dorsal background coloration similar to timber rattlesnakes, therefore I spray painted my models light gray. Following the protocol of Wills and Beaupre (2000), I then used a matte black spray paint to add three, evenly spaced, V-shaped markings to each model. These markings enhanced the realism of the model by representing the dark bands on the dorsum of a living timber rattlesnake. Thermochron iButtons (Dallas Semiconductors model DS1921G-F50, Dallas, Texas) served as miniature temperature loggers within each Te model. The particular thermochrons used in this study have a resolution of 0.5 °C and are guaranteed by the manufacturer to be accurate to within ±1.0 °C. However, when tested over a biologically-relevant range of temperatures (5-50 °C), Angilletta and Krochmal (2003) found that on average these devices deviated ± 0.3 °C from the actual temperature. Thermochrons were programmed to record a temperature at the beginning of each hour for all hours of the day, beginning on 02 June 2006 at 0100 hours. Because each device 158 can collect and store a maximum of 2048 temperature records, I was able to obtain hourly data through 0800 hours on 26 August. After programming, thermochrons were placed in small wire baskets made from XA in. hardware cloth. The baskets were designed to span the inner diameter of the copper tubes and anchor the thermochrons in the middle of each model (away from either end) and prevent the thermochrons from contacting the model's inner wall. This ensured that the thermochrons would only be measuring the air temperature within each model and not the temperature of the copper tubing. After the baskets were inserted into the Te models, the ends of each model were sealed with a plug of aluminum foil. Thirty models were placed in a woodland corridor and thirty were placed in an adjacent, uncut, grassy field. The locations were representative of the secondary forest and old field habitats frequented by telemetry snakes. In each habitat, models were placed approximately 2.0 meters apart in a 5 x 6 grid, and small wooden stakes were used to mark the location of each model. The grid determined model placement, therefore the deployment of models into the surrounding microhabitat was essentially random. Models were positioned to mimic how live rattlesnakes used that particular habitat. In woodland tracts and corridors, timber rattlesnakes coiled on the ground and were typically unconcealed by surface vegetation. Therefore, woodland models were placed on the ground and not covered with leaves or surface debris. By contrast, snakes using fields were almost always well hidden beneath the field vegetation. Field models were placed down into the vegetation accordingly and were generally well concealed. As suggested by Peterson et al. (1993), models were always oriented perpendicular to the path of the sun. 159 Accuracy of the remodels was checked by comparing model temperatures to those of live snakes. Two adult timber rattlesnakes implanted with temperature-sensitive transmitters were placed in a 2.44 x 1.52 m wire enclosure in an unshaded outdoor area. The snakes represented the yellowish-tan and silvery-gray color patterns found in the population. Calibration data were collected over multiple days and at different times of day. Weather conditions included clear and cloudy skies, variable air temperatures, occasionally high winds, and a rain event. Snakes were either exposed on the surface without shade, shaded from above with a tarp (simulating canopy shade experienced by snakes in woodland habitats), or made to coil beneath clumps of grassy vegetation (simulating field foraging). Once the model was placed next to a particular snake, I waited a minimum of 15 minutes before collecting data to ensure the model and snake would have time to reach thermal equilibrium. Properly designed Te models should be good predictors of snake TV The linear regression of this relationship should have a slope = 1.0 and an intercept = 0. Snake 7b was regressed on the Te of the model using major axis (model II) regression (box 15.6, Sokal and Rohlf, 1995). Calculations were first done by hand and later validated using the software program SMATR (Standardized Major Axis Tests and Routines) version 2.0. Thermal constraints on activity in both the woodland and field were evaluated by calculating overall and hourly percentages of thermally available habitat during two-week intervals spanning 02 June to 26 August 2006. For each habitat and time period of interest, the percentage of thermally available habitat was calculated by dividing the number of recorded TVs at or below the maximum voluntary 7b documented for a non- gravid timber rattlesnake in the study population (34.6 °C) by the total number of 160 recorded Te's, and multiplying the resulting quotient by 100. Because thermochrons have a resolution of 0.5 °C, I considered re's of 35.0 °C or greater to represent thermally unavailable habitat. Given that climate patterns and vegetation growth did not remain static during the three-month sampling period, data were arbitrarily partitioned into two- week intervals to investigate temporal differences in thermal constraints as the active season progressed. Results Body temperatures in woodlands, edges, and fields Median body temperature did not differ significantly among habitat types (Kruskal-Wallis test: Chi-Square = 0.09, df = 2, P = 0.9558). Snake TVs were similar in woodland (median = 28.0 °C, range = 16.4 - 34.2 °C, n = 63), edge (median = 28.6 °C, range = 16.7 - 33.9 °C, n = 56), and field (median = 28.2 °C, range = 20.2 - 34.4 °C, n = 61) habitats (Fig. 1). Additionally, I examined 126 Tb's of snakes found using either woodland, edge, or field habitat between ca. 1400 - 2030 hours (Fig. 2). During this sampling period, Tb's ranged between 16.4 °C and 34.4 °C (mean 29.1 °C, SD = 3.8). Patterns of T\, dispersion that could be attributed to time of day were not apparent, and snakes exhibited warm 7b's well beyond midday. Thermal constraints Analysis of the relationship between model temperature and snake It, using major axis regression yielded a slope of 1.191 (95% CI from 0.997 to 1.430) and an intercept of 0.821 (95% CI from -4.283 to 5.925), where Tb = 1.191(7;) - 0.821 (n = 51). The 161 regression line did not differ significantly from a line having a slope of one and an intercept of zero, indicating a properly designed Te model. Therefore, the models were determined to be good predictors of snake 7b. The deployment of 30 Temodels in both the woodland and field to log hourly re's during the 86 day sampling period (02 June to 26 August 2006) should have yielded 61,440 re records in each habitat. However, Te records in the field were reduced to 59,392 after one model was dropped from the study due to a thermochron malfunction. During each two-week interval, the overall percentage of thermally available habitat was high in both woodland (range 84.2 - 99.8 %) and field (range 78.1 - 94.9 %) habitats (Fig. 3). However, sharp decreases in habitat availability occurred in both the woodland and field during midday (Figs. 4, 5, 6). The lowest hourly percentages of thermally available habitat were recorded at 1400 hours in the field (19.7%) and at 1500 hours in the woodland (36.7%) during the two-week interval from 29 July to 11 August (Fig. 6, top). Body temperatures of females gestating within an inactive quarry Randomization analysis detected a significant difference in mean T\, between population segments, as the observed difference (Di) of 3.7 °C was in the top 0.01% of the differences generated from randomized distributions. Body temperatures of gestating females (median = 32.5 °C, range = 25.3 - 34.7 °C, n = 113) were warmer and less variable (Levene's test: F = 30.15, df = 1, P < 0.0001) than 7b's of males and non-gravid females (median = 28.3 °C, range = 16.4 - 34.4 °C, n = 180) (Fig. 7). 162 Discussion Habitat-specific thermal biology Direct thermal consequences associated with using disturbed habitat (i.e. elevated Ib's) were not apparent in this study, given that Ib's of timber rattlesnakes did not differ among woodland, edge, and field habitats. However, sample sizes did not allow me to investigate possible habitat-specific differences in It, related to time of day. Interestingly, habitat-specific differences in It, have been documented for other snake species. Blouin-Demers and Weatherhead (2002) found that during the day black rat snakes were significantly warmer in barns than in edge habitat, and significantly warmer in edges than in forest. At night, rat snakes in barns remained significantly warm while Jb's of snakes in edges and forests were no longer significantly different from one another. Temperature-sensitive telemetry was used to obtain It, data on timber rattlesnake populations in closed-canopy forests of northeastern New York and northwestern Arkansas. In New York, near the species northernmost range limit, Brown et al. (1982) reported an overall mean It, of 26.9 °C and a maximum voluntary Th of 33.3 °C. In Arkansas, Wills and Beaupre (2000) reported average Ib's ranging from nocturnal lows of 20.8 °C to diurnal highs of 26.7 °C during the month of August. In September, average Ib's dropped to nocturnal lows of 16.8 °C and diurnal highs of 24.0 °C. The authors report a maximum voluntary Tb of 37.4 °C. Unfortunately, directly comparing the thermal biology of snakes from my study population to those in New York and Arkansas is not possible due to large differences in sampling. For instance, Brown et al. (1982) used point sampling to obtain data on five snakes during a single season. The 163 overall mean It, reported by the authors included data from a gravid female that maintained a warmer mean It, relative to other snakes. Wills and Beaupre (2000) used both point and semi-continuous sampling to obtain hourly T\, profiles for twelve snakes during August and September of two consecutive seasons. Although automated (semi- continuous) data collection systems generate large sample sizes, my emphasis on habitat specific Jb's required that I physically locate each snake within the mosaic of habitats present at the study site. Taylor et al. (2004) suggest that point sampling, often a non random process, may produce skewed mean T\, values having a greater variance than those obtained via semi-continuous sampling. Questions regarding population-specific differences in thermal biology can best be answered by using automated data collection systems at each site to gather large samples of semi-continuous data over equivalent time periods (Beaupre and Beaupre, 1994). Thermal constraints on activity The overall percentage of Te models recording temperatures of 34.5 °C and below was high in both the woodland and field habitats during the 86 day sampling period (Fig. 3). By examining the data in approximately two-week intervals, we see that the woodland habitat remains slightly cooler than the field habitat in each of the six time periods (on average, the woodland had 6.8% more thermally available habitat than the field) (Fig. 3). The lowest overall thermal availability in both habitats occurred during the two week period of 29 July- 11 August, when 15.8% of woodland and 21.9% of field models recorded temperatures of 35.0 °C or higher (Fig. 3). Models, regardless of habitat, never recorded temperatures at or above 35.0 °C between 2000 and 0800 hours 164 (Figs. 4, 5, 6). Thus, as to be expected, snakes could freely forage in either habitat from late evening until early morning without encountering uncomfortably high temperatures. Thermal constraints on activity only occurred between 0900 and 1900 hours, and during this time the differences between the woodland and field were often pronounced (Figs. 4, 5,6). Hourly lows in thermal availability for both habitats were also recorded between 29 July and 11 August (Fig. 6, top). During this two week period, only 19.7% of the field habitat was thermally available at 1400 hours. At 1500 hours, only 36.7% of the woodland habitat could be accessed by surface active rattlesnakes. Operative temperature models were also used by Wills and Beaupre (2000) to quantify thermal constraints on timber rattlesnake activity at their study site in northwest Arkansas. Within a mature Ozark forest during the months of August and September, the amount of thermally available surface habitat never dropped below 78%. Habitat availability was found to be less than 100%) only during midday (between 1000 and 1600 hours). Wills and Beaupre concluded that the activities of timber rattlesnakes in mature, closed-canopy forests are not meaningfully constrained by the thermal environment. By comparison, my data suggest that snakes using small, secondary woodland tracts and corridors with less canopy closure may at times be considerably more constrained (Figure 6, top). Although I found no evidence that snakes in fields were warmer than those in woodlands or edges (Fig. 1), redata indicate that fields can be more thermally challenging during midday (Figs. 4, 5, 6). Therefore, it is reasonable to assume that timber rattlesnakes foraging in fields must either 1) make thermoregulatory movements to cooler microhabitats during these warm midday periods or 2) seek-out microhabitats 165 that remain within their thermal tolerance range throughout the day. Because most pit vipers are strictly ambush foragers, researchers have suggested that repeated movements for the purpose of thermoregulaton ("shuttling") might alert potential prey to their presence (Wills and Beaupre, 2000; Shine et al., 2002). However, Clark (2006) found that 81% of timber rattlesnake feeding events occurred at night. Therefore, I contend that short, midday movements within fields may not significantly impact an individual's foraging success, especially if warm temperatures have forced nocturnally active rodents into their daytime retreats. Furthermore, it has been shown that rattlesnakes can use their thermally sensitive facial pits to remotely sense cooler, more suitable microhabitats nearby (Krochmal and Bakken, 2003). Timber rattlesnakes may use these structures to quickly and efficiently relocate to cooler microhabitats within fields, thus reducing random movements and the possibility of alerting prey. Selecting foraging microhabitats that remain thermally available throughout the day may prevent snakes from making thermoregulatory-based movements, even though such microhabitat may only represent a small fraction (ca. 1/5) of that which is available. Snakes may benefit from foraging within small patches of relatively dense field vegetation that provide thermal refugia for themselves, as well as their prey. The mutual concentration of snakes and their prey in thermal refugia is not unknown, as Webb and Shine (1998) found that rocks with specific thermal attributes were favored by broad-headed snakes (Hoplocephalus bungaroides) and the velvet geckos (Oedura lesueurii) on which they feed. In theory, both shuttling and the selection of benign microhabitats may be feasible strategies allowing timber rattlesnakes to overcome midday thermal constraints within fields. However, telemetry observations of snakes at my study site indicate that 166 individuals remain in fields throughout the day and appear to move very little within a particular field between successive relocations. Interestingly, these snakes were often found foraging near the base of leafy vegetation or a cluster of thistle. These structures may be important in providing favorable thermal microclimates for foraging snakes. Similarly, Porter et al. (1973) found that by ascending into small bushes, desert iguanas (Dipsosaurus dor salts) could greatly increase their daily activity time in the Mojave Desert. Although I found no evidence that timber rattlesnakes abandon fields to avoid midday temperatures, Beaupre (1995) found that rock rattlesnakes (Crotalus lepidus) would forgo surface activity when the percentage of thermally available habitat dropped below 7%. A wide range of Ib's was recorded for snakes during mid to late afternoon hours (Fig. 2), suggesting that although diurnally foraging snakes must avoid lethal temperatures, they are by no means stenothermic. In fact, Wills and Beaupre (2000) predicted that timber rattlesnakes in thermally benign closed-canopy forests might trade thermoregulatory precision for the ability to forage in a specific location. However, the authors cited significant differences between 7b and Te distributions as possible evidence of thermoregulation. Although individuals do remain static for long periods and exhibit a wide range of 2Vs while foraging, this does not necessarily imply passive thermoconformity. According to Stevenson (1985a), terrestrial ectotherms between 0.1 and 1.0 kg have thermal inertias that allow for the greatest range of body temperatures. Presumably, postural changes that modify surface to volume ratio (i.e. tightly versus loosely coiled) would allow foraging rattlesnakes to alter their heat flux properties without alerting prey. However, Stevenson (1985b) used mathematical heat-transfer 167 models to estimate that changes in body shape would, at most, allow a 1.0 kg reptile to change its It, by 5 °C. Thermal biology of gestating females My finding that gestating female timber rattlesnakes maintained warmer and less variable JVs than their male and non-gravid female counterparts is consistent with data reported for this species in closed-canopy forests. In northwest Arkansas between 15 July and 22 August of a single active season, Gardner-Santana and Beaupre (2009) found that six gravid females exhibited warmer mean Jb's than five non-gravid females during early morning, late afternoon, and the first half of scotophase. Additionally, gravid female TVs were significantly less variable than those of non-gravid females between 1500 and 1800 hours. Gardner-Santana and Beaupre also use hourly T\, data obtained via a datalogger to calculate overall mean 7b's of 28.6 °C and 24.2 °C for gravid and non- gravid females, respectively. Thermal studies of other rattlesnake taxa have also found that gestating females exhibit warm and stable TVs when compared to other segments of the population. A telemetry study of free-ranging western rattlesnakes (Crotalus viridis viridis) found that gravid females maintained a mean It, of 26.5 °C, which was both warmer and less variable than T^'s maintained by non-gravid females (Graves and Duvall, 1993). Charland and Gregory (1990) also evaluated the thermal biology of gravid and non- gravid western rattlesnakes, but did so within an outdoor enclosure. The authors found that during a daily 7b plateau period (between 1230 and 1630 hours) mean Jb's did not differ significantly between gravid (31.7 °C) and non-gravid (30.7 °C) females. However, the mean temperature range tolerated by gravid females was significantly 168 narrower. Gier et al. (1989) found that gravid and non-gravid northern Pacific rattlesnakes {Crotalus viridis oreganus) both fluctuated around a mean overall T\, of 29.5 °C when placed in a laboratory thermal gradient. However, females exhibited a wanner and less variable mean T\, before parturition than after (gravid 30.7 °C, SD = 0.89; postpartum 28.5 °C, SD = 1.41). The trend of reproductive females maintaining warmer mean 7b's and / or a narrower range of TVs has actually been documented in snake families other than the Viperidae. In a laboratory thermal gradient, gravid diamondback water snakes (Nerodia rhombifera), family Colubridae, fhermoregulated more precisely and selected a warmer mean Th (27.7 °C) than non-gravid (24.8 °C) and postpartum (July and August 24.1 °C, November 22.3 °C) females (Tu and Hutchison, 1994). A study of free-ranging black rat snakes (Elaphe obsoleta), also in the family Colubridae, found that prior to egg deposition females maintain warmer 7Vs than both males and non-gravid females (Blouin-Demers and Weatherhead, 2001). In the family Elapidae, the eastern brownsnake (Pseudonaja textilis) deposits its eggs in burrows that are warmer and more thermally stable than other available subterranean retreats (Whitaker and Shine, 2002). A study of free-ranging diamond pythons (Morelia spilota), family Boidae, found that brooding females used spasmodic muscular contractions to maintain warm and stable TVs (shivering thermogenesis) while tending their clutch (Slip and Shine, 1988). Brooding females had a substantially warmer and less variable mean 7b (30.9 °C, SD = 2.15) than non brooding females (24.7 °C, SD = 4.43) and males (24.4 °C, SD = 4.99). Interestingly, some studies report that gravid females do not maintain warmer mean TVs or thermoregulate more precisely than other segments of the population. In 169 their study of the copperhead (Agkistrodon contortrix), family Viperidae, Sanders and Jacob (1981) compared the mean 7b of gravid females to males and non-gravid females from two different size classes. Gravid females had a mean 7b of 23.9 °C and were actually cooler than males and non-gravid females both greater (25.2 °C) and less (26.4 °C) than 50 cm SVL. The authors did suggest, however, that gravid females might thermoregulate more precisely than males and non-gravid females. Isaac and Gregory (2004) used an outdoor enclosure to study the thermal ecology of grass snakes (Natrix natrix), family Colubridae, at a high latitude site in the United Kingdom. Gravid females maintained a mean 7b of 23.1 °C prior to laying their eggs, while non-gravid females had a significantly warmer mean 7b of 25.3 °C. Gravid females also exhibited a more variable mean 7b (SD = 1.95) than non-gravid females (SD = 1.75), although this difference was not significant. Studies of reptilian embryogenesis have shown that temperature can affect length of the gestation period (Naulleau, 1986) as well as the viability, morphology, and sex of the offspring (Shine, 2004). Consequently, the 7t,'s of gestating females may differ from those selected by males and non-gravid females for activities such as foraging, digestion, and ecdysis (Peterson et al., 1993). Gier et al. (1989) predict that the thermoregulatory behavior of gravid females should diverge from the rest of the adult population when the optimum temperature for embryonic development differs from that of other activities. Increasingly, studies are finding that female snakes demonstrate shifts in their thermal biology when gravid (Peterson et al., 1993). However, the magnitude, precision, and direction of such shifts are quite variable. 170 I suggest that the variation reported in the literature pertaining to thermoregulatory patterns of gravid snakes may be attributed to several factors. First, studies vary widely from those conducted in laboratory cages or outdoor enclosures to those that monitor free-ranging snakes in nature. Furthermore, sample sizes of gravid snakes are low in many studies. For instance, Sanders and Jacob (1981) report T\, data from only two gravid copperheads. Determining if the described thermoregulatory patterns of gravid snakes are artifacts of experimental design or patterns that occur in nature can therefore prove difficult. Second, optimal temperatures for embryogenesis are species- and perhaps even population-specific (Shine, 2004). Third, Ij/s maintained by gestating females may be the result of tradeoffs or limitations associated with the local environment. By engaging in overt thermoregulatory behavior (i.e. basking), gravid reptiles may risk exposure to predators (Shine, 1980). As a tradeoff between thermoregulatory precision and survival, females may gestate at suboptimal TV s in order to avoid predators. Limitations occur when a snake's ability to thermoregulate is constrained by the range of temperatures in the surrounding environment. For example, as forests mature and overstory canopies close, warm microhabitats previously used by timber rattlesnakes for basking and gestation become shaded over and can no longer be used (Brown, 1993; Fitch, 1999). Thus, while laboratory studies may not recover the true thermoregulatory patterns of snakes in nature, TVs obtained from free-ranging females may not represent temperatures that optimize embryo development. Further research is needed in order to better understand how the above factors interact to produce the variable thermoregulatory patterns of gravid reptiles. 171 Use of rock quarry as a rookery Studies of timber rattlesnakes in eastern forests indicate that gravid females seek- out birthing rookeries that consist of rocky, sparsely wooded sites that provide thermally favorable habitat for gestation and eventually parturition (Keenlyne, 1972; Reinert, 1984b; Martin, 1993). Birthing rookeries are often located in close proximity to the hibernacula (Martin, 1993). Telemetry, mark-recapture, and observational data on the study population failed to locate any rookeries or hibernacula located beyond the boundaries of an inactive (ca. 25 years post mining) rock quarry approximately 32 hectares in total area (Wittenberg, chapter 2). Gestation occurred in sparsely wooded areas that featured little if any canopy cover, primarily amongst the many boulder piles that lined old access roads or atop boulder strewn ridges. During the gestation period, females did not actively forage and instead spent much of their time beneath one or more large maternal rocks until parturition occurred in late summer (Wittenberg, chapter 3). Graves and Duvall (1993) described similar use of rocks by gestating western rattlesnakes at rookeries in south-central Wyoming and suggested that the rocks insulated the snakes from excessively high and low environmental temperatures. Huey et al. (1989) found that thermoregulatory opportunities available for gravid female garter snakes (Thamnophis elegans) beneath rocks matched, and sometimes surpassed, those opportunities available to snakes on the surface. Gestating garter snakes exploited daily temperature cycles beneath rocks and were able to stay within their preferred It, range throughout a 24 hour period without becoming surface active. The quarry, although anthropogenic, provided gestating females the opportunity to select suitable maternal rocks beneath which they could maintain warm and stable 7b's. 172 Conclusions The present study found no direct, adverse effects of habitat fragmentation on the thermal biology of timber rattlesnakes. Although activity within old fields is constrained at midday during the warmest portion of the active season, snakes are never completely excluded from such habitat. Despite being considered a specialist of mature, closed canopy forest, this species is adept at using ground vegetation to maintain its thermal requirements in areas where no trees are present. Furthermore, gestating females derived thermal benefits through the exclusive use of an anthropogenic habitat as a birthing rookery. By modeling a single portion of woodland and field habitat, I undoubtedly failed to capture the complexity of thermal microclimates used by this population. Despite lacking canopy cover, tall crop fields and extremely dense, brushy fields may be thermally benign. When such fields are plowed or mowed, vast areas of the landscape are rendered thermally unsuitable for timber rattlesnakes (pers. obs.). I chose to investigate woodland and field microclimates because they represented the extremes with respect to canopy closure. The relationship between amount of solar radiation reaching ground-level vegetation and canopy cover is less clear when comparing woodland and edge habitats. For instance, two edge habitats with identical canopy cover may receive substantially different amounts of solar radiation depending on the direction they face. Studying the daytime thermal preferences of black rat snakes at high latitude (Ontario, Canada), Blouin-Demers and Weatherhead (2002) concluded that edge habitats were actually thermally superior to both forests (too cool) and open habitats (too warm). Clearly, there is much to be discovered by attempting to elucidate the mechanisms linking forest fragmentation, microclimate change, and the physiological ecology of 173 vertebrate ectotherms. Specifically, I urge researchers working to understand the impact of forest fragmentation on woodland reptile populations to consider thermal effects when designing their studies. Acknowledgments This research was conducted with the approval of the University of Arkansas Institutional Animal Care and Use Committee (protocol # 05001) and the Missouri Department of Conservation (collecting permits # 12005, 12367, 12715, 13101). I thank Steven J. Beaupre, Kimberly G. Smith, Ines Pinto, and Edward E. Gbur Jr. for comments on this manuscript. Joseph Agugliaro and Jeff Velie kindly provided technical assistance with various aspects of this project. I am indebted to Sara Wittenberg for her diligent proofreading and words of encouragement while this work was being completed. Numerous private landowners in Missouri allowed this work to be conducted on their property, and to them I am grateful. This research was supported in part by a University of Arkansas Causey Grant-in-Aid Award, a Harry Steinman Memorial Grant from the St. Louis Herpetological Society, and a grant from the Arkansas Audubon Society Trust. References Adolph, S.C., Porter, W.P., 1993. Temperature, activity, and lizard life histories. Am. Nat. 142, 273-295. Angilletta, M.J., Jr., Krochmal, A.R., 2003. The thermochron: a truly miniature and inexpensive temperature-logger. Herpetol. Rev. 34, 31-32. Bakken, G.S., Gates, D.M., 1975. Heat transfer analysis of animals: some implications for field ecology, physiology, and evolution. In: Gates, D.M., Schmerl, R.B. (Eds.), Perspectives of Biophysical Ecology. Springer- Verlag, New York, pp. 255-290. 174 Beaupre, S.J., 1995. Effects of geographically variable thermal environment on bioenergetics of mottled rock rattlesnakes. Ecology 76, 1655-1665. Beaupre, S.J., 2008. Annual variation in time-energy allocation by timber rattlesnakes (Crotalus horridus) in relation to food acquisition. In: Hayes, W.K., Beaman, K.R., Cardwell, M.D., Bush, S.P. (Eds.), Biology of the Rattlesnakes. Loma Linda University Press, Loma Linda, pp. 111-122. Beaupre, S.J., Beaupre, R.W., 1994. An inexpensive data collection system for temperature telemetry. Herpetologica 50, 509-516. Blouin-Demers, G., Nadeau, P., 2005. The cost-benefit model of thermoregulation does not predict lizard thermoregulatory behavior. Ecology 86, 560-566. Blouin-Demers, G., Weatherhead, P.J., 2001. Thermal ecology of black rat snakes {Elaphe obsoleta) in a thermally challenging environment. Ecology 82, 3025- 3043. Blouin-Demers, G., Weatherhead, P.J., 2002. Habitat-specific behavioural thermoregulation by black rat snakes {Elaphe obsoleta obsoleta). Oikos 97, 59- 68. Brattstrom, B.H., 1965. Body temperatures of reptiles. Am. Midi. Nat. 73, 376-422. Brown, G.P., Weatherhead, P.J., 2000. Thermal ecology and sexual size dimorphism in northern water snakes, Nerodia sipedon. Ecol. Monogr. 70, 311-330. Brown, W.S., 1993. Biology, status, and management of the timber rattlesnake (Crotalus horridus): a guide for conservation. SSAR Herpetol Circ No. 22, 1-78. Society for the Study of Amphibians and Reptiles. Univ. of Kansas, Lawrence. Brown, W.S., Pyle, D.W., Greene, K.R., Friedlaender, J.B., 1982. Movements and temperature relationships of timber rattlesnakes (Crotalus horridus) in northeastern New York. J. Herpetol. 16, 151-161. Charland, M.B., Gregory, P.T., 1990. The influence of female reproductive status on thermoregulation in a viviparous snake, Crotalus viridis. Copeia 1990, 1089- 1098. Christian, K.A., Baudinette, R.V., Pamula, Y., 1997. Energetic costs of activity by lizards in the field. Funct. Ecol. 11, 392-397. Clark, A.M., Moler, P.E., Possardt, E.E., Savitzky, A.H., Brown, W.S., Bowen, B.W., 2003. Phylogeography of the timber rattlesnake (Crotalus horridus) based on mtDNA sequences. J. Herpetol. 37, 145-154. 175 Clark, R.W., 2006. Fixed videography to study predation behavior of an ambush foraging snake, Crotalus horridus. Copeia 2006, 181-187. Collins, J.T., Knight, J.L., 1980. Crotalus horridus. Cat. Amer. Amphib. Rept. 253, 1-2. Dunham, A.E., Grant, B.W., Overall, K.L., 1989. Interfaces between biophysical and physiological ecology and the population ecology of terrestrial vertebrate ectotherms. Physiol. Zool. 62, 335-355. Fitch, H.S., 1999. A Kansas snake community: composition and changes over 50 years. Krieger Publishing Company, Malabar. Furman, J., 2007. Timber rattlesnakes in Vermont and New York: biology, history, and the fate of an endangered species. University Press of New England, Lebanon. Gardner-Santana, L.C., Beaupre, S.J., 2009. Timber Rattlesnakes {Crotalus horridus) exhibit elevated and less variable body temperatures during pregnancy. Copeia 2009, 363-368. Gibbons, J.W., Scott, D.E., Ryan, T.J., Buhlmann, K.A., Tuberville, T.D., Metts, B.S., Greene, J.L., Mills, T., Leiden, Y., Poppy, S., Winne, C.T., 2000. The global decline of reptiles, deja vu amphibians. BioScience 50, 653-666. Gier, P.J., Wallace, R.L., Ingermann, R.L., 1989. Influence of pregnancy on behavioral thermoregulation in the northern Pacific rattlesnake Crotalus viridis oreganus. J. Exp. Biol. 145, 465-469. Grant, B.W., Dunham, A.E., 1988. Thermally imposed time constraints on the activity of the desert lizard Sceloporus merriami. Ecology 69, 167-176. Graves, B.M., Duvall, D., 1993. Reproduction, rookery use, and thermoregulation in free- ranging, pregnant Crotalus v. viridis. J. Herpetol. 27, 33-41. Heard, G.W., Black, D., Robertson, P., 2004. Habitat use by the inland carpet python {Morelia spilota metcalfei: Pythonidae): seasonal relationships with habitat structure and prey distribution in a rural landscape. Austral Ecol. 29, 446-460. Hertz, P.E., Huey, R.B., Stevenson, R.D., 1993. Evaluating temperature regulation by field-active ectotherms: the fallacy of the inappropriate question. Am. Nat. 142, 796-818. Hood, G. M. (2008) PopTools version 3.0.6. Available on the internet. URL http://www.cse.csiro.au/poptools Huey, R.B., 1974. Behavioral thermoregulation in lizards: importance of associated costs. Science 184, 1001-1003. 176 Huey, R.B., 1982. Temperature, physiology, and the ecology of reptiles. In: Gans, C, Pough, F.H. (Eds.), Biology of the Reptilia. Academic Press, New York, pp. 25- 91. Huey, R.B., 1991. Physiological consequences of habitat selection. Am. Nat. 137, S91- S115. Huey, R.B., Peterson, C.R., Arnold, S.J., Porter, W.P., 1989. Hot rocks and not-so-hot rocks: retreat-site selection by garter snakes and its thermal consequences. Ecology. 70, 931-944. Huey, R.B., Slatkin, M., 1976. Cost and benefits of lizard thermoregulation. Q. Rev. Biol. 51,363-384. Isaac, L.A., Gregory, P.T., 2004. Thermoregulatory behaviour of gravid and non-gravid female grass snakes {Natrix natrix) in a thermally limiting high-latitude environment. J. Zool. Lond. 264, 403-409. Keenlyne, K.D., 1972. Sexual differences in feeding habits of Crotalus horridus horridus. J. Herpetol. 6, 234-237. Krochmal, A.R., Bakken, G.S., 2003. Thermoregulation is the pits: use of thermal radiation for retreat site selection by rattlesnakes. J. Exp. Biol. 206, 2539-2545. Mac Nally, R., Brown, G.W., 2001. Reptiles and habitat fragmentation in the box- ironbark forests of central Victoria, Australia: predictions, compositional change and faunal nestedness. Oecologia 128, 116-125. Manly, B.F.J., 1997. Randomization, bootstrap and monte carlo methods in biology. 2n ed. Chapman and Hall, New York. Martin, W.H., 1992. The timber rattlesnake: its distribution and natural history. In: Tyning, T.F. (Ed.), Conservation of the Timber Rattlesnake in the Northeast. Massachusetts Audubon Society, Lincoln, pp. 13-22. Martin, W.H., 1993. Reproduction of the timber rattlesnake {Crotalus horridus) in the Appalachian Mountains. J. Herpetol. 27, 133-143. Martin, W.H., 2002. Life history constraints on the timber rattlesnake {Crotalus horridus) at its climatic limits. In: Schuett, G.W., Hoggren, M., Douglas, M.E., Greene, H.W. (Eds.), Biology of the Vipers. Eagle Mountain Publishing, Eagle Mountain, pp. 285-306. Naulleau, G., 1986. Effects of temperature on 'gestation' in Vipera aspis and V. berus (Reptilia: Serpentes). In: Rocek, Z. (Ed.), Studies in Herpetology: Proceedings of the European Herpetological Meeting (3rd Ordinary General Meeting of the 177 Societas Europaea Herpetologica) Prague 1985. Charles University, Prague, pp. 489-494. Newmark, W.D., 2001. Tanzanian forest edge microclimatic gradients: dynamic patterns. Biotropica33, 2-11. Peterson, C.R., Gibson, A.R., Dorcas, M.E., 1993. Snake thermal ecology: the causes and consequences of body-temperature variation. In: Seigel, R.A., Collins, J.T. (Eds.), Snakes: Ecology and Behavior. McGraw-Hill, New York, pp. 241-314. Porter, W.P., Gates, D.M., 1969. Thermodynamic equilibria of animals with environment. Ecol. Monogr. 39, 227-244. Porter, W.P., Mitchell, J.W., Beckman, W.A., DeWitt, C.B., 1973. Behavioral implications of mechanistic ecology: thermal and behavioral modeling of desert ectotherms and their microenvironment. Oecologia 13, 1-54. Pringle, R.M., Webb, J.K., Shine, R., 2003. Canopy structure, microclimate, and habitat selection by a nocturnal snake, Hoplocephalus bungaroides. Ecology 84, 2668- 2679. Reinert, H.K., 1984a. Habitat separation between sympatric snake populations. Ecology 65, 478-486. Reinert, H.K., 1984b. Habitat variation within sympatric snake populations. Ecology 65, 1673-1682. Reinert, H.K., Cundall, D., 1982. An improved surgical implantation technique for radio tracking snakes. Copeia 1982, 702-704. Reinert, H.K., Zappalorti, R.T., 1988. Timber rattlesnakes (Crotalus horridus) of the pine barrens: their movement patterns and habitat preference. Copeia 1988, 964-978. Robinson, T.S., 1966. Effect of canopy density and slope exposure on the subcanopy microenvironment of a northern hardwood forest. Am. Midi. Nat. 75, 339-346. Rose, B., 1981. Factors affecting activity in Sceloporus virgatus. Ecology 62, 706-716. Rudolph, D.C., Burgdorf, S.J., Conner, R.N., Dickson, J.G., 1998. The impact of roads on the timber rattlesnake, {Crotalus horridus), in eastern Texas. In: Evink, G.L., Garrett, P., Zeigler, D., Berry, J. (Eds.), Proceedings of the International Conference on Wildlife Ecology and Transportation. FL-Er-69-98, Florida Department of Transportation, Tallahassee, pp. 236-240. Sanders, J.S., Jacob, J.S., 1981. Thermal ecology of the copperhead (Agkistrodon contortrix). Herpetologica 37, 264-270. 178 Schlaepfer, M.A., Gavin, T.A., 2001. Edge effects on lizards and frogs in tropical forest fragments. Conserv. Biol. 15, 1079-1090. Schwartz, C.W., Schwartz, E.R., 1959. The wild mammals of Missouri. University of Missouri Press, Columbia. Seebacher, F., Shine, R., 2004. Evaluating thermoregulation in reptiles: the fallacy of the inappropriately applied method. Physiol. Biochem. Zool. 77, 688-695. Shine, R., 1980. "Costs" of reproduction in reptiles. Oecologia 46, 92-100. Shine, R., 2004. Adaptive consequences of developmental plasticity. In: Deeming, D.C. (Ed.), Reptilian Incubation: Environment, Evolution and Behaviour. Nottingham University Press, Nottingham, pp. 187-210. Shine, R., Sun, L.X., Kearney, M., Fitzgerald, M., 2002. Thermal correlates of foraging- site selection by Chinese pit-vipers (Gloydius shedaoensis, Viperidae). J. Therm. Biol. 27, 405-412. Sims, P.L., 1988. Grasslands. In: Barbour, M.G., Billings, W.D. (Eds.), North American Terrestrial Vegetation. Cambridge University Press, New York, pp. 265-286. Slip, D.J., Shine, R., 1988. Reptilian endothermy: a field study of thermoregulation by brooding diamond pythons. J. Zool. Lond. 216, 367-378. Sokal, R.R., Rohlf, F.J., 1995. Biometry: the principles and practice of statistics in biological research. 3rd ed. W.H. Freeman and Co., New York. Stevenson, R.D., 1985a. Body size and limits to the daily range of body temperature in terrestrial ectotherms. Am. Nat. 125, 102-117. Stevenson, R.D., 1985b. The relative importance of behavioral and physiological adjustments controlling body temperature in terrestrial ectotherms. Am. Nat. 126, 362-386. Taylor, E.N., DeNardo, D.F., Malawy, M.A., 2004. A comparison between point- and semi-continuous sampling for assessing body temperature in a free-ranging ectotherm. J. Therm. Biol. 29, 91-96. Tu, M.C., Hutchison, V.H., 1994. Influence of pregnancy on thermoregulation of water snakes {Nerodia rhombifera). J. Therm. Biol. 19, 255-259. Waldron, J.L., Lanham, J.D., Bennett, S.H., 2006. Using behaviorally-based seasons to investigate canebrake rattlesnake {Crotalus horridus) movement patterns and habitat selection. Herpetologica 62:389-398. 179 Webb, J.K., Shine, R., 1998. Using thermal ecology to predict retreat-site selection by an endangered snake species. Biol. Conserv. 86, 233-242. Webb, J.K., Whiting, M.J., 2005. Why don't small snakes bask? Juvenile broad-headed snakes trade thermal benefits for safety. Oikos 110, 515-522. Whitaker, P.B., Shine, R., 2002. Thermal biology and activity patterns of the eastern brownsnake (Pseudonaja textilis): a radiotelemetric study. Herpetologica 58, 436- 452. Wills, C.A., Beaupre, S.J., 2000. An application of randomization for detecting evidence of thermoregulation in timber rattlesnakes (Crotalus horridus) from northwest Arkansas. Physiol. Biochem. Zool. 73, 325-334. 180 Figure 1. Comparison of It, distributions for timber rattlesnakes occupying three distinct habitat types. Snake 7b's did not differ between woodland (median = 28.0 °C, range = 16.4 - 34.2 °C, n = 63), edge (median = 28.6 °C, range = 16.7 - 33.9 °C, n = 56), and field (median = 28.2 °C, range = 20.2 - 34.4 °C) habitats. Box plots illustrate the 10th, th th th th 25 , 50 , 75 , and 90 percentiles of the Th distribution. 181 182 Figure 2. Temporal distribution of timber rattlesnake Tb records during well-sampled hours of the day (ca. 1400 - 2030 hours). Body temperatures (n = 126) were collected from males and non-gravid females using woodland, edge, and field foraging habitats. During the sampling period, rb's ranged between 16.4 °C and 34.4 °C (mean 29.1 °C SD = 3.8) and many individuals maintained warm 71,'s into late afternoon/early evening hours. 183 max. voluntary Tb 184 Figure 3. Overall percentage of habitat thermally available for timber rattlesnakes (percentage of Te models recording temperatures of 34.5 °C and below) in both a woodland and field between 02 June and 26 August 2006. Data are averaged across all hours of the day and presented in ca. two week intervals to depict temporal changes in climate and vegetative growth during the sampling period. 185 100.0 90.0 80.0 70.0 60.0 50.0 40.0 30.0 20.0 10.0 0.0 153-167 168-181 182-195 196-209 210-223 224-238 Day of Year 186 Figure 4. Percentage of habitat thermally available for timber rattlesnakes (percentage of Te models recording temperatures of 34.5 °C and below) in both a woodland and field for each hour of the day. Data are presented in ca. two week intervals for 02 June - 16 June (top) and 17 June - 30 June (bottom) to depict temporal changes in climate and vegetative growth during the sampling period. 187 Days 153-167 10 15 Hour of Day Days 168-181 100 90 80 •fi « 70 a 60 bl e — '3 50 <> _>> 40 CS 30 1E- 10 15 20 Hour of Day 188 Figure 5. Percentage of habitat thermally available for timber rattlesnakes (percentage of Te models recording temperatures of 34.5 °C and below) in both a woodland and field for each hour of the day. Data are presented in ca. two week intervals for 01 July - 14 July (top) and 15 July - 28 July (bottom) to depict temporal changes in climate and vegetative growth during the sampling period. 189 Days 182-195 100 ^ 90 5? ^-' CS 80 £* * cs 70 E 60 bl e CS —« 50 <> >^ 40 C5 30 &£• u H 20 -*-- field 10 -v— woodland 10 15 20 Hour of Day Days 196-209 100 ^ 90 s—s?' cs 80 +- £es 70 X 60 bl e '—3 50 ?• «U £> 40 a B 30 •4-* H 20 -*- field 10 -v— woodland 10 15 20 Hour of Day 190 Figure 6. Percentage of habitat thermally available for timber rattlesnakes (percentage of Te models recording temperatures of 34.5 °C and below) in both a woodland and field for each hour of the day. Data are presented in ca. two week intervals for 29 July - 11 August (top) and 12 August - 26 August (bottom) to depict temporal changes in climate and vegetative growth during the sampling period. 191 Days 210-223 10 15 Hour of Day Days 224-238 10 15 Hour of Day 192 Figure 7. Comparison of 7b distributions from two behaviorally distinct segments of the study population. Body temperatures of gestating females (median = 32.5 °C, range = 25.3 - 34.7 °C, n = 113) were warmer and less variable than Jb's of males and non-gravid females (median = 28.3 °C, range = 16.4 - 34.4 °C, n = 180) using woodland, edge, and th th th th th field habitats. Box plots illustrate the 10 , 25 , 50 , 75 , and 90 percentiles of the Th distribution. 193 40 - 35 - -S- + 30 - _/ 25 ^ 1 20 - 1 0 15 - 10 - 5 - n u 1 1 gestating females males / non-gravid females Population Segment 194 CHAPTER 5: The timber rattlesnake {Crotalus horridus) exhibits rapid growth in a fragmented habitat Abstract In the mature forests of eastern North America, habitat loss and fragmentation are considered a primary threat to timber rattlesnake {Crotalus horridus) populations. Yet surprisingly, little is known about the life history of timber rattlesnakes in fragmented habitats. In this study, mark-recapture data and rattle morphology were used to compare the birth size and early growth rates of timber rattlesnakes in a fragmented habitat of west central Missouri (MO) to those in a closed-canopy forest of northwest Arkansas (AR). Climatological data indicate MO snakes typically have 8.5% less time to acquire and assimilate prey than snakes in the AR population. Snakes in the MO population are also born slightly, but significantly, shorter than those from AR. Nevertheless, MO snakes increase in length more rapidly than their AR counterparts through their first eight ecdyses. Furthermore, males and females from MO diverge in size between the fifth and sixth ecdysis event, while growth trajectories of AR males and females remain indistinguishable through eight ecdyses. In the MO population, female reproductive data indicate sexual size dimorphism occurs at the onset of sexual maturity. Results of a common garden study were consistent with our hypothesis that differences in growth between MO and AR snakes are influenced by local resource environments. Because prey populations inhabiting fields and edges are unlikely to fluctuate with mast crop production, timber rattlesnakes in fragmented habitats may exhibit favorable life history characteristics such as rapid growth and early maturation. Our work suggests forestry 195 practices that provide rodents with a diversity of food sources may be an important tool to those working to conserve declining populations of timber rattlesnakes. Introduction Geographic variation in life history exhibited by wide ranging species can often be attributed to proximate factors in the environment (Ballinger 1977; Dunham 1978; Seigel and Fitch 1985; Ford and Seigel 1989; Beaupre 1995, 2008; Angilletta 2001). Conditions such as climate, habitat structure, and resource environment may vary widely across a species' range and produce changes in morphology, growth, and reproduction among populations (O'Conner et al. 2006). Distinct populations are often subject to a unique suite of physical and biotic factors not experienced by the species as a whole (Dunham et al. 1989). When geographic variation in life history is widespread and not canalized by local population genetics, it may indicate that the species has a plastic phenotype capable of responding to recent changes in the environment. By studying populations occupying novel or disturbed habitats, researchers can gain insight as to how a species may respond to environmental perturbation elsewhere throughout its range (Bury 2006). Habitat loss and fragmentation have been identified as important factors in the global decline of reptile populations (Gibbons et al. 2000). Nevertheless, reptiles are the least represented of any terrestrial vertebrate group in studies of organisms residing in fragmented habitats (Mac Nally and Brown 2001). One such reptile, the timber rattlesnake {Crotalus horridus), has vanished from many parts of its formerly widespread range throughout the deciduous forests of eastern North America (Brown 1993). Despite 196 being afforded a degree of protection in many states, this species has been extirpated from Rhode Island, Maine, and possibly the Canadian province of Ontario (Brown 1993). Most consider the timber rattlesnake to prefer closed-canopy ecosystems (Collins and Knight 1980; Reinert 1984; Reinert and Zappalorti 1988; Martin 1992a; Rudolph et al. 1998), and the loss and fragmentation of woodland habitat has been cited as a primary threat to the survival of the species (Martin 1992a; Brown 1993; Clark et al. 2003; Furman 2007). Although the loss of large, continuous tracts of woodland habitat is considered harmful to the timber rattlesnake, populations in the Midwest are known to inhabit areas devoid of large, continuous tracts of closed-canopy forest (Fitch and Shirer 1971, Fitch 1999; Fogell 2000; Fitch and Pisani 2002, 2005, 2006; Fogell et al. 2002; Fitch et al. 2004; Pisani and Fitch 2006). In fact, recent studies of this species in eastern Kansas (Fitch and Pisani 2006; Pisani and Fitch 2006) suggest that individuals are capable of rapid growth in diverse, fragmented landscapes that consist of both woodland and open habitat. To date, most discussions of timber rattlesnake growth have focused on the effects of climate (Brown 1991; Martin 1993, 2002). Researchers have concluded that the short active seasons experienced by individuals from high latitudes and elevation limit the time available for resource acquisition, therefore resulting in slow growth and delayed sexual maturity (Brown 1991; Martin 1993, 2002). Although such assertions are valid, the quality of the local resource environment may be just as important to individual growth as seasonal constraints on foraging (Beaupre 2002). Individuals in poor resource environments will spend a disproportionate amount of their activity budget foraging 197 (Beaupre 2008) and have less available energy to allocate to the competing functions of growth, maintenance, storage, and reproduction (Dunham et al. 1989). In this study, mark-recapture data and rattle morphology are used to compare the birth size and early growth rates of timber rattlesnakes in a fragmented habitat of west central Missouri (MO) to those in a closed-canopy forest of northwest Arkansas (AR). Latitudinal differences in active season length slightly favor the AR population, thereby providing individuals more time to assimilate resources and potentially grow faster. Therefore, if detected, superior individual growth rates of snakes in the MO population can likely be attributed to an enhanced resource environment, genetic differences at the population level, or both. Our findings suggest that 1) life history characteristics of local timber rattlesnake populations may be influenced by environmental factors other than active season length, 2) populations found outside of mature, closed-canopy forests can be quite viable and 3) those working to reestablish timber rattlesnakes in closed-canopy ecosystems should consider enhancing the local prey base through silviculture practices that promote habitat heterogeneity. Materials and methods Field sites The MO study population was located in a highland region of west-central Missouri in what is considered the prairie geographic region (Schwartz and Schwartz 1959). The highlands consist of hills containing hardwood forests that extend across a flat landscape of what was once tall grass prairie (Schwartz and Schwartz 1959; Sims 1988). Nearly every habitat component used by this timber rattlesnake population is 198 anthropogenic. An approximately 32 hectare inactive rock quarry provides surrogate hibernacula for over-wintering snakes, and remnant boulder piles are used as rookery habitat for gestating females. Following spring emergence, individuals disperse from the quarry to utilize a mosaic of woodland, agriculture, and residential habitat. Snakes forage in small woodlots, wooded fencerows, roadside vegetation, old fields, and row crops. A study of the movement patterns, habitat use, thermal biology, and foraging ecology of individuals in this population was conducted between the spring of 2003 and 2007. The AR study population is located at the Ozark Natural Science Center, Bear Hollow Natural Area (Arkansas Natural Heritage Commission) and adjacent Madison County Wildlife Management Area in the northwest region of the state. The rolling, heavily forested, rocky hills of the Ozarks characterize the habitat typical for this species across much of its range. Limestone outcroppings serve as hibernacula, while gestation often occurs in cedar glades that provide both exposed rock and openings in the tree canopy. Foraging occurs on the floor of this late succession forest, among a dense layer of leaf litter and deadfall (Cundall and Beaupre 2001). A study of the physiological ecology of individuals in the AR population has been underway since the fall of 1995 (Wills and Beaupre 2000; Beaupre and Zaidan III 2001; Zaidan III and Beaupre 2003; Beaupre 2008; Gardner-Santana and Beaupre 2009). Data collection All snakes encountered in the field were collected and processed as part of the mark-recapture studies being conducted on both populations. Snakes were restrained in a squeezebox (Quinn and Jones 1974) for sexing, marking, and obtaining morphological 199 data. Snout-vent length was measured to the nearest 1.0 mm with a flexible metric tape and sex was determined by caudal probing. The number of segments contained in each rattle string and the rattle condition (either intact or damaged) were recorded. Digital calipers were used to measure the width (± .001 cm) of each rattle segment. Potential effects of climate on resource acquisition Both MO and AR populations exhibit a similar length of active season, with egress typically beginning in April and ingress concluding by mid October. However, because the populations are separated by 260 km along a latitudinal gradient, National Weather Service freeze/frost probability tables for locations adjacent to each study site were used to estimate potential climate-based differences in the time available for snakes to acquire resources. Tables were derived from climatological normal data spanning the 30-year period between 1971 and 2000 and present first (fall) and last (spring) frost dates associated with 90, 50, and 10 percent probabilities. Spring and fall dates associated with the conservative 10% frost probability were chosen as a crude measure of resource acquisition period. The 10% probability means there is less than a 10% chance of the air temperature reaching 2.2°C on/after the spring date, or on/before the fall date. Rattle morphology as an indicator of growth Most rattlesnakes belonging to the genera Crotalus and Sistrurus possess a functional structure at the distal portion of their tail known as the rattle (Klauber 1972). The rattle consists of a series of interlocking keratinized segments that can be vibrated at a high rate of speed by special shaker muscles in the tail to provide an acoustic warning 200 to other species (Moon 2001). Although several evolutionary hypotheses have been proposed to explain the rattle's origin, behavioral studies on extant species indicate it may have evolved to deter predators and prevent snakes from being stepped on by large ungulates (Klauber 1972; Greene 1988). The rattle begins as a single structure known as the prebutton in newborn snakes. After young snakes undergo their first ecdysis event, known as the postnatal shed, the prebutton is lost and replaced by the button. In timber rattlesnakes, the postnatal shed typically occurs 8-16 days after birth (Martin 1992b). Each subsequent ecdysis event results in the formation of an additional rattle segment at the proximal portion of the rattle (referred to as the basal segment). The button remains the terminal segment on this growing string of rattle segments unless the rattle is physically damaged. The width of the basal segment is determined by the width of the tail at its most distal portion. Thus, as longer snakes will have thicker tails, snout-vent length (SVL) has been shown to positively correlate with basal segment width (Klauber 1972; Fitch and Pisani 1993, 2006; Beaupre et al. 1998). Because each segment in a rattle string was at one time a basal segment, snakes with unbroken rattle strings carry with them a complete history of their growth on a per shed basis. The rattles of individuals that are growing in length between successive sheds (i.e. young snakes) will have a distinctively tapered profile, while individuals not experiencing recent gains in length (i.e. old snakes) will possess a non-tapering rattle comprised of segments of relatively even width. As mark-recapture studies of snakes are often plagued by low recapture rates (Parker and Plummer 1987), the relationship between basal segment width and SVL has proven useful in studies of rattlesnake growth (Klauber 1972; Fitch and Pisani 1993, 2006; Beaupre et al. 1998). 201 By treating the width of each segment of an intact rattle string as a repeated measurement of size (SVL) taken through time, Beaupre et al. (1998) used a repeated measures analysis to study the onset of sexual size dimorphism within a population of western diamondback rattlesnakes {Crotalus atrox). To compare the early growth rates of MO and AR snakes in this study, we wished to apply a similar repeated measures analysis. However, to validate this approach we must first establish three things. First, because we are analyzing individual growth through time, we must determine if MO and AR snakes differ in size at birth. Second, a significant relationship between basal segment width and SVL must be detected before rattle segment width can be used as a reliable predictor of SVL. Third, in order to compare growth between sites, the relationship between basal segment width and SVL must not differ between the MO and AR populations. Common garden study In the fall of 2003, eight neonates from three different litters (3,3, and 2 neonates per litter) were retained from the MO population for a comparative growth study. Neonates from the AR population were not available in 2003, however, a single litter of eight neonates were taken from a private collector in 2004. The mother to this litter was collected in a county adjacent to the AR study site. Using this opportunity to make a preliminary investigation into genetically based differences in growth, all 16 individuals were maintained under common garden conditions at the University of Arkansas. Snakes were kept in a single room and housed individually in 10 gallon glass aquaria. Each aquarium contained a newspaper for substrate, an aluminum foil hidebox, and a large 202 glass bowl in which water was provided ad libitum. Snakes were never placed in artificial hibernation and kept active throughout the study at a room temperature of approximately 25 degrees Celsius. A window ensured that for much of the year, snakes were exposed to the natural AR photoperiod. During winter months, a light timer maintained a 12:12 light to dark cycle which was sufficient to prevent snakes from becoming aphagic. Prey masses were recorded at each feeding so that differences in consumption could be investigated if growth rates differed between populations. Approximately three years of growth data were obtained from both the MO and AR snakes. Using the measurement technique described above for field subjects, the SVL of each neonate was recorded at birth and at four subsequent points in time when gains in length became apparent. For all individuals, the interval between the initial and final SVL measurement ranged between 1,055 and 1,256 days. Data analysis A one-factor analysis of variance (ANOVA) was used to determine if SVL at birth differed between populations (fixed effect) or among litters nested within population (random effect). The analysis was restricted to litters containing a minimum of three offspring. We tested nine litters from MO (73 offspring) and twenty from AR (194 offspring). Litters were typically born in captivity to gravid females placed in the lab prior to parturition; however, post partum females and their newborn offspring were occasionally collected in the field. Using all snakes with complete rattle strings consisting of the terminal button and up to ten segments (MO = 68, AR = 147), we used analysis of covariance (ANCOVA) to 203 1) determine if basal segment width (covariate) and SVL were significantly correlated and 2) test the consistency of the relationship between basal segment width and SVL across both populations. Basal segment width and SVL were logio-transformed prior to using ANCOVA. Next, the repeated measures approach of Beaupre et al. (1998) was used to compare the growth of MO and AR snakes on a per shed basis. We investigated potential effects of both population and sex on rattle segment width. To do this, we conducted a repeated measures ANOVA using PROC MIXED because it handles missing data (complete rattle strings did not always contain the same number of rattle segments) and an unbalanced design (AR samples outnumbered MO samples). In order to obtain the most comprehensive analysis of growth from both populations, data were pooled across all years of study (MO = 4, AR = 12). Repeated measurements of SVL recorded through time were used to compare the growth rates of MO and AR snakes in the common garden study. A repeated measures ANCOVA, treating SVL measured at birth as a covariate, was used to determine if population significantly affected rate of growth. A type I error of 0.05 was applied for all hypothesis tests. All statistical analyses were performed using SAS (v. 9.3.1, SAS Institute). Results Potential effects of climate on resource acquisition A comparison of frost free intervals at the MO (16 May to 21 September, 129 days) and AR (08 May to 25 September, 141 days) sites revealed a slight difference of 12 204 days. Thus, AR individuals may potentially have an average of 8.5% more time each season to acquire and assimilate prey. Size at birth A one-factor analysis of variance indicated that offspring from MO litters were born significantly smaller than those from AR (P < 0.0001; MO = 29.2 cm ± 1.86 SD; AR = 31.3 cm ± 2.02 SD). The random effect of litter nested within population was also significant (P < 0.0001). While SVL at birth appears to be influenced by genetic variation and/or maternal effects in both populations, MO litters typically contain slightly smaller offspring. Basal segment width and SVL correlation Analysis of covariance detected a significant relationship between logio basal segment width and logio SVL which did not differ between populations as indicated by a non significant interaction between logio basal segment width and population (ANCOVA, model P < 0.0001; R2 = 0.97; logio basal segment width effect, P < 0.0001; population effect, P = 0.0145; logio basal segment width x population interaction, P = 0.3278; n = 215; Fig. 1). Therefore, basal segment width is a good predictor of SVL and direct comparisons can be made between MO and AR populations. A significant main effect of population on logio SVL indicates the mean SVL of snakes in the MO and AR samples differed. This was not surprising, as snakes were included in the analysis based on whether or not they possessed complete rattle strings, regardless of their SVL. 205 Comparison of individual growth rates In our comparison of rattle segment widths between MO and AR populations, a repeated measures analysis of variance detected a significant rattle segment number x population x sex interaction (P = 0.0110; Table 1). Plotting this three-way interaction shows that despite being born smaller, MO snakes increase in length more rapidly than AR snakes through their first eight ecdysis events (Fig 2). Furthermore, male and female growth trajectories in the MO population clearly diverge between the fifth and sixth ecdysis event (Fig 2). As expected, growth between ecdysis events resulted in a significant increase in width between successive rattle segments (segment number effect, P < 0.0001; Table 1). Because the analysis was designed to compare the response variable (segment width) at each specific segment number, significant main effects of sex (P = 0.0369) and population (P = 0.0046) on rattle segment width cannot be meaningfully interpreted. Common Garden Analysis After correcting for differences in initial size by treating birth SVL as a covariate, growth trajectories of MO and AR snakes in the common garden experiment did not differ significantly (ANCOVA, time effect, P < 0.0001; population effect, P = 0.0849; time x population interaction, P = 0.0684). Although MO snakes were born smaller, increases in SVL occurred at a rate similar to the AR snakes. Comparable growth demonstrated by these individuals supports the hypothesis that genetic differences in growth between populations are not meaningful. 206 Discussion At birth, timber rattlesnakes from the MO population are slightly, but significantly, shorter than those from AR. However, significant differences in rattle segment width suggest that MO snakes, regardless of sex, increase in length more rapidly than their AR counterparts through their first eight ecdysis events (Fig. 2). Furthermore, males and females from MO diverge in size between the fifth and sixth ecdysis event, while growth trajectories of AR males and females remain indistinguishable through eight ecdyses (Fig. 2). In western diamondback rattlesnakes (Crotalus atrox), sexual size dimorphism becomes apparent at the onset of sexually maturity, as males continue to grow while females begin allocating most incoming energy to reproduction (Beaupre et al. 1998; Duvall and Beaupre 1998). Seven MO females with complete rattle strings were known to be reproductively mature after having undergone 6, 8 (n = 3), 9 (n = 2), and 11 ecdyses. These females were either follicular (containing ovarian follicles), gravid, or post partum. Therefore, it is likely that the appearance of sexual size dimorphism in the MO population is associated with sexual maturity, indicating MO snakes mature before those in AR. Differential resource environments (prey abundance) may explain the rapid growth and maturation of snakes in the MO population. Although limited by sample size, our common garden study failed to detect a difference in individual growth rate between the populations. Other studies have shown that prey availability has a greater effect on snake growth than genetic differences among individuals (Madsen and Shine 1993, 2000; Forsman and Lindell 1996; but see also Bronikowski 2000). This is not surprising given that snake growth is indeterminate, meaning that snakes grow throughout their life and 207 food intake strongly influences the rate at which they grow (Andrews 1982). Furthermore, individually-based computer simulations that model time-energy allocation strategies in vipers suggest that a large degree of growth variation in timber rattlesnakes can be accounted for by variation in their resource environment (Beaupre 2002). At the MO study site, 157 capture events consisted of only eleven recaptures. However, one noteworthy recapture directly supports the results of our growth analysis using rattle morphology. In September of 2004, a female neonate measuring 33.9 cm SVL and weighing 25.19 g was collected from beneath a maternal rock where she had remained with her mother following birth. Upon recapture in September of 2006, this young female had attained a length of 70.8 cm SVL and a mass of 291.54 g after only two full active seasons. Her rattle profile consisted of four segments plus the button, indicating she had undergone ecdysis five times since her birth. The mean size of reproductive females in this population is 83.4 cm SVL (n = 30, ± 6.71 SD), however, the minimum size at reproduction is only 72.5 cm SVL. It is quite possible that this female would have acquired a reproductive body condition during her third active season and reproduced at the conclusion of her fourth. Additionally, timber rattlesnakes using a mixture of woodland and grassland habitat in nearby eastern Kansas exhibit growth rates quite comparable to our MO population (Fitch and Pisani 2006; Pisani and Fitch 2006). The growth exhibited by our MO female matches that of the species in west-central South Carolina, where Gibbons (1972) reports individuals reaching 65-75 cm SVL by the end of their second summer. The similarity in early growth rates between MO and South Carolina snakes is somewhat surprising, considering that timber rattlesnakes in the southeastern Coastal Plain may be active for 7-11 months each year (Martin 1992a). 208 Telemetry data confirms that MO snakes readily forage in edge and open field habitat (Wittenberg chapter 3). Additional observations of timber rattlesnakes foraging in open fields have been reported in Kansas (Fitch and Pisani 2006; Pisani and Fitch 2006) and Nebraska (Fogell 2000; Fogell et al. 2002). A dietary study revealed that prairie voles (Microtus ochrogaster) were the predominate species consumed by the MO population, and were trapped exclusively in uncut grassy fields (Wittenberg chapter 3). Interestingly, Pisani and Fitch (2006) suggest prairie vole abundance may explain the rapid growth of individuals in their KS population. Schwartz and Schwartz (1959) considered the mammalian fauna of Missouri's prairie geographic region to be very diverse and supported by both natural and anthropogenic factors. The authors suggested that the historical intermingling of two major vegetation types (grassland and forest), as well as the habitat heterogeneity resulting from human land use practices, maintained a landscape capable of supporting both forest and grassland species. Small mammal trapping at the MO site supported this assertion. Woodlots contained an abundance of white-footed mice {Peromyscus leucopus) and wooded fencerows separating fields supported large numbers of both white-footed mice and hispid cotton rats (Sigmodon hispidus) (Wittenberg chapter 3). Physiological and behavioral data suggest that snakes in the AR population endure periods of low food availability. Under these conditions, foraging time increases as snakes secure fewer meals and exhibit decreases in growth, lower field metabolic rates, and poor body condition (Beaupre 2008). Furthermore, reproductive activity declines as there is little evidence of mate searching, courtship, or follicular development among food-stressed individuals (Beaupre 2008). Woodland small mammal populations 209 have been shown to fluctuate with annual variations in acorn production (Wolff 1996; McShea 2000), while the agricultural practices associated with mixed habitats may weaken the relationship between mast production and rodent densities. In Virginia during the first half of the 20th century, Martin (1979) describes how rattlesnake populations responded when abandoned farms fell under government control during the establishment of Shenandoah National Park and the George Washington and Jefferson National Forests. The author remarks, "Old fields with their rock piles and stone walls became overgrown. Rodents multiplied and rattlers followed, in artificially restored favorable environments." Clearcutting within mature forests has indeed been shown to increase small mammal densities (Kirkland 1977, 1990), and may be an effective way to restore timber rattlesnake populations that are currently in decline. Ultimately, the response of timber rattlesnakes to the fragmentation of their forested habitats may be population-specific. A telemetry study in West Virginia found that individuals used anthropogenic habitat edges as well as clearcuts, leading the author to suggest that a degree of clearcutting may be beneficial to the species (Adams 2005). However, telemetry subjects in the Piedmont region of North Carolina tended to avoid open canopy habitats (Sealy 2002). In South Carolina, both males and nongravid females prefer to forage in forested habitat but primarily use fields during the mating season (Waldron et al. 2006). Because timber rattlesnakes evolving in geographic regions historically containing grassland may be pre-adapted to using non-forested habitat (Fogell 2000), researchers cannot assume that populations in mature forests of the eastern United States will respond favorably to changes in habitat structure. 210 Although climatic constraints on the time available for timber rattlesnakes to acquire and assimilate prey undoubtedly influence their rate of growth (Brown 1991; Martin 1993, 2002), we suggest other environmental factors, such as prey abundance, may be just as important. Because fragmented habitats may actually support a prey base that is unlikely to fluctuate with mast crop production, favorable life history characteristics such as rapid growth and early maturation may be promoted. In fact, Taylor et al. (2005) have shown that when supplementally fed, free-ranging female Western diamondback rattlesnakes {Crotalus atrox) exhibit greater gains in length and mass, reproduce more frequently, and have better postpartum body conditions than females not receiving supplemental feedings. These life history characteristics should positively impact the stability and viability of rattlesnake populations. However, we are not suggesting that habitat fragmentation (of which type, cause, and extent can infinitely vary) is beneficial to the timber rattlesnake. For instance, snakes in the MO population incur mortality from human sources that the AR snakes do not. MO snakes were struck by vehicles, destroyed by landowners, and four individuals in the telemetry study were killed by farm implements while foraging in fields (Wittenberg chapter 2). Still, we agree with Adams (2005) and suggest that timber rattlesnakes within mature, late succession forests may benefit from silvicultural practices that promote a diversity of rodent food sources. Our research underscores the importance of studying the life history characteristics of organisms not only in pristine habitats, but also in disturbed landscapes, should the species occur there. Acknowledgments 211 The study of MO Timber Rattlesnakes was conducted with the approval of the University of Arkansas Institutional Animal Care and Use Committee (protocol # 05001) and Missouri Department of Conservation (collecting permits # 12005, 12367, 12715, 13101). The study of AR Timber Rattlesnakes was conducted with the approval of, and all relevant permits from, the University of Arkansas Institutional Animal Care and Use Committee, the Arkansas Game and Fish Commission, the Arkansas Natural Heritage Commission, and the Ozark Natural Science Center. J. Agugliaro, L. Douglas, E. Gbur Jr., J. Ortega, I. Pinto, K. Smith, M. Smith, J. Van Dyke, and S. Wittenberg provided valuable comments on an early draft of this manuscript. We thank numerous private landowners in MO, the Ozark Natural Science Center, the Arkansas Natural Heritage Commission, and the Arkansas Game and Fish Commission for allowing this work to be conducted on their property. Research on the MO population was supported in part by a University of Arkansas Causey Grant-in-Aid Award, a Harry Steinman Memorial Grant from the St. Louis Herpetological Society, and a grant from the Arkansas Audubon Society Trust. Research on the AR population was supported by funds from the University of Arkansas Research Incentive Fund, the Arkansas Science and Technology Authority (grant # 97-B-06), and the National Science Foundation (grants # IBN- 9728470, and # IBN-0130633, and #IBN-0641117 to SJB). References Adams JP (2005) Home range and behavior of the timber rattlesnake (Crotalus horridus). Unpubl. Master's Thesis. Marshall University Andrews RM (1982) Patterns of growth in reptiles. In: Gans C, Pough FH (eds) Biology of the Reptilia. vol. 13 Academic Press, New York, pp 273-320 212 Angilletta MJ (2001) Thermal and physiological constraints on energy assimilation in a widespread lizard (Sceloporus undulatus). Ecology 82:3044-3056 Ballinger RE (1977) Reproductive strategies: food availability as a source of proximal variation in a lizard. Ecology 58:628-635 Beaupre SJ (1995) Effects of a geographically variable thermal environment on bioenergetics of mottled rock rattlesnakes. Ecology 76:1655-1665 Beaupre SJ (2002) Modeling time-energy allocation in vipers: individual responses to environmental variation and implications for populations. In: Schuett GW, Hoggren M, Douglas ME, Greene HW (eds) Biology of the vipers. Eagle Mountain Publishing, Eagle Mountain, pp 463-481 Beaupre SJ (2008) Annual variation in time-energy allocation by timber rattlesnakes (Crotalus horridus) in relation to food acquisition. In: Hayes WK, Beaman KR, Cardwell MD, Bush SP (eds) Biology of the Rattlesnakes. Loma Linda University Press, Loma Linda, pp 111 -122 Beaupre SJ, Duvall D, O'Leile J (1998) Ontogenetic variation in growth and sexual size dimorphism in a central Arizona population of the western diamondback rattlesnake (Crotalus atrox). Copeia 1998:40-47 Beaupre SJ, Zaidan III F (2001) Scaling of CO2 production in the timber rattlesnake {Crotalus horridus), with comments on cost of growth in neonates and comparative patterns. Physiol Biochem Zool 74:757-768 Bronikowski AM (2000) Experimental evidence for the adaptive evolution of growth rate in the garter snake Thamnophis elegans. Evolution 54:1760-1767 Brown WS (1991) Female reproductive ecology in a northern population of the timber rattlesnake, Crotalus horridus. Herpetologica 47:101-115 Brown WS (1992) Emergence, ingress, and seasonal captures at dens of northern timber rattlesnakes, Crotalus horridus. In: Campbell JA, Brodie Jr. ED (eds) Biology of the Pitvipers. Selva, Tyler, pp 251-258 Brown WS (1993) Biology, status, and management of the timber rattlesnake (Crotalus horridus): a guide for conservation. SSAR Herpetol Circ No. 22:1-78. Society for the Study of Amphibians and Reptiles. Univ. of Kansas, Lawrence Bury RB (2006) Natural history, field ecology, conservation biology and wildlife management: time to connect the dots. Herpetol Cons Biol 1:56-61 213 Clark AM, Moler PE, Possardt EE, Savitzky AH, Brown WS, Bowen BW (2003) Phylogeography of the timber rattlesnake (Crotalus horridus) based on mtDNA sequences. J Herpetol 37:145-154 Collins JT, Knight JL (1980) Crotalus horridus. Catalogue of American Amphibians and Reptiles 253:1 -2 Cundall D, Beaupre SJ (2001) Field records of predatory strike kinematics in timber rattlesnakes, Crotalus horridus. Amphibia-Reptilia 22:492-498 Dunham AE (1978) Food availability as a proximate factor influencing individual growth rates in the iguanid lizard Sceloporus merriami. Ecology 59:770-778 Dunham AE, Grant BW, Overall KL (1989) Interfaces between biophysical and physiological ecology and the population ecology of terrestrial vertebrate ectotherms. Physiol Zool 62:335-355 Duvall D, Beaupre SJ (1998) Sexual strategy and size dimorphism in rattlesnakes: integrating proximate and ultimate causation. Amer Zool 38:152-165 Fitch HS (1999) A Kansas snake community: composition and changes over 50 years. Kreiger Publishing Company, Malabar Fitch HS, Pisani GR (1993) Life history traits of the western diamondback rattlesnake (Crotalus atrox) studied from roundup samples in Oklahoma. Univ Kans Mus Nat Hist Occ Pap 156:1-24 Fitch HS, Pisani GR (2002) Longtime recapture of a timber rattlesnake (Crotalus horridus) in Kansas. Journal of Kansas Herpetology 3:15-16 Fitch HS, Pisani GR (2005) Disappearance of radio-monitored timber rattlesnakes. Journal of Kansas Herpetology 14:14-15 Fitch HS, Pisani GR (2006) The timber rattlesnake in northeastern Kansas. Journal of Kansas Herpetology 19:11-15 Fitch HS, Pisani GR, Greene HW, Echelle AF, Zerwekh M (2004) A field study of the timber rattlesnake in Leavenworth county, Kansas. Journal of Kansas Herpetology 11:18-24 Fitch HS, Shirer HW (1971) A radiotelemetric study of spatial relationships in some common snakes. Copeia 1971:118-128 Fogell DD (2000) Seasonal activity, habitat preferences, and natural history of the timber rattlesnake (Crotalus horridus) in southeastern Nebraska. Unpubl. Master's Thesis. The University of Nebraska at Omaha 214 Fogell DD, Leonard TJ, Fawcett JD (2002) Crotalus horridus horridus (timber rattlesnake) habitat. Herpetol Rev 33:211-212 Ford NB, Seigel RA (1989) Phenotypic plasticity in reproductive traits: evidence from a viviparous snake. Ecology 70:1768-1774 Forsman A, Lindell LE (1996) Resource dependent growth and body condition dynamics in juvenile snakes: an experiment. Oecologia 108:669-675 Furman J (2007) Timber Rattlesnakes in Vermont and New York: biology, history, and the fate of an endangered species. University Press of New England, Lebanon Gardner-Santana LC, Beaupre SJ (2009) Timber Rattlesnakes (Crotalus horridus) exhibit elevated and less variable body temperatures during pregnancy. Copeia 2009:363-368 Gibbons JW (1972) Reproduction, growth, and sexual dimorphism in the canebrake rattlesnake (Crotalus horridus atricaudatus). Copeia 1972:222-226 Gibbons JW, Scott DE, Ryan TJ, Buhlmann KA, Tuberville TD, Metts BS, Greene JL, Mills T, Leiden Y, Poppy S, Winne CT (2000) The global decline of reptiles, deja vu amphibians. BioScience 50:653-666 Greene HW (1988) Antipredator mechanisms in reptiles. In: Gans C, Huey RB (eds) Biology of the Reptilia. vol. 16 Alan R. Liss, New York, pp 1-152 Kirkland Jr. GL (1977) Responses of small mammals to the clearcutting of northern Appalachian forests. J Mammal 58:600-609 Kirkland Jr. GL (1990) Patterns of initial small mammal community change after clearcutting of temperate North American forests. Oikos 59:313-320 Klauber LM (1972) Rattlesnakes: their habits, life histories, and influence on mankind. University of California Press, Berkeley Mac Nally R, Brown GW (2001) Reptiles and habitat fragmentation in the box-ironbark forests of central Victoria, Australia: predictions, compositional change and faunal nestedness. Oecologia 128:116-125 Madsen T, Shine R (1993) Phenotypic plasticity in body sizes and sexual size dimorphism in European grass snakes. Evolution 47:321-325 Madsen T, Shine R (2000) Silver spoons and snake body sizes: prey availability early in life influences long-term growth rates of free ranging pythons. J Anim Ecol 69:952-958 215 Martin WH (1979) The timber rattlesnake in Virginia: its distribution and present status. Virginia Herpetological Society Bulletin 89:1-4 Martin WH (1992a) The timber rattlesnake: its distribution and natural history. In: Tyning T.F. (ed) Conservation of the timber rattlesnake in the northeast. Massachusetts Audubon Society, Lincoln, pp 13-22 Martin WH (1992b) Phenology of the timber rattlesnake {Crotalus horridus) in an unglaciated section of the Appalachian Mountains. In: Campbell JA, Brodie Jr. ED (eds) Biology of the Pitvipers. Selva, Tyler, pp 259-277 Martin WH (1993) Reproduction of the timber rattlesnake {Crotalus horridus) in the Appalachian Mountains. J Herpetol 27:133-143 Martin WH (2002) Life history constraints on the timber rattlesnake {Crotalus horridus) at its climatic limits. In: Schuett GW, Hoggren M, Douglas ME, Greene HW (eds) Biology of the Vipers. Eagle Mountain Publishing, Eagle Mountain, pp 285-306 McShea WJ (2000) The influence of acorn crops on annual variation in rodent and bird populations. Ecology 81:228-238 Moon BR (2001) Muscle physiology and the evolution of the rattling system in rattlesnakes. J Herpetol 35:497-500 O'Conner MP, Sieg AE, Dunham AE (2006) Linking physiological effects on activity and resource use to population level phenomena. Integr Comp Biol 46:1093-1109 Parker WS, Plummer MV (1987) Population ecology. In: Seigel RA, Collins JT, Novak SS (eds) Snakes: ecology and evolutionary biology. McGraw-Hill, New York, pp 253-301 Pisani GR, Fitch HS (2006) Rapid early growth in northeastern Kansas timber rattlesnakes. Journal of Kansas Herpetology 20:19-20 Quinn H, Jones JP (1974) Squeeze box technique for measuring snakes. Herpetol Rev 5:35 ReinertHK (1984) Habitat separation between sympatric snake populations. Ecology 65:478-486 Reinert HK, Zappalorti RT (1988) Timber rattlesnakes {Crotalus horridus) of the pine barrens: their movement patterns and habitat preference. Copeia 1988:964-978 Rudolph DC, Burgdorf SJ, Conner RN, Dickson JG (1998) The impact of roads on the timber rattlesnake, {Crotalus horridus), in eastern Texas. Pages 236-240 in G. L. 216 Evink, P. Garrett, D. Zeigler and J. Berry, editors. Proceedings of the International Conference on Wildlife Ecology and Transportation. FL-Er-69-98, Florida Department of Transportation, Tallahassee, Florida Schwartz CW, Schwartz ER (1959) The Wild Mammals of Missouri. University of Missouri Press, Columbia Sealy JB (2002) Ecology and behavior of the timber rattlesnake {Crotalus horridus) in the upper Piedmont of North Carolina: identified threats and conservation recommendations. In: Schuett GW, Hoggren M, Douglas ME, Greene HW (eds) Biology of the vipers. Eagle Mountain Publishing, Eagle Mountain, pp 561-578 Seigel PvA and Fitch HS (1985) Annual variation in reproduction in snakes in a fluctuating environment. J Anim Ecol 54:497-505 Sims PL (1988) Grasslands. In: Barbour, MG, Billings, WD (eds), North American terrestrial vegetation. Cambridge University Press, New York, pp 265-286 Taylor EN, Malawy MA, Browning DM, Lemar SV, DeNardo DF (2005) Effects of food supplementation on the physiological ecology of female Western diamond- backed rattlesnakes {Crotalus atrox). Oecologia 144:206-213 Waldron JL, Lanham JD, Bennett SH (2006) Using behaviorally-based seasons to investigate canebrake rattlesnake {Crotalus horridus) movement patterns and habitat selection. Herpetologica 62:389-398 Wills CA, Beaupre SJ (2000) An application of randomization for detecting evidence of thermoregulation in timber rattlesnakes {Crotalus horridus) from northwest Arkansas. Physiol Biochem Zool 73:325-334 Wolff JO (1996) Population fluctuations of mast-eating rodents are correlated with production of acorns. J Mammal 77:850-856 Zaidan III F, Beaupre SJ (2003) Effects of body mass, meal size, fast length, and temperature on specific dynamic action in the timber rattlesnake {Crotalus horridus). Physiol Biochem Zool 76:447-458 217 Table 1. Results of a repeated measures ANOVA testing the effects of population and sex on rattle segment width. 218 Effect Numerator Denominator F Value Pr > F df df Population I 200 8.22 0.0046 Sex ]I 200 4.42 0.0369 Population x sex i 200 0.07 0.7942 Segment # I 539 4499.32 <.0001 Segment # x population 539 27.52 <.0001 Segment # x sex 539 43.65 <.0001 Segment # x population x sex 539 6.52 0.011 219 Figure 1. Basal segment width correlates positively with snout-vent length (SVL), and the relationship is consistent between MO and AR timber rattlesnake populations. 220 • Arkansas o Missouri 1.4 1.6 1.8 2.0 Log SVL (cm) 221 Figure 2. Significant rattle segment x population x sex interaction indicates that MO timber rattlesnakes grow more rapidly and mature earlier than snakes from the AR population. Segment "0" represents the terminal button (the first ecdysis), and each segment thereafter corresponds with a subsequent ecdysis event. Sexual size dimorphism in snout-vent length becomes apparent in the MO population between the fifth and sixth ecdysis. By contrast, growth trajectories of male and female snakes in the AR population remain similar through the first eight ecdyses. 222 E 1.6 o S 1-4-1 "E 1-2 -I CD E S? i.o CO _CD AR Males S 0.8 AR Females CD MO Males MO Females S 0.6 H CD 0 4 6 8 10 Segment Number 223 CONCLUSION Although the Timber Rattlesnakes in this study inhabited an area fragmented by agriculture, their ecology was similar to that of forest-dwelling populations. Thermal constraints on activity were detected and several sources of human-induced mortality were identified. Woodlands, fields, and edges contained ample prey and snakes were capable of rapid early growth. Because this research was conducted in a prairie- deciduous forest ecotone region of the Midwest, the study animals may have descended from Timber Rattlesnakes adept at using open-canopy habitat. Nevertheless, results of this study suggest that enhancing small mammal densities through habitat management may be a more effective strategy for conserving Timber Rattlesnake populations than strictly attempting to preserve large tracts of mature forest. 224 g- c C« c 60 3 c" ca i^ -5 o s _ca o ca •s u c« E vi W ; "3 E ~3o U o ca coj ^i- /—^ >,-e select r ns " ee n grow t hy , fe e rpin g indivi d II 1 ^ «^ c/1 0>0. E 1? 60 .D O CJ _o -3 -2 3 PH "o ibe : c CO g J3 3 •- o f h e -3 surviva l ogra p thof i i bet w abita t 'seas o '** 'r"„ « If tn o 60 cj E 5 .c 60 u .3 ^ or p o. on g log y 'C 'o o ^o tl ca ca o E P o o o •a cj l/> J3 p. .is 3 K PH 5 •> , u £ cd ol« zfi O a ^ D. 3 •a W, O CO C u &0 td E a> co 3 £ >- K a! 60 sa -o c