PHOTOCOPY AND USE AUTHORIZATION

Transcription

1 PHOTOCOPY AND USE AUTHORIZATION In presenting these doctoral papers in partial fulfillment of the requirements for an advanced degree at Idaho State University, I agree that the Library shall make it freely available for inspection. I further state that permission for extensive copying of my thesis for scholarly purposes may be granted by the Dean of Graduate Studies, Dean of my academic division, or by the University Librarian. It is understood that any copying or publication of this thesis for financial gain shall not be allowed without my written permission. Signature Date i

2 TITLE PAGE CHANGES IN REPTILE POPULATIONS IN THE SNAKE RIVER BIRDS OF PREY AREA, IDAHO BETWEEN AND : THE EFFECTS OF WEATHER, HABITAT AND WILDFIRE By John Olen Cossel Jr. Doctoral Papers Submitted in partial fulfillment Of the requirements for the degree of Doctor of Arts in the Department of Biology Idaho State University May 2003 ii

3 Copyright (2003) John Olen Cossel Jr.

4 SIGNATURE PAGE To the Graduate Faculty: The members of the committee appointed to examine the doctoral papers of JOHN OLEN COSSEL JR. find them satisfactory and recommend that they be accepted. Major Advisor Committee Member Committee Member Committee Member Graduate Faculty Representative iv

5 ACKNOWLEDGEMENTS I would like to thank Almighty God for the wonders of His creation, the mind He has given me to ponder His works, the heart He has given me to know the love of my wife and family, and the patience He provided to allow me to see this challenge to fruition. I would also like to thank my wife Ronda for her unwavering faith, support, love and many hours of field assistance; and my children Johnathon, Jessie and Rebecca for their willingness to share their father with a laptop computer when they would ve rather had me playing ball with them. My parents John and Sandee Cossel and parents in law, Rebecca and Edward Hardaway also greatly deserve my appreciation for their support and belief in me. I would like to acknowledge the crucial role my advisor Charles Peterson has played in my development as a scientist, and the role the Doctor of Arts faculty played in my development as an educator. I would like to thank my committee members, Don Streubel, Sharolyn Belzer, Rod Seeley and Al Strickland for the time and effort they expended for the sake of my program. I would also like to recognize the late Jay Anderson for his service as a committee member and mentor. In addition, numerous faculty and colleagues have shared insights and encouragement from which I ve benefited. I extend my sincere thanks to Lowell Diller who graciously shared his data set and spent two blustery days bouncing around in my old Ford pickup to locate his original study sites. I would also like to thank Teri Peterson for her invaluable statistical advice. I appreciate the effort expended by Jim Clark to coordinate BLM s involvement with this research. Financial support of my research is gratefully acknowledged, and was provided by the Lower Snake River District Office, Bureau of Land Management Challenge Cost Share Grant #ID CCS-259Z, and the Graduate Research Committee Grant proposal # Finally, in the tradition of the Areopagus Greeks of Athens that offered a place of honor to the unknown god, I extend my appreciation to all unknown and/or un-named benefactors. v

6 TABLE OF CONTENTS SIGNATURE PAGE...iv ACKNOWLEDGEMENTS...v TABLE OF CONTENTS...vi LIST OF TABLES... vii LIST OF FIGURES... viii LIST OF APPENDICES...ix INTRODUCTION... 1 Importance... 2 Challenges in assessing change... 3 Goals of the study... 4 MATERIALS AND METHODS... 5 Study Area... 5 Study sites... 6 Trapping... 7 Data Gathered... 9 Data Analysis RESULTS Changes in reptile taxa status Reptile status and habitat correlations Correlation of habitat variables Change in lizard species status Change in snake species status Weather effects Wildfire effects DISCUSSION Temporal effects Weather Habitat Disturbance Other factors Implications and recommendations Conclusion LITERATURE CITED TABLES FIGURES APPENDICES vi

7 LIST OF TABLES TABLE 1: Habitat classification of trapping sites within the Snake River Birds of Prey Area (SRBPA) for the years (1970s) and (1990s). 83 TABLE 2: The presence and absence of reptile species (number of sites in parentheses) at trapping sites in the SRBPA.. 84 TABLE 3: Shannon-Wiener diversity index values of reptiles at trapping sites in the SRBPA TABLE 4: Species richness of reptiles at trapping sites in the SRBPA TABLE 5: Relative abundance (captures per 100 trap days) of reptiles at trapping sites in the SRBPA TABLE 6: Correlations of habitat variables measured in 1998 at 26 trapping sites within the SRBPA TABLE 7: Reptile status and habitat correlations measured at 26 trapping sites during 1998 in SRBPA TABLE 8: Correlations (Spearman s rho), which are significant at P < 0.1, between reptile species abundance (captures/100 trap days) and habitat variables measured in 1998 at 26 trap sites within the SRBPA...90 TABLE 9: A comparison of habitat variables measured in 1998, between burn conversion sites and unburned shrub sites (> 20 years since wildfire) in the SRBPA TABLE 10: A comparison of reptile status as measured in 1998 at burn-conversion sites and unburned shrub sites (> 20 years since last wildfire) in the SRBPA TABLE 11: Mean capture rates (per 100 trap days) at burn-conversion sites and shrub sites (unburned), measured in 1998 at 17 trapping sites in the SRBPA vii

8 LIST OF FIGURES FIGURE 1: Snake River Birds of Prey National Conservation Area (SRBPA) boundary and study site locations FIGURE 2: Examples of the nine cover types identified by Diller and Johnson (1982) in the SRBPA FIGURE 3: Photograph of funnel trap used in conjunction with drift fencing at study sites within the SRBPA, Idaho..79 FIGURE 4: Dimensions and configuration of drift fence trapping arrays used in the SRBPA...79 FIGURE 5: Shannon-Wiener diversity index for reptiles encountered at all trapping sites within the SRBPA during , 1997 and FIGURE 6: Species richness of reptiles encountered at all trapping sites within the SRBPA during , 1997 and FIGURE 7: Abundance of reptiles (captures/100 trap days) encountered at all trapping sites within the SRBPA during , 1997 and FIGURE 8: Lizard species abundance (captures per 100 trap days) at drift fence trapping sites in the SRBPA..81 FIGURE 9: Snake species abundance (captures per 100 trap days) at drift fence trapping sites in the SRBPA..82 viii

9 LIST OF APPENDICES APPENDIX Study site photos, maps and cover types APPENDIX 2. Codes used to represent reptile species encountered APPENDIX Capture rates per 100 trap days (north and south) APPENDIX 4. Ranked abundance of some shrub species, cheatgrass and Russian thistle. 132 APPENDIX 5. Habitat characterization as determined by point and line intercept, and transect census ix

10 ABSTRACT CHANGES IN REPTILE POPULATIONS IN THE SNAKE RIVER BIRDS OF PREY AREA, IDAHO BETWEEN AND : THE EFFECTS OF WEATHER, HABITAT AND WILDFIRE Doctoral Papers Abstract -- Idaho State University (2003) Southwestern Idaho has the highest state reptile diversity, yet there are indications that reptile declines may have occurred there. To assess this possibility, I determined if temporal changes in reptile populations had occurred in the Snake River Birds of Prey Area (SRBPA). Drift fence trapping arrays were used to measure reptile status during the summers of , and these data were compared to an historic data set collected in at the same 24 study sites. Habitat variables were also measured in 1998 and correlated with reptile status. I made comparisons of habitat and reptile status between shrub sites and sites that had burned and were converted to exotic annual dominated sites (burn-conversion sites). Contrasts between shrub sites and burn-conversion sites revealed that the abundance of lizards and a lizard-eating snake was significantly less at burn-conversion sites, suggesting that ground foraging lizards and at least one of their reptile predators (Masticophis taeniatus) may have declined in other areas throughout the SRBPA where comparable changes have occurred. In addition, lizards were negatively correlated with a number of habitat variables indicative of disturbance, such as increased herbaceous plant cover, greater abundance of cheatgrass and other exotic annuals, and decreases in shrub cover, open soil and cryptogamic crusts. Snakes demonstrated few habitat correlations with the exception of Coluber constrictor, which in contrast to lizards, was positively x

11 correlated with some of the same disturbance variables. Although there was little change in species occurrence for the SRBPA, significant differences existed between 1998 and the other two periods ( and 1997), and were likely due to the effects of weather. Awareness of wildfire-induced changes is important, as knowledge of the status of reptiles in southwestern Idaho and their response to large-scale losses of native vegetation was previously nonexistent. In addition, if biodiversity and species abundance continue to decline, it will become increasingly important to understand the factors that result in negative impacts. Knowledge of these factors may foster increased monitoring of populations that were formerly assumed stable, and may add impetus to efforts aimed at minimizing additional habitat loss. xi

12 INTRODUCTION Global biodiversity is being lost at an alarming rate (Wilson, 1992; Kiesecker et al., 2001), and perhaps the single most readily identifiable cause is habitat loss (Alford and Richards, 1999; Pough et al., 2001; Meegaskumbura et al., 2002). While much habitat loss is associated with forests and wetlands, there have been extensive changes in other ecosystems as well. For example, the widespread loss of native shrub habitat in the Snake River Plain in southwestern Idaho, particularly within the Snake River Birds of Prey National Conservation Area (SRBPA). Habitat loss in this area is due to a number of factors, including agricultural conversion, historic overgrazing and the synergistic influence of fire and invasive exotic annual plants (Yensen, 1980, 1981; Whisenant, 1990; D Antonio and Vitousek, 1992; Knick and Rotenberry, 1997). The conversion of native shrublands to areas dominated by exotic annuals is the most noticeable and critical conservation issue for the SRBPA (USDI, 1996). These wildfire-conversions are known to negatively affect a number of vertebrate species in the SRBPA, including birds (Knick and Rotenberry, 1995) and mammals (Yensen et al., 1992; Knick and Dyer, 1997; Van Horne et al., 1997), however, their effects on reptile populations have not been explored. In addition to the loss of habitat because of wildfire conversions, several other lines of evidence suggest reptile declines in southwestern Idaho are a possibility. First, there are anecdotal accounts of lizard declines following applications of pesticide (Dana Quinney, Idaho Army National Guard Environmental Service Biologist, pers. comm., 1999). Second, commercial collecting of reptiles and amphibians for the pet trade has occurred (Idaho Department of Fish & Game, unpublished data). Third, Beck and 1

13 Peterson (1995) failed to find Longnose Snakes (Rhinocheilus lecontei) and Ground Snakes (Sonora semiannulata) at a study site on the Snake River Plain where the two species formerly occurred (Diller and Johnson, 1982). Heightened awareness of biodiversity loss has generated much research worldwide, and although considerable attention has been given to assessing the changes associated with mammals, birds and amphibians, relatively little attention has been given to the status of reptiles (Gibbons et al., 2000). Importance Reptile declines have several important implications for management (Scott and Seigel, 1992). Land managers should be concerned about reptile populations because of the biodiversity that they represent. This is especially true for sensitive species such as the Longnose Snake (Rhinocheilus lecontei), Ground Snake (Sonora semiannulata) and Great Basin Collared Lizard (Crotaphytus bicinctores) (BLM, 1995). These species occur in the SRBPA and are considered species of special concern and sensitive species by the Idaho Department of Fish and Game and the BLM respectively (Engle and Harris, 2001). Land managers should also be concerned because reptiles have important functional roles in desert ecosystems as both predators and prey (Pough, 1980, 1983). For example, Diller and Johnson (1988) estimated the combined impact of Western Rattlesnake (Crotalus viridis) and Gopher Snake (Pituophis catenifer) predation as an 18% annual harvest of the juvenile Townsend Ground Squirrels (Spermophilus townsendii) in the SRBPA. In addition, Grothe (1992) found that snakes made up over 66% of the total observed 2

14 biomass of prey delivered to Red-tailed Hawk (Buteo jamaicensis) nestlings in the eastern portion of the Snake River Birds of Prey Area. Finally, land managers should be concerned with the status of reptiles because they may serve as indicators of environmental change (Gibbons and Stangel, 1999). For example, Rie and others (2000a, 2000b) found that a population of Painted Turtles (Chrysemys picta) sampled near an aquatic pollution plume on Cape Cod, Massachusetts had greater levels of hepatic biotransformation enzyme activity than turtles sampled from a control site. In another study, Hopkins (2000) found that Southern Water Snakes had elevated levels of potentially harmful elements in their bodies because of eating contaminated fish. Reptiles can be relatively long lived and often occupy higher trophic levels, consequently through the process of bioamplification, they may have concentrated levels of pollutants, making them useful as indicators of environmental health (Brisbin, Jr. et al., 1974). Challenges in assessing change There are at least two major challenges in determining if changes in reptile status have occurred. One is the availability of historic data. David Wake (1994) described the lack of baseline data against which to measure population changes as one of the most difficult problems in conservation biology. Likewise, Diamond and Case (1986) describe natural trajectory experiments, which require data obtained prior to the perturbation under examination. This often proves difficult because of the stochasticity of naturally occurring disturbances. However, because of a reptile survey conducted during by Diller and Johnson (1982), a rare opportunity 3

15 existed for quantitatively testing the hypothesis that reptile populations have changed in the SRBPA between the late 1970 s and 1990 s. A second challenge is the high annual variation that can be exhibited by herpetofauna populations (Pechmann et al., 1991; Pechmann and Wilbur, 1994; Reed and Blaustein, 1995). Continuous longterm data are needed to determine trends, however, these data are unavailable for reptile populations in the SRBPA. The absence of these data limited my ability to address this second issue directly. Goals of the study The primary goal of this study was to determine if changes in reptile populations have occurred in the Snake River Birds of Prey Area between and To accomplish this goal, I identified several specific objectives: (1) Evaluate changes in reptile status, such as occurrence, abundance, species richness and diversity by repeating the trapping portion of Diller and Johnson s study (1982). (2) Quantitatively assess habitat variables at each trapping site and determine if correlations exist with measures of reptile status. (3) Investigate factors that may have contributed to changes in reptile status, including weather, relative grazing intensity, and particularly vegetation changes caused by wildfire. 4

16 MATERIALS AND METHODS Study Area Location/Geology The study area is located in southwestern Idaho, primarily within the boundaries of the Snake River Birds of Prey National Conservation Area (see Fig. 1). The SRBPA contains 196,221 hectares (484,873 acres) and was established to protect the high density of nesting raptors, and the natural, environmental, scientific, cultural and educational resources of the area (BLM, 1995). A major feature of this area is the Snake River and its associated canyon that unequally bisects the area. The canyon consists of cliff walls ranging in height from 2 to 125m (USDI, 1996), and basalt talus slopes tapering towards the canyon bottom. The area to the north is characterized by deep loess or silty alluvium soils that form a large, relatively flat plain, which increases in elevation from south to north (Diller and Wallace, 1986; USDI, 1996). The area to the south of the Snake River is more varied topographically with basaltic buttes, and eroded badlands featuring soils that are primarily lacustrine sediments of sand, gravel and silt (Diller and Johnson, 1982; USDI, 1996). Climate The climate at the SRBPA is typical of the Great Basin Desert, having a marked difference between summer and winter in both mean temperature and precipitation. Summers are hot and dry with a July mean daytime temperature of 24 C and a mean precipitation of less than 10mm for Boise, Idaho (USDI, 1996). In 5

17 contrast, winters are cold and much of the annual precipitation occurs between November and April. The January mean daytime temperature for Boise, Idaho is 1 C with a mean precipitation of approximately 35mm (USDI, 1996). Flora/fauna The vegetation found in the area can be classified into two broad community types: big sagebrush (Artemisia tridentata)/grass, and salt desert shrub (BLM, 1995). There is a general transition from big sage/grass to salt desert shrub from north to south in the SRBPA following elevation and precipitation gradients (Yensen and Smith, 1984). The combination of desert shrub communities and deep soil provides for a high density of raptor prey species, specifically the Townsend s ground squirrels (Spermophilus townsendii) and black-tailed jackrabbits (Lepus californicus). The high availability of preferred prey species along with nesting habitat provided by nearby canyons allows this area to host a high density and diversity of nesting raptors as well as wintering and migrating raptors (USDI, 1996). Reptiles are an important component within these communities serving as prey for raptors (Grothe, 1992; USDI, 1996) and as competitors with raptors for small mammal prey (Diller and Johnson, 1982, 1988). Study sites To distribute their sampling effort, Diller and Johnson (1982) identified nine major cover types, big sagebrush, winterfat (Krascheninnikovia lanata), shadscale (Atriplex confertifolia), greasewood (Sarcobatus vermiculatus), grass (native 6

18 perennial bunch grasses and exotic annuals), riparian, canyon rim, talus slope, and sand (areas of relatively low plant cover with sandy soils). Refer to Figure 2 for example photographs of habitat types and see Appendix for site maps and photos. Due to wildfires and habitat conversion, cover type classification did not remain the same for all sites in (see Table 1). To minimize spatial differences between the historic and current surveys, I located the original sites sampled by Diller and Johnson (1982). The 24 sites consist of 12 sites north of the Snake River and 12 sites on the south side of the Snake River (see Fig. 1). In March 1997, Lowell Diller helped me locate 22 of the original 24 trapping sites. We accomplished this by using Diller s site descriptions and photos as well as information and annotated land-use maps supplied by Mike Kochert (USGS). I subsequently located the other two sites by relying on personal communications and archived topographical maps. After locating each of the sites, I used a GPS receiver (Trimble GeoExplorer ) to obtain rover files that were differentially corrected in order to measure site universal transmercator (UTM) coordinates. Trapping To ensure temporal trapping consistency, I replicated the trap design utilized by Diller and Johnson (1982), using their trap descriptions as well as original traps for patterns. I constructed 44 new traps built according to original descriptions, and these were used in combination with four original traps. The exterior of the traps was constructed out of 3.2mm (1/8-in.) hardware cloth. The wire mesh was attached to a central square wooden frame of approximately 4cm x 4cm diameter pine boards. 7

19 Finished traps were 1.2m in length, 0.6m in width and 0.3m in height. The traps had hardware cloth funnels in both ends with a wooden divider (plywood) forming right and left halves of each funnel. The body of each trap had an access opening on the top (approx. 10cm x 15cm) that was covered by a piece of hardware cloth larger than the opening, and was held closed with elastic bands (see Figure 3). I formed all of the hardware cloth seams by folding an overlapping 1-3cm margin over the adjoining edge then securing the unions with wire or 3.2mm (1/8-in.) pop rivets. To provide refuge within the traps, two plywood cover boards (approximately 10cm x 15cm) were placed in each trap along with approximately 3 to 5cm of soil. I also covered each trap with cardboard to minimize temperature extremes. In March of 1997, trapping arrays were installed at the original 24 sites surveyed by Diller and Johnson in 1978 and The trapping arrays consisted of a 15m section of 0.51m wide galvanized roof flashing, buried approximately 5-10 cm in the ground. This length of fence material was reinforced with wooden surveying stakes. A funnel trap was placed on each end of this stretch of fence. An additional 7.5m piece of galvanized roof flashing was placed on the exposed ends of the 2 traps and these were reinforced with stakes. This formed an approximately 32m length of drift-fence with the two traps located at breaks in the fence (see Figure 4). After the trapping arrays were installed, they were closed until 1 May 1997, at which time they were opened. To minimize mortalities, traps were checked during the survey periods approximately every three days. Weather occasionally made access to sites impossible and longer periods resulted. However, the cool, wet weather reduced both capture rates and high temperature extremes, so that increased 8

20 trap mortalities did not result. At the end of the 1997 survey season (24 July 1997), the traps were closed until the beginning of the 1998 trapping period (1 May 1998). During the winter and spring closure period, the trapping arrays were checked to ensure that the traps were still in place and that the trap openings were blocked. Wind damage necessitated the reinstallation of three of the trapping arrays. Subsequently, all of the arrays were reinforced with additional stakes and 5cm drywall screws were used to attach the fencing to the stakes. During the 1998 field season I established two additional trapping arrays to increase the sampling of big sagebrush cover. This was necessary because the original big sagebrush sites sampled in the 1970s had been converted to predominately exotic herbs (e.g., Bromus tectorum). The two new sites, one on each side of the Snake River, were located as close as possible to an original big sage array. I also matched the slope and aspect of each new location as closely as possible to the original array. After the new arrays were installed, their UTM coordinates were determined using a GPS receiver (Trimble GeoExplorer ). Data Gathered Reptile data To accomplish the study objectives, I collected data on both reptile status and habitat at each of the study sites. Reptile data gathered consisted of the date and time each site was visited. Each of the two traps (designated A and B) were opened, the cover boards overturned and the substrate stirred to check the contents. Captured reptiles were identified, and their scientific names were codified using the first two 9

21 letters of the genus and the first two letters of the specific epithet (e.g., "PICA" for Pituophis catenifer; see Appendix 2). The sex of each individual was determined by attempting to manually evert hemipenes and/or probing for their presence, except with species that are sexually dimorphic or dichromic, which were visually sexed. I determined size, snout-vent length (SVL) and total length (TL) by using calipers for lizards and a metric cloth tape for snakes. The mass of captured reptiles was determined using Pesola spring scales. The stomach of each snake was palpated to detect the presence of prey items. When prey was detected, regurgitation was induced to determine prey identity. Each individual was given a specific identification number by clipping toes on lizards or ventral scales on snakes, using methods similar to those discussed by Ferner (1979). When incidental captures occurred (organisms other than reptiles, e.g., rodents) their presence was noted and they were identified when possible. Data were recorded in a field notebook and then transcribed into a spreadsheet application (Microsoft Excel), (see Appendices 3.1 and 3.2 for capture rate data tables). Habitat data Quantitative habitat data describing the study sites during the 1970s survey are unavailable; consequently, I subjectively assessed changes by referencing Lowell Diller's field notes and site photographs. However, to provide quantitative data for future comparisons, and to evaluate correlations between habitat variables and reptile status, I obtained additional information using a combination of techniques, including walk through surveys, line intercept and point intercept sampling. To accommodate 10

22 each of these techniques, I established a 100m transect that ran perpendicular to most arrays. However, when the habitat was not continuous in a perpendicular direction, for example on the rim of a cliff, transects were established in a parallel position 10 meters from the array. To randomize the location of each transect, drift fences were sectioned into 10 equal segments and then a random number generator (Microsoft Excel) was used to determine the transect origin. The side of the drift fence from which, the transect originated was also randomly determined. Walk through surveys Walk through surveys provide a view of the plant community at a given site while simultaneously representing their relative abundance, and were accomplished as described by Anderson (1991). Each 100m transect was surveyed and select plant species (shrubs, cheatgrass and Russian thistle) encountered within a 10m wide path were given a rank of 1-4 (1 = rare, 2 = uncommon, 3 = common and 4 = abundant; see Appendix 4 for ranked vegetation data). Ranked values can be transformed by taking the squared value of each ranking (1, 4, 9, 16) to better represent the relative abundance of each species (Anderson, pers. com., 1997). However, although transformed values intuitively represent differences in abundance, the necessary nonparametric correlation analyses (Spearman s rho) are insensitive to transformed rank values, producing identical probability values. 11

23 Point and line intercept surveys To determine the percent cover of various habitat variables, I used a point intercept survey that consisted of lowering a plumb bob perpendicularly until debris (litter), rock, soil or vegetation was encountered at 0.5m intervals along the same randomly determined 100m transects. To obtain another measure of shrub relative abundance, a line intercept survey was employed, that consisted of quantifying the intercept length of all shrubs crossing the previously established 100m transect lines (see Appendix 5 for habitat data). Burrow availability and recent grazing intensity To assess two other factors that may influence reptile occurrence and relative abundance, I quantified burrow availability by recording the number of burrows encountered within the 10m wide path along the 100m transects at each array. Burrow abundance has been shown to be a reliable indicator of Townsend s ground squirrel abundance (Yensen et al., 1992) and this may well hold true for other rodent species as well. As a result, burrow abundance may reflect not only available refugia but rodent abundance as well. I also quantified the relative amount of grazing that occurred at each site by recording the number of domestic ungulate feces encountered in the aforementioned transect swaths. A similar method was described by Brooks (1999) as a means of estimating black-tailed jackrabbit abundance. This approach assumed a positive correlation between time spent in an area by livestock and the number of feces (Brooks, 1999). This may actually serve as a better indication of use than animal 12

24 month units (AMUs) because AMUs assigned to an area do not necessarily represent heterogeneous spatial use by livestock. Nearly all of the feces enumerated were bovine (cattle). However, some were ovine (sheep) and a cluster of pellets was scored as one. To account for grazing intensity that may vary between species, a grazing index was created where intensity was ranked 1-4, where 1 = minimal (0-9 feces/1000m 2 ), 2 = light (10-19), 3 = moderate (20-49), 4 = heavy (50-99). Data Analysis General approach Statistical tests were conducted using Statistical Package for the Social Sciences (SPSS) version Where data failed to meet parametric assumptions, I used non-parametric tests. Results were considered statistically significant at α = 0.05, however, to minimize the risk of committing type two errors, results approaching significance (α = 0.10) are also discussed. Significant values are indicated throughout using the following notation, = 0.05 *, = 0.01 ** and = ***. Taxonomic groups The reptile species encountered within the SRBPA were divided into hierarchical categories that reflect phylogeny. These categories were all reptile species, Lacertilia (Lizards), Serpentes (Snakes) and individual species. This approach allowed me to evaluate whether changes may have occurred, but were widespread or limited in their range of effect. I identified specific measures of reptile 13

25 status, two of which, were limited to the taxonomic groups (reptiles, lizards and snakes). The first was diversity, represented by the Shannon-Wiener function (H ), where s H = ( pi)(ln i= 1 pi) s = the number of species and p i = the proportion of the total sample belonging to the ith species. This function is expressed in the form N 1, where N 1 = e H, and is equal to the number of equally common species that would be required to produce the same value of H (Krebs, 1999). The second measure of status was species richness, which was the number of species trapped at each study site. I evaluated changes in relative abundance by partitioning the capture rate (number of captures per trap days multiplied by 100) among the different combined groups and for individual species. Changes in diversity, species richness and abundance were assessed using Repeated Measures ANOVA for normally distributed data and Friedman tests for data that did not meet the assumptions of normality. I conducted Wilcoxon signed rank contrasts on nonparametric tests that yielded significant differences, adjusting for experiment wise error via Bonferroni corrections. Individual species Measures of individual species status included occurrence (site occupancy) and relative abundance. Occurrence was determined by the presence or absence of a species at each site, as indicated by trapping encounters. I tested the change in the proportion of sites where each species was encountered and where they were absent from our sample, using Fisher s exact tests. Changes in the relative abundance of each species were assessed using Friedman tests due to heteroscedasticity of the data. 14

26 Significant differences were further explored by conducting Wilcoxon signed rank contrasts, adjusting for experiment-wise error with Bonferroni corrections. Habitat correlations Bivariate correlation analyses (Spearman s rho) were conducted on diversity, richness and relative abundance, in relation to habitat variables, ranked grazing intensity, fecal count, number of burrows and rodent abundance (determined by calculating capture rates similarly to reptiles, based on trap encounters per 100 trap days). The habitat variables included percent cover for shrub, herb (which included exotic annuals), moss, lichen, cryptogamic crust (moss and lichen combined), litter, rock and soil. Also included were ranked values for cheatgrass, Russian thistle and exotic annuals (mean rank of cheatgrass and Russian thistle). Bivariate correlations (Spearman s rho) were evaluated for environmental and habitat variables independent of reptile status, to determine if and how each of these variables were associated. Weather I obtained archived daily weather records reported for the Swan Falls Power Station and the Boise Airport weather stations from the Desert Research Institute ( These data included the maximum daily temperature and the daily precipitation for the both the 1970 s and the 1990 s study periods. Due to the proximity of the Swan Falls Power Station (within the SRBPA), I relied primarily on these data. However, I substituted values from the Boise station when data were missing from the Swan Falls data set. These substitutions were justified because of a 15

27 highly significant positive correlation between the Swan Falls and Boise values (r = 0.967, P < 0.001). In the Pacific Northwest, reptile activity levels change throughout their active season, with many species showing peak activity periods in the spring/early summer and again in late summer/early fall (St. John, 2002; pers. obs.). Because of these temporal differences, I assessed correlations between weather data (maximum daily temperature and daily precipitation) and capture rates for each month separately. These data did not meet the assumptions of parametric testing; consequently, I performed nonparametric bivariate correlation analyses (Spearman s rho). Finally, I determined the mean maximum daily temperature and precipitation for each month of the field season (May, June and July) for each of the study periods ( , 1997 and 1998). Wildfire - Vegetation change Based on Diller s site descriptions, historic site photographs, and direct evidence (charred shrub remains, etc.), I determined that four original shrub sites had burned and were subsequently converted to predominately exotic herb communities (see Table 1). In addition to these four sites, a fifth site had burned prior to sampling in the and most likely had burned in the interim period (personal observation of adjacent burns). To examine the effects of wildfire on shrub sites within the SRBPA I conducted Mann-Whitney U tests to compare the diversity, richness and abundance for all three taxonomic groups between the five burned sites and the remaining unburned shrub sites. Although 1970 s pre-disturbance data existed for four of the five burned shrub sites, the small sample size and high annual variation in 16

28 reptile status precluded meaningful statistical analyses. For these analyses, sites were considered unburned when there was a > 20 year period since the last wildfire event, as evidenced by the presence of intact shrub communities during my study period in 1997 and 1998, and by the absence of written, or photographic evidence of a wildfire during or immediately before Diller s sampling in the 1970s. Analyses were limited to shrub sites because no riparian sites had burned and consequently their presence in the control (shrub) sample could inflate some measures of reptile status, such as diversity, etc. I also eliminated sites with a major rocky component (rim or talus), because rock is not directly modified by wildfire. To assess impacts of wildfire on individual species, one-way ANOVA was used to compare species abundance between burned and unburned shrub sites. When assumptions of normality were violated, I used Mann-Whitney U tests to compare species abundance. Wildfire in the SRBPA often results in distinct changes in the plant community. To evaluate the differences in environmental/habitat variables between burned and unburned shrub sites, I conducted a one-way ANOVA on variables that had homogeneous variance (or nearly so), including all measures of percent cover, number of burrows and rodent abundance. However, ranked habitat variables were not evaluated. RESULTS Changes in reptile taxa status Reptile occurrence Simply comparing the species encountered between the two study periods suggests relatively little change of reptile status in the SRBPA between and 17

29 All of the reptile species encountered by Diller and Johnson (1982) were also trapped at one or more of the 26 sites surveyed (see Table 2). A slight positive change was indicated by the trapping of two additional reptile species that were not encountered during Diller and Johnson s (1982) survey. The Great Basin Collared Lizard (Crotaphytus bicinctores) was represented by a single capture at the Nahas Ranch site (S7) in The other new reptile species occurrence was the Terrestrial Garter Snake (Thamnophis elegans), represented by several captures at riparian sites on both sides of the river (N1 and S8). Interestingly, there was also a new amphibian species represented in the incidental captures, the Tiger Salamander (Ambystoma tigrinum). Reptile diversity Reptile and lizard diversity was lower in than during , while there was no apparent trend in snake diversity (see Figure 5). Comparisons of diversity among the three study periods reveal a significant difference for all reptile species grouped together (P = 0.029). However, pairwise contrasts with Bonferroni corrections, show only the difference between the 1970 s and 1998 was marginally significant (P = 0.062). Lizard species diversity followed a similar pattern, with a significant difference among the three study periods (P = 0.015), but again, only the 1970 s and 1998 contrast was significant (P = 0.012). There was no significant different in snake diversity among the study periods (see Table 3). 18

30 Reptile species richness Reptile species richness varied widely across the study area with a site (N9) low of 1 in the 1970s and a site (N2 and S12) high of 9 in the 1970s and 1997 (see Table 4). There were no apparent trends in species richness for reptiles at all sites among sampling periods (see Fig. 6). Lizard species richness also varied across study sites, (see Table 4). There was little difference in lizard richness between the 1970 s and 1997; however, there was significantly lower lizard species richness in 1998 than in the 1970 s (P = 0.03), and in 1997 (P = 0.009) (see Fig. 6). Although snake species richness varied among sites (from a low of 1 to a high of 6; see Table 4), it was not significantly different between the study periods (see Fig. 6). Reptile relative abundance Relative abundance was significantly different between the study periods for all three taxonomic groups, displaying a general trend towards an increase (see Fig. 7). Reptile abundance (for all species) was significantly different between the study periods (P = 0.002), but yearly contrasts suggest that there was no difference between 1970 s and The contrasts did however reveal a significantly greater capture rate in 1997 than in the 1970 s and 1998 (P = and P = respectively; see Fig.7 and Table 5). Lizards likewise varied significantly in relative abundance (P < 0.001) between the study periods across all sites (see Fig. 7 and Table 5). Yearly contrasts suggested no differences in lizard relative abundance between the 1970 s and both 1997 and However, lizards were significantly more abundant in 1997 than in 1998 (P < 0.001) (see Fig. 7 and Table 5). The relative abundance of snake 19

31 species varied significantly between the study periods (P = 0.002) (see Fig. 7 and Table 5), with a significant increase between the 1970 s and 1997 (P = 0.003), and an increase approaching significance (P = 0.069) between the 1970 s and Although there was a decrease in snake relative abundance between 1997 and 1998, this difference was not significant. Reptile status and habitat correlations Correlation of habitat variables The numerous correlations between habitat variables are summarized in Table 6. A major pattern existed between percent shrub cover, which was negatively correlated with several related variables including percent herb (r = ***) and litter cover (r = *), exotic annuals abundance (r = **), as well as both cheatgrass and Russian thistle abundance separately (r = ** and * respectively). Percent shrub cover was also correlated with a number of related variables in a positive fashion, including percent cover of moss (r = 0.389*) and bare soil (r = 0.441*). The variables correlated with shrub cover also seem associated with each other in similar fashion. For example, the negative correlations between herbaceous cover and cryptogamic crust cover (r = ***) and between percent herbaceous cover and percent bare soil (r = ***). Correlations between vegetation/habitat variables may help explain associations between measures of habitat and reptile status discussed below. 20

32 Reptile diversity and habitat Reptile diversity in 1998 was positively correlated with percent cover of rock (r = 0.416*). Correlations approaching significance included negative associations with cheatgrass (r = , P = 0.068) and percent cover of herbaceous plants, which included exotic species (r = , P = 0.076). In addition, there was a positive association with percent bare soil (r = 0.331) that approached significance (P = 0.098). Lizard diversity had a single habitat correlation that approached significance (P = 0.084), which was a negative correlation with cheatgrass abundance (r = ). Snake diversity also had a single correlation, a positive correlation with percent cover of rock (r = 0.424**) (see Table 7). Reptile richness and habitat Reptile species richness in 1998 was negatively correlated with percent herbaceous cover (r = *) and positively correlated with percent shrub cover (r = 0.399*; see Table 6). The richness of lizard species was negatively correlated with percent herbaceous cover (r = **), cheatgrass abundance (r = *) and exotic annuals abundance (-0.386, P = 0.051). Lizard species richness was positively correlated with percent moss (bryophyte) cover (r = 0.434*), percent cryptogamic crust (moss and lichen combined) cover (r = 0.427*) and percent bare soil (r = 0.416*). Correlations approaching significance exist between lizard richness and percent shrub cover (r = 0.363, P = 0.069), as well as percent lichen cover (r = 0.357, P = 0.073; see Table 6). Snake species richness was positively correlated with percent rock cover (r = 0.424*; see Table 7). 21

33 Reptile abundance and habitat The abundance of all reptile species trapped in 1998 was significantly correlated with grazing intensity (r = **) and fecal counts (r = *). A negative correlation that approached significance (P = 0.07) also existed between reptile abundance and Russian thistle abundance (r = ; see table 7). Lizard abundance was significantly negatively correlated with percent herbaceous cover (r = **), cheatgrass abundance (r = *) and exotic annuals abundance (r = *). Positive correlations between lizard abundance and habitat included percent moss cover (r = 0.451*), percent cryptogamic crust cover (r = 0.427*) and percent bare soil (r = 0.395*). A positive correlation approaching significance (r = 0.367, P = 0.065) between lizard abundance and percent shrub cover also existed. Snake relative abundance was negatively correlated with fecal counts (r = **), grazing intensity (r = *), percent bare soil (r = *), percent cryptogamic crust cover (r = *), percent moss cover (r = *) and a negative correlation (r = ) that approached significance (P = 0.06) with percent lichen cover (see Table 7). Change in lizard species status Great Basin Collared Lizard (Crotaphytus bicinctores) Crotaphytus bicinctores was not encountered during the 1970s, and was represented by a single capture in 1997 (see Table 2). The small number of encounters did not allow meaningful analyses for this species. 22

34 Longnose Leopard Lizard (Gambelia wislizenii) Gambelia wislizenii was found at 5 sites in the 1970s, 2 in 1997 and 2 in 1998 (see Table 2). The proportion of occupied sites did not differ significantly between the sample periods (P = 0.416, and 1.0 respectively). The mean capture rates for G. wislizenii were 0.68/100 trap days (1970 s), 0.099/100 trap days (1997), and 0.20/100 trap days (1998) (see Fig. 8). These differences were not significant. There were a number of habitat correlations with G. wislizenii abundance, including a positive correlation with percent cover of lichens (r = 0.392*). Positive correlations with percent cryptogamic crusts (r = 0.371, p = 0.062), percent moss cover (r = 0333, p = 0.096) and the abundance of burrows (r = 0.367) all approached significance (P = 0.062, and respectively). A negative correlation between G. wislizenii abundance and percent herb cover (r = ) approached significance (P = 0.058) (see Table 8). Desert Horned Lizard (Phrynosoma platyrhinos) Phrynosoma platyrhinos was present at 5, 7 and 3 sites in the 1970s, 1997 and 1998 respectively (see Table 2). The difference between the proportions of occupied sites was not significant. The mean capture rates for P. platyrhinos were 0.65/100 trap days, 1.3/100 trap days and 0.74/100 trap days for each period (1970s, 1997 and 1998) respectively (see Fig. 8). The differences in capture rates were marginally significant (P = 0.074). The relative abundance of P. platyrhinos in 1998 was negatively correlated with cheatgrass abundance (r = *) and exotic annuals 23

35 abundance (r = *). There was a significant positive correlation with percent moss cover (r = 0.417*), percent cryptogamic crust cover (r = 401*). Correlations approaching significance included a negative correlation with percent herb cover (r = , P = 0.072) and a positive correlation with the number of burrows present (r = 0.366, P = 0.066), (see Table 8). Western Fence Lizard (Sceloporus occidentalis) Sceloporus occidentalis was encountered at 4 sites in the 1970s, 3 sites in 1997 and 2 sites in 1998 (see Table 2). These differences were not significantly different. The relative abundance of the S. occidentalis was 0.32/100 trap days, 0.25/100 trap days and 0.15/100 trap days for the 1970s, 1997 and 1998 respectively, and did not differ significantly (see Fig. 8). The abundance of S. occidentalis was negatively correlated with percent cover of moss (r = *), percent cover of cryptogamic crusts (r = *) and the number of burrows (r = *). There was a negative correlation between S. occidentalis abundance and percent cover of bare soil (r = ) that approached significance (P = 0.097). Positive correlations with S. occidentalis capture rates and percent rock cover (r = 0.493**), and abundance of Russian thistle (r = 0.403*) existed and there was a positive correlation with the abundance of exotic annual herbs (r = 0.372) that approached significance (P = 0.061; see Table 8). 24

36 Side-blotched Lizard (Uta stansburiana) Uta stansburiana was detected at 12, 9 and 5 sites in the 1970s, 1997 and 1998 respectively (see Table 2). There was no significant difference in the occurrence of U. stansburiana between the 1970 s and 1997 or between 1997 and However, the difference between the 1970 s and 1998 approached significance (P = 0.069). U. stansburiana capture rates were 1.8/100 trap days, 1.6/100 trap days and 0.44/100 trap days for the 1970s, 1997 and 1998 respectively (see Fig. 8). The differences in abundance were significantly different between sample periods. However, contrasts revealed that the differences between 1998 and both 1997 and only approached significance (P<0.1). Uta stansburiana relative abundance in 1998 was negatively correlated with cheatgrass abundance (r = *). Negative correlations existed with percent herb cover (r = ) and exotic annuals abundance (r = ) that approached significance (P = and respectively). There were positive correlations with U. stansburiana abundance and percent moss cover (r = 0.471*), percent lichen cover (r = 0.430*), percent cryptogamic crust cover (r = 0.463*), and burrow abundance (r = 0.469*); and a positive correlation with percent bare soil (r = 0.349) that approached significance (P = 0.081; see Table 8). Western Whiptail (Cnemidophorus tigris) Cnemidophorus tigris were the most widely occurring lizard species, detected at 14, 14 and 11 sites during the 1970s, 1997 and 1998 respectively (see Table 2). 25

37 There was no significant change in the proportion of sites where this species occurred. Cnemidophorus tigris was also the most frequently captured lizard species, with a relative abundance of 3.5/100 trap days, 7.0/100 trap days and 3.8/100 trap days for the respective sample periods (see Fig. 8). There was a significant difference in capture rates for C. tigris between sample periods (P = 0.042), with significantly more captures in 1997 than in 1998 (P = 0.003). However, there was no significant difference between the 1970 s and 1997 or Cnemidophorus tigris relative abundance in 1998 was negatively correlated with percent herbaceous cover (r = **), cheatgrass abundance (r = **), exotic annuals abundance (r = **) and with Russian thistle abundance (r = *). There were positive correlations with percent moss cover (r = 0.479*), percent cryptogamic crust cover (r = 0.448*), percent shrub cover (r = 0.444*) and percent bare soil (r = 0.394*; see Table 8). Change in snake species status Racer (Coluber constrictor) This species did not occur widely in the 1970 s, being encountered at only 4 sites. However, C. constrictor were found at 11 sites in both 1997 and 1998, and this increase in occurrence approached significance (P = 0.060; see Table 2). Similarly, the capture rate for this species was low in the 1970 s, requiring nearly 200 trap days for a single capture (0.54/100 trap days). In contrast, the capture rate in 1997 was 4.6/100 trap days and was nearly as high in 1998, at 4.1/100 trap days. This increase was highly significant (P < 0.001; see Fig. 9). Coluber constrictor abundance in

38 was negatively correlated with percent bare soil (r = ***), percent lichen cover (r = ***), percent cryptogamic crust cover (r = **), percent moss cover r = **), and approached a significant correlation with fecal abundance (r = , P = 0.060). There were positive correlations between C. constrictor abundance and percent herb cover (r = 0.622***), and cheatgrass abundance (r = 0.433*; see Table 8). Night Snake (Hypsiglena torquata) Hypsiglena torquata was encountered at 9 sites in the 1970 s, 11 sites in 1997, and 6 sites in 1998 (see Table 2), and the differences in occurrence between years were not significant. Hypsiglena torquata was not abundant at sites where they occurred, with capture rates of 1.8/100 trap days, 1.5/100 trap days and 0.76/100 trap days for the 1970 s, 1997 and 1998 respectively. The difference in abundance between years was significant (P = 0.015; see Fig. 9). However, after Bonferroni corrections, contrasts demonstrated that there was no significant difference between and 1997, but the lower capture rate in 1998 than in either the 1970 s or 1997 approached significance (P = 0.084, and 0.093). Although there were no significant correlations with H. torquata abundance and measured habitat variables, sites with a rocky component did seem to have higher capture rates. Striped Whipsnake (Masticophis taeniatus) Masticophis taeniatus was widespread during all three study periods, occurring at 19 sites in the 1970 s and 18 sites during both 1997 and 1998 (see Table 27

39 2). This difference was not significant. Capture rates were also similar for the three study periods, with 4.5 captures per 100 trap days, 5.4/100 trap days and 4.8/100 trap days for the 1970 s, 1997 and 1998 respectively (see Fig. 9), and these differences were not significant. There were no correlations between M. taeniatus abundance and habitat variables. Gopher Snake (Pituophis catenifer) This species was the most commonly occurring snake during all the sampling periods. Pituophis catenifer were encountered at 22 sites in the 1970 s, 23 sites in 1997 and 24 sites in 1998 (see Table 2). These differences in occurrence were not significant. There were however, some differences among capture rates, with a P. catenifer abundance of 5.5/100 trap days during the 1970 s, 11.3/100 trap days in 1997 and 9.3/100 trap days in 1998 (see Fig. 9). These differences were highly significant (P < 0.001). Yearly contrasts determined that the differences between the 1970 s and 1997 and 1998 were both significant (P = and 0.027), however, there was no significant difference between 1997 and Pituophis catenifer was negatively correlated with fecal counts (r = *), and approached significant negative correlations with percent rock cover (r = , P = 0.055) and grazing rank (r = , P = 0.060; see Table 8). Longnose Snake (Rhinocheilus lecontei) Rhinocheilus lecontei was encountered at 9 sites in the 1970 s, 7 sites in 1997 and 6 sites in 1998 (see Table 2). These differences were not significant. Although 28

40 this species occurred at roughly 33% of the trapping sites, the mean number of captures were relatively low, with 1.4/100 trap days during , 1.0/100 trap days in 1997 and 0.40/100 trap days in 1998 (see Fig. 9). These changes were significantly different (P = 0.023), however contrasts revealed that only the difference between 1970 s and 1998 was significant (P = 0.039). Rhinocheilus lecontei abundance was not correlated with any habitat variables, most likely due to the small sample size. Ground Snake (Sonora semiannulata) Sonora semiannulata was not commonly encountered, being found at only 2 sites in the 1970 s and 3 sites during 1997 and again in 1998 (see Table 2). This sample size precludes meaningful statistical analyses. The mean capture rate for all three study periods was also low with 0.30 captures /100 traps days in the 1970 s, 0.15/100 trap days in 1997 and 0.34/100 trap days in 1998 (see Fig. 9). These differences were not significant. In spite of the small sample size, S. semiannulata abundance was positively correlated with the percent cover of rock (r = 0.435*; see Table 8). Terrestrial Garter Snake (Thamnophis elegans) Thamnophis elegans were not encountered during the 1970 s, and although they were trapped at 1 site in 1997 and 2 sites in 1998 (see Table 2), this sample size was too small for meaningful statistical analysis. The capture rate for this species in 1997 was 0.05/100 trap days, and 0.29/100 trap days in 1998 (see Fig. 9). These 29

41 differences also were not significant. Although T. elegans abundance in 1998 was negatively correlated with percent lichen cover (r = *), and had negative correlations approaching significance with percent cryptogamic crust cover (r = , P = 0.056) and percent moss cover (r = , P = 0.071), they were only encountered at riparian sites, and these correlations may simply reflect characteristics of riparian sites (see Table 8). Western Rattlesnake (Crotalus viridis) Crotalus viridis was encountered at a number of sites, occurring at a minimum of 33% of the 24 trapping sites. This species was trapped at 8 sites in , 14 sites in 1997 and 10 sites during 1998 (see Table 2), and these differences were not significant. The capture rate for C. viridis was 0.71/100 trap days during , 1.2/100 trap days in 1997 and 0.66/100 trap days in 1998 (see Fig. 9). These differences also were not significant. Crotalus viridis abundance was not correlated with any habitat variables. Weather effects Capture data reflected a high amount of variation between 1997 and1998. When the mean monthly capture rates were compared, I found that there was indeed a significant difference (P < 0.001). The month of May had the highest rate at a mean of 37.3 captures per 100 trap days, while June was second with a mean capture rate of 28.7/100 trap days, followed by a 17.3/100 trap days capture rate for July. Bonferroni corrected contrasts revealed that each of the monthly capture rates was significantly 30

42 different from the others (May-June P = 0.027, June-July and July-May P < 0.001). These results support the analyses of monthly rather than seasonal correlations between reptile capture rates and weather (maximum daily temperature and precipitation). During May, reptile capture rates were significantly correlated in a positive fashion with maximum daily temperature (r = 0.726***) and negatively with precipitation (r = ***). During June, the only correlation with capture rate is negative association with maximum daily temperature (r = ) that approached significance (P = 0.097). The SRBPA during July usually experiences less variation in maximum daily temperature. There were no correlations between weather and reptile capture rates during July. The May mean maximum daily temperature for the period of is 25.1 C (77.2 F), while for it was 24.1 C (75.3 Fahrenheit), and 1997 had a monthly average of 26.0 C (78.8 F). However, 1998 was cooler with a monthly mean of only 21.1 C (69.9 F). The mean precipitation for the month of May was also noticeably different. During the period of , the mean daily precipitation for May was 0.88mm (0.035inches), while May had a mean daily precipitation of only 0.22mm ( inches), May of 1997 had a mean daily precipitation of 0.58mm (0.023 in.), and May of 1998 had a much higher mean daily precipitation of 2.79mm (0.11 in.). Wildfire effects Wildfire and habitat variables Nearly all of the habitat variables representing percent cover were significantly different between burned and unburned shrub sites (see Table 8). 31

43 Unsurprisingly, percent shrub cover was significantly less (P < 0.001) with means of 2.5% (burned) compared to 22.2% (unburned). Similarly, percent herb cover was also significantly different (P < 0.001) with means of 60.1% (burned) and 24.0% (unburned). Changes in these variables also influenced the percent cover of bare soil, with significantly less (P = 0.018) at burned sites with a mean of 8.8%, and a mean of 23.2% at unburned shrub sites. Interestingly, the percent cover of moss and cryptogamic crust was also significantly less at burned sites compared to unburned shrub sites (P = and 0.02 respectively). The difference between burned and unburned site percent litter cover approached significance (P = 0.058), with means of 24.2% (burned) and 17.0% (unburned). Cheatgrass had a mean rank score of three or common even at unburned sites. However, at burned sites the average rank was 4 or abundant. There was no difference in the grazing rank between burned and unburned sites. Likewise, there were no significant differences in the number of burrows or rodent abundance at burned and unburned sites (see Table 9). Wildfire and reptile status Habitat variables were not the only variables significantly different between burned sites and unburned sites. When measures of reptile status for 1998 were evaluated between desert shrub sites that had burned and those that hadn t burned recently (> 20 years) some interesting patterns emerged. All nine reptile taxa status variables (diversity, richness and abundance for reptiles, lizards and snakes) were lower on burned sites compared to unburned shrub sites (see Table 10). Both reptile diversity and species richness were significantly less on burned shrub sites, with a 32

44 mean diversity of 1.87 (burned) and 3.27 (unburned) (P = 0.001). The mean reptile species richness for unburned shrub sites was 4.5, while burned sites only had a mean richness of 2.4 (P = 0.006). The mean reptile abundance on burned sites was captures/100 trap days versus captures/100 trap days for unburned shrub sites; however, this difference only approached significance (P = 0.082). Wildfire and lizard status Lizard richness and abundance were both significantly different (P = 0.006) between burned sites that had means of zero for richness and abundance, and unburned shrub sites, which had means of 1.42 for richness and captures/100 trap days. Although lizard diversity was less on burned sites, with a mean of 1.0, and a mean of 1.35 for unburned shrub sites, this difference was not significantly different (see Table 10). The only lizard species that had differences in occurrence and abundance approaching significance (P = 0.064) was Cnemidophorus tigris. These lizards were encountered at seven of the 17 sites included in these analyses, and all seven were unburned shrub sites. The remaining lizard species that occurred at burned and/or unburned shrub sites did not have statistically different capture rates, however, they were all absent from burned sites and were only found on unburned shrub sites (see Table 11). Wildfire and snake status Diversity was the only snake status variable that was significantly lower (P = 0.037) on burned sites (mean = 1.87) compared to unburned shrub sites (mean = 33

45 2.49). The mean values for snake richness were 2.4 (burned) and 3.08 (unburned). Mean snake abundance was 16.9 captures/100 trap days (burned) and 18.7 captures/100 trap days (unburned), (see Table 10). In contrast to the lizard species response of occurring only on unburned sites, snake species were more varied in occurrence and abundance relative to burn status. The only snake species having a negative significant difference in status between burned and unburned shrub sites was M. taeniatus. Masticophis taeniatus was encountered at 12 of the 17 sites included in these analyses, and only two of these occurrences were on burned sites. The capture rate of M. taeniatus was significantly lower (P = 0.019) on burned sites, with a mean abundance of 0.68 captures/100 trap days compared to 6.3 captures/100 trap days for unburned shrub sites. Coluber constrictor had an opposite relationship to burn status, being found at only six of 17 sites, five of which were burned. There was a significant difference (P = 0.006) in the abundance of this species at burned versus unburned shrub sites, with mean abundances of 4.98 captures/100 trap days (burned) and 1.11 captures/100 trap days. Pituophis catenifer was a ubiquitous species in the SRBPA and occurred at all 17 sites included in these analyses, and there was no significant difference in the capture rate of P. catenifer between burned and unburned shrub sites. Although the remaining snake species occurring at shrub sites or burnconversion sites did not have statistically significant different capture rates relative to burn status, they were all absent from burn sites (see Table 11). 34

46 DISCUSSION Although the dependent variables (measures of reptile status) have hereto been presented in a taxonomic and hierarchical fashion, an evaluation of each of the independent variables as they relate to reptile status will illustrate the relative importance of each variable. To accomplish this, I will first discuss the influence of temporal effects on reptile status. Second, I will consider the effects of weather. Next, relationships between habitat and reptile status will be evaluated, followed by a specific look at the effects of wildfire induced changes. Other factors that may have affected reptile populations will also be addressed. In conclusion, the management implications of this study, needs for future work, and the importance of these issues will be presented. Temporal effects Temporal changes of reptile taxa When the status of higher taxonomic groups (reptiles, lizards and snakes) was assessed for all sites, there were mixed results, with some measures indicating increases while others suggested declines. This may initially seem contradictory, however, under careful scrutiny, patterns and explanations become apparent. For example, the total number of reptile species encountered throughout the study area (at all sites combined) exhibited a slight increase. All of the species present in the 1970s were also encountered in the 1990s along with two new species, C. bicinctores and T. elegans. These two species however, almost certainly occurred in the SRBPA during the 1970 s sampling period, but were simply not detected. Trapping arrays were not 35

47 optimally located to detect C. bicinctores. Moreover, because C. bicinctores spend much of their time basking on rocks, they are not as likely to be captured by drift fence trapping arrays as are ground dwelling, actively foraging species such as C. tigris. Consequently, C. bicinctores was under sampled during both periods. Thamnophis elegans is effectively sampled via drift fence trapping arrays. However, within the SRBPA, they primarily occur in association with water, and although there were three riparian sites included in the study, they were all some distance from water (approx m). The new encounters of this species in the 1990s likely resulted from the flooding of the Snake River that resulted in the expansion of the water s edge to within meters of the two arrays where they occurred (sites N1 and S8). Thamnophis elegans occurs elsewhere along the Snake River, both upstream and downstream (personal observations), and there is little reason to suspect they were absent during the 1970 s along the portion of river within the SRBPA. As illustrated by the new encounter of C. bicinctores, it is obvious that this measure of reptile status is sensitive to the observation of a single individual from a formerly unencountered species and that a single approach to sampling (e.g., drift fences) may not be sufficient to accurately document all the species present in a study area. Likewise, there is the always the possibility that uncommon habitats and/or species will be under sampled, even though the method of detection is appropriate. In contrast to the slight increase in total number of reptile species encountered, evaluations of all sites for changes in diversity, species richness and abundance between the decades suggest that statistically significant temporal differences exist. However, when yearly contrasts were conducted, the differences 36

48 between the 1970s and 1998 were the most substantial for measures exhibiting a decrease (reptile diversity, and lizard diversity and richness); conversely, measures displaying increases (reptile, lizard and snake abundance) had significant contrasts between 1997 and the other sampling periods. The differing results between the two years in the 1990s suggest the likelihood that an independent variable differed between the two years and influenced measures of reptile status. The nature of the inter year variation makes it possible to draw entirely different conclusion depending on which year s data are used. Comparisons with 1970s data and 1997 suggest a general improvement of reptile status as indicated by the increase in reptile abundance; in contrast, a comparison between the 1970s and 1998 suggest a decline in reptile status. This demonstrates the need for multiple year studies so that trends can be detected. Just as limited or isolated temporal measures can produce unclear results, single measures of reptile status may also lead to different or incomplete conclusions. For example, at a den site in northern Utah, Brown and Parker (1982) compared snake status data measured from to data obtained 30 years earlier by Woodbury (1951). They found that measures of snake community displayed little change in diversity, with Shannon-Wiener H = 0.99 for , and H = 1.05 for ; and species richness was 7 in and 7 in However, when they inspected relative abundance, they discovered that the two dominant species in the 1940s, Crotalus viridis and Masticophis taeniatus, had been replaced by Coluber constrictor and Pituophis catenifer (melanoleucus). My data display a similar potential, with measures of snake diversity and richness suggesting relatively little 37

49 change between decades, yet differences in abundance indicate that snake species were more abundant in 1997, due largely to the increase in abundance of two species, C constrictor and P. catenifer. These results suggest that when attempting to determine changes in reptile status, multiple measures of status should be included to obtain a more complete understanding of community dynamics. Temporal changes in lizard species Although higher taxa exhibited some changes, an evaluation of individual species is necessary to determine the extent and nature of those changes. However, a difficulty in assessing temporal changes of many lizard species in the SRBPA was the infrequency of trapping events and/or the limited number of sites where they were encountered. For example, C. bicinctores was only trapped once at a single site out of 24 sites, and at a sampling period of over 1000 trap days. Three other lizard species including G. wislizenii, P. platyrhinos and S. occidentalis also had low abundance and site occurrence reducing analytical power, and increasing the possibility of a type II error. These issues make data interpretation difficult. Both G. wislizenii and P. platyrhinos are ground dwelling lizard species that should be sampled somewhat effectively by drift fence trapping arrays. However, they were still represented by a small sample size, precluding any strong conclusions. One other species, S. occidentalis, had at least two related factors contributing to its small sample size. First, in this region they are associated primarily with habitats having a rock component, limiting the number of sites where they were likely to occur. Second, because of their extensive use of rock, ground movements are minimized, 38

50 reducing the likelihood of drift fence captures. Although these limitations were present during data collection in the 1970s as well as during the 1990s, the problem of small sample size is not resolved. Not all lizard species had insufficient sample size for statistical analyses; both U. stansburiana and C. tigris were encountered often enough to draw some conclusions about temporal changes. Uta stansburiana demonstrated no significant differences in either occurrence or abundance except in 1998, which had a decline approaching significance in the number of sites where they were encountered, and a 75% decrease in abundance. This variation suggests an annual effect that will be discussed in the next section. Cnemidophorus tigris is an active widely foraging lizard, and is effectively sampled using drift fence trapping arrays (Rosen, 2000). Cnemidophorus tigris had the highest occurrence rate, being encountered at over half of the study sites. They also had the highest lizard capture rate with a high of 7.0/100 trap days in 1997, which was significantly greater than the 1998 rate, again suggesting change related to inter year variation. In addition, while not statistically significant, the 1997 rate was a 100% increase over the 1970 s rate. If the capture rate in 1997 reflected actual abundance then it is possible that there was an inter decade increase in the abundance of this species. Temporal changes in snake species Snake species that occur in the SRBPA utilize terrestrial habitats (Brown and Parker, 1982; Diller and Wallace, 1986; St. John, 2002) and generally travel greater distances during their active season than lizards. Both of these factors make drift 39

51 fence trapping arrays a suitable method for sampling snake species, and many of the SRBPA species were sufficiently sampled to allow meaningful statistical analyses. Although there was an increase in the number of sites at which eight lizard and snake species were encountered during compared to , the most notable was the Racer (Coluber constrictor). This species had an increase of seven sites at which they were encountered. Most of the other reptile species increases were the result of only one or two additional site occurrences (including C. bicinctores, P. platyrhinos, H. torquata, P. catenifer, S. semiannulata and T. elegans). This serves as a reminder that caution should be used when drawing conclusions about presence/absence data. Indeed, exact testing revealed no significant differences in occurrence for any species other than C. constrictor, which only approached significance (P< 0.1). However, abundance data for C. constrictor exhibited a highly significant increase, suggesting that this type of measurement may be more sensitive for detecting changes in reptile status. Unlike Racers, Gopher Snakes did not differ in the number of sites where they occurred, but similarly had a significant increase in abundance during the 1990s. Because these temporal changes were relatively consistent for both 1997 and 1998, they likely indicate actual inter decade changes. The remaining snake species did not exhibit significant temporal changes with the exception of two species, H. torquata and R. lecontei. Capture rates in 1998 were significantly lower for both of these species than in the 1970s; and for H. torquata, lower in 1997 as well, suggesting inter year difference, but not necessarily inter decade change. 40

52 Annual variability Many herpetofauna populations exhibit natural fluctuations and may be capable of significant changes in relatively short spans of time. For example, Pechmann et al. (1991) determined that the number of breeding females of four amphibian species at a pond in South Carolina exhibited fluctuations of over three orders of magnitude among years. Similarly, Germano and Hungerford (1981) found that May reptile abundance at their Sonoran desert study doubled between 1977 and In addition to actual population changes, behavioral changes in response to variations in weather can change detection probability, generating false perceptions of change (see discussion on weather below). Change can also seem apparent due to the low capture probability of secretive and/or rare species, which at times may be represented by single captures. For example, initial concern over the failure of Beck and Peterson (1995) to locate two secretive species of snakes, the Longnose Snake and the Ground Snake at sites where they were previously detected by Diller and Johnson (1982) is moderated by the awareness of the amount of annual variation possible. The data obtained in by Diller and Johnson (1982) were only for a single field season per site, consequently it is difficult to determine whether that particular year was representative of the average or general condition of reptile populations for that decade. The temporally isolated measures did not allow for the evaluation of annual population trends, and as a result, caution should be used when drawing conclusions based on changes between the two data sets (1970 s and 1990 s). 41

53 My data from the SRBPA illustrate the potential for annual variation. Depending on whether 1997 or 1998 data were compared with Diller and Johnson s 1970 s data (1982), different conclusions could be drawn. Contrasts of reptile status between the 1970 s and 1997 suggest no significant difference in reptile diversity, lizard diversity or lizard species richness. Yet, contrasts of these same measures between the 1970 s and 1998 all indicate significant decreases. These conflicting results clearly indicate the weakness of 2-point comparisons, and corroborate the claim made by Pechmann et al. (1991) that long-term studies may often be required to separate natural fluctuations from declines. I agree that long-term monitoring should be implemented when attempting to assess declines. When this is impossible, there may be another approach; although not as powerful, comparisons at a single point in time across sites (subjects) that have experienced different levels of disturbance (treatment) may shed some light on the impacts of the treatment. A geographical information system (GIS) could be used to extrapolate the results of these comparisons across landscapes/study areas. Weather Effects of temperature For many measures of reptile status, variation between 1997 and 1998 was equal to and often greater than variation between and Because weather can greatly influence reptile behavior, and there were considerable weather differences between 1997 and 1998, this seems the most likely explanation for the abrupt changes. Variation in temperature can affect reptiles in a number of ways as a 42

54 consequence of reptile ecothermy. Achieving and maintaining a suitable body temperature is essential to reptiles, as nearly all physiological processes are temperature dependent in some manner (Zug et al., 2001). For example, Gopher Snake prey capture is optimal at 27 Celsius (Greenwald, 1974); consequently, predation success is dependent upon appropriate thermal conditions. Likewise, lizard locomotion and hence their ability to forage and evade predators, is temperature dependent (Avery et al., 1982; Van Berkum, 1986). The needs to forage and to avoid predators have resulted in some species evolving under different selective pressures. Consequently, they may have unique threshold or voluntary minimum temperatures. For instance, in northern Utah, Brown and Parker (1982) report active body temperatures for a number of reptile species common to the SRBPA, ranging from 37.6 to 27.9 C (C. tigris and P. catenifer respectively). In spite of a range of suitable temperatures for the SRBPA reptile community, there are minimum temperatures below which, individuals are likely to remain inactive. Diller and Wallace (1986) suggested that their relatively low capture rates of H. torquata in the SRBPA during the spring and fall were likely due to night time temperatures that were too cool. Mushinsky (1985) attributed declines in herpetofauna diversity and abundance to cold weather conditions that occurred during the second year of his study. In addition to thermal minima, reptiles have lethal maximum temperatures and, as environmental temperatures rise beyond an individuals ability to thermoregulate, they will seek refuge (Zug et al., 2001). These behaviors immediately affect the activity level of individuals, but can also have population effects (Beaupre 1996). Adolph and Porter (1993) suggest that temperature can 43

55 influence annual survival and fecundity by affecting the extent of thermally suitable activity time. Weather extremes then, can cause short-term reductions in activity levels, and subsequent reductions in trapping numbers resulting in a perceived decrease in species abundance, as well as possible long-term population responses that may reflect actual changes in abundance. Reptile capture data from the SRBPA suggest that weather does indeed affect measures of reptile abundance. In the SRBPA, May has the highest rate of reptile activity (this study), making this month particularly sensitive to variations in weather. May of 1997 had the highest number of captures and also had the warmest mean May maximum temperature at 26.0 Celsius. The May 1997 temperature was the closest to the active temperature (T act ) of four species of snakes found in the SRBPA. Brown and Parker (1982) reported T act means of 31.8, 31.2, 27.9 and 28.3 Celsius for C. constrictor, M. taeniatus, P. catenifer and C. viridis, respectively, at a site in northern Utah. In contrast, May of 1998 had the coolest maximum daily temperature and 4.8 times greater precipitation. Correspondingly, the capture rate of nearly every reptile species was lower in 1998 than in 1997 and, for a number of species, lower than in the 1970s as well. Additionally, the negative correlation of capture rates with May precipitation and the positive correlation of captures and May temperature support the notion that weather affects capture rates and, consequently, perceived abundance. It is important to note that community status variables such as reptile diversity and lizard diversity and richness were also significantly lower in This is still likely attributable to weather conditions. Because a number of species were represented by 44

56 relatively few captures, occurrence at relatively few sites, or both, declines in capture rates can also result in substantial differences in richness and diversity. Effects of precipitation Another means by which, weather can influence reptiles is indirectly through precipitation. Primary productivity in desert systems is highly dependent on the amount and timing of precipitation, and increases in plant biomass in turn result in increased biomass of primary consumers (Brown and Parker, 1982; Durtsche, 1995). Rosen (2000) found that increased rain caused arthropod blooms that resulted in increases in lizard reproductive output and recruitment in the Organ Pipe Cactus National Monument. Whitford and Creusere (1977) also found that annual variation of lizard species in the Chihuahuan Desert seemed to be related to changes in precipitation, and Ortega and Hernandez (1983) determined that availability of lizard prey was correlated with increased precipitation. Positive effects of precipitation on arthropod abundance are not limited to lizards. Brown and Parker (1982) suggested that increased numbers of C. constrictor at their study site in northern Utah, was likely due to six years of consistent rainfall and adequate numbers of arthropod prey. During their study, Racers only declined in body weight during 1972, which had an unusually dry summer and apparent reductions in Orthopteran populations. Rodents are also influenced by precipitation levels and plant productivity. According to Rosen (2000), most small mammal taxa in his study area also had increases that at least initially, followed high levels of precipitation and plant productivity. A number of snake species occurring in the SRBPA prey on rodents and other small mammals, 45

57 including P. catenifer, C. viridis and to a lesser degree M. taeniatus (Brown and Parker, 1982; Diller and Johnson, 1988). If average or above average levels of precipitation in the SRBPA generate increased plant productivity and attendant arthropod and small mammal abundance, then most reptile species should show increases in abundance. However, this increase in abundance would have at least a one-year time lag.. It should be noted that although precipitation and prey abundance may explain changes in reptile abundance, Rosen (2000) suggests that predation can play an equal or greater role in altering the population size of lizards. Consequently, although not addressed in this study, predation in addition to precipitation and productivity should be considered when assessing community changes. Weather conditions such as maximum daily temperature and precipitation can have a significant impact on the behavior and movements of reptiles, the abundance of their prey and on their reproductive output and survival. Consequently, having access to historical data sets for two point comparisons may not be sufficient to separate weather induced natural fluctuations from long term population trends. Habitat Effects of habitat cover type Much of the variation in species occurrence, abundance and community structure across the SRBPA is explained by the distribution of resources within each community, and this in turn, is dependent on interactions between the physical properties of each site, the biota present and site histories. The availability of some resources can be determined by common measures of habitat such as percent cover of 46

58 shrubs and trees for arboreal species. Other aspects of resource availability are inferred from habitat properties, such as prey abundance, suitable refugia, etc. The broad cover types described and sampled by Diller and Johnson (1982) provide a coarse categorization of habitat and explain some of the variation between sites. For example, riparian sites indicate mesic or hydric environments that provide suitable conditions for fish and a number of amphibian species both of which are common prey items for T. elegans; not surprisingly, this species only occurred at riparian sites. Likewise, talus slope implies a high percent cover of rock in the form of loose boulders and scree, without reference to plant physiognomy. Species that are known to utilize rocks for some aspect of their life history are expected to occur at these sites. Species that are associated with rock cover include S. occidentalis, and C. bicinctores (St. John, 2002), both of which were encountered solely at talus sites. Although cover type helps explain the presence of some reptile species, other measures of reptile status such as abundance and community composition may require quantification and evaluation of site habitat characteristics beyond subjective categorical placement. Effects of habitat quantitative data Quantifying habitat variables at each site allowed the discovery of some interesting species-habitat relationships. Each of the habitat variables measured had a significant correlation with at least one measure of reptile status. Percent cover of rock was correlated with the saxicolous lizard S. occidentalis, and as mentioned above, the use of rock for basking and display sites by this species explains variation 47

59 in abundance much the same as rock cover type explained presence/absence. Sonora semiannulata was the only snake species that was correlated with percent rock cover. Rocks serve as refugia for this species (pers. observ.), and a number of their preferred arthropod prey as well, including spiders, centipedes, scorpions and Orthopterans (Nussbaum et al., 1983; St. John, 2002). Although there was no significant correlation between rock cover and H. torquata abundance, this species seemed to be more common at rocky sites. My observations were similar to those of Diller and Wallace (1986) who found H. torquata most frequently at canyon rim sites in the SRBPA. The associations between these species and rock may also explain the positive correlation between rock cover and reptile and snake diversity, because rock increases the likelihood of capturing species that would otherwise be absent, thus elevating measures of diversity. The percent cover of bare soil was differentially correlated with lizard and snake species, and this may be due to differing modes of locomotion, foraging techniques and thermal requirements. Snake abundance was negatively correlated with the amount of bare soil. However, this was largely due to the highly significant negative correlation between C. constrictor abundance and open soil patches, as C. constrictor was the only snake species correlated with percent bare soil. The explanation for this negative relationship may lie in the positive association between C. constrictor and percent herb cover, which increased as the amount of bare soil decreased. Because C. constrictor are able to traverse dense herbaceous cover and can utilize chemosensory cues, they may be able to effectively use areas with high herb cover. In contrast, lizard abundance was positively correlated with percent bare 48

60 soil, and this association was again driven predominately by a single species, C. tigris. Germano and Hungerford (1981) describe C. tigris as occupying sparsely vegetated desert shrub, quickly moving from bush to bush while foraging, and that dense grass (herbaceous cover) might hinder both movements and foraging behavior. Pianka (1970) demonstrated that C. tigris activity is related to soil temperature, and Brown and Parker (1982) list C. tigris as having the highest active body temperature (37.6 C) of four lizard species also found in the SRBPA. It seems possible then, that in addition to impeding movements and foraging, excessive plant cover may limit thermal opportunities associated with radiant heating of open soil. Future work should include efforts to compare thermal microhabitat availability to explore relationships among the amount of bare soil, temperature and lizard abundance. The percent cover of cryptogamic crusts was positively correlated with percent bare soil, and similarly to percent bare soil, lizard and snake species responded differently to varying amounts of cryptogamic crust. With the exceptions of S. occidentalis and C. bicinctores, every lizard species had significant positive correlations with some measure of cryptogamic crust (moss, lichen or both). Sceloporus occidentalis and C. bicinctores are the only two lizard species in the SRBPA that are not primarily ground dwellers, and this may explain the lack of correlation between cryptogamic crusts and the abundance of these species. Although one might expect the amount of open soil to decrease as cryptogamic crusts increased, the positive relationship may be due to a common factor that decreases both (see discussion on disturbance below). Snake response to cryptogamic crusts also mirrored those of bare soil cover. Snake species abundance was negatively correlated 49

61 with cryptogams but C. constrictor was the only species that demonstrated this association. The similarity of results for soil and cryptogamic crust cover suggests that species may be responding to a habitat or site characteristic represented by these variables. Percent shrub cover was significantly correlated with reptile and lizard richness, but was not associated with any snake species and the only lizard species positively associated with the amount of shrub cover was C. tigris. Based on discussions above, one would expect that C. tigris abundance might decrease with increased shrub cover. However, as with other habitat variables, correlations between habitat variables seem to explain this relationship; shrub cover was inversely related to herb cover. Consequently, the associations between C. tigris abundance, soil, cryptogamic crust cover, and herb cover also help explain the positive relationship with shrub cover. The percent cover of herbaceous plants, cheatgrass and other exotic annuals included in this category, had the most consistent negative association with measures of reptile status. In fact, every ground dwelling species of lizard had negative correlations with herb cover, cheatgrass or exotic annuals abundance that at least approached significance (P< 0.1), and C. tigris had highly significant negative correlations with these variables. Herbaceous cover and the abundance of exotic annuals may be associated negatively with reptile and lizard diversity/richness because they are indicative of disturbance and may limit movements and prey availability (Vander Haegen et al., 2001). 50

62 Many of the associations among habitat variables can be explained by plant dynamics and disturbance. For example, as shrubs are removed from a community, provided there are sufficient nutrients, precipitation and propagules, then some form of plant growth will utilize these vacated resources. In the SRBPA, herbaceous plants, particularly exotic annuals seem to have a competitive edge over native shrubs when filling these voids (Young and Evans, 1978), hence the inverse relationship between shrub and herb cover. The non-disturbed state of most communities in the SRBPA consists of a major shrub component. However, fire and other disturbances such as vehicle damage, grazing etc. can reduce shrub cover (Busack and Bury, 1974) and increase opportunity for herbaceous growth, particularly exotic annuals in highly disturbed areas (Yensen, 1981). As a result, as shrub cover decreases, herbaceous plant cover may increase. A less intuitive relationship explained by disturbance is the positive associations between percent shrub cover and the percent cover of cryptogamic crusts and soil. Relatively undisturbed shrub sites have well developed cryptogamic crusts (moss and lichens) that form under the canopies of shrubs and to some extent on the soil surface between shrubs. Disturbances such as grazing and fire can cause a decrease in the percent cover of cryptogamic crusts (Johansen et al., 1984; Johansen and St. Clair, 1986), and may give exotic annuals opportunity to proliferate by providing microsites for invasion (Hobbs and Huenneke, 1992). As a result, sites that are less disturbed have higher percent shrub cover and a greater percent cryptogamic crust cover. 51

63 Disturbance Disturbance can play a major role in changing or maintaining habitat structure (Dublin et al., 1990). Consequently, it is important to understand the nature of disturbance in the SRBPA and the effects it has on communities. Two common forms of disturbance that have occurred widely in the SRBPA are grazing and wildfire. Both forms have historically been natural components of community dynamics. However, during the last century, the frequency and intensity of these disturbances has changed dramatically (Yensen, 1980; Whisenant, 1990). Grazing Grazing can result in disturbance through the disruption of cryptogamic crusts (Johansen et al., 1984; Johansen and Smith, 1986) and the spread of exotic annuals (Yensen, 1980, 1981); these factors along with wildfire can have a synergistic influence on shrub loss (Knick and Rotenberry, 1997). Grazing also results in decreased herbaceous cover and shrub abundance (Bock et al. 1984). For example, Anderson and Holte (1981) found that the absence of grazing on sagebrushdominated rangeland in southeastern Idaho resulted in an increase of shrubs as well as the percent cover of perennial grasses. The most direct effects of grazing appear to be changes in the structural composition of the vegetative component at a given site. In fact, Jones (1981) determined that grazing reduced lizard abundance and species diversity only when structural composition of the habitat was changed. Romero- Schmidt and Ortega-Rubio (1999) found that grazing effects were greatest on lizards that use a sit-and-wait approach to foraging. They attributed declines in lizard 52

64 abundance to changes in vegetation cover that in turn resulted in altered conditions of surface and near surface microenvironments. In addition to changing microenvironmental conditions, grazing can have an impact on herpetofauna through increased predation (Janzen, 1976) and/or decreased prey abundance (Busack and Bury, 1974). Bock et al. (1990) suggested that grazing reduces herb cover, resulting in higher predation because of decreased availability of predator refuge sites. Grazing may likely result in an increase of predation on reptiles; however, the effects of grazing on prey abundance can be varied. For example, Phillips (1936) found that deer mice were most abundant on moderately grazed prairies in Oklahoma, while Heske and Campbell (1991) determined that rodents were more abundant in areas excluded to cattle grazing. Furthermore, Brooks (1999) found that livestock grazing can negatively affect rodents and he attributed this difference to changes in seed availability and invertebrate abundance. While the effects of grazing on rodents are mixed, invertebrate response seems to be more consistently negative. For example, Morris (1968) found that grasslands, which had not been grazed for 2-3 years, had greater numbers of invertebrates than grazed sites. In addition, Gibson and others (1992) found that changes in Araneae (spider) assemblages varied because of grazing effects on vegetative structure, while other invertebrate groups were impacted via grazing through altered plant species composition. Finally, Vitt and Ohmart (1974) described changes in perennial grasses that resulted in degradation of the invertebrate fauna, which in turn reduced the population size of C. tigris. Although a number of studies have found negative 53

65 effects of grazing, in some systems it can play a positive role in maintaining habitat heterogeneity (Ballinger and Jones, 1985; Ballinger and Watts, 1995). Much of the work done assessing the impacts of grazing on reptiles has focused on the response of lizards. However, this group oddly demonstrated no correlation with grazing intensity in the SRBPA. This may be due to insufficient variation in grazing intensity between sites, or that grazing at the observed levels of intensity simply does not have a significant impact on the lizard fauna in the SRBPA. Although there were no apparent effects of grazing on lizards, snake abundance was negatively correlated with grazing intensity; however, only two snake species appear to have lower abundance in association with grazing levels. Pituophis catenifer was the only species to have a significant negative association with grazing, while C. constrictor had a negative correlation that approached significance (P< 0.1). These relationships are difficult to explain, as no habitat measurements were correlated with grazing levels except the percent cover of litter and the abundance of Russian thistle, which both approached significance. Russian thistle can be associated with recent disturbance (Yensen, 1981), so perhaps livestock activity resulted in snake emigration from those areas and/or resulted in snake casualties or decreased production/recruitment (Janzen, 1976). Grazing induced differences in vegetation cover and attendant changes in rodent populations, as determined by abundance estimates and burrow densities, were not apparent. Consequently, although there were some associations between grazing and declines in snake abundance, other disturbance factors such as wildfire may have had a much more important effect on reptiles in the SRBPA. 54

66 Wildfire The potential for extensive and long lasting effects of wildfire, as well as its role in maintenance of community structure (Christensen, 1985) make this component of habitat dynamics worthy of study. Friend (1993) states Fire and its effect on the environment remains one of the most important issues being addressed in ecological research and management in Australia today. Likewise, The occurrence of wildfire in the SRBPA is a substantial disturbance factor considering that over 50% of the area burned between 1980 and 1988 (Kochert and Pellant, 1986; USDI, 1995; Knick and Rotenberry, 1997). When fire is a natural disturbance with which communities have evolved, then the role of fire can have a positive effect, maintaining spatial and temporal patchiness as well as successional change that favors herpetofaunal diversity (Lillywhite, 1977a, 1977b; Caughley, 1985; Braithwaite, 1987; Greenberg et al., 1994). Although community responses may generally be positive, the effects of fire are often species specific (Greenberg et al., 1994) and are likely dependent on fire frequency and intensity (Braithwaite, 1987). For example, Mushinsky (1985, 1992) found that increased fire frequency negatively affected Five-lined Skinks (Eumeces inexpectatus) but positively affected Six-lined Racerunners (Cnemidophorus sexlineatus). Regardless of species response to wildfire, the effects are mediated through direct or indirect mechanisms. The primary direct effect of wildfire would be deaths caused by temperature extremes. However, while there are likely casualties that occur from wildfire, Friend (1993) suggests that compared to mammals, reptiles show fewer acute effects of fires 55

67 in Australian woodlands. Kahn (1960) proposed that limited direct effects of fire in chaparral communities were because of reptiles using burrows or refugia under rocks. Patterson (1984) found that the immediate effects of burning a New Zealand grassland were a 28% reduction in the population density of a common skink (Leiolopisma nigriplantare). However, several individuals captured before the fire, were encountered a month afterwards uninjured. Because there were no wildfires at or near any of the trapping sites immediately prior to or during data collection in the 1970s or 1990s, any declines observed are not likely due to recent direct effects of fire. More important than the direct effects of wildfire on reptiles are the indirect effects that result from structural changes to the plant community (Mushinsky, 1985). Structural changes may in turn result in differences in microhabitat, thermal microenvironments, predation levels (Braithwaite, 1987), and prey abundance. Pianka (1966, 1967) suggested that the number of flatland desert lizard species is correlated with structural attributes of vegetation, and that both vertical and horizontal components of spatial heterogeneity are important. Changes in horizontal heterogeneity can have a positive effect, such as greater use by Phrynosoma at burned sites in Texas than at sites with accumulated litter and rank vegetation (Fair and Henke, 1997). Lillywhite and North (1974) described a temporary increase in lizard carrying capacity of a chaparral community as the result of openings created by fire. Although fire can induce positive changes, there is likely an optimal level of disturbance; for example, Lillywhite (1977a, 1977b) found that lizard abundance seemed maximal at intermediate levels of fire disturbance rather than in senescent 56

68 shrub communities or in converted grasslands. Lillywhite (1977a, 1977b) went on to describe how S. occidentalis and U. stansburiana were common in chaparral but were never seen on three different areas that were converted to grasslands devoid of shrub cover. Structural changes are likely the cause of observed differences in reptile status between burned and unburned sites in the SRBPA and these effects will be discussed below. Wildfire induced structural changes in plant communities result in changes in microhabitat (Simovich, 1979). For instance, the loss of shrubs may result in greater soil insolation, decreased soil albedo, and subsequent soil heating (Christensen, 1985). Without heterogeneity, soil heating and xeric conditions could make thermoregulation and water homeostasis difficult for some lizard species (Romero- Schmidt and Ortega-Rubio, 1999). Moreover, low amounts of cover following wildfire could open shrub and brush canopies, increasing predation (Lawrence, 1966; Simovich, 1979). In addition to differing predation pressures, foraging opportunities due to different fire regimes can also affect reptiles and prey species (Mushinsky, 1992). Because of the primarily phytophagus habit of many arthropods, changes in plant communities should be reflected in insect abundance. Indeed, D Antonio and Vitousek (1992) found that exotic grass invasion resulted in local declines of insect species. Rodents may also be susceptible to fire effects. Lillywhite (1977a, 1977b) found that rodent abundance was greater in chaparral than in sites converted to grass. Similarly, Yensen et al. (1992) described cheatgrass, an exotic plant strongly associated with post-burn sites, as an unstable food source for Townsend s ground 57

69 squirrels (Spermophilus townsendii) that results in widely fluctuating rodent populations. There is a relationship between reptiles and their shelter and foraging needs that is associated with the successional changes in vegetation resulting from fire (Friend, 1993). There can be negative effects of wildfire, but a number of the above studies found fire to play a positive role in maintaining reptile diversity (e.g., Mushinsky, 1985; Braithwaite, 1987). However, rather than maintaining habitat heterogeneity in the SRBPA, wildfires and the synergistic effect of exotic annuals are converting former shrublands to large tracts dominated by weedy herbaceous plants, with reduced structural diversity. This loss of shrublands, the attendant loss of structural heterogeneity, and the myriad of associated changes are likely the most important factors affecting reptile populations within the SRBPA. Conversion of shrublands to exotic annuals With the realization that at least 40 million hectares in the western United States are affected by the invasion of cheatgrass and the attendant fires, the gravity of the situation becomes apparent (Whisenant 1990). To understand how this much land could be affected so dramatically, it is necessary to visualize how several factors work together. Exotic species that have been present in the SRBPA since 1900 (Yensen 1981, 1982), such as cheatgrass (Bromus tectorum), Bur buttercup (Ranunculus testiculatus), Russian thistle (Salsola kali), and a variety of mustards, have all been involved with the conversion of native shrublands to sites dominated by exotic annuals. Another major factor in the conversion from primarily native species to predominately exotic species is the amount of disturbance. Fire has been the 58

70 primary disturbance factor (USDI, 1996) in the SRBPA. Wildfires reduce native vegetation, which was historically shrublands (Yensen, 1981, 1982) and provide opportunity for exotic annuals to out-compete native species. Although in chaparral, post-fire shrub growth can occur from crowns (Simovich, 1979), big sagebrush does not exhibit this ability (Young and Evans, 1978). Consequently, wildfires often result in substantial shrub death. The shrub eliminating effect of wildfire is even greater when there is repeated burning in an area that kills shrub seedlings, and allows the depletion of the native seed bank (Winward, 1984; Whisenant, 1990). Several characteristics of cheatgrass make it amenable to wildfire. For example, it germinates in the moist winters, grows, propagates early, and then dies leaving fine, dry herbage that conducts fire from shrub to shrub (Young and Evans, 1978; D Antonio and Vitousek, 1992). However, perennial bunchgrasses are not as likely to cause this effect and are much less likely to burn (Roberts, Jr., 1990). The fire tolerant properties of cheatgrass may well be due to the evolutionary history experienced in its native range. Plant species that have evolved with the selective pressures of recurrent wildfire may, in addition to having mechanisms that increase survival, have inherent flammable properties that perpetuate a fire-dependent plant community (Mutch, 1970). Although fire has historically been a part of community dynamics in the SRBPA, Whisenant (1990) has described a reduction in the interval between wildfires from years to approximately 3-10 years. Christensen (1985) states that shrublands with historically low fire frequency often have the greatest susceptibility to fire mortality. This apparently is the case for shrub communities in the SRBPA. While other desert shrub communities may respond to fire by going through 59

71 successional changes that increase temporal and spatial habitat heterogeneity (Simovich, 1979), the community response in the SRBPA is a novel stable state that resists successional change. Laycock (1991) describes most stable state communities in North America to involve exotic annual invasion, grazing pressure, and/or altered fire regimes, all three which have affected shrublands in the SRBPA. Post-fire habitat relationships The results of this study corroborate claims about the fire response of shrublands in the SRBPA. The contrasts between shrub sites and burn-conversion sites indicate that there are significant decreases in shrub cover, cryptogamic crust cover, and the amount of bare soil. Shrub mortality, seed and seedling casualties, and increased distances to intact shrub communities for propagule transfer, all explain the nearly one order of magnitude difference in shrub cover between shrub communities and burn-conversion sites. In addition to reducing shrub cover, wildfires have been known to cause decreases in cryptogamic crusts elsewhere in the Great Basin Desert (Johansen et al., 1984; Johansen and Smith, 1986), and this seems the most plausible explanation for the 282% difference in cryptogamic cover between burn-conversion sites and shrub sites in the SRBPA. Johansen et al. (1984) suggested that post-fire recovery of cryptogamic crusts is dependent on the proximity of propagules and the absence of further disturbance such as grazing. This again suggests the synergistic effect that the combination of disturbances has had on former shrublands in the SRBPA. This synergism may also explain the increase in herbaceous cover, which was 150% greater at burn-conversion sites. Surprisingly, there was not a significant 60

72 difference in the number of burrows or rodent abundance (as indicated by drift fence capture rates) between shrub sites and burn-conversion sites. This may be due to a response similar to the widely fluctuating ground squirrel populations observed by Yensen et al. (1992) at cheatgrass sites in the SRBPA. Their results suggest that long-term ground squirrel success may be less at cheatgrass dominated sites, but that temporary population abundance may occur. Wildfires have multiple effects on shrubland communities in the SRBPA, inducing changes in habitat structure and microenvironment, and through these mechanisms have likely had an effect on reptile populations. Post-fire reptile status Nine measures of reptile status were all lower at burn-conversion sites than at shrub sites. Although not all of these differences were significant, reptile and snake diversity, reptile and lizard richness, and lizard abundance were significantly less than at non-burned shrub sites. This suggests that habitat differences discussed above, translate into differences in reptile communities across the SRBPA. Although there may be a number of factors involved, they are likely related to changes in habitat structure and a decrease in habitat heterogeneity (Pianka, 1966, 1967). Different species generally vary in their response to disturbance. However, lizards seemed to have the most consistent response, with no lizard species being captured in 1998 at any of the five burn-conversion sites. Romero-Schmidt and Ortega-Rubio (1999) suggest that lizards rely strongly on microhabitat and substrate adaptation to carry out necessary life functions. Therefore, wildfire-induced changes that result in the 61

73 absence or an insufficient quantity of necessary microhabitat may explain the apparent group response of lizard species. For example, Simovich (1979) suggested that Phrynosoma coronatum prefer open spaces and would be expected to decrease as vegetation density increases. Phrynosoma platyrhinos utilizes similar foraging tactics and it is reasonable to assume they would display similar responses to increased herbaceous plant cover. Simovich (1979) also describes U. stansburiana as preferring open spaces and rocks, and that a decrease in lizard abundance would be expected as vegetation density increased. Lillywhite (1977a, 1977b) proposed that if areas were converted from shrublands to grasslands with little or no remaining shrub cover, lizards would essentially be eliminated from the area. My data are consistent with these hypotheses, and suggest that the status of lizards at burn-conversion sites may be the result of wildfire-induced changes. Although the lizard taxa data suggest a possible group response, C. tigris was the only species that had a difference in abundance that approached significance (P = 0.064), and this may be a result of small sample sizes. The absence of C. tigris on burn-conversion sites may be explained by excessive herbaceous cover and diminished patches of open soil. Mushinsky (1985, 1992) attributed the lower abundance of another Cnemidophorus species (C. sexlineatus) at some of his study sites, to the thick herbaceous layer that eliminated open spaces and probably interfered with foraging. Because of the active foraging mode of C. tigris and their relatively high active body temperature, loss of exposed soil and increases in herb cover may have similar effects. However, in addition to restrictions of movements and thermal gradients, factors such as prey abundance may also explain the absence 62

74 of this species at burn-conversion sites. Morris (1975) found that significantly greater numbers of some insect orders were recorded in unburned grasslands compared to burned areas. In addition, Brooks (1999) suggested that the abundance of C. tigris might be a good indicator of invertebrate abundance in the Mojave Desert, suggesting a close predator-prey association. Unfortunately, data necessary to assess correlations between burn status and invertebrate abundance in the SRBPA are unavailable. Unlike the lizard species, the status of snake species encountered at shrub and burn-conversion sites varied. The general condition was reduced abundance and occurrence at burn-conversion sites. For example, M. taeniatus were significantly less abundant at burn-conversion sites; while C. viridis, R. lecontei, H. torquata and S. semiannulata were absent from burn-conversion sites. Although little work has been done concerning the response of snake species to wildfire and subsequent habitat changes, it is likely that the observed differences are also due to changes in the vegetative structure and habitat heterogeneity. Similarly to lizards, snakes need to have appropriate substrate and microhabitat conditions to maximize fitness. However, prey abundance may be as important as habitat structure in determining snake status at burn-conversion sites. Considering the hypothesized effect of habitat change on lizard populations, snake species that utilize lizards as a significant prey source, would also be expected to have reduced occurrence and abundance at burnconversion sites. The species in this category include R. lecontei, H. torquata and M. taeniatus. Of these three species, two (R. lecontei and H. torquata) were absent from burn-conversion sites; however, they only occurred at a limited number of shrub sites, making meaningful statistical analysis difficult. Masticophis taeniatus was 63

75 encountered more frequently and even occurred at two burn-conversion sites, but was significantly less abundant at these sites than at shrub sites. These observations support the proposition that burn induced habitat changes may affect lizards, which in turn may affect lizard-eating snake species. This may not bode well for R. lecontei, a state of Idaho sensitive species (Engle and Harris, 2001). The observed relationships between C. tigris status and burn-conversion sites suggest that widespread wildfire and subsequent habitat conversion may result in reduced C. tigris numbers in the SRBPA. These negative consequences may in turn impact R. lecontei, as they have been described as dietary specialists preying primarily on Cnemidophorus species (Rodriguez-Robles and Greene, 1999). Hypsiglena torquata also prey primarily on lizards, particularly U. stansburiana, and declines in lizard abundance would be expected to affect Night Snakes. Fortunately, however, this effect might be moderated by the association with rocky habitats of H. torquata, (Diller and Wallace, 1986) and that U. stansburiana will use rocky sites as well as shrub sites (St. John, 2002). Kahn (1960) suggested that S. occidentalis would be less affected by losses of ground cover because of their heavy utilization of rock; the same is likely true for U. stansburiana. Moreover, Germano and Hungerford (1981) found that even areas that were mostly grass (hence largely unsuitable for U. stansburiana) had horizontal stratification because of the presence of rock and bare patches of soil. Thus, rock could minimize some of the losses of habitat heterogeneity, and may allow for the persistence of U. stansburiana at sites where H. torquata are also likely to exist. The effects of wildfire in the SRBPA may result in a bleak outlook for M. taeniatus as well. I hypothesize that although this species does prey on rodents and birds in 64

76 addition to lizards, they may be unable to compete with species that prey more heavily on mammals, such as P. catenifer and C. viridis, resulting in M. taeniatus declines. In contrast to lizard-eating snakes, snake species that do not rely heavily on lizards may show little or no difference in status between burn-conversion sites and shrub sites. Two snake species were actually more abundant at burn-conversion sites, P. catenifer and C. constrictor. Pituophis catenifer have been described as preying primarily on rodents (Brown and Parker, 1982; Rodriguez-Robles, 1998). There were no significant differences between the number of burrows, rodent abundance or P. catenifer abundance between burn-conversion sites and shrub sites. If rodent abundance actually is comparable between burn-conversion sites and shrub sites, this may explain why P. catenifer had such a widespread occurrence (occurred at all 17 sites in this comparison), and why their abundance did not differ significantly among habitat types. The apparent minimal effects of wildfire and habitat conversion on rodents and P. catenifer may explain, in part, the significant increase in abundance of this species between the 1970s and the 1990s. An additional explanation may be the reduction of C. viridis in the SRBPA because of human persecution. Brown and Parker (1982) suggested that increases in the abundance of P. catenifer at their study site in northern Utah, may have been due to competitive release as a result of substantial declines of the Western Rattlesnake population. There is dietary overlap in these two species (Diller and Johnson, 1988), and with increased human use of the SRBPA, it is conceivable that C. viridis has declined. However, contrasts between the 1970s and the 1990s data do not suggest such a decline. 65

77 The increased abundance of C. constrictor at burn-conversion sites is a little more difficult to explain, based on rodent prey abundance. Brown and Parker (1982) described C. constrictor at their study site as relying heavily on arthropod prey, although this species also eats rodents and other reptiles. If changes in habitat structure were affecting lizards via reduced arthropod prey abundance, then one would expect C. constrictor to exhibit declines as well. Unfortunately, as mentioned earlier, data on arthropod abundance related to burn-conversion sites are unavailable. It is possible that insects (particularly Orthopterans) are not limited at burnconversion sites, and C. constrictor are able to utilize this prey source. Another possibility is that C. constrictor have been able to shift their prey selection to take advantage of apparently adequate rodent populations that exist at burn-conversion sites. In addition, C. constrictor is described as occurring in a wide range of habitats (Nussbaum et al. 1983), and perhaps changes in habitat structure and microhabitat availability may not affect this species in the same manner as lizard species. Regardless of the mechanisms involved, C. constrictor demonstrated significant increases in occurrence and abundance between the 1970s and the 1990s. Based on the positive associations of C. constrictor with habitat variables that are consistent with burn-conversion sites, and their significantly greater abundance at these sites, their population increases may well be associated with the widespread disturbance that has occurred in the SRBPA. Habitat structure and heterogeneity are important aspects of reptile ecology, and uniform stands of grass or exotic annuals may offer little in the way horizontal variation and even less in the vertical direction (Pianka, 1966, 1967; Germano and 66

78 Hungerford, 1981). Consequently, the wildfire-induced habitat changes that are occurring in the SRBPA are likely to affect reptiles. In many communities, wildfire may act to maintain habitat heterogeneity and increase reptile diversity (Lillywhite and North, 1974; Lillywhite, 1977a, 1977b; Mushinsky, 1985, 1992; Braithwaite, 1987). However, wildfire and the subsequent vegetation changes that occur in the SRBPA may have negative impacts on reptile diversity, species richness, and for some groups, particularly lizards and lizard-eating snakes, on abundance. Although the direct mechanisms behind these decreases have yet to elucidated, the sobering story seems to be similar to the plight of shrub-obligate bird species that are declining as native shrublands are fragmented and lost (Knick and Rotenberry, 1995). Consequently, efforts to minimize additional native shrub loss will likely be important for the persistence of the relatively high reptile biodiversity historically present in southwestern Idaho. Other factors Pesticide applications Although wildfire induced habitat changes are likely to have the greatest longterm impact on reptile populations in the SRBPA, another potential influence is the widespread application of pesticides. Land managers may use pesticide applications to control infestations of arthropod pests when nearby agricultural grounds are at risk of suffering loss. Because many reptile species in the SRBPA, particularly lizards, prey directly on arthropods or on arthropod predators, an unfortunate consequence of widespread pesticide use is the potential for negative impacts on reptiles due to 67

79 decreases in arthropod abundance. Idaho Army National Guard biologist Dana Quinney noted an apparent decline in the number of lizards encountered within the Orchard Training Area (a component of the SRBPA) in the 1980s, after widespread spraying of pesticide (Dana Quinney, pers. comm., 1999). However, based on the apparent lack of widespread general declines, any effects of pesticide application were likely limited in their range and/or have not persisted. Collection of reptiles Another factor with the potential to cause reptile declines is the collection of reptiles. Reptiles have been collected at varying intensities for decades. Although some collecting has been for private use, a number of reptile species in Idaho have commercial potential. Enge (1993) suggests that factors influencing collection numbers include abundance, market value, market demand and ease of collection. An example of how these factors may interact to influence commercial collection is the fact that Side-blotched Lizards were commercially collected in large numbers over a 3 year period ( , years with tabulated data, n = 3557; IDF&G, unpublished data), even though this species has marginal market value and demand. However, historically they have been abundant and relatively easy to collect, which may explain the large harvest. A species with particular appeal to collectors is the robust and colorful Great Basin Collared Lizard. During two years of commercial collecting of this species ( ), only 72 individuals were reported as harvested. However, Enge (1993) describes that vulnerability to over-collection can occur when a species or population occupies a limited geographic range, occupies limited suitable habitat, 68

80 has low fecundity, or concentrations of individuals exist such as breeding areas or hibernacula. At least two of these criteria are likely for C. bicinctores. In response to conservation concerns, IDF&G only allowed commercial collection of this species for 2 years ( ). My data were insufficient to evaluate possible impacts on this species. The IDF&G allowed commercial collecting of native amphibian and reptile species between the years of 1994 and 1998, during which, over 8000 individual reptiles were reported collected for commercial purposes (IDF&G unpublished data). The majority of these were likely collected from southwestern Idaho, however, detailed collection locality information is unavailable, so the extent of collecting, if any, that occurred within the SRBPA is unknown. Again, based on my small sample size of 24 sites, the general lack of declines suggest that impacts of commercial collection were either localized in areas other than my sampling sites and/or were not persistent in their effect. Implications and recommendations Study limitations This study provides evidence concerning changes in reptile populations that have occurred in the SRBPA between and However, there were some inherent study design problems that I was unable to address entirely. One such problem is the difficulty of studying secretive species. Some species were captured so infrequently that meaningful statistics were impossible. For example, S. semiannulata had a mean capture rate for the 1970s and 1990s of only 0.26 captures/100 trap days; at that rate, a trap would have to be in place for over a year to 69

81 catch a single S. semiannulata. This became problematic when measures of reptile status such as species richness and to a lesser degree, diversity were dependent on presence/absence data because at times secretive species were represented by a single capture for the whole field season. Another limitation of this study was the relatively small sample size. Although sites seems an adequate sample, this was insufficient for some types of statistical analyses. However, small sample size was not easily remedied. Sampling reptile status with drift fence trapping arrays is very resource intensive, which prohibits large sample sizes; manpower was insufficient for any additional trapping arrays. Moreover, for comparisons between the 1970s and the 1990s, I was limited to the initial sampling scheme and number of sites utilized by Diller and Johnson (1982). Limited sample size became an even larger factor for the burnconversion and shrub comparisons. The sample size of burn-conversion sites was only five, which reduced statistical power and increased the potential for committing a type II error (Hayes and Steidl, 1997). In addition, some artifacts resulted from the small sample size of sites surveyed, for example the odd habitat correlations of S. occidentalis. Western Fence Lizards had a negative correlation with cryptogamic crusts and a positive correlation with Russian thistle, which in combination would suggest this species abundance, would increase with increasing disturbance. However, this artifact is likely due to the small number of talus slope sites (n = 3) where this species occurred. All three sites exhibited moderate to severe disturbance, as two of the sites experienced burn conversions and one was heavily grazed; as a result, the significant positive correlation with rock is more ecologically meaningful. 70

82 A different, but related problem is that the burn-conversion sites were primarily limited to former big sage shrub sites. Yensen et al. (1992) found that shadscale shrub communities in the SRBPA that had burned still clustered with unburned sites when a non-metric multidimensional scaling ordination and cluster analysis was done. They attributed this to the abundance of exotic annuals at shadscale sites even when unburned. In contrast, however, former big sage sites were predominated by exotics after experiencing wildfire and clustered separately. It is possible that because of greater shrub size, big sage canopies will generate wildfires that are less patchy and of more intense heat than shadscale sites, resulting in greater shrub death and increased likelihood of exotic herb dominance. A wider breadth of former shrub community types within the burn sample would make inferences concerning the effects of burn-conversion much stronger and more generalizable. Figure 1 illustrates another limitation of this study, and that is the restricted spatial distribution of sites within the SRBPA. The 26 study sites were clustered in the western portion of the SRBPA. As mentioned earlier, because of the use of historic data, I was limited to evaluating previously selected sites. However, even if a new sampling scheme had been devised, access to widely distributed sites across such harsh terrain would be difficult. Diller and Johnson s (1982) initial site selection was likely influenced by these limitations. The historic distribution of wildfire in the SRBPA has been widespread, and the unsampled southeastern portion of the area may have experienced as many or more burned acres as the northwestern portion (Knick and Rotenberry, 1997). In order to determine if the observed differences are a 71

83 local response, additional studies evaluating burn-conversion effects across the SRBPA in a more spatially distributed approach would be worthwhile. Habitat fragmentation, patch size, and spatial distribution of habitat patches are important aspects of ecology (Wiens, 1985). However, site selection by Diller and Johnson (1982) minimized these issues by positioning trapping arrays in homogeneous cover types at a distance from habitat edges or ecotones. Data from another study done in the SRBPA (Peterson et al., 2002) suggest that polygon size, and nearness to habitat edges (transitions from one dominant physiognomy to another) are associated with greater reptile richness, and the abundance of some species. Information regarding polygon size and nearness to edge were not available from the 1970s data set. Fortunately, however, Diller and Johnson s placement of polygons distant from edges, likely minimized edge and polygon size effects. The results of a number of other studies (e.g., Germano and Hungerford, 1981; Pechman et al., 1991; Rosen, 2000) suggest that long-term data are often necessary to detect trends and to separate natural fluctuations from changes due to some disturbance. Although the two consecutive years of sampling conducted during shed much light on the effects of weather and provided a more generalized view of reptile status, additional data would likely have proven very useful. In addition to pre- and post-disturbance data that are strong evidence for disturbance caused changes (Schlesinger and Shine, 1994), trend data are necessary. In spite of the present atmosphere of reduced funding and pressures to undertake short-duration projects, trend data should be obtained when possible. 72

84 Management recommendations Land and wildlife managers are charged with the task of preserving biodiversity in the face of numerous wildlife threats including loss of habitat, competitive pressures of exotics species, disease, and climate change to list a few. Additional concerns are the effects of disturbances such as grazing and wildfire. Although some work evaluating responses to wildfire has been done (e.g., Lillywhite, 1977a; Mushinsky, 1985), little information exists concerning the response of reptile populations to wildfire in the Great Basin Desert. Previous studies have generally involved systems in which wildfire occurs at natural or historic frequencies and intensities, helping to maintain habitat heterogeneity and reptile diversity. However, this study provides information on the largely negative responses of reptiles to the novel synergistic effects of wildfire and exotic annuals. Providing land managers with this understanding may generate the impetus for more additional studies aimed at exploring the mechanisms involved. Specific management actions that can be taken include measures that will minimize wildfire in the SRBPA. Because exotic annuals are widespread within the area and disturbance continues to occur at varying levels, fire prevention is not very realistic. A key focus may be on rapid response and containment of fires that start. Continued work should focus on rehabilitating exotic herb sites, although because these communities are likely at a stable state this may prove difficult or even impossible to achieve (Laycock, 1991; Knick and Rotenberry, 1995). 73

85 Future studies In order to maintain native herpetofauna, it will be necessary to understand how different foraging guilds respond to structural changes in vegetation (Brooks, 1999; Romero-Schmidt and Ortega-Rubio, 1999), and whether or not reptile species have natural histories that will allow predictions of their responses to wildfire (Friend, 1993). To accomplish this, future research needs to be directed at quantifying: 1) herpetofauna responses to both direct and indirect effects of wildfire, 2) structural characteristics of habitats, both vertical and horizontal components (Simovich, 1979), 3) food resource availability (arthropod, rodent and reptile), and 4) the gradient of thermal opportunities over time at burn influenced sites. Determining the mechanisms by which wildfire-conversion affects the above variables would allow a clearer understanding of what may have occurred in the past and the predicted response to disturbance for each species. This is important because responses to both disturbance and management will vary by species and managers must decide for which species to allocate resources. Future work should also include modeling efforts and the impact assessment of habitat fragmentation. Efforts aimed at modeling or estimating the extent of impacts/changes within the SRBPA would allow managers to assess the severity of the many issues facing managed lands and to prioritize accordingly. Also, with the realization that habitat fragmentation is a growing problem and one that is a result of wildfires and exotic annual invasion, efforts to explore the effect of polygon size, edge effect, polygon heterogeneity and other landscape characteristics are important to managers (Wiens, 1985; Knick and Rotenberry, 1997). 74

86 Conclusion Contrasts of reptile status between shrub sites and burn-conversion sites indicate significant differences in reptile diversity, richness and abundance, particularly that of lizards and lizard-eating snakes, suggest that wildfire-induced changes may have had negative effects on reptile populations. Negative correlations between a number of reptile species and habitat variables, which are indicative of disturbance, provide additional support for the assertion that disturbances such as wildfire and grazing may have negative effects on reptile communities within the SRBPA. Although much of the change associated with burn-conversion is negative, at least two species demonstrated little or positive changes, suggesting that species responses to disturbance vary. Future efforts should attempt to increase the sample size of burn-conversion sites, and to include a wider variety of former shrub types (i.e. shadscale, winterfat, etc.) in the burn-conversion sample. In addition, research is needed to determine the mechanisms by which the effects of wildfire are mediated towards reptile populations. This additional information may allow land and wildlife managers to predict the response of individual species towards management practices and disturbances, and to recognize the potential impacts of wildfire on reptile populations. Superficially, the results of this study provide little evidence to suggest reptile populations have declined in the SRBPA between and In fact, two species, C. constrictor and P. catenifer, were significantly more abundant in the 1990s. However, under closer inspection, significant changes were apparent between 1998 and the other two sampling periods (1970s and 1997). These 75

87 differences were likely due to the influence of the abnormally cold and wet weather during May of 1998, which resulted in fewer captures occurring. These results indicate that weather can have a major impact on the perceived abundance of herpetofauna, and should be considered when evaluating temporal change. This potential for annual variation also underscores the importance of having data from more than just two points in time. Although I had access to a historic data set, the conclusions drawn from a single comparison with 1997 or 1998 would have been entirely different. Consequently, long-term data should be obtained when evaluating changes in herpetofauna populations, to evaluate trends. The results of this study are important, as knowledge of the status of reptiles in southwestern Idaho and their response to large-scale losses of native vegetation was largely nonexistent. Finally, as global biodiversity and species abundance continue to decline, it will become increasingly important to understand the factors that may result in negative impacts. Knowledge of these factors may foster increased monitoring of populations that were formerly assumed stable, particularly after specific habitat change occurs, and may add impetus to efforts aimed at minimizing additional habitat loss. 76

88 LITERATURE CITED Adolph, S. C., and W. P. Porter Temperature, activity and lizard life histories. The American Naturalist 142(2): Alford, R. A., and S. J. Richards Global amphibian declines: a problem in applied ecology. Annual Review of Ecology & Systematics 30: Anderson, J. A Fieldwork to support the NPR Environmental Impact Statement. Final Report to EG&G Idaho. Anderson, J. A., and K. E. Holte Vegetation development over 25 years without grazing on sagebrush dominated rangeland in southeastern Idaho. Journal of Range Management 34(1): Avery, R. A., J. D. Bedford, and C. P. Newcombe The role of thermoregulation in lizard biology: predatory efficiency in a temperate diurnal basker. Behavioral Ecology and Sociobiology 11: Ballinger, R. E., and S. M. Jones Ecological disturbance in a Sandhills Prairie: impact and importance to the lizard community on Arapaho Prairie in Western Nebraska. Prairie Naturalist 17(2): Ballinger, R. E., and K. S. Watts Path to extinction: Impact of vegetational change on lizard populations on Arapaho Prairie in Nebraska Sandhills. American Midland Naturalist 134: Beaupre, S. J Field metabolic rate, water flux, and energy budgets of mottled rock rattlesnakes, Crotalus lepidus from two populations. Copeia, 1996: Beck, J. M., and C. R. Peterson Movements and habitat selection of the Longnose Snake (Rhinocheilus lecontei) in southwestern Idaho. Idaho Bureau of Land Management. Technical Bulletin No Bock, C. E., J. H. Bock, W. R. Kenney, and V. M. Hawthorne Responses of birds, rodents, and vegetation to livestock exclosure in a semi desert grassland site. Journal of Range Management 37: Bock, C. E., H. M. Smith, and J. H. Bock The effect of livestock grazing upon abundance of the lizard Sceloporus scalaris, in southeastern Arizona. Journal of Herpetology 24(4): Braithwaite, R. W Effects of fire regimes on lizards in the wet-dry tropics of Australia. Journal of Tropical Ecology 3:

89 Brisbin, Jr., I. L., M. A. Staton, J. E. Pinder, III, and R. A. Geiger Radiocesium concentrations of snakes from contaminated and non-contaminated habitats of the AEC Savannah River Plant. Copeia 2: Brooks, M Effects of protective fencing on birds, lizards and black-tailed hares in the western Mojave Desert. Environmental Management 23(3): Brown W. S., and W. S. Parker Niche dimensions and resource partitioning in a Great Basin Desert snake community. Pp In: Herpetological Communities Edited by N. J. Scott Jr. United States Department of the Interior, Fish and Wildlife Service. Wildlife Research Report 13, Washington, D. C. Bureau of Land Management (USDI) Snake River Birds of Prey Area Management Plan. Lower Snake River District Office, Idaho. 170pp. Busack, S. D., and R. B. Bury Some effects of off-road vehicles and sheep grazing on lizard populations in the Mojave Desert. Biological Conservation 6(3): Caughley, J Effect of fire on the reptile fauna of Mallee. Pp In: Biology of Australian Frogs and Reptiles. Editors: G. Grigg, R. Shine, and H. Ermann. Surrey Beatty, Chipping Norton, NSW. 527pp. Cristensen, N. L Shrubland fire regimes and their evolutionary consequences. Pp In: The ecology of natural disturbance and patch dynamics. Academic Press, Inc. D Antonio, C. M., and P. M. Vitousek Biological invasions by exotic grasses, the grass/fire cycle, and global change. Annual Review of Ecology & Systematics 23: Diamond, J., and T. J. Case Community Ecology. Harper and Row Publishers, NY. 665pp. Diller L. V., and D. R. Johnson Ecology of reptiles in the Snake River Birds of Prey Area. Final Report to USDI Bureau of Land Management. Snake River Birds of Prey Research Project. 107pp Food habits, consumption rates, and predation rates of Western Rattlesnakes and Gopher Snakes in southwestern Idaho. Herpetologica 44(2): Diller L. V., and R. L. Wallace Aspects of the life history and ecology of the desert night snake, Hypsiglena torquata deserticola: Colubridae, in southwestern Idaho. The Southwestern Naturalist 31(1):

90 Dublin, H. T., A. R. E. Sinclair, and J. McGlade Elephants and fire as causes of multiple stable states in the Serengeti-Mara woodlands. Journal of Animal Ecology 59: Durtsche, R. D Foraging ecology of the Fringe-toed Lizard Uma inornata, during periods of high and low food abundance. Copeia 4: Enge, K. M Herptile use and trade in Florida. Florida Game and Freshwater Fish Commission. Final Performance Report. Nongame Wildlife Program. Tallahassee, Florida. 102pp. Engle, J. C., and C. E. Harris Idaho species of special concern element state ranking reviews. Nongame and Endangered Wildlife Programs, Conservation Data Center. Idaho Department of Fish and Game. 104pp. Fair, W. S., and S. E. Henke Effects of habitat manipulations on Texas horned lizards and their prey. Journal of Wildlife Management 61(4): Ferner, J. W A Review of Marking Techniques for Amphibians and Reptiles. Circular No. 9. Society for the Study of Amphibians and Reptiles. 42pp. Friend, G. R Impact of fire on small vertebrates in Mallee woodlands and heathlands of temperate Australia: A review. Biological Conservation 65: Germano, D. J., and C. R. Hungerford Reptile population changes with manipulation of Sonoran desert shrub. Great Basin Naturalist 41(1): 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 The global decline of reptiles, Déjà vu amphibians. BioScience 50(8): Gibbons, J. W., and P. W. Stangel Conserving amphibians and reptiles in the new millennium. Proceedings of the Partners in Amphibian and Reptile Conservation (PARC) Conference Savannah River Ecology Laboratory Herpoutreach Publication #2. Aiken, SC. 54pp. Gibson, C. W. D., C. Hambler, and V. K. Brown Changes in spider (Araneae) assemblages in relation to succession and grazing management. Journal of Applied Ecology 29: Greenberg, C. H., D. G. Neary, and L. D. Harris Effect of high-intensity wildfire and silvicultural treatments on reptile communities in sand-pine scrub. Conservation Biology 8(4): Greenwald, O. E Thermal dependence of striking and prey capture by gopher snakes. Copeia 1974:

91 Grothe, S Red-tailed Hawk predation on snakes: the effects of weather and snake activity. M.S. Thesis, Idaho State University. 103pp. Hayes, J. P., and R. J. Steidl Statistical power analysis and amphibian population trends. Conservation Biology 11(1): Heske, E. J., and M. C. Campbell Effects of an 11-year livestock exclosure on rodent and ant numbers in the Chihuahuan desert, southeastern Arizona. Southwest Naturalist 36: Hobbs, R. J., and L. F. Huenneke Disturbance, diversity, and invasion: Implications for conservation. Conservation Biology 6(3): Hopkins, W. A Reptile toxicology: opportunities and challenges on the last frontier of vertebrate ecotoxicology. Environmental Toxicology and Chemistry 19: Janzen, D. H The depression of reptile biomass by large herbivores. The American Naturalist 110(973): Johansen, J. R., L. L. St. Clair, B. L. Webb, and G. T. Nebeker Recovery patterns of cryptogamic soil crusts in desert rangelands following fire disturbance. The Bryologist 87(3): Johansen, J. R., and L. L. St. Clair Cryptogamic soil crusts: recovery from grazing near Camp Floyd State Park, Utah, USA. Great Basin Naturalist 46(4): Jones, K. B Effects of grazing on lizard abundance and diversity in western Arizona. The Southwestern Naturalist 26(2): Kahn, W. C Observations on the effect of a burn on a population of Sceloporus occidentalis. Ecology 41: Kiesecker, J. M., A. R. Blaustein, and L. K. Belden Complex causes of amphibian population declines. Nature 410: Knick, S. T., and D. L. Dyer Distribution of black-tailed jackrabbit habitat determined by GIS in southwestern Idaho. Journal of Wildlife Management 61: Knick, S. T., and J. T. Rotenberry Landscape characteristics of fragmented shrubsteppe habitats and breeding passerine birds. Conservation Biology 9(5):

92 Landscape characteristics of disturbed shrubsteppe habitats in southwestern Idaho (U.S.A.). Landscape Ecology 12: Kochert, M. N., and M. Pellant Multiple use in the Snake River Birds of Prey Area. Rangelands 8: Krebs, C. J Ecological Methodology 2 nd Edition. Benjamin Cummings, New York. 620pp. Lawrence, G. E Ecology of vertebrate animals in relation to chaparral fire in the Sierra Nevada foothills. Ecology 47: Laycock, W. A Stable states and thresholds of range condition on North American rangelands: a viewpoint. Journal of Range Management 44(5): Lillywhite, H. B. 1977a. Animal responses to fire and fuel management in chaparral. Pp In: Proceedings of the symposium on the environmental consequences of fire and fuel management in Mediterranean ecosystems. Forest Service, U.S. Department of Agriculture. Washington D. C b. Effects of chaparral conversion on small vertebrates in southern California. Biological Conservation 11: Lillywhite, H. B., and F. North Perching behavior of Sceloporus occidentalis in recently burned chaparral. Copeia 1974(1): Meegaskumbura, M., F. Bossuyt, R. Pethiyagoda, K. Manamendra-Arachchi, M. Bahir, M. C. Milinkovitch, and C. J. Schneider Sri Lanka: An amphibian hot spot. Science 298:379. Morris, M. G Differences between the invertebrate faunas of grazed and ungrazed chalk grassland. II. The faunas of simple turves. Journal of Applied Ecology 5: Preliminary observations on the effects of burning on the Hemiptera (Heteroptera and Auchenorhyncha) of limestone grassland. Biological Conservation 7: Mushinsky, H. R Fire and the Florida sandhill herpetofaunal community: with special attention to responses of Cnemidophorus sexlineatus. Herpetologica 41(3): Natural history and abundance of southeastern five-lined skinks, Eumeces inexpectatus, on a periodically burned sandhill in Florida. Herpetologica 48(3):

93 Mutch, R. W Wildland fires and ecosystems a hypothesis. Ecology 51(6): Nussbaum, R. A., E. D. Brodie, and R. M. Storm Amphibians and Reptiles of the Pacific Northwest. University of Idaho Press, Moscow, Idaho. 332pp. Ortega, R. A., and L. Hernandez Abundancia relativa de insectos en un medio estacional: su influencia en la historia de vida de dos iguanidos simpatricos. Fol. Entomol. Mexi., 55: Patterson, G. B The effect of burning-off tussock grassland on the population density of common skinks. New Zealand Journal of Zoology 11: Pechmann, J. H. K., D. E. Scott, R. D. Semlitsch, J. P. Caldwell, L. J. Vitt, and J. W. Gibbons Declining amphibian populations: the problem of separating human impacts from natural fluctuations. Science 253: Pechmann, J. H. K., and H. M. Wilbur Putting declining amphibian populations in perspective: natural fluctuations and human impacts. Herpetologica 50(1): Peterson, C. R., J. O. Cossel Jr., D. Pilliod and B. M. Bean The occurrence, distribution, relative abundance, and habitat relationships of amphibians and reptiles on the Idaho Army National Guard Orchard Training Area, Ada County, Idaho. Final Report to IDARNG, Boise, Idaho. Phillips, P The distribution of rodents in overgrazed and normal grasslands of central Oklahoma. Ecology 17(4): Pianka, E. R Convexity, desert lizards, and spatial heterogeneity. Ecology 47(6): On lizard species diversity: North American flatland deserts. Ecology 48(3): Comparative autecology of the lizard Cnemidophorus tigris in different parts of its geographic range. Ecology 51: Pough, F. H The advantages of ectothermy for tetrapods. Am. Nat., 115: Amphibians and reptiles as low-energy systems. Pp , In: Behavioral Energetics: Vertebrate Costs of Survival. Edited by W.P. Aspey and S.I. Lustic. Ohio State University Press, Columbus, OH. 82

94 Pough, F. H., R. M. Andrews, J. E. Cadle, M. L. Crump, A. H. Savitzky, and K. D. Wells Herpetology 2 nd ed. Prentice-Hall Inc. New Jersey. 612 pp. Reed, J. M., and A. R. Blaustein Assessment of nondeclining amphibian populations using power analysis. Conservation Biology 9(5): Rie, M. T., K. A. Lendas, B. R. Woodin, J. Stegemen, and I. P. Callard. 2000a. Hepatic biotransformation enzymes in a sentinel species, the painted turtle (Chrysemys picta), from Cape Cod, MA: Seasonal, sex and location differences. Biomarkers 5(5): b. Multiple bioindicators of environmental pollution in a sentinel species, Chrysemys picta, on Cape Cod, MA. Marine Environmental Research vol. 50: Roberts, Jr., T. C Cheatgrass: Management implications in the 90 s. Pp In: Proceedings of a symposium on cheatgrass invasion, shrub die-off, and other aspects of shrub biology and management. Edited by E. D. McArthur, E. M. Romney, S. D. Smith and P. T. Tueller. Intermountain Research Station, Ogden, Utah. Rodriguez-Robles, J. A Alternative perspectives on the diet of gopher snakes (Pituophis catenifer, Colubridae): Literature records versus stomach contents of wild and museum specimens. Copeia 1998(2): Rodriguez-Robles, J. A., and H. W. Greene Food habits of the long-nosed snake (Rhinocheilus lecontei), a specialist predator? Journal of Zoology, London 248: Romero-Schmidt H. L., and A. Ortega-Rubio Changes in lizard abundance on protected versus grazed desert scrub in Baja California Sur, Mexico. Brazilian Archives of Biology and Technology 42(2): Rosen, P. C A monitoring study of vertebrate community ecology in the northern Sonoran Desert, Arizona. University of Arizona Doctoral Dissertation 307pp. Schlesinger C. A., and R. Shine Choosing a rock: perspectives of a bush-rock collector and a saxicolous lizard. Biological Conservation 67: Scott, N.J., and R.A. Seigel The management of amphibians and reptile populations: species priorities and methodological and theoretical constraints. Pp In: Wildlife 2001: Populations. Edited by D.R. McCullough and R.H. Barrett. Elsevier Applied Science, London. 83

95 Simovich, M. A Post fire reptile succession. Cal-Neva Wildlife Transactions St. John, A. D Reptiles of the Northwest: Alaska to California; Rockies to the Coast. Lone Pine Publishing, Washington. 272pp. U.S. Department of the Interior Effects of military training and fire in the Snake River Birds of Prey National Conservation Area. BLM/IDARNG Research Project Final Report. U.S. Geol. Surv., Biol. Res. Div., Snake River Field Sta., Boise, ID. 130pp. Van Berkum, F. H Evolutionary patterns of the thermal sensitivity of sprint speed in Anolis lizards. Evolution 40: Vander Haegen, W. M., S. M. McCorquodale, C. R. Peterson, G. A. Green, and E. Yensen Wildlife of eastside shrubland and grassland habitats. Pp: In: Wildlife habitat relationships in Oregon and Washington. Managing Directors: D. H. Johnson and T. A. O Neil. Oregon State University Press, Corvallis, Oregon. Van Horne, B., G. S. Olson, R. L. Schooley, J. G. Corn, and K. P. Burnham. 1997a. The effects of drought and prolonged winter on demography of Townsend s ground squirrels in different shrubsteppe habitats. Ecological Monographs 67(3): Vitt, L. J., and R. D. Ohmart Ecology and reproduction of Lower Colorado River lizards, Cnemidophorus tigris (Teiidae) with comparisons. Herpetologica 33: Wake, D. B In: Measuring and monitoring biological diversity: Standard methods for amphibians (Biodiversity Handbook). Edited by W. R. Heyer, M. A. Donnelly, R. W. McDiarmid, L. C. Hayek, and M. S. Foster. Smithsonian Institution Press. 384pp. Wiens, J Vertebrate responses to environmental patchiness in arid and semiarid ecosystems. Pp In: The ecology of natural disturbance and patch dynamics. Edited by S. T. Pickett and P. S. White. Academic Press, New York. Wilson, E. O The Diversity of Life. Norton W. W. Company, New York. Winward, A. H Fire in the sagebrush-grass ecosystem the ecological setting. Pp In: Rangeland fire effects: A symposium. Edited by K. Sanders and J. Durham. Idaho Bureau of Land Management, Boise, Idaho. Whisenant, S. G Changing fire frequencies on Idaho s Snake River Plains: ecological and management implications. Pp In: Proceedings of a symposium on cheatgrass invasion, shrub die-off, and other aspects of shrub biology and 84

96 management. Edited by E. D. McArthur, E. M. Romney, S. D. Smith and P. T. Tueller. Intermountain Research Station, Ogden, Utah. Whitford, W. G., and F. M. Creusere Seasonal and yearly fluctuations in Chihuahuan Desert lizard communities. Herpetologica 33: Woodbury, A. M Introduction a ten year study. Pp In: Symposium: a snake den in Toole County, Utah. Edited by A. M. Woodbury et al. Herpetologica 7:1-52. Yensen, D. L A grazing history of southwestern Idaho with emphasis on the Snake River Birds of Prey Area. Bureau of Land Management, Boise, Idaho. 82pp The 1900 invasion of alien plants into southern Idaho. Great Basin Naturalist 41(2): Yensen, D. L., and G. W. Smith Big Sagebrush- Winterfat and big sagebrush- Nuttall saltbush mosaic vegetation in southwestern Idaho. Pp In: Symposium on the biology of Atriplex and related chenopods. Edited by A. R. Tiedmann, E. D. McArthur, H. C. Stutz, R. Stevens and K. L. Johnson. Provo, Utah. Yensen, E., D. L. Quinney, K. Johnson, K. Timmerman, and K. Steenhof Fire, vegetation changes and population fluctuations of Townsend s ground squirrels. American Midland Naturalist 128: Young, J. A., and R. A. Evans Population dynamics after wildfires in sagebrush grasslands. Journal of Range Management 31: Zug, G. R., L. J. Vitt, and J. P. Caldwell Herpetology: An Introductory Biology of Amphibians and Reptiles, 2 nd edition. Academic Press, San Diego, CA. 630pp. 85

97 Table 1. Habitat classification of trapping sites within the SRBPA, for the years (1970s) and (1990s); shading indicates the sites classified as burn-conversion sites for the analyses of burn-conversion effects (*N13 and S13 were not established until 1998). Site 1970s Habitat Type N1 Riparian N2 Talus N3 Talus N4 Riparian N5 Canyon Rim N6 Shadscale N7 Winterfat N8 Shadscale N9 Winterfat N10 Canyon Rim N11 Big Sage N12 Big Sage N13 N/A* S1 Greasewood S2 Sand S3 Grass / Exotic annuals S4 Big Sage S5 Big Sage S6 Canyon Rim S7 Talus S8 Riparian S9 Shadscale S10 Sand S11 Shadscale S12 Greasewood S13 N/A* 1990s Habitat Type Riparian Talus Talus Riparian Canyon Rim Shadscale Winterfat Shadscale Winterfat Canyon Rim Exotic annuals Exotic annuals Big Sage Greasewood Sand Grass / Exotic annuals Exotic annuals Exotic annuals Canyon Rim Talus Riparian Shadscale Sand Shadscale Greasewood Big Sage 86

98 Table 2. The presence and absence of reptile species (number of sites in parentheses) at trapping sites in the SRBPA. Species 1970s Lizards Crotaphytus bicinctores (1) Gambelia wislizenii (5) (2) (2) Phrynosoma platyrhinos (5) (7) (3) Sceloporus occidentalis (4) (3) (2) Uta stansburiana (12) (9) (5) Cnemidophorus tigris (14) Talus (14) slope (11) Snakes Coluber constrictor (4) (11) (11) Hypsiglena torquata (9) (11) (6) Masticophis taeniatus (19) (18) (18) Pituophis catenifer (22) (23) (24) Rhinocheilus lecontei (9) (7) (6) Sonora semiannulata (2) (3) (3) Thamnophis elegans (1) (2) Crotalus viridis (8) (14) (10) 87

99 Table 3. Shannon-Wiener diversity index values of reptiles at trapping sites in the SRBPA (shading indicates burn-conversion sites). Site Reptile 1997 Reptile 1998 Reptile Diversity Diversity Diversity Lizard Diversity 1997 Lizard Diversity 1998 Lizard Diversity Snake 1997 Snake 1998 Snake Diversity Diversity Diversity N N N N N N N N N N N N N S S S S S S S S S S S S S Mean

100 Table 4. Species richness of reptiles at trapping sites in the SRBPA (shading indicates burned-conversion sites). Site Reptile 1997 Reptile 1998 Reptile Richness Richness Richness Lizard 1997 Lizard 1998 Lizard Richness Richness Richness Snake 1997 Snake 1998 Snake Richness Richness Richness N N N N N N N N N N N N N S S S S S S S S S S S S S Mean

101 Table 5. Relative abundance (captures per 100 trap days) of reptiles at trapping sites in the SRBPA (shading indicates burn-conversion sites). Site Reptile Abund Reptile Abund Reptile Abund Lizard Abund Lizard Abund Lizard Abund Snake Abund Snake Abund Snake Abund. N N N N N N N N N N N N N S S S S S S S S S S S S S Mean

102 Table 6. Correlations of habitat variables measured in 1998 at 26 trapping sites within the SRBPA (bold indicates P<0.1). Spearman's rho % Shrub % Herb % Moss % Lichen % Crypto. % Litter % Rock % Soil Cheat grass Russian thistle Exotic annuals % Shrub Corr. coeff Grazing rank Sig. 2-tailed % Herb Corr. coeff Sig. 2-tailed % Moss Corr. coeff Sig. 2-tailed % Lichen Corr. coeff Sig. 2-tailed % Crypto. Corr. coeff Sig. 2-tailed % Litter Corr. coeff Sig. 2-tailed % Rock Corr. coeff Sig. 2-tailed % Soil Corr. coeff Sig. 2-tailed Cheat grass Corr. coeff Sig. 2-tailed Russian thstl. Corr. coeff Sig. 2-tailed Exotic ann. Corr. coeff Sig. 2-tailed Grazing rank Corr. coeff Sig. 2-tailed. 0 Fecal count 91

103 Table 7. Reptile status and habitat correlations measured at 26 trapping sites during 1998 in SRBPA, Idaho (bold indicates P < 0.1). Nonparametric Correlations Spearman's rho % Shrub Cover % Herb Cover % Moss Cover % Lichen Cover % Crypto Cover % Litter Cover % Rock Cover % Bare Soil Cheat grass Russian thistle Exotic annuals Reptile Diversity Corr. Coeff Grazing rank Burrows Rodent abund. Sig. (2-tailed) Lizard Diversity Corr. Coeff Sig. (2-tailed) Snake Diversity Corr. Coeff Sig. (2-tailed) Reptile Richness Corr. Coeff Sig. (2-tailed) Lizard Richness Corr. Coeff ~ Sig. (2-tailed) Snake Richness Corr. Coeff Sig. (2-tailed) Reptile Abund. Corr. Coeff Sig. (2-tailed) Lizard Abund. Corr. Coeff Sig. (2-tailed) Snake Abund. Corr. Coeff ~ Sig. (2-tailed)

104 Table 8. Correlations (Spearman s rho), which are significant at P < 0.1, between reptile species abundance (captures/100 trap days) and habitat variables measured in 1998 at 26 trap sites within the SRBPA. Species % shrub %herb %moss %lichen %crypto. %litter %rock %soil cheat rnk. Russ. rnk. Exot. rnk. Graz. rnk. Fecal # Burrows C. bicinctores Corr. Coeff. ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ G. wislizenii Corr. Coeff. ~ ~ ~ ~ ~ ~ ~ ~ ~ Sig. (2-tailed) P. platyrhinos Corr. Coeff. ~ ~ ~ ~ ~ ~ ~ ~ Sig. (2-tailed) S. occidentalis Corr. Coeff. ~ ~ ~ ~ ~ ~ ~ Sig. (2-tailed) U. stansburiana Corr. Coeff. ~ ~ ~ ~ ~ ~ Sig. (2-tailed) C. tigris Corr. Coeff ~ ~ ~ ~ ~ ~ Sig. (2-tailed) C. constrictor Corr. Coeff. ~ ~ ~ ~ ~ ~ ~ Sig. (2-tailed) H. torquata Corr. Coeff. ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ M. taeniatus Corr. Coeff. ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ P. catenifer Corr. Coeff. ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ Sig. (2-tailed) R. lecontei Corr. Coeff. ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ S. semiannulata Corr. Coeff. ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ Sig. (2-tailed) T. elegans Corr. Coeff. ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ Sig. (2-tailed) C. viridis Corr. Coeff. ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ 93

105 Table 9. A comparison of habitat variables measured in 1998, between burnconversion sites and unburned shrub sites (> 20 years since wildfire) in the SRBPA (bold indicates means that are significantly different at P<0.05). Habitat Variables Site Status N Mean Std. Dev. Std. Error Sig. % Shrub burn no burn % Herb burn no burn % Moss burn no burn % Lichen burn no burn % Crypto. burn no burn % Litter burn no burn % Rock burn no burn % Soil burn no burn Cheatgrass burn ~ no burn Russian thstl. burn ~ no burn Exotic annuals burn ~ no burn Grazing Rank burn ~ no burn # Burrows burn no burn Rodent abund. burn no burn

106 Table 10. A comparison of reptile status as measured in 1998 at burn-conversion sites and unburned shrub sites (> 20 years since last wildfire) in the SRBPA (bold indicates means that are significantly different at P<0.05). Reptile Status Site Status N Mean Std. Devtn. Std. Error Exact Sig. Reptile Diversity burn no burn Lizard Diversity burn no burn Snake Diversity burn no burn Reptile Richness burn no burn Lizard Richness burn no burn Snake Richness burn no burn Reptile Abund. burn no burn Lizard Abund. burn no burn Snake Abund. burn no burn

107 Table 11. Mean capture rates (per 100 trap days) at burn-conversion sites and shrub sites (unburned), measured in 1998 at 17 trapping sites in the SRBPA (bold indicates means that are significantly different at P<0.1). Species Site Status N Mean Std. Dev. Std. Error Sig. G. wislizenii burn no burn P. platyrhinos burn no burn U. stansburiana burn no burn C. tigris burn no burn C. constrictor burn no burn H. torquata burn no burn M. taeniatus burn no burn P. catenifer burn no burn R. lecontei burn no burn S. semiannulata burn no burn C. viridis burn no burn

108 FIGURES Figure 1. Snake River Birds of Prey National Conservation Area boundary and study site locations. 97

109 Cover type page (landscape orientati Figure 2. Riparian Talus slope Greasewood Sand Shadscale Winterfat Canyon rim Grass (Exotic annuals) Big Sage Figure 2. Examples of the nine cover types identified by Diller and Johnson (1982) in the SRBPA. 98

110 Figure 3. Photograph of funnel trap used in conjunction with drift fencing at study sites within the SRBPA, Idaho. Figure 4. Dimensions and configuration of drift fence trapping arrays used in the SRBPA. 99

111 Diversity Index Reptiles Lizards Snakes s Figure 5. Shannon-Wiener diversity index for reptiles encountered at all trapping sites within the SRBPA during , 1997 and Species Richness s Reptiles Lizards Snakes Figure 6. Species richness of reptiles encountered at all trapping sites within the SRBPA during , 1997 and

112 Abundance (captures / 100 days) Reptiles Lizards Snakes s Figure 7. Abundance of reptiles (captures/100 trap days) encountered at all trapping sites within the SRBPA during , 1997 and Captures per 100 trap days CRBI GAWI PHPL SCOC UTST CNTI Figure 8. Lizard species abundance (captures per 100 trap days) at drift fence trapping sites in the SRBPA s

113 Captures per 100 trap days COCO COCO HYTO HYTO MATA MATA PICA PICA RHLE SOSE SOSE THEL THEL CRVI CRVI Figure 9. Snake species abundance (captures per 100 trap days) at drift fence trapping sites in the SRBPA s

114 N N = Site N1 -- Riparian Appendix 1.1. Site photos and map for N1 Riparian cover type, SRBPA. 103

115 N N = Site N2 Talus slope Appendix 1.2. Site photos and map for N2 Talus slope cover type, SRBPA. 104

116 N N = Site N3 Talus slope Appendix 1.3. Site photos and map for N3 Talus slope cover type, SRBPA. 105

117 N N = Site N4 Riparian Appendix 1.4. Site photos and map for N4 Riparian cover type, SRBPA. 106

118 N N = Site N5 Canyon rim Appendix 1.5. Site photos and map for N5 Canyon rim cover type, SRBPA. 107

119 N N = Site N6 Shadscale Appendix 1.6. Site photos and map for N6 Shadscale cover type, SRBPA. 108

120 N N = Site N7 Winterfat Appendix 1.7. Site photos and map for N7 Winterfat cover type, SRBPA. 109

121 N N = Site N8 Shadscale Appendix 1.8. Site photos and map for N8 Shadscale cover type, SRBPA. 110

122 N N = Site N9 Winterfat Appendix 1.9. Site photos and map for N9 Winterfat cover type, SRBPA. 111

123 N N = Site N10 Canyon rim Appendix Site photos and map for N10 Canyon rim cover type, SRBPA. 112

124 N N = Site N11 Big sage / Exotic annuals Appendix Site photos and map for N11 Big sage cover type ( 78), and exotic annuals cover type ( 98), SRBPA. 113

125 N N = Site N12 Big sage / Exotic annuals Appendix Site photos and map for N12 Big sage cover type ( 78), and exotic annuals cover type ( 98), SRBPA. 114

126 N = Site N13 Big sage Appendix Site photos and map for N13 Big sage cover type, SRBPA. 115

127 S S = Site S1 -- Greasewood Appendix Site photos and map for S1 Greasewood cover type, SRBPA. 116

128 S S = Site S2 Sand Appendix Site photos and map for S2 Sand cover type, SRBPA. 117

129 S S = Site S3 Grass / Exotic annuals Appendix Site photos and map for S3 Grass / Exotic annuals cover type, SRBPA. 118

130 S S = Site S4 Big Sage Appendix Site photos and map for S4 Big Sage cover type, SRBPA. 119

131 S S = Site S5-- Big sage / Exotic annuals Appendix Site photos and map for S5 Big sage cover type ( 79), Exotic annuals ( 98), SRBPA. 120

132 S S = Site S6 Canyon rim Appendix Site photos and map for S6 Canyon rim cover type, SRBPA. 121

133 S S = Site S7 Talus slope Appendix Site photos and map for S7 Talus slope cover type, SRBPA. 122

134 S S = Site S8 -- Riparian Appendix Site photos and map for S8 Riparian cover type, SRBPA. 123

135 S S = Site S9 Shadscale Appendix Site photos and map for S9 Shadscale cover type, SRBPA. 124

136 S S = Site S10 Sand Appendix Site photos and map for S10 Sand cover type, SRBPA. 125

137 S S = Site S11 -- Shadscale Appendix Site photos and map for S11 Shadscale cover type, SRBPA. 126

138 S S = Site S12 Greasewood Appendix Site photos and map for S12 Greasewood cover type, SRBPA. 127

139 S = Site S13 Big sage Appendix Site photos and map for S13 Big sage cover type, SRBPA. 128