Herpetologica, 68(2), 2012, 195–202 E 2012 by The Herpetologists’ League, Inc.

POPULATION DENSITY AND SURVIVAL ESTIMATES OF THE AFRICAN VIPERID, SCHNEIDERI

1,2 1 BRYAN MARITZ AND GRAHAM J. ALEXANDER 1School of , Plant, and Environmental Sciences, University of the Witwatersrand, P.O. Wits, 2050, Gauteng, South Africa

ABSTRACT: Although estimates of survival and population density are available for several populations, most of these are for populations in northern temperate environments. Here we present the results of a 3-yr mark–recapture study for an African , the Namaqua Dwarf Adder (Bitis schneideri). We estimated survival, recapture probability, and population density by marking 279 adders (121 males, 108 females, 50 juveniles) on two study sites along the Namaqualand coast of South Africa. We recaptured 54 individuals (27 males, 23 females, 4 juveniles), yielding absolute recapture rates of 20.4% and 17.9% at each site respectively. The adult sex ratio of our sample did not differ significantly from equality. We fitted eight models that included both individual-dependent and time-dependent covariates to our recapture data, and compared models using Akaike information criterion corrected for small samples. The best models included snout–vent length and search effort covariates. Mean (6SD) estimated monthly apparent survival was low (0.86 6 0.11 and 0.86 6 0.12), as were mean recapture likelihoods (0.06 6 0.04 and 0.06 6 0.01). Population density estimates were similar for the two sites (7.52 6 3.62 ha21 and 8.31 6 7.38 ha21). Juvenile exhibited higher survival than adult females, which, in turn, had higher survival than adult males. Juveniles had a lower recapture probability than adult males or females. The likelihood of an individual leaving the study area was 6.8% and 9.4% per month for the larger and smaller sites, respectively. On the basis of our measures of apparent survival and emigration, we estimate annual survival rates of 39% and 56% for the two sites respectively. Population characteristics of B. schneideri differ substantially from those of other viperids, highlighting the need for additional population studies of African snakes. Key words: Bitis schneideri; Mark–recapture; Namaqua Dwarf Adder; Succulent Karoo

ROBUST estimates of population size calcu- (Shine et al., 1998). Although such studies lated using mark–recapture techniques gen- have formed the basis for an increase in the erally require large samples and relatively use of snakes as ecological model organisms high recapture rates. For species that are (Shine and Bonnet, 2000), there is little cryptic and exhibit low detection probability, indication of whether they are representative such as many snakes, these data are difficult of African species, especially given the phylo- to collect. Additionally, working in difficult-to- genetic distinctiveness of African snake line- access or politically unstable areas makes ages (Cadle, 1994). In fact, Janzen (1976) fieldwork problematic. It is therefore not argues that snake abundance in Africa is surprising that Africa, despite its remarkable generally lower than that found in the snake diversity, remains underrepresented in Americas and Australia, but this idea remains the field of snake population ecology (Parker speculative. and Plummer, 1987; Shine et al., 1998; Bonnet Population characteristics such as abun- et al., 2002). Moreover, recent analyses de- dance, density, and survival rate are essential scribing global declines in snake populations in the development of an understanding of the (Reading et al., 2010) highlight the importance autecology of an organism. Although popula- of population-based studies, especially in tion estimates are available for several species regions where biodiversity loss remains poorly of snakes globally (Parker and Plummer, 1987), quantified. relatively few of these have been estimated Most of the theory governing snake ecology using mark–recapture models (Koons et al., has been derived from information garnered 2009). In Africa, the situation is confounded from a few well-studied model species (Siegel by the fact that very few species have been et al., 1987). Geographically, there is a strong investigated. Madsen and Osterkamp (1982) bias toward North America and Europe analyzed mark–recapture data for a population of Lycodomorphus bicolor on Lake Tangan- 2 CORRESPONDENCE: e-mail, [email protected] yika, whereas more recent work in West Africa

195 196 HERPETOLOGICA [Vol. 68, No. 2 has provided density estimates for a few large, topography and vegetation, differing only in forest-associated species such as Bitis gabo- area and distance from the coastline (0.8 km nica, Bitis nasicornis (Luiselli, 2006), and and 2.6 km). Surveys were conducted over Dendroaspis jamesonii (Luiselli et al., 2000). 17 mo between September 2007 and March Given the geographic extent and diversity of 2010. Search effort totaled 596 observer-h, Africa, studies of four species, restricted to and was biased toward summer months when particular biomes, are unlikely to be represen- Dwarf Adders are more active and more likely tative of snake populations across Africa. to be encountered. All individuals were cap- Here we present the estimates of popula- tured by hand and marked by clipping a tion density, apparent survival, and recapture unique sequence of ventral scales (Fitch, probability of an arid-adapted African viperid, 1987). For each snake, we measured mass the Namaqua Dwarf Adder, Bitis schneideri. (to the nearest 0.1 g) using a digital balance, This species is currently listed as vulnerable SVL (to the nearest 1.0 mm) using a ruler, and by the International Union for Conservation sex (as either male, female, or juvenile) on the of Nature (IUCN; World Conservation Mon- basis of male-biased tail dimorphism (males itoring Centre, 1996) and is very poorly have much larger tails than females). Sex- represented in museum collections (fewer based differences in tail length are not than 50 specimens globally). It has tradition- apparent in individuals smaller than 140 mm ally been considered rare (McLachlan, 1978) SVL and these were thus designated as and our study was thus aimed at evaluating juveniles (Maritz and Alexander, 2011). For the of the species. Here, each capture we recorded the geographic our primary objectives are to report on mea- coordinates of the location (using a handheld sures of capture probability (p), survival (Q), global positioning system device, Garmin and density (D) for two populations that occur eTrex and datum WGS1984). in the coastal dune system along the Nama- To provide more accurate estimates of qualand coast in South Africa. population parameters, we included two covariates in our analyses: SVL was included MATERIALS AND METHODS as an individual-dependent continuous vari- Bitis schneideri is an arid-adapted viperid able that ranged from 99 to 251 mm, and that inhabits coastal Sandveld habitats along allowed us to test for age effects on survival the west coast of southern Africa (Broadley, and capture probability. Total search effort 1983; Branch, 1998). It is the smallest viperid, invested during each month was applied to with adults averaging 200 mm snout–vent certain models as a time-dependent covariate length (SVL) and not exceeding 254 mm SVL (effort), calculated as the total duration of all (Maritz and Alexander, 2011). Our unpub- observations at a particular study site, during a lished field observations strongly support the particular capture occasion. We standardized assertion that it is an ambush predator and this covariate so that it ranged between zero that prey consists mainly of small vertebrates, and one. primarily lizards and frogs (Branch, 1998). We fitted Cormack–Jolly–Seber models to The species is currently listed as vulnerable by recapture data from each site separately using the IUCN as a result of habitat degradation the mark–recapture analysis package (Mac- over much of its distribution (Branch, 1988; Donald, 2010) in R (R Development Core World Conservation Monitoring Centre, Team, 2009). We started by fitting a fully 1996). parameterized model (i.e., Q[svl] p[effort, svl])to Our study was conducted on the farm the capture histories from the two popula- Noup, in the Northern Cape Province, South tions, and tested for goodness of fit using Africa (30u089S, 17u129E). The study site has Hosmer–Lemeshow (Hosmer and Lemeshow, been described elsewhere (Maritz and Scott, 2000) and Osius–Rojek (Osius and Rojek, 2010; Maritz and Alexander, 2011, 2012). We 1992) goodness-of-fit statistics. We then fitted searched for Namaqua Dwarf Adders at two an additional seven reduced models (Table 1) localities: a 27-ha north site and a 16-ha south and compared them using Akaike information site. The sites were similar with regard to criterion corrected for small samples (AICc; June 2012] HERPETOLOGICA 197

TABLE 1.—Relative performance of eight candidate mark–recapture models. Q 5 survival; svl 5 snout–vent length; p 5 monthly capture probability; effort 5 sampling effort; AICc 5 Akaike’s information criterion for small samples. Variables included in parentheses after parameters, (.) denotes that parameter was kept constant.

North site South site

Model No. parameters AICc DAICc AICc weight AICc DAICc AICc weight

Q(svl)p(svl, effort) 4 314.30 0.00 0.41 197.85 0.00 0.28 Q(svl)p(effort) 4 318.02 3.72 0.06 198.69 0.84 0.18 Q(svl)p(svl) 3 330.36 16.06 0.00 198.86 1.02 0.17 Q(svl)p(.) 3 333.80 19.50 0.00 199.80 1.95 0.10 Q(.)p(svl, effort) 4 314.32 0.02 0.41 201.56 3.71 0.04 Q(.)p(effort) 3 316.89 2.59 0.11 199.45 1.60 0.13 Q(.)p(svl) 3 330.37 16.07 0.00 202.63 4.79 0.03 Q(.)p(.) 2 332.70 18.40 0.00 200.55 2.70 0.07

Burnham and Anderson, 2002). To account movements with random turning angles. for variation in model outputs we averaged all Model movements were based on mean daily model estimates for Q, p,andN and weighted step lengths of 3.2 6 4.8 m d21 (Maritz and the contribution of each model using AICc Alexander, 2012). Although the estimate of weight scores. daily displacement underrepresents the move- To test for differences in the survival and ment of male snakes during spring, our radio- recapture probabilities of adult males, adult telemetry work (Maritz and Alexander, 2012) females, and juveniles, we averaged our esti- suggests that male snakes return to a smaller mates of Q and p for each individual, pooled central portion of their home ranges after the data for our populations, and tested for mate-searching forays. Thus, increased spring differences using single-factor ANOVA. We movements of males are unlikely to have a used Tukey honestly significant difference significant impact on our estimate of emigra- post hoc tests to assess the statistical signifi- tion. For the simulation we assumed that cance of paired contrasts. juveniles and adults experience similar rates of Apparent mortality (1 2 Q) comprises those mortality and emigration likelihood. Although that do not survive to the next sampl- this is unlikely to be the case, our objective ing occasion, as well as those that emigrate was to simulate emigration for the population from the study site. The relative contribution as a whole that would predominantly comprise of emigration to estimates of apparent mor- adults. We ran each cohort of 1000 individuals tality have been poorly investigated for most 10 times for each study site, and calculated snakes (Parker and Plummer, 1987) but mean (6SD) likelihood of emigration. remains important in understanding overall survival. We attempted to apportion apparent RESULTS mortality by simulating the relative contribu- Search effort totaled 346 and 250 observer- tions of emigration and mortality. To estimate h on the north and south sites, respectively. In the likelihood of an individual having emi- total, we captured 279 snakes, comprising 121 grated from our study sites after 30 d (i.e., the males, 108 females, and 50 juveniles. Of these, likelihood that an animal initially detected 162 snakes (70 males, 67 females, and 25 on the study site will be absent during the juveniles) were captured at the north site, and following sampling occasion), we produced a 117 individuals (51 males, 41 females, and 25 random walk simulation in Microsoft Excel juveniles) were captured at the south site. in which 1000 individuals were distributed Twenty-seven individuals (13 males, 11 fe- randomly within an appropriately sized rect- males, and 3 juveniles) from the north site angular plot, and allowed to move indepen- were recaptured once, and six individuals (two dently in a stepwise manner. Data collected males and four females) were recaptured from radiotelemetered individuals (Maritz twice. Nineteen individuals (10 males, 8 and Alexander, 2012) suggested no directional females, and 1 juvenile) from the south site biases in movements, allowing us to model were recaptured once, and two males were 198 HERPETOLOGICA [Vol. 68, No. 2

TABLE 2.—Model-averaged estimates of population parameters estimated for Bitis schneideri at two study sites in South Africa. Q 5 survival; p 5 monthly capture probability; N 5 number of individuals at site; D 5 population density (ha21).

North site South site

Q pNDQ pND Mean 0.86 0.06 203 7.52 0.86 0.06 133 8.31 SD 0.11 0.04 98 3.62 0.12 0.01 118 7.38 Range 0.57–0.91 0.01–0.24 99–388 3.7–14.4 0.33–0.96 0.03–0.11 22–357 1.4–22.3

recaptured twice. Absolute recapture rates juveniles (F2,275 5 116.34; P , 0.001), with from the north and south sites did not differ juvenile snakes (0.93 6 0.0063) exhibiting from one another (x2 5 0.016, P 5 0.68), significantly higher (mean 6 SE) survival than returning 20.4% and 17.9% of all marked adult females (0.84 6 0.0049), which in turn animals respectively. The adult sex ratio for exhibited significantly higher survival than captured adult snakes was 1:1.12 (female: adult males (0.82 6 0.0046). Recapture male), which was not significantly different probability also varied significantly among from unity (x2 5 0.74, P 5 0.39). adult males, adult females, and juveniles Goodness-of-fit tests of our most parame- (F2,275 5 295.11; P , 0.001); however, this terized model at each site provided no significant effect was a result of juvenile indication of a lack of fit at either the north snakes exhibiting significantly lower (0.036 6 site (Z 5 1411.19, P 5 0.99; H 5 3.13, P 5 0.0011) recapture likelihood than adult males 0.37) or south site (Z 5 1032.81, P 5 0.58; (0.066 6 0.00082) and females (0.064 6 H 5 4.75, P 5 0.19), which allowed us to 0.00078). continue to work with our suite of reduced The area of each site influenced likelihood models. of emigration from that site. At the larger 6 At both sites, the saturated model (Q[svl] north site, mean ( SD) likelihood of emigra- p[svl, effort]) performed best (i.e., attained the tion after 30 d was estimated to be 6.8% lowest AICc score; Table 1). However, several (60.5%), which was lower than the mean other models scored similar AICc values: one estimated likelihood of emigration for the model at the north site and four models at the south site, 9.4% (61.0%). Using these esti- south site attained DAICc scores of less than mates of emigration likelihood increases the two. We chose the saturated model from the estimated mean monthly survival to 0.92 and candidate models at both sites on the basis of 0.95 for the north and south sites, respective- their relatively low AICc scores, and for the ly, and these translate into annual survival sake of congruence in models across our study estimates of 0.39 and 0.56, respectively. sites. At both sites, the saturated model suggested minor influences of SVL on either DISCUSSION survival or recapture probability. Search effort Our findings represent one of the first was positively associated with capture proba- robust, empirical assessments of the survival bility at both sites. rate, capture probability, and population Averaged estimates of Q and p were very density of any African snake. Bitis schneideri similar for the two sites (Table 2). Because of study populations in coastal Namaqualand did differences in site area, the north site pro- not exhibit a biased sex ratio, experienced duced larger estimates of N than the south low apparent survival rates, and occurred at site. However, area-corrected density esti- relatively high population densities. Addition- mates for the two sites were similar (Table 2). ally, juvenile snakes exhibited relatively high Assuming that survival is fixed through the apparent survival and relatively low recapture course of a year, our apparent monthly sur- probabilities compared with adults. Adult vival estimate of 0.86 translates into an annual males exhibited significantly lower apparent apparent survival estimate of 0.19. survival than adult females, but similar Apparent monthly survival differed signifi- recapture probabilities. Finally, emigration cantly among adult males, adult females, and may account for less than 10% of apparent June 2012] HERPETOLOGICA 199 mortality. The general congruence among schneideri and other ambush-foraging snakes), model outputs for the two populations studied bouts of movement are likely to pose an support our findings. In general, capture rates increased predation risk (Bonnet et al., 1999). were low, and as a result many of the Given that mate-searching movements appear parameters that we have estimated are asso- to be the chief motivation for movement in this ciated with wide confidence limits. However, species (Maritz and Alexander, 2012) and that such problems are unavoidable for cryptic juvenile snakes do not mate-search, then organisms such as many snake species (Turn- juvenile snakes may face greatly reduced er, 1977; Parker and Plummer, 1987). predation risk and thus higher survival. This Studies of African snake population dynam- idea is supported by our concurrent estimates ics are rare, making direct comparison of our of recapture probability that are relatively low result with other African species difficult. To (see below), and relatively low measures of our knowledge, no other estimates of survival daily displacement by juvenile snakes (Maritz exist for an African snake population. Numer- and Alexander, 2012). ous studies on primarily European and North We calculated a mean recapture probability American populations suggest that (apparent) of <6% for our study population. Recapture survival varies widely among snake species probability was low compared with those (0.35–0.85; Shine and Charnov, 1992). More estimated for other snake populations (Diller recently, several mark–recapture modeling and Wallace, 2002; Lind et al., 2005; Brown analyses of snake populations have reported et al., 2007), was strongly influenced by search survival estimates for viperids including Vi- effort, but weakly influenced by SVL. We pera aspis (0.75; Flatt et al., 1997), Crotalus detected no difference in recapture probabil- oreganus (0.55–0.82; Diller and Wallace, ity of adult males and females, despite males 2002), Crotalus horridus (0.82–0.95; Brown moving much farther than female snakes et al., 2007), and Agkistrodon piscivorus (0.79; during the spring mate-searching season Koons et al., 2009); colubrids including (Maritz and Alexander, 2012). This finding Thamnophis radix (0.35–0.45; Stanford and supports the hypothesis that outside of the King, 2004) and Thamnophis atratus (0.56– brief mate-searching season, male and female 0.64; Lind et al., 2005); and elapids includ- snakes exhibited similar behavioral reper- ing Hoplocephalus bungaroides (0.82; Webb toires. Mean (6SE) juvenile recapture likeli- et al., 2003) and Rhinoplocephalus nigrescens hood was low (3.6%), supporting our hypoth- (0.74; Webb et al., 2003). esis of reduced movement during this life Our estimates of survival for B. schneideri phase. (0.39–0.56) are low in comparison with several Density estimates for African snake popu- of the estimates cited above, especially in lations are also rare, making generalizations comparison with measures for other viperids. regarding African snake populations difficult. Shine and Charnov (1992) demonstrated a Luiselli (2006) estimated that both B. gabo- strong relationship between mean annual nica (0.16 ha21) and B. nasicornis (0.10 ha21) survival and age at maturity for snakes, with occurred at much lower densities than our rapidly maturing species exhibiting lower measures for B. schneideri. Additionally, D. annual survival rates. Maritz and Alexander jamesonii, a large elapid snake, is reported to (2011) show that sexual maturity is reached at occur at low population densities (0.11 ha21; the age of 10–15 mo in B. schneideri. This Luiselli et al., 2000). Conversely, the aquatic rapid maturation and the low survival report- L. bicolor is reported to occur at densities ed here fit the trend described by Shine and of up to 380 ha21 (Madsen and Ostercamp, Charnov (1992). 1982), but this measure may be inflated by Our finding that juvenile B. schneideri have snakes being attracted to a focal feeding point. higher survival rates than do adults may be the Parker and Plummer (1987) reviewed snake first corroborating evidence of the prediction population densities, and found that estimates of Pike et al. (2008) that juvenile snakes exhibit of population density ranged from ,1ha21 behavioral traits that reduce predation risk. In (including the viperids Crotalus cerastes, C. many ambush-foraging species (such as B. horridus, and C. oreganus)to.1000 ha21 200 HERPETOLOGICA [Vol. 68, No. 2

(Diadophis punctatus and Regina alleni). If viewed as a proportion of apparent However, not all of the estimates reported mortality, our simulations suggest that emi- by Parker and Plummer (1987) are the result gration could have important implications for of robust mark–recapture analyses, and few the interpretation of survival. Our estimates of are the result of models that simultaneously emigration show that it could make up 49% of include estimates of recapture probability apparent monthly mortality at the north site (e.g., Koons et al., 2009). Generally, viperid and 67% at the south site due to the dif- populations do not occur at densities .10 ha21 ference in size of the sites. This finding also (Parker and Plummer, 1987). Two noteworthy has important implications for other studies exceptions are the island endemics Gloydius that have not factored emigration out of appa- shedaoensis (<273 ha21; Sun et al., 2001) and rent survival measures. However, our emi- Bothrops insularis (<55 ha21; Martins et al., gration-corrected measures of monthly and 2008). Although Bitis schneideri populations annual survival are relatively low, and do do not reach such extremes, our measures not change the interpretation that our study (7.52 ha21 and 8.31 ha21) are still high system experiences high mortality. Parker and compared with most viperid populations. Plummer (1987) summarized the population Peters and Wassenberg (1983) demonstrated traits of snakes, and described early-maturing that for most groups of animals, smaller- temperate colubrids (low adult survivorship, bodied species tend to occur at higher low longevity, annual reproduction, high densities. Although this pattern has not been fecundity). Bitis schneideri is more closely unequivocally demonstrated for snakes (Lui- allied with these early-maturing colubrids selli et al., 2005), our relatively high popula- than the late-maturing viperids of temperate tion density estimates for B. schneideri are regions. concordant with this generalization. Our study represents one of very few studies Partitioning the contributions of mortality that have empirically assessed population pa- and emigration on apparent mortality is rameters for an African snake population. We problematic, and remains a major challenge have demonstrated that B. schneideri exhibits facing mark–recapture analysis. Our analysis low recapture probability, high mortality rates, represents a first attempt at understanding and relatively high population densities. The these relative contributions. Most studies development of generalized theories that ap- report apparent survival, as we have, but with propriately describe African snake ecology will little or no discussion of the contribution of require the study of additional African species. emigration. Our random-walk model showed Until this situation is remedied, management that in the short term, emigration from actions related to African snake populations will the study site by a marked individual was remain speculative at best. , relatively low ( 10%). Bitis schneideri are Acknowledgments.—We thank E. Oppenheimer and relatively sedentary animals, occupying small Son, the Rufford Small Grants Foundation, and the South home ranges (Maritz and Alexander, 2012). African National Biodiversity Institute for financial Given the size of our study sites and the support. The Wessels family and numerous fieldworkers sedentary nature of our study animals, the provided valued logistic support. S. Nielsen provided valuable comments on this manuscript. All work was likelihood of a marked individual emigrating cleared under Northern Cape Province Department of from the study site is dependent upon the Tourism, Environment and Conservation (0914/07 and initial capture location of the animal. Accord- FAUNA 698/2009) and University of the Witwatersrand ingly, estimates of emigration are probably Animal Ethics Screening Committee (2007/68/1 and 2007/ 69/3) permits. Two reviewers improved an earlier version more sensitive to study-plot size and shape of this manuscript. than to the duration of the study, because increased duration gives the snakes the time to emigrate, but also gives them time to return LITERATURE CITED to the plot. We therefore surmise that our Bonnet, X., G. Naulleau, and R. Shine. 1999. Dangers of leaving home: Dispersal and mortality in snakes. estimates for a 30-d sampling occasion pro- Biological Conservation 89:39–50. vide an appropriate estimate of emigration Bonnet, X., D. Pearson, M. Ladyman, O. Lourdais, likelihood. and D. Bradshaw. 2002. ‘‘Heaven’’ for serpents? A June 2012] HERPETOLOGICA 201

mark–recapture study of tiger snakes (Notechis scuta- bicolor in Lake Tanganyika. Journal of Herpetology tus) on Carnac Island, Western Australia. Austral 16:185–188. Ecology 27:442–450. Maritz, B., and G.J. Alexander. 2011. Morphology, sexual Branch, W.R. 1988. South African Red Data Book: dimorphism, and growth in the smallest viperid, Bitis and Amphibians. South African National schneideri (Reptilia: : ). Journal of Scientific Projects Report 151. CSIR, South Africa. Herpetology 44:1–6. Branch, W.R. 1998. Field Guide to the Snakes and Other Maritz, B., and G.J. Alexander. 2012. Dwarfs on the move: Reptiles of Southern Africa. Struik Publishers, South Spatial ecology of the world’s smallest viper, Bitis Africa. schneideri. Copeia 2012:115–120. Broadley, D.G. 1983. FitzSimmons’ Snakes of Southern Maritz, B., and S.L. Scott. 2010. Assessing the threat of Africa. Delta Books, South Africa. avian predation on a small viperid snake. African Brown, W.S., M. Ke´ry, and J.E. Hines. 2007. Survival of Zoology 45:309–314. Timber Rattlesnakes (Crotalus horridus) estimated by Martins, M., R.J. Sawaya, and O.A.V. Marques. 2008. A capture–recapture models in relation to age, sex, color first estimate of the population size of the critically morph, time, and birthplace. Copeia 2007:656–671. endangered lancehead, Bothrops insularis. South Burnham, K.P., and D.R. Anderson. 2002. Model American Journal of Herpetology 3:168–174. Selection and Inference: A Practical Information- McDonald, T. 2010. MRA: Analysis of mark–recapture Theoretic Approach, 2nd Ed. Springer-Verlag, USA. data. R package version 2.5. Available at http://cran. Cadle, J.E. 1994. The colubrid radiation in Africa r-project.org/web/packages/mra/. (Serpentes: Colubridae): Phylogenetic relationships McLachlan, G.R. 1978. South African Red Data Book— and evolutionary patterns based on immunological Reptiles and Amphibians. South African National data. Zoological Journal of the Linnaean Society Scientific Programmes Report 23. CSIR, South Africa. 110:103–140. Osius, G., and D. Rojek. 1992. Normal goodness-of-fit Diller, L.V., and R.L. Wallace. 2002. Growth, reproduc- tests for multinomial models with large degrees of tion, and survival in a population of Crotalus viridis freedom. Journal of the American Statistical Association oreganus in north central Idaho. Herpetological Mono- 87:1145–1152. graphs 16:26–45. Parker, W.S., and M.V. Plummer. 1987. Population ecology. Pp. 253–301 in R.A. Siegel, J.T. Collins, and Fitch, H.S. 1987. Collecting and life history techniques. S.S. Novak (Eds.), Snakes: Ecology and Evolutionary Pp. 143–165 in R.A. Seigel, J.T. Collins, and S.S. Novak Biology. McGraw Hill, USA. (Eds.), Snakes: Ecology and Evolutionary Biology. Peters, R.H., and K. Wassenberg. 1983. The effect of McGraw Hill, USA. body size on animal abundance. Oecologia 60:89–96. Flatt, T., S. Dummermuth, and B.R. Anholt. 1997. Mark– Pike, D.A., L. Pizzatto, B.A. Pike, and R. Shine. 2008. recapture estimates of survival in populations of the Asp Estimating survival rates of uncatchable animals: The Viper, Vipera aspis aspis. Journal of Herpetology myth of high juvenile mortality in reptiles. Ecology 31:558–564. 89:607–611. Hosmer, D.W., and S. Lemeshow. 2000. Applied Logistic R Development Core Team. 2009. R: A language and Regression. Wiley, USA. environment for statistical computing. R Foundation Janzen, D.H. 1976. The depression of biomass by for Statistical Computing, Austria. ISBN 3-900051-07- large herbivores. American Naturalist 110:371–400. 0, URL http://www.R-project.org. Koons, D.N., R.D. Birkhead, S.M. Boback, M.I. Williams, Reading, C.J., L.M. Luiselli, G.C. Akani, X. Bonnet, G. and M.P. Greene. 2009. The effect of body size on Amori, J.M. Ballouard, E. Filippi, G. Naulleau, D. Cottonmouth (Agkistrodon piscivorus) survival, recap- Pearson, and L. Rugiero. 2010. Are snake populations ture probability, and behavior in an Alabama swamp. in widespread decline? Biology Letters 6:777–780. Herpetological Conservation and Biology 4:221–235. Shine, R., and X. Bonnet. 2000. Snakes: A new ‘‘model Lind, A.J., H.H. Welsh, Jr., and D.A. Tallmon. 2005. organism’’ in ecological research? Tree 15:221–222. Garter snake population dynamics from a 16-year study: Shine, R., and E.L. Charnov. 1992. Patterns of survival, Considerations for ecological modelling. Ecological growth, and maturation in snakes and lizards. American Applications 15:294–303. Naturalist 139:1257–1269. Luiselli, L. 2006. Site occupancy and density of sympatric Shine, R., W.R. Branch, P.S. Harlow, and J.K. Webb. Gaboon viper (Bitis gabonica) and Nose-horned viper 1998. Reproductive biology and food habits of Horned (Bitis nasicornis). Journal of Tropical Ecology 22: Adders, Bitis caudalis (Viperidae), from southern 555–564. Africa. Copeia 1998:391–401. Luiselli, L., F.M. Angelici, and G.C. Akani. 2000. Large Siegel, R.A., J.T. Collins, and S.S. Novak (Ed.). 1987. elapids and arboreality: The ecology of Jameson’s green Snakes: Ecology and Evolutionary Biology. Macmillan, mamba (Dendroaspis jamesoni) in an Afro-tropical USA. forested region. Contributions to Zoology 69:1–11. Stanford, K.M., and R.B. King. 2004. Growth, survival, Luiselli, L., G.C. Akani, L. Rugiero, and E. Politano. 2005. and reproduction in a northern Illinois population Relationships between body size, population abun- of the Plains Gartersnake, Thamnophis radix. Copeia dance and niche characteristics in the communities of 2004:465–478. snakes from three habitats in southern Nigeria. Journal Sun, L., R. Shine, Z. Debi, and T. Zhengren. 2001. Biotic of Zoology 265:207–213. and abiotic influences on activity patterns of insular pit- Madsen, T., and M. Osterkamp. 1982. Notes on the vipers (Gloydius shedaoensis, Viperidae) from north- biology of the fish-eating snake Lycodonomorphus eastern China. Biological Conservation 97:387–398. 202 HERPETOLOGICA [Vol. 68, No. 2

Turner, F.B. 1977. The dynamics of populations World Conservation Monitoring Centre. 1996. Bitis of squamates, crocodilians, and rhynchocephalians. schneideri. In IUCN 2006: 2006 IUCN Red List of Pp. 157–264 in C. Gans and D.W. Tinkle (Eds.), Threatened Species. Available at http://www.iucnredlist. Biology of the Reptilia, Vol. 7. Academic Press, UK. org. Downloaded on 11 June 2007. Webb, J.K., B.W. Brook, and R. Shine. 2003. Does foraging mode influence life history traits? A compar- ative study of growth, maturation and survival of two species of sympatric snakes from south-eastern Aus- .Accepted: 3 January 2012 tralia. Austral Ecology 28:601–610. .Associate Editor: Rulon Clark