ECOGRAPHY 26: 107–117, 2003
Interspecific patterns of species richness, geographic range size, and body size among New World venomous snakes Robert N. Reed
Reed, R. N. 2003. Interspecific patterns of species richness, geographic range size, and body size among New World venomous snakes. – Ecography 26: 107– 117. Many higher taxa exhibit latitudinal gradients in species richness, geographic range size, and body size. However, these variables are often interdependent, such that examinations of univariate or bivariate patterns alone may be misleading. Therefore, I examined latitudinal gradients in, and relationships between, species richness, geographic range size, and body size among 144 species of New World venomous snakes [families Elapidae (coral snakes) and Viperidae (pitvipers)]. Both lineages are monophyletic, collectively span 99° of latitude, and are extremely variable in body size and geographic range sizes. Coral snakes exhibit highest species richness near the equator, while pitviper species richness peaks in Central America. Species – range size distributions were strongly right-skewed for both families. There was little support for Bergmann’s rule or Rapoport’s rule for snakes of either family, as neither body size nor range size increased significantly with latitude. However, range area and median range latitude were positively correlated above 15° N, indicating a possible ‘‘Rapoport effect’’ at high northern latitudes. Geographic range size was positively associated with body size. Available continental area strongly influenced range size. Comparative (phylogenetically-based) analyses revealed that shared history is a poor predictor of range size variation within clades. Among vipers, trends in geographic range sizes may have been structured more by historical biogeography than by macroecological biotic factors. R. N. Reed (
[email protected]), Drawer E., Sa6annah Ri6er Ecology Lab., Aiken, SC 29802, USA.
Within many higher taxa, latitude appears to have profound effects on species richness, geographic range sizes, and body sizes. Species richness tends to peak in tropical regions and declines with increasing latitude (Dobzhansky 1950, Pianka 1966), and numerous mechanisms have been advanced to explain this phenomenon (Blackburn and Gaston 1997, Rohde 1999). One such mechanism is purportedly evidenced by Rapoport’s rule, the tendency for geographic range size of species within a given taxon to increase with latitude (Stevens 1989). Rapoport’s rule is often explained in terms of differential abilities of species to attain large range sizes. The high annual climatic variability experienced by organisms in temperate zones may set a high minimum threshold of conditions that an organism must be able to withstand in order to persist. Adaptation to
such a wide range of conditions would presumably result in the ability to disperse to a variety of habitats and attain large geographic ranges. In contrast, tropical organisms experience stable climates with a narrow range of environmental conditions. This may limit their ability to disperse through areas with more extreme conditions (Allee et al. 1949, Janzen 1967), resulting in small geographic range sizes in tropical areas. Unfortunately, conclusions on the generality of Rapoport’s rule are precluded by the uneven taxonomic and latitudinal representation of organisms examined thus far (Gaston and Blackburn 1999, Ashton 2001). Body size may complicate the relationship between latitude and geographic range size. In many higher taxa, geographic range increases with body size (Gaston and Blackburn 1996a, b). When combined with
Accepted 11 June 2002 Copyright © ECOGRAPHY 2003 ISSN 0906-7590 ECOGRAPHY 26:1 (2003)
107
Bergmann’s rule (which states that body size increases with latitude), the relationship between body size and geographic range area could conceivably bias tests of Rapoport’s rule. This would invalidate most ecological explanations for Rapoport’s rule, as the relationship would be incidental rather than explanatory. The effects of continental shape and/or area are only occasionally addressed in macroecological studies of geographic range (Rosenzweig 1995, Rohde 1998, Hecnar 1999, Chown and Gaston 2000), but may also obscure geographic range size gradients. In the New World, this concept may be envisioned by imagining North and South America as two downward-pointing triangles, one on top of each other, with the equator near the top of the lower triangle. In North America, increased geographic range sizes in northern biomes could be influenced more by the large longitudinal extent of the continent than by latitude per se. Conversely, range sizes may be highly constrained at southerly latitudes in South America. Potential relationships between geographic range size, body size, and continental shape mean that examining a single univariate or bivariate pattern is unlikely to be of great utility when asking questions about latitudinal gradients. Therefore, this paper investigates interactions between these variables, largely focusing on mechanisms structuring geographic ranges of New World venomous snakes. First, I investigate univariate latitudinal patterns of species richness, body size, and species – range size distributions, as well as relationships between body size, geographic range and effects of continental area. I am thus able to examine potential interactions between Rapoport’s rule, Bergmann’s rule, and the species richness gradient. Reptiles have received scant attention in the macroecological literature. Stevens’ (1989) original advancement of Rapoport’s rule was based on a number of higher taxa, including a sample of eastern North American reptiles and amphibians. However, of 24 studies published 1998 addressing relationships between geographic range and latitude, only Stevens’ (1989) study had included reptiles (Gaston et al. 1998a). Similarly, none of the 24 studies addressing interspecific body size – range size relationships reviewed by Gaston and Blackburn (1996a) dealt with reptiles or amphibians. Among turtles, latitudinal variation in geographic range size largely contradicts Rapoport’s rule at both continental and global scales (Hecnar 1999). Most germane to the present study, the viperid snakes of South America exhibit a positive relationship between body size and geographic range (Bonfim et al. 1998); this study was confined to 36 species ranging wholly within the South American continent, thus comprising a partial analysis (sensu Gaston and Blackburn 1996a). New World snakes of the families Elapidae and Viperidae provide moderately large (n = 169 total species) groups for examining factors influencing geo108
graphic range. These families encompass a latitudinal range spanning 99°, from 48°S to 51°N. None of these species reaches the absolute northern or southern boundaries of the New World, largely avoiding biases in range size analyses caused by latitudinal ‘‘hard boundaries’’ (Colwell and Hurtt 1994, Lyons and Willig 1997). These snakes primarily prey on vertebrates, and thus variation in geographic range due to differences in trophic level should be minimized, unlike other groups (e.g., turtles; Hecnar 1999). However, morphological, behavioral, and ecological differences between the families may have divergent effects on geographic range. New World elapids (coral snakes) are secretive, aposematically-colored, often semi-fossorial species, found primarily in moist tropical forests. These slender snakes actively search for elongate prey items (usually other snakes). Coral snake maximal adult body lengths range from 280 mm (Micrurus bocourti) to over 1600 mm (M. spixii), and the elevational ranges of some species reach 2400 m (Roze 1996). New World pitvipers (subfamily Crotalinae) are mostly cryptic terrestrial or arboreal species found in a wide variety of habitats, including wet and dry tropical forests, deserts, and high montane areas. These heavy-bodied snakes are chiefly ambush predators feeding on a variety of prey, although large species primarily eat endotherms (Campbell and Lamar 1989, Greene 1997). Maximal adult body lengths of New World viperids range from 460 mm (Crotalus trans6ersus) to over 3000 mm (Lachesis muta), and the elevational ranges of some species exceed 4300 m.
Methods and materials Data were gleaned from a variety of primary and secondary sources (Stebbins 1985, Campbell and Lamar 1989, Ernst 1992, Roze 1996, Zamudio and Greene 1997, Conant and Collins 1998, Jorge da Silva and Sites 1999). Island endemics and species known from only a few specimens were discarded from the data set, resulting in a solely continental fauna with reasonably wellknown geographic range areas. When literature sources conflicted for a variable (e.g., geographic range area or body size), the arithmetic average of the sources was used. The maximal recorded total body length for each species was used as a measure of body size. Many authors have examined the relationship between median latitude of a species’ range and area of the geographic range (e.g., Taylor and Gotelli 1994, Blackburn and Gaston 1996, Hughes et al. 1996). Often, however, they neglect to test the assumption that the latitudinal extent of geographic ranges [the foundation of Stevens’ (1989) claims] and actual range area are tightly correlated. If the east-west and north-south dimensions of geographic ranges are not approximately ECOGRAPHY 26:1 (2003)
equal, then tests of Rapoport’s rule using range area rather than latitudinal extent may be biased. I therefore used estimates of both geographic range area and latitudinal extent in my analyses. Numerous methods are commonly used to measure geographic range area, making comparisons between studies difficult. Area of occupancy (the area within a putative geographic range which is actually occupied) is rarely equal to extent of occurrence (the sum of all areas which fall within the limits of occurrence of the species; Gaston 1991). The latter is commonly used in the macroecological literature, with the understanding that it is probably an overestimate of actual area used. My measure of geographic range area corresponded to the extent of occurrence of each snake species, as follows: I overlaid a grid (6.3 grid squares/linear cm) on the published range map, counting the number of squares falling within the geographic range, and correcting for scale of the map. If \50% of a given square was occupied by a portion of the geographic range, that entire square was included in calculations. I also determined the southernmost and northernmost extents and latitudinal midpoint of geographic range for all species. To gauge the effect of continental area on range size, I calculated the width (km) of New World continental landmasses in increments of 2.5° of latitude. All species with the latitudinal midpoint of their geographic ranges falling within a given 2.5° increment were assigned the corresponding continental width as an estimate of available longitudinal land area at the latitudinal center of the range. Phylogeny may confound examinations of interspecific trends in geographic range areas, as geographic range has been considered a heritable phenotypic trait over evolutionary time in some lineages (Jablonski 1987). Snake phylogeny is poorly understood for many clades, such that rigorous analyses are problematic. However, existing phylogenies allow at least rudimentary assessment of the historical effects on patterns of geographic range among New World venomous snakes. I used network autocorrelation to calculate the effect of phylogeny on variation in geographic range area and body size; this method is not dependent on a completely resolved phylogeny. Network autocorrelation partitions variation within a trait vector into two components: variation due to shared evolutionary history of species, and variation due to the subsequent evolution of species independent of phylogeny (Cheverud et al. 1985, Miles and Dunham 1992). I computed autocorrelation coefficients using a program cited in Miles and Dunham (1992). The resultant autocorrelation coefficient (r) and R2 values indicate the proportion of variation in the variable of interest which is attributable to phylogeny. I followed the phylogeny of Slowinski (1995) when constructing autocorrelation matrices for the family Elapidae. This phylogeny is poorly resolved at the ECOGRAPHY 26:1 (2003)
species level, but is the sole available cladogram of relationships within the New World Elapidae (Sites and Slowinski pers. comm.). To Slowinski’s (1995) tree, I added the putatively monophyletic Micrurus frontalis species group (M. altirostris, M. baliocoryphus, M. brasiliensis, M. diana, M. frontalis, M. pyrrhocryptus, M. tricolor; Jorge da Silva and Sites 1999) nested within the ‘‘short-tailed’’ coral snakes, as well as the putatively monophyletic Micrurus psyches species group (M. circinalis, M. medemi, M. paraensis, M. psyches, M. remotus; Roze 1996) nested within the ‘‘long-tailed’’ coral snakes. A number of phylogenies of members of the Viperidae based on morphological and molecular characters have been published recently (Werman 1993, Kraus et al. 1996, Vidal et al. 1997, Gutberlet 1998, Parkinson 1999, Werman et al. 1999, Parkinson et al. 2000). Kraus et al. (1996) and Parkinson (1999) independently presented evidence that the New World pitvipers are monophyletic, resulting from a single invasion from Old World ancestors. However, these studies are discordant regarding the relationships of the taxa of interest in this study [see Wuster et al. (1997) for a synopsis of some recent phylogenetic hypotheses]. I therefore used an unpublished phylogeny assembled by C. Parkinson from four different mitochondrial genes (12S, 16S, ND4, and cytochrome B). I used Parkinson’s maximum-likelihood topology, which is well supported and includes representatives of all genera of New World pitvipers (Agkistrodon, Atropoides, Bothriechis, Bothriopsis, Bothrops, Cerrophidion, Crotalus, Lachesis, Ophryacus, Porthidium, and Sistrurus; N =36 species), plus a number of Old World outgroup species (N = 25). This analysis indicates that some genera (sensu lato) are paraphyletic (e.g., Bothrops, Porthidium, Sistrurus), and so my autocorrelation analyses use phylogenetic groupings rather than present taxonomy. For elapid snakes, log10-transformed range sizes were mapped onto a tree using the range sizes of all 63 species. Ranges were mapped on the viperid phylogeny in two ways: a) using the species values for the 36 species on Parkinson’s topology; and b) using mean range size values for each genus mapped on appropriate genus-level nodes (correcting for paraphyletic genera when necessary by grouping species with evolutionary relatives rather than traditional congeners). The latter approach utilized the entire dataset of viperid snakes rather than the subset represented by Parkinson’s topology. As discussed above, body size is often positively correlated with geographic range area. Autocorrelation results indicating that phylogeny accounts for a large proportion of variation in range size therefore could be misleading if body size is also strongly affected by phylogeny. Therefore, I repeated autocorrelation analyses using the residuals of a linear regression of range area on maximal total length (both log10-transformed). 109
This method removes the effect of body size and unmasks any remaining phylogenetically-induced interspecific trends in range size. Results are reported as mean plus or minus one standard error. Significance was set a priori at 0.05. Body size and geographic range area were log10-transformed in analyses except where noted. All statistical analyses were performed with Systat v. 9.
Results My analyses used 144 species of New World species of elapid and viperid snakes currently described (63 of 69 coral snakes, 81 of 97 pitvipers). Species richness of elapid snakes peaked in equatorial regions, but viperids exhibited a peak from 10 to 20° N (Fig. 1). Species – range size distributions were strongly right-skewed for both families on arithmetic scales (Elapidae: skewness =2.84, Viperidae skewness =2.24), and a large proportion of species exhibited geographic range areas B500 000 km2 (Fig. 2). Mean range area was 751 719 9171 140 km2 for elapids, and 981 569 9 186 872 km2 for viperids. Logarithmic transformation
Fig. 2. Distribution of geographic range sizes for New World species of the families Elapidae and Viperidae. Solid bars indicate elapid species, hollow bars indicate viperid species. A) Range areas displayed on arithmetic scale. B) Range areas displayed on log10 scale.
(base 10) failed to normalize these distributions, but eliminated most skewness (Wilks-Shapiro test of normality, Elapidae: p =0.037, skewness = − 0.051, Viperidae: p = 0.002, skewness = −0.095; Fig. 2). No interfamilial difference in mean range area was detected (t = 0.06; p = 0.95). Mean maximal total body length was 871 9282 mm for elapids and 1178 9662 mm for viperids. The relationship between body length and geographic range area (on log10 axes) was significantly positive for both families (Fig. 3), with body length explaining roughly
Fig. 1. Latitudinal patterns of species richness among New World snakes of the families Elapidae and Viperidae. Solid bars indicate elapid species, hollow bars indicate viperid species. Negative numbers on x-axis indicate latitudinal degrees south of the equator, positive numbers indicate degrees north of the equator. A) Histograms produced using only median latitude of each species’ latitudinal range, i.e. each species is represented only once. B) Histograms produced using species’ total latitudinal range, i.e. species may be represented on multiple bars, and a given bar indicates the total number of species in five degrees of latitude.
110
Fig. 3. Relationships between body size and geographic range area among New World species of the families Elapidae and Viperidae. Solid diamonds and solid regression line indicate elapids, hollow squares and dashed regression line indicate viperids. Both axes log10 transformed. Total length measured in mm, geographic range area in km2. ECOGRAPHY 26:1 (2003)
Fig. 4. Effects of latitude on geographic range area among New World species of the families Elapidae and Viperidae. Elapids indicated by solid diamonds, viperids by hollow squares. Negative numbers on x-axis indicate latitudinal degrees south of the equator, positive numbers indicate degrees north of the equator. For each species, median latitude is the midpoint of the total latitudinal range. Geographic range areas log10 transformed.
the same amount of variation among viperid (r2 = 0.23, p= 0.0001) and elapid snakes (r2 =0.18, p =0.0001). The slopes of regression equations were not significantly different between families (ANCOVA: interaction term p =0.643). Median latitude of geographic range was not related to body size (Elapidae: r2 = 0.015, p= 0.333; Viperidae: r2 B0.001, p =0.968). Latitudinal extent of geographic range was positively correlated with range area in both families (Elapidae: r2 =0.59, p =0.0001; Viperidae: r2 =0.68, p = 0.0001), and the slopes of these relationships were not significantly different among families (ANCOVA: interaction term p =0.228). These variables are therefore roughly equivalent as measures of range size for these taxa. Regardless of whether range area or latitudinal extent of range was used as the dependent variable, median latitude was not a significant predictor of range size for elapids (r2range area =0.02, p = 0.31; r2lat. extent B 0.001; p= 0.96; Fig. 4). Among vipers, there was a trend towards a positive relationship between range area and median range latitude (r2range area =0.04, p =0.07; r2lat. extent B0.001, p =0.97), but the smallest geographic ranges were generally observed between 10°N and 20°N rather than at the equator (Fig. 4). Continental width (at the median geographic range latitude of each species) was positively related to geographic range area for both families (Elapidae: r2 =0.23, p B0.001; Viperidae: r2 =0.25, p B0.001; Fig. 5). The combination of median range latitude, body size, and available continental width explained a total of 43% of variation in geographic range area (40% for elapids, 48% for viperids; Table 1). Most of the original data sets supporting Rapoport’s rule were composed of taxa from the northern hemisphere, especially North American endothermic vertebrates (Gaston et al. 1998a). Studies of the geographic ranges of species from areas nearer the equator and ECOGRAPHY 26:1 (2003)
Fig. 5. Effects of available land area on geographic range area among New World species of the families Elapidae and Viperidae. Elapids indicated by solid diamonds and solid regression line, viperids by hollow squares and dashed regression line. Continental width (in km) measured at the latitudinal midpoint of each species’ geographic range. Geographic range area log10 transformed.
from Australia have provided little evidence in support of Rapoport’s rule (e.g., Rohde et al. 1993, Smith et al. 1994, Blackburn and Gaston 1996, Hughes et al. 1996). As contrary evidence accumulates, authors have recently begun to refer to the increase of species’ range size with latitude as the Rapoport ‘‘effect’’, only valid in north temperate zones (Gaston 1999). I therefore analyzed the effect of median latitude on geographic range area for all species with median latitudes above 15° north (N =15 elapids and 38 viperids). Both elapids and viperids showed a strong positive relationship (Elapidae: r2 = 0.41, p =0.011; Viperidae: r2 = 0.76, p B 0.001). These trends remained significant after removal of the effect of continental width (by repeating Table 1. Multiple regression models of the effects of median latitude of geographic range, continental width at median latitude, and body size on geographic range area among New World venomous snakes (Families Elapidae and Viperidae). Overall regression statistics for all snakes together: F =34.65, pB0.001, r2 =0.426; regression statistics for elapids F = 13.28, pB0.001, r2 =0.403; regression statistics for viperids: F= 23.46, pB0.001, r2 =0.478. All variables log10 transformed prior to analysis. Variable All snakes together Continental width Median latitude of range Body size Elapid snakes only Continental width Median latitude of range Body size Viperid snakes only Continental width Median latitude of range Body size
Tolerance
t
p
0.900 0.837
5.458 3.179
B0.0001 0.002
0.956
5.380
B0.0001
0.677 0.668
4.342 0.950
B0.0001 0.346
0.968
3.785
B0.0001
0.900 0.951
5.458 3.593
B0.0001 0.001
0.944
4.416
B0.0001
111
Table 2. Results of network autocorrelation analyses of phylogenetic contributions to geographic range area among New World venomous snakes (families Elapidae and Viperidae). See text for detailed methodology. Matrix Viperidae Species Species Genera Elapidae Species Species
Vector
r
p
R2
Range area Residuals Range area
−0.26 −0.13 −0.10
\0.5 \0.5 \0.5
0.06 0.01 0.01
Range area Residuals
−0.10 0.31
\0.5 \0.5
0.01 0.13
the analysis with dependent variable defined as the residuals of a regression of geographic range area on continental width), but the regression models had much lower explanatory power (Elapidae: r2 =0.226, p = 0.042; Viperidae: r2 =0.163, p =0.007). Phylogeny accounted for a small amount of interspecific or intergeneric variation in geographic range area, regardless of the type of autocorrelation analysis performed. Most R2 values were B 0.05, and none of the results approached statistical significance (Table 2). Removal of the effect of body size using residuals did not appreciably alter autocorrelation results (Table 2).
Discussion Patterns of species richness As is true for many other taxa, species richness of both elapid and viperid snakes generally peaked in tropical or subtropical regions. A recent explanation for this phenomenon (Rohde 1998) revolves around two primary mechanisms. First, global gradients in species richness may be primarily determined by ‘‘effective evolutionary time’’, comprised of evolutionary speed and geological time during which an ecosystem has existed. This explanation assumes that tropical environments have remained stable over long periods of geologic time, and that evolutionary tempo increases with environmental temperature. Second, this stable environment and associated rapid evolutionary rates allow the tropics to act as a generator of new species that then invade more temperate zones. These invasions require local adaptation to cooler temperatures, which should be an increasingly difficult prospect as populations move away from the equator. Few lineages are expected to successfully adapt to the temperate zone, thus offering a simple explanation for diversity gradients. Viperid snakes apparently invaded the New World during an interglacial period via the Bering landbridge, subsequently dispersing south to invade the tropics [Brattstrom 1964, Gloyd and Conant 1990; New World elapids are also monophyletic and possibly arose from a similar landbridge invasion (Slowinski 1995)]. Tem112
perate genera (e.g., Agkistrodon, Crotalus) may have given rise to tropical lineages (Brattstrom 1964, Van Devender and Conant 1990), and Parkinson’s (unpubl.) phylogeny is consistent with this hypothesis. These ancestral viperids were thus presumably well adapted to moderately cool temperatures. Because these snakes invaded the tropics rather than vice versa, the second part of Rohde’s (1998) hypothesis is not compelling for snakes: There is little evidence that the New World tropics served as a generator of new species which subsequently invaded temperate regions. However, his (1998) concept of ‘‘effective evolutionary time’’ may have some validity for snakes. Temperate zone winters largely preclude snake activity, resulting in abbreviated activity seasons relative to tropical species. Short activity seasons may limit prey capture opportunities, thus affecting age at sexual maturation and ability to acquire energy reserves for reproductive purposes. Temperate zone snakes thus commonly exhibit low reproductive frequencies, with females producing young at biennial to quadrennial intervals (Brown 1991). Delayed maturation and infrequent reproduction increase the time between generations, and may limit the potential tempo of evolution in a population. If so, this could decrease ‘‘effective evolutionary time’’ for temperate populations of ectotherms, regardless of the geologic age of an area. A closer examination of latitudinal patterns of species richness indicated that viperid snake species richness peaks from 10 to 20°N, and elapid snakes similarly exhibit much higher diversity in this region than in equivalent southerly latitudes. These latitudes correspond to Central America and southern Mexico, which have complicated biogeographic histories. Orogenic episodes have repeatedly fragmented this region, offering many opportunities for allopatric speciation (Savage 1966). Thus historical biogeography may have a larger effect on regional species richness than that exerted by biotic factors, especially among viperid snakes. Ancestral lineages in this region may have been longitudinally constrained in potential range size by oceans, but in spite of this constraint have undergone extensive diversification (Crother et al. 1993). This represents a caveat to a prevalent assumption that species with large geographic ranges are more likely to speciate (e.g., Rosenzweig 1995, Tokeshi 1996).
Species–range size distributions Species – range size distributions are usually strongly right-skewed, taking the form of a ‘‘hollow curve’’, in which the vast majority of species appear to have small geographic ranges (Willis 1922, Brown et al. 1996). However, these distributions often approach normality after log-transformation, such that extremely small ranges are as rare as extremely large ones (Gaston 1996). An increasing number of studies document sysECOGRAPHY 26:1 (2003)
tematic departures from log-normality; significant left skews in multi-species body size distributions may emerge as a general pattern, although this phenomenon is poorly understood (Blackburn and Gaston 1996, Gaston 1998, Gaston et al. 1998b). My results indicated that species – range size distributions for both snake families are strongly right-skewed at arithmetic scales, generally matching the distributions previously posited. If one considers the possible addition of missing species, which are likely to have small geographic ranges, then it seems likely that the skew is even stronger. However, log-transformation failed to normalize these distributions: Elapids displayed a skew towards small range size classes, while viperids displayed a multimodal distribution of range sizes. The Neotropical rattlesnake (Crotalus durissus) has the largest geographic range area of snakes in this analysis ( 7.5×107 km2), with a latitudinal range from 36°S to 25°N. However, populations are discontinuous within the total range (Campbell and Lamar 1989), and it is likely that adoption of evolutionary or phylogenetic species concepts would split this taxon into several discrete species with smaller range sizes. Similar taxonomic revisions would seem to be most likely for those species with very large ranges, but may not yet have occurred because of the logistics and expense of adequately sampling populations over large tropical areas. Taxonomic splitting of this sort may be expected to reduce the right skew in range size distributions on arithmetic scales, but conversely may increase the left skew of these distributions on logarithmic scales. If so, a stronger trend for range size distributions to comprise more species with very small ranges than species with very large ranges would be expected.
Body size – range size relationships Positive relationships between body size and range area have been observed for many taxa (Gaston and Blackburn 1996a, b). The body size/range size relationship may best be understood by considering the overall pattern; the upper bounds of range sizes are constrained by available continental area, while lower bounds may be set by minimum viable population size or other factors. Residual variation should be decreased for large-bodied species because these species are extremely unlikely to have small ranges, resulting in a roughly triangular relationship which is truncated at large geographic ranges (Brown 1995, Brown et al. 1996). Various mechanisms may explain positive relationships between geographic range size and body size. Brown (1995) states that large animals usually require large activity ranges, in order to acquire sufficient resources. Large-bodied species may thus require larger total areas to maintain minimum viable populations, ECOGRAPHY 26:1 (2003)
resulting in large overall geographic ranges. Small-bodied species, on the other hand, are not subject to the same constraints and so may occupy either small or large geographic ranges. This hypothesis helps explain the triangular nature of bivariate body size – range size plots. However, this explanation was formulated using data from endotherms, animals with many times the mass-specific energy requirements of ectotherms. Contrary to the assumptions of Brown’s hypothesis, even the largest reptiles can attain locally high population densities (Auffenberg 1981, Madsen and Shine 1996), due to their low individual energy requirements (Pough 1983). I offer an alternative hypothesis for snakes, based on the observation that both clutch size and offspring number tend to increase with maternal body size among snakes (both within and between species; Dunham et al. 1988, Shine 1996). Large clutch sizes may allow populations of large-bodied species to attain relatively high local densities, reducing the risk of local extinctions during periods of stochastic environmental conditions or temporary decreases in a resource base. This effect could be augmented by higher survival rates among relatively large offspring. Higher population densities resultant from these mechanisms could, over the long term, provide many ‘‘source’’ populations able to colonize new areas, and eventually result in large species geographic ranges. A positive relationship between geographic range and body size thus would not entail the invocation of individual activity area size as the prime phenomenon affecting geographic range of a species. A second explanation for positive relationships between body size and range size involves the tendency for both body size and geographic range size of species within a species assemblage to increase with latitude [Bergmann’s rule and Rapoport’s rule, respectively (see below)]; if these trends result from the same mechanism, then they could produce a positive body size/ range size relationship. However, my results fail to support either of these ‘‘rules’’, thus invalidating these explanations. Third, the body size – range size relationship could result from differential tolerance to environmental variability. Larger animals are more capable of maintaining homeostasis over a range of climatic variables, and thus may be able to take advantage of larger geographic ranges that encompass greater variability (Root 1991). Similar to those arguments based on area needed for minimum viable populations, the environmental variability hypothesis was based on data from endotherms and thus may not be applicable to ectotherms. Maintenance of homeostasis is not necessarily a goal for these animals, which may drastically manipulate body temperatures (Peterson et al. 1993) and metabolic rates (Secor and Diamond 1998) in response to ecological or physiological stimuli. Many ectotherms also escape from cold temperatures via hibernation or torpor, such 113
that a large geographic range can be attained without exposure to debilitating seasonal temperatures. The hypothesis that increased body size allows expansion of range size via increased tolerance to environmental extremes is a logical extension of Bergmann’s rule, which is not supported for venomous snakes; most of the very large elapids and vipers are found in the tropics. Alternatively, relationships between body size and range size may be due to differences in realized versus potential geographic range size among snakes of different body sizes (Blackburn and Gaston 1996, Gaston and Blackburn 1996a). Large-bodied species may disperse and establish more readily (Hugueny 1990) and may be members of evolutionarily older lineages than lineages of small species (according to Cope’s rule, which states that body sizes in a lineage increase over evolutionary time). These ‘‘old’’ large-bodied species may thus occupy a greater proportion of their potential geographic range than smaller species that are still expanding. While considerable variation exists in the movement distances and activity ranges among snakes (MacCartney et al. 1988), body size appears to be a poor predictor of dispersal ability. Fairly small snakes may move great distances (Secor 1994), while the largest pitvipers appear to have very small activity ranges (Greene and Santana 1983). Evaluation of Cope’s rule is problematic given current phylogenies, but large-bodied snakes do not appear to be members of relatively ‘‘older’’ lineages. Venomous snake assemblages thus offer little support for the notion that organisms with good dispersal abilities and/or longer evolutionary ‘‘lifetimes’’ are likely to attain high range sizes. The plot of geographic range size as a function of body size generally conformed to the predicted triangular bivariate plot (discussed above; Brown 1995, Brown et al. 1996), in which small-bodied species may have ranges of various sizes, but large-bodied species occupy only large ranges. The left side of this relationship may be set by structural or functional limits on minimum body size, while the upper limit is set by the amount of land available. The largest geographic ranges of snakes, however, did not approach the total available land area on both continents, as is apparently the case for birds (Root 1988, Blackburn and Gaston 1996). The hypotenuse of body size – range size plots is predicted to be set by minimum viable population considerations and probabilities of extinction (Brown 1995, and see above). As if in defiance of the predicted shape, however, a few large-bodied vipers have small range sizes, and two species of the genus Lachesis (the bushmasters) offer the most obvious departures. This genus has recently been revised (Zamudio and Greene 1997), and includes the largest known members of the Viperidae. One species (L. melanocephala) has a very small range in a portion of Central America, which has 114
led to concern about its conservation status as its lowland rainforest habitat is cleared (Zamudio and Greene 1997). The threatened status and small range of this species would appear to lend credence to the hypothesis that large-bodied species are more likely to become extinct when occupying small ranges, due to minimum viable population considerations (Brown 1995). However, this constraint is unlikely to be of major importance for snakes due to their low mass-specific energetic requirements (see above). Indeed, the results of Zamudio and Greene (1997) indicate that L. melanocephala did not arise from a recent invasion of Central America (diverging from L. stenophrys between 11 and 4 myr), suggesting it has maintained stable populations despite its small geographic range. In the absence of anthropogenic habitat degradation, therefore, it seems unlikely that this species would be of particular conservation concern. Most explanations for a positive body size – range size relationship attempt to explain why large-bodied species can attain large ranges but rarely occupy small ranges. However, the tendency for some small species to have very small ranges is of equal interest, and may affect the relationship to a greater degree than do large-bodied species. I have argued above that the most often-cited hypotheses for this relationship may not be applicable to snakes, and that ectotherms are freed of many physiological limitations on range size. This freedom may allow a majority of snake species to assume large geographic ranges. What, then, may be keeping some small-bodied species from increasing their geographic range? The answer may revolve around ecological specialization and consideration of certain habitats as ‘‘islands’’ surrounded by dispersal barriers. Species inhabiting montane areas seem likely targets for investigation of this idea. I therefore calculated the midpoints of elevational ranges for all snakes in the dataset, and for each family I compared mean elevational midpoints of the 15 species with the smallest geographic range areas against the mean elevational midpoints of all other confamilial species. Species with small geographic range areas were consistently found at higher elevations (vipers: 1630 vs 1047 m, t = 2.64, p = 0.008; elapids: 958 vs 656 m, t = 1.61, p = 0.05). At least among vipers, this trend does not appear to be an artifact of phylogeny, as these 15 species comprised seven genera. These snakes were also of smaller body size (vipers: 971 vs 1220 mm, t = −1.78, p =0.04; elapids: 708 vs 918 mm, t= − 2.61, p =0.008). Unlike lowland areas, Neotropical montane habitats experience great temperature fluctuations, on both daily and annual cycles. Mean annual temperatures are B15°C at elevations \2000 m (Campbell and Solorzano 1993). This results in a lower annual total of degree-days in these environments, with obvious repercussions for the activity seasons of ectotherms. Annual energy budgets must be allocated between reproduction ECOGRAPHY 26:1 (2003)
and growth/maintenance, and small-bodied species should require less energy per year to meet these goals. In cool climes, the energy necessary for large-bodied snakes to attain a minimum threshold for successful reproduction may result in very infrequent breeding (Brown 1991). Small-bodied montane snakes would be better able to attain a minimum energetic threshold, resulting in more frequent reproduction and potentially higher lifetime fitness. Nonetheless, adaptations to cool temperatures or specific habitats may preclude successful invasion of lowlands by these snakes, thereby checking geographic range expansion. Occupation of small ranges thus may not simply be due to dispersal or other macroecological limitations of small body sizes, but because montane species are effectively castaways in their highland habitats.
genetic hypothesis. Phylogenetic contributions to geographic range were uniformly of low magnitude, such that closely related species are not likely to have similar range areas. This is in contrast to results obtained for some fossil and extant marine invertebrates, among which related species have similar geographic range areas (Jablonski 1987). The minor effect of phylogeny among elapid and viperid snakes in turn implies that adaptations to large range sizes are not likely to arise early in a lineage and persist, but are instead functional at a population level, perhaps responding dynamically to changing environmental conditions. Autocorrelation results could also imply that ranges are not static, and that phylogenetic effects are not to be seen until species reach their potential range size. This seems unlikely, given that range sizes of the ecologically better-known species are not known to have increased in historical times.
Effects of available land area Continental width explained a large amount of interspecific variation in range area. Rosenzweig (1992, 1995) stated that latitudinal decreases in species richness may occur simply because there is less land area in the temperate zones of the world, but this hypothesis was criticized as having little empirical support (Rohde 1998, Chown and Gaston 2000). I propose that Rosenzweig’s (1995) hypothesis could be applicable to snakes, largely due to oceanic boundaries to dispersal. Consider a new species that arises in a locality (due to vicariance or other isolating mechanisms), and then gradually expands its geographic range. This expansion should continue until populations reach the limits of the species’ potential range, which may be set by climate, competitors, prey availability, or other factors. Then consider a second species, expanding its range at the same rate, but with ocean blocking it from expanding in two directions (as would happen for species arising in Central America, for example). Absent other limiting factors, range expansion will proceed apace along unaffected dispersal fronts, but the total geographic range will increase much more slowly. Compounding this dilemma is the likelihood that related species in the same region are subject to the same limits on range expansion, perhaps resulting in higher interspecific competition for available resources along dispersal fronts. This will further decrease the rate of range size expansion and may eventually culminate in small range sizes.
Effects of phylogeny Interpretation of phylogenetic autocorrelation results must be tentative, given the unresolved nature of cladograms used in my analyses. This is especially true of New World elapids, which lack a well-supported phyloECOGRAPHY 26:1 (2003)
Latitudinal patterns in range size and Rapoport’s rule The total latitudinal range of all viperid snakes exceeds that of elapid snakes; a number of pitviper species in the northern hemisphere have median range latitudes up to ten degrees north of the most northerly coral snake. This could be due to a number of historical or ecological factors. Pitvipers are heavy bodied and thus have lower surface:volume ratios than do coral snakes of similar length, possibly conferring thermoregulatory advantages in northern climates by reducing rates of temperature flux (Peterson et al. 1993). Alternatively, steep declines in overall snake diversity with increasing latitude implies that prey items are generally rare or absent for ophidiophagous coral snakes in these regions, while not affecting pitvipers which largely consume mammals. On the whole, my data do not support Rapoport’s rule. Indeed, the relationship between range size and latitude has probably lost its status as an ecological rule, given the volume of recent criticism leveled against it (Rohde et al. 1993, Gaston et al. 1998a, Gaston 1999, Kerr 1999, Rohde 1999). Among New World birds, the smallest range sizes were found not at the equator, but at ca 17°N (Blackburn and Gaston 1996). Snake range sizes were also at a minimum in this region. Blackburn and Gaston (1996) postulated that small range size among birds were due to high faunal turnover in this region, as vegetation in this latitudinal zone undergoes a sharp shift from tropical to temperate. My analysis of body size, elevation, and range size (above) offers an alternative hypothesis: small geographic ranges of snakes in Central America result from the ultimate geologic (especially orogenic) history of the region, rather than proximate vegetative changes. 115
Although my analyses help lay to rest the generality of Rapoport’s rule for interspecific assemblages, they provide support for a local Rapoport ‘‘effect’’, restricted to north temperate zones (Rohde 1996, Gaston et al. 1998a). Available continental area accounts for only a portion of this phenomenon. Published work has placed this north temperate Rapoport effect between 40° and 50°N (Rohde 1996, Gaston et al. 1998a), beyond the latitudinal range of virtually all New World venomous snakes. Among snakes, I found the pattern to be strong from 15 to 40°N. Another reptilian taxon (turtles) exhibits a similar pattern, with range size increases beginning at 25°–30°N (Hecnar 1999). An absence of mechanisms for coping with extreme cold (e.g., long-distance migration, maintenance of homeostasis during hibernation, high metabolic rate, etc) among reptiles may explain the tendency for Rapoport effects to occur at relatively lower latitudes as compared to endotherms. Gaston et al. (1998a) proposed that geographic range sizes may be limited more by relative sizes of biogeographic provinces than by latitude. Areas of these provinces generally increase with latitude in the Nearctic, offering an explanation for the Rapoport effect. This seems a logical explanation for snakes, as the proportion of mountainous land area (with high environmental and biotic variability) steadily decreases north of Central America. Indeed, among vipers the largest geographic ranges in north temperate zones are similar in area to ranges in the Amazon basin (another extremely large area with relatively uniform environmental conditions; Fig. 4). Thus, if a Nearctic snake species can adapt to the environmental variability in a given locale, the climatic similarity within large Nearctic biogeographical provinces may allow achievement of a large range size. Acknowledgements – This manuscript was improved by comments from K. Ashton, S. Dobson, M. Eubanks, J. W. Gibbons, and C. Guyer. Long-term research opportunities and manuscript preparation were aided by Contract DE-AC0976SROO-819 between the U.S. Dept of Energy and the Univ. of Georgia’s Savannah River Ecology Laboratory and with Financial Assistance Award Number DE-FC09-96SR18546 from U.S. Dept of Energy to the Univ. of Georgia Research Foundation.
References Allee, W. C. et al. 1949. Principles of animal ecology. – Saunders. Ashton, K. G. 2001. Are ecological and evolutionary rules being dismissed prematurely? – Div. Distrib. 7: 289 – 296. Auffenberg, W. 1981. The behavioral ecology of the Komodo monitor. – Univ. Florida Press. Blackburn, T. M. and Gaston, K. J. 1996. Spatial patterns in the geographic ranges of bird species in the New World. – Philos. Trans. R. Soc. Lond. B 351: 897 –912. Blackburn, T. M. and Gaston, K. J. 1997. The relationship between geographic area and the latitudinal gradient in species richness in New World birds. – Evol. Ecol. 11: 195 – 204.
116
Bonfim, F. A., Diniz-Filho, J. A. F. and Bastos, R. P. 1998. Spatial patterns and the macroecology of South American viperid snakes. – Rev. Brasieira Biol. 58: 97 – 103. Brattstrom, B. H. 1964. Evolution of the pit vipers. – Trans. San Diego Soc. Nat. Hist. 13: 185 – 268. Brown, J. H. 1995. Macroecology. – Univ. Chicago Press. Brown, J. H., Stevens, G. C. and Kaufman, D. M. 1996. The geographic range: size shape, boundaries, and internal structure. – Annu. Rev. Ecol. Syst. 27: 597 – 623. Brown, W. S. 1991. Female reproductive ecology in a northern population of the timber rattlesnake, Crotalus horridus. – Herpetologica 47: 101 – 115. Campbell, J. A. and Lamar, W. W. 1989. The venomous reptiles of Latin America. – Comstock/Cornell. Campbell, J. A. and Solorzano, A. 1993. The distribution, variation, and natural history of the middle American montane pitviper, Porthidium godmani. – In: Campbell, J. A. and Brodie Jr, E. D. (eds), Biology of the pitvipers. Selva Press, pp. 223 – 250. Cheverud, J. M., Dow, M. M. and Leutenegger, W. 1985. The quantitative assessment of phylogenetic constraints in comparative analyses: sexual dimorphism in body weight among primates. – Evolution 39: 1335 – 1351. Chown, S. L. and Gaston, K. J. 2000. Areas, cradles, and museums: the latitudinal gradient in species richness. – Trends Ecol. Evol. 15: 311 – 315. Colwell, R. K. and Hurtt, G. C. 1994. Nonbiological gradients in species richness and a spurious Rapoport effect. – Am. Nat. 144: 570 – 595. Conant, R. and Collins, J. T. 1998. A field guide to reptiles and amphibians of eastern and central North America. – Houghton Mifflin. Crother, B. I., Campbell, J. A. and Hillis, D. M. 1993. Phylogeny and historical biogeography of the palmpitvipers, genus Bothriechis: biochemical and morphological evidence. – In: Campbell, J. A. and Brodie Jr, E. D. (eds), Biology of the pitvipers. Selva Press, pp. 1 – 20. Dobzhansky, T. 1950. Evolution in the tropics. – Am. Sci. 38: 209 – 221. Dunham, A. E. et al. 1988. Life history patterns in squamate reptiles. – In: Gans, C. and Huey, R. B. (eds), Biology of the reptilia Vol. 16 (Ecology B). Allan R. Liss, pp. 441 – 522. Ernst, C. H. 1992. Venomous reptiles of North America. – Smithsonian Inst. Press. Gaston, K. J. 1991. How large is a species’ geographic range? – Oikos 61: 434 – 438. Gaston, K. J. 1996. Species – range size distributions: patterns, mechanism and implications. – Trends Ecol. Evol. Biol. 11: 197 – 201. Gaston, K. J. 1998. Species – range size distributions: products of speciation, extinction and transformation. – Philos. Trans. R. Soc. Lond. B 353: 219 – 230. Gaston, K. J. 1999. Why Rapoport’s rule does not generalize. – Oikos 84: 309 – 312. Gaston, K. J. and Blackburn, T. M. 1996a. Range size – body size relationships: evidence of scale dependence. – Oikos 75: 479 – 485. Gaston, K. J. and Blackburn, T. M. 1996b. Conservation implications of geographic range size – body size relationships. – Conserv. Biol. 10: 638 – 646. Gaston, K. J. and Blackburn, T. M. 1999. A critique for macroecology. – Oikos 84: 353 – 368. Gaston, K. J., Blackburn, T. M. and Spicer, J. I. 1998a. Rapoport’s rule: time for an epitaph? – Trends Ecol. Evol. 13: 70 – 74. Gaston, K. J. et al. 1998b. Species – range size distributions in Britain. – Ecography 21: 361 – 370. Gloyd, H. K. and Conant, R. 1990. Snakes of the Agkistrodon complex: a monographic review. – Soc. for the Study of Amphibians and Reptiles, Contrib. to Herpetology 6. Greene, H. W. 1997. Snakes: the evolution of mystery in nature. – Univ. of California Press. ECOGRAPHY 26:1 (2003)
Greene, H. W. and Santana, M. A. 1983. Field studies of hunting behavior in bushmasters. – Am. Zool. 23: 897. Gutberlet Jr, R. L. 1998. The phylogenetic position of the Mexican black-tailed pitviper (Squamata: Viperidae: Crotalinae). – Herpetologica 54: 184 –206. Hecnar, S. J. 1999. Patterns of turtle species’ geographic range size and a test of Rapoport’s rule. – Ecography 22: 436 – 446. Hughes, L., Cawsey, E. M. and Westoby, M. 1996. Geographic and climatic range sizes of Australian eucalypts and a test of Rapoport’s rule. – Global Ecol. Biogeogr. Lett. 5: 128 –142. Hugueny, B. 1990. Geographic range of west African freshwater fishes: role of biological characteristics and stochastic processes. – Acta Oecol. 11: 351 –375. Jablonski, D. 1987. Heritability at the species level: analysis of geographic ranges of cretaceous mollusks. – Science 238: 360 – 363. Janzen, D. H. 1967. Why mountain passes are higher in the tropics. – Am. Nat. 101: 233 –249. Jorge da Silva, N. J. and Sites Jr, J. W. 1999. Revision of the Micrurus frontalis complex (Serpentes: Elapidae). – Herpetol. Monogr. 13: 142 –194. Kerr, J. T. 1999. Weak links: ‘Rapoport’s rule’ and large-scale species richness patterns. – Global Ecol. Biogeogr. 8: 47 – 54. Kraus, F., Mink, D. G. and Brown, W. M. 1996. Crotaline intergeneric relationships based on mitochondrial DNA sequence data. – Copeia 1996: 763 – 773. Lyons, S. K. and Willig, M. R. 1997. Latitudinal patterns of range size: methodological concerns and empirical evaluations for New World bats and marsupials. – Oikos 79: 568 – 580. MacCartney, J. M., Gregory, P. T. and Larsen, K. W. 1988. A tabular survey of data on movements and home ranges of snakes. – J. Herpetol. 22: 61 –73. Madsen, T. and Shine, R. 1996. Seasonal migration of predators and prey – a study of pythons and rats in tropical Australia. – Ecology 77: 140 –156. Miles, D. B. and Dunham, A. E. 1992. Comparative analyses of phylogenetic effects in the life-history patterns of iguanid reptiles. – Am. Nat. 139: 848 –869. Parkinson, C. L. 1999. Molecular systematics and biogeographical history of pitvipers as determined by mitochondrial ribosomal DNA sequences. – Copeia 1999: 576 – 586. Parkinson, C. L., Zamudio, K. R. and Greene, H. W. 2000. Phylogeography of the pitviper clade Agkistrodon: historical ecology, species status, and conservation of cantils. – Mol. Ecol. 9: 411 –420. Peterson, C. R., Gibson, A. R. and Dorcas, M. E. 1993. Snake thermal ecology: the causes and consequences of body-temperature variation. – In: Seigel, R. A. and Collins, J. T. (eds), Snakes: ecology and behavior. McGraw-Hill, pp. 241 – 314. Pianka, E. R. 1966. Latitudinal gradients in species diversity: a review of concepts. – Am. Nat. 100: 33 –46. Pough, F. H. 1983. Amphibians and reptiles as low-energy systems. – In: Aspey, W. P. and Lustick, S. I. (eds), Behavioral energetics: the cost of survival in vertebrates. Ohio State Univ. Press, pp. 141 –188. Rohde, K. 1996. Rapoport’s rule is a local phenomenon and cannot explain latitudinal gradients in species diversity. – Biodiv. Lett. 3: 10 –13. Rohde, K. 1998. Latitudinal gradients in species diversity. Area matters, but how much? – Oikos 92: 184 – 190. Rohde, K. 1999. Latitudinal gradients in species diversity and Rapoport’s rule revisited: a review of recent work and what can parasites teach us about the causes of the gradients? – Ecography 22: 593 –613.
ECOGRAPHY 26:1 (2003)
Rohde, K., Heap, M. and Heap, D. 1993. Rapoports’ rule does not apply to marine teleosts and cannot explain latitudinal gradients in species richness. – Am. Nat. 142: 1 – 16. Root, T. L. 1988. Energy constraints on avian distributions and abundances. – Ecology 69: 330 – 339. Root, T. L. 1991. Positive correlation between range size and body size: a possible mechanism. – Acta XX Congr. Int. Ornithol., pp. 817 – 825. Rosenzweig, M. L. 1992. Species diversity gradients: we know more and less than we thought. – J. Mammal. 73: 715 – 730. Rosenzweig, M. L. 1995. Species diversity in space and time. – Cambridge Univ. Press. Roze, J. A. 1996. Coral snakes of the Americas: biology, identification, and venoms. – Krieger Press. Savage, J. M. 1966. The origins and history of the Central American herpetofauna. – Copeia 1966: 719 – 766. Secor, S. M. 1994. Ecological significance of movements and activity range for the sidewinder, Crotalus cerastes. – Copeia 1994: 631 – 645. Secor, S. M. and Diamond, J. 1998. A vertebrate model of extreme physiological regulation. – Nature 395: 659 – 662. Shine, R. 1996. Life-history evolution in Australian snakes: a path analysis. – Oecologia 107: 484 – 489. Slowinski, J. B. 1995. A phylogenetic analysis of the New World coral snakes (Elapidae: Leptomicrurus, Micruroides, and Micrurus) based on allozymic and morphological characters. – J. Herpetol. 29: 325 – 338. Smith, F. D. M., May, R. M. and Harvey, P. H. 1994. Geographic ranges of Australian mammals. – J. Anim. Ecol. 63: 441 – 450. Stebbins, R. C. 1985. A field guide to western reptiles and amphibians. – Houghton Mifflin. Stevens, G. C. 1989. The latitudinal gradient in geographical range: how so many species coexist in the tropics. – Am. Nat. 133: 240 – 256. Taylor, C. M. and Gotelli, N. J. 1994. The macroecology of Cyprinella: correlates of phylogeny, body size, and geographical range. – Am. Nat. 144: 549 – 569. Tokeshi, M. 1996. Dynamics of distribution in animal communities: theory and analysis. – Res. Popul. Ecol. 34: 249 – 273. Van Devender, T. R. and Conant, R. 1990. Pleistocene forests and copperheads in the eastern United States, and the historical biogeography of New World Agkistrodon. – In: Gloyd, H. K. and Conant, R. (eds), Snakes of the Agkistrodon complex: a monographic review. Soc. for the Study of Amphibians and Reptiles Contrib. to Herpetology 6, pp. 601 – 614. Vidal, N. et al. 1997. Molecular systematics of pitvipers: paraphyly of the Bothrops complex. – C. R. Acad. Sci. Paris, Sci. de la Vie 320: 95 – 101. Werman, S. J. 1993. Phylogenetic relationships of Central and South American pitvipers of the genus Bothrops (sensu lato): cladistic analyses of biochemical and anatomical characters. – In: Campbell, J. A. and Brodie Jr, E. D. (eds), Biology of the pitvipers. Selva Press, pp. 21 – 40. Werman, S. J., Crother, B. I. and White, M. E. 1999. Phylogeny of some Middle American pitvipers based on a cladistic analysis of mitochondrial 12S and 16S DNA sequence information. – Contemp. Herpetol. 1999: 3 (online journal). Willis, J. C. 1922. Age and area. – Cambridge Univ. Press. Wuster, W., Golay, P. and Warrell, D. A. 1997. Synopsis of recent developments in venomous snake systematics. – Toxicon 35: 319 – 340. Zamudio, K. R. and Greene, H. W. 1997. Phylogeography of the bushmaster (Lachesis muta: Viperidae): implications for neotropical biogeography, systematics, and conservation. – Biol. J. Linn. Soc. 62: 421 – 442.
117
