SHORTER COMMUNICATIONS

Transcription

1 Jounal of Hepetology, Vol. 36, No. 2, pp , 2002 Copyight 2002 Society fo the Study of Amphibians and Reptiles Do Dietay Habits edict Scale Counts in Snakes? RICHARD SHINE, Biological Sciences A08, Univesity of Sydney, New South Wales 2006, Austalia; Although studies on the biomechanical basis of feeding in snakes have evealed many links between dietay composition and feeding stuctues (e.g., ough and Goves, 1983; Cundall, 1987), the stength and geneality of such links have aely been examined quantitatively. In this pape, I exploe a simple ecomophological question: can the type of pey eaten by a snake species be used to pedict two aspects of its scale counts (the numbes of supalabial scales and dosal scale ows)? The idea was fist expessed moe than 50 yeas ago by an innovative Geman scientist, Rudolf Mell (see Zhao and Adle, 1993). Based on his eseach in China, Mell noted that snake species eating lage pey (e.g., mammals athe than eptiles) had moe supalabial scales and moe midbody scale ows (Mell, 1929a,b). Mell suggested that an inceased numbe of scale ows povides geate flexibility, enabling the skin between the scales to stetch ove a lage pey item. Moe ecent authos have epeated this suggestion. Fo example, Gans (1974) intepeted the inceased numbe of scale ows on the neck of egg-eating snakes in this way. ough and Goves (1983) intepeted high supalabial and midbody scale-ow counts in vipeid snakes (compaed to colubids) in the same fashion. Mell s oiginal fomulation was simplistic, howeve, and ecent analyses of snake feeding povide moe sophisticated views on these topics. Fo example, Geene (1983, 1997) has pointed out that pey items can be lage in a numbe of ways; Mell s hypothesis elies upon pey being lage in diamete athe than (fo example) mass. Equally, it is a goss ovesimplification to ate all mammals that ae consumed by snakes as being lage (elative to the snake s size) than ae all eptiles. Last, elasticity of the body could be enhanced by changing the stuctue of the skin, o inceasing the amount of skin between scales, athe than by simply inceasing the numbe of scale ows (Jayne, 1988). Given these issues, was Mell coect in suggesting an association between scale counts and dietay habits in Chinese snakes? And if so, how can this be tue given the ovesimplifications embedded in his assumptions? I used ope s (1935) tabulation of data on scale counts and pey types to test Mell s hypothesis on Chinese snakes. To examine the validity of the idea fo a diffeent assemblage, I also collated data on scale chaactes, body sizes and dietay composition fo all teestial Austalian snake species. Infomation on scale counts was obtained fom publications by Wallach (1985), Sto et al. (1986), Bake and Bake (1994), O Shea (1996), and Gee (1997). I ecoded the mean (o modal, if no mean was epoted) numbe of dosal scale ows (the numbe of scales encicling the body, counted halfway between head and tail) and the numbe of supalabial scales (i.e., the scales along the top of the lip). Infomation on dietay composition and mean adult body sizes of snakes came fom dissection and measuement of museum specimens (Shine, 1991, 1994). Diets wee quantified in tems of the popotion of identifiable pey items falling into each majo goup (invetebate, fish, amphibian, eptile, bid, mammal). Mean adult body size was calculated as the aveage of mean snout vent lengths fo adult males and adult females. Thee kinds of analysis wee caied out on these data. Fist, I teated each species as a statistically independent unit and looked fo the pedicted coelations between the two types of scale chaactes (dosal scale ows and supalabials) and between these taits and dietay composition (e.g., popotion of mammals in the diet). Second, because I found stong coelations between scale counts and body size (see below), I also compaed dietay composition to size-coected scale counts. To calculate size-coected scoes fo each of the scale chaactes, I used esidual scoes fom the geneal linea egessions of scale count vesus mean body length. Thus, a species with a positive esidual scoe fo dosal scale-ows would be one that had moe scale-ows than would be expected fo a species of that mean body size. Last, I combined the data with a putative phylogeny fo these snakes (Fig. 1). This phylogenetic hypothesis was assembled by combining published suggestions about elationships among vaious subsets of the species involved (Kluge, 1991, 1993; Gee, 1997; Keogh et al., 2000; S. Keogh, pes. comm.). Hence, the esultant phylogeny (Fig. 1) is speculative and is not the esult of any explicit phylogenetic analysis. Then I used this composite phylogeny to conduct a compaative analysis using independent contasts (uvis and Rambaut, 1995). This method ovecomes the poblem that many taits ae highly consevative, so that coelations among taits may eflect phylogenetic inetia athe than functional elationship (Havey and agel, 1991). Evolutionay changes in one vaiable (such as the popotional composition of the diet) within a clade can be compaed to simultaneous changes in othe taits (such as scalation) within the same clade. In the absence of eliable data on banch lengths within phylogenies, I assumed constant banch lengths (i.e., punctuated speciation events). hylogenetically coected elationships among vaiables wee assessed by linea egession though the oigin (uvis and Rambaut, 1995). Statistical analyses wee pefomed using CAIC (uvis and Rambault, 1995) and Statview 5 on an Apple Macintosh G4 compute. io to analysis, data wee tested fo confomity with assumptions of nomality equied fo tests. A one-facto ANOVA on ope s 1935 data (his table LX) with pey type as the facto and numbe of scale ows as the dependent vaiable, confims Mell s (1929a,b) assetion that dietay habits coelate with scale counts within the Chinese snake fauna (F 5, , 0.003). osthoc (Fishes LSD) tests show that species that eat mammals have significantly moe

2 269 FIG. 1. hylogenetic hypothesis fo Austalian snakes, as used in the pesent study. See text fo list of souces fo this phylogeny. midbody scale ows than those that eat eptiles, fogs o invetebates ( 0.05 in each case). Data wee obtained fo 108 Austalian species, compising 11 pythonids, nine colubids, 76 elapids and 12 typhlopids. Mean adult snout vent length anged fom cm, midbody scale ows fom 13 67, and supalabial scale counts fom The popotion of the diet composed of vaious pey types also vaied widely (fom 0 100% fo each of fishes, eptiles, and mammals; fom 0 97% fo amphibians; 0 36% fo bids). Oveall, my analyses fo the Austalian taxa evealed the same geneal patten as noted by Mell fo Chinese snakes but cast substantial doubt on the mechanism that he suggested to explain this patten. Table 1 summaizes esults fom statistical tests of thee pedictions fom Mell s hypothesis. Numbe of Midbody Dosal Scale-Rows Is ositively Coelated with Numbe of Supalabial Scales. As pedicted, species with high numbes of midbody dosal scale ows also had many supalabial scales (Fig. 2, Table 1). Howeve, both of the scale counts also showed stong positive coelations with mean adult body size. These elationships diffeed among families. Fo example, at any given body size, pythons have moe supalabial scales than elapids, and colubids ae intemediate in this espect (Fig. 3). The coelation between dosal and labial counts emained high even afte the effects of body size wee emoved (Table 1). That is, the tend fo species with high numbes of scale-ows to also have many supalabial scales is not an indiect consequence of body size. Compaative analysis futhe suppoted the notion of a functional elationship between scale numbes aound the body vesus along the uppe lip. Inceases in midbody scale ows duing phylogeny wee consistently associated with inceases in supalabial counts (Table 1). Afte emoving body-size effects fom the compaative analysis, shifts in elative numbes of midbody dosal scale-ows (i.e., esidual scoes) wee associated with shifts in elative numbes of supalabial scales (i.e., esidual scoes; N 70, 0.65, ). Species that Eat Lage ey Types Have Moe Dosal Scale Rows. As in the Chinese fauna, Austalian snake species that feed on bids and mammals tend to have moe dosal scale ows than do othe species (Table 1; Fig. 4). When the effects of body size on scale counts ae emoved, howeve, none of the pobability values fo these analyses attain the conventional level of significance ( 0.05) afte Bonfeoni coections ae applied (Table 1). hylogenetically based analysis evealed a consistent association between dosal scale ows and the popotion of mammalian pey in the diet but not fo the othe pey goups (Table 1). Using size-coected scoes (i.e., esidual values fom scale counts vesus SVL) geatly weakened the appaent link between scalation and diet: no dietay popotions wee significantly coelated with size-coected scale counts ( 0.10 in all cases). Species that Eat Lage ey Types Have Moe Supalabial Scales. attens fo this vaiable ae simila to those fo dosal scale ows (above), as might be expected fom the high coelation between the two taits (Fig. 2). Species that feed on bids and mammals have moe supalabial scales than do species that feed on ectothems (Fig. 4), but this patten disappeas afte body-size effects ae emoved fom the analysis (Table 1). hylogenetic shifts in supalabial counts wee coelated (albeit weakly) with shifts in the popotion of mammalian pey (Table 1), but this effect disappeaed when size-coected values wee used in the analysis (all 0.10).

3 270 SHORTER COMMUNICATIONS TABLE 1. Relationsihps among scale chaactes and dietay composition in Austalian snakes. The table povides esults fom linea egession analysis of thee types of data. Fist, values fo each species wee used as independent datapoints ( aw data ). Then, because both types of scale counts wee highly coelated with oveall body size, the effects of size wee emoved by calculating esidual scoes fom the geneal linea egessions of the scale tait against mean adult snout vent length ( size-coected data ). Thid, compaative analysis using independent contasts was used to emove the effects of phylogenetic consevatism. Sample size was 108 fo the fist two types of tests and 70 fo the compaative analysis. The table shows the coelation coefficient () and its associated pobability (). Boldface shows esults that ae statistically significant ( 0.05) afte Bonfeoni coection. Taits being compaed Raw data Size-coected data Compaative analysis Dosal scale ows vs supalabial scales Dosal scale ows vs. % fishes Dosal scale ows vs % fogs Dosal scale ows vs % eptiles Dosal scale ows vs % bids Dosal scale ows vs % mammals Supalabial scales vs % fishes Supalabial scales vs % fogs Supalabial scales vs % eptiles Supalabial scales vs % bids Supalabial scales vs % mammals FIG. 2. The elationship between the numbes of dosal scale ows and supalabial scale count in an intespecific compaison among Austalian snakes. Each point epesents one species. See Table 1 fo esults of a test of the statistical significance of this elationship. The end esult of these analyses is that thee taits (body size, dietay habits, and scale counts) ae significantly intecoelated. Although it is difficult to distinguish diect fom indiect effects within such constellations of taits, patial coelation offes a patial solution. I used this technique to evaluate the coelation between dietay habits and one vaiable (such as scale-ow numbe) when anothe facto (such as body size) was held constant. Two sets of these analyses wee conducted, one on the aw data and one on phylogenetic contasts. In both cases, patial-coelation analyses suggested that dietay habits wee moe highly associated with body-size than with scale counts. Fo the aw data, the popotion of the diet composed of mammals was significantly coelated with mean adult SVL when eithe scale-ow numbes o supalabial counts wee held constant (N 108, 0.63, 0.71, espectively, in both cases) but neithe of the scale chaactes was coelated with dietay habits if SVL was held constant (N 108, 0.03, 0.05, espectively, 0.50). No othe patial coelations between dietay habits and mophological taits wee statistically significant ( 0.05) afte Bonfeoni coection fo multiple tests. The compaative analysis (independent contasts) povided an almost identical esult. The popotion of mammals in the diet was coelated with mean adult SVL when eithe of the scale chaactes was held constant (N 70, 0.56, 0.55, espectively, in both cases) but not with eithe of the scale counts when SVL was held constant (N 108,, 0.04, espectively, 0.50). These esults suggest that the coelations between scale counts and diet ae seconday consequences of body-size effects on both scalation and diet. If indeed we could find stong links between ecological taits (such as dietay composition) and mophological featues (such as the numbes of dosal scale-ows), such associations would be of consideable value. Fo example, we have quantitative infomation on dietay composition fo only a small popotion of snake species. Unfotunately, my study is not encouaging with espect to infeing diet fom scalation (o vice vesa). Thee ae at least thee pocesses that might geneate a coelation between dietay composition and scale counts in snakes. Fist, intespecific divegences in dietay habits (especially, elative pey size) might have imposed selection on scale-ow numbes as envisaged by Mell. Second, the

4 271 FIG. 3. The elationship between body size (mean adult snout vent length) and scale counts in Austalian snakes. Lowe gaph shows data fo the numbe of dosal scale ows, and uppe gaph shows the data fo the numbe of supalabial scales. Each point epesents one species. Linea egession fo these datasets shows: dosal scale ows, N 108, 0.65, ; supalabial scales, N 108, 0.62,. coelation between scale counts and dietay composition might be an atifact of phylogenetic inetia, because some lineages of snakes have low values fo all thee taits (supalabials, dosal scale ows, and popotion of the diet composed of endothems) wheeas othe lineages have high values fo all thee taits. This cicumstance might have aisen ealy in snake phylogeny, and equie no adaptive explanation in tems of selective foces duing the adiation of species within lineages. In pactice, phylogenetic shifts in the degee of eliance on mammalian pey wee consistently accompanied by shifts in scale counts (Table 1), suggesting a functional association between the taits. Nonetheless, phylogenetic consevatism undoubtedly amplifies the stength of these coelations (Table 1, Fig. 4). Last, coelation does not imply causation. A thid facto (mean adult body size) is known to coelate both with dietay habits (e.g., Fitch, 1960; Anold, 1993; Shine, 1994) and with scale counts (e.g., Klaube, 1956; Lindell, 1994). Thus, a shift in food habits towad lage pey could favo the evolution of lage body size, which in tun could favo an incease in scale counts (because lage snakes have moe scales: see Fig. 3). In keeping with this hypothesis, coelations between scale counts and dietay habits FIG. 4. Scale chaacteistics of Austalian snakes in elation to dietay composition. Species that consume a highe popotion of bids and mammals tend to have moe dosal scale ows and moe supalabial scales. See Table 1 fo tests of the statistical significance of these elationships. geneally disappeaed when body-size effects wee emoved fom the analysis (Table 1). In summay, Mell was pobably coect in ecognizing a geneal association between scale counts and diets in snakes, but wong about its cause. To fully undestand the patten identified by Mell 50 yeas ago, we will need to gathe data on taits that elate much moe closely to the actual selective pessues and functional challenges involved in pey ingestion. Fo example, many eseaches in snake ecology still epot only the species and mass of pey items, despite heatfelt appeals to take moe extensive data (e.g., Geene 1997). Maximum pey diamete is simple to measue (even with patially digested pey) and is likely to be a bette measue of the physical difficulty of ingesting the item than is pey mass. Until we have extensive datasets on such vaiables, many of the ideas espoused by visionaies like Rudolf Mell will emain untested. Acknowledgements. I thank K. Adle fo extensive infomation on the life and science of R. Mell, A. Schulmeiste fo tanslations, H. Geene fo an unpublished manuscipt on snake feeding, and M. Elphick fo data analysis. S. Keogh geneously povided advice on elapid phylogeny. The study was funded by the Austalian Reseach Council. LITERATURE CITED ARNOLD, S. J Foaging theoy and pey-sizepedato-size elations in snakes. In R. A. Seigel

5 272 SHORTER COMMUNICATIONS and J. T. Collins (eds.), Snakes. Ecology and Behavio, pp McGaw-Hill, New Yok. BARKER, D. G., AND T. M. BARKER ythons of the Wold. Vol. 1. Austalia. Advanced Vivaium Systems, Lakeside, CA. CUNDALL, D Functional mophology. In R. A. Seigel, J. T. Collins, and S. S. Novak (eds.), Snakes: Ecology and Evolutionay Biology, pp MacMillan, New Yok. FITCH, H. S Autecology of the coppehead. Univesity of Kansas ublications of the Museum of Natual Histoy 13: GANS, C Biomechanics: An Appoach to Vetebate Biology. J.. Lippincott, hiladelphia, A. GREENE, H. W Dietay coelates of the oigin and adiation of snakes. Ameican Zoologist 23: Snakes. The Evolution of Mystey in Natue. Univesity of Califonia ess, Bekeley. GREER, A. E The Biology and Evolution of Austalian Snakes. Suey Beatty and Sons, Sydney, New South Wales, Austalia. HARVEY,. H., AND M. D. AGEL The Compaative Method in Evolutionay Biology. Oxfod Studies in Ecology and Evolution, Oxfod Univesity ess, Oxfod. JAYNE, B. C Mechanical behaviou of snake skin. Jounal of Zoology, London 214: KEOGH, J. S., I. A. SCOTT, AND J. D. SCANLON Molecula phylogeny of vivipaous Austalian elapid snakes: affinities of Echiopsis aticeps (Sto, 1980) and Dysdalia coonata (Schlegel, 1837), with desciption of a new genus. Jounal of Zoology, London 252: KLAUBER, L. M Rattlesnakes. Thei Habits, Life Histoies and Influence on Mankind. Univesity of Califonia ess, Bekeley. KLUGE, A. G Boine snake phylogeny and eseach cycles. Miscellaneous ublications of the Museum of Zoology, Univesity of Michigan 178: Aspidites and the phylogeny of pythonine snakes. Recods of the Austalian Museum, Supplement 19:1 77. LINDELL, L. E The evolution of vetebal numbe and body size in snakes. Functional Ecology 8: MELL, R. 1929a. Beitage zu Fauna Sinica. IV. Gundzuge eine Okologie de Chinesischen Reptilien und eine Hepetologischen Tiegeogaphie Chinas. Walte de Guyte, Belin, Gemany b (1931). eliminay contibutions to an ecology of East Asiatic eptiles, especially snakes. Lingnan Science Jounal 8: O SHEA, M A Guide to the Snakes of apua New Guinea. Independent ublishing, ot Moesby, New Guinea. OE, C. H The Reptiles of China. Ameican Museum of Natual Histoy, New Yok. OUGH, F. H., AND J. D. GROVES Specialization in the body fom and food habits of snakes. Ameican Zoologist 23: URVIS, A., AND A. RAMBAUT Compaative analysis by independent contasts (CAIC): an Apple Macintosh application fo analysing compaative data. Compute Applications in the Biosciences 11: SHINE, R Austalian Snakes. A Natual Histoy. A. H. and A. W. Reed, Sydney, New South Wales, Austalia Allometic pattens in the ecology of Austalian snakes. Copeia 1994: STORR, G. M., L. A. SMITH, AND R. E. JOHNSTONE Snakes of Westen Austalia. Westen Austalia Museum, eth, Westen Austalia, Austalia. WALLACH, V A cladistic analysis of the teestial Austalian Elapidae. In G. C. Gigg, R. Shine and H. Ehmann (eds.), The Biology of Austalasian Fogs and Reptiles, pp Royal Zoological Society of New South Wales, Sydney, New South Wales, Austalia. ZHAO, E., AND K. ADLER Hepetology of China. Society fo the Study of Amphibians and Reptiles, Oxfod, OH. Accepted 25 June Jounal of Hepetology, Vol. 36, No. 2, pp , 2002 Copyight 2002 Society fo the Study of Amphibians and Reptiles Testing fo Equal Catchability of Tituus Newts by Dip Netting J. W. ARNTZEN 1, Cento de Estudos de Ciência Animal, Univesidade do oto (CECA/U), Campus Agáio de Vaião, Vila do Conde, otugal Testing the assumption of equal catchability is fundamental to undestanding amphibian populations, modeling population dynamics, and testing hypotheses concening life-histoy taits. Unequal catchability foms an impotant souce of potential eo in quantitative suveying, affecting the estimation of population paametes such as population size, in paticula when the captue-ecaptue method is used (Bohlin and Sundstöm, 1977; Caothes, 1979; ollock et al., 1990). Nonepesentative sampling has been epoted fo many animal species in a wide ange of taxa, and it has been questioned whethe populations consisting of animals with equal catchability actually exist (Caothes, 1973). In amphibian studies, sampling bias has been mostly ignoed, with some notable exceptions (Gelde and Rijsdijk, 1987; Shiose and Books, 1995; Wood et al., 1998; Wilson and eaman, 2000). The techniques available fo the study of amphibians include some sophisticated appoaches (Heye et al., 1994; Mölle and Kupfe, 1998). Howeve, the sampling of many diffeent habitats will be facilitated by the use of simple, small, and inexpensive tools, such as the taditional dip net. Amphibians that beed ove long peiods in lentic wate and those with laval stages of long duation ae especially amenable to netting. aleactic newts of the genus Tituus fall into this cat- 1 esent addess: Natualis-National Museum of Natual Histoy,. O. Box 9517, 2300 RA Leiden, The Nethelands; antzen@natualis.nnm.nl