UNWtRSlTY OF ILLINOIS LIBRARY pr URBANA CHAMPAIGN "* BIOLOGY m FIB 1 DIANA 590.5 Fl ,.Bl ^4-1987 Zoology ^TEW SERIES. NO. 34 :;^Z «% Patterns of Snake Evolution Suggested ®? by Their Proteins Herbert C. Dessauer John E. Cadle Robin Lawson ^ubrarv of the 'ESSE .0,3 A Contribution in Celebration of the Distinguished Scholarship of Robert F. Inger on the Occasion of His Sixty-Fifth Birthday May 29, 1987 Publication 1376 PUBLISHED BY FIELD MUSEUM OF NATURAL HISTORY Information for Contrihut* uld follow (botanical pa[>. rm: s, but within th< FIELDIANA Zoology NEW SERIES, NO. 34 Patterns of Snake Evolution Suggested by Their Proteins Herbert C. Dessauer Department of Biochemistry Louisiana State University Medical Center 1901 Perdido Street New Orleans, Louisiana 70112 Robin Lawson The Museum of Zoology Louisiana State University Baton Rouge. Louisiana 70803 Present Address: Department of Herpetology California Academy of Sciences Golden Gate Park San Francisco, California 94118 John E. Cadle Department of Biochemistry Louisiana State University Medical Center 1901 Perdido Street New Orleans, Louisiana 70112 A Contribution in Celebration of the Distinguished Scholarship of Robert F. Inger on the Occasion of His Sixty-Fifth Birthday Accepted for publication September 22, 1986 May 29, 1987 Publication 1376 PUBLISHED BY FIELD MUSEUM OF NATURAL HISTORY © 1987 Field Museum of Natural History ISSN 0015-0754 PRINTED IN THE UNITED STATES OF AMERICA Table of Contents VIII. Acknowledgments 28 IX. Literature Cited 28 Abstract 1 I. Introduction 1 II. Methods and Nature of the Evi- dence 2 III. Biochemical Genetics and Popula- tion Structure 7 A. Protein Inheritance 7 B. Population Genetics 7 IV. Species Formation 11 V. Differential Rates of Morphologi- cal Evolution 13 A. Radiations Illustrating Rapid Mor- phological Evolution 13 1 . North American Natricines ... 13 2. North American Colubrines ... 15 3. Xenodontines 15 4. Australian Elapids/Hydrophiids 15 B. Taxa Illustrating Covergence or Slow Morphological Evolution ... 17 C. Implications of Differential Rates of Evolution in Snakes 18 VI. Higher Levels of Relationship 19 A. Position of Snakes Among Rep- tiles 19 B. Relationships Within and Between Major Groups of Snakes 19 1 . Scolecophidia 21 2. Henophidia 21 a. Boidae/Pythonidae/Tropi- dophiidae 21 b. Uropeltidae/Aniliidae 21 c. Acrochordidae 22 d. Conclusions 22 3. Caenophidia 22 a. Viperidae 22 b. Elapidae 23 c. Atractaspis 24 d. Colubridae 25 VII. Summary 27 List of Illustrations 1 . Robert F. Inger preparing a liquid-N tank during a 1986 trip to southeast Asia 2 2. Natives loading donkeys with liquid-N in the Peruvian Andes 3 3. Field and laboratory scientists of the 1969 ALPHA HELIX Expedition to New Guinea 4 4. Professor G. H. F. Nuttall 6 5. Illustration of a rate test 8 6. Genetic differentiation among species within North American colubrid radia- tions 14 7. Genetic differentiation among species within two xenodontine radiations 16 8. Enhanced Ouchterlony double-diffusion tests of albumins 20 List of Tables 1 . Indices of genetic variability for snake populations 9 2. Immunological distances between the al- bumin of Boa constrictor and other heno- phidian albumins 21 3. Immunological distances and rate tests concerned with the albumins of the Vi- peridae 23 4. Immunological distances between the al- bumins of Atractaspis and other ad- vanced snakes, using an antiserum to Atractaspis bibroni albumin 25 5. Immunological distances involving Lyco- dontine/Boodontine albumins 27 in Patterns of Snake Evolution Suggested by Their Proteins Abstract Genetic variability and other data on snake pro- teins are reviewed in the context of population genetics, species relationships, and current phy- logenetic hypotheses. Protein diversity in snakes is comparable to that reported in other vertebrates, and protein polymorphisms are useful for iden- tifying individual snakes as well as for studies of breeding patterns, population genetics, and species formation. Such biochemical data suggest that many populations of natricine snakes, presently classified as subspecies, are already reproductively isolated or are "incipient" species. As molecular evolution is largely divergent and often regular over time, comparative protein evidence allows one to overcome many of the difficulties encoun- tered in estimating branching patterns of organ- isms. Such comparisons support the following conclusions: (1) many major lineages of snakes include one or more highly speciose radiations of relatively recent origin; (2) some genera are relics of ancient radiations; (3) lizards are the closest relatives of snakes, which are probably monophy- letic; (4) primitive snakes include a number of ancient lineages that are probably not monophy- letic; (5) vipers are the sister group of other ad- vanced snakes; (6) sea snakes are closely related to Australopapuan elapids; (7) natricines and col- ubrines are probably monophyletic but the xe- nodontine and lycodontine groups possibly are not; and (8) relationships among major clades of the Colubridae remain unresolved. I. Introduction Comparative protein studies have demonstrat- ed that molecular information is capable of solving many previously intractable problems concerned with the evolutionary biology of snakes. Such problems arise because of the extreme conver- gence, parallelisms, and specializations that typify snake morphology (Underwood, 1967). The suc- cess of the molecular approach in giving insight where traditional methods have failed is due in part to the fact that evolutionary processes at the molecular and morphological levels are largely in- dependent (Wilson et al., 1977). Morphological evolution is highly variable in rate, rapid within some groups but extremely slow in others, and often subject to homoplasy (Simpson, 1953; Gould & Eldredge, 1977). In contrast, protein evolution is largely divergent and evolutionary change can be measured in units of known quantity (amino acid substitutions) that are often regular over time. Because of this, evidence from proteins allows one to overcome many difficulties in evolutionary studies traceable to convergence, parallelism, and specialization (Fitch, 1982). Although many ques- tions remain regarding the evolution of snakes, general patterns concerned with their genetic di- versity, population structure, speciation, historical biogeography, and phylogeny are discernible in protein structure. In this paper we present an over- view of current comparative protein evidence bearing upon evolution within and among lineages of living snakes. We dedicate this paper to Robert F. Inger (fig. 1), who always has been highly supportive of such nontraditional approaches to evolutionary stud- ies. Bob Inger and other biologists whose primary focus is field and comparative anatomical study have generally originated the phylogenetic hy- potheses upon which molecular biologists base their experimental work. Realizing that many problems can be solved with molecular data, systematic bi- ologists are collecting tissues in the field with in- creasing frequency for use in such research. Be- DESSAUER ET AL.: SNAKE EVOLUTION Fig. I . Robert F. Inger preparing a liquid-N tank and other gear for storing tissues from specimens collected during a 1986 trip to Southeast Asia. (Photo by Harold Voris.) sides the usual strenuous activities and logistic problems of fieldwork, the collector seeking tissues for molecular studies must transport liquid nitro- gen tanks, centrifuges, and other unusual equip- ment into the held, many times into almost im- penetrable areas (fig. 2). The frozen tissue collections that result from these activities provide a valuable resource for research in evolutionary biology and a variety of other disciplines. "Often, the principal contribution to a molecular study is the field work of the naturalist who provides the tissue upon which the study is based" (Dessauer & Hafncr. 1984). We believe that collaborative interactions between field- and laboratory-orient- ed scientists (fig. 3) are producing more definitive phytogenies and a more comprehensive under- standing of evolutionary processes. II. Methods and Nature of the Evidence Many biochemical techniques have been used to acquire comparative protein evidence regarding snake evolution. The majority of these data were obtained by means of electrophoresis (Smithies, 1959) of a wide variety of proteins, microcomple- ment fixation (Champion et al., 1974), or other immunological comparisons (Goodman & Moore, 1971) of transferrins and albumins, and peptide fingerprints (Canfield & Anfinsen, 1963) of hemo- globins. Amino acid sequences of comparative value on snake proteins are rare, consisting prin- cipally of those for neurotoxins of elapid venoms (Strydom, 1973, 1979; Hseu et al., 1977; Yang, 1978; Mebs, 1985; Tamiya, 1985). The use of protein evidence in evolutionary FIELDIANA: ZOOLOGY Fig. 2. Natives loading donkeys with liquid-N tanks for transport to a camp site of an L.S.U. Museum of Zoology expedition high in the Peruvian Andes. (Photo by J. P. O'Neill.) DESSAUER ET AL.: SNAKE EVOLUTION -X FIELDIANA: ZOOLOGY studies depends upon the fact that homologous genes among species diverge in a continuous fash- ion from the time of reproductive isolation. As a result, the sequence differences between the pro- tein products of homologous genes among taxa represent an estimate of the degree of divergence among the organisms themselves (Zuckerkandl & Pauling, 1 965). Furthermore, some genes and their protein products can be shown to diverge in a clocklike manner. The variation in rate for a par- ticular protein is usually about twice that expected for a simple Poisson process such as radioactive decay (Wilson et al., 1977). Although amino acid sequences offer the max- imum evolutionary information encoded in the protein molecule, they are difficult ;to determine and require extensive expenditure of tissue, time, and money. The majority of questions confronting the evolutionary biologist can be answered sooner and far more economically with data sets on pro- teins based upon peptide fingerprinting and elec- trophoretic and immunological methods. Nuttall (fig. 4), for example, using the most primitive of immunological methods, had by 1 904 successfully predicted the relative affinities of major lineages of primates (Nuttall, 1904). Comparative immunological data sets on an ho- mologous series of proteins are highly correlated with amino acid sequence differences between the proteins. Wilson and his colleagues (1977; see also Benjamin et al., 1984) have shown that immu- nological distances (IDs) obtained by microcom- plement fixation titrations (MC'F) are directly pro- portional to sequence divergence of the same proteins. Antigenic distances obtained by im- munodiffusion, in turn, are directly proportional to the MC'F IDs on the same protein (Goodman & Moore, 1971; Schwaner & Dessauer, 1982). When applied to closely related organisms, even genetic distances between taxa estimated from electrophoretic data sets on proteins relate roughly to the IDs for transferrins or albumins of the same taxa (Sarich, 1977; Wilson et al., 1977; Maxson & Maxson, 1979; Wyles & Gorman, 1980). Each of these indirect methods differs in the nature of the evidence it furnishes and in the taxo- nomic levels at which it is most efficiently applied. Electrophoresis of proteins yields banding patterns of allozymes, the protein phenotype, that can be interpreted genotypically (Harris, 1975; Harris & Hopkinson, 1976; Dessauer et al., in press); con- sequently, the method has a wide variety of ap- plications, especially in studies of inheritance in- volving closely related organisms. These include the detection of allelic variation at structural gene loci, determination of genotypes of individuals at polymorphic loci, estimation of levels of genetic diversity within populations, and genetic distances between populations and species (A vise, 1974; Smith et al., 1982). Currently, specific staining techniques are available to detect proteins deter- mined by approximately 200 different loci, in- cluding nonenzymic proteins and enzymes of all major classes (Harris & Hopkinson, 1976; Hames & Rickwood, 1981). Immunological estimates of protein divergence are measures of structural changes at antigenic sites (Benjamin et al., 1984). As rates of divergence of proteins coded by different structural genes vary more than one hundred-fold (Wilson et al., 1977), one of the factors to be considered in selecting a protein for use in a taxonomic study is its rate of evolutionary change. For example, transferrin is more sensitive than albumin for estimating the affinities of closely related taxa, as it usually di- verges more rapidly than albumin, about twice as fast on the average in snakes (Cadle & Dessauer, 1985); however, albumin comparisons are valu- able over a wider range of taxa. Quantitative pre- cipitin and MC'F estimations of protein diver- gence can be used to construct phylogenetic trees showing the branching order of taxa and the amounts of protein divergence attributed to each lineage (Felsenstein, 1982). The semiquantitative immunodiffusion method (Goodman & Moore, 1971) is simpler but less sensitive than MC'F. The immunodiffusion ap- proach is ideal for survey studies and can also be used to construct phylogenetic trees (Dene et al., 1978), although their precision is generally less than with the more quantitative methods. Spur formation, the sign of protein divergence in the immunodiffusion reaction, does not occur with transferrins having MC'F IDs below 20 (Schwaner & Dessauer, 1982). It is generally not possible to use immunodiffusion to detect cross-reactions be- tween transferrins or albumins with antibodies to their homologs from widely divergent species; however, by adding polyethylene glycol to the gel to lower the solubility of the resultant antigen- antibody complex, reactions can be visualized be- tween antisera to snake albumins and albumins from the most distantly related snake taxa, even from albumins of some lizards (cf. fig. 8). Comparisons of fingerprints of purified proteins can furnish estimates of the minimum number of sequence differences between homologous pro- teins. The proteins to be tested are hydrolyzed with DESSAUER ET AL.: SNAKE EVOLUTION Fio. 4. Professor G. H. F. Nuttall, the pioneer worker in comparative immunology. (Photo from M. F. Shaffer.) trypsin or some other enzyme with a high speci- ficity for particular peptide bonds. The resultant peptide fragments are spread across sheets of filter paper by means of chromatography and electro- phoresis and then visualized with chemical stains. The sequence divergence of the proteins is esti- mated on the basis of differences in positions and staining properties of peptide fragments (Sutton, 1969; Dessauer, 1974; Mao et al., 1978, 1984). The relative timing of lineage separation can be inferred from quantitative estimates of sequence divergence between homologous proteins in dif- ferent lineages, such as provided by immunolog- ical distances. Estimates of absolute times of sep- FIELDIANA: ZOOLOGY aration of the taxa under study are possible if the rate of evolution of the protein can be estimated from fossil and/or biogeographic evidence (Max- son et al., 1 975) and if the rate is relatively constant among the lineages, as shown by the relative rate test (fig. 5; Wilson et al., 1977). Although rates of albumin and transferrin evolution are approxi- mately constant in most lineages, they can vary substantially within and especially among some vertebrate lineages (e.g., Sarich, 1985). Relative rate tests suggest that evolutionary rates for both albumins and transferrins differ somewhat among snake lineages (Cadle, 1982a,b; unpubl. data). For example, the rate of albumin evolution in some vipers appears to be at least 30% slower than in the majority of elapid and colubrid lineages (see sec. VLB., 3a). Despite these differences in rate, however, the molecular clock concept can still be used in interpreting aspects of snake evolutionary history as long as rate differences are recognized and measured. III. Biochemical Genetics and Population Structure A. Protein Inheritance High resolution electrophoresis of tissue ho- mogenates followed by the identification of elec- tromorphs for specific proteins has yielded con- siderable evidence on genetic diversity at structural gene loci, both within single populations and be- tween populations presumably undergoing species formation. At such close levels of relationship, the majority of allelic differences at a specific locus is traceable to point mutations. If these result in ami- no acid substitutions in the polypeptide deter- mined by the gene, the variant protein may be electrophoretically detectable. In snakes, as in other vertebrates, most proteins are inherited as the products of codominant al- leles. Direct evidence for codominance as the mode of inheritance of proteins in snakes has been shown in studies involving the breeding colony of king- snakes at the American Museum of Natural His- tory (Dessauer & Zweifel, 1981) and in laboratory- bred rat snakes of known parentage (Lawson & Dessauer, unpubl. data). The maternal contribu- tion to protein phenotypes of their offspring has been observed for rattlesnakes (Crabtree & Mur- phy, 1984; Murphy & Crabtree, 1985) and many species of natricine snakes (Schwaner et al., 1980; Dessauer & Lawson, unpubl. data). Considerable indirect evidence on snakes as well as on other vertebrates supports this general conclusion (Des- sauer et al., in press). Electrophoretic phenotypes for many proteins within a species or species group of snakes are identical, even in individuals from widely sepa- rated areas of the geographic range of a species. These invariant proteins are often useful markers for identifying species, or higher categories of snakes in a cladistic analysis. Most species, how- ever, have some proteins that are polymorphic; in snakes these commonly include transferrin, phos- phogluconate dehydrogenase, phosphoglucomu- tase, and esterase-D. Although most polymorphic loci are diallelic, three or more alleles are com- monly observed at the transferrin locus in snakes of the same population (Dessauer et al., 1962; Gartside et al., 1977; Lawson & Dessauer, 1979). Amino acid oxidases (Jimenez-Porras, 1 964a; Aird & Dessauer, 1977), proteases (Jimenez-Porras, 1964a,b) and toxins of venoms of viperid snakes (Schenberg, 1959) are so polymorphic that viperid venoms may have the most highly variable protein composition of any biological fluid (see Dessauer, 1974). Polymorphic proteins have been used to iden- tify individual snakes and to study their breed- ing patterns. Individuals in populations of Both- rops neuwiedi from southeastern Brazil could be identified by patterns of six venom antigens (Schenberg, 1963). Knowledge of transferrin and prolidase genotypes of individual kingsnakes (Lampropeltis getulus) in the American Museum of Natural History colony allowed Zweifel and Dessauer (1983) to plan matings that proved that kingsnake broods can be the result of insemina- tions by at least two males. Polymorphisms at the albumin, transferrin, superoxide dismutase, and esterase-D loci were utilized by Banks and Schwa- ner (1984) to show that a brood of Australian py- thons, conceived and hatched at the Melbourne Zoo, were progeny of a mating between Python spilotes and P. amethystinus, and that the P. ame- thystinns female, coiled about the clutch of eggs during their incubation, was not the mother of the brood. B. Population Genetics Alleles responsible for polymorphic proteins may be rare to moderate in frequency, widespread, or DESSAUER ET AL.: SNAKE EVOLUTION B 25 B 35 20 X y + 25 y + 10 y+10 Fig. S. Illustration of a rate test. We are interested in the relationships among ingroup taxa A, B, and C. Observed molecular distances are given in the matrix. A straightforward apportionment of these distances would result in the estimated phytogeny illustrated in b. A rate test shows this to be in error. To perform the test, outgroup-X is chosen on the basis of non molecular evidence (e.g., morphology), and the distances between X and taxa A, B, and C are measured (matrix row 4; variable y is that portion of the distance between the outgroup and ingroup that is the same for all members of the ingroup). The rate test shows that taxa B and C are conservative relative to taxon A and, thus, the distances should be apportioned as in a. The resulting phytogenies a and b differ in branching order. In this example, the fact that taxa B and C are both conservative suggested their apparent phylogenetic association, but the rate test can be used to properly assess their relationships (Cadle, 1984a). restricted to a specific population. The origin of geographic variation in allele distribution may be traceable to isolation by distance or by some nat- ural barrier to gene flow. For example, garter snakes (Thamnophis sirtalis) from the northeastern and western coasts of North America are fixed for al- ternative alleles at the cytosolic superoxide dis- mutase locus (Lawson, 1978). Proteases have dif- FIELDIANA: ZOOLOGY Table 1 . Indices of genetic variability for snake populations. Species Geographic location No. of speci- No.of loci mens tested Hf Source* Rhinophis philippinus Phyllorhynchus arenicolus Thamnophis proximus T. couchii atratus T. c. couchii T. c. hydrophilus T. c. hammondii T. elegans terreslris T. e. vagrans T. e. vagrans T. ordinoides T. ordinoides T. brachystoma T. rufipunctatus T. sirtalis sirtalis T. s. sirtalis T. s. sirtalis T. s. sirtalis T. s. parietalis T. s. parietalis T. s. parietalis T. s. pickeringi T. s. dorsalis Nerodia fascial a confluens N. f. compressicauda N. f. clarkii Central SRI LANKA Isla San Marcos, Baja California, MEXICO Vicinity of La Place, "St. John the Baptist" Parish, La. Isenberg Ranch, San Mateo County, Calif. Feather River, Butte and Plumas counties, Calif. Applegate River, Jackson and Jose- phine counties, Ore. Picnic Lake Park, Potrero, San Diego County, Calif. Samoa Peninsula, Humboldt County, Calif. Florida Mesa, La Plata County, Colo. Qualicum Beach, Vancouver Island, CANADA Port Orford, Curry County, Ore. Parksville, Vancouver Island, CAN- ADA Allegheny River Valley, Warren County, Pa. Rio Papigochic, near Ciudad Guer- rero, Chihuahua, MEXICO Bono, Ottawa County, Ohio Islesboro Island, Waldo County, Maine Baton Rouge, La. Illinois Southwestern Illinois Inwood, Manitoba, CANADA In wood, Manitoba, CANADA Parksville, Vancouver Island, CAN- ADA Rio Grande Valley, N. Mex. Baton Rouge, La. Boca Ciega Bay, Pinellas County, Fla. Grande Isle, Jefferson Parish, La. 34 26 19.2 4.3 1 4 34 10.3 3.8 2 40 26 23.1 3.8 3 38 31 16.1 8.2 4 13 31 12.9 3.4 4 14 31 22.6 7.8 4 8 31 6.4 1.6 4 10 31 12.6 3.0 4 42 33 9.0 1.7 5 12 26 3.9 2.6 6 23 33 24.2 8.9 5 37 33 21.2 5.3 5,6 25 26 3.8 0.6 6 8 26 7.7 1.0 6 52 14 28.6 8.3 7 20 27 29.6 8.0 6 17 27 25.9 8.2 6 11 27 25.9 7.8 6 13 27 25.9 7.7 6 56 15 6.7 2.0 8 14 27 18.5 4.4 6 11 27 11.1 2.7 6 8 27 25.9 7.4 6 31 35 11.4 2.8 9 35 35 8.6 1.0 9 16 35 17.1 4.8 * P = percent polymorphism where the frequency of the most common allele does not exceed 0.95. t H = percent heterozygosity by direct count. % Sources of data: 1 = Dessauer, Gartside & Gans (unpubl. data); 2 = Murphy & Ottley (1980); 3 = Gartside et al. (1977); 4 = Lawson & Dessauer ( 1 979); 5 = Lawson (1978); 6 = Lawson (unpubl. data); 7 = Sattler & Guttman (1976); 8 = Bellemin et al. (1978); 9 = Lawson (1985). ferent electromorphs in Bothrops nummifer populations from the Caribbean and Pacific slopes of the central mountain chain in Costa Rica (Ji- menez-Porras, 1964b, 1967). Variations in fre- quencies of polymorphic amino acid oxidases and proteases distinguish populations of the fer-de- lance {Bothrops asper) from the two sides of these same mountains (Jimenez-Porras, 1964a). Geo- graphic differences in venom proteins are so com- mon that venomologists emphasize the impor- tance of preparing regional types of antisera (Goncalves & Vieira, 1950). The magnitude of protein diversity in snakes (table 1) is similar to that observed in many other vertebrates (Selander & Johnson, 1973; Nevo et al., 1984). In populations of snakes that have been examined, polymorphism, the frequency of poly- morphic loci relative to the number of loci ex- DESSAUER ET AL.: SNAKE EVOLUTION amined, ranges from 0.04 to 0.30; and heterozy- gosity, the frequency of heterozygous phenotypes per individual over the number of loci examined, ranges from 0.01 to 0.09 (table 1). When data are obtained from different sources, much caution must be observed in comparing levels of polymorphism or heterozygosity across populations or between species. These indices of genetic variability are highly correlated with both sample size and the number and kind of loci tested. Exclusion of al- leles found at less than the 5% level corrects for some disparities in sample size, but the inclusion or exclusion of such highly polymorphic loci as nonspecific esterases and transferrins can raise or lower these indices considerably (table 1; e.g., compare sets of data for Thamnophis sirtalis sam- ples from Manitoba, Canada). Genetic diversity is not uniform across subspecies ranges. Snakes in zones of secondary contact between populations that have been disjunct for some time may have high levels of heterozygosity and often exhibit rare alleles not found in other areas of the species' range. Populations of Nerodia f. fasciata from the pan- handle of Florida exemplify this phenomenon in snakes (Lawson, 1 985), as it has been documented in population studies of other reptiles (Case & Wil- liams, 1984; Murphy et al., 1984). The majority of evidence on protein diversity in snakes concerns members of the genus Tham- nophis, natricine snakes that are widely distrib- uted throughout North America (Conant, 1975; Stebbins, 1985). The common garter snake {Thamnophis sirtalis) is one of the most wide- spread species; its range is continuous from the Atlantic to the Pacific and extends from Mexico into northern Canada. In a broad sense, this species can be divided into two color morphs. Those with lateral red markings occupy the western part of the range, and those without laterally distributed red pigment, the eastern part. The red-sided garter snakes along the West Coast have been divided into several subspecies, but a single subspecies (Thamnophis s. parietalis) occupies the extensive area between the western coastal states and the Mississippi Valley. In the East the subspecies Thamnophis s. sirtalis ranges from the Atlantic Coast westward to a narrow zone of contact with '/'. s. parietalis that generally follows a north-south line approximating the Mississippi Valley. The only subspecies that has a disjunct range is Thamnophis s. dorsalis. which is isolated in the Rio Grande Valley. As an ecological generalist (Fitch, 1965), Tham- nophis sirtalis has been able to occupy a vast geo- graphic range and has adapted to all but the driest habitats. If the complicating influences of demo- graphic factors and history on the genetics of a population could be eliminated, natural selection suggests that ecological generalists should possess above average genetic diversity, especially those populations from continuous areas of the species range (Nevo et al., 1984). Populations of Tham- nophis sirtalis from inland areas do have hetero- zygosities of about 8%, relatively high for verte- brates (table 1 ; Thamnophis s. sirtalis from Illinois, Ohio, and Louisiana, and T. s. parietalis from Illinois). In contrast, populations at the geographic pe- riphery of a species' range, on islands, or in other distributional disjunctions might be expected to have lower levels of diversity. Although true for some populations, this hypothesis does not hold as a generalization for snakes. Bellemin and col- leagues (1978) examined intrademic variability in Thamnophis sirtalis collected as they emerged from each of four hibernacula near Inwood, Manitoba, Canada. Of the 1 5 loci assayed, only one, xanthine dehydrogenase, was variable, with heterozygosi- ties ranging from 1 . 1% to 2.8%. The authors posed two non-mutually exclusive hypotheses to explain these low indices: bottleneck effect due to periodic frost kills, and strong directional selection at the periphery of the species' range. While these factors may partially explain their results, choice of loci tested may be the more important factor. Lawson (1978; unpubl. data) has found that the 14 invar- iant loci of the Bellemin study are largely invariant throughout the range of this garter snake species. Transferrin and cytosolic superoxide dismutase were among 27 protein loci that Lawson examined in a Thamnophis sirtalis sample also taken near Inwood, although not necessarily from one of the same dens. The majority of the individual snakes were heterozygous at one or both of these loci. Thus, Thamnophis sirtalis from the vicinity of Inwood actually falls in the midrange for percent polymorphism and for percent heterozygosity, considering populations of T sirtalis as a whole (table 1). Probably because Thamnophis sirtalis is semi- aquatic and a feeding generalist (Fitch, 1 965; Kep- hart & Arnold, 1982), it is found on many coastal islands, both in eastern and western North Amer- ica. Continuous recruitment from the mainland to continental islands should be the rule, so reduced genetic variability due to founder effect is not ex- pected. Garter snakes from Islesboro Island in Pe- nobscot Bay, Maine, are as variable as those from 10 FIELDIANA: ZOOLOGY inland populations; however, the population on Vancouver Island off the coast of western Canada does appear to have a low level of genetic diversity (table 1) A variety of studies on snakes shows that mor- phological and protein polymorphisms are gen- erally inherited independently. In laboratory-bred kingsnakes (Lampropeltis getulus), color pattern is highly polymorphic (Zweifel, 1981), but these morphological features are inherited indepen- dently of protein polymorphisms (Dessauer & Zweifel, 1981; Zweifel & Dessauer, 1983). In Thamnophis ordinoides, marked variability in ground color and in the pattern and color of dorsal striping is accompanied by high molecular vari- ability (Lawson, 1978). On the other hand, Ne- rodia fasciata compressicauda, a snake of highly variable color pattern with melanistic and ery- thristic morphs in many populations, shows very low protein diversity (table 1 ). Differentiation through isolation by distance ap- pears to be a factor in producing and maintaining a number of color morphs in Thamnophis; many of these are recognized as subspecies. Differentia- tion that may in part be due to isolation by dis- tance can also be demonstrated at the molecular level in Thamnophis sirtalis. Populations along the northeastern and western coasts of North America are fixed for alternate alleles at the cy- tosolic superoxide dismutase locus (Lawson, 1978), but in all inland populations except those of the Florida Peninsula both alleles are found in ap- proximately equal frequencies (Sattler & Gun- man, 1976; Lawson, unpubl. data). Factors acting to maintain fixation of these different alleles in the Atlantic and Pacific coast populations of T. sir- talis are unknown. Sattler and Guttman (1976) used electrophoresis in an attempt to determine whether reproductive isolation or localized natural selection is responsible for the maintenance of high levels of melanism found in some populations of garter snakes in Ottawa County, Ohio. Allelic fre- quencies at 1 4 loci from melanistic and normally colored garter snakes collected near the town of Bono on the southwestern shore of Lake Erie sup- ported the view that these snakes are freely inter- breeding. Based upon this finding, Sattler and Guttman hypothesized that selection for conceal- ing coloration is responsible for the high frequency of the melanistic morph endemic to that general geographic area. The high levels of genetic variability at the mo- lecular level observed in Thamnophis sirtalis have not been found in all species of ecological gener- alists with wide geographical and alt i t udinal ranges (table 1). Populations of wide-ranging T elegans sampled at sea level and at 10,000 feet all have relatively low variability indices. The converse of the correlation of high genetic variability with eco- logical adaptability and extensive range is that the ecological specialist with a small distributional range should have low genetic variability. For some snake species this certainly holds true. Thamno- phis brachystoma and T. rufipunctatus each have very low indices of variability. However, not all specialization results in reduced genetic variabil- ity; the fossorial uropeltid Rhinophis phillippinus (Dessauer et al., 1976; unpubl. data) and the xe- nodontine Carphophis amoenus (Lawson & Ax- tell, unpubl. data) show average and higher than average genetic diversity, respectively. Overall, these observations suggest that any re- lationship between levels of genetic diversity and ecological specializations is extremely tenuous at best; they also serve to focus attention on popu- lation structure and history as major determinants of genetic variation. IV. Species Formation Studies at the protein level are giving insight into many problems concerned with the processes of speciation. If unique alleles distinguish two taxa, genotypes of individuals in the geographic area of contact may offer irrefutable evidence for the pres- ence or absence of gene flow between them. For example, such data suggest that the morphologi- cally polymorphic and wide-ranging kingsnake Lampropeltis getulus is a single species. Eastern and western populations have different albumin and haptoglobin phenotypes; but where the east- ern and western forms meet in western Texas in- tergradation is apparent, as proteins of both forms are present in snakes from that region (Dessauer & Pough, 1975). Additionally, Elaphe bairdii (see Olsen, 1977) and E. obsoleta are distinguished by unique esterase-D alleles as well as by frequency differences of alleles at other structural gene loci. Populations in a contact zone between the two forms include snakes heterozygous for the two marker esterase-D alleles (Lawson & Lieb, unpubl. data). The evolutionary biology of the North Ameri- can radiation of natricine snakes presents many speciation problems that have been most exten- sively studied in members of the Nerodia sipedon DESSAUER ET AL.: SNAKE EVOLUTION 11 complex and in various groups of Thamnophis. Three groups of water snakes have ranges that come into contact along the coastal region of southeastern United States: (1) Nerodia sipedon. adapted to freshwater streams; (2) Nerodia fascia- ta, adapted to other freshwater habitats; and (3) Nerodia fascial a clarkii and N.f. compressicauda, adapted to saline environments. Conant (1963) concluded that Nerodia sipedon and N. fasciata were distinct. Later investigators (Schwaner & Mount, 1976; Blaney & Blaney, 1979) interpreted color pattern similarities as signs of intergrada- t ion. Protein studies suggest that Conant probably was correct, at least for populations in contact zones along the Tchefuncte and Bogue Chitto rivers in southeastern Louisiana (Schwaner et al.. 1980). Similarly, the freshwater and salt marsh forms of Nerodia fasciata are largely reproductively iso- lated across their long but narrow zone of contact along the coasts of the Gulf of Mexico and the Atlantic Ocean (Lawson et al., 1981; Lawson, 1985). Unraveling the taxonomy of garter snakes of genus Thamnophis. the most speciose genus of North American natricines, has been especially baffling. The molecular evidence suggests that many members of the genus either have become reproductively isolated only recently or are pres- ently in final stages of speciation. Transferrins of members of the genus differ by a maximum ID of only 15 (George & Dessauer, 1970; Mao & Des- sauer, 1971); their albumins differ by a maximum ID of 10 (Dowling et al., 1983). The ribbon snakes, Thamnophis sauritus and T proximus, are sibling species. Thamnophis s. sau- ritus inhabits a large area, stretching from the east- ern coast of the United States westward to a line running northward approximately from the Pearl River through western Indiana. Along its western boundary it contacts T. proximus to produce a narrow zone of parapatry. The Florida Peninsula is inhabited by T. sauritus sackenii, which may intergrade with T. s. sauritus in the region of the former Suwanee Straits (Rossman, 1962). Thamnophis sauritus and T. proximus show lit- tle differentiation either morphologically (Ross- man, 1962) or at the molecular level (Gartside et al., 1 977). No marker alleles have been found, and the Nei genetic distance between them is only 0.023 (Lawson & Dessauer, 1979), a level usually indic- ative of conspecific populations. Extensive field observations and morphological evidence, how- ever, convinced Rossman (1962) that gene flow does not occur between these taxa; similarly, mor- phological evidence suggests that T. s. sauritus and T. s. sackenii may also be reproductively isolated (Williamson & Moulis, 1979). Additional insight into the evolution of these snakes, not apparent in the phenetic analysis of gene frequency data, is revealed by the distribution of a derived allele at the cytosolic malate dehy- drogenase locus, based upon combined data on 250 individuals (Gartside et al., 1977; Lawson, unpubl. data). There are two common alleles at this locus in the ribbon snakes. One, representing the derived state as determined by outgroup com- parison (Watrous & Wheeler, 1981), is apparently fixed in Thamnophis s. sauritus. The alternative, primitive allele predominates in T. proximus and T sauritus sackenii as well as in other species of Thamnophis. Although interbreeding between T. sauritus and T. proximus in the zone of parapatry probably no longer occurs (Rossman, 1962), evi- dence for its occurrence in the recent past is pro- vided by the presence of a step cline coincident with the zone of parapatry in Louisiana, showing penetration of the derived allele typical of T. s. sauritus into T. proximus populations as far west as western Texas. Similarly, gene flow from T. s. sauritus into T. s. sackenii populations has oc- curred with the derived malate dehydrogenase al- lele detectable in sackenii as far south as Tampa Bay in central Florida. Protein data are insufficient at present to estimate whether or not interbreeding is still taking place between these subspecies in the putative zone of intergradation. Relationships among the West Coast garter snakes present another problem that has chal- lenged herpetologists for many years (Rossman, 1979). Based on a series of classical studies, Fitch (1940) and Fox (1951) concluded that the many morphologically distinct forms comprised aquatic and terrestrially adapted groups of races of one species, Thamnophis elegans, distributed over most of California as a ring of races. Those subspecies adapted to aquatic conditions were thought to in- tergrade with subspecies adapted to terrestrial con- ditions along the Klamath River valley in northern California. Protein electrophoretic studies have modified our concept of relationships within the complex. Phenotypes for marker transferrin alleles showed that gene flow does not occur between forms of the two ecological groups in the area that Fitch and Fox proposed as the site of intergradation (Fox & Dessauer, 1965; Lawson & Dessauer, 1979; see 12 FIELDIANA: ZOOLOGY also the morphological studies of Rossman, 1964, 1979). A matrix of Nei genetic distances between subspecies showed that the terrestrial and aquatic groups of subspecies were members of relatively widely divergent lineages. The terrestrial lineage, Thamnophis elegans, was found to consist of four very closely related subspecies. The aquatic lin- eage split into two subgroups: the atratus subgroup, consisting of the subspecies atratus, hydrophilus, aquaticus, and gigas; and the couchii subgroup, consisting of the subspecies couchii and hammon- dii (Lawson & Dessauer, 1979). Relationships within the aquatic lineage have been further clarified by data on marker alleles for specimens collected from zones of contact between the subspecies. Apparently, gene flow in nature is rare between hydrophilus and couchii. Of several dozen snakes examined from the area of parapatry, only three individuals were identified as putative hybrids. Morphological studies on the same in- dividual snakes provided a similar interpretation (Rossman & Stewart, 1979). Moreover, couchii appears not to intergrade with hammondii where they contact in the Tehachapi Mountains, a con- clusion also drawn independently on morpholog- ical grounds by Rossman and Stewart (1982). In southwestern California where the ranges of the subspecies atratus and hammondii overlap exten- sively, occasional hybrids have been identified on both molecular and morphological criteria, again indicating that reproductive isolating mechanisms can break down occasionally between members of the aquatic subgroups. The California giant garter snake, T. gigas, was formerly thought to be de- rived from the couchii subgroup; however, protein evidence (Lawson & Dessauer, 1979; Lawson, un- publ. data) clearly shows that its affinities are with the atratus subgroup (recognition of gigas as a distinct species follows Rossman & Stewart, 1985). A third species of garter snake endemic to the west coast of North America, Thamnophis ordi- noides, has ecological preferences similar to coast- al populations of T. elegans. Phenetic analysis of protein evidence suggested that its affinities were with the atratus subgroup (Lawson & Dessauer, 1979); however, cladistic analysis of the same al- lelic data has since shown that T. ordinoides is instead closer to T. elegans (Lawson, unpubl. data). Thus, current evidence taken in toto suggests that the complex of West Coast garter snakes may con- sist of six closely related species: Thamnophis ele- gans, T. couchii, T. atratus, T. hammondii, T. gi- gas, and T. ordinoides. In general, genetic evidence suggests that many populations of natricine snakes, presently classi- fied as subspecies, are either already reproduc- tively isolated or have at least attained the "in- cipient" species level. Often these forms contact parapatrically without interbreeding. Each ap- pears to be adapted to some unique feature or features of the environment (e.g., fresh vs. salt water); selection countering gene flow appears to maintain these habitat distributions. Yet some forms are so similar genetically that it is easy to visualize how changes in the environment due to geological or climatic events or to man-induced disturbances could easily alter population equilib- ria. V. Differential Rates of Morphological Evolution A. Radiations Illustrating Rapid Morphological Evolution Many major lineages of snakes include one or more highly speciose radiations that appear to be relatively recent in origin. Such characterized groups have been identified in the natricine, col- ubrine, xenodontine, and elapid lineages. Each ra- diation includes species that are so distinct in mor- phology and/or ecology that they are classified in different genera. Yet the divergence of such rapidly evolving proteins as transferrin and albumin is so small that electrophoretically generated evidence on alleles has been required to assess affinities of members of each radiation. Examples from four well-studied groups are discussed here. 1. North American Natricines (fig. 3, top)— The tribe Thamnophiini (Rossman & Eberle, 1977), the natricine snakes of North America, is the most thoroughly studied of these radiations. Morphological divergence within the group is con- siderable; taxonomists recognize nine genera and about 45 species. These species are distributed from Central America to Canada and show terrestrial, semiarboreal, semiaquatic, aquatic, and semifos- sorial adaptations. The diets of the different species vary, encompassing invertebrates such as worms, slugs, and crayfish, as well as most classes of ver- tebrates (Wright & Wright, 1957). Molecular divergence among thamnophiines is low. IDs among the transferrins of 22 species ranged from 4 to 28, with an average of 1 1 (George, 1 969; DESSAUER ET AL.: SNAKE EVOLUTION 13 i H* d Old World Natricinae ' Virginia ttrlatula • Clonophls Mrtlandll • AdelophntoKi • Virginia valarlaa ' Saminatrix pygaaa ' Regina septemvittata 1 Nerodia sipedon • Thamnophis couchli Regina grahamii Tropidoclonion linaatum Regina rigida Regina all em Sloreria dekayi Storena occipitomaculata Storeria storarioidas ^ _rE Ariiona alagans Elaphe guttata Elaphe obsoleta Pituophis melanoleucus 1 Lampropeltis triangulum 1 Elaphe vulpina ■ Elaphe bairdi ' Lampropeltis getulus ■ Lampropeltis pyromalana Elaphe rosallaa Elaphe suboculans Elaphe triaspis Lampropeltis calligaster Rhinocheilus lecontei Cemophora coccinaa Stilosoma extenuatum Elaphe quatuorlinaata Fici. 6. Genetic differentiation among species within North American colubrid radiations. These are UPGMA phonograms (Sneath &. Sokal, 1973) clustering Nei's unbiased genetic distances (Nei, 1978), which appear on the scale associated with each phcnogram. These diagrams show only the relative degree of genetic differentiation among taxa and should not be interpreted as phylogenetic trees. Top, The New World natricine (Thamnophiini) radiation (Lawson. 1 985); bottom, species of a North American colubrine radiation, along with Old World Elaphe quatuorlineata (Lawson St Dessauer. 1981). 14 FIELDIANA: ZOOLOGY Mao & Dessauer, 1971; Schwaner & Dessauer, 1 982); IDs for albumins of 2 1 species ranged from 1 to 19,* with an average of 7 (Dowling et al., 1983). Chromosomal morphology of the different species is very similar (Baker et al., 1972; Eberle, 1972; Rossman & Eberle, 1977). The phenogram (fig. 6, top) showing relative degrees of genetic dif- ferentiation among the Thamnophiini is based upon an electrophoretic survey of 27 loci and also indicates the low degree of molecular divergence within this group. All species cluster within a Rog- ers's genetic distance of about 0.4 (Lawson, 1 985). 2. North American Colubrines (fig. 6, bottom)— Molecular evidence is most extensive regarding the group that includes Lampropeltis. Minton and Salanitro (1972), using antiserum to plasma proteins of Elaphe vulpina, were unable to distinguish immuno-electrophoretic patterns of plasma proteins of Elaphe vulpina, E. obsoleta, E. guttata, Pituophis melanoleucus, Lampropeltis ge- tulus, and L. calligaster. Immunodiffusion com- parisons suggest that transferrins of species of these genera have IDs of less than 30 when compared to the transferrin of Elaphe obsoleta (Schwaner & Dessauer, 1 982; Lawson & Dessauer, unpubl. data). Most MC'F IDs for albumins in these taxa, ob- tained with antisera raised to albumins of Elaphe obsoleta and Lampropeltis getulus, are less than 20 (Dowling et al., 1983). The organisms within this radiation are so close genetically that electro- phoretic evidence was needed to assess affinities of individual species (fig. 6, bottom; Lawson & Dessauer, 1981, unpubl. data). All North American colubrines examined, with the exception of Elaphe subocularis, have very similar karyotypes (Baker et al., 1972; Bury et al., 1970); even the unique karyotype ofE. subocularis may be derived from the common colubrine pat- tern (Baker et al., 1971). Numerous instances of hybridization between New World Elaphe guttata and E. obsoleta in captivity and in the wild have been recorded (see Neill, 1949; Mertens, 1950; Lederer, 1950). On the other hand, an interspecific mating between E. obsoleta and Old World E. schrenckii produced an inviable clutch (Broer, 1978). 3. Xenodontines (fig. 7)— Two groups of xe- nodontine snakes appear to comprise relatively recent radiations. One, the pseudoboines, com- prising eight genera (sensu Bailey, 1967) and here excluding Saphenophis and Tropidodryas as pro- * Exclusive of Thamnophis mendax. posed by Jenner and Dowling (1985), is molecu- larly the most cohesive and geographically one of the most widespread groups of South American xenodontines (Cadle, 1984a). Immunological comparisons of albumins and transferrins and mul ti locus electrophoretic comparisons (fig. 7, top) suggest that these snakes share a long period of common ancestry relative to other South Ameri- can xenodontine genera. Despite the recent sepa- ration among these genera, they have radiated into habitats ranging from rain forests to savannas and deserts; representatives of four genera, Clelia, Ox- yrhopus, Tripanurgos, and Siphlophis, have dis- persed from South America into Central America (Cadle, 1985). This considerable geographic and habitat distribution has been achieved without ex- tensive speciation (approximately 25 to 30 species among eight genera). There is some morphological diversity in body size and form, dentition, and skull structure (Bailey, 1939, 1967). A different situation (fig. 7, bottom) is found in Central American xenodontines in which four "dipsadine" genera, Dipsas, Sibon, Tropidodipsas, and Sibynomorphus, plus Ninia and Geophis, form a closely related clade (Cadle, 1 984b); all albumin IDs are less than 25. For such closely related gen- era, and in contrast to the pseudoboines, this group is remarkable for its diversity of morphological specializations: most Geophis species are modified for a fossorial existence, whereas Ninia, Sibyno- morphus, and some Tropidodipsas are terrestrial, and Sibon and Dipsas are arboreal specialists. Dipsadines have developed various specializa- tions related to their gastropod-feeding habits. For reviews of these feeding and habitat specializa- tions see Downs (1967) and Peters (1960). This group provides the best example to date among snakes of the extraordinary morphological changes that may accrue with little molecular change among species. Indeed, the morphological specializations found in more highly modified species of Dipsas are extreme for snakes, and their origin has long been recognized as an intriguing evolutionary problem (Dunn, 1951). In addition, this group is very speciose (approximately 1 00 species) and has an extensive geographic distribution, encompass- ing the entire range of the Central American xe- nodontines (Cadle, 1985). From the biochemical data we may infer that the fossorial, arboreal, and trophic specializations of this group probably have arisen within the last eight to 15 million years (Cadle, 1982a). 4. Australian Elapids/Hydrophiids— Rapid morphological evolution has characterized the DESSAUER ET AL.: SNAKE EVOLUTION 15 Clelia tcyialtna Pseudoboa neuwiedn Pseudoboa nigra Pseudoboa coronata Clelia clelia Clelia occipitolutea Clelia rustica Oxyrhoput tor mot ut Oxyrhopus fitzingen Oxyrhoput petola Oxyrhoput melanogenys Oxyrhoput trigeminus Trlpanurgot comprettut Hehcops pattazae Helicops angulalus Tropidodiptat tarlorl Dip tat catetbyi Sibynomorphut turgidut Sibon nebulata Sibon annulala Fig. 7. Genetic differentiation among species within two xenodontine radiations. These are phenograms prepared as in Figure 3. Top, The South American pseudoboine radiation (Cadle & Dcssaucr, 1985); bottom, South and Central American dipsadines (Lawson, unpubl. data; see also Cadle, 1984b). elapid snakes of Australia. Their radiation appears to have followed an invasion of precursors from Asia, perhaps beginning in the middle Miocene. Within Australia, species in 16 genera related to the tiger snake genus Notechis appear to comprise a remarkable radiation of even more recent origin. Immunological distances among transferrins of species within the group range between two and 20. As compared to the transferrin of Notechis, those of A us tr claps, Echiopsis. Hemiaspis, Hoplo- cephalus. Suta. Tropidechis, and Unechis have IDs of less than 10, differences close to the limit of sensitivity of the MC'F method and indistinguish- able by immunodiffusion analysis. If species with transferrin IDs below 20 are included, members of Cryptophis, Furina. Parademansia, Simoselaps, Vermicella. and some species of Drysdalia and Denisonia also belong to the Notechis radiation (Schwaner et al., 1 985). Branching sequences based upon preliminary electrophoretic analysis of tissue proteins are broadly concordant with the trans- ferrin evidence (Mengden, 1985a). The close re- lationship of Denisonia and Notechis was sug- gested by Kellaway and Williams ( 1 93 1 ) in one of the first comparative immunological studies. Di- vergence in karyology (Mengden, 1985a), behav- ior, ecology (Schwaner, 1985), and external mor- phology is very great within the presumptive tiger snake radiation. Terrestrial forms alone are pres- ently classified as 32 species in 1 5 genera (Cogger, 1975), exclusive of the species of Denisonia and Drysdalia that have transferrin IDs above 20. The protein evidence also shows that sea snakes, which are generally classified either as a subfamily 16 FIELDIANA: ZOOLOGY of the Elapidae or as a distinct family, the Hydro- phiidae, are members of the Australian radiation of elapids (Minton & Da Costa, 1975; Minton, 1978; Cadle & Gorman, 1981; Mao et al., 1983), with at least some species possibly being members of the tiger snake group. The transferrin IDs be- tween Notechis and sea snakes of two of the most speciose genera, Aipysurus and Hydrophis, are less than 20. The low albumin and transferrin IDs be- tween sea snakes and Australian terrestrial elapids suggest that the morphological diversity within and between these two groups has arisen rapidly in geological time (Schwaner et al., 1985). B. Taxa Illustrating Convergence or Slow Morphological Evolution In contrast to those snake lineages showing great degrees of morphological differentiation relative to molecular divergence, others provide examples of conservative morphological evolution or ho- moplasy. We do not distinguish the latter two pro- cesses here, since more detailed phylogenetic hy- potheses are required to assess their relevance. Our examples derive from the use of molecular data to evaluate systematic arrangements of particular snake taxa. Most cases involve genera that were traditionally considered monophyletic, but whose para- or polyphyletic nature was demonstrated by biochemical data. Presumably, convergence or re- tention of primitive character states were respon- sible for the inability of classical taxonomic pro- cedures to partition these genera into natural units. The most thoroughly documented example of the application of protein taxonomy to such a problem concerns the species composition of nominal genus Natrix. Prior to Malnate's (1960) study, 86 species were included in what he called this "unwieldly and confusing assemblage." Al- though external morphology of these snakes was too similar to classify them effectively, by focusing attention on tooth and hemipenial structures Mal- nate was able to partition the 86 species into five genera. With 26 species retained in Natrix (sensu Malnate, 1 960), the genus still included snakes of North America, Asia, Europe, and Africa. Immunological comparisons of transferrins proved that the placement of species from these different regions in the same genus was artificial. Immunological distances between the transferrins of species from the four regions were relatively great, ranging between 40 and 6 1 . In contrast, those between species from the same region averaged about 10 and ranged from 2 to 23. Transferrins of the North American species, in fact, were more similar to those of other North American genera of natricines than to those of Natrix from other continents (George & Dessauer, 1 970; Mao & Des- sauer, 1971; Gartside & Dessauer, 1 977; Schwaner & Dessauer, 1982). Based upon the transferrin evidence, serological analyses of unfractionated plasma proteins (Pear- son, 1966; Minton, 1976), karyological findings (Buryetal., 1970; Baker etal., 1972;Eberle, 1972), and several sets of morphological characters, Rossman and Eberle (1977) repartitioned the as- semblage, retaining species from Europe and North Africa in Natrix, placing those from North Amer- ica in Nerodia as a member of the tribe Tham- nophiini, those from Asia in Sinonatrix, and those from Africa south of the Sahara in Afronatrix. Al- bumin immunological evidence (Dowling et al., 1983) and electrophoretic evidence on numerous proteins (Lawson, 1985, fig. 2a) have furnished additional support for considering thamnophiine snakes as a natural group. Comparative protein studies suggest that the colubrine genus Elaphe is also not monophyletic. Currently, the more than 50 species assigned to the genus are found in North America, Europe, Asia, and the East Indies. Antiserum to plasma proteins of North American Elaphe obsoleta reacts more strongly with sera of other North American colubrine genera than with sera of Eurasian con- geners (Minton, 1976). Immunological distances between the albumins of the different North Amer- ican species of Elaphe and the North American species of Cemophora, Lampropeltis, and Pituo- phis averaged 16, whereas the ID between the North American Elaphe and E. radiata of South- east Asia equalled 50 (Dowling et al., 1983). Im- munodiffusion analyses, using antisera to trans- ferrins of reference species from North America, Europe, and Asia, also attest to the wide diver- gence of forms from the different zoogeographical regions. Large spurs formed in cross-reactions in- volving serum and antiserum samples of species from the different regions (Lawson & Dessauer, unpubl. data). Electrophoretic evidence on proteins deter- mined by 1 5 structural gene loci also attest to the non-monophyletic nature of Elaphe, in concord- ance with the immunological findings. Genetic distances distinguishing North American species of Elaphe and other North American colubrine genera are usually less than those distinguishing New World and Old World Elaphe (see fig. 6, DESSAUER ET AL.: SNAKE EVOLUTION 17 bottom). Nei genetic distances between Elaphe from different geographic regions ranged from 0.24 to 0.78, the Asian E. moeUendorffi being the least distinct and E. scalaris of Southern Europe the most distinct from E. obsoleta (Lawson & Dcs- sauer, 1981; unpubl. data). The genus Coluber is another unnatural assem- blage of New and Old World species that has de- fied partition. The external morphological char- acters used traditionally in classification are either too uniform or too variable to offer useful char- acter states for developing a more natural classi- fication. Schatti (198S), in a preliminary report, noted that his observations on osteology, anato- my, and protein electrophoresis support the sep- aration and distinction of Old and New World species and clearly demonstrate the polyphyletic nature of Palearctic Coluber. Serological analyses also show that Coluber constrictor of North Amer- ica is more closely related to the North American colubrine genera Masticophis and Drymarchon than to Coluber jugularis of Israel (Minton, 1976). Other examples where molecular data have sug- gested or confirmed the polyphyletic nature of gen- era include Rhadinaea. a widespread and speciose genus of Neotropical xenodontines. Using albu- min immunological comparisons, Cadle (1984b) showed that members of the R. brevirostris species group were derived from South American xeno- dontines, whereas other species of Rhadinaea stem from a Central American stock. This provided strong evidence for earlier suspicions of their in- dependent origins based upon morphological evi- dence (Myers, 1 974). Molecular data also suggest that several mor- phologically similar genera of the xenodontines are not close relatives. These include the rather distant relationship of (icophis to A tract us, despite considerable convergence in morphological fea- tures related to fossoriality (Downs, 1967; Cadle, 1984b), and the very distant relationship between Heterodon and Xenodon (Cadle, 1984a; unpubl. data), which are also similar in many aspects of morphology (e.g.. Weaver, 1965). C. Implications of Differential Rates of Evolution in Snakes The examples discussed indicate that lineages of snakes vary considerably in rates of speciation and morphological evolution. Conservatively, we estimate that the lineages discussed here arose in the Middle Miocene or later, with much of the evolution within groups such as the thamno- phiines and Australian elapid/sea snake radiations having occurred since the late Miocene. The rea- sons why particular lineages may show either rapid morphological evolution or stasis has been an im- portant problem in evolutionary biology (Simp- son, 19S3). Rapid morphological evolution has usually been attributed to alterations in the control of gene expression or to changes in the sequence or timing of developmental events (Wilson et al., 1977; Alberch et al., 1979). The incorporation of changes produced by these mechanisms is influ- enced by selective pressures and aspects of pop- ulation structure and history (e.g., see Larson, 1 984). There has been little investigation of these parameters in snakes (but see Haluska & Alberch, 1983, for a possible example of heterochronic change). A fruitful area for future research on mechanisms of evolutionary change in snakes will be to analyze the distribution and developmental basis of morphological features in lineages for which detailed phylogenetic data are available. In some groups of organisms, rates of chro- mosomal evolution are correlated with rates of speciation and morphological evolution (Wilson et al., 1977; Larson et al., 1984). This correlation appears to be only weakly supported in snakes. Although karyotypic evolution is slow in snakes as compared to many other vertebrates (Wilson et al., 1975), there is some variability in rates among snakes. The remarkable diversity of karyotypes observed among species of the Australian elapid radiation is accompanied by high rates of specia- tion and morphological evolution (Mengden, 1985a,b). On the other hand, karyotypes within North American natricines and colubrines, re- spectively, are extremely uniform (Baker et al., 1972; Eberle, 1972; Rossman & Eberle, 1977) de- spite a diversity of species and morphological types. Thus, superficially, there appears to be only a ten- uous association between gross karyotypic evo- lution and evolution at the species level. Finer resolution of chromosomal structure and broader sampling among lineages will be necessary to eval- uate this association more fully. Based on the few lineages that have been exten- sively studied biochemically, rapid morphological divergence appears to occur commonly in snakes. Examples such as the Australian terrestrial elap- ids/sea snake radiation and the Central American genera allied to Dipsas demonstrate that dramatic trophic and habitat adaptations may occur among closely related forms during relatively brief pe- riods of evolutionary time. This observation is 18 FIELDIANA: ZOOLOGY perhaps one reason why it has proven so difficult to estimate phylogenetic relationships among snakes; derived characters linking various taxa may be transformed rapidly into new states. Without detailed knowledge of character state transfor- mations, which depends on an estimated phytog- eny (see Lauder, 1981; Alberch, 1985), the use of such characters in phylogenetic reconstruction is difficult. Compounding the difficulties is an ap- parently high degree of homoplasy in snake mor- phology (Cadle, 1982a). We believe that biochemical evidence, although not free of interpretative problems, can circum- vent some of the difficulties inherent in the use of morphological data to reconstruct snake phylog- eny. Ultimately, of course, any worthwhile phy- logenetic hypopthesis must be evaluated with re- spect to all comparative data. For the remainder of this paper we discuss the comparative biochem- ical data bearing on snake systematics and phy- logeny above the generic level. VI. Higher Levels of Relationship A. Position of Snakes Among Reptiles Biochemical data support the traditional view that lizards and snakes, comprising the Order Squamata, are closest relatives among extant rep- tiles (see Dessauer, 1974). Serologists since Gra- ham-Smith (1904) have obtained weak immu- nological cross-reactions in tests involving proteins from lizards and snakes but little or no cross-re- action in tests involving proteins of members of the Squamata and other orders of reptiles. Using MC'F, Gorman and colleagues (1971) demon- strated that heart lactate dehydrogenases of lizards and snakes were much more similar than were the lactate dehydrogenases of lizards compared to those of Sphenodon, crocodilians, turtles, or birds. Sim- ilarly, fingerprints of tryptic peptides of the hemo- globins of snakes and some lizards (e.g., Iguana) share numerous similarities, whereas few com- parable peptide fragments of snake hemoglobins are detectable on hemoglobin fingerprints for croc- odilians or chelonians (Sutton, 1969; Dessauer, 1974). Qualitative differences in the metabolic pathways involved in bile acid synthesis (Hasle- wood, 1978; Tammar, 1974) and nitrogen metab- olism (Cohen & Brown, 1 960) also distinguish the Squamata from other reptilian orders. Although biochemical studies show that snakes and lizards are more closely related to each other than to other reptiles, these studies have contrib- uted little evidence on the precise relationship among lizards and snakes (Dessauer, 1974). That is, does the divergence between lizards and snakes predate the separation of extant lineages within either of these groups, or are snakes derived from a particular lineage of lizards (e.g., the Angui- morpha; McDowell & Bogert, 1954)? Using anti- sera to snake albumins we have recently compared various lizard albumins in enhanced Ouchterlony double-diffusion tests (see sec. II). Strong cross- reactions were obtained with albumins of iguanids (fig. 8, sample Cr), anguids, amphisbaenids, and Heloderma (fig. 8, sample H); weaker reactions with albumins of agamids and teiids; and no re- action with albumins of Varanus (fig. 8, sample V), skinks, xantusiids, cordylids, chamaeleonids, pygopodids, or gekkonids. Because of the lack of rate tests for these albumins, results such as the strong reactions for Gerrhonotus and Heloderma and no reaction for Varanus (fig. 8, middle), all of which are anguimorphs, are difficult to interpret at present. The differential reaction of snake al- bumins with various lizard groups could reflect either variations in rates of albumin evolution among groups, or differences in their phylogenetic relationships. Because it is possible to obtain such cross-reactions, immunological and other molec- ular methods promise to offer valuable insights on the relationships of snakes and lizards. Although the evidence is not clear on the rela- tive placement of lizards and snakes, comparable molecular evidence supports the monophyletic status of snakes. In cross-reactions involving anti- sera to albumins of snakes, such as those for Lep- totyphlops and Boa (fig. 8, samples L and B), pre- cipitin arcs for snake albumins spur over reactions for lizard albumins, showing albumins of snakes to be more similar to each other than to albumins of lizards. This is true even for Typhlops, which McDowell and Bogert (1954) have considered to be a lineage distinct from snakes (fig. 8, sample T over Cr). B. Relationships Within and Between Major Groups of Snakes Snakes are divided into two major monophy- letic groups, the Scolecophidia and Alethinophidia (McDowell, 1974; Rieppel, 1979a). The latter group includes both the Henophidia and the Cae- nophidia of Underwood (1967). Beyond this area DESSAUER ET AL.: SNAKE EVOLUTION 19 Fig. 8. Enhanced Ouchtcrlony double-diffusion tests of albumins. Central wells contain antiserum to albu- mins of: top, Leptotyphlops humilis; center. Boa con- strictor, and bottom, Rhinophis phUlippinus. Peripheral wells contain plasma of: L - Leptotyphlops humilis; T - Typhlops (Rhamphotyphlops) braminus; Cr - Cro- taphytus collarisr, Ag - Agkistrodon bilineatus; H - Hel- oderma suspectum; V - Varanus varius; G - Gerrhon- otus multtcarinatus; R - Rhinophis phUlippinus; P - Pseudotyphlops phUlippinus; Cy - Cylindrophis rufus; B - Boa constrictor. An - Anilius scytalr, U - Uropeltis liura. of comparative agreement, taxonomists have many different opinions concerning further taxonomic subdivisions, as well as on the generic composition and inter-relationships of major subdivisions. The monophyly of the Henophidia has also been ques- tioned (see Groombridge, 1979a; McDowell, 1987; Cadle. 1987). Protein studies are offering new in- sights on such problems. Definitive evidence on xenodontine, natricine, and colubrine snakes il- lustrates the potential of molecular approaches for studies at the suprageneric level. Less conclusive but suggestive molecular data are also becoming available regarding affinities within and between other groups of advanced snakes and lineages of primitive snakes. Many major unsolved problems in snake phy- togeny concern the relationships among major clades. Although the molecular data available at this point are capable of resolving more recent separations among genera and, in some cases, subfamilial groups, there is a paucity of evidence bearing on the branching order of more ancient separations. At that level the resolving power of the most widely used molecular techniques is lim- ited because few shared derived states at the amino acid level are likely to be conserved for such long periods of evolutionary time (Sarich, 1985). For these reasons, we feel that it is premature to pre- sent a phylogenetic tree estimating relationships among major clades. Concerted efforts are cur- rently underway to gather such data. Thus, the discussions that follow concentrate on relation- ships within major clades and reflect our current interpretation of data bearing on intergroup rela- tionships. Immunological evidence on plasma proteins shows that the Scolecophidia (blind snakes), Hen- ophidia (primitive snakes), and Caenophidia (ad- vanced snakes) are the result of ancient radiations. Precipitin tests involving plasma proteins show low levels of cross-reactivity when antigens from a species of one infraorder are tested with anti- bodies raised to the plasma proteins of a member of a different infraorder (Graham-Smith, 1904; Pearson, 1968; Cadle, 1982a; Schwaner & Des- sauer, 1982). In enhanced Ouchterlony tests using antisera to the albumins of Boa and Rhinophis (Henophidia), weak reactions were obtained with albumins of Leptotyphlops (Scolecophidia; fig. 8, sample L), Boa and Cylindrophis (fig. 8, samples B and Cy), and Agkistrodon (Caenophidia; fig. 8, sample Ag). Quantitative precipitin tests also sug- gest that the Henophidia (represented by Boa and Python) are more closely allied to the Caenophidia 20 FIELDIANA: ZOOLOGY than to the Scolecophidia (Pearson, 1968). Trans- ferrins of species of the three infraorders are so different that they do not form precipitin lines in immunodiffusion tests with antibodies raised to the transferrin of a member of another infraorder (Schwaner & Dessauer, 1982). 1. Scolecophidia— Comparative immunolog- ical evidence suggests that the Leptotyphlopidae and Typhlopidae form a lineage relative to other snakes, consistent with their grouping in the Sco- lecophidia. Pearson ( 1 968), with whole-serum pre- cipitin tests, found that proteins of the Typhlopi- dae have only weak affinities to those of boids, pythons, colubrids, and viperids. Antiserum to Leptotyphlops albumin in enhanced Ouchterlony immunodiffusion tests, reacts strongly with albu- min of Typhlops, but yields only weak reactions with albumins of other snakes (fig. 8, top). In all comparisons of Leptotyphlops albumin with hen- ophidian albumins, the Typhlops precipitin arc spurs over that for the henophidians. The precip- itin line for Leptotyphlops albumin also spurs strongly over the arc for Typhlops albumin (fig. 8, top), illustrating that within the Scolecophidia the molecular divergence between the Typhlopidae and Leptotyphlopidae is substantial. 2. Henophidia— a. Boidae/Pythonidae/Tropi- dophiidae — Immunological comparisons of plas- ma proteins suggest that the Henophidia is a com- plex of ancient lineages. Immunological distances between albumins of a selection of henophidians, obtained with antiserum to the albumin of Boa, are given in Table 2. Rather than interpreting these data phylogenetically, as they have not been rate tested, we simply make the following observa- tions: ( 1 ) albumin IDs within this group approach the technical limits of the MC'F technique, im- plying very ancient, perhaps Cretaceous, separa- tions between several lineages; and (2) relative to Boa, several presumptive associations do not re- flect the current systematic arrangement of these taxa. For example, the MC'F data suggest that the albumins of Boa and Exiliboa are more similar to each other than are those of Boa and Tropidophis. Exiliboa and Tropidophis are currently placed to- gether in the family Tropidophiidae (McDowell, 1 975). Also, the albumins of Cylindrophis and Boa are more similar to each other than are those of Boa and Python. Cylindrophis is classified either in the Aniliidae (Underwood, 1967) or the Uro- peltidae (McDowell, 1987; Rieppel, 1979b), whereas Boa and Python are usually placed to- gether in the Boidae (Underwood, 1967; but see Groombridge, 1979b, and McDowell, 1979, 1987). Table 2. Immunological distances between the al- bumin of Boa constrictor and other henophidian albu- mins. The classification follows McDowell (1987). Albumin Antiserum to Boa albumin Boidae (Booidea) Boa constrictor Epicrates cenchria Corallus caninus 0 37 58 Python idae (Booidea) Python molurus 135 Tropidophiidae (Tropidophoidea) Exiliboa plicata Tropidophis greenwayi Tropidophis haetianus 54 126 113 Uropeltidae (Anilioidea) Cylindrophis rufus 94 Loxocemidae (Anilioidea) Loxocemus bicolor 169 Acrochordidae (Acrochordoidea) Acrochordus javanicus 152 Precipitin tests involving plasma proteins (Pear- son, 1966, 1968) and immunodiffusion studies of transferrins and unfractionated plasma proteins (Schwaner & Dessauer, 1981) also document the wide separation of boas and pythons. Both albu- min IDs (table 2) and immunodiffusion evidence on plasma proteins show that the New World gen- era Epicrates, Corallus, Charina, and Lichanura, as well as Candoia of New Guinea, group closer to Boa than to Python. Candoia, however, is widely divergent from the New World boas (Schwaner & Dessauer, 1981). b. Uropeltidae/ Aniliidae — Molecular studies are supplying evidence on affinities of the Uropelti- dae. Tests with antisera to albumins of Rhinophis and Cylindrophis appear to confirm a distant as- sociation of these genera, a conclusion consistent with some recent morphological evidence (Riep- pel, 1979b; McDowell, 1987). Electrophoretic pat- terns of tissue proteins (Dessauer et al., 1976) and comparisons with these antisera also suggest that the uropeltids are a rather compact radiation. Im- munodiffusion reactions with the Rhinophis an- tialbumin yield at most only weak spurs in cross- reactions with albumins of species of Rhinophis, Pseudotyphlops, and Uropeltis (fig. 8, samples R, P, and U). In contrast, the albumin of the aniliid genus Anilius reacts only weakly with the antial- bumins of both Rhinophis (fig. 8, sample An) and Cylindrophis. Although the biochemical evidence suggests that Anilius is widely divergent from the DESSAUER ET AL.: SNAKE EVOLUTION 21 uropeltids, it is not yet clear whether or not Anilius shares a common lineage with the uropeltids rel- ative to other primitive snakes. c. Acrochordidae — Immunological compari- sons confirm the distinctness of Acrochordus rel- ative to other snakes, although current data offer no insight on whether or not Acrochordus is a sister group of colubroids (Groombridge, 1979a,b; Riep- pel, 1979a). Using an antiserum to Acrochordus albumin, Agkistrodon (which has a conservative albumin) spurs over albumins of some henophidi- ans (e.g.. Python, Tropidophis, Loxocemus) but not others (Boa, Cylindrophis). The precise placement of Acrochordus among these lineages can be as- certained once relationships among henophidian lineages are better understood. We specifically re- ject hypotheses associating Acrochordus with either natricine or homalopsine colubrids (e.g., Dowling & Duellman, 1978; Dowling et al., 1983). In tests with antisera to transferrins of natricine and col- ubrine snakes, transferrins of Acrochordus gave MC'F IDs greater than 1 1 5 (George & Dessauer, 1970) and produced no detectable precipitin bands in Ouchterlony immunodiffusion analyses (Schwaner & Dessauer, 1982). Only weak im- munodiffusion reactions were detectable between albumins of Acrochordus and antisera to albumins of species of natricines and homalopsines. In en- hanced Ouchterlony tests using antiserum to Ac- rochordus albumin, only weak reactions were ob- tained with albumins of Rhabdophis (Natricinae) and Homalopsis (Homalopsinae), and the precip- itin arc for Boa albumin spurred over both of these. d. Conclusions— Collectively, the molecular evidence shows that the Henophidia is composed of a number of widely divergent lineages. Even a cautious interpretation of the data reveals several conclusions that are not concordant with current classifications. For example, the very large molec- ular distances between boas and pythons and rel- atively less between boas and some aniliids (Cy- lindrophis) are not predicted by most classifications. There exists much disagreement among taxono- mists concerning the definition and composition of taxa within the Henophidia (see Rieppel, 1977, 1979a; Groombridge, 1979b; McDowell, 1987). Groups such as the Boidae of Underwood (1967, 1978) are placed together because of primitive morphological features. As knowledge of molec- ular evolution in these primitive snake groups in- creases, we expect that a revaluation of many accepted phylogcnetic hypotheses for the heno- phidians will be necessary and that this group will be recognized as a paraphyletic taxon, as suggested by Groombridge (1979b). 3. Caenophidia— Molecular comparisons in- volving all groups of advanced snakes (Viperidae, Elapidae, Atractaspis. and Colubridae) include whole-serum precipitin studies of Graham-Smith (1904) and Pearson (1966, 1968), MC'F compar- isons of albumins (Cadle, 1982a,b), and an im- munodiffusion survey of transferrins (Schwaner & Dessauer, 1982). These studies corroborate the monophyly of the colubroids and suggest that the Viperidae is the sister group of the three other clades (Cadle, 1982a; unpubl. data). In the follow- ing sections we discuss molecular data bearing on relationships within each of these groups. a. Viperidae— Biochemical evidence on viperid relationships includes an immunoelectrophoretic study of venom proteins (Detrait & Saint-Girons, 1979), observations on bile acid synthetic path- ways (Haslewood, 1978;Tammar, 1974), and im- munological comparisons of plasma proteins (Ku- wajima, 1 953) and albumins (table 3). Current data are consistent with the view that the Viperinae and Crotalinae are monophyletic sister groups (Liem et al., 1971; Groombridge, 1979b, 1984). These two groups can be distinguished on the basis of bile acid synthetic pathways (Tammar, 1974) in which crotalines show a pattern common to many advanced snakes, whereas viperines are characterized by acids almost restricted to this group. The interpretation of the albumin immunolog- ical evidence (table 3) is complicated by the fact that the rate of albumin evolution appears to be variable within vipers. For example, relative to the albumin of an outgroup represented by Boa, the albumin of Bids has changed about 34 ID units more than that of Crotalus (table 3). Differential rates of albumin evolution are also evident in pre- cipitin tests involving unfractionated serum, in which albumin-antibody complexes compose much of the precipitate (Pearson, 1966, 1968), and in MC'F tests using another outgroup, Atrac- taspis (table 4). Using the rate test, most of the albumin divergence between Bit is and Crotalus is attributable to the Bitis lineage. Not only do rates of albumin evolution appear to be variable within vipers, but as a group they have more conservative albumins than elapids, colubrids, and Atractaspis (table 3; Cadle, 1982a). Additional reciprocal comparisons and rate tests for albumins from a wider selection of vipers are needed to support statements on intraviperid re- 22 FIELDIANA: ZOOLOGY Table 3. Immunological distances and rate tests concerned with the albumins of the Viperidae. The rate tests for advanced snake albumins used an antiserum to Boa albumin as an outgroup and are expressed as immunological distances relative to Crotalus enyo = 0. Note the marked conservatism shown by viperid albumins relative to those of elapids and colubrids. Antisera Crotalinae Viperinae Crotalus Bothrops Bitis Albumins albumin albumin albumin Rate tests Reciprocal Comparisons Crotalus enyo 0 22 66 0 Bothrops atrox 24 0 76 - 2 Bitis nasicornis 72 79 0 + 34 Unidirectional Comparisons Viperinae . . . Bitis arietans . . . 77 70 Vipera aspis 53 77 + 25 Vipera palestinae . . . 62 67 Causus resimus 64 62 89 Causus maculatus ... 75 HI Echis ocellatus 47 51 70 + 16 Echis coloratus . . . 44 76 + 14 Pseudocerastes fieldii . . . 38 47 Cerastes cerastes 40 74 Crotalinae Lachesis muta 30 32 92 + 4 Sistrurus catenatus 16 20 83 Sistrurus miliarius 33 96 + 8 Agkistrodon piscivorus 28 93 Agkistrodon contortrix 34 Agkistrodon bilineatus 39 98 + 19 Elapidae (5 species) + 86 ± 7 Colubridae (7 species) + 65+16 lationships. It is apparent, however, from avail- able comparisons that within crotalines Sistrurus and Crotalus are close relatives, and that diver- gence among the extant genera began early in the history of the lineage. Although the phylogenetic interpretation is not clear, Detrait and Saint-Gi- rons ( 1 979) determined that venoms of the African viperine genera Bitis, Echis, and Cerastes share more antigens with each other than with European Vipera, suggesting that there may be a basic phy- letic separation between African and Eurasian vi- perine genera (Cadle, 1 987). Without comparisons using additional antisera and rate testing, we can- not evaluate this hypothesis with respect to current albumin data (table 3). b. Elapidae — Molecular evidence on elapid phylogeny comes from a variety of sources, in- cluding MC'F and immunodiffusion comparisons of transferrins (Mao et al., 1977; Schwaner et al., 1985), MC'F comparisons of albumins (Cadle & Sarich, 1981; Cadle & Gorman, 1981; Mao et al., 1 983), peptide fingerprinting of hemoglobins (Mao et al., 1978, 1984), and venom protein sequences and antigenic structure (Strydom, 1973, 1979; Minton & Da Costa, 1975;Hseuetal., 1977; Coul- ter etal., 1981;Dufton, 1984;Tamiya, 1985; Mebs, 1 985). Most of these studies have addressed ques- tions concerning the phylogenetic position of sea snakes among elapids, and the relationships of the endemic Australian forms. The extensive body of albumin and transferrin immunological comparisons and studies of venom proteins strongly support the close relationship be- tween sea snakes (Laticaudinae + Hydrophiinae) and Australopapuan terrestrial elapids; in addition they suggest that these are derived elapid lineages. It is less clear as to whether the two sea snake groups were derived independently from different groups of terrestrial elapids (McDowell, 1 969). Ca- dle and Sarich (1981), Cadle and Gorman (1981), DESSAUER ET AL.: SNAKE EVOLUTION 23 and Mao ct al. (1983) concluded that there was no special association between laticaudines and New World Micrurus (McDowell, 1967, 1969), since the albumins of these groups were more distinct from one another than were albumins of laticau- dines, hydrophiincs, and Australian terrestrial elapids. Schwaner and colleagues (1 985) inter- preted transferrin immunological data as sup- porting a possible derivation of hydrophiines from the Notechis group of Australian elapids (sec. V.A.), whereas laticaudines were derived independently from an unspecified lineage. However, the trans- ferrin immunological data have not been rate- tested, and the association between Notechis and hydrophiines could be due to rate differences among transferrins of these groups (as suggested for Notechis in fig. 2, the unrooted tree of Schwa- ner et al., 1985). In the absence of rate test data, the problem of independent origins for the two sea snake groups cannot be resolved. Although all molecular studies confirm the as- sociation of sea snakes with Australopapuan ter- restrial elapids, there has been little attention di- rected to the rest of the Elapidae. The phylogenetic position of New World coral snakes (micrurines) within the Elapidae was addressed by Cadle and Sarich (1981), and their general conclusions were supported by morphological studies (McCarthy, 1 985). Subsequent immunological comparisons of albumins (Cadle, unpubl. data) have failed to dem- onstrate a close association between micrurines and specific Old World elapid groups, but many Old World lineages remain to be tested. Mao and his colleagues (Mao et al., 1977, 1978, 1983) have interpreted their albumin and transferrin immu- nological comparisons within a framework which, a priori, assumes a basic division of elapids into a "terrestrial" group and a "sea snake" group, the latter recognized in 1983 as including Australo- papuan terrestrial elapids. Consequently, they concluded (Mao et al., 1983) that the albumin of Naja was highly divergent from those of other elapids (showing, for example, three times the rate of evolution of Bungarus albumin, see Mao et al., 1983, fig. 1). An elapid phytogeny constructed by using antisera to albumins of a variety of African and Asian elapids and sea snakes (Cadle, unpubl. data) shows that the assumption of "terrestrial" and "sea snake" groups is unwarranted; that is, the terrestrial elapids do not form a clade relative to the sea snakes. Our rate test data using appro- priate outgroups (e.g.. see table 4) indicate that, although albumin has changed somewhat more in Naja than in some other elapids, it has not changed to the degree suggested by Mao and colleagues. A tree analysis of albumin immunological compar- isons using appropriate outgroups (Cadle, unpubl. data) shows that Naja is a derivative of an ancient lineage from the common elapid stock, and thus its albumin and transferrin are very dissimilar from those of other elapids. Sequence data for elapid venom proteins have not been used extensively in addressing problems of elapid phylogeny. Sequences vary considerably within and between species and are subject to ex- tensive length mutations and gene duplications; also, their interpretation may depend on specific models of toxin evolution (Hseu et al., 1977; Duf- ton, 1984). Three major classes of elapid venom toxins are recognized: long and short neurotoxins, and cytotoxins. These classes are apparently re- lated by gene duplication events, but the relation- ship among the three classes is not clear (Strydom, 1979; Hseu et al., 1977). Short and long neuro- toxins are found in all elapids examined to date, whereas cytotoxins have thus far been found only in cobras of the genera Naja and Hemachatus (Hseu et al., 1977). This suggests that, primitively, a sin- gle duplication gave rise to the long and short neu- rotoxins, and these have been retained in all ela- pids. Subsequently, a further duplication occurred in the lineage leading to Naja and Hemachatus and gave rise to the cytotoxins. The cytotoxins have not yet been found in the king cobra (Ophio- phagus), suggesting that it might belong to a sep- arate lineage from the other cobras or that the gene duplication giving rise to the cytotoxins occurred after the divergence of Ophiophagus from other cobras. Independent albumin immunological evi- dence (Cadle, unpubl. data) shows that the former interpretation is more likely. Naja is a very early branch of the elapid lineage, whereas Ophiophagus is a much later lineage, more closely related to several other genera of Asian elapids. c. Atractaspis— Hypotheses concerning the re- lationships of this enigmatic African genus were summarized by Cadle (1982b). Although Atrac- taspis was long considered to be an aberrant viper, more recent comparative anatomical studies have suggested that it is an "aparallactine" colubrid (Bourgeois, 1965; McDowell, 1987) or has elapid affinities (Kochva et al., 1967; Kochva & Woll- berg, 1970). MC'F comparisons of albumins (table 4) allow us to reject an association between Atrac- taspis and viperids (Cadle, 1982a). These results show that viperids are among those advanced snakes most distant from Atractaspis. In view of the conservative nature of many viperid (partic- 24 FIELDIANA: ZOOLOGY ularly crotaline) albumins, one would expect the Atractaspis-viperid distances to be less than those to other advanced snakes if there were a phylo- genetic association between these groups, but this is not the case. The separation of Atractaspis from the viperid clade is also demonstrated by a tree analysis of albumin immunological data involving all major lineages of colubroids (Cadle, 1982a). We are less confident in offering a definitive statement on the phylogenetic position of Atrac- taspis at present, because of difficulties concerning the development of a molecular phylogeny for all colubroids (see Cadle, 1987, for discussion). The available albumin immunological data indicate no special relationship between Atractaspis and some aparallactines (Amblyodipsas and Aparallactus; Cadle, 1 982b); but others remain to be tested, and the "Aparallactinae" may not be monophyletic (Cadle, 1982b; McDowell, 1987). Dowling and colleagues (1983) placed Atractaspis among the lycodontines on the basis of an albumin ID of 80 from Madagascar ophis\ however, this distance is typical of that between Atractaspis and many lin- eages of colubrids (Cadle, 1982a,b; unpubl. data) and does not specifically support a lycodontine association. Among all taxa to which Atractaspis has yet been compared, there is a possibly remote asso- ciation with the elapids (Cadle, 1983). The data are also consistent with an independent origin for these lineages at about the same time from a com- mon colubroid stock that later also gave rise to colubrids (Cadle, 1982a). The elapid association is apparent in Table 4, where immunological dis- tances between Atractaspis and elapids are lowest among all comparisons. That this association is not due to conservativeness of either elapid or Atractaspis albumins has been confirmed by rel- ative rate tests (Cadle, 1982a). Notably, the pos- sible phylogenetic association between elapids and Atractaspis is also indicated by certain aspects of venom composition (Minton, 1968; Parnas & Russell, 1967; Kochva et al., 1982), venom gland structure (Kochva et al., 1967; Kochva & Woll- berg, 1 970), and ectopterygoid shape (Lombard et al., 1986). d. Colubridae— Unraveling relationships with- in this large group poses many difficulties. Because homoplasy in morphological characters appears to be rampant in colubrids, molecular data, which generally do not exhibit strong convergence, ul- timately will help solve many of the more difficult phylogenetic problems. Currently, molecular data bearing on colubrid relationships consist of com- Table 4. Immunological distances between the al- bumins of Atractaspis and other advanced snakes, using an antiserum to Atractaspis bibroni albumin. Antiserum to Atractaspis bibroni Albumins albumin Atractaspis bibroni 0 Atractaspis dahomeyensis 16 Atractaspis microlepidota 32 Crotalinae Crotalus enyo 94 Bothrops atrox 98 Lachesis muta 97 Sistrurus catenatus 91 Sistrurus miliarius 104 Agkistrodon piscivorus 86 Agkistrodon contortrix 82 Agkistrodon bilineatus 100 Viperinae Bids arietans 132 Vipera aspis 106 Vipera palestinae 126 Cansus resimus 117 Causus maculatus 113 Echis ocellatus 122 Echis coloratus 119 Cerastes cerastes 112 Pseudocerastes field ii 95 Elapidae Micrurus spixi 74 Bungarus fasciatus 73 Laticauda semifasciata 72 Hydrophis melanosoma 72 Dendroaspis polylepis 79 Elapsoidea semiannulata 91 Naja haje 99 parisons of albumins and transferrins by MC'F and immunodiffusion. It is clear that the amount of molecular change separating major clades (e.g., natricines, colubrines) is relatively small for these two proteins, thus making it difficult to sort the lineages into a series of dichotomous branches. For example, Cadle (1982a, 1984a), using MC'F com- parisons of albumins, could not resolve a trichoto- my involving colubrines and two xenodontine lin- eages. Dowling and colleagues (1983) provided an estimated phylogeny for several colubrid lineages based on MC'F comparisons of albumins. How- ever, these data were not rate-tested using suitable outgroups, and involved few representatives of each lineage; thus, we question the details of branching order among major lineages presented in that work. In general, members of the various colubrid lin- eages are separated by about 70 to 90 albumin ID DESSAUER ET AL.: SNAKE EVOLUTION 25 units and 100 or more transferrin ID units (Cadle, 1982a,b; George & Dessauer, 1970; Mao & Des- sauer, 1971; Cadle & Dessauer. unpubl. data). This is estimated to represent between 30 and 60 mil- lion years of separation for the lineages (Cadle, 1982a, 1987). In the discussion below, we con- centrate on those areas where molecular data have contributed substantially to the phylogenetic anal- ysis of particular groups of colubrids. We do not review previously published data in detail. Xenodontinae— Albumins of numerous species of xenodontines have been compared by MC'F (Cadle, 1984a,b,c); multilocus electrophoretic studies and quantitative immunological compar- isons of transferrins are in progress (Cadle & Dessauer, 198S, unpubl. data). The albumin im- munological results are consistent with immuno- diffusion comparisons of transferrins (Schwaner & Dessauer, 1982). The biochemical data suggest a major dichotomy between two speciose Neotrop- ical lineages (Central and South American xeno- dontine lineages, using terminology in Cadle, 1 984a) and several essentially monotypic lineages that are well differentiated from the major lineages and from each other: the North American genera Farancia. Heterodon, Carphophis, Diadophis, and Contia; the Central American Conophis; and the South American Hydrops (Cadle, 1984c; unpubl. data). Members of these lineages differ on average by approximately 70 albumin immunological units. The molecular data were instrumental in unrav- eling relationships among genera within this com- plex group, and in interpreting historical biogeo- graphic patterns in the Neotropics (Cadle, 1985). Molecular studies will likely contribute substan- tially to the resolution of two major phylogenetic problems that still exist for xenodontines: ( 1 ) Are the various xenodontine lineages monophyletic relative to other colubrid lineages; and (2) what is the sister group or groups of the xenodontines? Lycodontinae/Boodontinae— There has been little agreement concerning relationships of snakes in these groups (e.g., compare Underwood, 1967; Dowling & Duellman, 1978; and McDowell, 1987). The molecular work that has been done on them indicates that, like the xenodontines, several an- cient lineages are involved. Semiquantitative im- munological comparisons of transferrins (Schwa- ner & Dessauer, 1982) and MC'F comparisons of albumins (table 5; Dowling et al., 1983) suggest that lycodontines may not be monophyletic. Clearly, the magnitude of albumin IDs within the lycodontines is equivalent to that between lyco- dontines and other colubrid lineages. These data are consistent with McDowell's (1987) view that the lycodontine/boodontine group includes many primitive snakes that are not clearly linked to one another by derived characters. Our present mo- lecular data suggest that the phylogeny of the ly- codontines will prove to be a complex series of lineages such as is seen in the xenodontines. Homalopsinae— This is a morphologically dis- tinctive radiation of primarily estuarine and aquatic snakes (Gyi, 1970; McDowell, 1987). Although they have sometimes been considered relatives of the natricines (e.g., Dowling & Duellman, 1978), biochemical studies show that they are an inde- pendent lineage (George & Dessauer, 1970; Schwaner & Dessauer, 1982; Dowling etal., 1983). Dowling and colleagues (1983) compared the al- bumins of Erpelon and Enhydris to Thamnophis and Madagascarophis by MC'F and found the IDs separating these (approximately > 70) equivalent to those generally separating colubrid lineages. Thus, homalopsines are molecularly well differ- entiated, but their relationship to other colubrid lineages is as yet unclear. We specifically exclude Acrochordus from the Homalopsinae (sec. VLB., 2c). Natricinae— Relationships among natricines have been extensively studied using immunolog- ical techniques (Pearson, 1966, 1968; Mao & Des- sauer, 1971; Schwaner & Dessauer, 1 982; Gartside & Dessauer, 1977; Dowling et al., 1983), peptide fingerprinting (Sutton, 1 969; Dessauer, 1974), and electrophoresis (Lawson & Dessauer, 1979; Law- son, 1985, 1986). These studies demonstrate the monophyly of North American genera (the Tham- nophiini) relative to Old World forms (see sec. V. A.). The albumins and transferrins of thamno- phiines are so similar when compared immuno- logically (Mao & Dessauer, 1971; Dowling et al., 1 983) that electrophoretic approaches are proving more useful in working out details of their rela- tionships (fig. 3, top; Lawson & Dessauer, 1979; Lawson, 1985, 1986, 1987). Among Old World genera, transferrin immunological comparisons (Mao & Dessauer, 1971; Gartside & Dessauer, 1 977; Schwaner & Dessauer, 1 982) show four ma- jor groups (Natrix, Afronatrix, Sinonatrix, and Xenochrophis-A mphiesma -Rhabdophis). The transferrin and albumin IDs separating these groups are about 50 to 60 and 40 to 50, respectively; however, the branching order among the major lineages is not resolved by currently available bio- chemical data. Numerous other genera from Asia and Africa are possible natricines (McDowell, 1987), but most of these have not been examined 26 FIELDIANA: ZOOLOGY Table 5. Immunological distances involving Lycodontine/Boodontine albumins. Antisera Albumins (1) (2) (3) (4) Lycodontinae/Boodontinae (1) Lamprophis fuliginosus (2) Rhamphiophis oxyrhynchus (3) Amblyodipsas polylepis (4) Mehelya crossi Aparallactus capensis Colubrinae Xenodontines South American Central American 0 79 88* 57 t 71(12) 103 92 b 121 89 t 87(11) 100 133 b 100 91* 88(4) t 111 b t 82 (10) 89 108 All intralycodontine comparisons are means of reciprocal titrations, except those marked with an asterisk (*), which are unidirectional comparisons. For the interlineage comparisons, numbers in parentheses give the number of com- parisons which were averaged to give the reported values; all other interlineage comparisons are based on a single titration. t No value available. biochemically (Dowling et al., 1983, included Na- triciteres here on the basis of an albumin ID of 46, as compared to Thamnophis). Biochemical data specifically refute the association of Acrochordus and homalopsines (Dowling & Duellman, 1978) with the natricines (see sec. VLB., 2c). Colubrinae— A substantial body of albumin and transferrin immunological comparisons and elec- trophoretic studies indicate that members of this lineage are relatively closely related worldwide (George & Dessauer, 1970; Minton & Salanitro, 1972; Dowling et al., 1983; Cadle, 1984c; Lawson & Dessauer, 1981; Schwaner & Dessauer, 1982). Many of the large number of genera are closely related genetically (Lawson & Dessauer, 1981, un- publ. data; Dowling et al., 1983); others, such as Elaphe and Coluber, are clearly polyphyletic (see sec. V.B.). Relationships within the colubrines are not worked out in detail, but the available molec- ular evidence identifies two major generic groups in the North American fauna: group 1 includes North American Arizona, Elaphe, Lampropeltis, Rhinocheilus, Pituophis, Stilosoma, and Cemo- phora; and group 2 includes North American Col- uber, Masticophis, and Opheodrys (Pearson, 1 966; George & Dessauer, 1970; Minton & Salanitro, 1972; Lawson & Dessauer, 1981; Dowling et al., 1983). Other genera can be associated with each of these groups, based on additional electropho- retic and immunological evidence (see sec. V). Among African forms, species of Bourgeois's (1965) Dispholidinae, Thrasops, Dispholidus, and Thelotornis form a tight cluster (Cadle, unpubl. data; Rhamnophis has not been tested). The close relationships among genera within the Colubrinae indicated by the molecular studies draw attention to the fact that, more than any other lineage of the Colubridae, this one has apparently radiated ex- plosively. The approximately 1 50 genera of col- ubrines are distributed among all habitable con- tinents, posing interesting questions concerned with their biogeography and dispersal (Cadle, 1987). Further discussion of molecular evidence on this group may be found in George and Dessauer (1970), Minton and Salanitro (1972), Minton (1976), Dowling and colleagues (1983), and Cadle (1984c). Although such molecular studies have been very effective in increasing our knowledge of relation- ships among the colubrines, it is clear that an enor- mous amount of work remains before a clear pic- ture of their phylogeny will emerge. Given the large number of taxa involved and the limitations of the techniques employed, most colubrine genera are too distantly related for electrophoresis to be effective and too closely related for the resolving power of MC'F. VII. Summary 1. Most electrophoretically studied structural genes of snakes are inherited as codominant al- leles. 2. Protein polymorphism in snakes is similar to levels observed in other vertebrates. 3. Polymorphic proteins are useful markers for identifying individual snakes and for studying their breeding patterns, population genetics, and prob- lems concerned with species formation. DESSAUER ET AL.: SNAKE EVOLUTION 27 4. Among North American species of Nerodia and Thamnophis. many populations presently classified as subspecies are either already repro- ductively isolated or have at least attained the "in- cipient" species level. 5. Comparative biochemical studies suggest that several widespread colubrid genera are not mono- phyletic (e.g., Natrix, sensu Malnate, 1 960; Elaphe, Coluber, Rhadinaea). 6. Many major lineages of snakes include one or more speciose radiations characterized by marked morphological and ecological diversity and minimal protein evolution. The best-documented examples of these are the Australopapuan elapid sea snake radiation and various groups of natri- cines, colubrines, and xenodontines. 7. A number of groups include some genera with few species that appear to be relics of ancient ra- diations (e.g., Acrochordus. Loxocemus, Farancia. Heterodon. Carphophis). 8. Among extant reptiles, the closest relatives of snakes are among the lizards. At present, the molecular evidence is not sufficient to determine the precise relationship between lizards and snakes. 9. Snakes are monophyletic, made up of at least two very ancient lineages, the blind snakes and the primitive snakes plus the advanced snakes. Their origins probably stem from the Mesozoic Era. The blind snakes and advanced snakes are monophy- letic; the primitive snakes (Henophidia) very likely are not. 10. Typhlops is an ancient sister group of Lep- totyphlops. 1 1 . The uropeltids are a compact radiation and appear to be the sister group of Cylindrophis. The relationship of Anilius to this group is remote. 1 2. Pythons and boids are not each other's clos- est relatives among primitive snakes; however, most classifications group these snakes together as either the Booidea or Boidae. Limited biochemical evidence also conflicts with other aspects of hen- ophidian classification, such as the monophyletic status of the Tropidophiidae. 1 3. Acrochordus is excluded from the Natrici- nae and Homalopsinae; its phylogenetic position cannot be resolved with present molecular data. 14. Vipers are the sister group to other lineages of advanced snakes. Albumin evolution is variable within vipers, and on the average is more conser- vative in this group than in other advanced snakes. 1 5. The sea snakes are a derived lineage of ela- pids, closely related to Australopapuan terrestrial elapids. Atractaspis is possibly a member of the elapid clade, but present data cannot exclude the possibility that it is the sister group of elapids and colubrids. 16. A major unresolved question in colubrid systematics is whether this group is monophyletic. Relationships among major clades is unresolved. Xenodontines and Lycodontine/Boodontines may not be monophyletic. Homalopsines and natri- cines do not clearly form a clade relative to other lineages. Colubrines are a highly speciose but mo- lecularly very cohesive group. VIII. Acknowledgments The list of scientists who have contributed spec- imens to the frozen tissue collections used in our studies is far too large to acknowledge individ- ually. Without their contributions, however, the research upon which this review is based would have been impossible. We also wish to thank the following colleagues for helping us with this manu- script: Dr. Vincent M. Sarich for the enhanced immunodiffusion method used to test cross-re- actions between proteins of widely divergent taxa; Drs. John P. O'Neill, Harold K. Voris, and Rich- ard G. Zweifel for allowing us the use of their photographs in Figures 1, 2, and 3; and Ms. Lisa Candelario for help with the preparation of the figures. Special thanks are extended to Drs. Harry Greene, Douglas A. Rossman, and to two un- known reviewers for evaluating our original manuscript. Whereas they do not necessarily agree with all of our interpretations, their suggestions certainly have strengthened the accuracy and qual- ity of the review. We also acknowledge with grat- itude many years of National Science Foundation grant support for research upon which this review is based, most recently Dissertation Research Sup- port to JEC (DEB 80-14101) and a joint research grant to HCD and JEC (BSR-84000166). IX. Literature Cited Aird, S. D., and H. C. Dessauer. 1977. Geographic variation in venom and blood proteins of Crotalus viridis. Abstract, Annual Meeting of the American So- ciety of Ichthyologists and Herpetologists. Alberch, P. 1 985. Problems with the interpretation of developmental sequences. Systematic Zoology, 34: 46- 58. Alberch, P., S. J. Gould, G. F. Oster, and D. B. Wake. 1979. Size and shape in ontogeny and phylogeny. Paleobiology, 5: 296-317. 28 FIELDIANA: ZOOLOGY A vise, J. C. 1974. Systematic value of electrophoretic data. Systematic Zoology, 23: 465—481. Bailey, J. R. 1939. Relationships and distribution of the snakes allied to the genus Pseudoboa. Ph.D. diss., University of Michigan, Ann Arbor. . 1 967. The synthetic approach to colubrid clas- sification. Herpetologica, 23: 155-161. Baker, R. J., J. J. Bull, and G. A. Mengden. 1971. Chromosomes of Elaphe subocularis (Reptilia: Ser- pentes), with the description of an in vivo technique for the preparation of snake chromosomes. Experien- tial: 1228-1229. Baker, R. J., G. A. Mengden, and J. J. Bull. 1972. Karyotypic studies of thirty-eight species of North American snakes. Copeia, 1972: 257-265. Banks, C, and T. D. Schwaner. 1984. Two cases of interspecific hybridization among captive Australian boid snakes. Zoo Biology, 3: 221-227. Bellemin, J., G. Adest, G. C. Gorman, and M. Alek- siuk. 1978. Genetic uniformity in northern popu- lations of Thamnophis sirtalis (Serpentes: Colubridae). Copeia, 1978: 150-151. Benjamin, D. C, J. A. Berzofsky, I. J. East, F. R. N. Gurd, C. Hannum, S. J. Leach, E. Margoliash, J. G. Michael, A. Miller, E. M. Prager, M. Reichlin, E. E. Sercarz, S. J. Smith-Gill, P. E. Todd, and A. C.Wilson. 1984. The antigenic structure of proteins: A reappraisal. Annual Review of Immunology, 2: 67- 101. Blaney, R. M., and P. Blaney. 1979. The Nerodia sipedon complex of water snakes in Mississippi and southeastern Louisiana. Herpetologica, 35: 350-359. Bourgeois, M. 1965. Contribution a la morphologie comparee du crane des ophidiens de l'Afrique Cen- trale. Publications del Universite Officielle du Congo, Lubumbashi, 28: 1-293. Broer, W. 1978. Bastarde bei zwei Elaphe- Arten (Reptilia: Serpentes: Colubridae). Salamandra, 14: 63- 68. Bury, R. B., F. Gress, and G. C. Gorman. 1970. Karyotypic survey of some colubrid snakes from west- ern North America. Herpetologica, 26: 461-466. Cadle, J. E. 1982a. Evolutionary relationships among advanced snakes. Ph.D. diss., University of Califor- nia, Berkeley, 274 pp. . 1982b. Problems and approaches in the inter- pretation of the evolutionary history of venomous snakes. Memoriasdo Instituto Butantan, 46: 255-274. . 1985. The Neotropical colubrid snake fauna (Serpentes: Colubridae): Lineage components and bio- geography. Systematic Zoology, 34: 1-20. 1987. The geographic distribution of snakes: . 1983. Phylogenetic relationships of African back-stabbing snakes, genus Atractaspis. Abstracts, Annual Meeting, Society for the Study of Amphibians and Reptiles and the Herpetologists' League: 49. -. 1 984a. Molecular systematics of Neotropical xenodontine snakes. I. South American xenodontines. Herpetologica, 40: 8-20. . 1 984b. Molecular systematics of Neotropical xenodontine snakes. II. Central American xenodon- tines. Herpetologica, 40: 21-30. . 1 984c. Molecular systematics of Neotropical xenodontine snakes. III. Overview of xenodontine phylogeny and the history of New World snakes. Co- peia, 1984: 641-652. Problems in phylogeny and zoogeography, pp. 77-105. In Seigel, R. A., J. T. Collins, and S. S. Novak, eds., Snakes: Ecology and Evolutionary Biology. Macmillan Publishing Company, New York. Cadle, J. E., and H. C. Dessauer. 1985. Biochemical evolution in South American xenodontine snakes. Ab- stracts, Annual Meeting of the American Society of Ichthyologists and Herpetologists: 47. Cadle, J. E., and G. C. Gorman. 1981. Albumin im- munological evidence and the relationships of sea snakes. Journal of Herpetology, 15: 329-334. Cadle, J. E., and V. M. Sarich. 1981. An immuno- logical assessment of the phylogenetic position of New World coral snakes. Journal of Zoology, London, 195: 157-167. Canfield, R. E., and C. B. Anfinsen. 1963. Concepts and experimental approaches in the determination of the primary structure of proteins, pp. 311-378. In Neurath, H., ed., The Proteins, vol. 1 . Academic Press, New York. Case, S. M., and E. E. Williams. 1984. Study of a contact zone in the Anolis distichus complex in the central Dominican Republic. Herpetologica, 40: 118- 137. Champion, A. B., E. M. Prager, D. Wachter, and A. C. Wilson. 1974. Microcomplement fixation, pp. 397-416. In Wright, C. A., ed., Biochemical and Im- munological Taxonomy of Animals. Academic Press, London. Cogger, H.G. 1975. Reptiles and Amphibians of Aus- tralia. A. H. and A. W. Reed, Sydney, 584 pp. Cohen, P. P., and G. W. Brown, Jr. 1960. Ammonia metabolism and urea synthesis, pp. 161-244. In Flor- kin, M., and H. S. Mason, eds., Comparative Bio- chemistry, vol. 2. Academic Press, New York. Con ant, R. 1963. Evidence for the specific status of the water snake Natrix fasciata. American Museum Novitates, 2122: 1-38. . 1975. A Field Guide to Reptiles and Amphib- ians of Eastern and Central North America, 2nd ed. Houghton Mifflin Co., Boston. Coulter, A. R., R. D. Harris, and S. K. Sutherland. 1981. Enzyme immunoassay and radioimmunoas- say: Their use in the study of Australian and exotic snake venoms, pp. 39-43. In Banks, C. A., and A. A. Martin, eds., Proceedings of the Melbourne Herpe- tological Symposium, Zoological Board of Victoria. Crabtree, C. B., and R. W. Murphy. 1984. Analysis of maternal-offspring allozymes in Crotalus viridis. Journal of Herpetology, 18: 75-80. Dene, H., M. Goodman, and W. S. Prychodko. 1 978. An immunological examination of the systematics of Tupaioidea. Journal of Mammalogy, 59: 697-706. Dessauer, H. C. 1970. An Alpha Helix Expedition to study proteins. Bulletin Serological Museum, Rutgers University, 44: 6-7. . 1974. Biochemical and immunological evi- dence on the relationships in Amphibia and Reptilia, DESSAUER ET AL.: SNAKE EVOLUTION 29 pp. 177-242. In Wright, C. A., ed., Biochemical and Immunological Taxonomy of Animals. Academic Press, New York. Dessauer, H. C. M. J. Braun, and L. D. Densmore. In press. Genetic interpretation of electrophoretic phenotypes of proteins. In Buth, D. G., and R. W. Murphy, eds.. Gene Expression in Reptilian System- atics. University of California Press, Berkeley. Dessauer, H. C, W. Fox, and Q. L. Hartwig. 1962. Comparative study of transferrins of Amphibia and Reptilia using starch-gel electrophoresis and autora- diography. Comparative Biochemistry and Physiol- ogy, 5: 17-29. Dessauer, H. C, D. F. Gartside, and C Gans. 1976. Protein evidence on the genetic diversity and affinities of uropeltid snakes. American Zoologist, 16: 268. Dessauer, H. C, and M. S. Hafner, eds. 1984. Col- lections of Frozen Tissues: Value, Management, Field and Laboratory Procedures, and Directory of Existing Collections. Association of Systematic Collections, University of Kansas Press, Lawrence, 74 pp. Dessauer, H. C, and F. H. Pough. 1975. Geographic variation of blood proteins and the systematics of kingsnakes (Lampropeltis getulus). Comparative Bio- chemistry and Physiology, SOB: 9-12. Dessauer, H. C, and R. G. Zweifel. 1981. Inheri- tance of transferrin, phosphoglucomutase, 6-phos- phogluconate dehydrogenase and prolidase in a breed- ing colony of kingsnakes. Journal of Heredity, 72: 453-455. Detratt, J., and H. Saint-Girons. 1979. Commu- nautes antigeniques des venins et systematique des Viperidae. Bijdragen to de Dierkunde, 49: 71-80. Dowung, H. G., and W. E. Duellman. 1978. Sys- tematic Herpetology. A Synopsis of Families and Higher Categories. HISS Publications, New York. Dowung, H. G., R. Highton, G. C Maha, and L. R. Maxson. 1983. Biochemical evaluation of colubrid snake phylogeny. Journal of Zoology, London, 201: 309-329. Downs, F. L. 1967. Intrageneric relationships among colubrid snakes of the genus Geophis Wagler. Miscel- laneous Publications, Museum of Zoology, University of Michigan, 131: 1-193. Dufton, M. J. 1984. Classification of elapid snake neurotoxins and cytotoxins according to chain length: Evolutionary implications. Journal of Molecular Evo- lution. 20: 128-134. Dunn, E. R. 1951. The status of the snake genera Dip- sas and Sibon. a problem for "quantum evolution." Evolution. 5: 355-358. Eberle, W. G. 1972. Comparative chromosomal mor- phology of the New World natricine snake genera Na- trix and Regina. Hcrpctologica, 28: 98-105. Felsenstein. J. 1 982. Numerical methods for inferring evolutionary trees. Quarterly Review of Biology, 57: 379-404. Fitch, H. S. 1940. A biogeographical study of the or- dinotdes artenkreis of garter snakes (genus Thamno- phis). University of California Publications in Zool- ogy. 44: 1-150. . 1965. An ecological study of the garter snake. Thamnophis sirtalis. University of Kansas Publica- tions, Museum of Natural History, 15: 493-564. Fitch, W. M. 1 982. The challenges to Darwinism since the last centennial and the impact of molecular studies. Evolution. 36: 1133-1143. Fox, W. 1951. Relationships among the garter snakes of the Thamnophis elegans rassenkreis. University of California Publications in Zoology, 50: 485-530. Fox, W., Jr., and H. C. Dessauer. 1965. Collection of garter snakes for blood studies. Year Book of the American Philosophical Society, 1964: 263-266. Gartside, D. F, and H. C Dessauer. 1977. Immu- nological evidence on affinities of African Natrix. Co- peia, 1977: 190-191. Gartside, D. F, J. S. Rogers, and H. C Dessauer. 1977. Speciation with little genie and morphological differentiation in the ribbon snakes Thamnophis prox- imus and T. sauritus (Colubridae). Copeia, 1977: 697- 705. George, D. W. 1 969. Immunological correspondence of transferrins and the relationships of colubrid snakes. Master's thesis, Louisiana State University Medical Center, New Orleans, 44 pp. George, D. W., and H. C Dessauer. 1970. Immu- nological correspondence of transferrins and relation- ships of colubrid snakes. Comparative Biochemistry and Physiology, 33: 617-627. Goncalves, J. M., and L. G. Vieira. 1950. Estudos sobre venenos de serpentes Brasileiras. Anais da Aca- demia Brasileira de Ciencias, 22: 141-149. Goodman, M., and G. W. Moore. 1971. Immuno- diffusion systematics of primates. I. The Catarrhini. Systematic Zoology, 20: 19-62. Gorman, G. C, A. C. Wilson, andM. Nakanishi. 1 97 1 . A biochemical approach toward the study of reptilian phylogeny: Evolution of serum albumin and lactate dehydrogenase. Systematic Zoology, 20: 167-185. Gould, S. J., and N. Eldredge. 1977. Punctuated equilibria: The tempo and mode of evolution recon- sidered. Paleobiology, 3: 115-151. Graham-Smith, G. S. 1904. Blood relationship amongst the lower Vertebra ta and Arthropoda, etc., as indicated by 2500 tests with precipitating antisera, pp. 336-380. In Nuttall, G. H. F., ed., Blood Im- munity and Blood Relationship. Cambridge Univer- sity Press, London. Groombridge, B. 1979a. On the vomer in Acrochor- didae (Reptilia: Serpentes), and its cladistic signifi- cance. Journal of Zoology, London, 189: 559-567. . 1979b. Variations in morphology of the su- perficial palate of henophidian snakes and some pos- sible systematic implications. Journal of Natural His- tory. 13: 447-475. . 1 984. The facial carotid artery in snakes (Rep- tilia, Serpentes): Variations and possible cladistic sig- nificance. Amphibia-Reptilia, 5: 145-155. Gyi, K. K. 1970. A revision of colubrid snakes of the subfamily Homalopsinae. University of Kansas Pub- lications, Museum of Natural History, 20: 47-223. Haluska, F., and P. Alberch. 1983. The cranial de- velopment of Elaphe obsoleta (Ophidia, Colubridae). Journal of Morphology, 178: 37-55. 30 FIELDIANA: ZOOLOGY Hames, B. D., and D. Rickwood. 1981. Gel Electro- phoresis of Proteins: A Practical Approach. IRL Press, Oxford, 290 pp. Harris, H. 1975. The Principles of Human Biochem- ical Genetics. North-Holland Publishing Co., Am- sterdam, 473 pp. Harris, H., and D. A. Hopkjnson. 1976. Handbook of Enzyme Electrophoresis in Human Genetics. North- Holland Publishing Co., Amsterdam. Haslewood, G. A. D. 1978. The Biological Impor- tance of Bile Salts. North-Holland Publishing Co., Amsterdam. Hseu, T. H., E. D. Jou, C. Wang, and C. C. Yang. 1977. Molecular evolution of snake venom toxins. Journal of Molecular Evolution, 10: 167-182. Jenner, J. V., and H. G. Dowung. 1985. Taxonomy of American xenodontine snakes: The tribe Pseudo- boini. Herpetologica, 41: 161-172. Jimenez-Porras, J. M. 1964a. Venom proteins of the fer-de-lance, Bothrops atrox, from Costa Rica. Toxi- con, 2: 155-166. 1 964b. Intraspecific variations in composition of venom of the jumping viper, Bothrops nummifera. Toxicon, 2: 187-195. . 1967. Differentiation between Bothrops num- mifer and Bothrops picadoi by means of the biochem- ical properties of their venoms, pp. 307-321. In Rus- sell, F. E., and P. R. Saunders, eds., Animal Toxins. Pergamon Press, Oxford. Kellaway, C. H., and F. E. Williams. 1931. The serological and blood relationships of some common Australian snakes. Australian Journal of Experimen- tal Biology and Medical Science, 8: 123-132. Kephart, D. G., and S.J. Arnold. 1982. Garter snake diets in a fluctuating environment: A seven year study. Ecology, 63: 1232-1236. Kochva, E., M. Shayer-Wollberg, and R. Sobol. 1967. The special pattern of the venom gland in Atractaspis and its bearing on the taxonomic status of the genus. Copeia, 1967: 763-772. Kochva, E., C. Viuoen, and D. Botes. 1982. Anew type of toxin in the venom of snakes of the genus Atractaspis (Atractaspidinae). Toxicon, 20: 581-592. Kochva, E., and M. Wollberg. 1970. The salivary glands of Aparallactinae (Colubridae) and the venom glands ofElaps (Elapidae) in relation to the taxonomic status of this genus. Zoological Journal of the Linnaean Society, 49: 217-224. Kuwajima, Y. 1953. Immunological researches on the main Formosan poisonous snakes, especially on the venoms. I. Classification of poisonous snakes in For- mosa by means of serological methods based on em- ploying snake blood-sera as antigens. Japanese Journal of Experimental Medicine, 23: 21-25. Larson, A. 1984. Neontological inferences of evolu- tionary pattern and process in the salamander family Plethodontidae, pp. 119-217. In Hecht, M. K, B. Wallace, and G. T. Prance, eds., Evolutionary Biology, vol. 17. Plenum Publishing Co., New York. Larson, A., E. M. Prager, and A. C. Wilson. 1984. Chromosomal evolution, speciation, and morpholog- ical change in vertebrates: The role of social behav- iour. Chromosomes Today, 8: 2 1 5-228. Lauder, G. V. 1981. Form and function: Structural analysis in evolutionary morphology. Paleobiology, 7: 430-442. Lawson, R. 1978. Molecular systematics and evolu- tion in the western garter snakes: Genus Thamnophis. Master's thesis, California State University, Hayward, 125 pp. . 1985. Molecular studies of thamnophiine snakes. Ph.D. diss., Louisiana State University, Baton Rouge, 183 pp. 1 986. Molecular systematics of some Old World natricine snakes, pp. 227-234. In Rocek, Z., ed., Stud- ies in Herpetology. Proceedings Societas Europaea Herpetologica, Prague. 1987. Molecular studies of thamnophiine snakes: 1 . The phylogeny of the genus Nerodia. Journal of Herpetology, 21(2): 140-157. Lawson, R., and H. C. Dessauer. 1979. Biochemical genetics and systematics of garter snakes of the Tham- nophis elegans-couchii-ordinoides complex. Occasion- al Papers of the Museum of Zoology, Louisiana State University, 56: 1-24. . 1981. Electrophoretic evaluation of the colu- brid genus Elaphe (Fitzinger). Isozyme Bulletin, 14: 83. Lawson, R., A. J. Meier, Jr., and P. Frank, Jr. 198 1 . Protein electrophoretic evidence on the status of sub- species of the banded water snake. Abstracts, Annual Meeting of the Southwestern Association of Natural- ists: 18. Lederer, G. 1950. Ein bastarde von Elaphe guttata (Linne)— mannchen x Elaphe quadrivittata quadri- vittata (Holbrook)— weibchen und dessen Ruckkreu- zung mit der mutterlichen Ausgangsart. Zool. Gart., 17: 235-242. Liem, K., H.Marx, and G.B.Rabb. 1971. Theviperid snake Azemiops: Its comparative cephalic anatomy and phylogenetic position in relation to Viperinae and Crotalinae. Fieldiana: Zoology, 59: 63-126. Lombard, R. E., H. Marx, and G. B. Rabb. 1986. Morphometries of the ectopterygoid in advanced snakes (Colubroidea): A concordance of shape and phylogeny. Biological Journal of the Linnaean Society, 27: 1 33- 164. Malnate, E. V. 1960. Systematic division and evo- lution of the colubrid snake genus Natrix, with com- ments on the subfamily Natricinae. Proceedings of the Academy of Natural Sciences of Philadelphia, 112: 41-71. Mao, S. H., B. Y. Chen, and H. M. Chang. 1977. The evolutionary relationships of sea snakes suggested by immunological cross-reactivity of transferrins. Com- parative Biochemistry and Physiology, 57 A: 403—406. Mao, S. H., B. Y. Chen, F. Y. Yin, and Y. W. Guo. 1983. Immunotaxonomic relationships of sea snakes to terrestrial elapids. Comparative Biochemistry and Physiology, 74A: 869-872. Mao, S. H., and H. C. Dessauer. 1971. Selectively neutral mutations, transferrins and the evolution of natricine snakes. Comparative Biochemistry and Physiology, 40A: 669-680. Mao, S. H., H. C. Dessauer, and B. Y. Chen. 1978. Fingerprint correspondence of hemoglobins and the DESSAUER ET AL.: SNAKE EVOLUTION 31 relationships of sea snakes. Comparative Biochemis- try and Physiology, 59B: 353-361. Mao, S. H., Y. W. Guo, F. Y. Yin, and B. Y. Chen. 1984. Hemoglobin fingerprint correspondence and relationships of Taiwan common venomous snakes. Comparative Biochemistry and Physiology, 78B: 85- 92. Maxson, L., and R. D. Maxson. 1979. Comparative albumin and biochemical evolution in plethodontid salamanders. Evolution, 33: 1057-1062. Maxson, L. R.. V. M. Sarich, and A. C. Wilson. 1975. Continental drift and the use of albumin as an evo- lutionary clock. Nature, 255: 397-400. McCarthy, C. J. 1985. Monophyly of elapid snakes (Serpentes: Elapidae). An assessment of the evidence. Zoological Journal of the Linnaean Society, 83: 79- 93. McDowell, S. B. 1967. Aspidomorphus. a genus of New Guinea snakes of the family Elapidae, with notes on related genera. Journal of Zoology, London, 151: 497-543. . 1969. Toxicocalamus. a New Guinea genus of snakes of the family Elapidae. Journal of Zoology, London, 159:443-511. . 1974. A catalogue of the snakes of New Guinea and the Solomons, with special reference to those in the Bemice P. Bishop Museum. Part I. Scolecophidia. Journal of Herpetology, 8: 1-57. 1975. A catalogue of the snakes of New Guinea and the Solomons, with special reference to those in the Bemice P. Bishop Museum. Part II. Anilioidea and Pythoninae. Journal of Herpetology, 9: 1-80. 1979. A catalogue of the snakes of New Guinea and the Solomons, with special reference to those in the Bernice P. Bishop Museum. Part III. Boinae and Acrochordoidea (Reptilia, Serpentes). Journal of Her- petology, 13: 1-92. 1987. Snake systematics, pp. 3-50. In Seigel, R. A., J. T. Collins, and S. S. Novak, eds., Snakes: Ecology and Evolutionary Biology. Macmillan Pub- lishing Co., New York. McDowell, S. B., and C. M. Bogert. 1954. The sys- tematic position of Lanthanotus and the affinities of the Anguinomorphan lizards. Bulletin of the Ameri- can Museum of Natural History, 105: 1-142. Mebs, D. 1985. List of Biologically Active Compo- nents from Snake Venoms. Zentrum der Rechtsme- dizin. University of Frankfurt, Frankfurt am Main, Federal Republic of Germany, 141 pp. Mengden, G. A. 1985a. Australian elapid phytogeny: A summary of the chromosomal and electrophoretic data. pp. 185-192. In Grigg, G., R. Shine, and H. Ehmann, eds.. Biology of Australasian Frogs and Rep- tiles. Royal Zoological Society of New South Wales, Surrey Beatty and Sons, Chipping Norton, New South Wales. 1985b. A chromosomal and electrophoretic analysis of the genus Pseudonaja, pp. 193-208. In Grigg, G., R. Shine, and H. Ehmann, eds.. Biology of Australasian Frogs and Reptiles. Royal Zoological So- ciety of New South Wales, Surrey Beatty and Sons. Chipping Norton. New South Wales. Mertens, R. 1950. Uber reptilienbastarde. Senken- bergiana, 31: 127-144. Minton, S. A., Jr. 1968. Antigenic relationships of the venom of Atractaspis microlepidota to that of other snakes. Toxicon, 6: 59-64. 1976. Serological relationships among some congeneric North American and Eurasian colubrid snakes. Copeia, 1976: 672-678. 1978. Serological relationships of some Phil- ippine sea snakes. Copeia, 1978: 151-154. Minton, S. A., Jr., and M. S. Da Costa. 1975. Se- rological relationships of sea snakes and their evolu- tionary implication, pp. 33-55. In Dunson, W. A., ed., The Biology of Sea Snakes. University Park Press, Baltimore. Minton, S. A., Jr., and S. K. Salanitro. 1972. Se- rological relationships among some colubrid snakes. Copeia, 1972: 246-252. Murphy, R. W., and C. B. Crabtree. 1985. Evolu- tionary aspects of isozyme patterns, number of loci, and tissue-specific gene expression in the prairie rat- tlesnake, Crotalus viridis viridis. Herpetologica, 41: 45 1— 470. Murphy, R. W., F. C. McCollum, G. C. Gorman, and R.Thomas. 1984. Genetics of hybridizing popula- tions of Puerto Rican Sphaerodactylus. Journal of Herpetology, 18: 93-105. Murphy, R. W., and J. R. Ottley. 1980. A genetic evaluation of the leafnose snake, Phyllorhynchus ar- enicolus. Journal of Herpetology, 14: 263-268. Myers, C. W. 1974. The systematics of Rhadinaea (Colubridae), a genus of New World snakes. Bulletin of the American Museum of Natural History, 153: 1- 262. Nei, M. 1978. Estimation of average heterozygosity and genetic distance from a small number of individ- uals. Genetics, 89: 225-233. Neill, W. T. 1949. A new subspecies of rat snake (genus Elaphe), and notes on related forms. Herpe- tologica, 5(suppl. 2): 1-12. Nevo, E., A. Hi ii i s. and R. Ben-Shlomo. 1984. The evolutionary significance of genetic diversity: Ecolog- ical, demographic and life history correlates. Evolu- tionary Dynamics of Genetic Diversity. Lecture Notes in Biomathematics, 53: 1 3-2 1 3. Nuttall, G. H. F. 1904. Blood Immunity and Blood Relationship. Cambridge University Press, London. Olsen, R. E. 1977. Evidence for the species status of Baird's ratsnakes. Texas Journal of Science, 29: 79- 84. Parnas, I., and F. E. Russell. 1967. Effects of venom on nerve, muscle and neuromuscular junction, pp. 40 1- 4 1 5. In Russell, F. E., and P. R. Saunders, eds., Animal Toxins. Pergamon Press, Oxford. Pearson, D. D. 1966. Serological and immunoelec- trophoretic comparisons among species of snakes. Bul- letin of the Serological Museum, Rutgers University, 36:8. . 1968. Immuno-taxonomic relationships among families of snakes. Bulletin of the Pennsylvania Acad- emy of Science, 42: 49-56. 32 FIELDIANA: ZOOLOGY Peters, J. A. 1960. The snakes of the subfamily Dip- sadinae. Miscellaneous Publications, Museum of Zo- ology, University of Michigan, 114: 1-224. Rieppel, O. 1977. Studies on the skull of the Heno- phidia (Reptilia: Serpentes). Journal of Zoology, Lon- don, 181: 145-173. . 1979a. A cladistic classification of primitive snakes based on skull structure. Zeitshrift fur Zoolo- gisch Systematic Evolutions-forschung, 17: 140-150. -. 1979b. The classification of primitive snakes and the testability of phylogenetic theories. Biolo- gisches Zentralblatt, 98: 537-552. Rossman, D. A. 1962. Thamnophis proximns (Say), a valid species of garter snake. Copeia, 1962: 741-748. . 1 964. Relationships of the elegans complex of the garter snakes, genus Thamnophis. Year Book of the American Philosophical Society, 1963: 347-348. -. 1979. Morphological evidence for taxonomic partitioning of the Thamnophis elegans complex (Ser- pentes, Colubridae). Occasional Papers of the Museum of Zoology, Louisiana State University, 55: 1-12. Rossman, D. A., and W. G. Eberle. 1977. Partition of the genus Natrix, with preliminary observations on evolutionary trends in natricine snakes. Herpetologi- ca, 33: 34-43. Rossman, D. A., and G. R. Stewart. 1979. Relation- ships of the Thamnophis couchii complex in northern California. Abstracts, Annual Meeting, Society for the Study of Amphibians and Reptiles and the Herpetol- ogists' League: 53. . 1982. Relationships of the Thamnophis couchii complex in southern California. Abstracts, Annual Meeting, American Society of Ichthyologists and Her- petologists: 24. 1 985. Taxonomic status of the California giant garter snake. Abstracts, Annual Meeting, Society for the Study of Amphibians and Reptiles and the Her- petologists' League: 73. Sarich, V. M. 1977. Rates, sample sizes, and the neu- trality hypothesis for electrophoresis in evolutionary studies. Nature, 265: 24-28. . 1985. Rodent macromolecular systematics, pp. 423-452. In Luckett, W. P., and J. L. Hartenberger, eds., Evolutionary Relationships Among Rodents: A Multidisciplinary Analysis. Plenum Press, New York. Sattler, P. W., and S. I. Guttman. 1976. An elec- trophoretic analysis of Thamnophis sirtalis from Western Ohio. Copeia, 1976: 352-356. Schatti, B. 1985. Morphological evidences for a par- tition of the snake genus Coluber L./Reptilia, Ser- pentes, Colubridae. Abstracts, Third Ordinary Gen- eral Meeting of Societas Europaea Herpetologica: 105. Schenberg, S. 1959. Geographical pattern of crot- amine distribution in some rattlesnake subspecies. Sci- ence, 129: 1361-1363. . 1963. Immunological identification of intra- species qualitative differences in snake venom com- position. Toxicon, 1: 67-75. Schwaner, T. D. 1985. Population structure of black tiger snakes, Notechis ater niger, on ofTshore islands of South Australia, pp. 35-46. In Grigg, G., R. Shine, and H. Ehmann, eds., Biology of Australasian Frogs and Reptiles. Royal Zoological Society of New South Wales, Surrey Beatty and Sons, Chipping Norton, New South Wales. Schwaner, T. D., P. R. Baverstock, H. C. Dessauer, and G. A. Mengden. 1985. Immunological evi- dence for the phylogenetic relationships of Australian elapid snakes, pp. 177-184. In Grigg, G., R. Shine, and H. Ehmann, eds., Biology of Australasian Frogs and Reptiles. Royal Zoological Society of New South Wales, Surrey Beatty and Sons, Chipping Norton, New South Wales. Schwaner, T. D., and H. C. Dessauer. 1981. Im- munodiffusion evidence for the relationships of Pap- uan boids. Journal of Herpetology, 15: 250-253. . 1982. Comparative immunological survey of snake transferrins focused upon the relationships of the natricines. Copeia, 1982: 541-549. Schwaner, T. D., H. C. Dessauer, and L. A. Landry, Jr. 1 980. Genetic divergence ofNerodia sipedon and N. fasciata in South Louisiana. Isozyme Bulletin, 13: 102. Schwaner, T. D., and R. H. Mount. 1976. Systematic and ecological relationships of the water snakes Natrix sipedon and N. fasciata in Alabama and the Florida panhandle. Occasional Papers of the Museum of Nat- ural History, University of Kansas, Lawrence, 45: 1- 44. Selander, R. K., and W. E. Johnson. 1973. Genetic variation among vertebrate species. Annual Review of Ecology and Systematics, 4: 75-91. Simpson, G. G. 1953. The Major Features of Evolu- tion. Columbia University Press, New York, 434 pp. Smith, M. W., C. F. Aquadro, M. H. Smith, R. K. Chesser, and W. J. Etges. 1982. Bibliography of Electrophoretic Studies of Biochemical Variation in Natural Vertebrate Populations. Texas Tech Press, Lubbock, 105 pp. Smithies, O. 1 959. Zone electrophoresis in starch gels and its application to studies of serum proteins. Ad- vances in Protein Chemistry, 14: 65-1 13. Sneath, P. H. A., and R. R. Sokal. 1973. Numerical Taxonomy. W. H. Freeman and Co., San Francisco, 573 pp. Stebbins, R. C. 1985. A Field Guide to Western Rep- tiles and Amphibians. Peterson Field Guide Series, 2nd revised ed. Houghton Mifflin Co., Boston, 336 pp. Strydom, D. J. 1973. Snake venom toxins: The evo- lution of some toxins found in snake venoms. System- atic Zoology, 22: 596-608. . 1979. The evolution of toxins found in snake venoms, pp. 258-275. In Lee, C.-Y., ed., Snake Ven- oms. Springer- Verlag, Berlin. Sutton, D. E. 1969. Fingerprint correspondence of hemoglobin applied to the taxonomy of reptiles. Ph.D. diss., Louisiana State University Medical Center, New Orleans, 85 pp. Tamiya, N. 1985. A comparison of amino acid se- quences of neurotoxins and of phospholipases of some Australian elapid snakes with those of other protero- DESSAUER ET AL.: SNAKE EVOLUTION 33 glyphous snakes, pp. 209-219. In Grigg, G., R. Shine, and H. Ehmann, eds.. Biology of Australasian Frogs and Reptiles. Royal Zoological Society of New South Wales, Surrey Beatty and Sons, Chipping Norton, New South Wales. Tammar, A. R. 1974. Bile salts in Reptilia. In Florkin, M., and B. T. Schecr, eds., Chemical Zoology. Vol. IX, Amphibia and Reptilia. Academic Press, New York. Underwood, G. 1967. A Contribution to the Classi- fication of Snakes. British Museum of Natural His- tory, London, 179 pp. . 1978. A systematic analysis of boid snakes, pp. 151-175. /nBellairs, A. d*A., and C. B. Cox, eds.. Morphology and Biology of Reptiles. Linnean Society Symposium Series, No. 3. Academic Press, London. Watrous, L. E., and Q. D. Wheeler. 198 1 . The out- group comparison method of character analysis. Sys- tematic Zoology, 30: 1-11. Weaver, W. G. 1965. The cranial anatomy of the hog- nosed snakes (Heterodon). Bulletin of the Florida State Museum, 9: 275-304. Williamson, G. K... and R. A. Mouus. 1979. Survey of Reptiles and Amphibians on Fort Stewart and Hunter Army Airfield. U.S. Army, Contract No. 21- 77-C-0 155, 343 pp. Wilson, A. C, G. L. Bush, S. M. Case, and M. C. King. 1975. Social structuring of mammalian populations and rate of chromosomal evolution. Proceedings of the National Academy of Sciences, USA, 72: 5061- 5065. Wilson, A. C, S. S. Carlson, and T. J. White. 1977. Biochemical Evolution. Annual Review of Biochem- istry, 46: 573-639. Wright, A. H., and A. A. Wright. 1957. Handbook of Snakes of the United States and Canada. Comstock Publishing, Ithaca, N.Y. Wyles, J. S., and G. C. Gorman. 1980. The albumin immunological and Nei electrophoretic distance cor- relation: A calibration for the saurian genus Anolis (Iguanidae). Copeia, 1980: 66-71. Yang, C.C. 1978. Chemistry and biochemistry of snake venom neurotoxins, pp. 261-292. In Rosenberg, P., ed., Toxins: Animal, Plant and Microbial. Pergamon Press, Oxford. Zuckerkandl, E., and L. Pauung. 1965. Evolution- ary divergence and convergence in proteins, pp. 97- 166. In Bryson, V., and H. J. Vogel, eds., Evolving Genes and Proteins. Academic Press, New York. Zweifel, R. G. 1981. Genetics of color pattern poly- morphism in the California kingsnake. Journal of He- redity, 72: 238-244. Zweifel, R. G., and H. C. Dessauer. 1983. Multiple inseminations demonstrated experimentally in the kingsnake (Lampropeltis getulus). Experientia, 39: 6- 13. 34 FIELDIANA: ZOOLOGY MM] UNIVERSITY OF ILLINOIS URBANA 590 SFIN t C001 FIFLDIANA ZOOLOGY $ NEW SERIES JCHGO 29 36 1988 87 0112 009378719