U o o -l0) P< o -f Lf rQ <-- ri: r 0'' aq d < l-\ T] U! AE tJ o UJ6 o ct-m xs tT I +.^ )- Q< 6H'n x'r' sin O -' .u,J O rn :_UJ 0-'q o! Qm7 m m ll >H ;#[r =io f o 0)- A-ds 6 f o (! a EE-II 0 CctNTENTEi
€jc,7 I. INTFIOEtUCTIclN 607 II. TEFIMINOLctGtY
VIVIPAFIITY 60E| III. HYPOTHE€tEg] ctN FTEPTILIAN
A. TyPes of HyPolrheses' 608 Factons' 603 B. Hypotheses Based on Envinonrnent'al 61O C. Hypotheses Eased on Species Chanacl'enisl'ics' 13 ctF HYPOTHESES €i IV. AEiSUM]'TION€I ANEI LctGTC
A. Genenal, 613 613 B. Ovipanity as the Pnimitive Condition' C. The Cos1, of ViviPanitY' 6'1 4 615 D. The lrnpontance of Intenmediate Stages' 6'1 6 E. Causes of Egg Montality in the Wild' F. Thenmal Felations of Eggs' 617 G. The Fole of Unpnedictabilitv' 618
HYPCTTHEE;EA 619 V. EMPIFTICAL SUPPGIFTT FOFI THE
A. Genenal, 5'1 g B. The Cold-Climate Hypol;hesis' 619 Hypothesis' 62''l C. The Envinonmental UnpnedicEability D. Othen Envinonrnental Factons' 621 E. Defensive AbilitY' 622 F. Othen Specres Chanactenistlcs' 623 G. Ovenview, 624
\/IVIPAFTITY IN VI. EVGILUTIGINAFTY ctFIIGINEi ctF Ei25 ETOUAMATE FIEPTILE€i
A. Genenal, 625 B. AmPhisbaenia, 625 C. Saunia' 626 tr}. SenPentes, 645 66c, VII. EVALUATIGIN CtF CASE HTEiTclFIIE€i 660 A. Fnequency of Evolution of Vivipanity' E}. Taxonomic Biases, 663 C. Evaluation of HYPotheses' 665 677 vilr. coNcLuEtloNEt Eiao ACKNC'vt,LEEtGIMENTCI 6E'l FTEFEFIENCEEi TEFIMINOLclGY
I. INTFIc'ElUCTIclN
(turtles €j07 Of the four orders of living reptiles, three are entirely oviparous and crocodilians),but approximately one-fifth of all speciesof lizards, 6c,7 snakesand amphisbaeniansare viviparous. Most egg-laying squamates 604 studiedto dateretain eggs in uterofor almostone-half of the totalperiod of embryonic developmentand, thus, have evolved part of the way toward viviparity (shine, 1983a).Clearly, viviparity has arisenmany times (e.g., 60s Tinkle and Gibbons, 7977;Shine and BuLl, 7979),and the dichotomy be- : Al fl tween oviparity and viviparity is a puzzling aspectof reptilian reproduc- 613 tion. Consequently,it has attractedconsiderable attention. (For recent re- views seePackard et al., \977; Tinkle and Gibbons,7977; Shine and Berry, 1978;Shine and Bull, 7979;Pllorge and Barbault,1981; Blackburn, 7982') severalhypotheses have been proposedto explainwhy live-bearinghas evolved;most of theseinvoke factorsthat kill eggs in the nest, but not irl utero,andhence favor the evolution of prolongeduterine retentionof eggs. Theory also suggeststhat certaintypes of speciesshould be most likely to evolve viviparity; specifically,groups in which the survivorship or food intake of the reproducing female would not be markedly affectedif she 619 retained eggsii utero.The present review examinesthe validity of these ideas in tJrms of the arguments and inherent assumptionsin each hy- pothesis,summarizes the empiricalsupport for eachhypothesis (based on publishedliterature), and testspredictions from thesehypotheses by iden- iiryir,g the squamatelineages in which viviparity has evolved-andby look- ing for the ecologicalcorrelates predicted by theory. For this purpose/ u'u"uilubl"phylogenetic and reproductive data on taxa containing both oviparous and viviparous forms are reviewed.
II. TEFIMINClLG,G;Y
Most reptilesreproduce by laying shelledeggs that containrelatively unde- veloped embryos;they are termed oaiparousor egglaying.However, many squamatespecies retain eggsin uterountil embryonicdevelopment is com- pl"tu. tn these casesthe young are fully formed at birth and are capableof independent movement and feeding; this reproductive mode is referred to as liae-bearingor oiaiparitY. Some authors have restricted the term viviparity to species with an intimate physiological connectionbetween the reproducing female and her 677 uterine ybttttg; this may involve complex placentation, absenceof calcified ciao egg-membrar,es,o. maternal-fetaltransfer of nutrients (Weekes, 1935; Bauchot, 1965;Spellerberg, 1976). Although the distinction between ooo- 6Cl1 aiaiparity (no maiernal-fetal nutrient transfer) and euaiaiparitymay be at- tractiveionceptually, it is difficult to apply to reptiles. The availabledata Eiog THE EVGILUTIG'N OF \/IVIF'AFIITY IN suggest that most, but not all, live-bearing reptiles are ovoviviparous (Thompson,7987; Yaron, Chapter 7, this volume). Exceptionsto this state- ment include the New World skinks of the genus Mabuya, n'hich show reduced ovum size and extensivematernal-fetal nutrient transfer (Vitt and Blackburn, 1983). Throughout this chapter, I use the term oaiparitrlin those cases where shelled eggs are laid, and uiaiparity where yourlg are either born alive or are deposited in thin-walled membranous sacs from which they emerge within a few days of parturition. This terminology follon's the recommendations of Smith et al. (1973), Guillette (1982a), and Yaron (Chapter 7, this volume) and corresponds to that used by most authors (e.g., Weekes, 1933; Rahn, 1939; Dumas, 1964; Neill, 1964; Greer, 1966; jenkins and Simkiss , 7968; Greene, 1970; Yaron , 7972; Huey, 1977; Sexton and Claypool, 1978; Thompson, 7987, 7982).
III. HYPGITHESEsi clN FIEPTILIAN VIVIPAFIITY
A. Types of Hypotheees
The selectiveforces underlying the transition from oviparity to viviparity have been a subject of speculationfor many years. The resulting hypoth- "benefits" "costs" esesmay be framed in terms of the relative and of the two reproductive modes to a female reptile, that is, the probablelifetime production of offspring (: fitness)of an oviparous femaleis comparedto that of a viviparous femaleunder a variety of parallelecological conditions. Most authors have emphasizedthe benefits of viviparity in terms of an increasein the number of surviving offspring (e.9., Weekes,1933; Neill, 1964).Eggs retained in uteromay be protectedfrom many sourcesof mor- tality, which they would normally experiencein the nest: for example, extremesof temperature or humidity, fungal attack, and predation. The benefit from viviparity could equally well be the loss of a cost associated with oviparity: for example, the need for reproductive femalesto make long and possiblyhazardous journeys to suitablenesting areas (e.g., Neill, 1964;Fihch, 7970). However, it is misleadingto look only at the benefitsof viviparity and to ignore the associatedcosts. A more balancedview considersthat viviparity may increasethe fitness of a femaleonly under a limited set of condifions (e.g., Fitch, 1970;Tinkle and Gibbons, 7977;Shine and 8u11,1979). Al- though viviparity may benefit survivorship of the offspring, it also may confer a cost to the food intake of the reproducing female,to the probabil- ity of her survival, and to subsequentfecundity. The nature of thesecosts can be used to frame predictions about the types of speciesor habitats in which viviparity would be most likely to evolve (Fitch, 1970;Shine and Bull,7979). In summary, both viviparity and oviparity confer costsand benefits.The HYPGITHESES clN FTEPTILIAN VIVITAFIITY
; are ovoviviparous advantages of viviparity may lie primarily lvith increasing survivorship of eptions to this state- the offspring, whereas its disadvantages may lie with the effects of physi- labuya,which show cally burdening the female with eggs for a prolonged period. The relative rnt transfer(Vitt and importance of these advantagesand disadvantageswill depend upon both rm ouiparityin those environmental conditions and speciescharacteristics. published hypoth- rre young are either esesregarding possible selectiveforces are revierved belorv. ,us sacsfrom which rinology follows the B. Hypotheses Elased on Environm€ntal Factors (1982a),and Yaron ed by most authors 1. COLD CLIMATES ,7964; Greer,7966; The earliest and the most widely accepted hypothesis of reptilian viviparity Huey, 7977;Sexton is that it evolved as an adaptation to cold climates. Tinkle and Gi6bons (7977)traced this idea to Mell (7929) and cite several examples that indicate its currently widespread acceptance. The argument usually runs as fol- lows. In cold climates, behavioral thermoregulation allows the body tem- VIPAFIITY perature of females to be much higher than that of the soil. Thus, eggs retained in utero will develop at higher temperatures (and hence more rapidly) than will eggs deposited in the soil. Early hatching, due to ac- celerated development, may be adaptive in at least the three following iparity to viviparity ways: (1) Eggs may hatch prior to the onset of lethal autumn frosts. (2) e resulting hypoth- ' "costs" Eggs spend less time in the soil in which they may be vulnerabre to factors and of the such as predation and desiccation. (3) Early hatching enables the offspring te probablelifetime to feed and accumulate energy reserves before hibernation, thus increising nale is comparedto their chances of survival over winter and subsequent growth rates. "cold-climate" ologicalconditions. An alternative and simpler version of the hypothesis is :rity in terms of an that soil temperatures at the time of ovulation will be y'eekes, so low as to be lethal 1933;Neill, to developing eggs. Thus, unless body temperatures of gravid females falr rny sourcesof mor- to soil temperatures, uterine retention will protect eggs from these lethal nest: for example, extremes (Shine and Bull, 7979). (The same argument can be applied to rnd predation. The very hot climates in which uterine retention might protect eggs from le- rf a cost associated thally high temperatures; Shine 'e and Bull, 1979). Also, high elevations may females to make play a role in stimulating the evolution of placentation, because low am- ig areas(e.9., Neill, bient oxygen concentrations might favor more efficient oxygen delivery systems to the embryo (Guillette et al., 1980). of viviparity and to Ldersthat viviparity 2. ENVIFIONMENTAL UNPREDICTA.E}ILITY :d set of conditions nd Bull, 7979). Al- It has been suggested that prolonged uterine retention of eggs would be ;pring, it also may most likely to evolve in highly variable environments (Tinkle and Gibbons, "experience ile, to the probabil- 7977). rn such areas, a reproducing female would difficulty in ature of thesecosts predicting, at the time of egg deposition, whether the site chosen would speciesor habitats remain favorable throughout the period of incubation and early life of the :h, 1970;Shine and hatchlings. In such environments selection might favor females which held their eggs through some part of this period of developmental uncertainty. s and benefits.The Cold environments may exacerbate this problem of predictability by in- ci 1Cl THE EVOLUTIGIN CtF VIVIPAFIITY IN FIEtrTTLEA
creasing the length of the incubation period and making it less likely that the egg deposition site chosen by the parent will remain favorable until after hatching. The more unpredictable the environment (whether for rea- sons of climate, predation, or resource availability) of the eggs and hatch- lings the more likely that the complete transition to viviparity will be favored" (Tinkle and Gibbons, 7977).
3. OTHEFI ENVIRONMENTAL FACTC]FIS
Any factor that kills eggs in the nest, but not in utero, might establish a possible selective pressure for the evolution of prolonged oviducal reten- tion of eggs and might favor viviparity. several authors have referred to factors, such as extremes of soil moisture, that may kill eggs and, hence, favor the evolution of viviparity (e.g., Sowerby,7930; Neill, 1964).Labora- tory studies show that substrate moisture levels are important determi- nants of hatching success and also may affect hatchling size (e.g., Muth, 1980;Packard et al., 1982). On the other hand, extreme aridity may prevent the evolution of viviparity, because thinning of the eggshell (assumed to be a necessary precondition for viviparity) is impossible in these environ- ments (Weekes, 1933; Packard, 7966). Another factor is egg predation, the impact of which may be reduced by uterine retention of eggs (Neill, 1964), but this factor has to be balanced against increased predation upon females during pregnancy.
C. Hypotheses Elasad on Speeies Ghanacteristics
1. DEFENSIVE ABILITY It has been suggestedthat viviparity should evolve most often in large or venomous species,because gravid femalesof theseforms are less vulner- able to predation;eggs inside of such femaleswould have a greaterchance of survival than eggs in the nest (Neill, 7964;Piorgeand Barbault, 1981). "costs" Also, such reptileswould have reduced for egg retention in terms of female survivorship (shine and Bull, 7979).rn less formidable species, the physical burden on the female during gestation might lower ]emale survivorship to the point that natural selectioncould not favor prolonged uterine retention of eggs.
2. NONDEPENDENCE ON SPEED OF MOVEMENT viviparity may be more likely to evolve in speciesthat do not depend on rapid movement for feeding or for escapefrom predators(Fitch, 1SZO1.tn these species,the additional physical burden on the gravid female should have less impact on survivorship or food intake. Ambush predators or sluggish herbivorousspecies may fit such criteria.However, lhe common cessationof feeding by gravid reptiles may reducethe importanceof this 'AFIITY IN FIEPTILEEI HYFGITHE€IEET c'N FIEtrTILIAN VIVITAFIITY €11 rg it lesslikely that factor (and the previous one) by tending to equatethe costsof egg reten- .ain favorableuntil tion among differentspecies (Shine and Bull, 1979). rt (whether for rea- he eggsand hatch- 3. ARBOFIEAL OFI AGUATIC HABITS viviparity will be Females of some species migrate to egg-laving sites. This migration mav be hazardous or energetically expensive, and viviparity may confer an advan- ]RS tage, for instance, in aquatic or arboreal reptiles or in habitats where suit, able nest sites are rare. Hence, viviparity might be likely to evolve under , might establish a these conditions (Neill, 1964;Fitch,1970). This argument has been attacked ;ed oviducal reten- strongly (Packard et al., 1977; Shine and Bull, 7979), because it offers no rs have referred to selective advantage for the intermediate stages of prolonged egg retention tl eggs and, hence, leading up to the evolution of viviparity. There are many conditions in leill, 1964).Labora- which viviparity will be favored over oviparity, but in which viviparity will mportant determi- not evolve because no advantage accrues to the necessary intermediate g size (e.g., Muth, stages. Additionally, viviparity seems less likely to evolve in certain ar- rridity may prevent boreal species that rely upon agility in climbing and may be greatly disad- hell (assumed to be vantaged by the weight and volume of the clutch (Shine and Bull, 1979). in these environ- egg predation, the 4. FOSSC]FIIALOF] SECRETIVE HABITS f eggs (Neill, 1964), ation upon females Females of fossorial or secretive species may be rarely exposed to preda- tion; hence, the prolonged physical burden of viviparity might not de- crease their survivorship (Fitch, 1970).This might facilitate the evolution of live-bearing. However, this prediction may be challenged on the following ecteristics three grounds: (1) The assumption of low predation intensity on fossorial species may be false, because it is based on the inability of herpetologists, not natural predators, to find these animals. (2) Fossorial species may st often in large or depend upon slender bodily form for burrowing, and hence the volume of ms are less vulner- developing young would impede the female or would affect the size of the ve a greater chance tunnels needed. (3) The thermal advantages of egg retention may be less nd Barbault,7981). evident in fossorial species, because maternal body temperatures may be retention in terms similar to soil temperatures, at least in some species (e.g., Brattstrom, 7965; ormidable species, Shine, 1983b; but see Huey, 1982). right lower female ot favor prolonged 5. MATERNALCAREOFEGGS
Egg-guarding ("brooding") behavior by females has usually been seen as an alternative to viviparity (e.9., Fitch, 7970; Packard et al., 7977). How- )VEMENT ever, prolonged uterine retention of eggs may be more likely to evolve in do not depend on egg-guarding species; certainly, females of these species face little addi- "cost" rrs (Fitch, 1970).ln tional in retaining eggs in utero (Shine and Bull, 1979). Because egg- lvid female should guarding females generally ceasefeeding, prolonged retention of eggs in- bush predators or creases the period of time during which the female can continue to feed. rever, the common However, the common cessation of feeding in gravid females (Shine, importance of this 1980a) reduces the force of this argument. crF VIVIPAFIITY IN FTEtrTILESI 61e THE EVOLUTIGIN
6. THEFMOF]EGULATOFIY STRATEGY ,,cold-climate,, selection of high The hypothesis relies upon behavioral squamatesmay body temperatures bythe ovigerous female' Heliothermic the substrate especially in maintain body temperatures iruch higher than mav not (Brattstrom' 1965; cold climates, but some thigmotherm]c forms body temperatures actu- Shine, 1983b). In some nociurnal species' female (Huev' 1982); hence' fe- uffy *uy average lower than soil iemperatures rates by being oviparous mites could acc-elerateembryonic developmental more likely to evolve in rather than viviparous! Viviparity should be heliotherms.
7. F]EPtr(]DUCTIVE FFIEC]UENCY "cost" reduced time available to A major of egg retention may be the (Sergeev, 1940;Tinkle and proJ,i." a second clutil within the same season Gibbons,1977).Hence,viviparityshouldevolvemorereadilyinsingle- clutching sPecies.
A. PHYSIOLOGICAL CONSTFIAINTS
Thelackofviviparityinturtlesandcrocodiliansmaybedueto'theirinabil- itytoeffectadaptivereductionsinthethicknessoftheeggshell(Packardet in calcium metabo- al.,7977).This hypothesisis supported by differences lismofdevelopingembryosbetweensquamateandnonsquamatereptiles' implausibleon io*"rr"r, tneiyiotnesis has been critiiized as intrinsically thegroundsthat^somealternatemechanismofcalciumsupplytotheem- retention of devel- Uryo".o.rtahave evolvedif natural selectionhad favored Highly calcifiedshells opr.,g"^U.y os in ut:ero(Tinkle and Gibbons' 7977)' may prevent the evolu- also may reduce gu, "^.huttg e in uteroand' thus' tion of viviparity (Blackburn,1982)' viviparity are possl- other physioiogicalconstraints on the evolution of ble.Forexample,turtteer,tUryosarekillediftheeggs.arerolledoverearly 1984);squamate eggs are in developm"nt 1e.g.,Limpui et al',1979;Ewert' to seehow natural more resilient (Marcelliniana Davis, 1982)'It is difficult embryos in tur- selectioncould favor intra-uterine retention of developing tles;asSoonaStt'"eg8'werelaidtheresultingmovementwouldkillthe to prevent em- "^Utyot. Indeed, turiies aPp91rto have some mechanism from ovipt:lti"g bryogenesisin utero;if gravia femalesare prevented lo-t et al' 1977; ' M ' severalweeks, embryogi,,esisdoes not proieed (Packard ' J L4)' It is difficult to Legler, personal comrirr,,icution; Ewert, 7979'Vol' some other reason, diiinguish causeand effect in such evidence;if., tor there is no selective selectiondoes not favor embryogenesisin utero'then of the eggs advantagein resistanceof devel6pingembryos to overturning (because-thlssituation almost never occurs)' which groups Other types of physiologicalconstraints might determine PAFTITY IN FIEPTILES ci13 AA€|UMPTIGINA ANE! LG,G'IC ctF HYPGITHESEB
EGY among squamatescould evolve viviparity' For example,uterine retention most likely to be favored in taxonomicgroups with rela- rl selection of high of egg"smight be (e.g.,those capable of developingsuccessfully only mic squamatesmay tlveili"deticate"eggs range-oftemperature or humidity). It alsois plausible bstrate especially in under a restricted be barred from evolving egg retention because 't (Brattstrom, 1965; that some squamatesmay temperatureof the femaleis much higher than the soil temperatures actu- the preferreauoay at which-the eggs normally develop.For example,in the v, 1982); hence, fe- temperatures groiiotrt, the preferred temperaturesof individ- bv being oviparous *ontun" lizard,,Sceloporus (Brattstrom,1965), but eggsincubated in labora- likeli' to evol-"e in uals averageover 35"C _the tory fail to develop at temperaturesabove 29"C (Fergusonand Brockman, tstio; Shine, unpublished dutu;. A similar situation occurs in several oviparous Austrilian skinks; however, the mean temperatureof females Y (i.e., allowing for a nighttimedecrease) is within the rangeof embryonic 'd time available to tolerance,ut d rn..urrful oviducal retention of eggs is observed(Shine, ev, 1940;Tinkle and 1983b). re readily in single- Certain sex-determiningsystems may also constrain the evolution of viviparity. For example,prolonged uterine retentionof eggsat maternal body temperaturesmay not evolve if incubation temperaturedetermines (Bull, 1980).Temperature sex determinationhas t- the sex of th" hatchling beenrecorded in a variety of crocodilians,testudines, and squamates(Bull, : due to their inabil- 1980).Similarly, female heterogametymight constrainviviparity because :ggshell (Packardet maternalhormones would tend to influence embryonicsexual differentia- in calciummetabo- tion (Mittwoc}l.,1.975). nsquamatereptiles. callyimplausible on r supply to the em- IV. ASSUMPTIG,NS ANE! LclGiIC c'F HYPclTHESEsi I retentionof devel- ghly calcifiedshells A. General prevent the evolu- Hypothesesconcerning the selectivepressures responsible for the evolu- viviparity are possi- tion of a trait are diffi< to test becausethey concernpast eventswhich rre rolled over early cannotbe directly observed.However, such hypothesescan_be evaluated squamateeggs are and testing their i by examining-Th" their assumptionsand.logic, and deriving t to seehow natural piedictions. fouo-ing section first considers some underlying as- ,ing embryos in tur- iumptions and the logic oflurrent hypotheseson the evolution of vivipar- ment would kill the ity. It then considers-assumptionsand theory relevan-tto specifichypoth- dsm to prevent em- eies. Basicto such considerationsis that the samecondition may havebeen rom ovipositing for produced by different selectivepressures and that presently observedef- Lrdet al.,7977;1. M. iects may have been produced in responseto pressuresno longer opera- 4). It is difficult to tive. some other reason, there is no selective Et. Oviparity as the Prirnitive Gondition :turning of the eggs There is general agreement that viviparity is derived from the primitive 'mine which grouPs conditioriof ovipar:ity, rather than vice versa. This is based on the follow- 61ul THE EVclLUTIctN clF VIVIPAFTITY IN FtEtrTILESt
ing evidence: (1) The vertebrate progenitors of reptiles (amphibians, fishes) are primarily oviparous. (2) Neonates of many viviparous squamates pos- "egg sesssmall teeth"; these occur in all oviparous squamatesand are lost shortly after hatching. They enable the hatchlings to cut or chip the eggshell, but lack functional significance in live-bearers (although one could argue that they are needed to tear the embryonic membranes). If they are nonfunctional, they offer some evidence that the present-dav live- bearers had an oviparous ancestry. (3) It has been argued on rogical grounds that it is simpler to lose a structure (the eggshell) than it is to evolve a new one (Gans, 1974). Hence, an evolutionary transition from oviparity to viviparity may be more likely than the reverse. Phylogenetic trees of squamate groups in which viviparity has evolved may be examined to determine the direction of the suggested transition. If oviparity is the primitive condition, it should be found in species with "primitive" morphological characteristics within each group, whereas "derived" viviparity should be restricted to forms. Among the squamate taxa reviewed later in this chapter, there are 16 casesin which this is clearly true (viviparity derived from oviparity). There are only four cases (in Lerista, Gerrhonotus-Barisia,Platysaurus, and vipera) in which the reverse might be true. However, in these casesthe data are dubious because com- parisons of relatively distantly related species are involved. overall, the evidence supports the assumption that oviparity evolves to viviparity rather than vice versa.
C. The GosG of Viviparity
A critical assumption underlying virtually all life-history theory, including ,,cost,, the hypotheses considered here, is that reproduction imposej a to the reproducing animal. This cost is anything that decreasesResidual Re- productive Value (Fisher, 1958; Williams,1966); for example, the cost may be increased vulnerability to predation or a decrease in feeding ability (and hence in the energy gathered, which could be used. for subiequent clutches). some theoretical , predictions on the evolution of viviparity are based directly on consideration of these costs (e.g., most of the-"species charac- teristics" hypotheses discussed above), whereas others (e.g., the cold- climate hypothesis) do not involve such reproductive costs inlny obvious way. However, all of these ideas assume that reproduction has some cost: specifically, that viviparity is assumed to be more costly to the female than is oviparity. If this is false, all of these ideas on the supposed benefits of viviparity to egg survival would yield the same inaicurate prediction: viviparity should evolve in all reptiles, because it confers a benefit at no cost. The possible costs of viviparity have been couched in terms of female survivorship and subsequent fecundity (Tinkle and Gibbons, 7977). Cost- ACI€IUMPTIG,NEI ANEI LclGIC CtF HYPGITHEEIEEI 615
amphibians, fishes) benefit models have been used to provide an algebraic formulation of the ous squamates pos- problem (Shine and Bull, 1979).Recently, the reasonablenessof the con- "reproductive ramatesand are lost cept of costs" in reptiles has been assessedby studies on to cut or chip the scincid lizards and by a review of published literature (Shine, 1980a). rers (although one Gravid female skinks are physically slowed down by the weight of the ,nic membranes). If clutch. Females with full-term clutches lose about 25Vaol their nongravid he present-day live- running speed. Probably because of this effect, gravid females have been argued on logical shown to be particularly vulnerable to predation by small elapid snakes in gshell) than it is to laboratory trials. Gravid females also bask more often than nongravid ones; ary transition from this behavior may increase vulnerability to avian predators (Shine, 1980a)' /erse. The mobility of gravid Lacertauiuipara is similarly reduced, but they rely iparity has evolved more on crypsis than on flight to escape predation (Bauwens and Thoen, gested transition. If 1981). Although the gravid skinks did not reduce their food intake during nd in species with reproduction, a review of published literature suggested that this is a com- :h group, whereas mon phenomenon (Shine, 1980a). Also, metabolic rates may be much nong the squamate higher in gravid than in nongravid females (Guillette, 1982b). Hence, rep- which this is clearly tilian reproduction probably often involves costs, although their nature rnly four cases (in may vary (Shine, 1980a). which the reverse bious because com- Et. The lmportance of Intermediate Stagee olved. Overall, the 'olves to viviparity The evolution of viviparity must almost certainly have been a gradual process (e.g., Packard et al., 7977; Shine and Bull, 1979). ln the early stages, females carry the eggs long enough to permit some embryonic development in utero but lay the eggs prior to the completion of em- bryogenesis. Continued selection for progressively longer egg retention y theory, including might eventually result in complete intrauterine incubation (live-bearing). "cost" mposesa to The need for intermediate stages is supported by the argument that reasesResidual Re- many physiological and anatomical changes in parent and offspring must mple, the cost may accompany the transition from oviparity to viviparity. These requirements "iu : in feeding ability probably prevent any rapid evolutionary P" from oviparity (without sed for subsequent in utero embryogenesis) to viviparity. For example, retention of the embryo requires vascular structures in the oviduct for gaseous exchange, as well as iviparity are based for thinning of the eggshell. Changes likely to have accompanied the evo- "species he charac- lution of viviparity are reviewed by Weekes (1935), Bauchot (1965), Yaron :rs (e.9., the cold- (1972, this volume), Packard et al. (1977), and Guillette (1982a). osts in any obvious There is relatively little information concerning the extent of embryo- :tion has some cost: genesis in utero in oviparous reptiles. Many authors merely comment that "very r to the female than eggs contain small" embryos when laid (e.9., Ditmars, 1942). How- rpposedbenefits of ever, available evidence suggests that prolonged uterine retention of eggs ccurate prediction: is the rule, rather than the exception, in egg-laying squamates (Shine, fers a benefit at no 1983a).Generally, eggs are retained in utero for almost one-half of the total period of embryonic development, to the equivalent of stage 30 for I'acerta in terms of female aiuipara in the table of Dufaure and Hubert (1,961). In contrast, turtles rbons, 1977).Cost- appear to retain eggs only briefly between ovulation and oviposition. At 616 THE EVC,LUTIGIN ClF VIVIPANITY IN FIEPTILECI the time of laying, embryos are no further advanced than the gastrula stage (Shine, 1983a; Ewert, Vol. 14). Further studies on this topic would be of great value. In order for viviparity to arise from oviparity, some advantage must accrue to an intermediate reproductive strategy in which females retain eggs in utero for progressively longer periods of time. Hence, theory pre- dicts that the conditions under which viviparity evolves should be those favoring a progressive increase in the duration of intra-uterine retention of eggs. Sucl-rconditions may differ from those under which viviparity, as such, is favored. In particular, several situations might confer an advan- tage to viviparity, but not to intermediate stages of egg retention (Weekes, 1935;Tinkle,1967; Packard, 1966;Packard et al., 7977;Tinkle and Gibbons, 1977; Shine and Bull, 1979). For example, viviparity might be advantageous to females under conditions in which suitable nesting sites or calcium supply are limited. However, these circumstances do not confer an advan- tage to the intermediate strategy of prolonged intra-uterine retention of eggs. The same argument applies to the avoidance of nesting costs, for instance those imposed on ovipositing females that make long migrations to nesting areas (energetic costs) or are vulnerable to predators (survivor- ship cost). Hence, it is important to distinguish between the evolutionary origin of viviparity (i.e., circumstances under which selection favors intra- uterine retention of eggs in oviparous forms) and the subsequent radiation of viviparous species into new habitats (i.e., those under which selection favors viviparity over oviparity), Some authors have failed to make this distinction. Consequently, their hypotheses relate to reasons for the present-day distribution of viviparity, rather than to reasons for its origin (e.g., the hypothesis that viviparity is favored in aquatic or arboreal species). Hypotheses that are equally appli- cable to both phenomena include the cold-climate hypothesis, unpredict- ability, defensive ability, nondependence on speed of movement, and fos- sorial or secretive habits. Intermediate stages of egg retention may not be favored merely because of high rates of egg predation. In this case, the time course of mortality is critical. If predation is concentrated in the period just after oviposition, and is relatively independent of the duration of time eggs spend in the nest, brief oviducal retention of eggs would reduce incubation period but would not reduce mortality due to predation (Shine and Bull, 1,979).Unf.ortunately, few data are available on egg suwivorship in natural nests, so that the importance of this objection is difficult to judge.
E. Gauses of Egg Mortality in the Wild
There is remarkably little information on rates or determinants of egg sur- vivorship in natural nestsof squamatereptiles; more data are availablefor turtles (e.g., Moll and Legler, 1971.;Burger,7977)and for crocodilians(e.g., Webb et al.,7977). /II'AFIITY IN FIEPTILEEi 617 ASSIUMPTIG'NS ANE, LOGIC clF HYPG'THESEST ranthe gastrulastage Incomplete development and death by frost is suggested,but not re- is topic would be of quired, by the cold-climatehypothesis for eggs depositedin marginal habitats.Such reduced survivorshiphas been documentedin three Euro- ,me advantagemust pean reptiles,Podarcis muralis, Lacerta uiridis (Cooper, 1965), and Nafrlr vhich females retain natrix (srnith, 7973).when eggsof an oviparouslizard (Phrynosoma)were . Hence,theory pre- transferredto the cold-climatehabitat of a viviparous congener,the em- ves should be those bryos died due to low temperaturesbefore completing development r-uterineretention of (Dumas,1964). Eggs of severalcold-climate species of Australianlizards which viviparity, as only hatchshortly before the onsetof lethallylow temperaturesin autumn ;ht confer an advan- (Shine,i983b). g retention(Weekes, Squamateeggs are vulnerable to many other sourcesof mortality, in- Tinkle and Cibbons, cluding predation;several groups of fossorialsnakes are obligateegg eaters ght be advantageous (Broadley,7979; Shine,1984). A single speciesof snake, Saluadoragrahami ing sites or calcium Iineata,is the most important causeof egg mortality in the lizard Sceloporus not confer an advan- oliaaceus(Blair, 1960).Other causesinclude desiccation,which has been uterine retention of shown to be important in an unusually dty year (Blair, 1960).Eggs of the cf nesting costs, for desertiguana, Dipsosaurusdorsalis, develop successfullyunder a restricted rakelong migrations range of conditions of temperatureand humidity; suitableconditions are predators(survivor- found only in a limited geographicarea and in specific microhabitats :en the evolutionary (Muth, 1980;Porter and Tracy, 1983).Thermal sensitivity of embryosalso :lection favors intra- may constrain nesting sites and seasonsof other iguanids (Rand, 1972). ,ubsequentradiation predationby ants may kill many eggsof anoline lizards (Andrews, 1982). rder which selection Although hatching successin natural nests may often be high (Van Devender and Howard, 1973;sexton and Claypool, 1978),there is little Consequently,their doubt that there are many differentcauses of mortality. Information on the bution of viviparity, relativeimportance of eachcause, and the time-courseof mortality would rsisthat viviparity is allow assessmentof the probable effect of a slightly shorter incubation at are equally appli- period on egg survivorship. All other things being equal, selectionwould pothesis,unpredict- iavor oviduial retention of eggsif a reduced incubation period increased movement,and fos- egg survivorship. :tention may not be rn. In this case,the F. Thermal Flelations of Eggs rtrated in the period the duration of time The cold-climatehypothesis requires that eggsretained within the body of eggs would reduce the femaleare kept warmer than eggslaid in the nest and that this temper- to predation (Shine ature differet ce has a significanteffect on the rate of embryonic develop- on egg survivorship ment and, hence, on tolal incubation period. The former assumption is :ction is difficult to likely to be valid for most heliothermicspecies in cold climates,but it is unlii Gi. The Flole of lJnPredictability The unpredictabilityhypothesis suggests that viviparity should evolve in cold climates, becausecold climates are variable, exposing the eggs to many unpredictablesources of mortality; cold climatesalso lengthen the period of incubation, so that the eggsare vulnerablefor a longer period of time (Tinkle and Gibbons,1977).Although the unpredictabilityhypothesis is highly original, I have doubts about its validity. The level of predictabil- ity should not affect selectionfor egg retention unless the femalehas the €t19 EMPIFTICAL €IUPPOFIT FOFI THE HYPOTHEAES if it is 'd embryogenesis in option of adopting alternative egg-laying strategies. For example, where will be great enough going to rain, the female must deposit the eggs in higher locations they will not be inundated; during drought she should deposit them on low ground where they may stay moist. A female must then be pro- ne, 1983b) suggests "knows" rrencebetween eggs grammed to predict the probability of rain or retain eggs until she may 7"C in heliothermic whether or not it will rain; in this case, environmental unpredictability favor egg retention. However, if some important et'ent occurs both infre- ; retention for about retention. In- rbation period from quently and unpredictably, selection is unlikely to favor of nic development at creased predictability of the event is likely to increase the advantage may some- ceas rapid as at nest facultative retention of eggs. This means that unpredictability times select against retention instead of favoring it. to evolve in re- ds on the tempera- In any case, long durations of retention are unlikely The longer re nest. It has been sponse to this kind of unpredictability (Guillette et ai., 1980). must pay for the select the warmest the mother retains the eggs, the higher the cost that she of the event ol, 1978);such nest- advantage of knowing whdt to do. unless the occurrence the mother ion on the duration makes a vast difference in the suitability of a given nest site, soon after is has the problem would benefit from a compromise strategy and lay the eggs as such re warmer than the ovulation. There is no compelling reason why unpredictability unpredict- allow egg survival. should be important in the evolution of live-bearing, although rm microhabitats as ability may favor short intervals of egg retention' "ir in decomposing likely that most se- SUPPG,FTT FG'FI THE HYPG,THESES 'ise suitable for egg V. EMPIFIICAL as high as the body A. General esis is that the rela- have some degree of empirical support, but in rr any given species All the listed hypotheses merely point to a particular viviparous species and 3rwise, selection in many cases authors hypothesis explains why viviparity confers a selective evelopment at low suggest that their This approach has been criticized as providing :ed without uterine advantage to that species. for the hypothesis; an array of oviparous species can be development rates only weak support ecological characteristics similar to those of members phibians; the rapid assembled that show array (Tinkle and Gibbons, 7977) so that the criticism is s perhaps the most of a viviparous problem stems from failure to consider the origin :celerated develop- valid. Another common of viviparity separately from the subsequent maintenance of the trait. Gold.Climate Hypothesis ,Y El. The climate and the incidence of viviparity is well should evolve in The correlation between ry cases,some variable related to temperature(either rosing the eggs to documented. In most has been shown to correlatewith the proportion of s also lengthen the latitude or elevation) which is viviparous. Viviparity is proportionately more r a longer period of the squamatefauna latitudes in Eurasia, China (Sergeev,1940), North ctability hypothesis common at higher (Tinkleand Gibbons, 7977).Also, the proportion of level of predictabil- America,and Australia increaseswith increasingelevation in Austra- the female has the viviparous squamatespecies THE EVCILUTIGIN c,F VIVItrAFIITY IN FIEPTILEA lia, Europe (Weekes, 1933; Sergeev,1940), East Africa (Greer, 1968)and Mexico (Duellman, 1965;Greene,7970). Within the lizard genus Sceloporus, viviparous forms tend to be found at higher elevations (and perhaps higher latitudes) than oviparous ones (Guillette et al., 1980).This type of analysis has been extended by comparison of the incidence of viviparity with a climatic measure (average summer temperature) rather than simplv with latitude. As expected, viviparous species comprise a higher proportion of the squamate fauna in colder areas (Tinkle and Gibbons, 1977). The inci- dence of viviparity also has been correlated nitl-r a variety of climatic esti- mates (Shine and Berry, 1978). In both Australia and North America, there is little relationship between temperature and percentage of viviparity over a wide range of relatively warm climates, but the proportion of viviparous species rises rapidly in extremely cold areas, in which mean midsummer minimum air temperatures average lower than 15'C (Shine and Berry, 7978). Great emphasis has been placed on the fact that, although the proportion of viviparous species is highest in cold areas, the nbsolutenumber of vivipa- rous species is higher in the species-rich warm climates (Tinkle and Gib- "Proportions bons,7977).It is suggestedthat can be misleading becausefar northern or southern and high altitude faunas are exceedingly impover- ished . . an explanation for the evolution of viviparity should also con- sider those situations in which the majority of viviparous species occur" (Tinkle and Gibbons, 7977). This does not seem to be an important criti- cism. Theory attempts to explain reasons for the successof viviparity rela- tive to oviparity. The tendency for species richness to increase toward the kopics is a common phenomenon in many groups, and I doubt that it has much to do with the evolution of viviparity. Several authors have pointed out that the overall present-day distribu- tion of viviparity may tell us very little about factors favoring the origin of viviparity (references above). An alternative approach has restricted atten- tion to (1) oviparous species showing prolonged oviducal retention of eggs and to (2) live-bearing species within genera with both oviparous and viviparous forms (Shine and Bull, 1979). This procedure was designed to "recentf' find origins of viviparity. The approach assumes that the ecol- ogies of these live-bearers (or oviparous egg-retaining species) may still represent the environments that favored the evolution of viviparity. One "recent" problem with this criterion for evolution lies in the assumption that speciation rates are similar in different taxonomic groups. Unless this is true, viviparity may be of very ancient origin within some genera; the group may have remained relatively unchanged over a long period. On the other hand, divergences between wholly viviparous and wholly oviparous genera may be of recent origin in a rapidly speciating group. "cold" The proportion of species inhabiting climates (using a subjective "cold") judgment of among squamates in general (447o) is lower than "recently among evolved" live-bearers (28 of 13 cases, or 85Vo)or among /IFAFIITY IN FIEPTILESI EMPIFIICAL EiUT'PGIFIT FclFI THE HYPGITHESES 6e1 :a (Greer, 1968) and oviparous specieswith prolonged egg retention in utero (72%) (Shine and lrd genus Sceloporus, 8u11,7979). A similar analysis on 75 taxa (Blackburn, 1987, 7982) reached (and perhaps higher the same conclusions. These data support the hypothesis that viviParitv "colc1" Tl'ristype of analysis evolves in coid regions. However, the subjective evaluations of "prolonged" of viviparity n'ith a climates, and oviducal retention of eggs represent major er than simply rvith weaknessesof the analyses.A stronger test of the prediction that viviparity righer proportion of evolves in cold climates is to compare climatic conditions in the habitats of tns, 7977). The inci- closely related oviparous and viviparous forms. This procedure was ap- rietv of climatic esti- plied to sceloporinelizards (Guillette et al., 1980)and to taxa involved in 71 "cold Jorth America, there evolutionary origins of viviparity (Shine, 1981), and the climate" rgeof viviparity over hypothesis was supported in both cases.The present revierv provides the rortion of viviparous detailed results and analvsis of my earlier abstract (Shine, 1981). h mean midsummer l (Shine and Berrv, C. The Environrnental Unpredictability Hypothesis hough the proportion The unpredictability hypothesis is a very general one and is of little use in ute numberof vivipa- deriving predictions unless one is willing to specify the particular variables tes (Tinkle and Gib- for which predictability is to be important. The variability of many sleadingbecause far significant factors, for example, rainfall, is likely to be no greater in cold E. Elefensiwe AbilitY Viviparity should evolve in large and venomous species, because the pro- longed retention of offspring would not lower female survivorship, and eggs in utero would be very safe (Neill, 1964; Shine and Bull, 1979)' There are many examples of species that combine viviparity with toxic venom or large body size (Neill, 7964; Fitch, 1970)' However, it is possible to counterpose an array of oviparous species with the same characteristics (Tinkle and Gibbons, 1977); examination of selected examples does not provide any useful test of such an idea. ,,deiensive The ability" hypothesis applies to the maintenance as well as to the origin of viviparity. The hypothesis predicts an association between viviparity and venomosity in snakes. Seventy-two percent of the 212 non- .reno*o.,s snakes for which reproduction data have been presented (Fitch, 1970) are oviparous, whereas only slightly more than half (54%) of the 139 venomous species are oviparous. This difference is significant (with the : assumption that each species represents an independent data point: r 951,x2: 10.8, L d.f., p < .07). These data support the hypothesis that viviparity is likely to be favored in species in which the gravid females are relatively invulnerable to predation. IPAFIITY IN FIET'TILEA EMPIFIICAL SIUPPclFIT FG'FI THE HYPOTHE=|ES| r the predicted char- F. Other Species Gharacteristics imate" prediction is "Recently-evolved" :limates. A detailed viviparous speciesdo not exhibit a higher frequer-rcvof rs that its viviparous arboreal or aquatic habits than do squamates in general (Shine and Bull, nore mesic habitats 1979). However, a greater fraction of the viviparous species of Sceloporusis :r, climatic variables arboreal or saxicolous than is true for the oviparous species, but such 1 (Shine and Berry, arborealitymay have evolved as adaptation to viviparity (for thermoregula- tats tend to be moist tion), rather than the reverse (Guillette et al., 1980). ated factors are best The hypothesis that maternal brooding facilitates the evoiution of .stureare not associ- viviparity is supported by a high frequency of origins of viviparity in gen- ity should evolve in era that also contain brooding speciesand by an apparent tendency for s well as cool moist prolonged uterine retention of eggs in brooding squamates (Shine and "prolonged" variable, viviparity BulI, 7979). However, it is difficult to define retention of eggs tats. The consistent and this associationmay not be valid (Shine, 1983a).Also, the link between rtly independent of brooding and the evolution of viviparity is open to another interpretation. ture is more impor- Uterine retention of eggs, viviparity, and maternal care of eggs ali involve rort the hypotheses increased parental investment. The environmental conditions or species rls or high tempera- characteristics favoring such investment might favor anv of these strate- at these hypotheses gies. Hence, the correlations between brooding and viviparity may relate e factors may not be to common causation rather than to a single factor (brooding) that proto- z and of high preda- adapts for another (viviparity). The hypothesis that viviparity evolves primarily in species that produce "recently- a single clutch per year is consistent with the tendency for evolved" live-bearers to inhabit cold climates, because single-clutching is usual in these areas (Fitch, 1970). However, single-clutching also is com- mon over most regions except the tropics (Fitch, 1970;Tinkle and Cibbons, rs, becausethe pro- 79n). ff single-clutching were an important constraint, viviparity should r survivorship, and evolve over a wide variety of climatic conditions and not be restricted to I Bull, 1979).There severely cold areas. The other hypotheses on species characteristics (non- r with toxic venom dependence on speed, fossoriality, secretive habits, heliothermy) are more :r, it is possible to difficult to quantify and have not been tested. ramecharacteristics It is difficult to derive specific predictions from the hypothesis that phys- examplesdoes not iological constraints may prevent the evolution of viviparity in certain taxa. One approach has been to look for a taxonomic correlation between intenanceas well as the incidence of viviparity and the incidence of prolonged retention of eggs rssociationbetween in utero on the grounds that a physiological inability to retain eggs should :ent of the 212non- be evident in both sets of data. Hence, one might expect that viviparity tn presented(Fitch, would be most frequent in taxonomic groups in which egg retention has alf(54Vo)ofthe 139 often appeared (Blackburn, 1982). This prediction was confirmed by an .gnificant (with the analysis of lizard families (Blackburn, 1982).However, Blackburn's conclu- : "prolonged" :nt data point: n sion may be falsified as it relies upon subjective evaluations of ;he hypothesis that egg retention. Objective data on embryonic developmental stages invali- rgravid femalesare date the differences upon which the analysis relies (Shine 1983a). The mode of sex determination is another physiological constraint that also e,24 THE EVctLUTIctN OF VIVIPAFIITY IN FIEPTILECT may play some role in the evolution of viviparity; however, data are inade- quate to reach any conclusions at present (Bull, 1980;Blackburn, 1982). G. Clverview The selectivepressures responsible for the evolution of reptilian viviparity have been a subject of vigorous debate for over half a century (e.g', Mell, 1929;Weeke s, 7933;Sergeev, 1940). Some of the ideas proposed lack gener- ality, or require a quantum phenotypic change before they are adrranta- geous, with the intermediate stagesneutral at best. The hypothesis that reptilian viviparity has evolved as an adaptation to cold climates is the onlV one that has achieved general acceptance (Tinkle and Cibbons, 7977). There are at least three reasonswhy it may be more powerful than alterna- tive ideas. L. Uterine retention of eggs speeds embryonic development in cold cli- mates because the body of the female can be warmer than the soil. A relatively short retention period can greatiy decrease the total period required for incubation. This would reduce the mortality from those factors that would kill eggs in the nest. 2. Geographic clines of temperature are common and usually reflect latitude or elevation. If a species invaded colder areas, there would be an advantage to individuals that tended to retain their eggs longer, facilitating the evolution of viviparity. Clines in other variables, such as intensity of egg predation, may be less common. 3. Any uterine retention of eggs would be advantageous in cold climates; hence, the intermediate stagesas well as full viviparity would be adap- tive. "cold-climate" The hypothesis also is supported on empirical grounds. Data on montane Australian skinks suggest that the major assumptions of the hypothesis are realistic (shine, 1983b). Present-day distributions of viviparous species are highly correlated with environmental temperatures (e.g., Tinkle and Gibbons,1977; Shine and Berry, 1978). Prolonged reten- tion of eggs may be more common in colder areas (Huey, 7977; Shine and BluJl,197i; Guillette et al., L980; but see Shine, 1983a). Recently-evolved viviparous species also tend to be found in cold areas (shine andBuIl,7979; Blackburn, 1982), although this conclusion is weakened by its reliance on "cold" subjective evaluations of and on broad comparisons among squa- mate groups that are only distantly related. Speculations on species characteristics likely to favor the evolution of viviparity have rarely been tested. Available data suggest that maternal egg brooding may protoadapt for the evolution of viviparity (Shine and nu,ll, tOZSland that there is an association between venomous capacity and viviparity among present-day snakes (see above). IPAFIITY IN FIEPTILE€I EVGILUTIGINAFTY GIFTIGINS clF VIVIPAFIITY IN SGIUAMATE FIEPTILEEi ever, data are inade- VI. EVOLUTIGINAFTY GIFIIGINS clF VIVIPAFIITY IN Blackburn, 1982). SGIUAMATE FIEPTILES A. General rf reptilian viviparity The preceding sectign shows that there are mat-Iy alternative hvpotheses century (e.g., Meli, and predictions on the selectiveforces favoring the evolution of viviparitv, :roposed lack gener- but few satisfactorv attempts to test these ideas. This chapter reanalvzes 'e they are advanta- the data, first identifying the reptilian taxa in which viviparity has evolved The hvpothesis that and next comparing these groups with the remaining squamates to see if I climatesis the only they show the characteristicspredicted by theory. This approach has been ,nd Gibbons, 7977). applied previously to origins of viviparity occurring within squamategen- rwerful than alterna- era (Shine and Bull, 7979) and to all identifiable origins of squamate viviparity (Shine, 1981; Blackburn, 1981, 1982. The present analysis was performed independently of the latter two studies; discrepanciesare dis- rlopment in cold cli- cussed below). mer than the soil. A The analysis starts with a review of all squamate taxa in which viviparity :ase the total period has evolved. Lineages for which few data are available, or within which nortality from those considerable speciation has occurred subsequent to the evolution of viviparity, are noted without discussion. The approach is conservative: it is and usually reflect designed to assess the minimum number of independent evolutionary reas, there would be origins of viviparity. n their eggs longer, Geographic distributions have been determined for most taxa, as have rer variables, such as climatic conditions over the range of each species. Judgments as to whether a species occupies a colder or hotter climate than another are ous in cold climates; based on mean midsummer temperatures over the geographic range of uity would be adap- each species (data from Arakawa, 1970; Gentilli, 1971; Wallen, 7971; Griffiths, 1,972;Bryson and Hare, 1974; Schwerdtferger, 7976; Landsberg and Wallen, 7977; Lydolph, 1977; Takahashi and Arakawa, 1981).Judg- "wet" "dry" empirical grounds. ments of or soil were based on habitat preferences of the ,ajor assumptions of species, in combination with broad climatic averages. These procedures lay distributions of offer only a rough approximation to conditions of temperature or moisture nental temperatures under which eggs develop, but they may be useful to compare geographic 3). Prolonged reten- distributions of related species (rather than as any absolute estimate of soil rcy,7977;Shine and temperatures or humidity). r). Recently-evolved ihine and Bull,1979; E!. Amphisbaenia ld by its reliance on risons among squa- Reproductive modes are known for only a few species (Gans, 1978). Viviparity has been recorded in Trogonophiswiegmawri (Trogonophidae: avor the evolution Bons and Saint Girons, 7963),whereas both oviparity and viviparity occur rggest that maternal in the Amphisbaenidae. oviparity characterizes Rhineurafloridana (Gans, viparity (Shine and 7967),Dalophinellenbergerl(BroadleyetaL.,7976),D.pistillum,andChirindia omous capacity and ewerbecki(Loveridge, 1941), whereas Monopeltis capensisis viviparous (Vis- ser, 7967). Prolonged uterine retention of eggs may occur in Rhineurct 6AGi THE EVGILUTIG,N c,F VIVIPAFIITY IN FIEPTILEEI (Gans, 7967).At least two origins of viviparitv are suggested, one in the Trogonophislineage and another in the Dalophia-Monopelfisgroup. Data are insufficient for further analysis. C. Sauria 1, AGAMIDAE Only two agarnid genera contain viviparous specics.Co\thotis includes the oviparous C. sumatrannand the slow-moving arboreal viviparous C. ccy- lanicaof the Sri Lankan mountains (Willey, 1906;Smith, 1935;Tryon,7979)' The latter may be phylogenetically closer to the other earlessdragons of Sri "C. " Lanka Gyriocephalusand Cerqtophora)than to sunntrana (S. Moodv, personal communication). In Sri Lanka, the oviparous genela inhabit lower elevations than the viviparous form (up to 2100 m; Smith, 1935). "toad-headed" The small agamids of the genus Phnlnocephalusinhabit steppes and deserts of central Asia (for ecology see Rustamov and Sham- makov, 7967; Shammakov et a1.,7973).Most are oviparous (i.e', P, eup- tilopus, P. guttatus,P. helioscopus,P. interscapularis,P. Iuteoguttatus,P. maculatus,P. mystaceus,P. ornatus,P. reticulstus,P. scutellatus-Pope,7935; Smith, 1943; Terentev and Chernov, 1965; Minton,7966). However, P/rry- nocephalustheobaldi is found at higher elevations than any other reptile (above 5000 m, Daan, 1968) and appears to be oviparous at low elevations but viviparous at high ones (sergeev, 1940). Other high-mountain species of Phrynocephalusmay be viviparous also (Daan, 7968), suggesting at least two origins of viviparity in the genus. However, the present review treats this as a single origin, because detailed data are not available. P. ANGUIDAE The subfamily Anguinae consists of two genela that are limbless, primarily terrestrial, and often associated with grassy habitats. The European genus Anguis has only a single and viviparous species, whereas the ten species of the Eurasian and North American genus Ophisaurusare oviparous. Mater- nal brooding behavior is common in Ophisaurus (Noble and Mason, 1933); this may have protoadapted the anguines for the evolution of viviparity (shine and Bull, 7979).The viviparous Anguis occupies cooler climates than its oviparous relatives (Fig. 1); it is often found at high elevations (up to 2400 m, Petzold, 1968a). The elongate galliwasps (Diploglossus)of Central America have both oviparous, egg-guarding species (D. bilobatus,D. delasagrdand at least one viviparous form (D. pleei, Greer, 7967a). Other viviparous diploglossines recently have been transferred to the genus Celestes(C. costatus,C' uus' culus, C. curtissi;Greer, 1967a;Strahm and Schwartz,7977)' We lack data on the reproductive modes of most species. Phylogenetic reconstructions suggest that the viviparous genera Celestesand Ophiodesarose indepen- VIPAFIITY IN FIEPTILES ,uggested,one in the ryeltis group. Data are =t) Co1tltot is includes the ij- :al viviparous C. cetl- h, 1935;Tryon,1979). earlessdragons of Sri ;umatrana(S. Moody, ; generainhabit lower imith, 1935). )hrynocephalusinhabit Rustamovand Sham- ;1 i ziparous(i.e., P. eup- -N , P. luteoguttatus,P, t* utellatus-Pope, 7935; 'lcNbo 966).However, Phry- ran any other reptile :ousat low elevations igh-mountain species 'tPo.> 3), suggestingat least .,e presentreview treats O:P available, r' !oa ! v -J t_ o 6- =No rrelimbless, primarily .90A *.; P The Europeangenus :.! .x reasthe ten speciesof xLoFSE. areoviparous. Mater- rl-- @ 1933); :le and Mason, i^F:3E volution of viviparity bo: > rscooler climates than righ elevations(up to I America have both ;agra)and at leastone farous diploglossines ; (C. costalus,C. crus- :, 1977).We lack data :netic reconstructions hiodesarose indepen- IHE EVclLUTIclN clF IN ]:IEPTILES dently from a pro-Diploglossasstock (Strahm and Schwartz, 1977), suggest- ing two further origins of viviparity within the diploglossine radiation. The gerrhonotine anguids contain both oviparous and viviparous forms witl-rin Gerrhonotus,and the viviparous Altronig (Stebbins, 195t1;Fitch, 1970).Mode of reproduction has been used to separatethe oviparous sub- genus Gerrhonotus(G. cedrosensis,G. khryi, G. multicarintttus,G. parnntitt- tinus, G. paucicarinatus)from the viviparous Barisis(G. coertlaus,G. gadoui, G. inrbricatus,G. monticolos,G, rnoreleti-Fitch, 1970).Taxonomic studies suggest that G. coeruleusis the viviparous form most closely related to the oviparous ones (e.g., Criley, 1968;Waddick and Smith, 1974).This vivipa- rous speciesinhabits cooler and moister areasthan do its oviparous conge- ners (Fig. 1). It may be that viviparity arose at least twice within the gerrhonotiform lizards. Barisia is thought to be the most primitive subgenus, from which Gerrhonotushas been derived (Waddick and Smith, 7974). AII present-day Barisia are viviparous, leading to the following possible postulates: (1) oviparous Gerrhonotusare derived from viviparous ancestors (a position considered unlikely), (2) ancestral Barisio were oviparous (viviparitv evolved within the genus after which the oviparous speciesdisappeared), or (3) the phylogenetic hypothesis of Waddick and Smith (7974)is false. An alternative phylogeny based on electrophoretic and morphological studies of the gerrhonotines also suggests two independent origins of viviparitv (D. A. Good and |. W. Wright, personal communication). 3. ANNIELLIDAE, XENOSAUtrIDAE Both of the small families, Anniellidae and Xenosauridae, consist entirely of viviparous species (Fitch, 1970). Probably, both are derived from primi- tive anguids (McDowell and Bogert, 1954; Scherpner, 1968; Bezy et al., 1977). Because viviparity is found in several anguid lineages, it may have been inherited rather than evolving separately in the Anniellidae and Xenosauridae. Hence, the conservative approach is to omit these taxa from further analyses. 4, CHAMAELEONIDAE The genus Chamaeleocontains approfmately 70 species distributed through Africa, Madagascar, and western Europe. Several authors have attempted to reconstruct the phylogeny of the group, mainly on lung anatomy, cytology, and cranial morphology (e.g., Hillenius, 7959; Rand, 1963; Visser, 1972; Klaver, 7973, 1,977;Raw, 7976, 1978). Reproduction is described by Bons and Bons (1960) and Blanc (7974), but detailed data are lacking for most species. Most chameleons are oviparous, but the eastern African C. bitaeniatusgroup and the southern African C. pumilis grouP (Bradypodionof Raw, 7976,7978) are viviparous. These two grouPs are only distantly related to each other (Klaver, 1977,1981), so it seems that vivipar- ity arose twice. 'AF|ITY IN FIEtrTILES VIVItrAFIITY IN SIGIUAMATE FIEFTILEA €t29 rtz, 7977),suggest- The viviparous C. bitaeniatusgroup (C. hitaeniatus,C. jacksonii,C. rudis, lssine radiation. C. fuelleborni,C. tempeli,C. werneri, C. hoehnelii)may derive from the C. d viviparous forms johnstoni group (C. monachus,C. namaquensis,C. melleri, C. cristatus,C. rbins, 1958;Fitch, montium, C. wiedersheimi,C. oueni, and C. johnstoni),all of rvhich probably the oviparous sub- are oviparous (Klaver, 1977).The viviparous forms occupy high elevations inatus,G. panamin- (C. bitaeniatilsis found at over 3000 m), whereas most of the C. johnstoni :oeruleus,C. gadorti, group live beiow 1700 m (Klaver, 7977;Fig. 2). A more rigidly cladistic Taxonomicstudies analysis (Klaver, 1981) suggeststhat a subset of the C. bitscniatusgroup oselvrelated to the above (the C. werneri group) is closest to the lineage in which viviparity "group 1974).This vivipa- evolved (Klaver's, 1981, E"). :soviparous conge- The second group of interest contains eight apparently oviparous species(C. nasutus,C. t'allax,C. gallus,C. boettgeri,C. guibei,C. Iinotusfrom the gerrhonotiform Madagascar, and C. tenuis and C. spirtosusfrom eastern Africa), plus the "C." genus, from which viviparous pumilis group from southern Africa. The pumilis group has 4). All present-daY recently been divided into 11 species and elevated to generic level (Brady- ible postulates:(1) podion:Raw,7976,1978),but appearsrelated to the C. nasutusgroup on the rcestors(a position basis of lung structure and karyology (Klaver, 1973). However, the C. iparous (r'iviparitY nasutusgroup may itself be polyphyletic (Klaver, 1981).The viviparous eciesdisappeared), forms range further south than their oviparous relatives (Fig. 2; Schmidt "the h (1974)is false.An and Inger, 7957). It has also been suggested that perils of a descent to rrphologicalstudies the ground (for egg-laying) are escaped by the few chamaeleons that are rrigins of viviparitY viviparous and are therefore able to carry out all reproductive functions in )n). the relative safety of bushes and trees" (Schmidt and Inger, 1957). Al- though viviparity could be advantageous in arboreal species, it remains \E unlikely that it evolved because of arboreal habits (Shine and Bull, 1979). lae, consist entirely lerived from primi- 5. CORDYLIDAE , 1968; Bezy et al., The two subfamilies (Gerrhosaurinae and Cordylinae) of cordylids are reages, it may have small- to medium-sized heavily armored African and Madagascan lizards. re Anniellidae and The gerrhosaurines are all oviparous, whereas the cordylines include lmit these taxa from both oviparots (Platysaurus) and viviparous (Cordylus, Pseudocordylus, Chnmaesaura)genera (Broadley, 7974). Chamaesaurais elongate and almost limbless. The geographic range of the viviparous speciesencompasses that of the oviparous forms (Fig. 2). species distributed :veral authors have 6. GEKKONIDAE ,p, mainly on lung lenius, 1959;Rand, The gekkonid subfamily Dplodactylinae is distributed in Australia, New 8). Reproductionis Caledonia and the Loyalty Islands, and New Zealand (Kuge, 1967). rut detaileddata are Viviparity has evolved at least twice in this subfamily; once in the ancestors lus, but the eastern of the three endemic and viviparous New Zealand genera Heteropholis, Ho- tn C. pumilis grouP plodactylus, and Naultin as (see Fitch, 7970, for references), and once in the two groups are onlY large endemic geckoes of New Caledonia. The climate of New Zealand is much colder than the climates inhabited by the oviparous diplodactylines. lseemsthat vivipar- "cold-climate" In keeping with the hypothesis, these viviparous geckoes THE EVOLUTIG'N ctF VIVItrAFIITY IN FIEPTILEA (vertical Fig. 2. Geographic distributions of oviparous (horizontal lines) and viviparous (B) the Chamaeleo lin"esyspeclesi *itt i., 1e; t1.eChamaeleo biiaeniatus group, Chamaeleonidae; pumilis group; (C) the lizard subfamily Cordylinae' Whitaket,7978; are unusual in being diurnal and heliothermic (Werner and Thomas,1980). (R. The New Caledonian Rhacodactylusincludes both viviparous (R. trachyrhynchus,Bartmann and Minuth, 7979) and oviparous species Ieachianis,Roux, 1913; H. Cogger, personal communication; Mertens, be 1964, suggestedwithout specifii data that some populations might viviparorii R. auriculatus,personal observation). Distributions and habitats of New CaledonianRhacidactylus are too poorly known to analyzecorre- lates of viviparity. Interestingly, gekkonid viviparity occurs only within EVC'LUTIG'NAFIY ('FIIG'INST clF \/I\/IFAFIITY IN SIOUAMATE FIEFTILES 6e.| FAFIITY IN FIEPTILEE| the diplodactylines, rather than in the more widespread gekkonines or eublepharines that produce calcareous-shelledrather than parchment- shelled eggs. 7. IGUANIDAE The helmetediguanids of centralAmerica consist of two oviparousspecies (Corytophanescristatus, C. hernandesii)and one viviparous one (C. per- carinatus;McCoy, 1968).The oviparousC. cristatusis restrictedto humid lowlands (Stuart, 1948),whereas C. percarinatusoccupies cold high- elevationregions (McCoy, 1968; Fig. 3). The medium-sized, heavy-bodiediguanids of the genus Ctenoble1tltoris are fossorialand live in sandy soilson the westernslopes of the Andes (Cei,7974).The speciesC. adspersus,C. stolzmanni,C. reicheiand C. nigriceyts form a closelyrelated group (Cei, 7979).The oviparous C. reicheiinhabits the Tarapacaand Antofagostodeserts of northern Chile (Donoso-Barros, 1966;Peters and Donoso-Barros,7970; Ce| 1979),whereas the viviparous C. nigricepsoccurs in the Atacamadesert of Chile, reachinglatitude 33"S (Cei, 7974).The modes of reproduction of the other speciesare unknown (|. Cei, personalcommunication). The smooth-throatedlizards (genusLiolaemus) of South Americaresem- ble the Central American Sceloporus,both in morphology and in the great diversity of species.The Chilean speciesinclude two speciesgroups with oviparous and viviparous forms (Donoso-Barros,7966). In Group A, viviparity is known in the speciesL. darwinii,L. cyanogaster,L. schroederi,L. graaenhorstii,L. alticolor,and L. bibronii,whereas L. lemniscatusand L. t'uscus are egg-layers.In Group B, viviparous forms include L. dorbignyi,L. fitzingeri,L. kingii, L. nigroairidis,L. ornatus,and L. pictus,whereas L. mon- ticola,L. platei,and L. tenuisare oviparous.In both groups, the rangeof the viviparous forms is more extensivethan, and encompasses,the range of the eggJayingspecies. However, the speciesin the coldestclimates tend to be viviparous (Fig. 3); this applies to all of the far southern and high montane Liolaemus(Fitch, 1970;Donoso-Barros and Cei, 1971). and viviparous (vertical "horned 'onidae; (B) the Chamaeleo Most speciesof toads" (genusPhrynosoma) of North and Cen- tral America are oviparous(P. asio,P. cornutum,P. modestum,P. platyrhinos, P. solare,P. mcalli), but those from the Meican highlands are viviparous and Whitaker,7978; (P. braconnieri,P. ditmnrsi,P. douglassi,P. orbiculare,Fitch,7970; Mon- tanucci, 1979; Pianka and Parker, 1975). The early phylogeny of Reeve (7969), oth viviparous (R. (1952)has been replaced by that of Presch which is based primarily 'iparous species(R. on osteology. The oviparots coronatumand the viviparous orbiculareare unication; Mertens, very similar to each other and apparently to the ancestralPhrynosomastock. rpulations might be Hence, one may infer that viviparity evolved in an animal similar to these (P. butions and habitats two recent forms. The live-bearer orbiculare)inhabits elevatedcountry in vn to analyzecorre- Mexico (the SierraMadre Occidental),whereas the oviparousP. coronatum occurs only within is found in warmer lowland regions of California (Fig.3). THE EVOLUTIG,N ctF VI\/IPAFIITY IN FIEPTILESI Fig. 3. Geographic distributions of oviparous (horizontal lines) and viviparous (vertical lines) iguanid lizard species of (A) the gents Corytophanes; (B) the species Phrynosoma coronatum and P. orbiculare; (c, D) the genus Liolaemus, Group A (C), the genus Liolaemus, Group B (D). The viviparous iguanids of the Senus Phymaturuslive in the Andes of Chile and Argentina at elevationsof 2800to 4200m (Cei, 1980).Osteolog- ical data suggest that Phymaturusis most closelyrelated to the oviparous Leiocephalusof South America and the West Indies (Paull et al., 1976; Etheridge, 1966;Fitch, 1970).Hence, Phymaturusoccupies colder regions than do its oviparous relatives. IAFIITY IN FIEPTILEA EVOLUTIGINAFIY G'FIIGINS| OF VIVIPAFIITY IN si(lrUAMATE FIEPTILESi 633 Sceloporusis a large Central and North American genus of active diurnal Iizards in which the modes of reproduction, and their correlates have been extensively reviewed (Guillette et al., 1980).The live-bearing specresoccur at higher elevations than the egg-laying ones (Smith, 7939). Phylogenetic reconstructions for Sceloporusgenerally agree with each other. One of the species groups defined by Smith (7939), the scalaris group/ contains both egg-laying and live-bearing species. There are trvo other casesof closely related oviparous and viviparous species.These are S. grammicus (viviparous) with S. megalepidrrus(oviparous), and the large S. spinosusgroup (oviparous) with the viviparous S. formosus and S. poitt- seffi groups. The viviparous S. acanthinus,which was initially placed with the oviparous sTtinosusgroup (Smith, 1939), was later transferred to the viviparous formosus group (Smith and Taylor, 1950). The S. torquatus as- semblage probably has evolved viviparity independently of the above, but it is not closely related to any present-day oviparous species. Subsequent taxonomic work has supported the analysis of Smith; thus, viviparity arose at least four and probably six times within Sceloporus. The evolution of viviparity within the Sceloporusscalaris group is the most clear-cut case of independent origins. The close relatedness of the three species within this group (5. aeneus,S. goldmani, S. scalaris)is indi- cated both by conventional morphological measurements (Smith, 1939) and by karyotypes (Cole, 1978). S. scalarisis oviparous and S, goldmani is viviparous (Smith and Hall, 1974). However, the mode of reproduction of S. aeneushas been the subject of debate. There is no doubt that the high- elevation (3000-4500 m) subspecies S. a. bicanthalisis viviparous (Smith, 1939; Guillette, 1981). The subspecies S. a. aeneusoccurs at lower (2500- 3500 m) elevations, and it has been reported to be both oviparous (Davis and Smith, 1953; Thomas and Dixon, 7976) and viviparous (Smith and Hall, 1974).Reproductive bimodality in this species has been confirmed by Guillette (1987, 7982a);two separate origins of viviparity in different popu- lations may be involved (L. f. Guillette, lr., personal communication). In addition to the evolution of viviparity in S. aeneus,the closely related S. goldmani may have evolved viviparity separately (see phylogeny of Larsen and Tanner,7975). Hence, viviparity arose two or three times within the scalarisgroup. nd viviparous (vertical Viviparity also arose within the Sceloporusgrammicus and 5. me- :he species Phrynosoma '), the genus Liolaemus, galepidurus groups. The viviparous S. grammicus is thought to be closest phylogenetically to the oviparous S. megalepidurus(Smith,1939; Larsen and Tanner, 1975). Both species inhabit high elevations in Mexico or Central ve in the Andes of America, with the viviparous form occurring at higher elevations but lower :i, 1980). Osteolog- latitudes (Guillette et al., 1980; Fig. a). The viviparous S. grammicus also d to the oviparous inhabits more mesic areas, and it is more arboreal than its oviparous rela- iPaull et al., 1976; tive (Guillette et al, 1980). pies colder regions A further origin of viviparity has occurred within Group III of Larsen and Tanner (1975, their Fig. 5). The phylogenies of these authors and of i\2 .97e 5 6- Y i-^r ) o.: v Oo o a! 'daS ii boi tr !.! t= x9"E ^:* ;\ 5 rL o{ bO i d^o Y !-q 'aYr o-8 ots ""E3 'trtst:>: 'o< sz ., -.: o '= c ,*< Eo3ae 9 oco ! u:-'s.Y'O - i .! i irl <.^ x-9 .bp€i gouaMATE ci35 EVGILUTTGINAFIY ctFilGTTNSctF VlVrrtaFtlTY lN FtEPTILEei smith (1939;his Fig. 3) suggestthat the sceloporusspinosus and s. formosus is re- speciesgroups are closel!-related.The oviparousspinosus group group siricted 6 elevationsbelow 2000m, whereasthe viviparoLLsfornl7sus inhabits elevations up to 3000m (smith, 7939;Fig. 4). The live-bearers generallyoccupy moie mesicareas (Guillette et al'' 1980)' " in Apari from'ihese five or six independent origins of "'iviparity repro- sceloporus,two further casesare suggestedby reportsof intraspecific "microlepidotus" is ductivebimodality. The evidencetoi S. (: S.'grammicus) The other weak (Smith,L939; L. J. Guillette,Jr', personal communication)' most of its case is that of S. aariabilis,which alihough oviparous over Mexico, Niciragua' C9:lu^ Fltch' 7970) o bO9 range (Guatemala, lto?'. teograpl',i. in Veracruz, fiur"U""t r"pori6d to be viviparousat high elevations(2500 m) 3 G- "q9i dissection Mexico (Werler, 1951).This record of viviparity is based ? i ;.U unreliable (Tinkle and Gib- O*o o of a single female, and has been reiectedas .F€ O specimen' bons,1.iZT1onthe grounds of possiblemisidentification of the 'oI3 Vivipariiy in thJ Patagonian Vilcuniamay represent-an independent u primitive origin of tn" trait especially if this qtglp has arisen fl:T [i bos hneageratherihan fiom so*eZioliemus-typeline (Ceiand Scolaro, f tr@: ilrr"ur,ia t for drawing t"y to this example') -Y! t"lSz;.1I thani D. G. Blackburn 1tt:"1ion ! o-( m),areasof the lizaias of this group inhabit high elevatiron(1000-1400 analysisbecause .IFo Cordilleras.Howevei, the group is omitted from further E{: no related oviparous form can be identified' o-! t0 5 F^ .'" A *o dtu o,* I El, LACEFTIDAE >Eo Most are The genus Eremiascomprises about 34 speciesof desert lizards. bSF E' .: - ovipirous (E. argus,E. irguta, E. grryyii1, E' namaquensis'E' intermedia' 9 F F; E' x s.o; tiniooceltata.,E.Iigubris, El.neumaini, E. nigrocellata,E. nikolskii,E. persica, Terentev :5.Et pteskei,E. regeli,i. scripta,E. strauchi,l"d E uelox'Fitch' 1970; te:> personal ;E:6 and Chernov, 7965;FiizSimons, 1943;Minton, 1966;R. Huey, !< s4 from Mongolia areviviparous (E' ., v.= o communication).However, three species E.E 'ii Sergeev,1940; Terentev and Cher- O.!o/ kessleri,E, multiocellata,E, przewalikii, F^? - o nov, 1965;Klemmer, 1968). ry;aH that viviparity 9dv= of the phylogeny of Eremiassuggests (1 i .. * A reconstruction within tr'" g"..,'' (Shcherbak,1971). one of Shcher- ..1 Ft'r may have arisen twice <,^F.g bai's speciesgroups includesboth the oviparous E'-argus'the viviparous .bp€H relatedto E' argus,Sowerby, lr=(') bo E. mulioceltnti, an| E. brenchleyi(most closely 1930)for which the mode of reproduction is unknown. Another species groupcontainstheviviparousE.przewalskii,plus,E'buechneri'.E.'quadrit'ons' (1977,p' 54) i"d f . uermiculataofunknown reproductivemode. Shcherbak made the following suggestion: ex- The poverty of the herpetofaunaof central Asia can undoubtedlybe plamld by ihe very difficultenvironment' The rapid worseningof the climate fact lvidently'occurrei relativelyrecently, and perhapswas relatedto the TN EIEPTILE9 THE EVGILUTIG'N G'F VIVIPAFIITY 636 there appeared a mountain barrier' that because of powerful orogenic forces moistei neighbouring country. The which isolated this area f.o*"*a.,,1", and a number of Eremiasspecies (E' ecological conditions of Central Asia forced several others) to change ttl multiocellata, E. przeulalskil, and possibly Astan Eremias (Translation of ovoviviparity, otherwise unknownln Central N. APouchting.) ThelargeoldWorldgenusLrtcertacontainsonlyoneviviparousspecies, its populationsseem to be L. aiuipara;earlyreports"of egg-laying in.some of Gibbons' 1977)' All other erroneous (Packard et al.,"iSZi; T]nkle and apparently are oviparous-.(Fitch'1970)' In his ;;.;t wltirln the genus Arnold (1973)concludes that the review of its taxono*ic ,"tutionships, ,,subgroup of his 2,,, affinitiesof L. uiutparalie with the northern members t-t"pr"rrr*ublyoviparousspeciesL'qrmeniaca'L'caucasica'L'chlorogaster' rosto.mltekotti'L' L. dahti,L. deriugini,L' horaithi,L' monticola'L' praticoln'.L' with earlierhypoth- rudis,L. saxicola,unJf . unisexualis;Arnolddisagrees derjugitriot L praticola' esesthat L. aiaiparaisparticularly closeto eitheil. range' extending to se- L. aiuiparais unusual in its enormous geographic verelycold areas(Fig. 5)' 9. SCINCIDAE Thescincidlizardsmaybedividedintofoursubfamilies(Greer,1970)of contain both oviparous which two, the Scincinae and the Lygosominae' and Acontinae are small and viviparo.,, ,p".i"r, whereas the Feyhninae are.specialized burrowers and entirely viviparous' The latter two groups the scincines of sub-Saharan thought to have ",roirr"j independently i'rom have. evolved jn both these Africa (Greer, 1970)- Althougir viviparity may that they may be descended subfamilies, the conservative view is adopted from viviParous scincines' TheScincinaecontainaboutequalnumbersofoviparousandviviparous are.insufficient to species (Greer, 1970), but available phylogenetit-d1t1 of viviparity' Apart from determine the number of evolutionary origins genera are exclusively Eumecesandperhaps Scincus(seebelow)' allicincine is k"9ll. in Brachymeles oviparous o, ,rirriparo.., 1Ct""t, 1970f Viviparity Minton' (Philippines; Brown and Alcala, 1980)' Ophiomorus.^(Pakistan; iluhn' 1968; Fitch' 1970)' and 1966), Chatcides (Airica,, Europe and Asia; (Africa; Fitch, proscelotes, Scelotes,iipslna, iyphtacontias and Melqnoseps closely to each other 7970) Greer,1970). The five latler genera are Tl3t"-d "Scelotes"of Madagascar (Greer, (FitzSimons,7943;C*"t, 1970) anito the 1970).AtleastoneMalagassyspecies'scelotesigneocaudatas'isoviparous has.evolved in this (Blanc and Blanc, 1967).itr""oi clearthat viviparity and.viviparous forms lineage. Thus, the mort lto'"ty related oviparous "Scelotis" Proscelotes(Greer' 1970)' -uy t" the Madaga srcan andthe Airican analysis' However, data are too few for a quantitative EVGILUTIONAFIY ctFTIGINst c,F VIVIPAFIITY IN gGIUAMATE FIEPTILEA mountain barrier, rring country. The \-) lremiasspecies (E. :s) to change to os. (Translation of vlvlparousspecles/ ulationsseem to be ;, 7977).All other ritch, 1970).In his :oncludes that the his "subgrovp2," sica,L. chlorogaster, , L. rostombekoui,L. 'ith earlier hypoth- ryini or L. praticola. e, extendingto se- es (Greer, L970) of rin both oviparous .continae are small cialized burrowers nes of sub-Saharan rlved in both these may be descended Fig. 5. Geographic distributions of oviparous (horizontal lines) and viviparous (vertical "Subgroup f,us and viviparous lines) lizards within (A) 2" of the genus Lacerta, Lacertidae; (B) the scincid species are insufficient to Ablephnrus biait tatus. ?arity. Apart from lera are exclusively genera may well have evolved viviparity rwn in Brachymeles The other viviparous scincine 'Pakistan; Eumeces,Brachymeles Minton, independently. For example, except for the oviparous is the only scincine in eastern Asia. However, the present distribution of ; Fitch, 1970), and (Greer, and hence Brachymelesmay have ;eps (Africa; Fitch, scincines may be a relict one 1970) distant ancestor. The lated to each other inherited viviparity from an extinct or geographically is likely to represent one ladagascar (Greer, viviparity of Brachymeles,Ophiomorus and Chalcides datas, is oviparous or more origins of viviparity. the large scincine Eumeces, ras evolved in this Viviparity also has evolved within Senus the North Temperate Zone except Europe. I viviparous forms which occurs throughout all of Many show maternal brooding behavior; pro- zlotes(Greer,7970). Most Eumecesare oviparous. longed retention of eggs in uterois known in E. brnilineatus, E. callicephalus, Cl3A THE EVGILUTION C,F VIVIT'AFIITY IN FIEPTILEEI and E. fasciatus(Campbell and Simmons,L961; Werler, 1951; Fitch, 1954). Viviparity has been recordedin the montane Mexican speciesE. breairo- stris,E. colimensis,E. copei,E. dugesi,E.lynxe and E. ochoterenai(Axtell, 1960; Taylor, 1936;Webb, 7968;Van Devenderand Van Devender,1975; Guil- lette, 1983).A phylogeneticreconstruction for Eumeces(Taylor, 1936)sug- gests that viviparity evolved three separatetimes: once in ochoterertai- "t'ttrcirostris") breairostris-colimensis-dugesl,once in lynxe (including and once in copei.However, E. copeimay belong with the breuirostris€,roup (Dixon, 1969)leaving two origins of viviparity within Eumeces,both occur- ring in the high Mexicanplateau (Taylor, 1936). Most oviparousEunrcces live at lower elevations,and hencein warmerclimates, than do the vivipa- rous ones (Fig.a). The sand-swimmingskinks of northern Africa and Arabia (Scincus)ap- parently include both oviparous and viviparous forms (Fuhn, 1968;Mer- tens, 1972;Arnold and Leviton, 7977).However, reproductive data are lacking for most species,and this caseis included only tentatively. The Lygosominaeincludes more than 40 genera and over 600 species and is the most numerousand diversescincid lineage (Greer, 1970). Ap- proximately one-third of speciesare viviparous (Greer, 1970),and many separateorigins of viviparity are involved. Viviparity and oviparity co- occur in thirteen genera.Phylogenetic reconstructions also point to inde- pendent origins of viviparity in the ancestryof severalwholly viviparous genera.For example,the large viviparous lizards of the Australasiangen- eraCorucia, Egernia, and Tiliquaare thought to have evolvedfrom an ovipa- rous Mabuyaof southeastAsia (Greer, 1970).Similarly, Isopachysof Thai- land (fossorialand viviparous, Taylor, 1963)probably is derived from the Sphenomorphus"laturense group," at least one member of which is ovipa- rous (Greer, 1977,and personalcommunication). The fossorialviviparous Hemiergisof Australia may share a common ancestry with Lerista and Sphenomorphus"solomonis group"; the taxa of both are primarily oviparous (Greer, 7967c;Greer and Parker,7974). Although all of thesecases indicate independentorigins of viviparity, the transitionof reproductivemodes has occurredtoo early in phylogeny to permit useful ecologicalcomparison of present-dayoviparous and viviparous forms. Not all of the wholly viviparous lygosominegenera are likely to repre- sent independent origins of viviparity; some may have inherited viviparity from a live-bearing ancestor. An example is the African Eumecia,which probably has evolved from Mabuya(Greer, 7967b),most speciesof which are viviparous. Hence,there is no needto postulatean independentevolu- tion of viviparity in Eumecia.A lack of phylogenetic information makes it difficult to determine whether viviparity in the semiaquatic southeast Asian Tropidophorus(Brown and Alcala, 1980)and Ophioscincusof Thailand (Taylor, L963) should be regarded also as independent origins of live- bearing. Some subgroups of the Lygosominae(Greer, 7974,7979b) consist en- IN ETOUAMATE FIEPTILES EISE IPAFTITY IN FIEPTILE€I EVoLUTIoNAFIY CIFIIG;INS OF VIVI]'AFIITY ',1951;Fitch, 1954). tirely of oviparousspecies (e.g., Dasia-Lamprolepis; Lampropholis subgroup n speciesE. breuiro- of the Eugongylusgroup; Greer, 7974),whereas others include viviparous tterenai (Axtell, 1960; ones (especiallythe sphenomorpfutsgroup, n'hich containstaxa showing at rvender, 7975;Cull least17 independent origins of viviparity).The l3lygosominegenera con- ; (Taylor,1936) sug- taining both oviparous and viviparous species(Ableplurus, Anomalopus, rnce in ochotererni- Leiolofisma,Lerista, Lipinia, Lobulia, Lygosoma, Mobuya, Prasinohaema, Scin- "furcirostris") and Tribolonotus)are now consideredin more 5 and cella,saiphos, sphenomorphus, rc breairostrlsgroup detail. )wneces,both occur- Ablepharuscontains five speciesof small skinks from centralAsia (Fuhn, t oviparousEumeces 1969).Oviparity has been reportedin A. deserti(Terentev and Chernov, . than do the vivipa- 1.965),A. kitaibeli(Arnold and Burton, 1978),and tentatively in A, pan- notticus(smith, 1935).The mode of reproductionof A. Srayanusappears Arabia (Scincus)ap- unknown. Most interestcenters on the remaining form, A, biaittatus.The s (Fuhn, 1968;Mer- subspeciesA. b. biaittatusis oviparous,whereas the subspeciesA. b. alqicus oroductive data are (given full specificstatus by some authors) is viviparous (Terentevand y tentatively. Cherr,on, 1965).The oviparous subspeciesis found below 2000m in the rd over 600 species ussR, whereasthe live-bearingsubspecies occurs up to 3800m (Terentev : (Creer, 7970).Ap' and Chernov,1965;Fig. 5). ,r, 1970),and many Anomalopuscontains burrowers with elongate bodies and greatly re- (4. y and oviparity co- duced or no limbs. Two speciesof Australia's central eastern coast ; also point to inde- lentiginosus,A. oerreauxil)are oviparous, whereas the more southern A. I wholly viviparous swansoni("species 3": Cogger, 7975)is viviparous' The live-bearer oc- ie Australasiangen- cupies coolerclimates than its egg-layingrelatives (Fig. 6). rlvedfrom an ovipa- Both oviparity and viviparity have been reported in the Australian "sand y, Isopachysof Thai- swimming" skinksof the genus Eremiascincus,but the data are unre- is derived from the liable(Greer,7979a), r of which is ovipa- Leiolopismacontains small, active,diurnal Australasianskinks with well- fossorialviviparous developed limbs. Live-bearinghas evolved at least three times. The first containingthe viviparous L. coaentrylof the Australian ry with Lerista and caseis in the group primarily oviparous southernhighlands and the oviparousL. ziaof mideasterncoastal Australia casesindicate (Greer, 1982and personal communication; Ingram and Ehmann, 1981)' these "baudini" oductivemodes has The secondcase is in the speciesgroup (Greer, 1982);thus, L. ,gicalcomparison of duperreyi,L. platynotumand L. trilineatumare oviparous, but L. entrecas- teauxi and L. metallicumare live-bearers.In both of these groups, the r are likely to repre- viviparous forms occupy cooler climates than the egg layers (Fig' 6). A inherited viviparity third origin of viviparity is probably represented by four other Australian can Eumecia,which species of Leiolopisma,belonging to the spencerispecies-group; these are rst speciesof which viviparous also (Greer, 1982). independentevolu- Viviparity may have evolved also in the speciesof Leiolopismainhabiting formation makes it two island groups: New Zealand and New Caledonia' However, both niaquatic southeast casesare doubtful becauseof insufficientphylogenetic information. Within 'oscincus of Thailand New Zealand, the northern (: warm climate) L. suteri is oviparous ent origins of live- (Towns, 1975),whereas all other speciesare live-bearers.It seemslikely that L. suteri representsan independent invasion of New zealand, and is , 1979b)consist en- not closely related to the other endemic scincid lizards (Hardy, 1977).How- THE EVG'LUTIGIN OF VIVIPAFIITY IN FIEPTILES Fig. 6. Geographic distributions of oviparous (horizontal lines) and viviparous (verticai lines) scincid lizards within (A) the gentts Anomalopus; (B) Saiphosand related oviparous forms; (C) the coaentryi species-group within kiolopisrna; (D) the baudini species-group n'ithin Leiolopisma. ever, the viviparous New Zealand skinks, together with some viviparous Australian forms (L. spencerigoup), may be derived from an oviparous lineage (Hardy, 1977,7979; Greer, 1982). The other leiolopismid assemblage containing both eggJayers and live- bearers occupies New Caledoria. kiolopisma tricolor is viviparous (Roux, 1913), and several other species are oviparous (R. Sadlier, personal com- munication). This suggests that live-bearing has evolved within the group' but this condusion may be in error if, as in the New Zealand example, the group is polyphyletic. The taxonomy of these lizards is too poorly known to permit any definite condusions. Lerista includes many burrowing skinks, some with reduced limbs, of the arid regions of Australia. The southerly L. bougainailleicontains both oviparous and viviparous forms (A. Greer and P. Robertson, personal com- munication; Shine, personal observation); as viviparity occurs in two al- lopatric populations it may have evolved independently in each case. 641 EVoLUTIGINAF|Y ctFIIGINst OF VIVIPAFTITY IN SGrUAMATE FIEPTILES PAFIITY IN FIEPTILEET However, this is listed here as a singleorigin. Eight other speciesof Lerista are known to be oviparous,but the southwesternL. microtisis live-bearing (Greer, 7967cand personalcommunication). Morphological characters ren- der it unlikely thai L. microtisis descendedfrom L, bougainuillei(A. Greer, personalcommunication), so it seemsprobable that viviparity has evolved of viviparous independently in this form. Both in the comparison lhe L. microtisto oviparousspecies and of the viviparousL. bougainailleito its oviparous conspecifics,the viviparous forms occur in cooler regions (Fig.7). Lioinioincludes small slenderarboreal skinks that are distributedwidelv through the island archipelagosof the southwest Pacific (Greer, 1,974). Modei of reproduction are known for ten of its 20 species(A. Greer, per- sonal communication).The speciesL. int'ralineata,L. pulchella,L. quadriuit- tata, and L. uittigeraare oviparous. The speciesL. auriculatn,L. noctua,L. rabori,L. relicta,L. semperi,and L. uenemaiare viviparous. In the Philip- pines, the viviparous L. auriculatqocatrs at higher elevationsthan the bviparousf . pilchellaor L. quadriaittata(Brown and Alcala, 1980),but data are insufficient for a thorough comparison. Lobulittcontains five small robust-bodiedskinks from New Guinea. Re- productive modes are known for two of them. The oviparousL. stanleyana is terrestrialand lives between 1200and 2000m, rarely to 2500m, whereas the viviparous L. elegantoidesis arborealand lives at 1500to 3500m (F. Parker, personal communication).Both speciesare diurnal heliotherms, and L. sianleyananests communally (F. Parker,personal communication). A review of reproductive modes and probable phylogenetic relation- ships within the scincid genus Lygosomasuggests two independent origins nd viviparous (vertical of viviparity (Greer, 7977).T\e oviparous afer group (L. afer, L' fernandi,L' and related oviparous guineeise,L. sundeaalti,and L. pembanum)are closelyrelated to the vivipa- within ni species-group ious L. laeaicepsand L. ainciguerra.The oviparous L. bowringiandL. punc- tatum form a group with the viviparous L. tanae(Greer, 7977).The vivipa- All three live-bearers are h some viviparous rous Lygosomado not inhabit cool clirnates: Kenya (Greer, 1977). From an oviparous endemiC to arid regions of Somalia and eastern within the afer group, the egg-layersare widely distributed over the Afri- extend into West :gg-layersand live- can continent. For example,L. fernandiand L. guineense and easternparts viviparous (Roux, Africa, and L. sundeaalliis distributed through southern the two lier, personal com- of the continent. Within the bowringi-punctatum-tanae8roup, t within the grouP, oviparous speciesoccur in Asia. 'ih" great variety of ground- aland example, the "o"*opolitan tropical Mabuyacontains a speciesoccur both in Asia too poorly known dwelling skinis. Oviparous as well as viviparous and Africa; however, all South American forms are viviparous (at least 1"983).one African reduced limbs, of some with elaborate placentation, vitt and Blackburn, different modesof repro- oillei contains both species(M. quinquetaeniota)was reportedto show (Fitch, 1970);subsequent investiga- son, com- duction in different parts of its range Personal 1976).Authorities disagree / occursin two al- tion disproved this (Visser, L975;Spellerberg, (Smith, 1935;Badhuri, 1943).Both ntly in each case. on the rbproductivemode of M. carinafa THE EVctLUTIctN OF \/IVIPAFIITY IN FIEPTILE€I and viviparous (vertical lines) Fig. 7. Geographicdistributions of oviparous'krista, (horizonal lines) (B\ the species L' scincid lizards within (A) the genus excludtng L. bougainaillei, "fasciatus bougainaillei;(C) the Sphenomorphus grouP"' com- oviparity and viviparity occur in M' capensls(W' Haacke' Personal -,rr,icuiior,; and M. suicata(D. Horton, personalcommunication)' (Horton, ir," genus Mabuyamay be divided into three speciesgroup: American forms 1973),o:ne of which "o.,rirt, entirely of viviparous.south M'.perroteti and "pp"L"afy derived f-* u lineage including-the oviparous at least one tie viviparo us M' brnicollls (Hoiton , t973);hence' it represents (one.African' one oJgrt, oi viviparity. Each of ihe other two speciesgrouPs forms' has eriu"l contain both oviparous and viviparous -Viviparity M' bensoni' evolved in a group containingthe oviparois M' aureopunctata' E\/CILUTIONAFIY GIFIIG'INSI OF VIVIPAFIITY IN SGIUAMATE trEFTILEB M. Iacertiformis,M. maculilabris,and M. quinquetaeninta,and the viviparous M. bayoni,M. irregularis,M. striata,M. sulcata,M. uaria,and M. aittata.All of these speciesare African exceptthe Iranian M. uittata.The live-bearers occur in more southerly, and hence cooler, climates in both of these viviparous origins. A phylogeny for Asian Mabuya(Horton, 1973)derives the viviparousM. auratafrom a lineageincluding the oviparousM, macularia;the viviparous M. multifasciatars tentatively thought to be related to the oviparous M. Iongicaudata,Overall, a total of six evolutionary origins of viviparity may have occurredwith Mabuya. Prasinohaemaincludes arboreal skinks of small to medium body size, with prehensiletails. Its speciesare remarkablein having the blood plasma and other tissuescolored green (Greer and Raizes, 1969).All five speciesin the genusoccur on New Guinea,but one speciesextends its rangeinto the Solomons(Greer, 7974).The small P. airensis oviparous(Greer, 7974)and occursfrom sealevel up to 820m elevation.Three other speciesare known to be viviparous and are found at higher elevations(flauipes, 1070 to 2500 m; prehensicauda,7200 to 2300m; semoni,sea level to 1800m). Scincellais distributed widely in the Old World and to a lesserextent in the New World (Greer, 1974).Modes of reproduction are known only for sevenof 32 recognizedspecies. S. barbouri,S. bilineata,S. formosa, S. Iadacen- sis, S. modesta,and S. sikkimensisare oviparous (Smith, 7935;A. Greer, personal communication), whereas S. himalayanais viviparous (Smith, 1935).Analysis is precludedby the lack of reproductivedata. However, S. himalayanaoccurs between 1300 and 4000m (Smith, 1935),suggesting that cold climatesmay have played a role in the evolution of viviparity. Most other membersof the genus live at lower elevationsthan does S. himalay- ana,but at least one oviparous species(5. ladacensis)ranges even higher (Smith, 7935;Greer, 1974). The monotypic fossorialSaiphos equalis is unusual in displayingoviparity and viviparity in different geographic areas. In the relatively mild climate of the Australian east coast,it is oviparous and its poorly calcifiedeggs "iviparous (vertical lines) hatch within a week or two of being laid (Bustard, 1964;Cogger, 7975).In nllei, (B) the speciesL. the cold highlands of the New England Plateau, young are born alive (Shine and Thompson, unpublished; Greer, 1983).It is of interest that those Sphenomorphusthat are the closest relatives of Saiphos(Greer, 1983) cke, personal com- and inhabit warmer climates (Fig. 6) are oviparous and lay normally nunication). calcified eggs with a much longer incubation period than the oviparous es groups (Horton, populations of Saiphos(Shine, personal observation). LthAmerican forms Sphenomorphusincludes a large but probably monophyletic group of cus M. perrotetiand Australasianskinks. Analysisreveals that viviparity originatedat leastfive "crassicaudus rresentsat leastone times within the genus. The Sphenomorphus group" contains rs (one African, one several elongate fossorial species including the oviparous S. crassicaudus ns. Viviparity has from northern Australia and PapuaNew Guinea. This speciesis morpho- unctata;M. bensoni, logicallysimilar to S.fragilis of southeasternNew Guinea,which produces IN FIEPTILES! g.44 THE EVG'LUTIGIN ctF VIVIPAFIITY "shell" fully developedyoung, although with a thicker and more opaque than the thin tianspirent membranesseen in most viviparous species (Greerand Parker, iOzol' Spttrrornorphusfragilis is found f19m s3alevel to ozo .r-,,probably at a climate similar to that experiencedby the related S. crassicaudus' oviparous "t'asciatus in" +Ospecies of the fossorialSphenomorphus group" are dis- tributed in New Guinea, the solomon Islands, the Philippines,northern Australia,and the Lessersunda Islands(Greer and Parker,1967). In many casesmodes of reproduction are unknown, but both oviparous and vivioarous forms occur within Australia (oviparous:s. brongersmai,s. douglasi,S. isolepis,S. pardalis,S, punctulatus,S' fuscicaudus;viviparous: S' within New Guinea(oviparous: S' ni- 'grilineatus,grriiliprr, S. mu;rrayi, S. tenuis)and derooyae,s. nigriaentris,s. oligolepis,s' "undulatus,s. brunneus,s. cranei,s. s. schultzei;viviparous: s. longicaudatus,s. cinereus,s. Ieptofas- ciatus).Analysis of the Australian speciessuggests that the viviparous specieso..,ri ir-rslightly cooler climatesthan the oviparousones (Fig. 7). Viviparity also originated in the elongate fossorial Sphenomorphusni- gricaudus,which is ciosely related to both of the Sphenomorphusgroups iiscussedabove. New Guineaspecimens of S. nigricaudusare oviparous, (A. but Australian specimensfrom CapeYork, Queensland,are viviparous Greer, personal communication).Within New Guinea, this speciesis re- strictedto the south coastsavannah belt, reaching600 m elevationon the sogeri plateau near Port Moresby (F- Parker, personal communication). specimensoccur from sealevel to over 1000m (Cogger,1975). Au"stralian "uariegatus The bodlform of the sphenomorphus grouP' is more robust " of reproduction than that oi the t'asciatusgroup," discussedabove. Modes are known for 29 species,ind it leasttwo separateorigins of viviparity are indicated (A. Greer, personal communication).one origin lies within a (from China and Taiwan)and $oup coniaining the SviparousS, boulengeri eastern Asia)' ihe ,rirripa.ou" {. lor*otensis and S. indicus(China, Taiwan, Anothei group containingboth oviparousand viviparous speciesoccurs in New Gui"neuis. tti"trti, &iparous), Borneo (5. multisquamatus,oviparous) and the solomon Islands (s. concinnatus,viviparous). There are no data on ecologyor elevationaldistribution of S. multisquamafas.However, behavior ana iiUitut of the other two speciesare similar to each other (F. Parker, personalcommunication). The bviparousS. stickelioccurs from sealevel up io 700 m (F. parker, personal communication), whereas the viviparous S' concinnatu.soccursfrom sealevel up to 1500m (on Guadalcanal'M' McCoy' communication). personal "undoubtedly Tribolonotashas been describedas one of the most bizarte taxa of lizards" becauseof its abdominal glands, volar pores, and aberrant squamation (Greer and Parker, 1968)' It occurs through New Guinea and the Solomon Islands. T. qnnectens,T. blanchardi,T. gracilis,T. noaaguineae, and T. pseudoponceletiare oviparous, butTribolonotusschmidti from the sol- omon Islands is viviparous (Greerand Parker, 1968).Recent studies pro- FIEPTTLES B'I5 /IPAFIITY IN FIEPTILEg EVC'LUTIC'NAFIY crFIIGld$g GIF VIVIPAFTITY IN SI(IUAMATE "shell" (McCoy' moreopaque vide valuable information on the Solomon Islands Tribolottotus occurs t viviparous species 1980,and personal comrnunication). Only the viviparousT'1chmidfi timber in rnd from sealevel to on Guadaicanal, where it is usually found under rotting fallen ovipa- :nced by the related forests and from sea level to over 1000 m. The superficially similar group. on rous T. blanchardireplaces it on other islands of the solomon moist creek iatus group" are dis- Nggela (: Florida liland.), T. blanclurdi is usually four-rd in lt has been 'hilippines,northern be"d"s,ofien in drifts of dead leaves, and in other forest debris' hills on Nggela), rrker,1967). In manY collected from sea level to 500 m (the tops of the highest islands' roth oviparous and but may occur at higher elevations on more mountainous ;'. S. brongersmai,S. ,1 audus;viviparous: S. O. XANTUSIIDAE rea(oviparous: S. nr discontinu- The small viviparous lizards of the Xantusiidae are distributed zntris,S. oligolePis,S. united ously in the west Indies, Central America, and the southwestern . cinereus,S. IePtofas- Statesandapparentlyarederivedfromgeckoes(McDowellandBogert' that the viviParous geckoes are 1954;Underwood, 1971;Northcutt, 1978)' The only viviparous )arousones (Fig. 7). Zealand) and' in the southern hemisphere (New Caledonia and New al Sphenomorplusni- inferred that hence, are unlikely ancestors for the xantusiids. It can be thenomorphusgrouPs viviparity evolved in the ancestors of the xantusiids' audusare oviParous, (A. d, are viviparous D. Si€FP€ntes ra, this speciesis re- the 0 m elevationon 1. ANILIIDAE AND UtrOPELTIDAE nal communication). "primitive are vivipa- )0 m (Cogger,1975)' These so-called snakes" of south America and Asia (Fitch 7970; 'oup" is more robust rous; they *uy L" relatively dosely related to each other ' Anilioidea con- odesof reProduction McDowell, 1975; Rieppel,'Xenopeltidae t 4. BOIDAE The large Boidae are widely distributed throughout the tropics and sub- tropics; a few speciesinvade temperate areas. Subfarnilialciassifications ars a source of disagreement(e.g., McDowell, 7975),but oniy the two largest subfamiliesare relevant for this discussion.One subfamily, the Boinaeor "true boas," areviviparous. They are confinedto the New World tropics with the exceptionof the MalagasianAcrantophis and Sanziniaand the South PacificCandoia. The boinesare similar in body form and habits to the Old World pythons (subfamily Pythoninae),except that the latter are oviparous. Maternal egg-guardingbehavior is widespread, possibly uni- verial, in the pythons (Shine and Bull, 1979).Morphological characters suggestthat the pythonines gave rise to the boinesrather than vice versa (underwood,7967). This is consistentwith the assumptionthat viviparity evolved from oviparity. The origin of boid viviparity is so far back in evolutionary history that analysisof ecologicalcorrelates is impossible.A recordof oviparity in the normally viviparous Boaconstrictor is probably an error (Tinkle and Gibbons,1977). 5. TFIOPIDOPHIIDAE Severalunique morphologicalfeatures suggest that the tropidophiids are only distantly related to the boids with which they usually have been combined (McDowell, 1975).Apparently, all tropidophiids are viviparous (e.g., Dowling,1974). The structure of the skull is similar to that in boly- erids (McDowell, 1975), one of which recently has been shown to be oviparous (Anonymous, 1982).Hence, viviparity has evolved within the hopidophiid lineage 6. COLUE}FIIDAE Viviparity has evolved several times within this large and-diverse lineage, but interpretation is confused by differing phylogenetic schemes.In many cases,relationships are obscure.For example,the viviparous Asiatic tree- snakeAhaetutla (Fitch,1970) may be closely related to the oviparous African Thelotornis(Underwood , 1967, 1979). If the presumed relationship is valid, viviparity originated in the group. A similar problem occurs with xenodon- tine-colutridJof the New World. Viviparity hasbeen described inTomodon, OFIIGIINEi ctF VIVIPAFTITY IN SICIUAMATE FIEPTILEA IPAFIITY IN FIEPTILES EVOLUTIG'NAFIY (Balley, 7966, e Opheodrysaernalis, Tachymenis,Ptychophis, Pseudotomodon, and Thamnodynastes On the basisof ;uch prolonged egg 1981;Gudynas, 1981; da Cunhaand do Nascimento,1981). (1981)suggests that lrlopsinclude 30-42 their phylogeneticrelationship to thesegenera, Bailey small-to-medium- for RamphotyPhloPs Calamodontophisprobably are viviparous also. All are Tachymenismay be rginsof viviparitY in sized snakes wiih a generally southern distribution. Leptodeira,whereas Tomo- f phylogeneticinfor- relatedto the oviparousImantodes, and perhaps don and Thamnodqnasfesmay be related to the oviparous Conophis(under- wood, 7967).Yet another interpretation links Tachrlmeniswith the ovipa- rous Philoclryasof subtropicalsouth America (Duellman, 1979).In each case, the egg-layersoccupy much warmer climates than do the live- he tropicsand sub- bearers.Underwood's taxonomy suggeststwo independentorigins of milial classifications viviparity, whereas Bailey's requires only one; parsimony dictates the ), but only the two adoption of the latterview. Recentstudies using microcomplementfixation One subfamily, the are consistent with Bailey's conclusions, but do not reveal any close :d to the New World xenodontinerelatives of the Tomodon-ThamnodynastesgrouP (cadle, 1984). his and Sanziniaand A problem arises with the viviparorts Psammodynastespuluerulentus, y form and habits to which is widely distributed in southeasternAsia (Fitch, 1970).lt has no pt that the latter are clear taxonomic affinities with any other colubrids; it is placed in the rread, possibly uni- boigine subfamily only provisionally (Underwood, 7967).Viviparity may hological characters have evolved independently in this case,as it seemsunlikely that Psam- ther than vice versa modynastesis closelyrelated to Ahaetulla,the only other viviparous Asiatic ption that viviParitY boigine. The situation is less clearin the caseof the subfamily Homalop- ty is so far back in sinae,all membersof which areviviparous (Gyi, 1970).Morphological data "boigines" rtesis impossible.A suggest that the homalopsines may be derived from the trictoris probably an (Underwood, 7979), so that it is conceivable,although unlikely, that homalopsinesevolved from a viviparous ancestor (Ahaetulla,Psammody- nastes,Ptychophis, Tachymenis, Thamnodynastes, Tomodon, Pseudotomodon, or their ancestors).However, biochemicaldata indicate that the homalop- "boigines" sines are related only distantly to other colubrids, including re tropidophiids are (Schwaner and Dessauer, 7982), suggesting that viviparity may have ' usually have been evolved independently in this subfamily. Nonetheless,this casewas not rhiids are viviparous analyzedfurther. nilar to that in bolY- Viviparity occurs also in the small Mexican Conopsisand Toluca(Greer, been shown to be 1966).bueilman (1961)considered that these live-bearersare closely re- r evolved within the lated to the oviparous genera Ficimia and Gyalopion,but Hardy (1975) at- gued against this view. It is difficult to identify the nearest relatives of Conopsii and Toluca.Underwood (1967)suggested that a grouP of ovipa- rous genera (including sonora,Tantilla, and stenorhina)might be closeto and diverselineage, the live-bearers.Whichever of thesetaxonomies is accepted,the viviparous ic schemes.In manY forms occur in much cooler climates and higher elevations (up to 3000 m) iparous Asiatic tree- than do their oviparous relatives. he oviparous African The small marsh-dwellingAfrican Amplorhinusmultimaculatus is vivipa- relationshipis valid, rous (Broadley and Cock, 1975),but its taxonomic affinities are problemat- ccurs with xenodon- ical (Branch, 1982).It may be included with the boigines (e.g., Broadley tescribedinTomodon, and Cock, 7975)or may be relatedto the natricines(Branch , 1982).Ineither 6UTB THE EVOLUTIGIN C,F VIVIPAFIITY IN FIEPTILE€| case,it is likely to representan independentorigin of viviparity. The only other rear-fangedAfrican snakesknown to be viviparousare Psammophylax a. aariabilisand ATtarallactusjacksonl; neither is likely to be ancestralto Amplorhinus.Another boigine, the oviparousHemirhagerrhis notntsenia, was formerlyincluded in the genusAmplorhinus, but it probablyis not closely related(Bogert, 1940). H. notatneniais a lowlandform, whereasAmplorhinus multimaculatrsis montane(Vessey-Fitzgerald, 1958; Pienaar, 1978; Broad- ley and Cock, 1975).An alternativephylogenetic hypothesis is that An- plorhinusis relatedto the natricinesof centraland westernAfrica (Branch, 1982),which probably are oviparous(Fitch, 1970;Rossman and Eberle, 1977).In each of these putative phylogeniesthe live-beareroccupies a much cooler climatic region than its oviparousrelafives. One cautionary note in acceptingan independent origin of viviparity for Amplorhinrsis its unexpectedanatomical resemblance to the viviparous Neotropical Tham- nodynastesand Tomodon(Underwood, 7967).Hence it is possible,although zoogeographicallyimprobable, that the viviparity of Atnplorhinasis inher- ited from a viviparous Neotropical ancestor. Three speciesof the East African Aparallactus(A. capensis,A. guentheri, A. Iunulqtug)are oviparous, one (A. jacksoni)is viviparous (FitzSimons, 7962;Pitman, 1974;Spawls, 7973)and the mode of reproduction of the others remainsunknown. All are small and fossorial;the range of A. jack- soniis included within those of its oviparouscongeners (Fig. 8). Coronellsis a genus of moderate-sizedcolubrids from Europe and India. Oviparity has been recordedin the Indian C. brachyura(Deoras, 7977) and the European C. girondica(Street, 1973).The other European species,C. austriaca,is viviparous (Smith, 7973);its ecologyis well known (Duguy, 1961;Andren and Nilson, 7976a;Spellerberg and Phelps,7977). It is found at higher latitudes, in cooler climates,and over a wider latitudinal range than its nearestoviparous congener,C. girondica(Fig. 8). The large colubrid genus Elapheof North America, Europe, and Asia is almost certainlypolyphyletic, but no satisfactorysubdivision has been de- vised (seecomments by Pope, 1935).Oviparity hasbeen recorded in many speciesof Elaphe,including E. bimaculata,E. carinata,E. climacophora,E. conspicillata,E. dione,E. flauolineata,E. guttata, E. helenae,E. hohenackeri, E. longissima,E. mandarina,E. moellendorffi,E. obsoleta,E. porphyracea,E. prasina,E. quadriairgats,E. quatuorlineata,E. radiata,E. scalaris,E. schrencki, E. situla, E. subocularis,E. taeniura,and E. aulpina(Chang and Fang, 1931; Deoras,1.971; Fukada,1956; Kopstein, 1938;Pope, 1935; Smith, 1943;Stew- ard, 197'1,;Fitch, 1970).Maternal care has been reported in three species (Fitch, 1970).The Chinese and Korean E. rufodorsatais viviparous (Sow- erby, 1930;Sura, 1981)and also semiaquatic,unlike the other Chinese Elaphe.Biochemical data suggestthat this speciesis only distantly related to any other Elaphespecies (Lawson and Dessauer,1980). It has been sug- "living, gested that viviparity in E. rufodorsataevolved because as these snakesdo, in open swamps and marshes,their eggs would stand a poor /IPAFIITY IN FIEPTILEA EVGILUTIGINAFTY CIFTIGINS crF VIVIPAFIITY IN siGlUAMATE FIEPTILES f viviparity. The only lus are Psammophylax Lyto be ancestralto yrrhis notataenia,was robablyis not closely rvhereasAmplorhinus )ienaar, 1978;Broad- 'pothesisis that Ara- sternAfrica (Branch, Lossmanand Eberle, rre-beareroccupies a ves. One cautionary for Amplorhinasis its stsNeotropical Tham- \ is possible,although Amplorhinusis inher- ---- 1 1- A\\ :apensis,A. guentheri, rparous(FitzSimons, reproduction of the the range of A. jack- ers(Fig. 8). m Europeand India. z (Deoras,1977) and iuropeanspecies, C. vell known (Duguy, ps, 7977).It is found der latitudinal range . 8). Europe, and Asia is livision has been de- in many en recorded Fig. 8. Geographic distributions of oviparous (horizontal lines) and viviparous (r'ertical t, E. climacophora,E. lines) cofubrid snakes within (A) the genus Aparallactus;(B) the genus Coronella. Ienae,E. hohenackeri, :a, E. porphyracea,E. scalaris,E. schrencki, chance of hatching" (Sowerby, 1930). However, many aquatic and semi- ang and Fang,1931; aquatic Asian snakes are oviparous (Pope, L935): nontheless, viviparity 5;Smith, 1943;Stew- arose in the aquatic Sinonatrix (see below). The viviparity of E' rufodorsata led in three species could also be due to its inhabiting a cold climatic region (northeastern is viviparous (Sow- China and Korea). Although a few oviparous Elaphe occur even farther : the other Chinese north (e.g., E. dione, E. schrencki),most European and Asian Elapheoccupy nly distantly related much warmer climates (Fig. 9). 80). It has been sug- The aquatic and semiaquatic colubrids of the genus Grayia inhabit trop- "living, rse as these ical Africi. Four species, cqesar,ornata, smithii, and tholloni, are generally would stand a poor recognized. Grayia smithii is oviparous (Angel et al., 7954; Pitman, 7974). 65cl THE EVG'LUTIGIN CrF VIVIPAFITTY IN FIEPTILEA t.t" 7-" I Fig. 9. Geographic distributions of oviparous (horizontal lines) and viviparous (vertical lin"es)colubriJ snakeswithin (A) Old World speciesof the genusHaphe; (B) the genusSlnona- trix. Developingembryos in uterohave been reported in G. tholloni(Loveridge, 1933),iug[esting-that this speciesmay be viviparous. Both specieshave similar geographic distributions (Pitman, 1974).Modesof reproduction are unknorin for ttre other species. The possibility of viviparity in Grayia ts supportedby the presumedrelationship of this genuswith_the-viviparous afiican colutrids buberria and Pseudaspis(Bogert,1940; Fitch, 1970).I infer that viviparity has evolved in this lineage, either within Grayiaor in the ancestori of buberria and Pseudaspis.Both of the latter two generaextend /IFAFIITY IN FIEPTILES EVOLUTIGINAFIY OFIIGINC| ctF VIVIPAtrITY IN A(IUAMATE FIEPTILESI 451 into cool regions;in Zimbabwe,D.Iutrix is restrictedto elevationshigher than 1400m, and P. cannoccurs up to 1500m (Broadleyand Cock,1975)' There are 15 speciesof South American watersnakesin HelicoTts;al- though Amaral (1.977)suggested that all are viviparous, reproductive modes are known definitely for only a few of them. Helicopscarinicaudus from central Brazil, Uruguay, and Argentina is viviparous (Mole, L924; Shine,personal observation), and so are H. Ttolylepisand H. triaittatus(da Cunha and do Nascimento,1981) and probablyH.leopardlnus (Shine, per- sonal observationfrom dissectionof gravid female). Both oviparity and viviparity have been reportedin the more northern H. angulatus:A female from Colombialaid eggsthat hatchedin only 16days (Rossman,1973), and a specimenfrom Trinidadwas oviparous,whereas ones from Peruand the Amazon were viviparous (Rossman,1984; da Cunha and do Nascimento, 1981).Data on this genusare insufficient for detailedanalysis, but suggest two origins of viviparity. The North American Opheodrysuernalis has a remarkablvshort incuba- tion period in the northern (coldest)part of its range.Incubation averages 30 days in the Chicago region (Stille, 1954),but only four to 23 days in northern Michigan (Blanchard,1933; MacGregor, 1975; Sexton and Clay- pool, 1978). Blanchard (1933)suggested that occasionalviviparity was iikely, but was not observedby him. Also, maternalcare of the eggsmay be shown (Blanchard, 1933).A four-day incubation period is only slightly longer than the time taken for the young of severalviviparous speciesto emergefrom the egg sac after parturition (e.g., Saiphosequalis, Pseudechis porphyrincus,Greer, 1983; Shine, personal observation; Trrmeresurus okinaaensis,Koba et a\.,7970);hence this caseis included as an exampleof viviparity. The medium-sized African Psammophylaxoccur in grasslands.P. tri- taeniatusand P. rhombeatusare oviparous,with the latter speciesshowing maternal egg-guarding and prolonged oviducal retention of eggs (Fitz- simons, 7962;Broadley,7977).P. aarinbilisshows geographicvariation in mode of reproduction: the lowland P. a. multisquamisis oviparous with and viviparous (vertical egg-guarding,whereas the montane (> 1800m) P. u. aariabilisis viviparous ryhe;(B) the genus Sinona- (Broadley,197n . P. uariabilisis a slow-moving speciesthat doesnot rely on speed to escapepredation; in contrast, P' rhombeatzsis an agile, fast- moving snake(Fitzsimons, 1962). The viviparous P. a. aariabilisis found in cooler,more southernareas than the oviparousP. a. multisquamis(distribu- . tholloni(Loveridge, tion maps in Broadley, 7977). s. Both specieshave on th,ebasis of biochemical,karyological, scutellation, and cranialdata, s of reproductionare the large cosmopolitan snake genus Natrix has been divided into four iviparity in Grayin is genera(Rossman and Eberle,1977).Five Asian speciescomprise sinonatrix; ; with the viviparous Malnate (1960)had previously pointed out the closerelationships between ); Fitch, 1970).I infer these forms. oviparity has been reported in s. trianguligera(Kopstein, thin Grayia or in the 1938)and the egg-guardings. percarinata(Pope, 1935), whereas s. annularis ,r two generaextend is viviparous (Pope, 1935).There aPpear to be no records of mode of EVGILUTI(IN c'F VIVIPAtrITY IN FIEPTILEEI 65e THE reproductionin s. acquilasciataor s. bcllula.All the sinonalrixare semi- (Pope, aquatic, and are rarely found far from streams or flooded fields ro'3s1.Hatchlings of the oviparous s. percarinstttshow obvious egg teeth, whereasnewborn s. annularisdo not (Pope,1935). This obsen'ationsug- geststhat viviparity may not be too recent_anacquisition in s, annularis's innulctrisisfound firther north than any of the oviparousforms (Fig. 9; but seeTinkle and Gibbons,7977). All New world natricines(e.g.,l,lerodia, Regina, storeria, Thamnophis) are that viviparity evolYedduring viviparous.^migration Many authors have suggested the of oviparous old world natricinesacross the Bering strait (e.g.,lieill, 1964).Hbwever, Malnate (1960) notes that the viviparousOld W&ta S. annulorisis closeto the stock which gave rise to the New World natricines. Perhaps, viviparity arose only once in this group (in S' an- nularis). Somerecords of the evolution of viviparity in colubrid snakesmav be in error. Two reports of viviparity in the North American Diadophispunctatus (Ditmars, 7942.;Peterso", tqSb) are based on the sudden appearanceof neonatesin cagescontaining adult D. punctatus;the actualprocess of live birth was not observed.As extensivestudies by other workers revealedno evidenceof viviparity (reviewby Tinkle and Gibbons,7977), this caseis not considered further. Similarly, early reviews claimed that viviparity is shown by, and has evolved in, the colubrid genera Bo-iga(Mell' 1929)' Dendrelaphis,Hemirhagerrhis, and Meizodon (Neill, 1964).subsequent work indicatesthat all fourlenera are exclusivelyoviparous (Pope, 1935; CamP- den-Main, 7970;Broadleyand Cock, 1975)' 7. ELAPIDAE The only viviparous African elapid is the spitting cobta HemachTtus hnemachitus(Fitch, 7970);it is thought to be related to the true cobrasof the genus Ncia, based on similarity of overall morphology (Fitzsimons,7962) and venom fractions (strydom, 1,979).Naja are oviparous, and maternal egg-broodinghas been reported (Kopstein, 1938;Fitch, 1970)'Hemachatus is found in more southerly (and hencecooler) climates than are most of its over oviparous relatives 6ig. i0), and it inhabits areasfrom sealevel to 2500m elevation (Branch,1979). Viviparity characterizesall the hydrophiine seasnakes, but most or all laticaudine sea snakesare oviparous (Fitch, 7970).Laticauda colubrina has been reported to show both oviparity (Smedley, 1931; Dunson' 1975; shine, personal observation) and viviparity (smith, 793y T,aylor, L965; semper, cited in smedley, 1931).Two of the latter cases(Taylor, semper) *ere basedon observationsof femaleswith newborn young: thesemay be due to prolonged egg-broodingrather than viviparity (smedley, 7931). Howevei, the lecorl-which iJ based on dissection of a gravid female (smith, 1930)is difficult to dismiss, although it is surprising that subse- IVIPAFIITY IN EIEPTILEEI EVG,LUTIGINAFIY ()FTIGIN€T (fF VIVItrAFIITY IN SOUAMATE FIEPTILES : Sinonstrixare semi- flooded fields (Pope, \ ,w obvious egg teeth, This observationsug- tion in S. annularis.S. 'ous forms (Fig. 9; but trcria, Thnmntrpfii-s) are >arityevolved during ross the Bering Strait at the viviparous Old se to the New World this group (in S. an- rrid snakes may be in n Diadophispunctatus T rdden appearanceof \ actual process of live 'workers revealedno t977), this caseis not ed that viviparity is 'a Boiga (Mell, 7929), '4). Subsequent work s (Pope, 1935;Camp- rg cobra Hemachatus the true cobrasof the ;y (FitzSimons,7962) ,arous,and maternal :h,1970). Hemachatus s than are most of its Fig. 10. Geographic distributions of oviparous (horizontal lines) and viviparous (vertical om sea level to over lines) elapid snakeswithin (A) the genera Nal and Hemachatus; (B) Australian species of the genus Pseudechis. akes,but most or all aticaudacolubrina has quent workers have not discovered viviparous populations of laticaudines [931; Dunson, 7975; (-SolomonIslands native peoplesuggest thatLaticauda ctockeri is viviparous: 1930;Taylor, 7965; H. Heatwole,personal comrr,unication). This caseis not consideredfurther ses(Taylor, Semper) in the analysis. young: thesemay be Whethei or not the viviparity of hydrophiine sea snakes is likely to ity (Smedley, 1931). representan independentoiigin dependson whether they derive from the of a gravid female teirestrial Austrafian elapids lmany of which are viviparous) or from the rrprising that subse- laticaudines,most of which are oviparous.Morphological and biochemical THE EVOLUTIGIN ClF data (McDowell, 1969;Minton, 7979;Cadle and Gorman, 7987;Mao et al., 1977, 7983)are ambiguous on this point, so the caseis not considered further. Until recently, all the large venomous snakesof the genus Pseudechis, which are widely distributed in Australia and New Guinea,were thought to be viviparous (Worrell, 1963).However, oviparity is now known for P. australis(Fitzgerald and Pollitt, 1981;Shine, unpubl.), P. guttatus(Charles et aL,7979),P. colletti(Charles et al., 1983),and P. papuanus(E. Worrell, personalcommunication; Shine, unpubl.). The only viviparousspecies, P. porphyriacus(Shine, 7977), rnhabits cooler and wetter areas than do its oviparouscongeners (Fig. 10). Viviparity occurs in at least 13 additional generaof Australian elapid snakes(Acanthophis, Austrelaps, Cryptophis, Denisonia, Drysdalia, Echiopsis, Hemiaspis,Hoplocephalus, Notechis, Rhinoplocephalus, Suta, Tropidechis,Un- echis;Cogger, 7975, and Shine, unpublished), whereas ten genera are known to be oviparous (Cacophis,Demansia, Furina, Glyphodon,Neelaps, Oxyuranus,most Pseudechis,Pseudonaja, Simoselaps, Vermicella). These gen- era have been divided into four groups on the basisof venom gland mus- culatureand the morphology of the hemiPenes(McDowell, 1967)'All four groups contain both oviparous and viviparous genera,suggesting at least four origins of viviparity. However, the classificationis not consistentwith karyotypic data (G. Mengden, personalcommunication). Data on scalation (Cogger, 7975)are significantin this respect:All of the viviparous genera except Acanthophispossess undivided subcaudals,whereas all of the oviparousgenera display divided subcaudals(Shine, 1985). The correlation between subcaudal scalation and reproductive mode suggeststhat the oviparous and viviparous taxa represent distinct phylogeneticlineages. Acanthophismay representan additional independentorigin of viviparity, but only two other origins are likely in the Australian elapids; one in Pseudechisand one far back in the phylogeny of the group. There are conflictingreports on reproductivemode in the elapid genus Cacophis(Tinkle and Gibbons, 1977), but a recent study demonstrates oviparity in all three species(Shine, 1980b).Viviparity in the Oriental genus Calliophiswas suggested,but not documented,by Mell (1929)and Neill (1964).Oviparity has been reported in C. japonicus(Fukada, 7965), C. maculiceps(Phelps, 1981) and C. melanurus(Deoras, 1971). A gravid female C. macclellandilcontained oviducal eggs with embryos up to 38 mm long (Pope,1935), but this recordalso has been interpretedas indicatingovipar- ity (Campden-Main, 1970;Whitaker, 1978).Further data are needed to clarify reproductivemodes in this genus. A. VIPEFIIDAE In the "Agkistrodon" group, large and deadly pit vipers of Asia and the Americas, two speciesare known to be oviparous (Deinagkistrodonacutus, /IFAFIITY IN FIEPTILEEI EVG'LUTIG,NAFTY C!]IIGIINCr ctF VIVI]'AFIITY IN €IOUAMATE FIEPTILES| ran, 1981;Mao et al., Calloselasmarhodostoma, Pope, 1935),whereas seven others are viviparous se is not considered (Agkistrodonbilineatus, A. contortrix,A. hnlys,A. himalayanus,A. pisciuorus, Hypnalehypnale, H. nepa,Chang and Fang, 1931;Pope, 1935;Kopstein, :he genus Pseudechis, 1938;Fukada,7962; Telford, 1980). The modeof reproductionis unknown "apparently iuinea, were thought in A. caliginosus;A. strauchiis ovoviviparous," and the sameis is now known for P. suggested for A. monticola(Pope, 1935).Taxonomists disagree on phy- , P. guttatus (Charles logeneticrelationships within thesepit vipers. Brattstrom's(i96a) phy- )apuanus(E. Worrell, logeny suggeststhat viviparity has evolved independently at least twice, 'iviparous species,P. and possibly four times, within the genus. Burger (1971)removes both er areas than do its oviparous forms to the single genus Calloselasma,which suggeststhat viviparity has evolved only once. Recent revisions (Gloyd, 1977, 7979) "Agkistro- of Australian elapid incorporatean earlier suggestion(Chernov, 7957)and partition Drysdalia, Echiopsis, don" into four genera.The two oviparousspecies occupy monotypic gen- iuta, Tropidechis,Un- era (Calloselasmarhodostoma, Deinagkistrodon acutus). Two small viviparous reas ten genera are snakesof Sri Lanka and peninsularIndia (A, hypnaleand A. nepa)ate placed Glyphodon, Neelaps, in the genusHypnale, and the remainingseven viviparous species remain rmicella). These gen- in Agkistrodon. rf venom gland mus- In the wider sense,whether one or two separateevolutionary origins of 'Agkistrodon" twell,7967). All four viviparity have occurredwithin dependson the relationship r, suggesting at least betweenthe two oviparousforms, D. acutusand C. rhodostoma.Ifthey are s not consistent with closerto each other than to any of the viviparous forms (as suggestedby n). Data on scalation Burger, 7971; and R. Conant, personal communication),then viviparity te viviparous genera may have arisen only once. If they are related more closelyto viviparous whereas all of the forms than to eachother (as in Brattstrom's1964study, or the serological 985). The correlation analysisof Kawamura, 1974),multiple origins of viviparity are suggested. "Agkistrodon" e suggests that the The conservativeposition is that provides only a single case rylogenetic lineages. of the evolution of viviparity. origin of viviparity, The analysisis confusedby geographicdistribution: Both oviparity and "Agkistrodon", lian elapids; one in viviparity occur in Asian but all three American speciesare "that ;rouP. viviparous. A review concluded no correlationbetween habitat pref- :in the elapid genus erenceor latitude of distribution and method of production of young can study demonstrates be found" (Pope, 1935).However, it seemsclear that the viviparous Old rity in the Oriental World "Agkistrodon" tend to have a more northerly distribution than the , by Mell (1929) and oviparous ones. The viviparous A, halysoccupies much colder climates "Agkistrodon" rs (Fukada, 1965), C. than doesany other (Fig.11).Female Calloselasma rhodostoma 71). A gravid female guard their eggs after laying (Smith, 1943);this behavior may facilitate the s up to 38 mm long evolution of viviparity (Shine and Bull, 1979).Considerable embryonic as indicating ovipar- developmentoccurs prior to ovipositionboth in this speciesand in Deinag- data are needed to kistrodonacutus (Pope, 1935;Smith, 1943;Wall, 7927).The incubation pe- riod of the former is only 30 days (Phelps,1981). Cerustes(: Aspis) contains two small heavily built vipers occupying similar geographic ranges in sandy areasof North Africa and southwestern Asia. Cerastescerastes is oviparous (Petzold, 1968b;Dmi'el, 1970;Saint Gi- rers of Asia and the rons, 7962),and C. aiperais probably viviparous, despite conflicting state- inagkistrodon acutus, ments in the literature (Werner,1930;Vogel, 7963;Petzold,1968b). Cerastes THE EVG'LUTIGIN ctF VIVIPAFIITY IN FIEPTILEEI \,'-- ,a\ -----1_-__- L$Fo - "B-a' r- l \YK /t, r\"\ Fig. 11. Geographic distributions of oviparous (horizontal lines) and viviparous (vertical lines) viperid slnakeswithin (A) the old World speciesof the genus Agtlstrodon;(B) the species Echisurinatus. viviparous in both the eastern(Mendelssohn, 1963) oiperais known to be "Biskra and central (Domergue,1959) parts of its range, as well as in the 1951). Oasis" (Schetty, "E. Echisis closlly related to Cerastes,and the catch-allname carinatus" has recentlybeen shown to include multiple species(Hughes, 1'976;Cher- lin, 1981).Echis coloratus of southwesternAsia is oviparous(Mendelssohn, 1965;Dmi'el,1970), but the E. carinatusgroup shows geographicvariation in the mode of reproduction (Fig. 11). Specimensfrom Russia,India, and pakistanare vivipirous (Petzold,1,968b; Deoras, 1971; Mendelssohn, 1965; Stemmler, 1969, 1'970;Stemmler-Gyget, 7965;Minton, 1966; Whitaker, EVGILUTIG'NAFIY ctFTICiINS crF VIVIPAFIITY IN SiGIUAMATE FIEPTILESi 1g78;Isemonger, 7962; Kramer and schnurrenberger,1963; Terentev and Chernov, 1965;,whereas those from North Africa are oviparous(Mendels- sohn, 1965;Isemonger, 1962; stemmIer,7971; Kramer and schnurrenber- ger, 7963;Duff-Ma&ay, 1965),as are those from Turkey and Asia Minor (Kru-". 1963).Tinkle and Gibbons (7977)ques' and Schnurrenberger, ','E. iioned Duff-Mackay's(1965j record of oviparityin African carinotus," but notedthat oviparityhad alsobeen recorded by otherauthors. Unfortu- nately, the mode of reproductionis unknown for west African specimens, upu.i f.o^ an unsubsiantiatedcomment in a popularbook on the region "E. (-ansdale, 1961)that csrinstus"(one of at leasttwo local species)is viviparous. The boundary betweenoviparous and viviparous populations of '8E.chrinalus" occurs in the mountainouscountry of Iran and Afghanis- "E. tan; carinatus"occurs up to 1800-melevation (Petzold,1968b)' A nen' species(E. multisquamatus..Cherlin, 1981) has recently been described from nbrthern and easternIran and southernAfghanistan: Data on reproduction of this form would be of great interest. The transition from oviparity to viviparity apparentlyo...t.i in this region of severelycold climate.In other areasof its-range, ihe speciesgroup occurs under much milder climatic popula- conditions€ig.1f y. fhiJ is true of both oviparousand viviparous tions. Trimeresurusis a large group of oriental pit vipers relatedto the Ameri- can genus Bothrops;the latter group is not consideredhere, becauseall are apparently viviparous (Fitch, 7970),and it is likely to be a monophyletic gio"p. Tiimereiurusincludes both terrestrialand arborealforms, many of *ni.i-t are highly venomous. Modes of reproduction are known for a few T. species:T, elegans,T. flaaoairidis,T. monticola,T. mucrosquamatus,and toknrensisare oviparous (werler and Keegan, 7963;Koba, 7977;de Rooij, 7917;Pope, 1935;Nakamoto and Sawai, 1982),whereas T' albolabris,T' gramineui, T. jerdoni, T. okinaaensis,T, popeorum,T. puniceus,T' purpu- Teomaculatus,T. stejnegeri,T. trigonocephalus,and T. wagleriare viviparous (abovereferences and Kopstein, 1938;Fitch, 1970;Stettler, 1971;Nicker- son, 1974;Phelps, 1981;Reitinger, 1978).Extensive ecological and repro- ) and viviparous (vertical Agkistrodon; (B) the species ductive data are availablefor some species(Fukada, 1965)' A reconstruction of the phylogeny of Trimeresurussuggests that vivipar- ity may have evolved independentlymany times within this group (Bratt- (Mendelssohn,1963) ,t.orr,,- 1964). However, most inferred pathways are tenuous; the two "Biskra vell as in the strongestcases are small speciesgroups, each containing oviparous and vivipirous members.The first of thesegroups comprisesthe oviparousT. "E. lname carinatus" (both from and the viviparous T. 'Hughes, flaaoairidisand T. toknrensis Japan) ierdoni 1976;Cher- from eastern China and Indochina. The second grouP, placed within rous (Mendelssohn, ooophisby Burger (7977),contains the stout-bodied terrestrial forms T. geographic variation monticola,u., o'oipurousspecies from southeastChina and Malaysia,and T. m Russia,India, and okinauensis,a live-bearer itom the Ryukyu Islandsnear Japan. Hence, both ;Mendelssohn,1965; casesinvolve a Japaneseform (or forms)and a Chineseform; the difference on, 1966; Whitaker, is that in one caseit is the Chinesespecies that is viviparous, whereasin ctF VIVI]'A' TTY IN FIEPTILEE| 6EA THE EVG,LUTIC,N is the other caseit is the Japanese(Fig. 12)' Egg-guardingby the female a known inT, monticola,and. eggs of this speciesare retainedin uterofor long period prior to oviposition (Pope,1935); this specieshas evolved part of the way toward viviParitY' Vivipaiitv of the New world crotalinesBothrops, Ctotalus, and sistrurus to interpret. BothroTtsmay b9 derived from (Fitch,' 1970) is difficult "Agkistrodon" Trimeresurus,and. Crotalus and Sistrurusfrom the group (Brattstrom, 7964;Koba, 1973);in each casethe ancestralgroup contains viviparous forms, so thesetaxa will not be treated as independentorigins of viviparity. The members of the genus vipera are old world in distribution, and almostall are terrestrial.viperalebetina, v. (formerlyPseudocerastes) persrca' and V. xanthinaareoviparous. A11 have fragmentedhead shields (Marx and Rabb, 1965)and are piobably closely related to each other. Viperaammo- dytes,V. aspis,V. beris, V. Iqtastei,V. russellii,and V. ursiniiare viviparous have been claimed to be 1iit.lr, 19Zb).Both V, lebetinaand V. xanthina viviparous as well as oviparous (Terentevand chernov, 7965;Mendels- sohn, to63;. The following review suggeststhat v. Iebetinais oviparous throughout its rangebut that the V. xanthinagroup doesshow both modes of reproduction. The large v.Iebetinaoccurs through northwesternAfrica and southwest- ern Asia (5teward, 1977).Asingle riport of viviparity for the entire specie,s (phelps, 1981)remains undocimented. Five subspeciesare recognized' ih" Afri.u., iiprro I. mauritanica(V. mauritanlcaof someauthors) has been shown to be oviparous (Kratzer, 1968),despite earlierre_cords to the con- trary (Werner, 7b30;Vogel, 1963).Vipera l. obtusafrom the central part of the speciesrange also iJ oviparous (Y..ratzer,1968; Muskhelishvilli, 1971), contrary to an Jarher speculation(Terentev and Chernov,1965).Vipera l. lebetina'isoviparous (vogel, 1963;Kratzer,1968;Arnold and Burton, 1978; contrary r".oid by Petzold, 1968b),as are 1z.l. schweizeri(Schweizer, 1956; Kratzei,1968) and V. I. turanica(Terentev and Chernov, 1965;Kratzet, 1968;Kornev:a, 7973).In summary, there are well-documentedcases of oviparity in all subspeciesof V.lebetina, and there are no authenticated casesof viviparity. "vipera The situation is more complex in the caseof the xanthina'ro]up" of Asia Minor and Israel. viperapalaestinae from Israelapparently is ovipa- rous (Mendelssohn,1963; but seePhelps, 1'981),whereas V' xanthinafrom Turkey and surrounding areasis viviparous (Kratzer,1968; Diesenet,1974; Andren and Nilson, I9/6b). A closely related viper, V. raddeifrom Turkey, Iran, and central Asian parts of the USSR is viviparous (Y,.ratzer,7968; Darevsky,1966; Terentev and Chemov, 1965).The oviparousv,. palaestinae occupiesa warmer and drier region than do its viviparous relatives(Fig' 12). Eggs of Vipera palaestinaecontain advanced embryos when laid been regardedas sub- 6och;; 1963).All three of the above specieshave speciesof V. xanthina,but a recent reviiion (Nilson and Sundberg, 1981) tF VIVIPAFIITY IN FTEPTILE€i is 4uardingbY the female are retained in utero for a "at s specieshas evolved Part nts,Crotalus, and Sistrurus rs may be derived from "Agkisfrodort" i^ the grouP gru : ancestrll group contains ed asindePendent origins ano ,Voridin distribution, *\ rerlvPscudoc erasf es) Persicn' -.9 shields(Marx and tediead O9E amtno- tr9b! r each other. ViPera = 4.^ td V . ursinii are viviParous qos to be have been claimed N.-i I Chernov, 1965;Mendels- h -i< is oviParous nt V. lebetina oPF :up doesshow both modes E" P c9: Africa and southwest- o:i.L ;tern * ;or for the entire sPecie.s .parity - ;,8 are recognized' subspecies 'E 'the "ViPeraxanthina grouP" m IsraelaPParentlY is oviPa- 1),whereis V, xanthinafrom '.tatzer,1968; Diesen et, 1974; viper, V. raddeifromTurkeY' ,suirriouto.tt (Y'.ratzer,7968; . The oviParousV' Palaestinae its viviParousrelatives (Fig' ranced embrYos when laid ; have been regardedas sub- (Nilson and Sundbetg' 798I) ()F IN FIEPTILES THE EVGILUTIG'N VIVIPAFIITY Bcicr As might be expected from elevatesall of them to full specific status' y. are more closelyrelated to reproductivemode, v. raddeiind rhnthina than to V. eaih--vi"ip"ri,y other Palaestinae' A11of the muy hu,r! evolved more than once n'ithin vipera. present-dayoviparousformsshowfragmentedheadshields'a"derived" many viviparous relative to the unfragmented head shields of character *ipers (Marx and Rabb, 1g65i'Hence' the present-dayovipar.us il;.i;t iot*s. Either (1) oviparity has are unlikely to be ancestralto the viviparous in most present-day vipers evolved from viviparity, or (2) viviparity ,,rirtii'", forms retained the evolved from the group,,, uut tne viviparous whereastheir oviparous nrimitive feature of unfragriented head shields' in most euotu"a fragi-rentedhead shields, or (3) viviparity ;;;;;;;; hy- ir";a now-extinct o'iparous ancestor..rhe third fi:;'h;;;;;;;J at leasttwo evolutions oothesisseems the most parsimoniousandimplies bf viviparity in ViPera' b'v.aproposed Finally, at leastone more origin of viviparity is suggested 1965;their Fig' 46) in which phylogeny of 'lpe.lJs.ute, llviuo and Iiabb' thetwoviviparousAfricangenera(AtherisandBitis,Fitch,l9T0)are thought to hive evolved from oviparous forms' VII. EVALUATIG'N c,F CAsiE HISTG'FIIES A. Frequency of Evolution of Viviparity origins of.viviparity The precedinganalysis reveals at least95 independent to identify vivipa- among the living squamategto-"!t Previousattempts 7977)and 38 (Shine and rous origins ,".og.u"Jat liintte and Gibbons' 1981)and75 (Blackburn' V.rllt,OVOI origini within gL""'u, or 71 (Shine' suprageneric ileil .ur", if u'ttorigl.,t of"viviparity,-includingthose.at.the. family.is viviparous) are level (i.e., where an entire g"ttt', iubfamily' or communication)' includea. Reviews by Blacfburn (1982,1984' personal increasedthis total to 76 "r,tirufy independent'of the present.study' have (4Tinlizards,23insnakes,linamphisbaenians)'ascomparedto95(63in present.study'The two lizards, 30 in snakes,2 in amphisbaenians)in the in the evolution of analyses generally ugr"" o^^the.lineages involved unpublished recordsof il;"",y:;u.,y or th"ediscrepancies are due (1) to for the present study, ,"piJ"iit"e bimodality and^phylogeny^obtained gtactbum, in interpretation uifi.,o"gt unavailable to ind (2) to differences example' Blackburn autl'on reproductive mode and phylogeny' For. ol of (personal communicalion) argues thai phylogenetic reconstructions ,t*" tu*u (e.g., Liolaemus,Eremias) are unreliable' identified in Although the number of independent evolutionary origins reviews' it is the preseit study is much higher than revealed in previous the conservative fifceil.to be well below the uittul number. This reflects VIT'AFIITY IN FIEPTILEA EVALUATIGIN clF GASIE HIEITC'F|IEA 6€i 1 approach, which is coupled with inadequate information on the reproduc- ;ht be expectedfrom rore closely related to tion and phylogeny of most reptiles. However, two weaknesses of the current estimate should be noted. The number may be reduced if oviparity thin Vipera.All of the has often evolved from viviparity (rather than the reverse) or if available "derived" I shields,a phylogenetic reconstructions are seriously in error. Nonetheless, addi- s of many viviparous tional information is more likely to increase than decrease the total. -day oviparousvipers Of the 95 identified origins, some are less well-documented than others. lither (1)oviparity has The data are least reliable in the 30 cases in which relatively distantly it present-dayvipers related species with different reproductive modes are compared. Examples 'us forms retained the are comparisons at the suprageneric level, for example, between genera rereastheir oviparous (Hemachatusversus Naja) or subfamilies (Boinae versus Pythoninae). It may i3) viviparityin most be that these groups are not as closely related to each other as we currently rcestor.The third hy- believe and that the viviparous group has inherited viviparity from an rt leasttwo evolutions unrecognized viviparous lineage. Perhaps the best example here is the viviparous Psammodynasfes;I have considered it an example of the evolu- tion of viviparity because its phylogenetic affinities probably lie with ;gestedby a proposed their Fig. 46) in which oviparous colubrids rather than with any recent viviparous species. llfls, Fitch, 1970) are The evidence is more reliable in the 65 instances in which congeneric species differ in reproductive mede. Most striking are the ten documented casesin which both viviparity and oviparity occur within a single speciesor small species-group (Table I). Only the single species Sceloporusaeneus was iToFlrEs accepted as containing both oviparous and viviparous populations in a recent review (Tinkle and Gibbons, 1977), but this clearly is an underesti- iviparity mate. Reliable documentation of reproductive bimodality is available for at least ten such taxa, and data on a further eight species suggest that the nt originsof viviparity same phenomenon occurs (Table I). It is interesting that all eight of the pts to identify vivipa- doubtful cases belong to genera in which viviparity is known to have 77) and 38 (Shine and evolved. This may reflect either multiple origins of viviparity in related 1) and 75 (Blackburn, species, or misidentification of similar species differing in reproductive se at the suprageneric mode. Further data on these taxa would be of great interest. nily is viviparous) are Many taxonomists would view mode of reproduction as a taxonomic cnal communication), character of sufficient importance to warrant elevation of related oviparous :reasedthis total to 76 and viviparous forms to full specific status. Recent studies have suggested comparedto 95 (63in such divisions in the Vipera mnthina group (Nilson and Sundberg, 1987) and the Echis carinatus group (Cherlin, 1981). Future taxonomic studies resentstudy. The two "subspecies" I in the evolution of may well divide many of the oviparous and viviparous listed .npublishedrecords of in Table I. Nonetheless, these taxa provide exceptionally clear-cut exam- or the present study, ples of the evolution of viviparity within small groups of closely related rncesin interpretation animals. For six of the ten taxa, data are available to show that prolonged r example, Blackburn retention of eggs in utero is common in oviparous populations (Helicops, lic reconstructionsof Rossman, 1973; Lerista, A. Greer, personal communication; Opheodrys, Blanchard, 1933; SceloporusAeneus, Guillette, 7982a; Saiphos, Bustard, 1964; Greer, 1983; Vipera, Kochva, 1963). ry origins identified in previousreviews, it is Despite the many evolutionary transitions from egg-laying to live- lects the conservative bearing within the squamates, no single population of any species shows TABLE I SquamateSpecies Reported to Show Both Oviparity and Viviparity Species Source RecordsConsidered Relittble AbIeph ar us biuittatus" Terentevand Chernov, 1965 Echiscarirntus" Kramer and Schnurrenberger,1963 Helicopsangulatus Rossman,1973; da Cunhaand dos Nascimento, 1981;D. A. Rossman,personal communicatton Leristn ltou ga inttillei A. E. Greer,pcrsonal communication; Shine, per- sonalobservation uplltodrys11htcfitaLts' Blanchard,1933 Psammophylax z,ariolt ilis Broadlev,1977 Sceloporusaeneus Guillette,198i, 1982a Saiphosequnlis Shineand Thompson,unpublished; Creer, 1983 S1irenomorphus nigr ica ud u s A. E. Greer,personal communication Viperaxanthina" Kratzer, 1968 Recordsof UncertainValidity Amblyodipsasconcolor Broadley,1983 Mabuyacapensis W. Haacke,personal communication M. carinata Smith, 1935;Badhuri, 1943 M. sulcata D. Horton, personalcommunication Phryno cephalus theobal di Sergeev,1940 Pseudechisguttatus Charleset al.,7979 Rhacodactulus leachianus Mertens,1964 Sceloporusuariabilis Werler,1951; Fitch, 1970 Sphenomorphuspardal is Rankin, 1978 Typhlopsdiardi Wall, 1918;Smith, 1943 RecordsThat Are ProbablyInrralid Aspiduradrummondhayi Gans and Fetcho,1982 Boaconstrictor Hoover, 1936;Tinkle and Gibbons,1977 Cacophiskreffti Cogger,1975; Shine, 1980b Diadophispunctatus Ditmars, 1.942;Tinkle and Gibbons,1977 Incertaaiaipara Lantz, 1927;Tinkle and Gibbons, 1977 Lachesismuta Mole,].924 Iaticaudacolubrina Smith, 1930;Smedley, l93t; Taylor, 1965;Fltch' 1970 Mabuya quinquetaeniat a Fitch, 1970;Spellerberg, 1976 Pythonregius Pope,1967 Sceloporusgrammicus Smith, 1939 Trimer e sur u s okinaaensis Fukada,1965 Xenodermusjaaanicus Taylor,1965 ,oviparous and viviparous subspecies elevated to full specific status by some authors. bviviparity not seen but incubation period variable and sometimes so brief (4 days) that it "viviparous" resembles other reptiles. EVALUATION ctF CASE HIgiTclFIIEEi y and Viviparity both oviparity and viviparity. Even when a speciesexhibits both modes of reproduction, the egg-layersand live-bearersinhabit different geographic areas. Facultative shifts of reproductive mode are unknown. El. Taxonornic Eliases , 1963 The 95 identified origins of viviparity encomPassall three squamate sub- I dos Nascimento, orders, and are distributed among 20 families (Table II). A compari- rl communication son of the squamate suborders shows that viviparity has arisen at ieast 63 nication; Shine, per- times in lizards, 30 times in snakes, and two times in amphisbaenians' However, these apparent differences among suborders are due largely to different numbers of speciesin each group. If the data are calculatedas the number of origins of viviparity per oviparous species in the suborder, lizards averageone origin per 47 species,snakes one origin per 54 species, lished;Greer, 1983 and amphisbaeniansat least one origin per 40 species.[In this and subse- nication quent calculations, species numbers are taken from the checklist of the U.S. National Fish and Wildlife Laboratory (7978); the proportion of oviparous forms in each taxa is taken from Fitch (1970), and other sources.] Chi-square analysis demonstrates that the relative frequency of evolution among the three suborders (ex- rication of viviparity does not differ significantly pected: g5originsin4693 species : .020;d.f. : 2,X' :1.2,p>.5). One corollary of this result is that the higher proportion of viviparous species in ication snakes, as compared to lizards (Tinkle and Gibbons, 1'977),is not attribut- able to a higher frequency of the evolution of viviparity in snakes. Appar- ently, it reflects a greater speciation rate, or lower extinction rate, in viviparous as opposed to oviparous snakes, a phenomenon that is less pronounced in lizards. The minimum number of separate origins of viviparity per squamate family ranges from zero to 34 (Table II). As is true in the comparison among suborders, the apparent differences among families in the frequency of evolution of viviparity are partly due to different species numbers in each account, large differ- bons, 7977 family. However, even when this factor is taken into ences among families remain (Table II). For example, the small family Anguidae (10 oviparous genera, 75 species) shows at least six origins of bons,1977 viviparity, whereas the larger Teiidae (38 genera, 230 species) is entirely ns, 1977 oviparous. Among the lizards, the relative frequency of evolution of viviparity is highest among the anguids, skinks and cordylids, and lowest 1970 rylor, 1965;Fitch, among teiids, agamids, and geckoes. Among the snakes, viviparity has arisen more frequently among vipers, boids, and elapids than among lep- totyphlopids, typhlopids, or colubrids. Statistical analysis convincingly re- jects the null hypothesis of equal probabilities of the evolution of viviparity (per oviparous species) in the different families of lizards and snakes (n : 93 origins, 18 families, excluding caseswhere entire families are viviparous or have less than 20 oviparous species;X' : 167.2,77 d'f., p << .001)' us by some authors. Similar taxonomic biases are evident at the generic level. The present rs so brief (4 days) that it I l THE EVclLUTIctN clF VIVIPAFIITY IN FIEPTILE€I TABLE II Minimum Estimatesof the Number of EvolutionaryOrigins of Viviparity in EachFamily of SquamateReptiles" Nurnberof Originsof Viviparitv Absolute Numberper Oviparous Suborder Family Number Speciesin the Taxon Amphisbaenia Amphisbaenidae 1 0.01 Trogonophidae 1 Sauria Agamidae 2 0.01 Anguidae 6 0.08 Chamaeleonidae 2 0.03 Cordylidae 1 0.03 Dibamidae 0 0 Gekkonidae 2 0.003 Helodermatidae 0 0 Iguanidae 12 0.02 Lacertidae 3 0.01 Pygopodidae 0 0 Scincidae 5+ 0.05 Teiidae 0 0 Varanidae 0 0 Xantusiidae 1 _b Serpentes Aniliidae, Uropeltidae 1 _b Acrochordidae 1 _b Typhlopidae 1 0.01 Boidae I 0.03 Tropidophiidae 1 _i, Colubridae 14 0.01 Elapidae 3 0.02 Leptotyphlopidae 0 0 Viperidae 8 0.27 'Entirely viviparous families ornitted, unless phylogenetic data suggest an independent origin of viviparity in ancestors of that group. bCannot calculate: all present day representatives are viviparous. cCannot calculate: reproductive data missing for most species. review has identified 65 origins of viviparity within a genus among the 619 squamategenera containing oviparous species,so that the probability of viviparity evolving within any one genusis 65 in 619,or 0.10.The probabil- ity of two origins occurringby chancewithin a single genus is 0.10', or 0.01.Hence, multiple origins of viviparity within a singlegenus should be extremely rare. In practice, the data reveal that more than half of the within-genus origins (36 of 65) occur in genera with multiple origins of viviparity. Thirteen generaaccount for these36 origins. This figure proba- VIPAFIITY IN FIEPTILEA EVALUATIGIN OF CASE HIEITcltrIE€I bly underestimates multiple origins, because they are difficult to detect igins of Viviparity in without strong phylogenetic data. Even allowing for the fact that many genera would be unlikely to evolve viviparity because thev contain onlv one or two species,this concentration of the evolution of viviparity within Origins of ViviparitY a few genera is certainly nonrandom. Number per Oviparous C. Evaluation of Hypotheses Speciesin the Taxon 0.01 1. GUESTIONS ,A.ND C)ONCEPTS _c The present-day distributions, morphologies, ecologies, and behaviors of 0.01 indication of the conditions 0.08 the species reviewed above may provide some 0.03 under which viviparity evolved. Any selective pressure important in the 0.03 evolution of viviparity could be revealed at any or all of the following three 0 taxonomic levels: 0.003 0 1. Viviparity has evolved more often in some squamate families than in 0.02 others. Do the former groups show the characteristics predicted by 0.01 theory, whereas the other families do not? 0 2. Do the oviparous members of the genera in which viviparity has 0.05 to a greater 0 evolved, show the theoretically predicted characteristics 0 degree than do other squamate genera? _1, 3. Do the oviparous and viviparous forms within a genus(or species)differ _t, from each other in the characteristics predicted by theory? _b 0.01 Any trends seen at the familial or generic level (categories 1 and 2) are 0.03 likely to reflect factors that protoadapt a speciesto the evolution of vivipar- -b ity. They will be characteristics shared by all or most members of a phy- 0.01 logenetic lineage. Of the hypotheses reviewed earlier, those concerning 0.02 species characteristics (e.g,, defensive ability, arboreal, aquatic, fossorial, 0 habits; physiological constraints) are likely to 0.27 or secretive egg-guarding; belong to this category, because closely related speciesare unlikely to differ ;gestan indePendent origin markedly in these respects. In contrast, consistent differences among oviparous and viviparous con- geners (category 3) are likely to reflect factors that are directly involved in the transition from oviparity to viviparity. Hypothesized environmental influences (e.g., climate, habitat) may vary among related species and, genusamong the 619 hence, may be involved at this level. The comparison among closely re- hat the probability of lated oviparous and viviparous species is a powlerful one, because of the or 0.10.The probabil- -^ general similarity of the species being compared. This similarity controls . -^) mode. However, it may be lle genus ls u.ru-, or for many variables unrelated to reproductive ngle genus should be difficult to distinguish between intraspecific differences that have favored rore than half of the the evolution of viviparity and the differences that have arisen subsequent h multiple origins of to the acquisition of viviparity. The change from oviparity to viviparity may ns. This figure proba- in itself select for a suite of related adaptations, for example, changes in THE EVGILUTIG,N clF FIEPTILEsl thermoregulatory strategy (Guillette et al., 1980), adoption of secretive habits by gravid females, invasion of colder areas,even an increasein the size of offspring (Shine, 1978).Year-to-year variance in egg survivorship is likely to be decreased,thus affecting selectivepressures on optimal repro- ductive effort and age at sexual maturity (Murphy, 1968). Differences in size of offspring and reproductive seasonality among oviparous and viviparous subspecies of Sceloporusaeneus (Guillette, 1982a), confirm that the evolution of viviparity may be accompaniedby other changes. The following analysis examines whether the characteristics predicted by theory are exhibited by the taxa in which viviparity has arisen; either by all or most species within a family or genus or by the viviparous as op- posed to the oviparous species. 2, ANALYSIS AT THE F,A.MILIAL LEVEL Each squamate family containing more than 20 oviparous species was rated with respect to whether the majority thereof exhibits the characteristics predicted by theory (Table III). Hypotheses concerning environmental un- predictability, intensities of egg predation, frequency of reproduction (sin- gle versus multiple clutches), and thermoregulatory mode could not be tested, either because of difficulty in framing specific predictions from these ideas or because of lack of the necessary data. The number of charac- ters scored as positive for each family ranged from zero to five. This score was not correlated with the relative frequency at which viviparity had evolved (combining data from Tables II and IlI, n : 78, r' : 0.04, n.s.). Indeed the combined score for the nine families with the highest frequencv of viviparous origins (20) was almost the same as that for the nine families with fewest origins fn : 18, Table III]. Some trends are apparent from an analysis of individual factors. Most taxa tend to occupy hot rather than cold climates, as would be expected from latitudinal gradients in reptilian species diversity. Very few families include primarily aquatic, arboreal, or slow-moving species or forms that are largely restricted to very dry or very moist soils (Table III). Maternal care occurs in several of the families, especially among snakes, but is not clearly correlated with the evolution of viviparity at this level of analysis. Within snakes, viviparity may have arisen most often in large and venomous species (Table III). The relative frequency of viviparous origins is highest in the Boidae, Viperidae, and Elapidae, is intermediate in the primarily nonvenomous Colubridae, and is lowest in the small fossorial Leptotyphlopidae and Typhlopidae (Table II). Although venomous species comprise fewer than one-fourth of the colubrid genera (Underwood,1979), they represent six of the 14 identified origins of viviparity within the Co- "ad- lubridae. Overall, viviparity has arisen at least 25 times within the vanced snakes"; eight of these origins occurred within the approximately 170 nonvenomous oviparous genera of the Colubridae (percentage of ori- .I]'AF|ITY IN FIEPTILEB -9 l = doption of secretive N-O---a6NNN6#6-OO t o :n an increasein the F=z r egg survivorship is a 'es on optimal rePro- o ='a 1968).Differences in 3 o .ong oviparous and qJ z 7982a), confirm that .: t !; ther changes. I& d a racteristics predicted E & has arisen; either by G ? 6l L g'E he viviparous as oP- r x^.XXXX^X (, o ;U r E3, + la> e E OJ u L 2' q, .!> q xx EVEL I ;59 x)< n o 33 roi qo .-> rus specieswas rated >' o -t !c ts the characteristics ,90 € -J 3:"1 g environmental un- a a c- ;:-2< k \ 7.: (sin- A of reproduction o .9 d i,d mode could not be ,t o ific predictions from H! eo a4 oF o<> he number of charac- ! srh sb >-: rro to five. This score ;x o E A- 4 ..E o vhich viviparity had '? Fr u-* .E !wr! 78,l:0.04,n.s.). m-o he highest frequency o s : for the nine families + 2'' U u=o L ividual factors. Most :6 o XXX XXXXX XXX I cs s would be expected I u9. q U families ty. Very few (')H 3 >q species or forms that I k d6- Maternal tr$ (Table III). / I"-o ng snakes, but is not U ocn / 3 =^y lhis level of analysis. I s P t 5!b'io. l> / =! t often in large and e tr - E## t e f; q g* +ie+ ", of viviparous origins - .=: a gr ; intermediate in the lql * { t n€e+e+ *t *t r tF*:*€; H '2' frt n the small fossorial bo F!**Fi Ei 3!s5 E i E3 5q-s s iEie a) gh venomous species n t= 2 i r (Underwood,1979), 3.i; h d .qIEE parity within the Co- "ad- Lrneswithin the .!-:o in the approximatelY 3, ANALYSIS AT THE GENEFIIC LEVEL Factors that protoadapt a speciesfor the evolution of viviparity also may be looked for at the level of the genus (or group of genera) in which viviparity has arisen. Many of these groups show the characteristics predicted by theory fiable IV), but the problem is to determine whether the frequency of any factor among the groups is higher than would be expected by chance (i.e., from a random sample of squamate genera). Such "expected" frequencies from a sample of 7077squamate species are given by Shine and Bull (1979;their Table 3). Analysis indicates that most characteristics tested are found about as frequently in the taxa with viviparous relatives as in "random" Shine and Bull's (1979) sample. The proportion of arboreal taxa "random" is 0.14 in Table IV and 0.10 in the sample; the difference is not significant (X2 : 0.06, 7 d.f ., n.s). Corresponding figures are 0.05 and 0.05 for aquatic taxa (12 : 0.0, 1 d.f., n.s.), 0.20 and O.fSlor taxa inhabiting dry areas(1' : 0.3, 1 d.f ., n.s.), and 0.1,6and 0.12 for taxa in areasof moist soil (X2 : 0.9,1 d.f., n.s.). In contrast to the theoreticalprediction, the propor- "cold" tion of taxa inhabiting climates is lower among the groups in Table IV Qn than among squamates in general (0.44), a difference that is statisti- cally significant (12 = 9.3, ! d.f., p < 0.0\ but is probably due to the "random" inclusion of viviparous species in the sample of Shine and Bull (1979).Considerably more genera were found in hot (0.43) rather than cold ! ! ong the 70 venomous q , >,o a 3.:tsi ids and scincids, and s> E - tJl i and agamids (Table Q,) o ;, x.q , it may be significant p.>_ rls adapted to moist, qt H6 c r agamids, varanids, t dry areas. The lizard i.: 'E is origins (Iguanidae, rl r be intermediate on >t-c HF> 5 hc -..F 6 .1 0, 0J -EVEL F E.6 dviparity also may be Et Fr qJ a) in which viviparity oZ :eristics predicted by ol! hether the frequency ruld be expected by E(,) "expected" a^ E ' - q r ra). Such Eo ri 0J 'e given by Shine and i . F.ii ?i ir F.3t rt characteristics tested 96 -c 0J rarous relatives as in >E a gSi; rtion of arboreal taxa sssgss ;s 3gsgsis:iii s, the difference is not t' rresare 0.05 and 0.05 66 0l .a eE. rr taxa inhabiting dry 3" " I q, €i r t fE t ; E t ts nX in areas of moist soil 3 =4&4;J : 'ediction, the propor- is 5 6 33 -4 J ;the groups in Table erence that is statisti- o probably due to the o 6 o :le of Shine and Bull ? c a a b ; i6y, XXXX x/.xxx q 3Y c .+ &-! c -'z < c F $r E = JC ! U U z- gr F.] F -o L +=LP- 5! bo -aL6 o i +- i Bo ^ ! ; q F "! E Siti a " *i**s .1i ! O 3i *!.: Tsii ia["q$iEsisrt $iii ss g$s3s+S c*!SSSF it i s€ sss3!i: €: i Si : s: iiSi E F ! R\RRg6ig sgBhgsgn$s$9$55988G8 o il ,",1 t €AF i ts t!! tr d;E - F-: z t /.xx//lx :,2;. Z 'i - :,': L 2t+t -- llxxx i Y;rr g e&-++ ; i,- i ^, - c ::F;.1 RbAi L i? i ","E _eEY.: i EPE;; Qo;P ^ -** i ! U5soi E.: E- : : xo: c EF.ff: t aP; t ? - o !S o E! 0 ! q 7 E: g^e.6 d k;= I 9 9. F^ Ei.:;; qPqltt ea + i 3 {E i €ia $ FE;i1.? 3Mt: i.Fr e,F :, !3 EE E F":F.H F tS.!r [T b.E::AF"aF . ;*s EEiv2 ssF:€.-' +b:= €F: ffli:;*lut. : > UE .. ! i F{€ €s I sI F lsitgt.=U 6g < ;rsSE;g*EE:. EF€E 5 $€r Er is$g F: t { g $i t =Fi F I r a >:: >> I; FF;; ;; s r s3 €.'s i,!E r 5s;t &i F*Ef s: SSE; g !5S::.^> x e h:EE E 5 ! Es $;: a *q iEi v v F >:: - :; !uS€i'i ;g!13.EE q90-eEE H; S5 :: i ! inr i! atEI ee s r .5 o - i s 9EE 9 & fi63hnfl ?Et?s€ sbs%RFR(RKRRKRSa 3iBEtrE 671 67e THE E\/OLUTIG,N C,F VIVIPAFIITY IN FIE;'TILEA (0.27) areas, probably reflecting the higher numbers of reptilian species in warm climates. Maternal care is almost twice as common (0.7) among the oviparous relatives of live-bearers than among a sample of 23I oviparous squarnate genera (Shine and Bull, 1979), although the difference falls short of statisti- cal significancewhen calculated in this way (X2 : 3.5, 1 d.f., p : 0.06). However, closer analysis confirms the associationbetween maternal care and the evolution of viviparity, The probability of viviparity evolving within a genus of squamates(i.e., a genus with both oviparous and vivipa- rous members) is 65 in 619 or 107o.The probability of viviparity evolving in a genus in which brooding also has been recorded is 12 in 31 (39%). The difference is significant (Xt : 25.7, 7 d.f., p < .001).Also, there are many brooding species in genera that are believed to be the closest oviparous relatives of viviparous genera (e.9., Ophisaurus,Ilnja, Python). No published estimates of expected frequencies are available for the other characteristics listed in Table IV, but the observed frequencies of fossoriality (0.28), large body size and/or venomous capacity (O.Zt), ana nondependence on speed (0.17) seem to be higher than one might expect among a random sample of squamates. However, it is difficult in these cases to distinguish an ecological bias from a taxonomic one. For example, 76 of 23 fossorial genera are scincid lizards, and seven of the 14 genera containing species that do not depend on speed for feeding or for escape from predation are vipers. The high proportion of venomous taxa in Table IV reflects the high frequency of viviparous origins in venomous snakes. 4. ANALYSIS AT THE SPECIES LEVEL The preceding sections have searched for protoadaptations that result in viviparity evolving more often in certain taxa than in others. However, the selective forces actually driving the transition from oviparity to viviparity may best be investigated by looking for consistent differences between closely related oviparous and viviparous species. Such differences are here examined with respect to hot or cold climates, moist or dry soils, invulnera- bility to predation, nondependence on speed, and fossorial, secretive, aquatic, or arboreal habits. The prediction that maternal care favors the evolution of viviparity cannot be tested by a comparison of oviparous and viviparous forms, because maternal egg-guarding is impossible for a live- bearer. The prediction concerning rates of nest predation cannot be tested because of insufficient data. An immediate problem is to decide on the criteria by which a prediction "cold-climate" can be considered supported. With regard to the hy- pothesis, for example, is the prediction confirmed only (1) if all live-bearers occupy cooler habitats than all egg-layers, (2) if live-bearers tend to be in cooler areas, or (3) if the species inhabiting the coldest climate is vivipa- rous? The decision is difficult, because the only direct prediction from E\/ALUATIC,N c,F CAstE HI€ITG,FIIEA 679 of reptilian speciesin theory is that viviparity evolved in the coldest part of the geographic range occupied by the taxon af that time. The pattern may well have been ob- lmong the oviparous scured by subsequent speciation and shifts in distribution, and possibly . oviparous squamate even by reversalsin reproductive mode. e falls short of statisti- Table V presents the results of an analysis of the viviparous origins with "cold-climate" 1.5,1d.f.,p-0.06). respect to the three criteria previously given for the hy- etween maternal care pothesis. Of the 95 squamate taxa in which viviparity has evolved, 35 cases ,f viviparity evolving are accompanied by such poor data, or have apparently been followed by rviparous and vivipa- such extensive speciation, that they do not permit any comparison among viviparity evolving in the climates occupied by live-bearers and by egg-layers. In a further seven s 12 in 37 (39%). The cases,the available data do not reveal any clear differences between ovipa- Also, there are many rous and viviparous forms. Most of the remaining 53 cases show a strong the closest oviparous bias toward viviparous species being found in colder climates. The most , Python). stringent criterion (all live-bearers in colder areas than all egg-layers) is met are available for the in 40 of the 53 taxa (74Vo),whereas the least stringent (the single coldest ierved frequencies of climate is occupied by a live-bearer) holds in 48 cases (977a).The similarity ; capacity (0.21), and of these results using different criteria reflects the high proportion of "ranone might expect groups in Table V that contain only a single viviparous species or sub- it is difficult in these species and for which the different criteria yield identical results. Only in nic one. For example, two casesdo the data negate the prediction that viviparous speciesoccur in ven of the 14 genera colder climates; both cases involve the scincid genus Lygosoma,for which feeding or for escape the live-bearers occupy hotter, drier climates than their congeners. Overall, nomous taxa in Table the bias toward viviparity in cold climates is very strong. Using the most n venomous snakes. stringent criterion (all live-bearers occupy colder climates than all related egg-layers), and treating the seven cases lacking clear climatic differences as definite exceptions (i.e., live-bearers assumed to be in warmer areas), .EVEL the observed bias toward viviparity in cold (39 of 60 cases) still differs rtations that result in significantly from the null hypothesis of 50Vo(y2 : 5.4, I d.f ., p < .05). An others. However, the independent review of viviparous origins by Blackburn (7982,and personal "cold" 'viparity to viviparity communication) also concluded that most origins had occurred in differences between climates. r differences are here Although quantitative analysis is difficult, the data in Table V offer little r dry soils, invulnera- support for the generality of alternative hypotheses on environmental con- fossorial, secretive, ditions that favor the evolution of viviparity. The strong association among ernal care favors the viviparous taxa and unusually cold climates (Table V) in itself casts doubt ;on of oviparous and on the generality of other hypotheses. This result is certainly inconsistent impossible for a live- with the hypothesis that viviparity commonly evolves in hot climates or "unpredictable" tion cannot be tested environments with only a slight bias toward cold. However, the result is consistent with the idea that single clutching is an ry which a prediction important protoadaptation for viviparity; reproductive frequenry is likely "cold-climate" : hy- to be lower in cooler climates (Fitch, 1970). A problem with this hypothesis y (1) if cll live-bearers is the lack of data on reproductive frequencies for almost any of the taxa bearers tend to be in listed in Table V; available information would suggest that double clutch- est climate is vivipa- ing is likely to be rare, even in oviparous forms (e.g., Fitch, 1970). This rect prediction from simply may reflect inadequacies of the data (Tinkle and Gibbons, 1977). 0 ql cc 2i OJ o 4 L E g! c q I g i ?.E .+ -_i 6 o 2 a E a 'i 0.) 'ia c oo q) qJ q .9 U .a c .E -O >.9 i ,__ ! H6 a. t t I 1 X... X Y d x X I X X /. d X X I I /. l t bl Q 41 = L rl E x X X X X \ Z I x X I I /. 7. /. I X x ^. d. I X I I 6 >. c k xxtxxxf xr" 22ttxxlr.- " tttt 1i q 3e i. ;a i* ?*$ f : ta 6 o @ i ; ti,$:,f-ri;"Fs*it= i : f,g*EtE+Ei;;E:iFF::Xii! -F gfRF i€: $F€iSiiggs Fl: !!5 & t6 vL 32€3: : i r';;; ; : :s i -+ € ; _o!; H P g *!P E ! i I : ci E = E EC i c, F. F E -!i e k { i U U'i.q, F o H G ! Irr o Fd6$6AN€60ido+rsN@oo 674 ? :* o.; oi +; id a E+ CF €* ..E it ar !Z .'9 /. x. i= ;. "x)*.. ttxt -. ?ax x x x t t t z x r t x /.. t t t t t x x |).x Qx I fa 'x 9a E; G^ rz_ :o..i!3 v5 k;Y- €!i ! ! j ai i+-tr;!irroE J^\C; p Y i r $ gtE F i F l:S ^ * .- ll o : $-iPi.q I : iil:=39*te ,i€iT9:.i:iri F:r: il =:i o;N .- E : < S < 9 . .::Gy{ rotiF i =Ei€. FFETS : i 6 E: a+ F i g E - ij n s Slruisr*!l"t{+= *i: gIi+ €c:;tis i :; :_ 0a s ! Fe $E"tiFi.EFsB ?Fs* 3'i t*s'* ! ! ": FY*s< E: ; Ei s E*r Fi Fi E EE F x { i E = $ i.;;-d*5d-i { j -j *: -- :f,:'- t t (: 6i.ii.l,3.r 6 6 v ( u c .. l: - 0d H" ! 6.9q 6> b i- q ooN€60ndos g F FNC idddNd R 8 \ K 3.P; | ;7 5 P,> 5 p S ; S 6 S 5 € 5 + 3; F r THE EVclLUTIclN clF IN FIEI'TILEA Another objection to the clutch-frequency h1'pothesis is the observation that many of the viviparous forms in iable"V occur in severelycold regions; the hypothesis would predict the evolution of viviparity in regions slightlv too cool for double clutching, and these environments probably would be relatively mild and temperate rather than severely cold ones. Nonetheless, the data do not falsify the clutch-frequency hypothesis. The number of taxa fitting the predictions that viviparous forms pre- dominate in areasof wetter soll (7, or 13Vc\or drier soil (5, or 9%) are much iower than the number of taxa fitting the cold-climate trend. The casesin which live-bearers are found in more mesic areas all involve taxa for which the live-bearers are also in cooler areas and may reflect a correlation be- tween temperature and moisture rather than an independent effect of moisture (as suggestedby Guillette et al., 1980). There are no substantial data in support of the predictions that vivipa- rous species are likely to be more formidable (large or venomous), more slow-moving, or more fossorial or secretive (Table V). One of the vivipa- rous cordyline genera Ch(unaesaurais elongate but not truly fossorial; the oviparous cordylines have normal-sized limbs. The opposite trend is seen "baudini in Australian Leiolopisma,in which the oviparous group// are more elongate and fossorial than their viviparous relatives. The prediction that viviparous species rely more on crypsis than on speed to escape predation may hold for the slow-moving viviparous Psnmmophylaxaariabilis; the oviparous P, rhombeatusis more active (FitzSimons, 7962). However, no data are available on motor patterns of the oviparous subspecies of P. aariabilis. The association of aquatic habits and viviparity is strongly confirmed for Elapherufodorsata, but the species inhabits small streams rather than large bodies of water; hence, females would not be faced with a long migration to land for eggJaying. Thus, E. rut'odorsatais not aquatic in the manner envisaged in Neill's (196a) hypothesis; it is probably more in accord with "moister the soil" prediction (Sowerby, 1930). A greater tendency toward arboreality in viviparous forms is evident within Lobulia, the Sceloporus " " grammicus grovp ," and perhap s the Sphenomorphusfasciatus group" (of 5 oviparous and 3 viviparous members, all are fossorial except the arboreal viviparous S. tenuis and perhaps S. murrayi). The prediction that viviparous species will be more heliothermic than their oviparous relatives is not tested in Table V, because the necessary detailed data are lacking for most groups. However, data on six groups suggest that this prediction may often be verified: viviparous species are more diurnal and heliothermic than their oviparous relatives (Coronella, Steward, 1977; Hemachatus,FitzSimons, 7962;Anguis, Schmidt and Inger, 1957; Naultinus, Y{erner and Whitaker, 7978; Pseudechis,Cogger, 1.975; Shine, 7979; Leiolopismaentrecasteauxii, Cogger, 1975; Shine, 1980a).In the last four taxa listed, the data indicate that the gravid viviparous females are most clearly heliothermic; presumably, basking is prolonged in order to IPAFIITY IN FIEPTILEA coNcLusroNsl 677 s is the observation accelerateembryonic development. Interpretation of this thermoregulatory everelycold regions; differenceamong oviparousand viviparous relativesis difficult; three pos- ty in regionsslightly sibilitiesare that (1) heliothermy favored the evolution of viviparitv, that ; probably would be (2) heliothermy evolved as an adaptation to viviparity, and that (3) I ones.Nonetheless, heliothermyis an adaptationto the coolerclimates occupied by the vivrpa- S. rous membersof all these taxa. Similarly,a tendencyfor arborealityin viparous forms pre- viviparous sceloporinelizards mav have evolved for thermoregulatory (5, or 97o)are much enhancement(Guillette et al., 1980).Undoubtedly, the oviparousand trend. The casesin viviparous relativeslisted in Table V exhibit many ecologicaldifferences volve taxa for which not discernedin the present analysis;the documentationof such differ- ect a correlation be- encesis an important task, but it is a task for biologistswith first-hand lependent effect of knowledge of the relevant group. dictions that vivipa- )r venomous), more vlll. coNcLUsloNs . One of the vivipa- t truly fossorial; the A summary of hypotheses,predictions, and tests of predictionson reP- "cold-climate" rposite trend is seen tilian viviparity (TableVI) revealsa complex picture. The "recently 'ini group" are more hypothesis has strong empirical support, in that evoived" The prediction that viviparous speciesoccur more consistentlyin coolerclimates than do their to escapepredation oviparous relatives. Weaker evidence for the role of coid climates also rhylax aarinbilis; the comes from the present-daydistribution of viviparous species(e.g., t962). However/ no Weekes, 1935;Sergeev, 1940) and from a tendency for recently evolved "cold" us subspecies of P. viviparous speciesto inhabit climatesthat are subjectivelyjudged as (Shine and Bull, 7979;Blackburn, 1982).Although the hypothesis that ongly confirmed for viviparity evolvesin cold areasis supported,the exactnature of the selec- "cold-cli- ns rather than large tive force remains unclear. Severalalternative versions of the ith a long migration mate" hypothesishave been proposed (Shineand Bull, 7979),and rePro- ratic in the manner ductive frequencyalso correlateswith this factor. nore in accord with Cold climates seem to have been the most important single selective er tendency toward agent favoring viviparity, but it is equally obvious that not all casesare bulin, the Sceloporus explicableby this single force. For example, the origins of viviparity in nciatus group" (of 5 Lygosoma,Cerastes, Sphenomorphus fragilis, and S. nigricaudusare difficult to "cold-climate" except the arboreal reconcilewith the hypothesis.Other hypotheseson envi- ronmental influenceseither have not been tested (e.g., nest predation)or e heliothermic than areunlikely to have generalimportance (hot climates,unpredictability, soil :ause the necessary moisture). data on six groups Among factorsthat protoadapta taxonomicgroup to evolve viviparity, 'iparous species are the most important may be venom in snakesand maternalbrooding behav- relatives (Coronella, ior. Minor biasesmay be introduced by other factors.There is no evidence Schmidt and Inger, to support the hypotheses that arboreality or aquatic habits are major chis, Cogger, 1975; protoadaptationsfor viviparity (TableVI). The significanceof physiological hine, 1980a).In the constraintsremains untested. 'iparous females are The broad correlationalanalysis of the Present study is designed to rlonged in order to identify maior and consistentbiases in the evolution of viviparity' The o a a. x -o 6 L L (, U H a) frl o G il- i o. sS!?.s 6 N N s-: = fr.l a.-"! .; .d .5 t\> o' 0, .- t L --,4 r.(\ € a' I d := + 3 ^ : : :tavJ-,JJJ q d€r z z . z z z 6 -.0 >F ble ba Ftr trl '- I qa "o 6 o L UU 'F -q !'o >3 UU q al 6 0) a b' €E E E H ,r, E Fi?:.ooJE.s xl9-r-o s cr E F Ex E E tiO.^Oq1OUOOu h x :ijege€.! .t ^FL boi^ bo ^^ ho.i bo'! L 6 q) bO(! boi bO.= bO: bOX bOd or qJ { rrl - lrl I 111 - Irl L trl ;: rrl * fr e E- r SAdNcq-+'ro\oN x .X oo -t is P- - a) A> a =;utn'a /d. - a .i +.; 3 3 o@ oo.^ oo,^ O =Ft ,Fd l)-^r, l. l; i!= - jyi: i .Yi; ooCi66E P>. a 29aL- e' ?T.T rd r3 '/F i'o: F i": :19 t.-^ - I ;,- oc C't l= o haiz h;,( ZZ ZZ 6! Zc. Z a J - o. o G 'tr 'o - 6 o' aq a c. ^^ o z zzz ).c e.). Z ZZ 6 I U tr P,.I !y :-F ., I p. 6. d > 0r- e = p '!'7^rlaqro-E:: 5 Eq - 6JJ .o -/d _o F . o o .- =y + E E bo; o @ t; H .9 t p.E! b= ?s . E X bo z1 - ) > - .E 4.. 9r 9 '=tc€ o b €':bEe 35 'tq -c o !g P>, \ c I F. : k E I o L .H R - ^ 3cs:t u"d,,Ee i € A f;: 'F* SEi#.8 ;':f '!Ei: d r tr€ 6 9.c/.! 6=; ! * U FF ; €*€i+ r€; €€ AA qq!!L.n : r:f Fei€B"e 83.'EEE n o 0) ! o/ 9.,1 o- t ;PyEtc€g:E 9U t sgH ti:€ip*tO !k-J p. o O., 9"a -Y'= Li EE$; *il € +*E t ; r€ ;;tE i'p @ 8oP 8o; E; bDi; bo(E 0') 0i g r'r ): rrl t ri e t $E E: e g gE J* g E s g E F E F* F€ f;:$-- == s *s s e o ,FE 6 THE EVOLUTIG'N c,F VIVIPAFIITY IN FIEPTILEEI "cold-climate" ability of the hypothesis to explain most cases makes it unlikely that the alternative predictions have generality. However, correla- tional analysesare incapableof detecting factorsthat are important only in a few cases.Hence, none of the factorshypothesized to be important in the evolution of viviparity (Table VI) can be dismissed; any one of them may have been crucial in a particular case.Nonetheless, only the few hypoth- eses supported by data (Table VI) are likely to have general importance. They are capable of explaining strong biases in the taxa in which viviparitv has evolved. For example, the probability of viviparity evolving within any given genus of snakes averages 6.7Va for all snakes (18 within-genus ori- gins in 294 genera with oviparous species), but this rises to 75% among genera with venomous speciesand to 35% among genera in which mater- nal care has been reported. Where both factors combine (venomous and brooding), the probability rises to 66% (six origins among nine genera). Within any genus in which viviparity has arisen, live-bearers usually oc- cupy colder environments than egg-layers (Table Y; 87% of cases if all uncertain casesare taken as exceptions; 96Voof all casesfor which sufficient data are available to make a clear judgment). As these results are based on correlations, causation is inferred but cannot be demonstrated. Even those casesthat accordwith prediction may do so for a reason other than the one suggested. For example, a particular viviparous taxon may occupy a cold climate simply by chance or because of a rapid radiation into that environment after viviparity evolved. Also, re- versals of reproductive mode (viviparity to oviparity) or errors in phy- logenetic hypotheses may confound specific cases in the above analyses. Hence, it remains useful to continue tests by considering groups for which no breeding data are yet available. However, any evidence based on corre- lations is open to many alternative interpretations. Nonetheless, the data do strongly support the predictions of several hypotheses and falsify others; they provide the strongest evidence to date on the origin of vi- viparity. The logical next step is to search for additional cases and to undertake detailed ecological and physiological studies on closely related oviparous and viviparous squamates. Such studies can overcome many of the prob- lems associated with broad correlational analyses; for example, by directly measuring variables such as temperatures of nests and gravid reptiles, "costs" sources of egg mortality, and of egg retention (e.g., Guillette, 1987, 7982a,1982b;Shine, 1980a,1983b).This information should serve to illumi- nate the selective forces involved in the transition from oviparity to vivipar- ity and to provide tests of many of the hypotheses and conclusions of the present studv. ACKNclWLEElGMENTS I thank Carl Gans and the late D. W. Tinkle for their encouragement.The analyseson scincidlizards-the most important single family in terms of 6A,l 4PAFIITY IN FIEPTILES FIEFEFIENCES the assis- nost cases makes it the evolution of viviparity-would have been impossible without 1980.I am grate- y. However, correla- tance of Allen E. Greer (Australian Museum) in 7979 and data on rre important only in ful also to the many workers who provided me n'ith unpublished S' A' Min- r be important in the specific groups, especiallyW. R. Branch, J. M. Cei, S. Moodl', Valuable ny one of them may ton, Jr.,-M. McCoy, F. Parker, D' A. Rossman, and R. Sadlier. W' nly the few hypoth- comments on the manuscript were provided by C. Gans, F. Billett, A' E' general importance. Branch, L. J. Guiliette, Jr., J. Cadle, R. B. Huey, R. Larnbeck, Bibiio- a in which viviparitv Dunham, G. R. Zug, D. G. Blackburn, G. Grigg, and G. Mengden. and D' evolving within any graphic assistancewas kindly provided by G. Mengden, R. Sadlier, assistance 18 within-genus ori- Kent. Finally, I thank Terri, James, and Cooper shine for their rises to 15% among and encouragement, and Sylvia Warren for typing the manuscript. rera in which mater- >ine (venomous and mong nine genera). FIEFEFIENCESi :-bearers usually oc- 87Voof cases if all Amaral, A. do (7977). serpentesdoBrazil. Editora da Universidade de sio Paulo, Brazil - ; for which sufficient Andren, C. and Nilson, C. (7976a). Hasselsnoken (Coronellaaustriaca) en utrotningshotad ormart! Fauna Flora 2, 61-76. tion is inferred but Andren, C. and Nilson , G. (1976b). observations on the herpetofauna of Turkey in 1968- with prediction may 1975.Brit. I. Herpet.5, 575-584. rxample,a particular Andrews, R. M. (1982). Spatial variation in egg mortality of the lizard Anolis lintifrons. Her- chanceor becauseof petologica 38, 165-170. M., and Roy, R. (1954). La r6s6n'e naturelle integrale du Mont y evolved. Also, re- Angel, F., Guib€, J., Lamotte, Nimba. Mim. lnst. frang. d'Afrique noire 40,381-402. l) or errors in phy- Anonymous. (1982). Boas breed (Casareadussumieri) Wildlife (December), 460' the above analyses. Arakawa, H. (1970). Climatesof Northernand EasternAsla. Vol. 8 in Worltl Suraeyof ClimatologV. for which ng groups Elsevier, Amsterdam. on corre- ence based Arnold, E. N. (1973). Reiationships of the Palaearctic lizards assigned to the genera Lacerfa, (nat, onetheless, the data Algyroidesand Psammorlronus(Reptilia, Lacertidae). Bull. Brit. Mus. Hist.) Zool.25, rotheses and falsify 291-366. and on the origin of vi- Arnold, E. N. and Burton, J. A. (1978). AField Guitle to the Reptilesand Amphibiansof Britain Europe. Colllns, London. (Reptilia: es and to undertake Arnold, E. N. and Leviton, A. E. (1977). A revision of the lizard genus sclncus (nat. Zool. 3l' 787-248' ly related oviparous Scincidae). BuIl. Brit. Mus. Hist.) of Eumecesdicei from the Sierra Madre of northeastern e many of the prob- Axtell, R. W. (1960). A new subspecies Mexico. Copeia 1960' 79-26. example, by directly Badhuri, L. (1943). Notes on the viviparity of the common Indian skink Mabuya carinata md gravid reptiles, J. (Schneider). l. Bombay Nat. Hist' Soc.M, 130-134. e.g., Guillette,7987, Copeia 1966' 885- Bailey, I. R. (1966). A redescription of the snake Calamodontophis paucidens. ,ould serve to illumi- 886. lntetnacional s6bre oviparity to vivipar- Baitey, J. R. (1981). Notes on the genus Thamnodynastes.Primeiro Symposio d conclusions of the ierpentes.Instituto Butantan, Sio Paulo, Brazil, 16-18 Nov 1981 (Abstract) Bartmann, W. and Minuth, E. (1979). Ein iebendgebdhrender Gecko, Rhacodactylustrachyrhyn- chus Bocage 1873, aus Neukaledonien (Reptilia: Sauria: Gekkonidae). Salamandra15,53- 60. Bauchot,R. (1965).La placentationchez les reptiles. Ann. BioL4,547-575' )ncouragement. The Bauwens, D. and Thoen, C. (1981). Escape tactics and vulnerability to predation associated Lacertaaiuipara. anim. EcoL 50,733-743' le family in terms of with reproduction in the lizard I' €iAA THE EVGILUTIG,N ctF VIVIFAFTITY IN FIEPTILEEI Bezy, y. R. L., Gorman, G. C., Kim, J., and Wright, J. W. (1977). Chromosomal and genetic divergence in the fossoriallizards of the farnily Anniellidae. syst.20o1.26,57 77. Blackburn, D. G. (1981).An evolutionary analysisof vertebrateviviparitv. Anrcr.Zod. 2-1,936 (Abstract). Blackburn, D. G. (1982).E'olutionary origins of ir.iparity in the Reptilia. I. sauria. AttlthiLtio- ReTtilin 3, i85-205. Blackburn, D, G. (1984) Evolutionarv origins of ','iviparitv in the Reptilia. IL Serpentes, Anrphisbaenia, and Ichthyos auria. AnTlilIti n_ Rep t ilia, in press. Blanchard,F. N (1933).Eggs and young of tht'smooth green snake, Liapdtisitlririli-s (l larlan). Pap. Michigan Acsd. Sci. Arts. Lett. 17,493-508. Blair, W. F. (1960)' The Rusty Lizard. A Population Slldy. University of Texas Press, Austin. Blanc, c. P. and Blanc, F. (1967). observations biologiques (adultes, oeufs, jeunes) sur quel- ques sauriens du Mont Bit1,.Annl. Lhtra. Mddagascar,Ser. Sci.S,67_74. Blanc, F. (1974). Table de developpement de ClrotnaeleolLtteralis Cray,, 1g31. Ann. Etiltn1ol. M'trphol.7,99 -115. Bogert' C. M. (1940). Herpetological results of the Vernay Angola Expeclition $,ith notes on African repfiles in other corlections. Buil. Amer. Mus. Nat.- Hist rbs, t-zsg. Bons, J. and Bons, N. (1960). Notes sur la r6production et le ddv6loppement de Chanaeleo chamaeleonL.Bull. Soc. Sci. nat. phys. Maroc. 40,323_935. Bons, and saint-Girons, J. H. (1963). Ecologie et rycle sexuel des amphisbaeniens du Maroc. Bull. Soc. Sci. nat. phys. Maroc. 43,717-tSB. Branch, w. R. (1979). The venomous snakes of southern Africa. part 2. Elapidae and Hv- drophidae. The Smke 71,199-225. Branch, W' R. (1982). Venomous snakes of southern Africa. 3. Concluding part: Colubri Brattstrom, B. H. (1964). Evolution of the pit vipers. Trans. san Diego soc. nat. Hist.73,7B5- 268. Brattstrom, B. H. (1965). Body temperatures of reptiles. Am. Mirll. Nat. 73, g76-422. Broadley, D. G. (1974)- Reproduction in the genus platysaurus (sauria: Cordylidae). Her petologica 30, 379 -380. Broadley, D. C. (1977). A revision of the African snakes of the genus PsammophylaxFitzinger (Colubridae). Occ. Pap. Nat. Mus. ModesiaB 6, 1_49. Broadley, D. c. (1979). Predation on reptile eggs by African snakes of the genus prosymna. Herpetologica 35, 338-341. Broadley, D. (1983). G. Fitzsimons' snakesof southern Africa. Delta, Johannesburg. Broadley, D. G. and Cock, E. V. (1975). Snakesof Rhodesra.Longman, Salisbury Broadley, D. G., Gans, C., and Visser, J. (1976). studies on amphisbaenians (Amphisbaenia, Reptilia). 6. The genera Monopeltis and Dalophia in southern Afnca. Bult. Amer. Mus. nat. Hist. 157,313-485. Brown, W. C. and Alcala, A. C. (1980). Philippine lizards of the family Scincidae. Silliman Unia. Nat. Sci.Monogr. 2. Bryson, R. A. and Hare, F. (1974). Climates of North America. Vol. 11 in world suraw of Climatology. Elsevier, Amsterdam. Bull, J. J. (1980). Sex determinarion in reptiles. e. Rro. Biol. SS,3_21. Burger, I. (J977). Determinants of hatching success in diamondback terrapin, Malaclemys terrapin. Am. MidL Nat. 97, 444-464. "Genera Burger, w. L. (1971). of pitvipers (serpentes: Crotalidae)." ph.D. thesis, university of Kansas, Lawrence. BE|s /IIIAFIITY IN FIEIITILEET FIEFEFIENCE€I skinks, Saipftosequalis and lhromosomal and genetic Bustard, H. R. (1961). Reproduction in the Australian rain forest :yst. Zool, 26, 57-71. Sphenomorphustryoni. Copeia1964, 775-716. snakes.I. South Ameri- paity . Amer. Zool. 21, 936 Cadle, J. E. (1984).Molecular systematicsof neotropicalxenodontine can xenodontines. Herpetologica40, 8-20. eviclenceand the relation- ptilia. I. Saurra. Anrythibia- Cadle, J. E. and Gorman, G. C. (1981).Alburnin immunological ships of sea snakes. I. Herpetol. 15,329-334. of Eumecescallicephalus re Reptilia. II. Serpentes, Campbell, H. and Simmons, R. S. (1961).Notes on the eggs and young Bocourt. Herpetolttgica 17, 212-213. of South Viettnm. Smithsonian lnst., , Liopcltislerrralis (IIarlan). Campden-Main, S. M. (1970). A FieldGuide to the Snakes Washington, D.C. Handbooks, Longmans, 'of Texas Press, Austin. cansdale, G. s. (1961). West African sirakes.west African Nature London. s, oeufs, jeunes) sur quel- (Reptilia, Iguanidae) from the arid envi- J, O/ -/+. Cei, J. M. (1974). Two nerv species of Ctenoblelthnris ronments of the central Argentina (Mendoza Province). l. ITerpt:to\8,71-75' iray, 1831.Arm. Lnfuryol. Iguanidae) wrth Cei, J. M. (1g7g). A reassessmentof the genus C.tenoblephnris(Reptilia, Sauria, a description of a new subspecies of Liolaemusmultimtcu[atus from western Argentina. /. Expedition with notes on 13,297-302. t.709,1-238. Herpetol. M. (1980). New endemic iguanid lizards from the Famatina Mountains of Western 6loppement de Chamaeleo Cei, J. Argentina. l. Herpetol. 14, 57-&. and scolaro, A. (1982). A new species of the Patagonian genus vilcttnia, with nphisbaeniens du Maroc. Cei, J. M. J. remarks on its morphology, ecology and distribution. J Herpetol 16, 354-363' T. H. and Fang, P. W. (1931). A stud1. of the ophidians and chelonians of Nanking. Part 2. Elapidae and Hy- chang, Contr. Biol. Lab. Soc.ChinaT,249-288. N., Watts, A. and Shine, R. (1983).Captive reproduction in an Australian elapid rcluding part: Colubridae. Charles, snake, Pseudechiscolletti. Herp Rezt.14, 76-78. N., whitaker, P. and shine, R. (1979). oviparity and captive breeding in the spotted 'goSoc. nat. Hist.1.3,185- Charles, blacksnake, Pseudechisguttatus(Serpentes, Elapidae). Aust Zool 20,367-364' (1981). The new saw-scaled viper, Echls multisquamatussP. nov. from south- Nat.73,376-422. cherlin, v. A. western and middle Asia- Proc.Zool. htstitut.101,92 95 [In Russian]' Sauria: Cordylidae). Her- Chernov, S. A. (1957). Systematic position of the poisonous snake Ancistrotlon rlutdostonta (Boie) (Serpentes, Crotalidae) in connection with its craniology. Zool. Zhur.36,790-792. s Psammophylar Fitzinger Cogger, H. C. (1975). Reptitesand Amphibians of Australia. A. H. & A W' Reed, Sydney' C. Karyotypes and systematics of the lizards in the aariahilis, jalapae, and ,s of the genus Prosymna. Cole, J. Og78). scalarisspecies groups of the genus sceloporus.Amer. Mus. Noxitates(2653), l-13. Cooper, S. (1965). Notes on fertilization, the incubation period and hybridization in Lacerta Johannesburg. J. Br. l. Herpet. 3,218*220. rn, Salisbury. criley, B. B. (1968). The cranial osteology of gerrhonotiform lizards. Am. Mittl. Nal. 80, 199- baenians (Amphisbaenia, 2\9. ica. BuIL Amer. Mus. nat. da Cunha, o. R. and do Nascimento,F. P. (1981).Ophidia of Amazonia,72. Bol.Mus. Para. EmilioGoeldi Noua Ser. Zool. 109,7-20. :ami-lyScincidae. Silliman Daan, s. (1968).Agamids. In Grzimek'sAnimal Life Encyclopedia. Yol. 6, Reptiles(8. Grzimek, ed.). Van Nostrand-ReinholdCo., New York, pp.205-227. rl. 11 in lNorld Suruey of Darevsky, I. S. (1966).Ecoiogy of rock-viper (Viper xanthinaraddei Boettger) in the natural surroundingsof Armenia. Memor.lnst. Butantan(Simp. Internac ) 33, 81-83' )1 Davis, W. B. and Smith, H. M. (1953).Lizards and turtles of the Mexicanstate of Morelos. 'ack terrapin, Malaclemys Herpetologica9, 100-108. ' Deoras,P. J. $971). Snakesof indra.National Book Trust, New Delhi' Ph.D. thesis, University de Rooij, N. (1917).The Reptilesof thelndo-Australian Archipelago. Vol. 2, ophidia.E. J. Brill, Leiden. 7 I THE EVCILUTIGIN clF VIVIPAFTITY IN FIEPTILEEI Diesener,G. (1974).Beitrag zur Biologieder Bergotter.Aqua. Terra-Z 27,2q9_251. Ditmars,R. L. (1942).The Reptiles of NorthAmerica. Doubleday Doran, New york. Dixon, J. R. (1969).Taxonomic review of the Mexicanskinks of the llneces lrrelrro-strlsgroup. LosAngeles Cottnty Mus. ContriLt.Sci.168, 1-30. Dmi'el,R. (1970).Growth and metabolism in snakeernbryos. J. Enrbryol. axp, MorTfiol. 23,76L- 772. Domergue,c. A. (1959).ovoviparit6 et o'iparit6 chez deux Viperidaedu gdnre A.spis(L.). Bull.Soc, Sci. nat. Tunis.9,73-74. Donoso-Barros,R. (1966).Reptiles de Cftile. Edicion de la Universidadde Chile, Sarrtiago. Donoso-Barros,R. and Cei,J. M. (1971).New lizarcisfrom the r.olcanicPatagonian plateau of Argentina.l. Herpetol.5,89-95. Dowling, H. G. (1974).The nearctic snake fauna. ybk. Herpetol l, 790-202. Duellman, W. E. (1961).The amphibiansand reptiles of Michoacan,Mexic6. Li1ii,. Kcrr-sn-s Publ.lvlus. Nal. Hisf. 15, 1-148. Duellman,W. E. (1965),A biogeographicaccount of the herpetofaunaof Michoacan,N{exico. Uniu. KansasPubl. Mus. Nat. Hist.1,5,627-709. Duellman,W . E. (1979).The herpetofaunaof the Andes: its origins,evolution, and dispersal. Uniu. KansasMonogr. 7, 371-460. Dufaure, J. P. and Hubert, J. (1961).Table de d6v6loppementdu l6zard vivipare:Lacerta (Zootoca)uittipara Jacquin. Arch. Anat. Microsc.Morphol. exp,50,309_32g. Duff-Mackav,A. (1965).Notes on the biology'of the carpetviper, Echlscarinntus pyrdmidunt (Geoffroy),inthenorthernfrontierprovinceofKenva. l.eastAfr,nat.Hist.Soc.natl.ltlus. 25,28-40. Duguy, R. (1961).Le rycle annuald'activit6 de CoronellaaustriscaLaur. Terre et Vit 1961.107- 436. Dr-Tl P. c. (1964).species-pair allopatry in the genera Ranaand phrynosoma.Ecology 4s, 178-787. Dunson,w. A. (1975).Adaptations of seasnakes. rn TheBiology of seasnakes (w. A. Dunson, ed.). University Park Press,Baltimore, Maryland, pp. 3_19. Erasmus,H. and Branch, w. R. (1983).Egg retention in the south African blind snake Typhlopsbibroni. I . Herpetol.t7, 97-99. Etheridge,R. (1966).The systematicrelationships of West Indian and SouthAmerican lizards referred to the iguanid genusLeiocephalus. Copeia 1966,79_gl. Ewert, M. A. (1979).The embryo and its egg: developmentand natural history. rn Turtles, Perspectiuesand Research(M. Harlessand H. Morlock, eds.).John wiley & sons, New York, pp. 333-413. Ewert, M. A.,(1984). Reproduction and development in turtles. ln Biologyof the Reptitia(C. Gans,P. F. A. Maderson,and F. Biilett,eds.). Vol. 14,John Wiiey & Sons,'Inc.,New york. Ferguson'G. W. and Brockman,T. (1980).Geographic differences of growth rateof Sceloporus lizards (Sauria:Iguanidae). Copeia1980, 259-264. Ferguson,M. w. l. (1984).Development of crocodilians.rnBiology of the Reptilia(C. Gans, P. F. A. Maderson,and F. Billett,eds.). Vol. 14, John Wiley & Sons,Inc., Nerv york. Fisher,R. A. (1958).The Genetical Theory of Naturalselection.2ndEd,., revised. l)over press Publ.,Inc., New York. Fitch, H. s. (1954). Life history and ecology of the fivelined skink, Eumecest'asciatus. publ. Mus. nat. Hist. Unia. Kansas8,l-156. Fitch' H. S. (1970).Reproductive cycles in lizards and snakes.Unlzr. I(ansas Mus. nat. Hist., Misc. Publ. 52, 1-247. IPAFIITY IN trEtrTILEA FIEFEFIENCE€I -z 27, 249-251. Fitzgerald, M. and Pollitt, c. (i981). Oviparity and captive breeding in the mulga or king australis(Serpentes; Elapidae) Aust Herpetol l',57-60' 'an, New York. brown snake Pseudechis l. (1943).The lizards of South Africa. Transi'aalMus Mem. (1), 1 528. Eumeces breuirost r is gr oup. FitzSimons, V. F. M. FitzSimons, V. F. M. (7962). Snakesof SouthernA/rirz. Macdonald, London. yol. exTt. Morphol. 23, 761- Fuhn, I. (1968). Scincidae.ln Crzimek'sAnimal Life Encycltt1tedia.Vol. 6, Rcplift:s(B. crzimek, ed.). Van Nostrand-Reinhold, New York, pp. 242-264. 'idae du gdnre ,4spi-s(L.). Fuhn, I. (1969). Revision and redefinition of the genus AbleTtharirsLichtenstern, 1823 (Reptilia, Scincidae).Rea. Roum. Biol., Ser.Zool.14,23-47. ad de Chilc, Santiago. Fukada, Il. (1956). Biological studies on the snakes. III. Observations on hatching of Elnpht (Boie), and Natrix uibakari uibakari (Boie). BuIL Kyoto LnicPatagonian plateau of climacophora(Boie), E. conspicillata GakugeiUnia., Ser. B. 9,21'-29. on the snakes. IX. Breeding habits of Agkistrodon halys 190-202. Fukada, H. (1962). Biological studies blomhoffii(Boie). Brl1. Kyoto GakugeiUnitt . B 20,72-17 an, Mexico. Llnit,. Kansas Fukada, H. (1965). Breeding habits of some Japanese reptiles (critical reviewl. Bull. Kyoto GakugeiUniu. B 27, 65-82. na of Michoacan, Mexico. Gans, C. (1967). Rhineuraflorida. CatalogueAmer' Amphib. Rept 43,l-1. (7974\. An Approach to VertebrateBiologv.I.P. Lippincott, Philadelphia evolution, and dispersal, Gans, C. Biomechanics: (1980, University of Michigan Press)' (1978).The characteristics and affinities of the Amphisbaenia. Trans. zool. Soc.Lond. u l6zard vivipare'. Lacerta Gans, C. , 309-328. 34,347-416. Aspidura (serpentes, Reptilia, Colu- Echiscarinatus pyranirlunr Gans, C. and Fetcho, J. R. (1982). The Sri Lankan genus r. nat. Hist. Soc. natl. Mus . bridae). Ann. CarnegieMus. 5'1,271-316. Gentilli, I. 0971). Climatesof Australia and New Zealand.Vol. 13 in World Suraeyof Climatology. ur. Terreet Vie 1961, 407- Elsevier, Amsterdam. Gloyd, H. K. (1977). Descriptions of new taxa of crotalid snakes from China and Ceylon (Sri ) Phrynosoma.Ecology 45, Lanka). Proc. biol. Soc.Wash.90, 1002-1015. Gloyd, H. K. (1979).A new generic name for the hundred-pace viper. Proc. biol. soc.wash. 91, eaSnakes (W. A. Dunson, 963-9&. Greene, H. W. (1970). Modes of reproduction in lizards and snakes of the Gomez Farias ruth African blind snake region, Tamaulipas, Mexico. Copeia1970, 565-568. Greer, A. E. (1966). Viviparity and oviparity in the snake genetaConopsis, Toluca, Gyalopion, Helicops.Copeia 1966, 377-373' d SouthAmerican lizards and Ficimia, with comments on Tomodon and Greer, A. E. (7967a). Notes on the mode of reproduction of anguid lizards. Herpetologica23, atural history. In Turtles, 94-99. ohn Wiley & Sons, New Greer, A. E. (1967b). The generic relationships of the African scincid genus Eumecia. Breuiora (276), 7-9. Biologyof the Reptilia(C. Greer, A. E. (1967c). A new generic arrangement for some Australian scincid lizards. Breaiora g & Sons,Inc., New York. (26n,1,-79. I growth rate of Sceloporus Greer, A. E. (1968).Mode of reproduction in the squamatefaunas of three altitudinally correlatedlife zones in east Africa. Herpetologica24, 229-232. Bull. Mus. comp.Zool' 1 of theReptilia (C. Cans, Greer,A. E. (1970).A new subfamilialclassification of scincidbzard,s. Sons,Inc., New York. Haru.139,151-183. ld., revised.Dover Press Greer, A. E. (1974).The generic relationships of the scincid lizard genus Leiolopismaand lts relatives.Aust. l. Zool.Suppl.3l, l-67. relationshipsof the scincid lizard genus <, Eumecesfasciatus. PubL Greer, A. E. (1977).The systematicsand evolutionary Lygosoma.l. nat. Hist.11, 515-540. t. KnnsasMus. nat. Hist., Greer,A. E. (7979a).Eremiascincus, a new genericname for someAustralian sand swimming skinks (Lacertilia:Scincidae). Rec. Aust. Mus. 32,327-338. 6ACi THE EVOLUTIGIN clF VIVIPAFIITY IN FIEPTILEEI Greer, A. E. (1979b).A phylogenetic subdivision of Australian skinks. Rec.Aust. Mus 32,339- 371. Greer, A. E. (1982). A nerv species of l-eiobpivna (Lacertilia: Scincidae) from Western Austra- Iia, rvith notes on the biologv ancl relationships of other Australian species.Ilcc. Alsf. Mtrs. 34, 549-573. Greer, A. E. (1983). The Australian scincid lizard genus CalrTplofii-;resurrection of the name, description of four new species, and discussion of the relationships. Ilcc. Ausf. A{tts. 35, 29-59. Greer, A. E. and Parker, F. (1967).A new scincid lizard from the northern Solomon Islands. Breaiora(275), 1-20. Greer, G. E. and Parker, F. (1968). A new species olTribolonotus (Lacertilia: Scincidae) from Bougainville and Buka, Solomon Islands, with comments on the biology of the genus. Breuiora(297\,1-16. Greer, A. E. and Parker, F. (1974). The lasclatrrs species group oI Sphenonorphrrs (Lacertilia: Scincidae): notes on eight prer.iouslv described species and descrtptions of three nen' species. Proc. Papua New Cuinea Scl. Soc. 25, 31-61. Greer, A. E. and Parker, F. (1979). On the identity of the New Guinea scincid lizard Lqgosorrro fragile Macleay 1877, with notes on its natural history. I. Herpetol. 13,221-225. Creer, A. E. and Raizes, G. (1969).Green blood pigment in lizards. Science,N, Y. 156, 392-393. Griffiths,J.F.(1972).ClimatesofAfrica. Vol. l0inWorldSurucyof Climatologv.Elsevier,Amstcr- dam. "colubra Gudynas, E. (1981). Confirmacion de la viviparidas en la sepia" Thannodynastes slrryills(Serpentes: Colubridae). C.E.D. OrioneCont. Biol. 5, 1-7. Guillette, L.J.,Ir. (1981).Ontheoccurrenceof oviparousandviviparousformsof theMexican lizard Sceloporusqeneus. Herpetologica 37, 17-15. Guillette, L J , Jr.(1982a). The evolution of viviparity and placentation in the high elevation, Mexican lizard Sceloporusaefieus. Herpetologica 38, 94-103. Guillette, L.J., Jr.(1982b). Effects of gravidity on the metabolism of the reproductivell' bimodal hzard, Sceloporusaeneus. J. exp. Zool. 223, 33-36. Guillette, L J.,Jr. (1983).Notesconcerningreproductionof themontaneskink, Etniccescopci. I. Herpetol. 17, 1M-148. Guillette, L.J., Jr., Jones, R. E., Fitzgerald, K. T., and Smith, H. M. (1980).Evolution of viviparity in the lizard genus Sceloporus.Herpetologica 36,201-275. Gyi, K. K. (1970). A revision of colubrid snakes of the subfamily Homalopsinae. Uniu. Kansas PubL Mus. nat. Hist.20, 47-223. Hardy, G. S. Q9n). The New Zealand, Scincidae (Reptilia: Lacertilia); a taxonomic and zoogeographic study. Near ZealandI. Zool.4,227-325. Hardy, G. S. (1979).The karyotypes of trvo scincid lizards, and their bearing on relationships in genus Leiolopismaand its relatives (Scincidae: Lygosominae). Neur ZealandJ. ZooI.6, 609-672. Hardy, L. M. (1975).Comparative morphology and evolutionary relationships of the colubrid snake genera Pseudofcimia,Ficimia, and Gyalopion.l. Herpetol.9,323-336. Hillenius, D. (1959). The differentiation within the genus ChamaeleoLaurenti, 7768. Beaut'ortia 8,7-92. Hoover, E. E. (1936).Onthebirth of Constrictorconstrictorimperator,incaptivity. Copeia1936, 62. "Evolution Horton, D. R. (1973). in the Genus Mabuya." Ph.D. thesis, University of New England (New South Wales, Australia). Huey, R. B. (7977). Egg retention in some high-altitude Anolis lizards. Colteia1977, 373-375. Ci47 /IPAFIITY IN FIEPTILEEI FTEFEFIENCEA Limpus, C. J., Baker, V., and Miller, J. D. (1979).Mo','ement induced mortalitt' of loggerhead eggs. Herpetologica35, 335-338. Loveridge, A. (1933). Report on the scientific results of an expedition to the southrvestern highlands of Tanganyika territory. VII. Herpekrlogy. Bull. Mus. cottrTt.Zod. 79,197,476. Loveridge, A. (1911).Revision of the African lizards of the family Amphisbaenidae. Bull. Mus. comp.Zool. 87, 351-451. Lydolph, P. E. (1977). Climates of the Sodet Llnion, Vol. 7 in Worlti Surr,ry of Climntologv. Elsevier, Amsterdam. McCoy, (1968). C. J. Reproductive cvcles and viviparitv in Guatemalan Coruthofhnnesltcr- carlnatus (Reptilia: Iguanidae). Herpetologica24, 175-778. McCoy, M. (1980). Reptilesof the Solomonlslands. W au Ecology Institute, Solomons, Handbook No. 7. McDowell, S. B. (1967).Aspidomorphus, a genus of New Guinea snakes of the famih'Elapidae, with notes on related genera. l. Zool. (Lond.) l5l, 497-543. McDowell, S. B. (1969). Notes on the Australian sea-snake Ephato1thisgreyi M. Smith (Ser- pentes: Elapidae, Hydrophiinae) and the origin and classification of sea-snakes. Zool, /. Linn, Soc. 48,333-349. McDowell, s. B. (1975). A catalogue of the snakes of New Guinea and the solomons, with special reference to those in the Bernice P. Bishop Museum. Part II. Aniloidea and Pythoninae. l. Herpetol.9, 1-80. McDowell, S. B. (1979). A catalogue of the snakes of New Guinea and the Solomons, with special reference to those in the Bernice B. Bishop Museum. Part lll. Boinae and Acrochordoidea (Reptilia, Serpentes). J. Herpetol. 13, 1-92. McDowell, s. B. and Bogert, C. M. (1954). The systematic position of Innthanotus and the affinities of the anguinomorphan ltzards. Bull. Amer. Mus. nat. Hist. 1,05, L-142. MacGregor, J. $975). An observation concerning the incubation period of the eggs of Opheo- drys uernalis. Kentuclcy Herpetologist 6, 181. Malnate, E. V. (1960). Systematic division and evolution of the colubrid snake genus Nafrir with comments on the subfamily Natricinae. Proc. Acad. nat. Sci.Philadelphia 112,41-71. Mao, S-H., Chen, B-Y., and Chang, H-M. (1977).The evolutionary relationships of sea snakes suggested by immunological cross-reactivity of transferins. Cotrtp.Biochem. Physiol.57A, 403-406. Mao, S. H., Chen, B-Y., Yin, F-Y., and Guo, Y-W. (1983). Immunotaxonomic relationships of sea snakes to terrestrial elapids. Comp. Biochem.Physiol. 74A, 869-872. Marcellini, D. L. and Davis, S. W. (1982). Effects of handling on reptile egg hatching. Herp. Rro.13, 43-44. Marx, H. and Rabb, G. (1965).Relationships and zoogeography of the viperine snakes (Family Viperidae). Fieldiana, Zool. 44, 1.61.-206. Mell, R. (1'929). Beitrage zur Fauna Sinica. lV. Grundzilge einer Okologie der chinesischenReptilien und einer herpetologischenTiergeographie Chinas. Waltet de Gruyter, Berlin. Mendelssohn, H. (1963). On the biology of the venomous snakes of Israel. l. lsraell. Zool. 12, 743-770. Mendelssohn, H. (1965). On the biology of the venomous snakes of Israel. II. lsraell. Zool.'1,4, 185-212. Mertens, R. (1964). Neukaledonische Riesengeckos (Rhacodactylus).Zool. Gart. 29, 49-57. Mertens, R. (1972). Uber den Senegal-skink (Scincusalbifasciatus) und seine Verwandten. Salamandra 8, 717-722. Minton, S. A., Jr. (1966). A contribution to the herpetology of West Pakistan. Butl. Am. Mus. nat. Hist.134.29-784. /I!'AFIITY IN FIEPTILECT nEFEFIENcEel EIBS' of Experimental rd mortality of loggerhead Minton, s. A., Jr. (1979).Common antigensin snake venoms. ln Handbook Pharmacologrl.snakevenoms (C-Y. Lee, ed.). Vol. 52,springer-Verlag, Berlin, pp. 847-862. in theAninnl ition to the southwestern Mittwoch, U. (1975).Chromosomes and sex differentiation.In Inter-sexuality ;. contp.Zool. 79, 797-476. Kingdom(R. Reinboth,ed.). AcademicPress, Nerv York, pp' 438-'146' nphisbaenidae.Bull. Mus. Mole,R. H. (1924).The Trinidadsnakes. Proc zttttl.Soc. Lond. !\,235-278' Moll, E. O, and Legler,J. M. (1971).The life history of a neotropicalslider turtle, Pseudenvs Scl.11,,7-102. vld Suruey of Clinatologrl. scrlpin(schoepff), in Panama.BulL LosAngeles Countrl Mus. nat. Hlsf. Montanucci,R. R. (1979).Notes on systematicsof horned lizardsallied lo Phrynosomaorbtcu- 'malan ConTfhoTrhanrspcr lcre(Lacertilia: Iguanidae). Herpetologicn 35' 116-124- Murphy, G. (196S).Pattern in life history and the environment.Amer. Natur.102' 397-404. lte, Solomons, Handbook Muskhelishvili,T. A. (1971).The biology of the reproductionof the aspVipera lebetina obtusa Dw. Akad. NaukGruz. Soobsch.62' 201-203. Pienaar, U. de V. (1978). The ReptilcFauna of the Kruger National Park. National Parks Board of South Africa. Pretoria. Pilorge,T. and Barbault,R. (1981).La vi','iparit6chez les lezards evolution et adaptation. Actrr Oecol., Oecol., Gener.2, 387-397. Pitman, C. R. S. (1971). A GuiLlc to the Snakesof UgarLda.Rev. Ed. Wheldon and Wcsley, London. Pope, C. H. (1935). The Reptilesoi China. American Museum of Natural Historv, Ne*, york, Pope, C. H. (1961).Tht Giant Snakes.Routledge and Keegan paul, London. Porter, W. P and Trafl', C. R. (1983).Biophvsical anall'ses of energr'tics,time-space utiiiza- tion, and distributional limits. In Lizard Ecology(R. B. Huey, E. R. pianka, and r. w. Schoener, eds.). Harvard University Press, Cambridge, Massachusetts, pp. 55_g3. Presch, w. (1969). E'olutionary osteology and relationships of the homed lizard genus phry- nosoma.Copeia 1.969, 250,275. Rage, J. (1982).L'histoire des serpentes.Pour la Science(April),16-27. Rahn, H. (1939). Structure and function of placenta and corpus luteum in viviparous snakes. Proc. Soc.exp. Biol. Med.40,381-382. Rand, A. s. (1963).Notes on the Clnmeeleobitaenntus complex. BuIl, Mus. com.st.Zool. 1g0,7- 2g. Rand, A. S. (1972). The iemperatures of iguana nests and their relation to incubation optima and to nesting sites and season. I{erpefologica28,252-253. Rankin, (1978). P. R. Notes on the biology of the skink Sphenomorphusltardalis (Macleay), including a captive breeding.iecord. Herpetot'auna10, 4-7. Raw, L. R. G. (1976). A survey of the dwarf chameleons of Natal, South Africa, with descrip- tions of three new species (sauria: Chamaeleonidae). Durban Mus. Nortit. 11,739-761. Raw, L. R. G. (1978). A further new dwarf chameleon from Natal, South Africa (Sauria: Chamaeleonid ae). Durban Mus. Nottit. 13, 265-269. Reeve, w. L. (1952). Taxonomy and distribution of the horned lizards, genus phrynosoma. KansasSci. Unia. Bull. 34,877-960. Reitinger, F. F. (1978). Common Snakesof SoutheastAsia and Hong Kong. Heinemann, Hong Kong. Rieppel, O. (1979). A cladistic classification of primitive snakes based on skull structure. Z. Zool. Syst. Eaol. Fortsch. 17,140-150. Romer, J. D. (1963). Notes on the white-lipped pit viper, Trimeresurusalbolabris Gray, in Hong Kong. In Venomousand PoisonousAnimals and Noxious Plants of the PacificReglon (H. L. Keegan and W. V. Macfarlane, eds.). Macmillan Co., New york, pp. 379-284. Rossman, D. (1973). Miscellaneous notes on the South American water snake genus Helicops. HISS Nears-/ 1., 189-191. Rossman, D. A. (1984). Helicopsangulatus (South American water snake) reprodu ction. Herp. Reaiap 15,50. Rossman, D. A. and Eberle, C. W. (1977). Partition of the genus Natrix, witt. preliminary observations on evolutionary trends in natricine snakes. Herpetologica 99,34-42. Roux, J. (1913). Les reptiles de la Nouvelle-Caledonie et des iles Loyalty, Nova Caledonia. In Rech?rches Scientifiquesen Nouaelle Caledonieet aux lles Loyalty (F . Sarasin and J. Roux, eds. ). Kreidels Verlag, Weisban. Rustamov, A. K. and Shammakov, S. S. (1967). Ecology of the takyr Phrynocephalus(phry- nocephalushelioscopus helioscopus Pallas) in Turkmenia. Zool. Zh. 46,747-248. saint Girons, H. (1962). cycle reproducteur de la vipdre a cornes, Cerasfescerasfes (L.), dans la nature et en captivit€. Bull. Soc. zool. Fr. 87, 47-Sl. Scherpner, C. (1968). Anniellidae, Xenosauridae. In Crzimek'sAnimal Life Encyclopedia.Yol. 6. Reptiles(8. Grzimek, ed.). Van Nostrand-Reinhold, New york, pp. 31.6-920. 891 VIVIPAFIITY IN FIEPTILEEI FIEFEFIENCEg schetty, P. (1951).Beobachtungen an vipera aspisaspis und cerasfesuipera. Aqua Tura-Z. 4, 83. Park. National Parks Board of schmidt, K. P. and Inger, R. F. (7957). Liaing Reptilesof the world. Hanover House, Garc'len Cify, New e'r,olutionet adaPtation,Acla Jersey. Schnurrenberger, Il. (1958). Het leven van A.splscercsfes (Linn6) in de vrije natuur. Lacerta16, 58-60. . Ed. Wheldon and Weslel', schwaner, T. D. and Dessauer, H. C. (1982). Comparative immunodiffusion survey of snake transferrins focused on the relationships of the natricrnes. Copeia1982,547-549. Natural Hiskrrr', Nen' York Schweizer, H. (1956). Die levante-otter von nordwet-Afrika-Vipara lebetita mauritaricn rul, London. (Guichenot). Aqua Terra-L. 9, 790-193. :nergetics,time-sPace utiliza- W. (1976). Climates of Central and South America. Vol. 12. \Norld Sutuatl ol Le,v,E. R. Pianka, and T W. Schwerdtferger, Climatology. Elsevier, Amsterdam. rssachusetts, pp. 55-83 A. M. (1940). Researches on the viviparity of reptiles. Mosc. Soc. Nat. (lubileelssue)' the horned lizard genus Phry Sergeev, 7-34. and Claypool, L. (1978). Nest sites of a northern population of an oviparous 16-27. sexton, o. J. snake, Opheodrysuernalis (Serpentes, Colubridae). l. nat. Hist 12, 365-370' luteum in viviparous snakes shammakov, s. s., Ataev, c. A., and Kamalova, z. Y. (1973). Ecology of the transcaspian lizard Phrynocephalusin the Turkmen SSR. Ekolo3iyn 6, 80-83. ull. Mus. comp.Zool. 130,7-29. Shcherbak, N. N. (1971).Taxonomy of the genus Erenlas (Sauria, Reptilia) in connection with relation to incubation oPtima the focuses of the desert-steppe fauna development in Palearctic. Vest. Zool.2, 48-55. Shine, R. (1977). Reproduction in Australian elapid snakes. II. Female reproductive cycles. nomorphuspardalis (MacleaY)' Aust. l. Zool. 25, 655-666. "safe shine, R. (1978). Propagule size and parental care: the harbour" hypothesis. l. theot. rl, South Africa, with descriP- Biol.75, 417-424. Mn Mus. Noait. 11,739-767. shine, R. (1979). Activity patterns in Australian elapid snakes. Herpetologica35, 1-11. tSauria: Natal, South Africa "Costs" Shine, R. (1980a). of reproduction in reptiles. Oecologia46' 92-100' Shine, R. (1980b). Comparative ecology of three Australian snake species of the genus :d lizards, genus Phrynosoma. Cacophis(Serpentes : Elapidae). Copeia1980, 831 -838. (1981). The evolution of reptilian viviparity: a review of hypotheses and case- ong Kong. Heinemann, Hong Shine, R. histories. Abstr. for ssAR-HL Annual Meeting, Memphis, Tennessee. Circulated 15 Janu- ary1981. s based on skull structure Z. shine, R. (1983a) Reptilian reproductivemodes: the oviparity-viviparitycontinuum. Her' petologiu39,1-8. 'esurusalbolabris GraY, in Hong viviparity in cold climates:testing the assumptionsof an evolu- nts of the Pacific Region (H. L. Shine,R. (1983b).Reptilian 57' 397-405. v York, pp. 379-384. tionary hypothesis. Oecologia of small fossorial Australian snakes of the genera Neelapsand rn water snake genus HelicoPs. Shine, R. (1984). Ecology simoselaps(serpentes, Elapidae). spec. Publ. Mus. nat. Hist. unia. Kansas10, 173-783. on elapid phylogeny.In Biologyof AustralasianFrogs and ter snake) reproduction. HerP. shine, R. (1985).Ecological evidence Reptiles(G.Grigg,R.Shine,H.Ehmann,eds')'SurreyBeatty,Sydney,inpress' Shine, R. and Berry, F. (1978).Climatic correlatesof live-bearingin squamatereptiles. enus Nafrix, with Preliminary ]. lerp et olo gica 33, 34- 43. Oecologia33,267-268. evolutionof iive-bearingin lizardsand snakes.Am. Natur' :s Loyaity, Nova Caledonia. In Shine,R. and Bull, J. J. Qg79).The r13,905-923. r (F. Sarasinand J. Roux. eds.1 Smedley,N. (1931).Oviparity in a sea-snake,Lnticauda colubrlna (Schneid)' BulI. Raffles Mus.5' he takyr PhrynocePhalus(Phry' 54-58. Zool. l. zh. 46,74L-748. Smith, H. M. (1939).The Mexicanand CentralAmerican lizards of the genusSceloporus. Hist. 26, L-397. zs, Cerastescerastes (L.), dansla Ser.Field Mus. Nat. smith, H. M. andHall,w.P.(1974). contributionstotheconceptsof reproductivecyclesand group of the lizard genusSceloporus. Great Basin Nat . 34, 97- Animal Life EncYcloPedia.Yol. 6. the systematicsof the scalaris York, pp. 316-320. 104. ci9e THE EVctLUTIclN OF V]VIPAFIITY IN FIEPTILEEI Smith, H. M. and Tavlor, E. H. (1950).An annotated checklist and ket' to the reptiles of Mexico exclusiveof the snakes.Bul/. U.5. Naf. Mus. i.99,7-799. Smith, H. M., Sinelnik, C., Farvcett,J. D., and Jones, R. E. (1973).The occurrence and taxonomic significance of a protracted reproductivc scason with alternation of ovulation in lizards. tsull. Canad.Amphih. Rept.Consertt. Soc. 11, 1-5. Smith, M. A. (1930).Ovoviviparity in sea snakes. Nature,Lond.126,568. Smith, M. A. (1935). The Faunaof Rritish lndia. Rcptilinnnd Amphibia. Vol. II, Sairrln.Taylor and Francis, London. Smith, M. A. (1943). The t"aunnof British lndta. Rqttilia anLlAtt1ltiltit. \'ol. IIl, SL'r;ri,n1i's.Tavlor and Francis, London. Smith, M. A. (1973). The British Amphibiansand Reptiles.5th Ed. Collins, London. Sowerby, A. de C. (1930). IV. The reptiles and amphibians of the Manchurian region. In The NattLralistsin Manchurh. Tientsin Press, Tientsin, pp. 1 41. Sparvls,S. (1973).Fossorial snake gives live birth. tlull. LastAfr. ntlt. Hist. Soc.,Scptetn[ter 1973, 722. Spellerberg, I. F. (1975). The grass snake in Britain. Oryx 13, 289-300. Spellerberg, L F. (1976). Adaptations of reptiles to cold. In Morphologvand Biologyo.f Reptiles (A. Bellairsand C. B. Cox, cds.). Linn. Soc.Syntp. \er.3,261-285. Spellerberg, L F. and Phelps, T. E. (1977). Biology, general ecologv and behaviour of the snake, Coronellaaustriaca Laurenti. Biol. l. Linn, Soc.9,733,764. Stebbins, R. C. (1958).A new alligator lizard from the Panamint Mountains, lnyo County, California. Amer. Mus. Nouitates0883),1-27. Stemmler-Cyger, O. (1965). Zur Biologie der Rassen von Echrscarinatus (Schnei Eumeceswith an account the distribution and relationships of its species. Uniz,. Kanses and key to the reptiles of of Sci.Bull.23,1-643. 99. 1973).The occurrence and Taylor, E. H. (1963).The lizards of Thailand. Unitt. KansasSci, Bull. 44,687-1077. ith alternation of ovulation Taylor, E. H. (1965).The serpentsof Thailand and adjacentrvaters. Urrli,. Kansas Scr. Bull. 45, 609-1096. 26, 568. Telford, S. R., III (1980). Notes t'tn Agkistrodonhimnlaynntts from Pakistan's Kaghan Valler'. 1980, 15,{-155. . Vol. il, Sarrla. Taylor and Copeia Terentev, P. V. and Chernov, S. A. (1965). Key to Antlthibiansnnd Reltiles. Translation, Israel Program for Scientific Translations, l. \','1.lll. S(rl" ril,:.Tr\ lur Jerusalem. Thomas, B. W. (1980). New Zealand Gekkonidae: evolutionary trends and taxa. In Near ZealandHerpetology (Newman, D. G., ed. Occ. Pap. New ZealandWild. Serzt.(2), 333-337. lollins, London. ). Thomas, R. A. and Dixon, R. (1976).A re-evaluation of the Sceloporusscalaris group (Sauria: Manchurian region. ln fhe J. Iguanidae). Southwest.Nat. 20, 523-536. Thompson, (1981). A study of the sources of nutrients for embryonic development in a t. Hist. Soc., Seyttcn*er1973, J. viviparous lizard, Sphenomorphusquoyii. Comp. Biochent.Physiol.70A,509-518. -300. Thompson, J. 0982). Uptake of inorganic ions from the maternal circulation during develop- ment of the embryo of a viviparous bzard, Spthenotnorphusquoyii. Contp. Biochem.Physiol, nlogv and Broloqaof Rcpttles 7tA, r07-112. -285. Tinkle, D. W. (1967). The life and demography of the side-blotched lizard, Uta stansburiana, krgv and behaviour of the Misc.Publ. Mus. Zool., Uniu. Michigan L32,7-182. 54. Tinkle, D. W. and Gibbons, J. W. (7977). The distribution and evolutiorr of viviparitv in Mountains, lnYo Countl , reptiles. Misc. Publ. Mus. Zool., Unia. Michigan 154, 1-55. Towns, D. R. (1975). Reproduction and growth of the black shore skink, Leiolopismasuteri caritatus lSchnciderl 1801. (Lacertilia: Scincidae), in north-eastern New Zealand. New Zealandl, Zool.2, 409-423. Tryon, B. W. (1979). Notes on reproduction of the lizard Cophotisceylonica. Brit.I. Hcryrt. i, natus sochurekisubsP. nov u5-u7. Underwood, G. (1967). A Contribution to the Classificationof Snakes.Briiish Museum (Natural asselotter(Lchis carinalus) History), London. "Classification Underwood, G. (1971). A modern appreciation of Camp's of the lizards." In Llerund Sochurek 1969, der Camp's Classificationof the Lizards. SSAR Facsimile Reprint, pp. 7-17 . ., Leipzig 40,200-210. Underwood, G. (1979). Classification and distribuhon of venomous snakes in the world. In errarium. Aqua Terra-Z.24, Hnndbookof ExperimentalPhnrmacology. Snake Venonts (C-Y. Lee, ed.). Vol. 52, Springer- Verlag, Berlin, pp. 15-40. .iversity Press, Rutherford Van Devender, T. R. and Howard, C. W. (1973).Notes on natural nests and hatching success >ufion of the green snake, in the regal horned lizard (Phrynosoma solare) in southern Arizona. Herpetologica 29, 238 239. lizard subfamily Diploglos- Van Devender,T. R. and Van Devender,W. (1975).Ecological notes on two Mexicanskinks (genusEumeces). Southwest. Nat. 20,279-282. ooth snake (Coronellagiron- Vessey-Fitzgerald,D. F. (1958).The snakesof northem Rhodesiaand the Tanganyikaborder- lands. RhodesianSci. Assoc. Proc. Trans. 46,17-102. oms. ln Handbookof LxPeri Visser,J. (1967).First report of ovoviviparityin a southernAfrican amphisbaenid,Monopeltis pringer-Verlag, Berlin, PP. c. capensis.Zool. At'ric.3, 111-113. Visser,J. C.I. Q972).Ontogeny of the chondrocraniumof the chamaeleon,Mictosaura pumila rz, Guatemala. Misc. Publ. pumila(Daudin). Ann. Unia. Stellenbosch47A, 1-68. Visser,J. (1975).Oviparity in two South African skinks of the genusMabuya, with noteson is tigrinus from the Korean hatching. ZooL At'ric. 10,215-217. Vitt, L. J. and Blackburn,D. G. (1983).Reproduction in the lizardMabuya freafful (Scincidae): a World rd Western Asia. ln commentary on viviparity in New World Mabuya.Canad. l. Zool. 61, 2798-2806. Yogel, Z. (1963).Aus demLeben rler Reptilien.Verlag Artia, Prag. [" IN FIEPTILEE| B9.l THE EVG,LUTIGIN OF VI\/IPAFTITY scale characters in evaluation of waddick, J. w. and smith, H. M. (1974). The significance of the lizard genera Gerrhottotus,Elgaria, and Barisia. Graat Basin Nat. 34,257-266. Botnbaynat' Wall, F. (1918). A popular treatise on the common Indian snakes. Part 25. ltntr. Hist. Soc,25,375-382. Wall, F. (7921). Snakesof Ceylon. Cottle, Colombo' of Clinntol- wallen, c. c. (1971). Climates of Northern and western Europe. ln world surt'al ogy. Vol. 5, Elsevier, Amsterdam. olCtocodyluspo,'osus in webb, G. J. w., Messel, H. and Magnusson, w. (1977).The nesting Arnhem Land, northern Australia. Copeia1977, 238-249' Publ Mus. webb, R. G. (1968). The Mexican skink Eumeceslyrre (squamata, scincidae). Michignn Stnte Unia. 4,7-28. European Weekes, H. (1933). On the distribution, habitat and reproductive habits of certain viviparitv' and Australian snakes and lizards, with particular regard to their adoption of Proc. Linn. Soc.Nezl South Wales58, 270-274' regard to the weekes, H. c. (1935). A review of placentation among reptiles with particular function and evolution of the placenta. Proc. zool' Soc Lontl' 1935,625-645' and young of Texas and Mexican Werler, J. E. (1951). Miscellaneous notes on the eggs reptiles. Zoologica36, 37-48. of the Pacific area. In venomousancl werler, J. E. and Keegan, H. L. (1963). Venomous snakes V- Macfar- PoisonousAnimals and Noxious Plantsof the PacificRegion (H. L. Keegan and w- lane, eds.). Pergamon Press; Oxford, pp' 219-325' "Die In- werner, F. (1930). Lurche und Kriechtiere, Brehm's Tierleben." Bibliographisches stitut, Leipzig. temPera- werner, Y. L. and whitaker, A. H. (1978). Observations and comments on the body tures of some New Zealand reptiles. Nau Zealandl ' Zool' 5' 375-393' Whitaker, R. (1978). Common Indian Snakes.A Fietd Guide, Macmillan Co., Delhi Spol' Willey, A. (1906). Viviparity of Cophotisceylanica and oviparity ol Ceratophorastoddartii' Leul. J, zJ3-zJ/. williams, G. c. (1966). Adaptationand Natural Selection.Princeton University Press, Princeton, New Jersey. Worrell, E. (1963). Reptitesof Australia. Angus and Robertson, Sydney' Endoctinol' Yaron, Z. (1972). Endocrine aspects of gestation in viviparous reptiles. Gen. Comp. Suppl. 3, 663-674. emphasis zwrefel, R. G. (1960). Reproductive biology of anurans of the arid southwest, with |_64. on adaptation of embryos to temPelatule. BuIl. Amer. Mus. nat. Hist, 740,