WIXLIB

HISTORICAL PATTERNS IN LIZARD ECOLOGYteiids, and 6 lacertids), but we commenton how they compare with lizards ingeneral. We also assembled a data set onlizard body temperatures from ourstudies. Because we were only interestedin comparisons at higher taxonomiclevels, we did not examine potentialdifferences among taxa in relationshipsof body temperatures to substrate or airtemperatures. For perspective, weincluded data on gekkonoid andiguanian lizards. Thus, data for thisanalysis include 3889 individual bodytemperature measurements on 67 lizardspecies.Dietary comparisons presentproblems due to variation among studiesin data reported. Thus we restrict ourdietary analysis to our own data becausethey were collected similarly. Wecombined data on desert and rainforestlizards (see VITT et al., 2003 for details)and restrict the analysis to the sevenmost important prey categories byvolume. We examine differences amongthe three major lizard clades, Iguania,Gekkota, and Autarchoglossa.Finally, to demonstrate limitationsof basing conclusions on numericalversus volumetric dietary data, weprovide analysis of a data set on 448Ameiva ameiva, containing 27 preycategories, 165 Anolis fuscoauratuscontaining 22 prey categories, and 58Plica umbra containing 16 preycategories. These species were chosenbecause they represent different dietstrategies. One (A. ameiva) is a generalistthat eats a lot of termites numerically,another (A. fuscoauratus) is a dietarygeneralist that eats a lot of ants andsmall beetles numerically, and the third(P. umbra), is an ant specialist. Methodsfor dietary analysis can be foundelsewhere (e.g., VITT & ZANI, 1996).RESULTS AND DISCUSSIONSetting the evolutionary stageHigh activity levels, jaw prehensionof prey, a tongue designed to transmitchemical information from the externalenvironment to the lizard’s sensorysystem, and an active foraging modemust have given ancestors to autarchoglossans a competitive advantageover all small diurnal vertebrates with asit-and-wait foraging mode in a worldrich in invertebrates and smallvertebrates (PIANKA & VITT, 2003; VITTet al., 2003). Jaw prehension dates backto the ancestor of scleroglossans andprovided two primary advantages: 1) thetongue could be freed up for otherfunctions and 2) larger prey could bemanipulated. In autarchoglossans, thetongue is used to transmit chemicalsfrom the external environment to thelizard’s vomeronasal sensory system.Even though scleroglossans, andautarchoglossans in particular, arecapable of eating large prey, manycontinue to eat small prey (VITT et al.,2003). Sit-and-wait foraging lizards inthe Iguania depend upon vision for preydetection and pursue moving insects(HUEY & PIANKA, 1981) and prey arecaptured by lingual prehension(SCHWENK, 2000). Consequently, mostnonmobile or hidden invertebrates andvertebrates are likely undetected byiguanian lizards, and large-sized prey,143

VITT, L. J. & PIANKA, E. R.(ANCOVA slope test, F1, 4308 22.1, P 0.0001), species, and in some casessexes, the all-species relationship isrelatively tight. Some exceptions exist,but a vast majority of lizards in theLacertoidea are terrestrial active foragers.Teiids and Gymnophthalmids arerestricted to the new World and, withthe exception of a handful of species,they are terrestrial. The strikingexceptions are at opposite ends of a sizegradient. A few small gymnophthalmidslike Bachia (AVILA-PIRES, 1995; COLLI etal., 1998), Psilophthalmus, andProcellosaurinus (RODRIGUES, 1991a, b)are nearly limbless and liveunderground, and two of the largestteiid genera, Dracaena and Crocodilurusbask from elevated perches in trees overwater and forage in swamps, streams,and rivers (AVILA-PIRES, 1995; PIANKA& VITT, 2003).particularly vertebrates, would bedifficult to manipulate with the tongue.Poorly developed chemosensory systemsin iguanians did not allow thesophisticated prey discriminationexhibited by most scleroglossans, andautarchoglossans in particular (COOPER,1994a, b, 1995a, b; SCHWENK, 1993a,b; 1994). Nevertheless, several cladeswithin Iguania developed chemosensoryprey discrimination in response to aswitch to herbivory (C OOPER &ALBERTS, 1990, 1991).Teiids, gymnophthalmids, andlacertids comprise the Lacertoidea(Figure 1), a group of lizards generallycharacterized as elongate and fusiform.The conservative morphology of thesethree taxa is apparent in SVL to massregressions (Figure 2). Althoughsignificant differences in the SVL-massrelationships exist between clades4GymnophthalmidaeLacertidaeTeiidaeLog10 total mass3210–11.01.21.41.61.82.02.22.42.62.8Log10 snout-vent lengthFigure 2. Relationship between total body mass and snout-vent length in teiid, gymnopthalmid, and lacertid lizards.The outlier points falling below the regression are nearly limbless gymnophthalmid lizards in the genus Bachia144

HISTORICAL PATTERNS IN LIZARD ECOLOGYTeiid morphology and its consequences4-legged squamates so restrictive thatlarger deviations in length-massrelationships simply do not occur? If so,did a major physiological shift in theancestor of snakes open up newphysiological and ecological options?Even though the SVL - massrelationship in 4-legged squamates isconservative, interesting and in somecases surprising variation exists.Comparison of Teioidea with Iguaniareveals a significant difference in slopesof the SVL - mass relationship withTeioidea increasing in mass more as SVLincreases (F 1,7152 0.07, P 0.0142).Power of the slope test is high (0.69). Tobriefly summarize, gekkotans haveshifted away from an overall lizardBauplan by becoming relatively light inmass and Teioidea have shifted theopposite direction becoming slightlyheavier. This result may seem surprising,because one is immediately struck by theapparent streamlined nature of Teioideamorphology. Coupled with streamlinedmorphology in Teioidea is elongationand in many instances apparentthickening and lengthening of the tail,which can be important in locomotion(e.g., B ALLINGER et al., 1979). Thisprobably contributes to the added sizespecific mass. Higher size-specific massin Teioidea may partially account for thenearly clade-wide restriction to terrestrialmicrohabitats. Cost of locomotioncombined with an active lifestyle maylimit abilities of Teioidea to climb, ahypothesis that remains to be tested.Conservatism in morphologywithin Teioidea is reflected both incomparisons among subclades (Figure 2)and in comparisons among genera ofWe first comment on lizardmorphology in general because a broadperspective is necessary to interpretpatterns within the Lacertoidea. Weconcentrate on the SVL to total massrelationship even though someinteresting differences occur in othermorphological variables. Among thosespecies for which we had both SVL andmass data, a significant difference inslopes (ANCOVA with clade as the classvariable and log10 SVL as the covariate,F 2,9353 148.6, P 0.0001) existsamong Iguania, Gekkota, andAutarchoglossa. These clades do not fallon the same regression line. Slopes ofthe log-log relationships are: Iguania3.22, Gekkota 2.93, and Autarchoglossa3.22. Comparison of autarchoglossansand iguanians yields no difference inslopes or intercepts (F1,7754 0.07, P 0.7858; F1, 7754 1.09, P 0.2972) butthe power in both tests (0.058 and0.171, respectively) is low. Thusautarchoglossans and iguanians fall on asimilar but highly variable regression.Most of that variation is amongiguanians. An important, but notobvious point in the Iguania Autarchoglossa(lizardsonly)comparison is that SVL-specific mass iniguanians is attained primarily by arobust body whereas similar SVLspecific mass in many autarchoglossansis attained by a combination of bodyand tail elongation with mass of the tailcontributing greatly to overall mass. Weonly point that out here, but this raisesseveral interesting questions. Forexample, are physiological processes of145

VITT, L. J. & PIANKA, E. R.successful in terrestrial microhabitats.The Bauplan of many lacertids is similarto that of Teioidea (e.g., Figure 6), likelyreflecting success of the ancestralBauplan of lizards in the Lacertoidea.The most striking difference betweenLacertids as a group and Teioidea is thatteiid lizards (Figure 3). Whether one isexamining ecologically distinctgymnophthalmids (Figure 4) or teiids ofdrastically different body sizes (Figure5), the overall Bauplan is similar. Oneconclusion that can be drawn from thisis that the Teioidea Bauplan has provenLog10 total nambis21011.522.53Log10 snout-vent lengthFigure 3. Relationship between total body mass and snout-vent length in 5 genera of teiid lizards representingboth subclades (Teiinae and Tupinambinae)Figure 4. A. Vanzosaura rubricauda, a gymnophthalmid that lives in open habitats in central and southern SouthAmerica. B. Cercosaura ocellata, a gymnophthalmid that lives in leaf litter of the southern Amazon rainforest andnorthern Cerrado of Brazil146

HISTORICAL PATTERNS IN LIZARD ECOLOGYFigure 6. Male and female of the lacertid Podarcis siculafrom Menorca, Spain(e.g., P IANKA , 1994), resembling inmany ways, some of the larger teiids likeAmeiva, Tupinambis, Crocodilurus, andDracaena. The only Varanoidea thathave survived in the New World are inthe genus Heloderma (Helodermatidae)and their lifestyle is quite different fromthat of varanids, and their distribution islimited. Absence of small-bodied teiidsmay result from historical interactionswith the sister clade Gymnophthalmidae, most of which are small.Within Lacertidae, some species haveevolved small body size (e.g.,Takydromus) but in terms of the numberof small species, lacertids have not beennearly as successful at generating smallbodied species in the Old World asgymnophthalmids have been in the NewWorld. One hypothesis for thisdifference is that gymnophthalmids didnot encounter a diverse skink faunaduring their early evolutionary historywhereas lacertids did. Small-bodiedskinks have been very successfulthroughout the Old World and onPacific islands. Considering thatgymnophthalmids have done well insome relatively cool environments, oneFigure 5. A. The large-bodied teiid Tupinambislongilineus. B. The relatively small-bodied teiidKentropyx altamazonica. Both are from the Rio Ituxi inAmazonas, Brazil. Note the overall similarities inBauplan even though each is in a different subclade(Tupinambinae versus Teiinae) and mass differs by anorder of magnitudeno extant giant species of lacertids existand no diminutive species of Teiidaeexist. The smallest is Cnemidophorusinornatus, which, at a maximum SVL ofabout 69 mm, is still larger than mostgymnophthalmids. A few islandlacertids are large compared to others(e.g., Gallotia), but no morphological orecological equivalents of Tupinambis,Dracaena, or Crocodilurus exist in theLacertidae. The most likely explanationfor this may be the presence of varanidsthroughout much of the evolutionaryhistory of lacertids in the Old Worldand their absence throughout recenthistory of New World lizards. Varanidsare highly active and voracious predators147

VITT, L. J. & PIANKA, E. R.(1.006 and 0.904, respectively; F1, 4333 145.8, P 0.0001). Teiids haverelatively longer hindlimbs and thedisparity appears to increase withincreasing body size. Reasons for thisremain unknown, but several hypothesescan be proposed. Diversification ofteiids in the absence of varanids mayhave released them from evolutionaryconstraints on limb length allowingthem to diverge considerably from theTeioidea ancestor. Alternatively, lacertidhindlegs may have shortened in responseto a lack of iguanian diversity providingopportunities to use microhabitats thatdid not contain competitors historically.has to wonder why they did not colonizeNorth America. One possible reason isthat the only potential distributioncorridor, Central America, islongitudinally small and has onlyperiodically provided a connection toNorth America.Several other aspects of morphologyprovide insight into potential similaritiesand differences in ecology of lacertidsand teiids. We consider one here.Relative length of hindlimbs variesamong subclades within Lacertoidea(Figure 7). Slopes differ in therelationships between hindlimb lengthand SVL (F 2,5159 57.8, P 0.0001)amongsubcladeswithgymnophthalmids having the shortesthindlimbs, lacertids intermediate, andteiids the longest. Even though relativehindlimb lengths of teiids and lacertidsappear superficially similar in Figure 7,the regressions have different slopesTeiid activity and body temperaturesAll known lizard species in theLacertoidea are diurnal. Compared withother clades, teiids and lacertids tend tohave slightly higher body 2Log10 hindleg 2.62.8Log10 snout-vent lengthFigure 7. Relationship between SVL and HLL in families of the Lacertoidea148

HISTORICAL PATTERNS IN LIZARD ECOLOGYteiid-like in their thermal biology andactivities. In other environments,particularly high elevation environments, lacertid thermoregulationappears similar to that of high elevationiguanians (e.g., CARRASCAL et al., 1992;M ARTIN & S ALVADOR , 1993; V ANDAMME et al., 1987, 1990; BAUWENS etal., 1990, 1996). They maintainrelatively low body temperatures andbask considerably to gain heat. Thus,some lacertids have diverged from theirancestors in thermal physiology allowingthem to persist in Old World habitatssimilar to those dominated by iguaniansin the New World.and gymnophthalmids tend to haveslightly lower body temperatures (Figure8). Teiids in general and lizards in thegenus Cnemidophorus in particular, tendto have among the highest active bodytemperatures known in lizards, withsome exceeding 40 C (e.g., ASPLUND,1970; C ASAS -A NDREU & C URROLA HIDALGO, 1993; PIANKA, 1970; SCHALL,1977). Nevertheless, some desertlacertids appear to have equally highbody temperatures while active. Forexample body temperature of Nucrastessellata averages 39.2 2.8 C (PIANKA,1986). In some environments(particularly deserts), lacertids are veryDiets and foraging modeThe shift from lingual to jawprehension in scleroglossans withconcomitant enhancement of chemicalsensing (olfactory in Gekkota andvomeronasal in Autarchoglossa) set thestage for a major dietary shift (see VITTet al., 2003) that ultimately led to theevolution of extreme modifications ofthe jaw allowing ingestion of huge preyas seen in many varanoids (in particular,most snakes; GREENE, 1997). Chemicaldiscrimination of prey (e.g., COOPER,1990; 1994a, b; 1995a, b; 1997a, b)allowed 1) identification of highlyprofitable prey and 2) identification ofpotentially toxic prey. The obviousprediction is that scleroglossans shouldinclude less biomass of noxious insectssuch as ants in their diets and replacethose with insects of higher energycontent or those not producing noxiouschemicals. In general, this appears to bethe case (Figure 9). We were unable toFigure 8. Body temperatures of lizards in 6 clades149

VITT, L. J. & PIANKA, E. R.40NEOTROPICAL LIZARDSMean percent by volume3020100G TLS A B H40G TLS A B HG TLS A B HDESERT LIZARDS3020 —Note: delete"hymenopterans"hymenopterans100G TLS A B HG TIGUANIALS A B HGEKKOTAG TLS A B HAUTARCHOGLOSSASCLEROGLOSSAantsbeetlesotherFigure 9. Diets of Neotropical lizards (upper) and desert lizards (lower). Only the seven most important prey categoriesare shown. A dietary shift occurs in Scleroglossa with a reduction in proportions of ants and other noxious insects. Antsare replaced by a combination of grasshoppers, insect larvae, and spiders in Neotropical lizards and termites and spidersin desert lizards. Prey types are: (G) grasshoppers and crickets, (T) termites, (L) insect larvae, pupae, and eggs, (S)spiders, (A) ants, (B) beetles, and (H), non-ant hymenopterans. Many beetles, ants, and other hymenopterans producenoxious chemicals for defense (e.g., BLUM, 1981; EVANS & SCHMIDT, 1990). Modified and reprinted with permissionfrom the American Naturalisttest this hypothesis for lacertids ingeneral because, aside from ERP’s desertlizard studies, a vast majority of data ondiets of lacertids does not includevolumetric data. Most such studiesprovide counts of prey, and ants (andother hymenopterans) do constitute asignificant portion of the diet for manylacertids based on counts (e.g., CASTILLAet al., 1991; P ILORGE , 1982; P ÉREZ M ELL ADO et al., 1991; P ÉREZ M ELL ADO , 1998). One study thatcontained numeric and volumetric data(R OBINSON & C UNNINGHAM , 1978)revealed that two Namib lacertids haddiets in which ants and otherhymenopterans were not dominant. Inone, Meroles (formerly Aporosaura)anchietae, ants comprised 0.42% of thediet by volume and all hymenopteranscomprised only 5.19%. In the other,Meroles cuneirostris, ants comprised6.9% of the diet by volume and allhymenopterans comprised 16.5%. Thepossibility exists that ants eaten by manyEuropean lacertids do not contain150

HISTORICAL PATTERNS IN LIZARD ECOLOGYnoxious chemicals and this would beworth looking into further.We now diverge slightly to providean example of how importantvolumetric data are in lizard studies. Weselected data on three lizard species thatdiffer ecologically and in their use ofprey; the large-bodied teiid Ameivaameiva, the small-bodied polychrotidAnolis fuscoauratus, and the mediumsized tropidurid Plica umbra (Figure10). In A. fuscoauratus, large numbers ofants and small beetles contribute verylittle to the diet volumetrically and,consequently reliability of numeric dataas a predictor of volumetric data is poor(52.3%). In A. ameiva, high numbers oftermites contribute to lack ofconcordance between numerical andvolumetric data, and reliability ofnumerical data as a predictor ofvolumetric data is also low (17.6%). In atrue ant specialist, P. umbra, numericand volumetric data are nearly identicaland reliability of numerical data as apredictor of volumetric data is high(99.3%). Use of numerical data only isunlikely to allow discriminationbetween lizard species that eat a lot ofants contributing little to their dietvolumetrically, and true ant specialists.Use of volumetric data alone is likely toreveal little about specific foragingbehaviors. To exemplify the latter,consider A. ameiva, a species that eats al

porción independiente de sus historias evo-lutivas. Dentro de Teioidea, la mayor diver-gencia se prudujo en tamaño corporal, dando lugar a Gymnophthalmidae (pequeño tamaño) y Teiidae (gran tamaño). El peque-ño tamaño corporal de los gimnoftálmidos afectó a su ecología de modo diferente a como lo hizo el tamaño corporal grande en