HistoryandFunctionofScaleMicroornamentation inLacertidLizards



ABSTRACTDifferencesinsurfacestructure(ober- mostfrequentlyinformsfromdryhabitatsorformsthat hautchen)ofbodyscalesoflacertidlizardsinvolvecell climbinvegetationawayfromtheground,situations size,shapeandsurfaceprofile,presenceorabsenceoffine wheredirtadhesionislessofaproblem.Microornamen- pitting,formofcellmargins,andtheoccurrenceoflongi- tationdifferencesinvolvingotherpartsofthebodyand tudinalridgesandpustularprojections.Phylogeneticin- othersquamategroupstendtocorroboratethisfunctional formationindicatesthattheprimitivepatterninvolved interpretation.Microornamentationfeaturescandevelop narrowstrap-shapedcells,withlowposteriorlyoverlap- onlineagesindifferentordersandappeartoactadditively pingedgesandrelativelysmoothsurfaces.Deviations inreducingshine.Insomecasesdifferentcombinations fromthisconditionproduceamoresculpturedsurfaceand maybeoptimalsolutionsinparticularenvironments,but havedevelopedmanytimes,althoughsubsequentovert lineageeffects,suchaslimitedreversibilityanddifferent reversalsareuncommon.Likevariationsinscaleshape, developmentalproclivities,mayalsobeimportantintheir differentpatternsofdorsalbodymicroornamentationap- peartoconferdifferentandconflictingperformancead- genesis.Thefinepitsoftenfoundoncellsurfacesare vantages.Theprimitivepatternmayreducefrictiondur- unconnectedwithshinereduction,astheyaresmaller inglocomotionandalsoenhancesdirtshedding,especially thanthewavelengthsofmostvisiblelight.J.Morphol. inground-dwellingformsfrommoisthabitats.However, 252:145–169,2002. ©2002Wiley-Liss,Inc. thissmoothmicroornamentationgeneratesshinethat maycompromisecrypticcoloration,especiallywhenscales KEYWORDS:;;scalemicroorna- arelarge.Manyderivedfeaturesshowcorrelationwith mentation,homoplasy;lineageeffects;friction;crypsis; suchlargescalesandappeartosuppressshine.Theyoccur shine;dirt-shedding

Phylogeneticinformationmaypermitthehistory turesadditionaltothecell-likeenclosures.Also,the ofmorphologicalfeaturesthatvarycomplexly oberhautchenandunderlyinglayersmayallbe throughacladetobeatleastpartlyreconstructed. ruckedtoproduceridgesonthescalesurface(Har- Thehistorycanthenbeusedtoexploretheevolu- vey,1993).Theoverallstructureoffeaturesofthe tionaryoriginsofthevariation,inparticularbyen- oberhautchensurfaceandepidermalfoldingisre- ablingcorrelationsbetweentheappearanceofpar- ferredtohereasmicroornamentation(Ruibal,1968) ticulartraitsandchangesinselectiveregimetobe buthasalsobeentermedultradermatoglyphics recognized.Thisapproachisappliedheretothefine (Larsenetal.,1973),dermatoglyphics(Bursteinet surfacestructureofthescalesoflacertidlizards. al.,1974),microdermatoglyphics(Dowlingetal., Thescalesofsquamates(lizardsandsnakes)have 1972),microstructure(PerretandWuest,1983),and arigidouterepidermallayerof␤-,the microarchitecture(Peterson,1984a). ␤-layer,whichisunderlainbythemesosandthen Squamatemicroornamentationiseasilystudied the␣-layers.Alltheseareformedfromcellspro- byscanningelectronmicroscopy(SEM)andthereis ducedbythelivingbasallayeroftheepidermis,the nowasubstantialliteratureonthesubject.Forin- stratumgerminativum.The␤-layeriscoveredby stance,publicationsdealingwithdorsalbodyscales theoberhautchen(theanglicizedformoftheoriginal include:Bryantetal.(1967),MonroeandMonroe GermanOberha¨utchen,recommendedbyIrishet (1967),Ruibal(1968),StewartandDaniel(1972, al.,1988).Bythetimethe␤-layerandoberhautchen 1973,1975),Bursteinetal.(1974),Coleandvan mature,coherentcellboundariesarenotapparent withinthem(Madersonetal.,1998).Thereareoften cell-likeenclosuresvisibleonthesurfaceoftheober- *Correspondenceto:Dr.E.N.Arnold,NaturalHistoryMuseum, hautchen,althoughitisnotcertainthattheseare CromwellRoad,LondonSW75BD,UnitedKingdom. alwaysderivedfromindividualcells.Thesurfaceof E-mail:[email protected] theoberhautchenfrequentlyexhibitsacomplex,mi- Publishedonline00Month0000in croscopical,three-dimensionalstructurefirstnoted WileyInterScience( byLeydig(1872,1873),whichusuallyincludesfea- DOI:10.1002/jmor.1096

©2002WILEY-LISS,INC. 146 E.N. ARNOLD Devender (1976), Sammartano (1976), Gans and Natural History Museum, London. At least three Baic (1977), Gasc and Renous (1980), Groombridge specimens of each were examined. (1980), Perret and Wuest (1982, 1983), Price (1982, The ␤-layer of the epidermis of individual scales 1983, 1989), Peterson (1984a,b), Peterson and Bezy was removed with forceps, washed in 80% alcohol, (1985), Renous et al. (1985), Bea (1986), Bowker et and, in a few cases where it was necessary, cleaned al. (1987), McCarthy (1987), Stille (1987), Bezy and further by brief ultravibration in chloroform. The Peterson (1988), Irish et al. (1988), Vaccaro et al. samples were then dried and mounted with Araldite (1988), Chiasson and Lowe (1989), Lang (1989), on scanning electron microscope stubs. After coating Price and Kelly (1989), Renous and Gasc (1989), with gold, the scales were examined using a Hitachi Harvey (1993), and Harvey and Gutberlet (1995). 2500 scanning electron microscope at 15 kV and at In spite of these extensive studies, no broad assess- magnifications from ϫ35 to ϫ10,000. Dorsal scales ment of the evolutionary factors that may cause the of 95 species were examined (Table 2) as well as development of different patterns of microornamenta- basal caudal scales of 20 species and belly scales of tion has been made. Although convincing functional 15. Estimates of cell size were made by measure- interpretations have been put forward in restricted ment of individual cells on micrographs of sur- instances (see Factors That May Cause Evolutionary faces enlarged ϫ5,000. Micrographs of the examined Change, below), microornamentation does not in gen- scales, together with the accession numbers of the eral correlate closely with known environmental pa- individual from which they came, are depos- rameters (Price, 1982; Peterson, 1984a,b). Nor does it ited in the library of the and Amphibian seem to be a particularly good general indicator of section of the Natural History Museum. relationship, although it may include some phyloge- netic signal (see, for instance, Harvey and Gutberlet, Experimental Assessment of Dirt-Shedding 1995, on cordylid and gerrhosaurid lizards). Ability and Reflectivity Although microornamentation has sometimes been surveyed across whole taxonomic groups (e.g., The relative ability of different microornamenta- Peterson and Bezy, 1985; Lang, 1989; Harvey and tion patterns to shed dirt was tested by painting dry Gutberlet, 1995), sampling is often limited or the detached scales of large-scaled species of lacertids assemblage concerned is small, relatively uniform in with fine wet silt, produced by differential flotation ecology, or both. Lacertid lizards exhibit substantial of garden soil in water. This was allowed to dry and variation in microornamentation and, with some 23 the scale then subjected to controlled gentle wiping genera and 250 species spread over a wide range of with a truncated primary of a house sparrow environments, they provide an opportunity to ex- (Passer domesticus), the number of strokes neces- plore this variation more fully in a historical and sary to clean the scale giving some idea of its ten- functional context. As only a few published descrip- dency to retain dirt. Light scattering by microorna- tions of lacertid microornamentation are available mentation was assessed by directing a narrow (for instance, Lacerta vivipara [Bryant et al., 1967]; parallel beam of light at various angles at the exter- Lacerta viridis [Sammartano, 1976; Peterson, nal surfaces of detached scales of large-scaled lac- 1984a]; Podarcis hispanica [Bowker et al. 1987]), a ertids, which had been glued flat onto a plane sur- systematic survey was undertaken as the first stage face with Araldite, and noting the degree of of the investigation. dispersal of the reflected beam and whether coher- ent shine was produced. MATERIALS AND METHODS Areas of Body Examined Approach to Data Analysis Differences in squamate scale microornamenta- The very wide range of microornamentation en- tion occur not only between taxa but also on differ- countered in lacertids has a complex taxonomic dis- ent parts of the body of individual and even tribution. It is analyzed as follows. 1) Beginning on individual scales (Cole and Van Devender, 1976; with the microornamentation of the dorsal scales, Peterson, 1984b). In consequence, the survey con- variable features are identified and separated into ducted here has been restricted to specified areas of characters with two or more states; the distribution the . Most observations were made on the para- of the states of each character is then plotted on the vertebral mid-dorsal surface of the posterior body, phylogeny of the family. 2) The history of the indi- as this is often typical of a large area of the trunk vidual characters is assessed, including the direc- dorsum, but a smaller number involve the dorsal tail tion and stages of their evolution and the frequent base and the belly. multiple origin of derived states. 3) Possible corre- lations with features of the microornamentation of Specimens and Their Examination dorsal scales are looked for. These correlations may include other aspects of microornamentation, other Material was obtained from alcohol-preserved intrinsic aspects of the species concerned, or envi- specimens in the permanent reptile collection of the ronmental parameters. Even distinctly imperfect SCALE MICROORNAMENTATION OF LACERTID LIZARDS 147 correlations may be significant and do not necessar- the hind margin projects backwards to overlap the ily have to apply just to derived states. Correlations cell (or cells) behind (Fig. 1a; see also longitudinal with other aspects of microornamentation may in- section of Lacerta vivipara scale illustrated by Bry- volve single features or groups of them associated by ant et al., 1967, plate XI). The amount of imbrication some common factor; for instance, being derived is very variable. In some instances, the projecting rather than primitive or possessing a particular margins are set at a steeper upward angle (Fig. 1e) physical characteristic. 4) These various correla- and in these cases are also often particularly exten- tions are used to generate a hypothesis about the sive. The most extreme examples of this condition, possible function of different types of dorsal body for instance, in Poromera (Fig. 1f) and some Gallo- microornamentation. 5) So far as is feasible, the tia, have the raised margins projecting almost per- hypothesis is tested by simple experiment and con- pendicularly from the general cell surface. trolled observation. 6) It is further tested by using it In Nucras boulengeri and N. tessellata the areas of to make predictions about the kind of microorna- contact between adjoining cells are depressed to mentation to be expected on other areas of the body form grooves (Fig. 2a). As noted above, the borders of surface of lacertids living in particular situations, polygonal cells in particular may be raised into welts and then checking to see if the predictions hold. This or ridges, for example in Heliobolus spekii, Ichnot- kind of testing is also extended to other taxa. 7) ropis (Fig. 1c), Pseuderemias (Fig. 1d), some Pedio- Observed apparent violations of the predictions of planis, and Ophisops (Fig. 1b). the hypothesis are considered to see if they can be Denticulation of posterior cell margins. The reconciled with it. posterior borders of strap-shaped cells are some- times rather wavy or slightly notched and, in a few RESULTS cases, where the borders are steeply angled or per- Dorsal Scales of Body pendicular they are denticulated. The denticula- tions may be rather sparse and irregular (Algyroides The variations in microornamentation encoun- moreoticus, Fig. 2b), or may form very distinct tered on lacertid dorsal body scales (Figs. 1–3) can groups, being coordinated in succeeding cells to form be resolved into differences in a few main characters tracts running mainly anteroposteriorly along the that often vary independently. This variation is dis- scale (Poromera fordi, Fig. 1f ) In other cases den- cussed below and the characters and their different ticulations are abundant and widespread, forming a states are listed in Table 1. Differences in these field of spikes ( stehlini, Fig. 2c). features between species examined in this study are Detailed structure of cell surface. This is not shown in Table 2. always easily seen, especially where strap-shaped Cell shape. As noted, boundaries on the ober- hautchen surface may not necessarily represent the cells have long, almost perpendicular posterior margins of actual cells, but units with continuous edges. When visible, the cell surface frequently ap- perimeters made up of welts or grooves are usually pears quite smooth (Fig. 1a.), even at magnifications referred to by that term in discussions of microorna- of ϫ8,000 or more, but in other cases it exhibits an mentation and that convention will be followed here. array of pits that are often about 0.5 ␮m in diame- Such cells are frequently visible in the microorna- ter. In some instances these pits are shallow and mentation of lacertids and are often narrow and scattered; for instance, in Nucras (Fig. 2a) and Phi- strap-shaped, their longer axes running trans- lochortus (Fig. 2d), but in many taxa they are rather versely to the main, approximately anteroposterior larger, irregular, and densely packed (Fig. 1b–d). axis of the scale (Fig. 1a). In a minority of cases, the When this is so, the surface keratin may be reduced microornamentation includes a reticulation of welts to a filigree, with the enclosed cavities making up or ridges enclosing polygonal areas (Fig. 1b–d) that about half the total cell surface (Fig. 2e). have also frequently been termed cells (see Problems Longitudinal ridges. In scales with strap- of , pg. 7). shaped cells, the epidermal surface may be rucked to Profile of cell surface. The surface of a cell is often produce ridges that run essentially longitudinally, more or less flat (Fig. 1a) but in polygonal ones it may either roughly parallel to the main axis of the scale be centrally depressed so that it is dished (Fig. 1b,c). or converging posteriorly (some Takydromus, Fig. Cell dimensions. Strap-shaped cells vary from 3a) or diverging in this direction (vertebral scales of 1–4 ␮m in the length of their shorter axes. In polyg- Philochortus hardeggeri, Fig. 3c). The ridges are onal cells this axis is about 4–6 ␮minHeliobolus usually quite long (Fig. 3a), but may sometimes be spekii, inornata, P. namaquensis and short (Fig. 3c) and may anastomose with each other. Ophisops jerdoni, about 10 ␮minP. lineoeoocellata, At lower magnifications, Poromera appears to have P. inornata, P. undata, and O. elbaensis, and 15–20 ridges arranged similarly to those in Takydromus ␮minIchnotropis, Pseuderemias mucronata, P. stri- (Fig. 3a,b), but in reality these are tracts of denticu- ata, and Pedioplanis rubens. lations (Fig. 1f). Cell margins. In strap-shaped cells, the cell sur- Pustular projections. Strap-shaped cells may face often slopes slightly upwards posteriorly and be interrupted by large pustular projections, a fea- Fig. 1. Microornamentation on dorsal scales of lacertid lizards. a: Lacerta monticola cantabrica (ϫ4,000). b: Ophisops jerdoni (ϫ3,500). c: capensis (ϫ4,000). d: Pseuderemias mucronata (ϫ2,000). e: montana (ϫ5,000). f: Poromera fordi (ϫ2,000). Anterior of scale is in the direction of upper left corner of photograph a, to the right in e, and to the left in the remainder. Scale bar ϭ 5 ␮m. Fig. 2. Microornamentation on scales of lacertid lizards; a, b, d, and e show dorsal body scales. a: Nucras boulengeri (ϫ5,000). b: Algyroides moreoticus (ϫ4,000). c: Gallotia stehlini (ϫ4,000), proximal tail scale. d: Philochortus spinalis (ϫ3,500). e: Pseuderemias mucronata (ϫ15,000). f: Pseuderemias mucronata (ϫ1,500), ventral body scale. Anterior of scale is in the direction of the bottom left corner of photograph a, bottom right corner in photograph b, top left corner in photograph d, and to the left in the remainder. Scale bar ϭ 5 ␮m. 150 E.N. ARNOLD

Fig. 3. Microornamentation on dorsal scales of lacertid lizards. a: Takydromus septentrionalis (ϫ200). b: Poromera fordi (ϫ200). c: Philochortus hardeggeri (ϫ350). d: Algyroides nigropunctatus (ϫ400). Anterior of scale is in the direction of top left corner of photograph, in a and d where it is towards upper margin in b, and towards bottom right corner in c. Scale bar ϭ 50 ␮m. ture only encountered in Algyroides (Figs. 2b, 3d) blage there are differences, but these are relatively and Adolfus africanus. slight in the studied cases. For instance, L. monti- Most derived features involve increased sculptur- cola has the edges of the strap-shaped cells more ing and roughening of the scale surface relative to raised on the tail scales than on the body, and in the 0 states in Table 1. Gallotia stehlini denticulations become more florid. Similarity of the microornamentation of the scales of Dorsal Scales on Tail Base the dorsal body and tail also occurs in a scattering of forms among the more derived members of the Er- Among many Lacertinae and the more basal Er- emiainae, including Philochortus spinalis, Pseuder- emiainae (see Fig. 5), microornamentation on the emias striata, , and Acanthodac- large dorsal scales of the proximal part of the tail is tylus haasi. However, substantial differences similar to that present on the dorsal body scales. between body and tail scales are common here. In This is true in Lacerta oxycephala, L. praticola, Nucras boulengeri, N. tessellata, Heliobolus speki, L. dugesii, Podarcis taurica, Adolfus africanus, and H. lugubris the non-overlapping strap-shaped A. vauereselli, Tropidosaura cottrelli, T. essexi, and cells on the dorsal body scales are replaced on the Poromera fordi. In some other forms in this assem- tail by ones in which the posterior imbrications are SCALE MICROORNAMENTATION OF LACERTID LIZARDS 151 TABLE 1. Characters and states of dorsal body scale ined. In and Pedioplanis husa- microornamentation in the Lacertidae bensis the main surface of the scale is quite smooth, 1. Cell shape. 0 ϭ narrow, transversely strap-shaped; 1 ϭ not without visible cell boundaries. narrow, polygonal. Surface structure may be different on the ante- 2. Profile of cell surface. 0 ϭ more or less flat; 1 ϭ dished, the rior, basal region of a ventral scale and sometimes center lower than the edges. also on the posterior margin of the scale where it 3. Cell dimensions, anteroposterior length. 0 ϭ 2–10 ␮m; 1 ϭϾ10 ␮m; 2 ϭϽ2 ␮m. States 1 and 2 appear to be curves upwards. Ichnotropis capensis and Pediopla- independently derived from state 0. nis husabensis have polygonal cells in the basal re- 4. Cell margins. 0 ϭ hind margins posteriorly imbricate with gion and Ophisops elegans strap-shaped, overlap- at most a shallow upward angle; 1 ϭ hind margins ping ones. posteriorly imbricate with steep upward angle; 2 ϭ no clear imbrication, borders between cells are grooves; 3 ϭ no clear imbrication, borders forming raised welts. States 1, 2, and 3 appear to be independently derived from state 0. DISCUSSION 5. Denticulation of posterior cell margins. 0 ϭ margins more or Problems of Homology less smooth; 1 ϭ margins with some denticulation; 2 ϭ denticulation abundant and widespread, forming tracts Harvey 1993, and Harvey and Gutberlet 1995; or a field of spikes. gives reasons why strap-shaped cells in squamate 6. Detailed structure of cell surface. 0 ϭ smooth; 1 ϭ scattered microornamentation may not be homologous with pits; 2 ϭ pits densely packed, often comprising around half the scale surface. the polygonal units that are also often called cells 7. Longitudinal ridges. 0 ϭ absent; 1 ϭ present. but that he terms “macro-honeycomb.” This author 8. Pustular projections. 0 ϭ absent; 1 ϭ present. points out that in some xenosaurid lizards, where the borders of the polygonal units are raised and the 0 usually indicates the primitive state within the family based on its distribution in the family itself and in outgroups. units themselves dished, the ridge-like borders in- volve not only the oberhautchen but also the ␤-, mesos, and ␣-layers of the epidermis underlying it. Furthermore, in Xenosaurus grandis agrenon strap- strongly raised and there is increased pitting on the shaped cells overlie a larger polygonal pattern of scale surfaces. In and ridges. In this case the two elements thus appear to P. rubens, polygonal cells are present on the tail as fail the conjunction test for homology (Patterson, well as the body, but are concave rather than convex 1982). in this situation. An increase in cell concavity also In lacertids, no cases have been encountered of occurs on the tail of Ophisops elbaensis.InAcantho- discrete strap-shaped cells and polygonal structures dactylus micropholis and A. robustus, the imbricate being superimposed, and intermediate conditions posterior edges of the strap-shaped cells are more occur between strap-shaped cells and deeply dished raised on the tail scales. polygonal units. In some species there may be gra- dation on the same body scales from relatively strap- Belly Scales shaped cells to polygonal ones (Pedioplanis nama- quensis), and from flat polygonal units on the dorsal The large, smooth, shiny belly scales of most lac- body to dished polygonal units on the dorsal tail ertids are often heavily scratched and no cell mar- base (P. rubens). The fact that, where the borders of gins may be apparent on their surface. However in polygonal units are raised into ridges, other deeper Takydromus and some Lacertinae and basal Eremi- layers of the epidermis may be involved does not ainae, the pattern common on dorsal scales of strap- automatically negate homology with borders where shaped cells with low overlapping posterior borders this is not the case, especially if there are interme- and smooth surface is also present on the belly diates between the two conditions. Because of this, scales, the cells having an anteroposterior length of strap-shaped and polygonal units in lacertid lizards about 2–5 ␮m. This was observed in Takydromus are regarded here as provisionally homologous. In kuehnei, Gallotia galloti, Lacerta monticola, Adolfus Ichnotropis squamulosa (Fig. 1c) the polygonal net- africanus, and Tropidosaura cottrelli and is also re- work of raised ridges is accompanied by a staggered ported in Podarcis hispanica (Bowker et al., 1987). polygonal system of thin, unraised lines. Similar In contrast, several of the more terminal members of patterns of thin lines in other lizards have been the Eremiainae have a pattern on the main surface interpreted as resulting from the cell borders of the of the ventral scales, consisting of large polygonal clear layer that separates the oberhautchen from the cells with simple abutting margins. It is found in previous exuvium during development (Stewart and Pseuderemias mucronata (Fig. 2f), Acanthodactylus Daniel, 1972; see also discussion by Irish et al., 1988). scutellatus, and Ophisops elegans and also occurs in boscai, where the cells are less expanded anteroposteriorly than in other cases and are lightly History of Dorsal Microornamentation pitted. Polygonal cells on lacertid ventral scales are Relationships within Lacertidae. Recent in- about 10 ␮m in length anteroposteriorly in Latastia vestigations of the morphology (Arnold, 1989a,b, boscai and around 20 ␮m in the other species exam- 1991) and mitochondrial DNA sequence (Harris et 152 E.N. ARNOLD

TABLE 2. Variation in microornamentation in 95 species of lacertid lizards 1234 5 6 7 8 9 Cell Cell Cell Cell Cell Long. Scale shape profile size margins Denticulate texture ridges Pustules size hispanicus 0 — 21 1 — 00? 0 — 01 0 1 1 0 1 Gallotia atlantica 0 — 21 1 0 0 0 0 Gallotia galloti 0 — 21 1 — 000 Gallotia stehlini 0 — 21 2 — 000 Lacerta vivipara 0 0 0 — 0 — 001 Lacerta lepida 0 0 0 0 0 0 0 0 0 Lacerta princeps 0 0 0 0 0 0 0 0 0 Lacerta agilis 0 0 0 0 0 0 0 0 0 Lacerta monticola 0 0 0 0 0 0 0 0 0 Lacerta mosorensis 0 0 0 0 0 0 0 0 1 Lacerta oxycephala 0 0 0 0 0 0 0 0 0 Lacerta brandti 0 0 0 0 0 1 0 0 0 Lacerta praticola 0 0 0 0 0 0 0 0 0 Lacerta graeca 0 0 0 0 0 0 0 0 0 Lacerta cappadocica 0 0 0 0 0 1 0 0 0 Lacerta laevis 0 0 0 1 0 0 0 0 0 Lacerta fraasi 0 0 2 0 0 0 0 0 0 Lacerta parva 0 0 2 1 0 0 0 0 0 Lacerta andreanszkii 0 0 2 0 0 0 0 0 0 Lacerta dugesii 0 0 0 0 0 0 0 0 0 Podarcis hispanica 0 0 0 0 0 0 0 0 0 Podarcis melisellensis 0 0 0 0 0 0 0 0 0 Podarcis taurica 0 0 0 0 0 0 0 0 0 Algyroides nigropunctatus 0 — 21 0 0 0 1 1 Algyroides moreoticus 0 0 2 1 1 0 0 1 1 Algyroides fitzingeri 0 — 21 0 0 0 1 1 Algyroides marchi 0 0 2 1 0 0 0 1 1 Takydromus amurensis 0 — 21 0 — 101 Takydromus septentrionalis 0 0 0 1 0 ? 1 0 1 Takydromus sexlineatus 0 0 0 1 0 2 1 0 1 Takydromus kuehnei 0 0 0 1 0 — 101 Takydromus toyamai 0 0 0 0 0 0 1 0 1 Australolacerta australis 0 0 0 0 0 1 0 0 0 Omanosaura cyanura 0 0 0 0 0 0 0 0 0 Omanosaura jayakari 0 0 0 0 0 1 0 0 0 Adolfus jacksoni 0 0 0 0 0 0 0 0 0 Adolfus alleni 0 0 0 1 0 1 0 0 1 Adolfus africanus 0 0 0 1 0 1 0 1 1 Adolfus vauereselli 0 0 2 0 0 0 0 0 1 Holaspis guentheri 0 0 0 0 0 0 0 0 1 Gastropholis echinata 0 0 0 1 0 1 1 0 0 Gastropholis tropidopholis 0 0 0 0 0 1 1 0 1 Gastropholis vittata 0 0 0 0 0 1 1 0 1 0 — 21 0 0 0 0 1 Tropidosaura gularis 0 0 2 1 0 0 0 0 1 Tropidosaura cottrelli 0 0 2 1 0 0 0 0 1 Tropidosaura essexi 0 0 2 1 0 0 0 0 1 Poromera fordi 0 — 21 2 0 ? 0 1 Nucras boulengeri 0 0 0 2 0 1 0 0 0 Nucras tessellata 0 0 0 2 0 1 0 0 0 Nucras lalandei 0 0 0 0 0 1 0 0 0 Philochortus hardeggeri 0 0 0 0 0 1 1 0 1 Philochortus spinalis 0 0 0 0 0 2 0 0 0 Latastia longicaudata 0 0 — 00— 000 Latastia johnstoni 0 0 0 0 0 2 0 0 0 Latastia neumanni 0 0 0 0 0 2 0 0 0 Heliobolus nitida 0 0 0 0 0 2 0 0 0 Heliobolus speki 1 0 0 3 0 2 0 0 0 Heliobolus lugubris 0 0 0 0 0 2 0 0 0 Ichnotropis capensis 1 1 1 3 0 2 0 0 1 Ichnotropis squamulosa 1 1 1 3 0 2 0 0 1 Pseuderemias mucronata 1 0 1 3 0 2 0 0 0 Pseuderemias striata 1 0 1 3 0 2 0 0 0 Meroles knoxi 0 0 0 0 0 — 000 Meroles suborbitalis 0 0 0 0 0 1 0 0 0 Meroles ctenodactylus 0 0 0 0 0 1 0 0 0 SCALE MICROORNAMENTATION OF LACERTID LIZARDS 153

TABLE 2. (Continued) 1234 5 6 7 8 9 Cell Cell Cell Cell Cell Long. Scale shape profile size margins Denticulate texture ridges Pustules size Pedioplanis lineoocellata 1 1 1 3 0 2 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 1 0 0 0 Pedioplanis laticeps 0 0 0 3 0 1 0 0 0 Pedioplanis inornata 1 0 0 3 0 2 0 0 0 Pedioplanis namaquensis 1 0 0 0 0 2 0 0 0 Pedioplanis rubens 1 0 1 3 0 2 0 0 0 1 0 — 30— 000 Eremias velox 0 0 0 0 0 0 0 0 0 Eremias fasciata 0 0 0 0 0 1 0 0 0 Eremias grammica 0 0 0 0 0 1 0 0 0 Acanthodactylus boskianus 0 0 0 1 0 — 00? Acanthodactylus micropholis 0 0 0 0 0 1 0 0 0 Acanthodactylus schmidti 0 0 0 0 0 2 0 0 0 Acanthodactylus tilburyi 0 0 0 0 0 1 0 0 0 Acanthodactylus haasi 0 0 0 0 0 1 0 0 0 Acantho. gongrorhynchatus 0 0 0 0 0 1 0 0 0 Acanthodactylus robustus 0 0 0 0 0 0 0 0 0 Acanthodactylus erythrurus 0 0 0 0 0 1 0 0 0 Acanthodactylus pardalis 0 0 0 0 0 — 000 Acanthodactylus scutellatus 0 0 0 0 0 1 0 0 0 Mesalina balfouri 0 0 0 0 0 1 0 0 0 Mesalina rubropunctata 0 0 0 0 0 1 0 0 0 Ophisops elegans 0 0 2 1 0 0 0 0 1 Ophisops leschenaulti 0 0 0 1 0 — 101 Ophisops elbaensis 1 1 1 3 0 2 0 0 1 Ophisops jerdoni 1 1 1 3 0 2 0 0 1 Ophisops minor 0 0 2 1 0 — 101 Teiids 1 0 2 3 0 ? ? ? ? States of characters given in columns 1–8 are listed in Table 1. Column 9 indicates size of dorsal scales (0 small, 1 large, see section on Nonancestral Resemblance). Dashes indicate lack of data. Data for Holaspis and Philochortus, refer to the enlarged vertebral body scales. al., 1998) of lacertid lizards indicate that principal conditions in outgroups clearly resolve the polarity relationships within the family are as shown in Fig- of these features, which include small cell size, pos- ure 5. The Gallotiinae occur in the West Mediterra— terior cell edges markedly raised, and these edges nean area and the Canary Islands, the Lacertinae denticulated. This is because microornamentation is principally in the West Palaearctic, and the Eremi- very varied in some of the Scleroglossan groups con- ainae mainly in the Afrotropical region, but with cerned and not all the interrelationships of these four terminal genera in North Africa and southwest groups are unequivocally resolved; for instance, the and central Eurasia. The precise position of Takydro- schemes of Estes et al. (1988) and of Lee (1998) mus of East Asia relative to the last two groups is exhibit significant differences. It is, however, gener- unresolved. Many species occur in generally mesic ally agreed that the sister-group of the Lacertidae is habitats but Psammodromus and the more derived the Teiioidea, which often exhibit large cell size and members of the Eremiainae (Nucras and its apparent posterior cell edges that are not markedly raised or sister group) are mainly found in more arid situations. denticulated (Stewart and Daniel, 1975; Peterson, Phylogenetic distribution. The MacClade pro- 1984a; Peterson and Bezy, 1985; Vaccaro et al., gram (Maddison and Maddison, 1996) was used to 1988). The last two features are uncommon in estimate character states on internal branches of Scleroglossans as a whole, although they are fre- the lacertid phylogeny by parsimony (see Figs. 4–9). quent in the Scincidae. Given that pitting occurs in This indicates that many states were derived within the Teiioidea, it is most parsimonious to consider the family, including polygonal cells, cell dishing, low levels of pitting as the primitive state in the large cell size, borders consisting of grooves or welts, Lacertidae if detailed structure of the cell surface is high levels of pitting, longitudinal ridges, and pus- treated as an ordered character, while polarity is tular projections. unresolved if it is not. With some other traits, it is not clear from their It appears from these considerations that the prim- distribution in the Lacertidae that they always itive dorsal microornamentation pattern for the Lac- arose within the family, although they did so on ertidae is likely to be a combination of strap-shaped occasion and are also usually minority states. Nor do cells of intermediate size with imbricate posterior bor- 154 E.N. ARNOLD

Fig. 4. Left. Phylogenetic distribution of strap-shaped cells (0) and polygonal cells (1). Right. Phylogenetic distribution of dished cell surfaces (1). Here and in Figures 5–9, 12 and 13 not all the species in Table 2 are included. ders that are not sharply raised, and lightly pitted or found in lacertids have evolved more than once and smooth cell surfaces in which denticulation, large often on several occasions. For instance, if the polar- ridges, and pustules are absent (Fig. 1a). ities assumed for the Lacertidae are accepted, strong The primitive lacertid pattern also occurs widely, pitting of cell surfaces has originated 3–4 times, although sporadically, in other lizards, being found in small cells 8–10, and raised posterior cell edges basal xantusiids (Stewart and Daniel, 1975; Peterson 9–12. For the 13 different forward transformations and Bezy, 1985), gerrhosaurids, the cordylid Platysau- listed in Table 1, there are a total of 43–61 indepen- rus (Harvey and Gutberlet, 1995), and some anguids dent cases. In the more restricted context of the (Gerrhonotus multicarinatus [Stewart and Daniel, Eremiainae, where polarity for most features can be 1973]). Scale microornamentation of Sphenodon, the determined without recourse to groups outside the living outgroup of the squamates, is also reported to be Lacertidae, high levels of parallelism are again ap- similar (Peterson, 1984a), although this has been dis- parent (Table 3). puted (Maderson et al., 1998). Given that only minimal numbers of origins can Use of MacClade also helped determine transforma- be estimated (since estimates are based on a parsi- tion series in multistate characters (Table 1). Espe- mony analysis and not all lacertid taxa are includ- cially large and especially small cells appear to have ed), this degree of parallelism is very striking. It arisen from the more widespread intermediate condi- suggests that, in developmental terms, the features tion occurring in the primitive microornamentation. may be quite easily produced. This could be partly a Similarly, raised imbricate posterior cell margins, characteristic of small-scale structures, where there grooves between cells, and borders forming ridges or is a high degree of self-organization, and possible welts between deeply dished cell surfaces all seem to developmental paths are quite limited. Perhaps the have been derived independently from the primitive different derived features may be “switched on” by condition. In these cases decisions were supported by natural selection if they promote performance ad- other clues as to proximity of states in transformation vantage in particular situations, but there seem to sequences, including relative similarity of states and be limits to the variations that can be produced. their occurrence together, often with intermediates, in Although conditions on some internal branches the same individuals. are equivocal, with parallelism and reversal both being possibilities, in the great majority of microor- Nonancestral Resemblance in Dorsal namentation features there is no overt evidence of Body Microornamentation reversal to primitive states taking place within the Lacertidae. The principal exception is the presence It can be seen from Figures 4–9 and Table 3 that of strong pitting on the scale surface, for which par- most of the derived features of microornamentation simony indicates at least one reversal leading to SCALE MICROORNAMENTATION OF LACERTID LIZARDS 155

Fig. 5. Phylogenetic distribution of large (1) and small (2) cells. Allocation of taxa to the three subfamilies of the Lacertidae is also shown. more primitive states in one or more members of Among main squamate clades there may be gen- Pedioplanis, Eremias, Acanthodactylus, Mesalina, eral trends in microornamentation. Thus, iguanians and Ophisops (Fig. 8). Reversal may possibly also and gekkotans typically have polygonal cells, while occur in the development of dished cell surfaces (see strap-shaped ones are usual in scincids (Perret and Lack of reversal after loss of function?, below). Such Wuest, 1982, 1983). But even at this taxonomic an asymmetry in frequency between parallelism and level, there may be striking parallels across large reversal may also occur outside the Lacertidae, as systematic distances. The microornamentation pat- absence of reversal in microornamentation has been tern found in Ichnotropis (Fig. 1c) is very similar to suggested in the phrynosomatid Sceloporus that occurring in some iguanians, such as Polychrus (Burstein et al., 1974). (Peterson, 1984b; Fig. 1c). Subdigital setae, which In spite of the rarity of overt reversal, the high are another form of microornamentation, are wide- incidence of homoplasy that exists makes microor- spread in gekkotans but also occur in the iguanian namentation overall a poor indicator of relationship genus Anolis and in one species of the autarchoglos- among lacertids, at least at higher taxonomic levels, san scincid genus Prasinohaema (Williams as is true in other taxa. Sometimes, however, par- and Peterson, 1982). Such close derived resem- ticular groups of lacertids are characterized by rel- blances among very distant relatives supports the atively rare features that distinguish the assem- hypothesis that developmental pathways for micro- blages concerned from their closer relatives. Such ornamentation may be restricted, and the discovery cases include substantial denticulation of posterior of just one case of digital setae among the 1,700 or so cell margins in Gallotia and the presence of pustules species of the autarchoglossan lizards warns against in Algyroides. making sweeping statements about the limits to the 156 E.N. ARNOLD

Fig. 6. Phylogenetic distribution of raised posterior cell margins (1), groove-like cell borders (2), and welt- or ridge-like borders (3). kinds of microornamentation that can occur in par- and the different derived features may appear in a ticular groups. range of sequences (see Fig. 11). In some in- stances, one feature may occur before another on a Patterns of Association and Order of particular lineage, while appearing after it on an- Change in Dorsal Body Ornamentation other one. The lack of a simple pattern of order of change in the features of microornamentation in- The principal derived features of lacertid microor- dicates that they are developmentally substan- namentation occur together in two main groupings tially independent. (Fig. 10). In one, microornamentation may include small cell size, raised posterior cell edges, denticu- Comments on Microornamentation of lation, longitudinal ridges, and pustules. In the Dorsal Tail Base and Belly other, possible components are polygonal cells, dished cell surfaces, large cell size, welt-like borders In general, the widespread primitive pattern of and dense pitting. microornamentation of the dorsal body scales is also Derived features of up to five characters of lac- found on the dorsal tail scales of many of the more ertid microornamentation may sometimes occur basal forms of the Lacertinae, while Gallotia and together. If derived features in general were as- members of the Eremiainae have more derived con- sembled in a fixed order during evolution, it would ditions on the tail. Where it is different from the be possible, given the degree of homoplasy microornamentation on the dorsal body scales, that present, to recognize repeated linear sequences of on the tail base may be more derived or less so. It is varying length resulting in increasingly complex more derived where the edges of strap-shaped cells microornamentation. In fact, this is not the case are raised in Lacerta monticola and Acanthodacty- SCALE MICROORNAMENTATION OF LACERTID LIZARDS 157

Fig. 7. Phylogenetic distribution of denticulated raised posterior cell margins: weak (1) and strong (2). lus and where the edges are more denticulated, as in nal, and polygonal ones where the main surface is Gallotia stehlini. In contrast, it is more primitive in featureless. The basal areas of ventral scales are the species of Nucras, Heliobolus, and Pedioplanis less exposed to the environment than the main sur- examined. In all these cases where microornamen- face, since they are protected by the posterior imbri- tation on the tail scales is different from that on the cation of the edges of the scales lying immediately in body, it is also more sculptured and three- front of them. They are therefore presumably likely dimensional. to be under less pressure to change in response to Distinctive ornamentation at the base of the belly alterations in external aspects of the selective re- scales found in some lacertids has also been noted on gime. the dorsal scales of (McCarthy, 1987; Price Frequently, the predominant microornamentation and Kelly, 1989). In adult snakes, the basal pattern on the ventral scales is different from that of the is often more primitive than that on the greater part dorsals, being often simpler and smoother. However, of the scale and, in cases where there is ontogenetic in Lacerta monticola and Gallotia galloti the primi- change in the latter region, the basal area tends to tive lacertid pattern of microornamentation that oc- remain more similar to the condition in neonates curs on their dorsal scales is also present on the (Price and Kelly, 1989). It is not known whether the ventrals and, in a few forms where the dorsal scales latter phenomenon is found in lacertid ventral scales have complex derived microornamentation, this too but, in the few cases where a different basal pattern is repeated on the belly scales, something that oc- has been observed, it is indeed primitive relative to curs in Poromera fordi, some Takydromus, and Gas- that found on the main surface of the scale. Thus, tropholis tropidopholis (see below). the scale base has overlapping strap-shaped cells in Dorsal and ventral scale microornamentations cases where those on the main surface are polygo- can clearly evolve independently of each other. Dor- 158 E.N. ARNOLD

Fig. 8. Phylogenetic distribution of lightly pitted (1) and heavily pitted (2) cell surfaces. The primitive condition may be either lightly pitted or smooth surfaces. sal microornamentation is more derived than ven- advanced pattern on the dorsal scales in G. tropi- tral in Takydromus kuehnei, Psammodromus algi- dopholis, but not in other members of the genus, rus, Adolfus africanus, Tropidosaura essexi, and including the more basal G. echinata (relationships Pseuderemias mucronata, while the pattern most discussed by Arnold, 1989b). widely distributed on each ventral scale is more advanced than that on the dorsals in Latastia boscai, Ichnotropis capensis, Pedioplanis husaben- Anatomical and Environmental Correlations sis, and Ophisops elegans. In some cases where dor- Association with large scale size. Distribution sal and ventral microornamentation is similar, phy- of various aspects of microornamentation was com- logenetic information indicates that they have pared with that of scale size. Where dorsal scaling is reached their final condition at different times. This homogeneous in lacertids, the number of scales in a can be seen in the genus Takydromus (phylogeny transverse row at mid-body is roughly correlated discussed by Arnold, 1997). As noted above, dorsal with relative scale size. Low mean transverse counts microornamentation is more derived than that on of about 18–40 were therefore taken as indicating the ventral scales in Takydromus kuehnei, a rela- large scales. In other cases, like Holaspis and Phi- tively primitive member of its genus, but ventral lochortus, where the scales of the vertebral region microornamentation has advanced to match that on are much bigger than the remaining more lateral the dorsal scales in more terminal members of the dorsals, large size was judged by direct inspection. group, such as T. toyamai, T. sauteri, and indepen- Distribution of large scale size in the Lacertidae is dently in the lineage leading to T. sexlineatus. The shown in Figure 12 and the number of cases where same phenomenon is also found in Gastropholis, particular derived scale ornamentation features are where the ventral microornamentation matches the associated with large scale size in Table 3. SCALE MICROORNAMENTATION OF LACERTID LIZARDS 159

Fig. 9. Phylogenetic distribution of longitudinal ridges underlying microornamentation (1) and of pustular projections (2).

Concentrated changes tests were carried out us- and large scale size on internal branches of the ing the MacClade program (Maddison and Maddi- lacertid phylogeny. The test can give only a rough son, 1996) to assess the association between the relative idea of the likelihood of correlations being appearance of derived microornamentation features due to chance because not all lacertids could be

TABLE 3. Number of instances in which particular derived microornamentation features have apparently evolved separately Number overtly arising with or Number of separate origins after large scale size Lacertidae Eremiainae 1. Cells polygonal (Fig. 4) 2–72–72–3 2. Cell surface dished (Fig. 4) 3–43–42–3 3. Cells dimensions (Fig. 5) large 4 4 2–3 small 8–10 4 6 4. Cell edges (Fig. 6) raised posteriorly 9–12 6–76 forming grooves 1 1 0 forming welts 2–72–72–3 5. Cell edges denticulated (Fig. 7) 3–41 2 6. Cell surfaces strongly pitted (Fig. 8) 3–42–31–2 7. Longitudinal ridges (Fig. 9) 6 4 6 8. Pustules (Fig. 9) 2 2 2 160 E.N. ARNOLD

Fig. 10. Association of derived states in microornamentation. Thick lines join two states that occur together in at least five or more cases, thin lines there that occur in just one or two. Small figures show number of instances. States mainly associate in two groups: left group includes small cell size (3/2), raised posterior cell edges (4/1), denticulation (5), longitudinal ridges (7), and pustular projections (8); right group includes polygonal cells (1), dished cell surfaces (2), large cell size (3/1), welt- or ridge-like borders (4/3), and heavy pitting (6). included in the test. Also, because the test cannot be The notional probability of chance correlation carried out when there are polytomies in the phy- with large scale size was also assessed for cases in logeny, these had to be arbitrarily resolved. which a number of derived microornamentation fea- Bearing these limitations in mind and including tures have evolved (Fig. 13). For four or more fea- the species shown in Figures 5–9, the probability of tures, three or more, and two or more, the probabil- chance correlation with large scale size is Ͻ0.02 for ities are Ͻ0.02. This indicates that there may well small cells, raised edges, and longitudinal ridges, be a correlation between more derived microorna- and Ͻ0.1 for dished cell profile, large cells, denticu- mentation and scale size in lacertids. An association lation of posterior cell margins, strong surface pit- between complex derived patterns of microornamen- ting, and pustular projections. The rather greater tation and large scale size has also been noted infor- probability of chance correlation in these latter mally in xantusiid lizards (Bezy and Peterson, cases may partly result from the features concerned 1988). having relatively few origins. The probability of Association with environment. Apparent cor- chance correlation was substantially greater in the relations also occur between dorsal patterns of mi- case of polygonal pits cell shape and welt-like cell croornamentation and the general nature of the hab- margins. itats occupied by the species concerned. The

Fig. 11. Order of origin of derived states in microornamentation. Lines connect derived states that appear sequentially on at least one lineage; arrows point towards the state that arises later. Clearly, there is no single sequence of state assembly. Derived states: cells polygonal (1), cell surfaces dished (2), cell size large (3/1), cell size small (3/2), posterior cell margins raised (4/1), posterior cell margins welt- or ridge-like (4/3), denticulation (5), heavy pitting (6), longitudinal ridges (7), pustules (8). SCALE MICROORNAMENTATION OF LACERTID LIZARDS 161

Fig. 12. Phylogenetic distribution of large dorsal scales (1). (see Nonancestral Resemblance in Dorsal Body Microornamentation). primitive pattern is predominant in the essentially been suggested that interdigitation between the mesic habitats of the Palaearctic and in the less arid sculptured surface on the newly matured ␤-layer parts of Africa, although it also occurs at lower fre- and the clear layer above it may be beneficial in quencies elsewhere. Among derived states, strongly preparing the old overlying epidermis for shedding raised posterior cell edges, their denticulation, lon- and in holding the immature epidermis together gitudinal ridges, and pustules are all found in liz- while it is developing (Maderson, 1966, 1970; Mad- ards occupying relatively mesic situations with some erson et al., 1998). As all squamates shed their pre- vegetation, particularly in climbing forms. The two vious epidermis, such a hypothesis would not ex- occasions where pustules have evolved, in Algy- plain the extensive interspecific variation found in roides and Adolfus africanus, are associated with microornamentation. In fact, this particular sugges- occupation of forest-floor habitats (Arnold, 1987, tion is not fully tenable as a general functional ex- 1989b). The remaining derived states are typical of planation of microornamentation because ecdysis more xeric conditions. Because no objective classifi- takes place successfully even where the epidermal cation of lacertid habitats exists, it is not possible at surface is very smooth, such as on the body scales of present to formally test these associations. laticaudine sea snakes (McCarthy, 1987) and on large areas of the belly scales of some lacertids. It Factors That May Cause has also been proposed that some aspects of lizard Evolutionary Change microornamentation may increase the mechanical strength of the ␤-layer or parts of it (Ruibal and A range of hypotheses about possible benefits of Ernst, 1965; Maderson et al., 1998) and that it func- microornamentation have been put forward. It has tions as an aid to capturing, dispersing, and retain- 162 E.N. ARNOLD

Fig. 13. Phylogenetic distribution of microornamentations with different numbers of derived features. ing pheromones (Smith et al., 1982). Again, these croornamentation, perhaps helping to make the skin hypotheses do not explain the great variability in waterproof (Chiasson and Lowe, 1989). Digital setae structure involved. in many geckoes and some other lizards facilitate Taxonomically more restricted performance ad- adhesion while climbing. No such very specialized vantages of squamate microornamentation patterns uses of microornamentation are apparent in the La- have been suggested and, in some cases, demon- certidae. As they have limited distributions in squa- strated. Particular patterns were thought to encour- mates as a whole, and are sometimes also confined age transport of water contacting the skin towards to restricted areas of the body, these uses do not the mouth in the agamid, Moloch (Bentley and explain the bulk of variation in the surface structure Blumer, 1962) but movement was subsequently of scales. found to take place in capillary channels between In most studies involving varied microornamenta- the scales (Gans et al., 1982). The smooth scales of tion, no obvious broad correlations with environ- laticaudine sea snakes may reduce the possibility of mental parameters have been discerned that would the skin being colonized by marine algae and other suggest function. Such lack of simple correlations organisms (McCarthy, 1987) and the very rough does not, of course, imply absence of performance scale surfaces on the tail of uropeltid snakes encour- advantage for the different patterns. All sorts of ages the accumulation of a plug of earth which helps factors, singly or together, may prevent such associ- prevent predators following the snakes into their ations making themselves apparent. The precise mi- burrows (Gans and Baic, 1977). Some partly aquatic croornamentation pattern may be influenced by his- natricine snakes in the genera Nerodia and Tham- torical factors and possible cases are discussed nophis have pores on their dorsal body scales that below, but previous events are unlikely to be respon- exude lipids that collect in hollows in the scale mi- sible for the changes themselves. Alteration may SCALE MICROORNAMENTATION OF LACERTID LIZARDS 163 sometimes possibly be caused by particular occur- most and of gerrhosaurids, and also of the rences, but patterns may persist even though the cordylid Platysaurus (Harvey and Gutberlet, 1995), factors that caused their appearance no longer act. which frequently retreats into very narrow crevices. Again, different factors may conflict, with some of The primitive microornamentation pattern in lac- them being more important in some situations than ertids would be expected to limit friction, too, at in others, so that none of them show simple correla- least in forms that use narrow spaces. For example, tion with environmental parameters or particular Holaspis guentheri, which regularly enters cracks in features of the animals concerned, and there may wood and under bark (Arnold, 1989b), where its very also be hierarchies of priority. Finally, a particular big vertebral scales usually contact the internal sur- morphology may confer advantages in more than face of such crevices, is one of the few lacertids with one very different way, so it appears in very differ- very large dorsal scales that retain the smooth prim- ent situations. itive lacertid pattern of microornamentation on In fact, the range of possible performance advan- them. Other crevice-using forms with fairly large tages for different patterns of microornamentation scales, such as Lacerta mosorensis, also show this that has been broadly considered in the literature is feature. quite small and discussion is often limited to re- Dirt shedding. Most but by no means all lac- stricted taxonomic groups. Two factors that deserve ertids spend a lot of time in close contact with the more consideration are the frictional and light- soil and are at potential risk of picking up dirt on reflecting properties of the epidermal surface, the their scales. This is likely to sometimes obscure former being potentially significant in locomotion cryptic coloring and markings used in intraspecific and dirt-shedding and the latter in the control of communication. Dirt may also clog the scales and shine, as a means of enhancing camouflage. It is the interstices between them, reducing ease of proposed here that these factors are of substantial movement and, in small lizards, which have a large importance in determining microornamentation surface area relative to their volume and muscle patterns in lacertid lizards and at least some other mass, it may weigh them down, reducing pursuit squamates. and escape speeds. Lizards, of course, cast off all dirt Locomotion. Some patterns of surface structure when they shed their , but this is a relatively are likely to enhance locomotory capacity of squa- infrequent event and will not maintain a clean sur- mates either in increasing purchase, for instance, by face between sheddings. providing attachment points, or by reducing friction. Soiling is a problem especially in moist situations Thus, while gecko setae permit adhesion of the toes, where water facilitates the spread of dirt particles the very smooth body scales of uropeltid snakes over the skin surface. As the liquid evaporates, sur- probably minimize friction when burrowing (Gans face tension brings the small particles into very close and Baic, 1977) and the smooth belly scales of many and extensive contact with the skin, increasing ad- other snakes and lizards, including most lacertids, hesion by weak molecular forces and tending to pull may serve the same role in surface locomotion. Con- particles into any concavities that may be present. A versely, complex microornamentation on the body relatively smooth, even scale surface limits such and tail is potentially likely to increase locomotory adhesion and permits dirt to be easily wiped off the friction. scales as the lizard brushes against objects in its The benefits of friction reduction in locomotion environment. In contrast, dirt particles are likely to may also explain the observation that lizard dorsal become lodged in the concavities of complex micro- body scales that project from the skin are often ornamentation, where their surface contact may be smoother in their more exposed areas than else- increased and their displacement by objects sweep- where, something that is not caused by wear (Irish ing across the scale surface during locomotion is less et al., 1988; Maderson et al., 1998). Such distal likely. smoothness is apparent in lacertids: whatever the When detached scales of large-scaled lacertids general microornamentation, the tips of scales and with different microornamentations were coated any strongly raised keels on them are nearly always with fine silt, this could be easily wiped away in the much smoother than other scale regions (Fig. 3b). case of lizards with smooth ornamentation, such as As strong microornamentation is absent on the Adolfus alleni and Holaspis. In contrast, where mi- most exposed parts of the body scales of lacertids, it croornamentation is more complex and three- is unlikely to have much importance in gaining pur- dimensional, as in Psammodromus algirus, Ichno- chase, something not unexpected in these lizards, in tropis, and Ophisops, some of the silt persisted which locomotion mainly involves the limbs. How- through several wipings, leaving numerous particles ever, general smoothness may permit significant re- still lodged on the scale. It is consequently not sur- duction in friction when passing through vegetation prising that most lacertids that spend time close to or through narrow cavities. Lizards in other groups the soil in moist places retain the primitive pattern that habitually make close lateral or dorsal contact of smooth microornamentation on their dorsal scales with their environment during locomotion do often and most shifts to derived states occur in dry situa- have relatively smooth scale surfaces. This is true of tions, where dirt is likely to adhere less tenaciously, 164 E.N. ARNOLD or in habitual climbers in vegetation, where expo- suddenly appears or disappears as an moves. sure to soil and other kinds of dirt is considerably For example, it is common in Arabia to glimpse an less. abrupt glint of light caused by the sun reflecting off Reflection of heat and light. The skin is impor- a Sand (Scincus mitranus) as it dives into the tant in both absorbing and reflecting electromag- slip-face of a dune. Similar disruption of camouflage netic radiation and there are interspecific variations may occur when a hunting predator, such as a cruis- in infrared reflectivity that could be significant in ing raptor, moves relative to its potential prey, even thermoregulation (Porter, 1967; Bowker, 1985). It when this is static; shine may be intermittently seen has also been suggested that projections of the ober- by the predator as the positions of one or both ani- hautchen may reduce the amount of visible and ul- mals change with respect to the sun. It would con- traviolet radiation penetrating the body cavity, sequently not be surprising if, in many circum- where it may damage the viscera. Refraction within stances, there were benefits in suppressing projections of the microornamentation has been hy- epidermal shine and particular microornamentation pothesized to lengthen the path of radiation passing patterns may be one of the means of accomplishing through the body wall, thus increasing its absorp- this, rather as the shiny surface of glass can be tion before it reaches the visceral cavity (Porter, frosted by etching it to produce a fine-scale three- 1967). However, such possible effects are still dimensional structure. largely uninvestigated and discussion here will be Importance of gross scale shape. In lizards with almost entirely confined to the effects of scale micro- strongly reflective scale surfaces, the extent of shine ornamentation on the appearance of lizards. that may be visible, to an observer situated on the The squamate epidermis absorbs some visible other side of the lizard from the sun, is determined light but is largely transparent to it, permitting partly by scale size and form. This may be a largely transmission to the upper dermis, where a propor- unconsidered selective factor in determining the tion is reflected back through the epidermis by chro- shape of squamate dorsal scales; for instance, matophores. Some light, however, may be reflected whether they are flat or raised, smooth or keeled. directly from the oberhautchen, something which is Small strongly convex scales are potentially shiny most obvious when it strikes obliquely. The combi- over their whole surface but, when they are illumi- nation of transmission and reflection by the epider- nated by the parallel rays of the sun, light is scat- mis is directly analogous to the way a sheet of glass tered in different directions by the areas of curva- not only transmits oblique light but also reflects ture (Fig. 14b). Consequently, shine will only be some of it from its upper surface. Reflection involv- produced from the very small area of a convex scale ing some kinds of relatively fine microornamenta- that reflects light directly to an observer. Because of tion occasionally contributes to interference colors this, there is no continuous area of bright reflection on the scales of snakes, for instance, in the colubrid on the skin, just a stipple of small shining spots , Drymarchon (Monroe and Monroe, 1967) and which is not usually very conspicuous. This phenom- in uropeltid snakes (Gans and Baic, 1977), although enon can be seen in such lacertids as Heliobolus such colors may at least sometimes be an incidental lugubris and Eremias arguta. Strongly carinate effect of selection for mechanical performance ad- small scales also produce a discontinuous shine that vantages (Gans and Baic, 1977). is largely confined to the scale keels. In contrast, More frequently, reflection of the sky and espe- sunlight is reflected off large, flat scales as parallel cially the sun may produce shine, where a smooth rays that are likely to be perceived as a large area of surface gives rise to coherent reflection rather than shine (Fig. 14a). If some patterns of microornamen- scattering the light rays in many directions. To- tation reduce shine, it may be no accident that most gether with shape, color, and shadows cast, shine is derived states of this are often found in species with acknowledged in military and other contexts to be such scales. one of the factors likely to attract attention and The distribution of different scale shapes on the needing to be hidden or suppressed if camouflage is bodies of squamates also reflects the requirements of to be achieved. Shine is sometimes quite striking in camouflage. Some lacertid lizards, such as Lacerta lizards, for instance, in many scincids, and may sig- agilis, L. viridis, and L. praticola have nificantly reduce crypsis in forms that are otherwise on their upper surfaces that limit shine, but scales camouflaged by their coloring and sometimes body are smoother on the sides of the body, which are less form. Components of many common lizard habitats, visible from above, and smoothness is even more such as most earth, sand, and bark, and sometimes marked on the belly, which is usually concealed. A rocks and leaves of plants, too, have matte surfaces similar distribution of scale types occurs in many and substantial shine on the skin of lizards occur- snakes. ring on such backgrounds is likely to attract the Effect of microornamentation on shine. Examina- attention of predators. Such coherent reflection also tion of living lacertids and other lizards, and of often provides a clear visual cue that an inconspic- alcohol-preserved ones in which the skin surface has uous, perhaps countershaded, object is in fact three- been allowed to dry, indicates that microornamen- dimensional. Shine is especially noticeable when it tation does affect the amount and intensity of shine. SCALE MICROORNAMENTATION OF LACERTID LIZARDS 165

Fig. 14. Effect of curvature of reflective surfaces on light scattering. a: Sunlight is reflected off flat scale surfaces to produce parallel rays that are often perceived as a large area of shine. b: Convex scales scatter parallel light rays in different directions, so shine is only perceived from a very small area. c: Dished microornamentation also scatters light, restricting or eliminating shine.

As expected, shine is very marked on smooth belly light source. However, perceived shine will not be scales and, among forms with relatively large dorsal wholly suppressed because some normal reflection scaling, it is conspicuous in species that have the still takes place from the raised edges themselves. primitive lacertid pattern of microornamentation, Other microornamentation features often associ- where cell surfaces are largely smooth and their ated with raised edges reduce coherent reflection imbricating posterior margins are not angled up- further. Thus, the pustules of Algyroides also inter- wards. For example, this can be seen in Lacerta fere with coherent reflection from the scale surface. vivipara, L. mosorensis, and particularly Adolfus Denticulations have the same effect and, where they alleni, where the scales are especially big. In the are particularly developed and abundant, as on the latter species, shine is relatively subdued when light dorsal scales of the tail base of Gallotia stehlini (Fig. strikes the skin and, is reflected from it, at steep 2c), they absorb light in their interstices rather like angles, but more conspicuous when the angles are velvet or plush does. shallower. Large ridges contribute to light scattering by In forms with strap-shaped oberhautchen cells where the posterior borders are turned obliquely varying the orientation of different parts of the upwards, such as Tropidosaura, shine is present but scale surface, and the combination of raised pos- tends to be reduced compared with forms with prim- terior edges of cells and large ridges on the scale itive microornamentation. Again, this reduction is surface is particularly effective in reducing coher- greater at steeper angles of incidence and observa- ent reflection, as, for instance, on the large scales tion. Even if scales are still shiny at shallow angles, of many Takydromus (Arnold, 1997). The combi- there are likely to be real benefits in reducing shine nation of raised cell edges and ridge-like tracts of at steeper ones, as in many situations this is how denticulations in Poromera (Fig. 3b) acts simi- many predators view lizards. Presumably there is larly. Finally, the large dished cells of Ichnotropis less shine at steeper angles because some of the light and some Ophisops appear to represent a different falls into the gaps between the raised cell edges and means of preventing coherent reflection by scat- is absorbed or reflected directly back toward the tering light (Fig. 14c). 166 E.N. ARNOLD Conflicting benefits. The above factors suggest companied by an appropriate change in microorna- that lacertid microornamentation is controlled by mentation. Thus, although the sculptured derived conflicting benefits. Relatively smooth scales with dorsal patterns of Psammodromus algirus and Ich- low-friction characteristics may aid locomotion and notropis capensis are still present on the flanks, they dirt shedding, while reduction of shine through are attenuated. This parallels what happens with matte surfaces enhances camouflage. The two prop- general scale architecture (see Importance of Gross erties cannot be combined, for low-friction surfaces Scale Shape, above). tend to shine and matte surfaces produce high levels The relationship between more three-dimensional, of friction, but some situations select one property usually derived, microornamentation patterns and over the other. The smooth primitive pattern of mi- large scale size is corroborated by its development on croornamentation is apparently selected in forms the large scales of the tail of forms with small body that often live on the ground in moist places, where scales that otherwise lack sculptured microornamen- soiling is a particular problem, or which use narrow tation. A further indication that smooth microorna- crevices, where friction would impede movement. mentation may be important in making scale surfaces Conversely, more matte surfaces are generally se- easily cleaned is that, even in forms with generally lected in forms with large dorsal scales that are matte surfaces, like Poromera, the scales bordering inclined to shine. In situations where these factors the mouth, which are regularly smeared with prey conflict, low-friction considerations appear to pre- residues, retain the primitive smooth pattern. vail. Large-scaled forms develop matte ornamenta- Observations on other squamate groups. Mi- tions in dry habitats or when they live away from croornamentation reduces reflection in many other the ground, both situations where soiling is unlikely squamate groups, thus improving camouflage. This to be a great problem, but this does not occur in is particularly so in the production of the strikingly large-scaled forms that would benefit from low fric- matte and velvety dorsal surfaces of geckos and of tion. For instance, this is true for Adolfus alleni, vipers; for instance, in the genera Bitis and Both- which is ground-dwelling in moist habitats and, as rops. Indeed, the Fer de Lance (Bothrops atrox)is noted above, for crevice-using Holaspis guentheri. called Terciopelo (ϭvelvet) in Spanish. In these Natural tests of performance advantage. The cases the dense perpendicular projections of the mi- hypothesis that smooth microornamentation of dor- croornamentation act exactly like velvet in enabling sal body scales is functionally important in reducing light to be absorbed and to some extent randomly dirt adhesion and friction during locomotion and reflected without producing any coherent shine. In that three-dimensional microornamentations re- other groups, the more three-dimensional microor- duce shine can be tested further. Other areas of the namentations also tend to reduce coherent reflec- body where these selective pressures are also likely tion. This can be seen in the Mauritian skinks of the to occur but with different levels of severity can be genus Gongylomorphus. Gongylomorphus bojeri, examined to see if their microornamentation varies with smoother microornamentation, is very shiny, in ways predicted by the hypothesis. Thus, the belly while G. fontenayi, in which there are stronger of most lacertids is in much more extensive contact raised and denticulated projections from the cell with soil than the dorsum is, when lizards are rest- edges, is more matte (pers. obs.). ing and during slow locomotion, and is largely out of Other groups also show the same kind of regional the sight of predators. As might be expected from differences over the body encountered in lacertids. this, the belly microornamentation is usually either Thus, many snakes have greater ornamentation on of the primitive pattern or even smoother (see Com- the dorsum than the flanks and little if any on the ments on Microornamentation of Dorsal Tail Base belly. Again, these shifts repeat on a microscopic and Belly, above). There are, however, four indepen- scale those often seen in general scale morphology. dent cases where the smooth ventral surface has been lost and replaced by a complex microornamen- Apparent Contraventions of the tation like that on the dorsum. This has occurred in Functional Hypothesis the more terminal members of the two main lin- eages of Takydromus (Arnold, 1997), in Gastropholis Although the distribution of the various kinds of tropidopholis, and in Poromera. All these forms microornamentation fits quite well with the hypoth- climb extensively in vegetation matrixes and are esis, that patterns confer different advantages in likely to be out of contact with the ground much of terms of smoothness and reducing shine, with the time and, because they are not regularly close to smoothness having precedence in cases on conflict, the substrate, their bellies are also potentially much there are still a number of aspects that require fur- more visible than is usual. ther explanation. Flanks are somewhat intermediate between dor- Pitting. As noted, many lacertids have minute sum and belly in the extent of their visibility and pits on the surface of scales that may take up around contact with habitat. Consequently, there is a shift half the surface area. These might be thought of as in the balance between the advantages of reducing yet another means of reducing reflection. However, shine and having low friction and this may be ac- unlike most other kinds of potential reflection spoil- SCALE MICROORNAMENTATION OF LACERTID LIZARDS 167 ers, pitting in general is not strongly associated with tailed habitats of the Ophisops species involved to the large scales that are likely to produce the most tell if this explanation is likely to apply to them. extensive shine. Indeed, even densely pitted scales Lack of reversal after loss of function? Al- with otherwise smooth surfaces are often shiny. This though a shift to polygonal cells that are separated is true of Pseuderemias and the Pedioplanis undata by welts is understandable in terms of light scatter- group (including, among others, P. undata itself, ing when the cells are dished, there are cases where P. inornata, P. namaquensis, and P. rubens) Here, such dishing is not apparent. Nor are such instances although overall shine is greatly reduced by the associated with large matte scales, the scales con- convex shape of the small scales, these are actually cerned being small and shiny. This occurs in He- quite glossy when examined closely. liobolus spekii, Pseuderemias mucronata, and In fact, pits are very small, often being only about P. striata and in most members of the Pediplanis 0.5 ␮m across. As such, their diameter is less than undata group. It is most parsimonious to assume the wavelength of most visible light (0.4–7.0 ␮m), that dishing arose after this situation, but such a which means that, in the context of light reflection, sequence would be difficult to understand in func- such pitted surfaces are likely to act as if they were tional terms. An alternative, less parsimonious, pos- entirely smooth. Consequently, large relatively sibility is that the pattern of polygonal cells with matte scales bearing pits probably get their dull raised borders did in fact arise in association with finishes from the grosser features of microornamen- dishing in the context of reducing shine on large tation that they bear, such as raised cell borders, scales, but later lost this role with change in scale large ridges, and dished surfaces. Dense pitting is size and shape. Dishing then disappeared but polyg- most common in dry habitats, and possibly it can onal cells and raised borders remained, not revers- only be sustained in such situations where adhesion ing to the primitive condition in spite of loss of is less of a problem, because pitted surfaces are function. perhaps more prone to hold dirt. What positive ben- Pedioplanis provides some evidence that this may efit pitting may confer is unclear. Perhaps it simply have been the case. Pedioplanis lineoocellata has makes epidermis cheaper to produce by reducing the somewhat enlarged dorsal body scales that have amount of ␤-keratin needed. dished polygonal cells with raised edges. An esti- Different solutions in different situations? mate of phylogeny based on 35 morphological char- Within the homogeneous genus Ophisops, there are acters places this species basal to all other members two different microornamentation patterns. Mem- of the genus (Arnold, 1991). However, this hypothe- bers of one clade have strap-shaped cells that are sis of relationship is not very robust and, when five small with raised posterior borders, while dished variable microornamentation characters (1–4 and 6 polygonal cells with ridge-like margins have arisen in Table 1) are also included in the phylogenetic in other species, perhaps even twice. Both arrange- analysis, P. lineoocellata shifts to become basal to ments reduce coherent reflection and do not differ the P. undata group alone (although admittedly this obviously in the extent to which they suppress shine, arrangement too is not strongly substantiated). something that also probably applies to frictional Members of the P. undata group may consequently properties. One possible explanation for the exis- have passed through a stage with relatively large tence of these two different methods of reducing scales that had dished cell surfaces. Although dished shine is that they may have different countervailing cells do not occur on the dorsal body scales of the costs that may be more important in some habitats P. undata group, they are present on the tail of at than others; for instance, those costs involving the least one of its members, P. rubens. absorption of electromagnetic energy. Visible light is Retention of a derived microornamentation pat- converted to heat when absorbed and must contrib- tern in situations where it does not necessarily con- ute to the heat load of lizards and snakes. Such costs fer its original benefit may occur elsewhere. may not be important in vegetated habitats where Ophisops microlepis retains the steeply angled pos- temperatures are lower as a result of shade and terior cell borders that reduce gloss in large-scaled evaporation from the plants, but they are likely to be species of its genus, but it itself has secondarily more significant in dry open situations, where veg- developed small dorsal scales which do not seem to etation is sparse. Perhaps this is reflected in the merit reduction of shine. kind of derived pattern of microornamentation that Why are there so many derived states? If de- reduces shine. Arrays of densely packed ridges or rived patterns of microornamentation are generally denticulations that limit reflection partly by direct- important in producing matte surfaces on large ing light inwards are more likely to result in light scales, why are there so many different “solutions”? absorption and consequent heating than dished po- One reason is simply that lacertid epidermis ap- lygonal cells, where coherent reflection is limited by pears capable of producing a wide range of struc- scattering the light by reflection. It is perhaps no tures. Also, as noted in Ophisops, different patterns accident that in other lacertid genera dished polyg- may sometimes possibly represent adaptations to onal cells occur in species living in open situations. different environmental conditions. The various de- Unfortunately, not enough is known about the de- rived states may also act in an additive way to 168 E.N. ARNOLD reduce shine and additional derived states are fre- LITERATURE CITED quently incorporated into microornamentation on lineages (see Patterns of Association and Order of Arnold EN. 1987. Resource partition among lacertid lizards in southern Europe. J Zool Lond B 1:739–782. Change in Dorsal Body Ornamentation, above). Arnold EN. 1989a. Towards a phylogeny and biogeography of the Lineage effects other than lack of reversal after Lacertidae: relationships within an Old-world family of lizards loss of function may also have influenced outcome. derived from morphology. Bull Br Mus (Nat Hist) Zool 55:209– For instance, Takydromus and Poromera have both 257. evolved strikingly similar microornamentations at Arnold EN. 1989b. Systematics and adaptive radiation of equa- torial African lizards assigned to the genera Adolfus, Bedria- middle magnifications (Fig. 3a,b) and both have lon- gaia, Gastropholis, Holaspis and Lacerta (Reptilia: Lacertidae). gitudinal arrays of apparent ridges. However, in J Nat Hist 23:525–555. Takydromus the ridges are in the scale itself and Arnold EN. 1991. Relationships of the South African desert liz- underlie the surface cells, whereas in Poromera they ards assigned to Aporosaura, Meroles and Pedioplanis (Reptil- ia: Lacertidae). J Nat Hist 25:783–807. are made up of tracts of denticulations produced Arnold EN. 1994. Investigating the origins of performance advan- from the free posterior edges of the cells (Fig. 1f ). tage: adaptation, exaptation and lineage effects. In: Eggleton P, Presumably, the latter option is open to animals Vane-Wright RI, editors. Phylogenetics and ecology. London: that have some capacity to produce denticulation (a Academic Press. p 123–168. rare feature) and might substitute for the generally Arnold EN. 1997. Interrelationships and evolution of the east Asian grass lizards Takydromus (Squamata: Lacertidae). Zool J more frequently produced ridges. Linn Soc 119:267–296. Bea A. 1986. A general review of the dorsal scales’ microorna- mentation in Vipera species (Reptilia: ). In: Roc˘ek Z, CONCLUSION editor. Studies in , Proc 3rd Ordinary General Meet- ing of the Societas Europaea Herpetologica. Prague: Charles The pattern of epidermal microornamentation University. p 367–372. found in particular lacertid lizards appears to result Bentley PJ, Blumer WFC. 1962. Uptake of water by the lizard Moloch horridus. Nature 194:699–700. from a complex system of influences. Evolutionary Bezy, RL, Peterson JA. 1988. The microstructure of scale surfaces changes occur in a variety of components that ap- in the xantusiid lizard genus Lepidophyma. Herpetologica 44: pear largely independent of each other. Particular 281–289. changes may take place many times in the Lacerti- Bowker RG. 1985. The infrared reflectivity of the desert lizards Cnemidophorus velox and Sceloporus undulatus. J Thermal dae, but rarely reverse. Consequently, it seems that Biol 10:183–185. microornamentation involves a number of develop- Bowker RG, Spindler HS, Tilden A, Bairos VA, Murray R. 1987. mental pathways that are substantially unidirec- Reflections on lizard skin: the ultrastructure of the scales of tional. Natural selection appears to control such Cnemidophorus exsanguis and Podarcis bocagei. In: Van Gelder changes but more than one selective factor is in- JJ, Strijbosch H, Bergers PJM, editors. Proc 4th Ordinary Gen- eral Meeting of the Societas Europaea Herpetologica. Nijme- volved. The benefits of low friction in locomotion and gen: Societas Europaea Herpetologica. p 83–86. dirt-shedding promote the selection of primitive Bryant SV, Breathnach AS, Bellairs A d’A. 1967. Ultrastructure smooth microornamentations, while the benefits of of the epidermis of the lizard (Lacerta vivipara) at the resting reduction in reflection in camouflage elicit more stage of the sloughing cycle. J Zool Lond 152:209–219. three-dimensional derived ones that can be achieved Burstein N, Larsen KR, Smith HM. 1974. A preliminary survey of dermatoglyphic variation in the lizard genus Sceloporus. J Her- in different ways. Effective compromise between the petol 8:359–369. two basic kinds of micromorphology is not possible Chiasson RB, Lowe CH. 1989. Ultrastructural scale patterns in and, where selective forces conflict substantially, Nerodia and Thamnophis. J Herpetol 23:109–118. low-friction structure prevails. The balance of the Cole CJ, Van Devender TR. 1976. Surface structure of fossil and recent epidermal scales from North American lizards of the main selective factors varies not only between spe- genus Sceloporus (Reptilia, Iguanidae). Bull Am Mus Nat Hist cies and populations, but also on different areas of 156:451–513. the skin of the same individual. Some microorna- Dowling HG, Gilboa I, Gennaro JF, Gennaro AL. 1972. Microder- mentation features, like fine pitting, are not con- matoglyphics: a new tool for reptile (Abstract). Her- trolled by these selective factors and different ways petol Rev 4:200. Estes R, de Queiroz K, Gauthier J. 1988. Phylogenetic relation- of producing three-dimensional microornamenta- ships within Squamata. In: Estes R, Pregill G, editors. Phylo- tions may carry different incidental costs. It is likely genetic relationships of the lizard families. Stanford: Stanford that the epidermal microornamentation of other University Press. p 119–281. squamates is also a product of these varied factors. Gans C, Baic D. 1977. Regional specialisation of surfaces: relation of texture and biologic role. Sci NY 195:1348– 1350. Gans C, Merlin R, Blumer WFC. 1982. The water-collecting ACKNOWLEDGMENTS mechanism of Moloch horridus re-examined. Amphibia-Reptilia 3:57–64. Nathan Hart produced many of the micrographs. Gasc J-P, Renous S. 1980. Les reliefs microscopiques de l’e´pi- C.G. Jones and L. Jones of the Electron Microscope derme des squamates () et le mode de locomotion. 105e Unit, Natural History Museum London, gave help Congre`s national des Socie´te´s savantes, Caen. Sciences 3:23– 33. and advice on techniques. I thank Colin McCarthy Groombridge BC. 1980. A phyletic analysis of the viperine for helpful discussion and two anonymous reviewers snakes. PhD Thesis. London: Council for National Academic for a range of useful comments. Awards. SCALE MICROORNAMENTATION OF LACERTID LIZARDS 169

Harris DJ, Arnold EN, Thomas, RH. 1998. Relationships of lac- Peterson JA. 1984a. The scale microarchitecture of Sphenodon ertid lizards (Reptilia: Lacertidae) estimated from mitochon- punctatus. J Herpetol 18:40–47. drial DNA sequences and morphology. Proc R Soc Lond B 265: Peterson JA. 1984b. The microstructure of the scale surface in 1939–1948. iguanid lizards. J Herpetol 18:437–467. Harvey MB. 1993. Microstructure, ontogeny and evolution of Peterson JA, Bezy RL. 1985. The microstructure and evolution of scale surfaces in xenosaurid lizards. J Morphol 216:161–177. scale surfaces in xantusiid lizards. Herpetologica 41:298–324. Harvey MB, Gutberlet RL. 1995. Microstructure, evolution and Porter WP. 1967. Solar radiation through the living body walls of ontogeny of scale surfaces in cordylid and gherrhosaurid liz- vertebrates with emphasis on desert reptiles. Ecol Monogr 37: ards. J Morphol 226:121–139. 273–296. Irish FJ, Williams EE, Seling E. 1988. Scanning electron micros- Price RM. 1982. Dorsal microdermatoglyphics: eco- copy of changes in epidermal structure occurring during the logical indicator or taxonomic tool? J Herpetol 16:294–306. shedding cycle in squamate reptiles. J Morphol 197:195–206. Price R. 1983. Microdermatoglyphics: the Liodytes-Regina prob- Lang M. 1989. The morphology of the Oberhautchen with the lem. J Herpetol 17:292–294. description and distribution of scale organs in basiliscine igua- Price R. 1989. A unified microdermatoglyphic analysis of the nians. Amphibia-Reptilia 10:423–434. genus Agkistrodon. Snake 21:90–100. Larsen KN, Burstein N, Smith HM. 1973. Phylogenetic trends in the ultradermatoglyphics of the lizard genus Sceloporus. HISS Price RM, Kelly P. 1989. Microdermatoglyphics: basal patterns News 1:12. and transition zones. J Herpetol 23:244–261. Lee MSY. 1998. Convergent evolution and character correlation Renous S, Gasc J-P. 1989. Microornamentations of the skin and in burrowing reptiles: towards a resolution of relationships. the spatial position of the Squamata in the environment. Biol J Linn Soc 65:369–453. Fortschr Zool 35:597–601. Leydig F. 1872. Die in Deutschland lebenden Arten der Saurier. Renous S, Gasc J-P, Diop A. 1985. Microstructure of the integu- Tu¨ bingen: Laupp. mentary surface of the Squamata (Reptilia) in relation to their Leydig F. 1873. Uber die a¨usseren Bedeckungen der Reptilien spatial position and their locomotion. Fortschr Zool 30:487– und Amphibien. Arch Mikrosk Anat Entwick 9:753–794. 489. Maderson PFA. 1966. Histological changes in the epidermis of the Ruibal R. 1968. The ultrastructure of the surface of lizard scales. Tokay (Gekko gecko) during the sloughing cycle. J Morphol Copeia 1968:698–703. 119:39–50. Ruibal R, Ernst V. 1965. The structure of the digital setae of Maderson PFA. 1970. Lizard glands and lizard hands: models for lizards. J Morphol 117:271–294. evolutionary study. Forma Functio 3:179–204. Sammartano F. 1976. Microornamentazioni delle scaglie di alcuni Maderson PFA, Rabinowitz T, Tandler B, Alibardi L. 1998. Ul- sauri. Animalia 3:195–205. trastructural contributions to an understanding of the cellular Smith HM, Duvall D, Graves BM, Jones RE, Chizar D. 1982. The mechanisms involved in lizard skin shedding with comments on function of squamate epidermatoglyphics. Bull Phil Herpetol the function and evolution of a unique lepidosaurian phenom- Soc 30:3–8. enon. J Morphol 236:1–24. Stewart GR, Daniel RS. 1972. Scales of the lizard Gekko gecko: Maddison WP, Maddison DR. 1996. MacClade 3.06. Sunderland, surface structure examined with the scanning electron micro- MA: Sinauer Associates. scope. Copeia 1972:252–257. McCarthy CJ. 1987. Sea snake puzzles. In: Van Gelder JJ, Stri- Stewart GR, Daniel RS. 1973. Scanning electron microscopy of jbosch H, Bergers PJM, editors. Proc 4th Ordinary General scales from different body regions of three lizard species. J Meeting of the Societas Europaea Herpetologica. Nijmegen: Morphol 139:377–388. Societas Europaea Herpetologica. p 279–284. Stewart GR, Daniel RS. 1975. Microornamentation of lizard Monroe EA, Monroe SE. 1967. Origin of iridescent colours on the indigo snake. Sci NY 159:97–98. scales: some variations and taxonomic correlations. Herpeto- Patterson C. 1982. Morphological characters and homology. In: logica 31:117–130. Joysey K, Friday AE, editors. Problems of Phylogenetic Recon- Stille B. 1987. Dorsal scale microdermatoglyphics and rattle- struction. Systematics Association special volume 21:21–74. snake ( and ) phylogeny (Reptilia: Viperidae: Perret J-L, Wuest J. 1982. La structure fine des e´cailles de Crotalinae). Herpetologica 43:98–104. quelques Lygosominae africains (Lacertilia, Scincidae), re´ve´le´e Vaccaro O, Uriondo A, Filipello AM. 1988. Microornamentaciones par le microscope e´lectronique a` balayage. Rev Suisse Zool de las ce´lulas de Oberha¨utchen en Tupinambis teguixin (Linne´, 89:269–280. 1758) y Tupinambis rufescens Gu¨ nther, 1871) (Sauria, Teiidae). Perret J-L, Wuest J. 1983. La microstructure des e´cailles de Iheringia Se´r Zool 67:65–75. quelques Scincide´s africains et pale´arctiques (Lacertilia), ob- Williams EE, Peterson JA. 1982. Convergent and alternative serve´e au microscope e´lectronique a` balayage, II. Rev suisse designs in the digital adhesive pads of scincid lizards. Sci NY Zool 90:913–928. 215:1509–1511.