A REVIEW OF AFRICAN-MADAGASCAN GEKKONID LIZARD PHYLOGENY AND BIOGEOGRAPHY (SQUAMATA)
ARNOLD G. KLUGE
AND
RONALD A. NUSSBAUM
MISCELLANEOUS PUP'.ICAT'3NS MUSEUM OF ZOOLOGY, UNIVERSITY OF MICHIGAN NO. 183 MIS(XLLANE0lTS PUBI,I<:ATIONS MUSElJM OF %OOI,OGl: LrNIVEKSI'lY OF MICHIGAN NO. 18?1
.I . lnc ~x~l)lic;~lionol'tl~e Muselur~ of Zoology, The Univrl-sity of Michigan, co~~sistprim;trily of' two series-tllc Occasional I'aprl-s and the Mi>.cclla~icorrsPublic;~tions. Both series welt foundetl by 1)n Rt-yant Walker, Mr. Bt.adshaw H. Swales, ancl Ilr. W.W. Ncwcon~h.Occasionally the Muscun~put)lisIncs contl-il~~~tionsor~tsidc ol these sel-irs; heginning in I990 thcsc at-c titled Special Publications and arc nr~mbercd.All submitted rnan~rscriptsreceive cxtet-nal review. The bliscellancous Publications, which incluclc nlonographic stltdies, papers on firltl and mu- seum techniqries, iintl othcl- contributions not within [he scopr of the Occasional Papers, are pub- lished separately. It is not inlendctl that they be gl-ouped into vol~~n~cs.Each number has a title page ;und, when ileccssrry, ;I table of contcnts. The C)cc;lsional Papel-s, publication of w11ic.h was begun in 19 13, serve as a ~nediu~nfor original studies based principally upon the collections in the Muserrm. They are isstled separ;rtely. Wl~cna slrfficirnt nrrn~her-of pages has bc.t:~i pri~~tetlto makc a volr~~rte,a title page, table ofcontet~ts,and ;III indvx arc s~rpplictlto lil~rariesand indivitl~r;tlson the rnailing list for tl~eseries. Publicatio~~of. the Occasional Papers has heen strspentled. A conrplete list of p~~hlic;itionson Rirds, I:isl~es, I~lsects,M;ur~n~als, Moll~rsks, Reptiles alrd A~lnpl~il)- ims, ;uncl otl~crtopirs is ;rvailable. Atltlt-rss inq~~iriesto tl~c Ilirccto~; Museum of Loology, The Univer- sity of' Michigan, An11 ArI)ot; Miclnig;~~~48 109-1 079.
Kol~ll',F.1. K. F.1,. 13ookstei11(ctls.). 1990. I'rocccdi~~gsof' the Michigan Mo~.~~l~o~r~ct~-icsM~I-ksl~op. Spec. I'l~bl.2. :480 pr). With softw;i~-c,$25.00. Witho~rtsoftware, $17.50. i\lc=~~~tlcl;R.D. 1990. kIow tlitl hurn21ns e\,olvc? licflcctions on the rrniq~rrlyIIII~~II~ species. Spec. P11hl. I. :?A pp. $4.00. liaxwol-thy, (:,j. & 1i.A. Nlrsst)aum. A rrvirw of' tl~cxMadaptscan SII;I~~(;cnc~-;r I'cc~r~clo~t.lro/)r~c, I'(~i.ccrhor/i,rtcrc.cc, i~ndff~/~roliodoir (Stl~~aniirtir: (:olrrb~-idae). Misc. Puhl. 182. 37 pp, 25 figs. $1 1.50. (-Tosl~nc, . . W.A. 1993. A survey ol'upl~rrjaw IIIIIS(.II~;~~III-~ill highel. telcostcai~fishes. Occ. Pap. 724. 2ti pp. 9 Jigs. $2.20. I)t~ell~nan,W.E. &,!.A. (:i~mpbcll. 1992. Hylicl 11.ogsof'tl~c gc1111s I'/f~c./tiiltyln: Systrnnatics ;i~~tlpltyloge- lietic- rcl;rtionsl~ips.Misc. I'uhl. 181. 38 pb), 21 figs. $9.10. Mclcitl-ick, b1.C. I!)!)I. Phylogcncric analysis of ;IV~:III I~indlirnbrn~rsc~~la~urc. Misc. 1'11hl. 179. 8') PI), 3 figs. $9.60. I I3ca11cBaker, AIIIIArbor I'ar~l W. Kro\v~~,Mackinac lsl;~nd Sl~il-IcyM. McFee, Kiittle (:leek Neil1 D. Niclscn, Brighton l'hilip H. I'owc~; All11 Arl)o~- Vcronic;~1,;itta Sn~itl~.(;rosse IIc Nellie M. Va1.ne1; Ilctl-oit ,];~nles I.. Waters, Muskegon 1;rrnesJ. I)~~dc~.sradt,c.?i oflit io A Review of African-Madagascan Gekkonid Lizard Phylogeny and Biogeography (Squamata) Arnold G. Kluge and Ronald A. Nussbaum Museum of Zoology and Department of Biology The University of Michigan Ann Arbor, Michigan 48109-1079 MISCELLANEOUS PUBLICATIONS MUSEUM OF ZOOLOGY, UNIVERSITY OF MICHIGAN NO. 183 Ann Arbor, October 23,1995 ABSTRACT Kluge, A.G., and K.A. N~~ssDaun~.1995. A Hminu OJAJi-ican-Mccdagascan Geltkonid Liznrd Phylogen? And Biog(:ography (Squamala). Aksc. Pu61. Mus. Zool. Univ. Micl~igan,183:l-20, 11 figs. The phylogenetic and biogeographic history of African-Madagascain gekkonid lizards is assessed with 34 morphological charac- ters. The cladistic analysis employs 22 i~lgroupand seven outgroup genera as terminal taxa, and the best- fitting hypothesis leads to the following corlclusions: (1) the ingroup forms a clade; (2) the generally rccog~lizedPachydactylus assemblage of genera forms a clade, as do the northern and southern African subgroups, (C:rYc;lton%a,Tarentola) and (Chondnxiactylus, (;olopus, Pachydactylus, I'almatogecko, lUloptrol,us), re- spectively; (3) the Madagascan-Seychelles sa~mpleforms a historical entity, except for the ambiguously related Gedtokpis; (4) the sample oS Madagascan-Seychelles endemics is derived from African gekkos, and some of the African fauna has a Madagascan ancestry; (5) a Lygodactylus-Phrlsuma clade (including Rhoptrof~ella)is delimited; (6)Blaesodaclyl~~s and Homopholis are not sister taxa. Adding other "padded-toed" gekkos, Gehj~ra,Geklio, and EZemidar.lylus, to the ingroup leads to a substantially different set of phylogenetic and biogeographic conclusions, and reveals the tenuous nature of most of' these conclusions. The close phylogenetic relationship between Ge/~,yraand Paragehym implied by Angel (1929) is not confirmed. It is clear that synapomorphies are more import;ant than geographic proximity in recovering the histoly of' gekkos. These results are largely at odds with Joger's (1985) and Rauer's (1990a) recent findings. ICcy words: I PAGE Introduction ...... 1 Recent Research on the Phylogeny ol' Gekkonids from the Ethiopian Region ...... 1 Methods. Ingroup Taxa and Characters ...... 3 Outgroup Taxa ...... 4 Character Descriptions ...... 6 Other Potential Evidence ...... 12 Phylogenetic Results ...... 13 Biogeographic Results ...... 15 Status ol' Pmagef~ym...... 15 Conclusions ...... 18 Acknowledg~nents...... 18 Rel'erences ...... 18 11.1. USTKATIONS FIGURE PAGE Joger's (1985.fig. 3) hypothesis of phylogenetic relationships ...... 2 1Sallcr's (1990a.fig. 1) hypothesis ol'phylogenetic relationships ...... 3 Phylogenetic hypothesis reslilting fro111 a reanalysis of Bar~er's(1990a: appendix 1) data ...... 4 1Sony palate oS gekkos ...... 10 Pectoral girdle of gekkos ...... 11 Phylogenetic hypothesis ol'gekkos restricted to the Ethiopian Region ...... 12 l'hylogenetic hypothesis of gekkos restricted to the Etliiopian liegion. and i~lcludillgthe cxtralimital Gehyrc~.C.eliltko. and Hemidnctylus ...... 13 Area cladogram Lor the gekkos analyzcd hyJoger (1985.fig. 3) ...... 16 Area cladogram for the gekkos analyzed by Bauer (19tlOa.fig. 1) ...... 16 Area c:ladograrn Sor the gekkos restricted to the Ethiopian Kegion ...... 17 Area claclogranl for the gekkos restricted to the Ethiopian Region. and includi~lgthe rxtralimital Geh.yra. Gekko. and Hmzidactyl~us...... 17 TABLE PAGE 1 (:haracter By Taxon Data Matrix ...... 5 2 Individr~alCharactel . Performa~ices...... 14 3 Individual Cllaractcr Pcrlorma~lccs...... 14 INTRODUCTION internal morphological variation, including diploid chroinosome number, to the aforerneiltioned regional endemics, as well as The most recent phylogenetic and biogeographic reviews of Geckonia. Unfortunately, none of this potential evidence was gclikoilid lizards1 from thc Ethiopian Region by Jogcr (1985) discussed in detail, and it is difficult to decipher several ofJoger's and Bauer (1990a) did not include Puragehyra, a Madagascan characters (fig. 3, p. 487), such as "fusion orrostra1 halves," "char- endemic. The omission was understandable because that taxon acteristic pedal morphology," "distinct body proportions," and was kiiown only from the holotype of I? pelili. Recently, "blood proteins." Those putative synapomorphies that he re- Nussbaulli and Raxworthy (1994) collected several specimens ferred to only as "other characters" (fig. 3, p. 487) are impos- of a new sister species of I? petiti, and the availability of that ma- sible to understand. In addition, much of the variation that tcrial has provided us with the opportunity to investigate the Joger considered was sufficient only to distinguish subgroups of relationships of Paragehyra. the aforemeiltioned terminal taxa, not intergeneric relation- Joger's (1985) and Bauer's (1990a) studies were examined ships. carefi~llyfor information that might be used to document the Joger's (1985: fig. 3) inethod of hypothesis formation must aCfinitics oP l'arngehyra. UnLort~~nately,their phylogenetic propo- be questioned as well. He began with the pattern of relation- sitions dill'er significailtly (compare Figs. 1 and 2). Moreover, as ships provided by the immu~lologicaldistances, onto which he thc following review emphasizes, the questiollable analytical then mapped some of the character data in an attempt to real- methods and data they cmployed makes it difficult to judge the ize collfirmatioil and increased resolutioil (Fig. 1; his fig. 3). basis for their differences and to idcntify general agreement Such a protocol assumes the truth of the immunological dis- (Bauer, 1993: 251). Consccluently, we foulld it necessary to re- tance topology, which prejudices the evaluation of the charac- cxamiiie the phylogenetic relationships of the Ethiopia11Region ter data. For example, Joger necessarily had to claim the inde- gckko F'I~II~~,including l'f~ragelzyra. pendent evolution of certain morphological character states such as liypcrphalangy (Haacke, 1976) and the re-evolution of "ad- hesive pads" in Plyodnclylus (his fig. 4). Sy~lapomorphiessuch as IIECZENT RESEARCI-I ON TI-IE PHYLOGENY OF these can never be judged homologous with Joger's method, GEKKONIDS and it is not clear why he used ally data other than immunologi- FROM THE ETHIOI'IAN REGION cal distances. Thus, we conclude that Joger (1985) made only a limited Joger (1985) emphasized the immunological distances he ob- coiltribution to our understanding of the history of the tained froin precipitin tests becarrse, as he asserted, such dis- gekkonids horn the Ethiopian Region. His sample of the en- tances arc "not subject to convergence or parallelism" (pp. 486- demic gekko fauna froin the Region is small, and the amount 487) and therefore givc tlle "correct phyletic brailching order" and quality of the evidence and the analytical methods he em- (11. 480). tlowevei-,Joger Pound it necessary to apply various ployed are disputable. His analysis of only Ethiopian Region "correction" lactors to the distances (p. 481), in order to be bring gekkos presumed a radiation of those endemics from a single them in line with his assumption of a "clocklikc behaviour of most recent common ancestor, an assumption that is as yet un- albunii~lin the evolutiorrary process" (p. 482). That he had to supported by any data. make those adjustments suggests this source of'evidence for phy- Bauer (1990a: 276) recog~lizedfive geographic kinds of Atiri- logenetic rclationsliips may exhibit error. Further; Joger's im- can-Indian Ocean gekkonids: (a) southern outliers of primarily rn~~ilologicaldistances iilcluded only a few representatives of northern (tropical) radiations, (b) populations of pan-tropical Ethiopian Region cndcrnics, i.c., fIomol)holis, 12ygodactylus (I,. species, (c) forms of broad distribution in the Indian Ocean gutturalis, I,. yiclu.ralzrs), I'achydaclylus, l'almalogecko, Phekumcc, region, (d) Iildiail Occan (primarily Madagascan) endemics, I'lyodaclj~kus, lihoplropi~s,Slenodaclylus, and Turenlola. These taxa and (e) southern African endemics. Bauer examined only those represent very few of the lineages usually assumed to have origi- taxa of the last three types (c-e), and in doing so he presumed nated in thal Rcgion, and with the exception of Pl~ekuma(I? (likeJoger) that the Ethiopian Regio11endemics evolved froin a ~kub%aSi-om Madag.ascar and 1-1 v-nigrafi-omthe Coinoro Islands), single most recent common ancestor. Bauer's data set consisted all of his samples came Gom the Afi-ican continent. Moreover, of 23 morphological characters observed on representatives of Jogei- provided only average values for one-way precipitin tests, the following taxa (his appendix 1): Afroedura, Ailuronyx, and the assumptioli of constancy of evolutionary rate and the Chonrlrodactylu.~,Colopus, Ebenauia, Geckolepis, Homopholis, accuracy of the "corrcctioil bctors" Joger eniployed cannot be Kaokogecko, Lygodactylus, Micro.scalabote.~,Millolisaurus, Narudasia, critically cvaluatccl in thc abscncc of a complctc, reciprocal Pachydactylus, Palmatogecko, Paroedura, Phelsu.ma, Phyllodactylus, matrix oP distances. I'crhaps most importantly, immunological Ptenopus, IQ~ol,lrol,elln,Rlzo$trol,us, Urocotykdon, and Urol~latus.Ex- distance data cannot be examined enipirically to determine the cept for the claw reduction characters (see below), he seems to specific iliclepe~lde~ltevolutionary cvcnts responsible for alter- have included all of the decipherable and potentially informa- native phylogenetic hypotheses, as is possible with character data tive characters employed by Joger. Although Rauer's character wheil investigated in ternls of character coilgrueilce (IUuge, descriptions are brief, they are much more explicit than Joger's. 1989; T S = 44, CI = 0.55). In Pact, there are 42 cladograms one step and the highly incongruent nature of the character states inves- shorter than he reported, the strict consensus of which is illus- tigated forced us to use heuristic algorithms in our attempts to trated in Figure 3. Several taxa, in addition to those empha- discover the most parsimoilious cladogram, and under those sized by Bauer (p. 278; G(:ckolepis,N~~rudasia, and Ptenopus), are circumstances the limited (m*) and extended (bb*) branch- rcsponsible for the multiple equally parsimonious hypotheses swapping routines are generally recommended (Farris, 1988). he round. Ballel- not only failed to find the best-fitting A posteriori iterative weighting (the xs w algorithm in Hennig86; hypothcsis(es), but he also did not present the most parsimoni- Farris, 1988) was employed in choosing among two or more ous distribution of character slates on his published cladograin equally most parsimonious cladograms (the secondary cla- (see Fig. 2 for corrections). Nonetheless, it is noteworthy that dograms of Kluge and Wolf, 1993). The combination of mq', both hypotheses based on Rauer's data (compare Figs. 2 and 3) bb*, and xs w is iterated until the ensemble consistency (CI) delimited the Pac/1ylaclylus clade (see also R~rssell[1972], Haacke and retention (RI) indices, rnetrics widely employed in phylo- [19'76],Joger 11985: fig. 31, and Iiluge [1987: fig. 12]),as well as genetic inference to assess fit to data, do not change. Individual the (Ailuronyx, IXomo$h,olis), (Lygodactylus, Microscalabotes, character consistency (ci) and retention (ri) indices are used to Millolisaurus) and (Paroedura, Ebenauia, Urol,lalus) groups, the evaluate each variable's performance. The minimum number latter being predicted by Bauer and Russell (1989a). of steps a character can exhibit on a particular cladogram is s; that number summed over all characters is S, or the total length of the cladistic hypothesis. A character's weight (w) is the prod- METHODS, INGROUP TAXA AND CFFARACTERS uct of its rescaled consistency and retention indices (times lo), where the smallest s is used. A more detailed discussion of this Wc used the olltgroup rooting method and heuristic algo- general approach can bc found in Kluge (1993~). ritlllns m* and bb* providccl by I-lennig86 (Farris, 1988) in our The following Ethiopian Region endemics constituie the ma- attempts to discover the best-fitting hypothesis of sister group jority of the gekkonid ingroup terminal taxa surveyed: Afroedura, relationships. Thc outgroup methodology not only establishes Ai1uron.y yx, Blaesodacly lus, Chon.drodacty lus, Colopus, Eb'l,enauia, the globally most parsimonious description of the evidence, but Geckolepis, Geckonia. Homopholis, Lygodactylu.~, Pach.ydactylus, tcsts the historical individuality of the ingroup (Clark and Palmatogecko, Paragehyra, Paroedura, Phelsuma, Phyllodactylus, Curran, 1986; see also Nixon and Carpenter, 1993). Such a bcst- Ptyodactylus, Rhoptr@ella, lihoptropus, Taren,tola, Urocolyleclon, fitting cladistic Ilypothesis(es), in minimizing homoplasy, maxi- Uroplatus. Three geographically widespread taxa, Gehyra, Gekko, mizes the explanatory power or the data (Kluge, 199313). It is in and Hemidactylus, were added to the ingroup in a second analy- this context that the homoplasy and taxonomic generality of sis, in order to further test the historical individuality of the Ethio- each 01' the characters are ,judged. The large number of taxa pian Region endemics and the robustness of their intergeneric Fig. 3. A strict consensus of thc 42 cq~~allylnosi parsimonious cladogranls (S = 43, CI = 0.55, RI = 0.78) recovered fro111 the data published by Bauer (1990a:appendix 1). (:ompare to Figtlrc 2. Note thr I-educed resolution in this I~ypothesis,but Urocol~~/rdon'sirllarnbiguous sister group rclatio~lsllips. relationships. All of' the Ethiopian Region genera consitlered of'all Lygodaclylusspecies to other illvcstigators (e.g., liaxworthy, in detail by Joger (1985; Fig. I) and Ra11er (1990a; Fig. 2) are in progress). For the purposes of the presenl reanalysis, we fol- included as ingroup or outgroup taxa (see below). All c)f the low Bauer (1990a; conlra Russell, 1977a) in considering the ter~ninaltaxa analyzed in this study are represented by one or monotypic Rh,optrof~ella(R. ocellatc~)aild P/i~elsu,mnasseparate taxa. more specimetls (see Appendix), and each terminal tax011 is The f'orrner taxon is an Africa endemic, whereas the latter is assu~nedto be monophyletic. With thc exception oC a largc very largely Madagascan, and their taxonomic separation in this number of uncatalogued southern Atrican Plzyllodrcclyl~u.~,the study provides a basis for testing the relationships between those species and skeletal material that we examined are listed in the two biotas. The Pli,yllodactylusanalyzed herein are southern At- Appendix, all of which are housed in the University of Michigan rican species only (I;: ansor ro Ail~i- Blaeso- Chondro- Colopus Ebennuin Grclto- Gecltonia Gehyra Geltko Hemi- r(1zirrr ronyx dnclylus dact)llus Zc1)is dacty lus 1. 1 1 I 0 0 1 0 0 0 0 0 2. 1 1 I 0 0 0 0/ 1 0 0 0 0 3. 0 0 0 0 0 0 0 0 0 0 0 4. 0 0/1 0 1 0 0 0 0 0 0 0 5. 0 1 0 0 0 1 0 1 1 1 0 6. 0 1 I 0 0 1 0 0 0 0 0/1 7. 0 1 1 0 0 0 I 0 1 1 1 4. 0 0 0 0 0 1 0 0 0 0 0 9. I 0 0 1 1 0 0 1 0 0 0 10. 1 0 ? 1 1 0 0 0 0 1 0 11. 0 0 0 0 0 0 0 0 0 0 0 12. 0 0 0 I I 0 0 1 0 0 0 IS. 0 0 0 1 1 0 0 0 0 0 0 14. 0 0 0 0 0 0 0 0 0 0 0 15. 0 0 1 0 0 0 1 0 1 0 1 16. 0 0 0 0 0 0 0 0 0 0 0 17. I 1 I I 1 1 1 I 1 1 1 18. 0 0 0 1 1 0 0 0 0 0 0 10. 0 0 0 0 0 0 0 0 0 0 0 20. 0 1 0 0 0 0 1 0 0 0 0 21. 0 0 0 0 1 0 0 0 0 0 0 22. 0 0 1 1 1 1 1 I 0 0 0/ 1 23. 0 0 0 0 0 0 0 0 0 0 0 24. 0 0 0 0 0 0 0 0 0 0 0 25. 0 0 0 0 0 0 0 I 0 0 0 26. 0 0 1 1 1 1 1 1 0/ 1 0/ 1 0 27. 0 0 0 1 1 1 0 1 0 0 0 28. 0 0 0 1 1 I 0 0 0 0 0 29. 0 0 0 0 0 0 0 0 0 0 0 30. 0 0 0 0 0 0 1 0 1 0 0 31. 1/2 0 0 0 0 2 1 1 1 0 0 32. 0 0 0 0 0 0 0 0 0 0 0 33. 0 0 0 0 0 0 0 0 0 0 0 31. 1 I 1 1 1 1 1 1 1 1 0/ 1 FTornn go Nor~i- Pncl~y- Pulmalo- I'nragr- Par- I'ILPI- southern African I'i1.~- Ptcn,o- f~hfl/i.\ dficlyl~~s dnsin dnclj~zr, gecf~o 1) o~!dui-cc sumn Pliyllodectylus lurus pu~ Mrsc. PUBL.MIJS.. ZOOL., UNIV. MICH., NO. 183 Tablc 1 (cant.) Ptyo- Q~ic,rlo~- otro- K/~o,/)lro- Saz~re Steno- Xlren- 'Iimto- Lfroco- liro- rlnclylri \ Ji.ldliir p(~lla pus rlncfylus daclyl~u lol(~ scincus lyhdo?, plnlus I. 0 0 1 0 0 0 0 0 0/ 1 I 2. 0 0 1 0 1 0 0 0/ 1 0 0 3. 0 0 I 0 0 0 0 0 0 0 4. 0 0 0 0 0 0 0/ 1 0 0 0 5. 1 0 ? 0 0 0 0 0 1 1 6. 0 0 0 0 0 0 0 0 1 1 7. 0 0 0 0 0 1 0 0 1 1 8. 0 0/1 1 0 O/ I 0/ 1 0 0 1 0 9. 0 0 0 0 0 1 1 0/ 1 1 0 10. 1 I 0 0 1 0/ 1 1 1 0 0 11. 0 I 0 0 1 0 0 0 0 0 12. 0 0 0 I 0 0 1 0 0 0 18. 0 0 0 0 0 1 0 0 0 0 14. 0 0 1 0 0 0 0 0 0 0 15. 0 0 0 0 0 0 0 0 0 1 16. 0 0 I 0 0 0 0 0 0 0 17. I 0 I I 0 0 1 0 I 1 18. 0 0 0 0 0 1 0 I 0 0 19. 0 0 0 0 0 0/ 1 0 0 0 0 20. 0 0 0 0 0 0 0 0 0 0 21. 0 0 0 0 0 0 0 0 0 0 22. 1 1 0 0/1 1 0/1 1 I 1 1 23. 0 0 ? 0 0 0 0 0 1 0 24. 0 I 0 0 1 0 0 0 0 0 25. 0 0 0 0 0 0 1 0 0 0 26. 0 0 1 1 0 0 1 0 0 0 27. 0 0 1 1 0 0 1 0 0 0 28. 0 0 1 1 0 0 0 0 0 0 20. 0 0 0 0 0 0 0 0 0 0 SO. 0 0 1 0 1 0 0 0 1 1 31. 0 1 1 1 O/ 1 0 0 0 2 2 32. 0 0 0 0 0 I 0 1 0 0 33. 1 0 0 0 0 0 0 0 0 0 3.1. 1 1 1 1 1 1 1 0 I 1 Gekkonidac, and the rnonophyly of that taxon is accepted, in Prom he Ethiopiar~Region, 01- are belicvcd to be close relatives the absencc of evidence to the contrary. Unfortunately, aside to thosc African lineages (Kluge, 1967a, 1987), and they include from the sphaerodactyl cladc (Kluge, 1995), no other large sub- the taxa investigated by.Joger and/or Bauer. The importance group of gekkonids has been distinguished unambigrlously by of considering additional "naked-toed" forms as outgroups (e.g., two or marc synapomorphies. Obviously, ti~rtherresearcl~ is Cnernaspis, Nnctus and Tropiocolotrs) requires further SLLI~~;how- required iSwe arc to understand the history of this most speciosc ever, in terms of the characters we employed, their presence part ofgelikota~lphylogeny. Even the sister group relationships would probably not change our findings, becat~sethey are so 01' much smaller taxa, like I'crc~gelzym, are dependent on this similar to the outgroups we did include. morc gelieral hypothesis. For the purposes of the present study, we assllmc the absence ol'a hyperextensive mechanism in the digits (sensu Russell, 1972), CHARACTER DESCRIPTIONS a11d being "naked-toed" (or "padless"), is plesiornorphic at the level or Gckkonidae. Wl~ilethe siatc or the digit cannot bc op- 1. Nasal bones paired (0) or fused (1). Roth states are widely timized ~~narubigt~ol~slyfor Pygopodidae, the mechanisni is :tb- distributed among gekkonoids (Kluge, 1987;J. R. Dixon, pers. sent in EltblepharicL~cand ?Iw~~toscinni,s,and more generally in comm.), and we review that information because it provides the Sq11ama~1.Thus, the lhllowing "naked-toed" terminals are des- reader with a general sense ofthe conservativelless of the char. ignated as outgroups: Na,rud(c.sin, Pristur~i,~,I'tenoptr.~, acter (the same considerations apply to characters 48, 11, 13, Q?tedei?fbldtin,Sa~crodactykus, Stenodactylus, and 'limtoscincu.~.Like 1!5,20,22-24,and 26). The [used condition is typical of Ajroedurn wise, Rauer (1990a) assl~rncclthe "naked-toed" condition to be (Coggel-,1964), Ailuronyx, Apmsin (all species, except A. repens), plesiomorpliic (his character 17), and the only two such t~:~aBlnesodactylus, C:hristiwu.s, (C:rrernaspis (fused in C. aflynis, C. africnna, that he incl~tdedin his study, Nai-i~dasiaandPtenopus, werejudgc:d C. Doulmgeri.i, C. kandinna, C. nipidius, C. qun/tuors&nta, C. to be sister lineages relative to all of the other terminals that lie siarnensis, and C,'. uy~zadensz.~;paired in C. indica, C. ken,dallii, and investigated on the basis oi'this single character (Figs. 2-3). Jo~er C. ornccta; see Dring [I9791 and Inger et al. [I9841 for reviews of (1985) described the sister group relatioilships of only one "na- Cn,~,maspis),Ehennvia, H~rnifihyllodactylus, LePidodactylus, ked-toed" Laxon, Stenodnctylus; however, he postulated that it Zdygodactylus (including Ilomergurlln), Micropclio (sensu Kluge, evolved from a "padded-toed" ancestor (Fig. 1). The afoi-emu- 1983: 472; contra Leviton et al., 1992: 47), Micro.scalaDotes, tioned "naked-toed" outgroup taxa employed in our study a]-e Millotisaurus, Nrcctus, I'aragehym (only paired posteriorly, like Uroj,lat,cr.s),Parocd~trq l'(~roc1,iru.s Pl~~lsz~rna(lilsed in all species UMMZ 201505; B. sakalava, UMMZ 192323) and Homopholis examined [see I~owev~l;Bauei-, 1990a1, inchtding I! abbotti, l? ualberpi (UMMZ 127699). The fused state is typical of tlre re- ccskicettr, I? harl~ouri,I! cepedicma, F1 c(l):rr~orensis,I! duhia, I? guenthpri, maining two Homopholis species (H.Jasciata and H. mulleri), and I? g~cimbeaui,F1 hticnud(~,I? lineata, i? madagascarien.sis, l? ornnta, is obviously correlated with the deep sculpturing that covers the and I? su,nrlOergi), so~rtlrci-11African Phj~llodactj~lt~s(also fused in parietals. Bauer (1990a: appendix 1) over-generalized the fused I? ~rieOecliii,hut not in other Old World species currently consid- condition to all species in the Bluesodactylus-Homopholiscomplex. ered I-'lr,yllodactyl~ii.s,such as I? europneus and P siamensis), 5. Stapes perforate (0) or imperforate (I). According to I's(!udogl.llko, RIi~o/)lro/)ella,sorne tirocotyledon (fused in U. inex/~ectata Underwood (1971; see also Underwood, 1957), the facial artery and U. palntatn; paired in I/. ?uoltcrstor/fi; U. 7ueiler-i to be dele]-- passes behind the stapes in all lizards, except in gekkonoids and mined), ant1 Iiroplntns. Tlre nasals are fused throughout most Dibamidae (Anelylrol~sirand Dibamus; Greer, 1976). Tlre two or all oftheir lcngtli in all Blae~~odactylus(13. centongibnsis, B. Ooiui~ii, conditions in gekkos and pygopods, both of which are consid- and B. saltnlavn), wlrcrcas the supposedly closely related ered derived, are (a) the stapes is imperforate, the artery passes FIomof,koli.s (H. ,Jascintct, H. mmull(~i,and H. wall~crgii)has paired anterior to the stapes, and (b) the artery passes through a fora- elements. Tlrc paired nature of the nasals is even ol>viousin l? men located distal to the footplate. The stapedial Soramen is Jn.scintn, wl1ic11 has the most extensive sculpturing of tlie bones absent in the following gekkonoids (see also Kluge, 1987): of thc snout of all six Ulaesodaclylus-Ho7ne)j)I~olisspecies. iiiluronyx, Cnem.aspis, 1?'benauia, Geckonia, Gehyra, Gekko, 2. Frontal bone single (0) or paired (I). The paired condi- Hemiphyllodactylus, Hoinonota, Lepidodactylus, Paroedura, Pmochirw; tion is present (Kluge, 1967a, 1987; see also Jogel-, 1985) in some soutlrern Afi-ican Ph~llodactylus(foramen present in l? iAfie(l.urn, Ai11~:ronyx(the posterior 2/3 of the frontal), some nn,sor,qi and l? lineatus, but absent in l? porphyreus, according to C;c~clr,ol(lbis(single in G. maculata, paired in (;. tyj~icn),Homopholis, J. R. Dixon, pers. comm.), Ptenopus, Ptychozoon, Ptyodactylus, solne I~jgo(l~ectj~lu.s,some Plielxumte, IO~oplropellce,Saurodactylus, and Pygopodidae (both diplodactylines and pygopods) , Thecadactylus, solnc 7irnrosci.n~cu.s(paired in 7: scincus, single in 7: rn.%crole/~~s, Urocotyledon (only U. inexfiectala examined; contra Bauer, 1990a), and otic specilnen of T I)rezrualskii has the posterior region of and Ur@latus. Although Bauer (1990a) recorded Rhoplropella tlre bone paired, b111 the anterior portion is single). The varia- as having a perforate stapes, the single cleared and stained speci- tion observed in 12ygodac/ylusand Phelsuma may be, at least in men of R. ocellata available to us (UMMZ 127760) has the im- pill-1, due to comparing specimens oL' cliffcrent sixes and ages; perforate condition. However, in that individual there is a con- Iiowr\lcr, that does not explain unambiguous cases of intraspe- spicuous notch on the anterior side of the pedicel of the stapes, cilic variatioti wlirre the spccinie~isexhibiting tlie alternative near the footplate, which suggests tlre artery passed within the condi~ionslravc nearly identical snout-vent lengths (e.g., in L. developmental template of the stapes and that one of the ltlrcgei, UMMZ 143389 exhibits the single state, whereas UMMZ foramen's enclosing walls failed to materialize. Thus, we score 143390 has the pairctl condition). In contrast to Bla/;.sodaclylu.s, Rhoptropella as lunknown. all adult I-lomofih,olisexhibit state 0. Only a sltbadult l? fi~scsciata 6. Marginal tooth positions moderate in number (0) or many (UMMZ 127698) has the paired conditio~r. The fact tliat the (1). Although Bauer (1990a: 278) employed this character, he largest species of Trratoscinc.us (7: sc.incus) exhibits the paired state did not specify a particular set of teeth (i.e., premaxillary, max- airti the smallcst species (7: in;icrolel~i,s)the single condition indi- illary or dentary), nor did he precisely define the conditions cates that, if this variation is to be interpreted as a heterochronic "moderate" and "many." Fortunately in gekkonoids, the num- patt(:rti, t1rc11 it does not always accompany niiniaturization (e.g., ber of teeth on the maxilla and dentary are highly positively as in C;c~ckole/~is).As suggcstccl by I(111ge (1987: 34), tlrc phyloge- correlated, and likewise vary with age (body size) (e.g., Kluge, netic infor~nativrnessof this character is questionable, because 1962). Thus, we have arbitrarily chosen the adult maxilla as the of the extreme variation that occurs within some species. source of information on marginal tooth positions, and we in- Stcplienson (1962: fig. 5; see also MoSSat, 1973) stated tliat fer those numbers from the states Bauer attributed to particular Plethokeex had a paired Sroirtal; however, no such condition is evi- taxa. For example, he listed only Ebenauia, Paroedura, Urocotyledon, dent in tlic specimens irt hand (UMMZ 13121 5,131232,173966). and Uroplatus as haviiig state 1 (his appendix l),and given that 3. Frontal bone broadly participating in orbital rim (O), or these taxa have 36 or Inore marginal maxillary tooth positions excluded from orbital rim by pre- and postfrontal contact (1). per side (see below) we will assume that range defines Bauer's Tllc 1°C- and 1x)stfrontals arc widely separated in almost all state 1, and that 35 or less marginal teeth defines state 0. Our gckkonoids. In most pygopods, the contact is complete, or nearly observations are summarized as follows, as modal or median so, and P11,eOumtrand liho/)lro/)~llaare also exceptions in that the values per side, determined on adults of as many species as are Srontal is cxcluded fi-om the orbit by tlie union of thc prc- and available, with exceptional material (those lying outside the nar- ~xjstfrontalboncs. row range exhibited by conspecifics and congeners) noted sepa- 4. Parietal bones paired (0) or single (1). This character var- rately: AJroedura (27), Ailuronyx (39), Blaesodactylus (39), ics little in gcltkonoids. Asingle parietal is present only in some Chondrodactylus (32), Colopus (25), Ebenauia (44), Geclrolepis (33), Ailicroiiyx, C~/r~oizo'rodacLylu.s,cubleplrarids, Lepidobleph,aris festae Geclionia (24), Gehyra (29, except for one of three specime~rsof (Parkcl; 1926),I,inlis, sorrie P(e(:I~ydnctylzr.s(e.g., tlre l? bibroniicom- G. ocean,ica which has 36 [UMMZ 185913]), Gekko (33), plex), P(!,ochir.u,s, some Pli~elsuina(e.g., I? ~ur:nlhrri),and some Hemidarlylus (29, except for H. garnolii and H. giganteus which ?./nr~illo!ce. Tlrc paired state may bc obscnred in the I'acl~ydactylus have 37 and 36, respectively), Homopholis (H.fasciata has 25 [I8 hibronii complex by the deep scl~lpturingwhich covers tlre su- in subadult, UMMZ 1276981, H. walbergii 35), I,ygodactylu.s (17), ~xrlicialpf~storbital bones. Also, the paired state is exhibited by Mil1otisnuru.s (16), Narudasia (22), Pachydactylus (25), except l? a11 l3leee.sorlaclykus (l3. (~ntoiigilonsi.~,UMMZ 192820; B. bo%uin.i, biOroniicomplex which has 32), Palmatogeclto (22),Paragehyra (38), 8 M~sc.PUBL. M~s.. ZOOI.., UNIV. MICI-I., No. 183 Pmrn'ura (36), PIIPLSU~IL~(25), southern African Pli~llodnrtyk~.~ comm.). It appears tllat the depth and width of the fenestra (30),I'ristrstrtnls (20),I-'ls.nof)~~.s (I8), I'tyodnctyki.t (30), Q~~ed(~r~felrltinwere measured along (anteroposteriorly) and Si-om (laterally) ( IS), Xho/)troJ)rlla (16) , Rko/)tiopus (27), Soul-odnctylus (24), the midline, respectively. Howevei-, that characterization niay S/c~ri~odnc.lj~li~.s(26), R~rentoln(30) , 7i Fig. 4. Velill.al view of tlre palate. A. The plesiomorphic colldicion in gekkos (I , Geckolepis Geckonia Hornopholis Lygodactylus Narudasia Pachydactylus rnaculata chazaliae walbergii klugei festiva bibronii Palrnatogecko Paragehyra Paroedura Phelsurna Phyllodactylus Pristurus rangei gabriellae pictus barbouri porphyreus carteri / Ptenopus Ptyodactylus Quedenfeldtia Rhoptropella Rhoptropus Saurodactylus garrulus hasselquistii moerens ocellata bradfieldi fasciatus Saurodactylus Stenodactylus Tarentola Teratoscincus Urocotyledon h' ~roplatus rnauritanicus sthenodactylus rnauritanica scincus inexpectata firnbriatus Fig. 5. Vt.11~1-alview of ~ht.111id-vcntral poi-tion of tht. pectoral girdlc of a repl-c-scntative of'each terlrlillal taxon irlvcstigated in clctail in this sltrtly (r~ottll-awn io sc~ilc:).For variatiol~wiihirl genera (e.g., Sn~urodacly1u.s)see text, clla~-acLer.30. Abbreviations: c = clavicle; i = interc-l;tvic-lc(black); r, = stet-n;rl rib; I-, =xiplrist~l-nal rib; s = stei-n~rm. Fig. 6. Sistrr- grollp rel>ttio~rshipsamong thc: cndrn~icEthiopian licgion ingl-oup taxa. Tl~rlrypothesis is a strici consensus of the 72 equally ~rrostparsi~noniolls cladogl-ams (S = 98, CI = 0.85, RI = 0.94), altcr four iterations of'dif'fbrc~~ti:~lcll>~racier weighting. rTl~cIcngt11 (S) ol the consensus cl;~dog~-amis 100. Tlre "nakccl- coecl" taxa, Norltdnsin. PI.~\/ILTLIJ,I'~PIIo/)~)(IJ, Q~~~cl~n,/(,lcl/ia, Sn?srodnrtylu.\, St~nodcrttylz~.\,and T~~n~osrJ~r~r~~~,were tlesig- natcd ol~tgroups. Tlre raw tlata al-c listed in Table 1. Tlre nutnbers I-cfcrto tllosr cha~actersand states (thc lattcl-in p:u-c~ltlleses)which rlnanrbigrlo~~slytlcli~nit sister groups. Each asterisk t1cnotc.s a 1111iqueand rl~lreve~.sed syrrapo~rrorphyin this study (see text lor fill-rlrer.discllssion). Comp;i~-cto Figul-r 7. to extren~elylong 211ld II~I-I-owarnls, the former lying sorrlle dis- rI (mloscincus. tance from the clavicles, whereas the latter make extensive. con- 33. Splenial present (0) or absent (1). According to Kluge Lact with the clavicles. In state 1, the vertical body of the (l987), the splcnial is present in all grkkonids, except i~iterclaviclecan be narrow (allnost rod-like) to extremely broad, Cobodaclylu.~,Connlod~,, L~f)l,idobkphnrk, Pri~lu.rus, 13s(,udogo~7,atodes, (either oval to nearly triang~~lar).Figllre 5 surn~narizestlhe in- Plyodaclylus, and Sphaerod(ict~!lu.s. tergerieric variation observed in the tam surveyed for this str~dy 34. Calcified postcranial endolymphatic sac absent (0) or (the two Snurodactylzcs illustratetl, S. /'n.scint.usand S. mn~rritanicu~, present (1). Among those gekkonids sllrveyed herein, only doclunent the high degree of variation that can occur among Tm(ito,scinc.u,sand "many" Hemidact~~lus(Si~nkiss, 1967) are said sister species). Altlrot~ghwe scored Pl~yllodrcctylz~,~as state 1, tlrcl-c to exhibit state O (Kl~rge,1987). is some lota able variation. For example, P florflh,j~rc..u.sl~asno arliis wliatsoevcl; I? linecctus has short, but obvious, projection.., and the coildition in I? e~c)^of)cceusis i~~ter~nediatebetween these ex- OTHER POTENTIAI. EWDENCE tremes. We have no doltbt that li1tlu.e research will make it pos- sible to score additional pl~ylogeneticallyinformative descrip- Underwood (1954; 1970) emphasized the importallcc of pu- tors, particularly idendfjing the daggcl--shape variation whic:h is pil shape in the classification of geklios. Although it is extremely interniediate behveen states O and I. diflicult to acc~rratelydetermine some ofthe different pupil types 31. Number of attached sternal and xiphisternal ribs 5-6 (O), that he recognized (Kluge, 1967a: 14),certain of Unde~wood's 4 (I), or 3 (2). Only tliosc ribs ar? co~mtedwhicl~ make a conl- characterizations may eventually prove informative. For ex- plctc connection to the sternllm and xiphisternum (Fi,y. 5). ample, he referred to the Rhoplropustype (Underwood, 1954: Although we assume this transforrnatio~lto be additive, so that 471) as occurring in Chondrodnctylus, Colof~us,Pc~lmmtogecko, it is reasoilable to apply n floste~or7iterative weighting, that prc- Ptenopus, Kkof)trof~rll(c,and Rhoptropus. In testing the generality sumption has no effect on the phylogenetic hypothesis that is of this synapomorphy, future investigators \vould be well advised finally realized (Fig. 6). to rccxa~lii~~ethose taxa ~vhichUnderwood listetl as llavi~ing 32. Relatively few (0) or many (I) scleral ossicles. These data straight-vertical and round pupils. Furthel; those tests should are taken SI-om Kluge (1987: table 1). The only ~axathat arc be made on living orgariisrns under similar conditions (particu- judged to have statr 1 in the present study are Stenodnctykl~a~td larly light intensity). Fig. 7. Sister group relationships among the e~ldernicEthiopian Region (see Fig. 6) and extralimital ingroup taxa ((;rl~yrn,Geltko, H/.nlidnclylus). The hypothesis is a strict consensus of ~hc128 cqually most parsimonious cladogr-ams (S = 101, CI = 0.82, RI = O.Y3), after Tour iterations of difkrcntial character weighti~lg.The length (S) of the consensus cl;~dogramis 103. The "nakcd-tocd" taxa, Nnrudasia, PI-isturu~,I'lvlnopus, Qucdenfrlllia, Sccurorloct~ilus,Stenodaclylus, and 7i~rcrl/~.sczn,cu,\,werc dcsignatecl outgrotrps. The raw data are listed in Table 1. The ~lu~nbersrcfkr to those charactcrs and states (the latter-in paren~lleses)which umarnbiguously delimit sister groups. Each asterisk denotes a unique and unrcvcrscd sytlapornorphy in this st~tdy(see text for Curther discussion). Compare to Figure 6. The variability Biihnie (1988) described in the gckkonid the historical individuality of the ingroup was confirmed, and llcrrlipcnis suggests another source of potentially informative Ptyodaclylus is the sister lineage to all other parts of that clade; characters. However, it is obvious that future research niust pro- (3) the Pachydactylus group of Russell (1972; see also Haacke, vide Inore precise descriptions of that variabiliiy and survey many 1976;Joger, 1985; Kluge, 1987; Bauel-, 1990a) is delimited; (4) Inore tam (Bohme, 1988: 67). Unfortunately, even the clearest northern and southern African subgroups are identified within of Biihme's (1988: 160) conclusions, that "[tlhe hemipenis or- the Puchydactylu.c radiation (as per Russell, 1972; see also Joger, na~nentationof lJroj)l(~tusis so unique that a higher categorial 1985); (5) no taxa known from Madagascar-Seychelles are in- [sic] rank ofthis grollp seclns justified," is without merit in the cluded within the Pachydactylus clade, and the Madagascan- context ol'the monophyletic taxonomy employed herein. Seychelles assemblage forms a historical entity in its own right, except Ibr the ambiguously placed GeBo1ej1i.s; (6) a Lygodactybs- Phelsurna group (including Rho$trofiella) is delimited; (7) PHYLOGENETIC RESULTS Paragehyra is part of a highly derived assemblage, the compo- nents of which are restricted to Madagascar and nearby islands Our analysis ofthe relationships of the ingroup terminal taxa (the only exception is Urocotyledon); and (8) Chondrodactylus and endemic to the Ethiopian Region (see p. xx above) involved Colo$u~are interpreted as being secondarily padless (see char- four iterations (1: CI = 0.39, RI = 0.66; 2: CI = 0.67, RI = 0.83; 3: acter 17; also Russell, 1972). Two of Bauer's clades, (Ailuronyx, CI = 0.74, RI = 0.88; 4: CI = 0.85, RI = 0.94). Seventy-two equally Homopholzs) and (Eben,avia, Puroedura, Uroplatus) (1990a; see also most parsimonious secondary cladograms (S = 98) resulted from Bauer and Russell, 1989a), are not suppor~ed,nor is the recent the bb* :rpplication in the last run, thc strict consensus of which claim by Volobouev and Ineich (1994) that Ailuronyx seychelbnsis is illustrated in Figure 6 (S = 100). Several conclusions can be is more closely related to Hom,opholis wahlbergii than it is to extracted [I-om that conservative hypothesis: (1) relationships Phelsuma cepediana (Fig. 6; see also Fig. 7). Geckonin and Tarenlola among the outgroups are unresolved, with the exception of continue to be identified as sister taxa (Russell, 1972:95;Joger, Nrr.rzcdusicr, Pristu~us,Quedenfildlia and Saurodactylus which form a clade (which also probably includes sphaerodactyls; ICluge, 'Sphaerodactyls wcrc excluded because they form a highly de- 1987; IClugc, in press)L,and the sister taxa Stenodaclylus and rived assemblage which does not affect the polarity decisions in 7i,1-(~to,sci~i,(:us(for an alternative opinion see Huge, 1987); (2) this study. 14 Mrsc. PUBL.Mus.. ZOOL.,UNIV. MICH., NO. 183 1985), in spite ofthe "profound differences" Biihme (1988: 160) data set is weaker than the individual character consistencies discovered between the hemipenes of these two taxa. I,asr.ly, we (c) might suggest (Table 2), because some independent evolu- do not agree with Riihme's (1988) conclusion that the unique tion is ignored by coding polymorphic data as missing (Platnick henlipenis ornamentation of Uroplatus justifies a higher rank et al., 1991). For example, the absence of cloacal bones (char- for that taxon alone, given the highly derived and consistent acter 11, Table 2; see also cloacal sac, character 24) is reported sister group relationship between Zlrocotyledon and U~qlatusthat as a unique and unreversed diagnostic feature of the (Narudasia, we discovered (Figs. 6-7). Quedenjeldtia, Pristurus, Saurodactylus) clade; however, the analy- Extreme caution must be exercised when the cladogram in sis does not take into account the absence of those bones (and Figure 6 is used in hypothesizing trends in character evolution sacs) in some Lygodactylus (Table 1) . and historical biogeographic events (see below). Although the The sensitivity of the cladogram illustrated in Figure 6 to taxo- cladistic hypothesis maximizes the explanato~ypower of the data nomic sampling is easily demonstrated by adding Gehyra, Geltko, at hand (Table I),and "globally" so over in- and outgrollps, it is and Hemidaclylus to the ingroup of Ethiopian Region endemics sensitive to changes in characters and/or taxa. and rerunning the analysis. Although only three iterations (m:';; It should be borne in mind that the proposition in Figure 6 is bb'g; xs w) are required to achieve maximal fit to data (S = 101; the result of three successive applications of character weight- CI = 0.82, RI = 0.93), the strict consensus of the resulting 128 ing (xs w iterations 2-4; see Table 2 for final weights). The strict equally most parsimonious secondary cladograms (S = 103) is consensus hypothesis from [he first run is largely unresolved substantially different from the ingroup of Ethiopian Region (there are 220 equally most parsimo~liouscladograms) . hilore- endemic taxa analyzed alone (compare Figs. 6 and 7). The three ovel-, less than half (8/19) of the clades in the final result are extrali~nitaltaxa are "padded-toed" (see character 17),and their corroborated by two or more synapomorphies (Fig. (i),and few addition substantially changes the relationships among the parts of those phylogenetically informative characters can be inter- of the clade predominated by Madagascan forms (the sister preted as unique and unreversed (Table 2). Still further, the yroup to the Pacl~ydaclylusradiation). Of particular interest is Tablc 2. Inclividual character pevformailces for hypothesis 01' Iigure (5. Abbreviations: s = number of steps; ci = consiste~lcy index; ri = retention index; w = weighting coefficient. Table 3. Intlividual charac~erpel-formances for hypothesis of Figul-e 7. Abbreviations: s = number of steps; ci = consistency index; ri = retenti011 index; w = weighting coefficient. the rrnstable sister group al'finities of Puragehyra. More gener- Madagascan radiation (including that of the Comoros and ally, Figure 7 calls into question Baucr's assumption (1990a) that Mascarenes) being derived from the southern gekkonid clade. his ingroup terminal tam form a natural group, and the addi- Although the available data do not support Bauer's (1990a: tion of those three extralimital "padded-toed" taxa de~nonstrates 280) conclusion that most of the Madagascan (including the the need for much more research on the higher classification Comoros and Mascarenes) gekkonid fauna (~iluronyx~, of Gekkonidae. Bauer's claim (p. 277) that "the outgroups as- Bluesodactylus, 1 3 7 7 $' 7 c7 @ a 3 $$a@ q b0 $ 8 8 s74 ‘!?a 7$.$ 3 $37 ? 06 P,o?b 91039 488 $ 2 $ 3 $8 + &? 3 b"+ $PI$9.~2 -23 $g C'!? 3 ."$$ " $? 2 284 $2 .*Po.*$ @ $' ,.$3 10 $' $8Q,ocese4e4 z,,o G 8 B PPP P P P PPP P.P P. ST*^ P' 4 4 4 Fig. 8. An area cladogfi~rnI~asrd on the [;lxolro~nicrelatio~lsllips ill~~stratrd in Figul-c 2. Ab11revi;~tions:A = Africa (illcluding tllc Ar;~l)iilnPeninsula); M = Madagascal- (including Seychelles and other nearby islalids in thc wcstcl-n Inclian 0cc;un). Fig. 9. A11 arca cladogmm b;~scdon the taxo~iorr~ticrelationships il1ustr;lted in Figure 3. Abbreviations: A = Mi-ica (iricl~tcling[lie AI-abianPeninsula); M = Madagascar- (i~lcludingScyclielles and other nearby islallds in the Iilclia~iOccan). 7 $ 7 $, g 2 -3 ,,s.S*g,g $ sebp&>s"4'" Q 0 ', 0 d $$$q* ~8~ < P, $&'-$$a$ $ e*$$QQ& PPPPPPP P PPPPPPP Fig. 10. iln ;rrc;r cl;rclog~-;rnlbasrtl on thc taxo~io~nicrelationships illustrated in Figure 6. ill~hrcviations:A = Africa (includ- ing the i\rabian Peninsula); M = M'idagascar (incliiding Seychelles and othcr ncarby isla~idsin thc Indian Ocean). Fig. 11. 1\11 arcs cladogra~nbasctl on the t;rxononlic. rrlaiiorisliips illr~stratedill Figure 7. Ahhrrviations: A = Africa (inclurling the AI-ahianl'cnins~ila); M = Madagascar (inclurling Scychcllcs and othcl- nearby islands in the Indian Occan); W = gcogi.;q)liically ~vitlcsprcad. 419. Pasteur, G. 1964. Recherches sur l'kvolution des lygodactyles, ICzards Illgel-, R. F., I-I. Marx, and M. Icoshy. 1984. An undescribed species of afro-malgaches actuels. Travaux de 1'Inst. Sci. ChCrifien, Zool. gckkonid lizard (Cnemasj)is) ftom India with comments on the sta- 29:l-132. tus of C. tropidogastet: I-Ierpetologica 40:149-154. Perret, J.-L. 1986. Rkvision des especes africaines du genre Cnemaspis Jogcr, U. 1985. The African gekkonine radiation - preliminary phy- Strauch, sous-genre Ancylodactylus Miiller (Lacertilia, logenetic results, based on quantitative immunological compari- Gekkonidae), avec la description de quawe especes nouvelles. Rev. sons of serum albumins. Proc. Intern. Symp. African Vert., Bonn: Suisse Zool. 93:475-505. 479-494. Platnick, N. I., C. E. Griswold, andJ. A. Coddington. 1991. On miss- King, M. 1987. Monophyleticisrn and -polyphyleticism - in the ing entries in cladistic analysis. Cladistics 7:337-343. Gekkonidae: A chronlosomal perspective. Australian J. Zool. Russell, A. P. 1972. The foot of gekkonid lizards: A study in com- 35541-654. parative and functional anatomy. Ph.D. dissertation, Univ. Lon- IUuge, A. G. 1962. Comparative osteoloby of the eublepharid lizard don, England. genus Cokonyx Gray. J. Morph. 110:299-332. . 1976. Some comments concerning interrelationships 1967a. Higher taxonomic categories ofgekkonid lizards and amongst gekkonine geckos. In A. d'A. Bellairs and C. B. Cox their evolution. Bull. Amcr. Mus. Nat. Hist. 135:l-60. (eds.), Morphology and Biology of Reptiles. Linn. Soc. Symp. . 1967b. Systematics, phylogeny, and zoogeography of the lip 3:217-244. ard genus Dij~lodactylusGray (Gekkonidae). Australian J. Zool. . 1977a. The genera Rhoptropus and Phelsuma (Reptilia: 15:1007-1108. Gekkonidae) in southern Africa: A case of convergence and a --- . 1976. Phylogenetic relationships in the lizard family reconsideration of the biogeography of Phelsuma. Zool. Africana Pygopodidae: An evaluation of theory, methods and data. Misc. 12:393-408. Publ. Mus. Zool. Univ. Michigan 1521-72. . 1977b. The phalangeal formula of Hemidactylus Oken, 1817 . 1982. Cloacal bones and sacs as evidence of gekko~loidliz- (Reptilia, Gekkonidae): A correction and a functional explana- ard relationships. Herpetologica 38:348-355. tion. Zentralbl. Veterinaermed. Reihe C. Anat. Histol. Embryol. . 1983. Cladistic rclationships among gekkonid lizards. Copeia 6332.338. 1983:465-475. . 1978. The status of the lizard genera BlaesodactylusBoettger . 1987. Cladistic relationships in the Gekkonoidea (Squamata, and Homopholis Boulenger (Reptilia: Gekkonidae). Copeia Sallria). Misc. Publ. Mus. Zool. Univ. Michigan 173:l-54. 1978:25-29. . 1989. A concern for evidence and a phylogenetic hypothesis . 1979. Parallelism and integrated design in the foot struc- of relationships among l$picrates (Boidae, Serpentes). Syst. Zool. ture of gekkonine and diplodactyline geckos. Copeia 1979:l-21. 38:7-25. , and A. M. Bauer. 1988. Paraphalangeal elements of gekkonid . 1993a. Gekkonoid Lizard Taxonomy. International Gecko lizards: A comparative survey. J. Morph. 197:221-240. Society, Spec. Publ., San Diego, (A. 245 pp. --- , and A. M. Bauer. 1990. Hypertrophied phalangeal . 19931-3. Three-taxon transformation in phylogenetic infer- chondroepiphyses in the gekkonid lizard genus Phelsuma: Their ence: Ambiguity and distortion as regards explanatory power. Cla- structure and relation to the adhesive mechanism. J. Zool., Lon- distics 9:246-259. don 221:205-217. . 1993c. Asl,idit(?.$and the phylogeny of pythonine snakes. Simkiss, K. 1967. Calcium in Reproductive Physiology: A Compara- Records A~tstralianMus., Suppl. 19:l-77. tive Study oivertebrates. Reinhold Publ. Corp., NY. 264 pp. . 1994. Principles of. phylogenetic systematics and the infor- Stephenson, N. G. 1962. The comparative morphology of the head mativcncss of the karyotype in documenting gekkotan lizard rela- skeleton, girdles and hind limbs in the Pygopodidae. J. Linn. ions ships. I-Icrpctologica 50:210-221. Soc., London, Zool. 44:627-644. . 1995. Cladistic relationships of sphaerodactyl lizards. Amer. Steyn, W., and W. D. Haacke. 1966. A new webfooted gekko Mus. Novitates 3139:l-23. (Kaokogecko uanzyli gen. et sp. nov.) from the northwestern South , and A. J. Wolf. 1993. Cladistics: What's in a word? Cladistics West Africa. Cimbebasia 18:l-23. 9183-199. Swofford, D. L. 1985. Phylogenetic analysis using parsimony. Ver- Leviton, A. E., S. C. Anderson, K. Adlcr, and S. A. Minton. 1992. sion 2.4. Illinois Natural History Survey, Champaign, IL. Handbook to Middle East Amphibians and Reptiles. Society for Underwood, G. 1954. On the classification and evolution ofgeckos. the Study of Amphibians and Reptiles, Contr. Herpetol. 8, Ox- Proc. Zool. Soc. London 124:469-492. ford, OH. 252 pp. . 1957. On lizards of the family Pygopodidae: A contribution Loveridge, A. 1947. Revision of the African lizards of the family to the morphology and phylogeny of the Squamata. J. Morph. Gekkonidae. Bull. Mus. Comp. Zool., Harvard Univ. 98:l-469. 100:207-268. . 1957. Check list of the reptiles and amphibians of East Af- . 1970. The eye. In C. Gans and T. S. Parsons (eds.), Biology rica (Uganda; Kenya; Tanganyika; Zanzibar). Bull. Mus. Comp. of the Reptilia, 2:l-97. Academic Press, NewYork. Zool., I-Iarvard Univ. 117:153-362. . 1971. A modern appreciation of Camp's "Classification of Meriens, R. 1964. Der Eidechsenschwanz als I-Iaftorgan. Senckenb. the lizards." Pretace to the Society for the Study of Amphibians Riol. 45:117-122. and Reptiles facsimile reprint of C. L. Camp's "Classification of Moffilt, I,. A. 1973. The concept of primitiveness and its bearing on the Lizards": iii-xvii. the phylogenctic classification of the Gekkota. Proc. Linn. Soc. van Eijsden, E. H. T. 1983. Der haftfahige Greifschwanz des New South Wales 97275-301. Europaischen Bla~~ngcrgekkosPhyllodactylus europa~usGenC. Nixon, I<. C., and J. M. 1993. On outgroups. Cladistics Salamandra 19:l-10. 9:413-426. Vinson, J., and J.-M. Vinson. 1969. The saurian fauna of the Nussbaurn, R. A., and C.,J. Raxworthy. 1994. The genus Paragelzym Mascarcnc Islands. Mauritius Inst. Bull. 6:203-320. (Rcptilia: Sauria: Gekkonidae) in sou~hernMadagascar. J. Zool., Visser, J. 1987. A new Homopholis (Sauria, Gekkonidae) from north- London 232:37-59. ern Transvaal with a discussion of some generic characters. South Parker, H. W. 1926. The Neotropical lizards of the genera African J. Zool. 22:110-114. I~~)i~,idobl@l~aris,Pseudogonatodes, I.athrogecko, and Sphaerodactylu,~, Volobouev, V., and I. Ineich. 1994. A chromosome banding study of with the description of a new genus. Ann. Mag. Nat. Hist. 17:291- Ailuronyx seychellensis (Reptilia, Gekkonidae) and its phylogenetic 301, 624. affinities. J. Herp. 28:267-270. API'ENDIX g(ru(liclrrrirclrr 1276l)J-94; H. lromcln 127695-96, 127(i97 (4), 135413-14, 187538-39; JiTornopIrolz~/nr(:ialo 127698; 11. irwlbrr@r 127699; Hoplod(ccly11~s duvnctcrliz 127572 (3);H. ~mnul~lu.\127573 (4), 183505-506;II. nrcrcululus 183497, 183504; H. prrcrficu~ 127574 (2), 127575, 142501-529; iU1 of the followiilg skeletal ~llalerialis housed in the UDdMZ. L(#~idoblepl~rrr~~~rnion1epi.s 127804805, 13 1865-6(i; L. .srtnctaencarlae 12504 3, Each catalogue number represents onc spccime~l,unless indi- 125045, 127806 (3) ; L(:/)idodnc/y/u~,yu$)f?lri 127700; L. lugubr-is 12770 1-705, cated otherwise in parentlleses. Nomenclature follows Huge 127706 (4), 127707 (2), 127709,176636 (4), 187621-22; I..p?imilu\ 127708 ( 1993a). (2); 1.~~1~Durlonr\ 131149-51, 1:11154, 131190, 148819; L. jzenri 131189; 12ygodac/ylu,sccrprnsis 127710-1 1; 1.. conr-(rui 127712; I>.klrrg~i 143389-90; L. A(~Ir~ro.scoloho/(:.~/i,lznlr.\: 127404, 146749, 148898, 185901; A/,oe(Iu~-n pictut-atus 127713; Micr-og~!clu~pmicus 127786-87; Millo/isnunr mil-nhzlis portrlolicr 127610; A. tr(rrr.\71(~alrca12761 1, 176624; Agntnurrr prrsico 127612; 127714; N(l(:tu~a1:riorrxir 127633; N(~r7rdnsza,f~.\til,n127715; N(rullinu eIr:g(rn.s Ailrrmnyx .s(yclrcllr.nri.\127604, 148949, 186002; A/)m~racirri-ila 1312!24; 11. 127576 (2), 142536; N. gv7nmrus 142538; N.slrllc~tu, 142539, 187534-37; N. ~trr~~/)rtlcl~c~Nrrlr131157,131193; A. p\c~~do/)ulchrlln131208, 131210; A. pulch(~1ln lubercul~ctus127571; Nephrrrru.~aspn- 127577-79; N. lncw~znius127580; N. 13 1235; A. ~$)CILJ 129978-79, 137573, 173965; A. \fIlolntn 131176, 131180; lnjis 127581 (5); N. iuheele~r127582; N. milii 127591-94, 173749-55; N. Al:i.~trllig(~vcoch~(rnne127605,172721, 185902; A. g(:orx~e?rsb127606, 148 1 17; s/)/r,yrrLrus 127597; (k?/lur/lk(,,\lrPlrnr 12758485, 172585; 0. 117,/Lr~Inorat/L127583; 4. pvne.\i~rni.s127(i07-fi09, 172722-23, 185924; Bavr~yincy(:lurn 12'7507; 0. nionil~s127586; 0.lobu~to 127588-89; 0. tymni 127587; I~achydact~~lzrs I~ln(i\o(lnc/ylrl.\:ylr~boznini 127614; Hdtn b?o,\rlrnnn127615; Bu?ropu~lubol~~lrc/u.\ hibr(~?riz12771&19, 151105, 187801-802; I? cr~f~(,n.\b127720; I? g(,iljr 127721; 127613; Crilodactylodes arrrprrs 127616; Cnrplrod(~(lylu.\lneurr 127508; I?rricrcrclatcr\ 176637, 187803; I'(~l?natog(