molecular phylogenetic analysis of relationships of the ... · phylogenetic relationships of many...

q 2000 by the American Society of Ichthyologists and Herpetologists

Copeia, 2000(1), pp. 42–70

Molecular Phylogenetic Analysis of Relationships of the TropicalSalamander Genera Oedipina and Nototriton, with Descriptions of a New

Genus and Three New Species

MARIO GARCIA-PARIS AND DAVID B. WAKE

Sequences of two mitochondrial genes (385 base pairs of cytochrome b and ap-proximately 520 base pairs of 16S DNA) were gathered for 26 taxa of the MiddleAmerican plethodontid salamander genera Nototriton and Oedipina and from threeoutgroup members of the tribe Bolitoglossini. Phylogenetic analyses of these datareveal well-supported cladistic structure and demonstrate the paraphyly of the mosssalamanders of the genus Nototriton, which includes two well-defined clades. Oneclade, the sister taxon of Oedipina, corresponding to the Costa Rican and Honduranspecies of the picadoi, richardi, and barbouri groups, retains the name Nototriton. Anew name is required for the second clade, the sister taxon of Oedipina plus Noto-triton (sensu stricto). This clade, which we name Cryptotriton, is well supported mor-phologically and includes the species of the nasalis and adelos groups. A new speciesof Nototriton from Monteverde, Costa Rica, is described as Nototriton gamezi. Speciesof Oedipina fall into two clades that we treat as subgenera. Oedipina (sensu stricto)includes the longer-bodied, generally more slender and darker colored species andis the more speciose clade. Oedopinola includes the shorter-bodied, generally morerobust and lighter-colored species. Two new species of the latter clade are described,Oedipina maritima from the lowlands of northwestern Panama, and Oedipina savagei,from uplands of southwestern Costa Rica.

THE evolutionary radiation of bolitoglossinesalamanders in tropical America resulted

in an enormous diversification of morphologyand ecology, ranging from large, robust terres-trial forms (e.g., Pseudoeurycea bellii) to diminu-tive moss dwelling forms (e.g., Nototriton abscon-dens). One of the most strikingly derived mor-phologies within the bolitoglossines is displayedby the genus Oedipina, which comprises 18 cur-rently recognized species distributed from Chia-pas and the Yucatan Peninsula in Mexico tonorthwestern Ecuador. Oedipina is a well-sup-ported monophyletic group (Wake and Elias,1983), characterized as being the only tropicalbolitoglossine salamanders in which body elon-gation is a consequence of an increase in thenumber of vertebrae. All bolitoglossine sala-manders with the exception of Oedipina have 14trunk vertebrae, whereas Oedipina has from 18to 23 (Wake, 1966; Brame, 1968). Members ofOedipina typically are fossorial, and consequent-ly they are rarely seen. They range from sea lev-el (throughout their range) to elevations in ex-cess of 2200 m in Costa Rica. The status andphylogenetic relationships of many of the spe-cies in the genus were clarified by Good andWake (1997) based on analysis of an extensiveallozyme dataset. Their analysis resulted in ageneral phylogenetic hypothesis for relation-ships among the species of Oedipina (members

of the genus Nototriton were used as outgroups,following the phylogenetic hypotheses of Wakeand Elias, 1983). Two major clades were found.The first includes the species O. alleni and O.parvipes, which corresponds to the short-bodiedand often white-headed species of the O. parvi-pes group, for which the name Oedopinola(which we treat here as a subgenus) is available(Hilton, 1949). A second group includes the re-maining species (subgenus Oedipina), all elon-gated, black or brown attenuate forms compris-ing the uniformis group of Brame (1968). Theelectrophoretic evidence (large Nei genetic dis-tances, ranging from 0.37 to 2.67 among spe-cies) suggests that most species of Oedipina havehad a long independent evolutionary history.Most of the species of Oedipina studied show alarge number of autapomorphic alleles.

Phylogenetic relationships within the super-genus Bolitoglossa remain largely unresolved, butNototriton and Oedipina appear to be sister taxa(Wake and Elias, 1983). Oedipina and Nototritonshare a strongly heteromorphic X/Y sex chro-mosome system (Kezer et al., 1989; Sessions andKezer, 1991), also present in Thorius and Den-drotriton (Sessions and Kezer, 1991), as well asseveral osteological synapomorphies (Wake andElias, 1983). Nototriton includes diminutive, slen-der animals which share with Oedipina the pres-ence of a long tail that considerably exceeds

43GARCIA-PARIS AND WAKE—NEW TROPICAL SALAMANDERS

their snout–vent length and some osteologicalfeatures (reviewed later in this paper). The spe-cies of Nototriton typically inhabit hanging mossmats, epiphytic bromeliads, or surface litter andare distributed in two disjunct regions: (1) thegeologically ancient core region of nuclear Mid-dle America along the mountains of the Oaxa-can-Chiapan-Guatemalan-Honduran corridor;and (2) the Cordillera Central and northernTalamancan region of Costa Rica. Good andWake (1993) revised the Costa Rican species ofthe genus. The genus has been suspected to beparaphyletic (following Wake and Elias, 1983),and relationships among the picadoi, richardi,nasalis and adelos species groups (Papenfuss andWake, 1987) remain unresolved. Nei’s geneticdistances among Costa Rican species range be-tween 0.14 and 1.18, and phylogenetic analysisof the allozyme data supports recognition of apicadoi group (with four species) and a richardigroup (with two species) for the Costa Ricansegment of the genus (Good and Wake, 1993).All Honduran, Guatemalan, and Chiapan spe-cies (only four were known to Papenfuss andWake, 1987) were included in the nasalis group.

In this paper, we generate new phylogenetichypotheses for Nototriton and Oedipina based onpartial DNA sequences from two mitochondrialgenes, the cytochrome b (cyt b) encoding geneand 16S rDNA. We examine the data for the twogenes separately and analyze congruenceamong hypotheses generated from the two mi-tochodrial datasets. A combined mitochondrialhypothesis is used to test previous allozyme andmorphological hypotheses. We show that Noto-triton as presently constituted is paraphyletic,and we place some northern species of the cur-rent genus in a new genus. Phylogenetic analysisreveals several new candidate populations forspecies status, and three species are describedas new.

MATERIALS AND METHODS

Isolation, amplification, and sequencing of DNA.—A total of 78 specimens representing most ofthe known geographic range of Oedipina andNototriton were included in the DNA study.These represent 17 taxa (two described herein)of Oedipina and 12 taxa (one described herein)of Nototriton. Many of these were used in theallozymic studies of Good and Wake (1993,1997). We have determined that the taxon iden-tified as O. ignea by Good and Wake (1997) ismore appropriately assigned to O. stenopodia.Localities of origin, museum collection num-bers and GenBank accession numbers are givenin Appendix 1. Genomic DNA was extracted

from small amounts of frozen tissue or proteinextracts using NaCl following a protocol modi-fied from Miller et al. (1988).

Sequences of the first portion (385 basepairs) of the cyt b gene and of the large (16S)ribosomal subunit gene (rDNA) were obtained.Those regions of mtDNA were selected to re-cover a maximum of phylogenetic informationat all phylogenetic depths. We expect that cyt bevolves more rapidly than 16S; thus, it is ex-pected to be more useful in determining rela-tionships of terminal taxa, whereas 16S shouldbe more useful for resolving basal relationships(Mindell and Honeycutt, 1990; Hillis and Dix-on, 1991; Tan and Wake, 1995).

Fragments extending from the third positionof codon 7 through codon 135 of the Xenopuscyt b gene (Roe et al., 1985), and approximately520 bp of the 16S rDNA gene corresponding topositions 2510–3059 in the human mitochon-drial genome (Anderson et al., 1981), were am-plified via the polymerase chain reaction (PCR;Saiki et al., 1988) using the primers cyt b-2 (Ko-cher et al., 1989) and MVZ15 (Moritz et al.,1992) for cyt b, and 16Sar-L and 16Sbr-H (Pal-umbi et al., 1991) for 16S. PCR reactions con-sisted of 38 cycles with a denaturing tempera-ture of 92 C (1 min), annealing at 48–50 C (1min) and extension at 72 C (1 min) in a TechnePHC-1 thermocycler. PCR reactions were run ina total volume of 25 ml, using 0.6 units of Taqpolymerase (Cetus) in tubes containing 0.5pmol of each primer, 0.75 mM dNTPs, and 1.5mM MgCl2 in a pH 8.4 buffer with 50 mM KCLand 10 mM Tris HCl (final concentrations).Both heavy and light primers were used for PCRamplifications and sequencing.

Double-stranded templates were cleaned us-ing MicroSpin S-300 HR columns (PharmaciaBiotech). Four ml of double-strand productwere used as the template for cycle sequencingreactions in 10 ml total volume with the PerkinElmer Ready Reaction Kity to incorporate dye-labeled dideoxy terminators. Thermal cyclingwas performed using standard conditions. Cyclesequencing products were purified using etha-nol precipitation and separated by electropho-resis on a 6% polyacrylamide gel using an ABI377 DNA sequencer (Applied Biosystems).

Sequence alignment and phylogenetic analysis.—Par-tial sequences of cyt b were read from bothstrands and aligned to each other by eye in theprogram Sequence Navigatory (vers. 1.0.1, Ap-plied Biosystems). The resulting partial 16S se-quences were checked and aligned using CLUS-TAL in the program Sequence Navigatory(vers. 1.0.1, Applied Biosystems). Computer-

44 COPEIA, 2000, NO. 1

generated alignments were refined by eye andby comparing them to published secondarystructure models for 16S (Guttel and Fox, 1988;Guttel et al., 1993; Ortı et al., 1996; Ortı andMeyer, 1997). Sequence divergences were esti-mated using the Kimura 2-parameter (K2p) dis-tance (Kimura, 1980; determined using PAUP4.0b1a, D. Swofford, Smithsonian Institution) tocorrect for multiple hits. Corrected sequencedivergence within and among taxa is shown inAppendices 2–6.

All phylogenetic analyses were run using atest version of PAUP 4.0b1a. Phylogenetic hy-potheses were generated by maximum parsi-mony (MP), using the heuristic algorithm. Heu-ristic searches were done by stepwise randomaddition of taxa, with 10 replications and TBRbranch swapping with the MULPARS option ineffect and by collapsing zero-length branches.The minimum number of character changessupporting each branch, consistency index (CI,Kluge and Farris, 1969), and the retention in-dex (RI, Farris, 1989) were calculated. Trans-version (TV) to transition (TS) weights 1, 3, and10 were used for the analysis to determinewhether resolution was thereby increased, es-pecially at the base of the tree (Moritz et al.,1992), and to examine the potential effects ofhomoplasy at more rapidly evolving sites. Weperformed 100 bootstrap (bs) replicates to doc-ument support for individual nodes (Felsen-stein, 1985). Two schemes were applied to usethe 16S positions affected by gaps: (1) gaps weretreated as missing data; and (2) each INDELwas treated as a character, independent of itssize, thus adding 19 characters (9 parsimony-in-formative) to the dataset, with each gap treatedas a single evolutionary event (present or ab-sent). We also generated neighbor-joining (NJ;Saitou and Nei, 1987) trees based on K2p dis-tances, using 1000 bootstrap replicates. Maxi-mum likelihood (ML) was used, with the heu-ristic algorithm, and the Hasegawa-Kishino-Yano model (Hasegawa et al., 1985). ML set-tings used base frequencies, proportion of sitesassumed to be variable, the gamma distributionshape parameter, and transition:transversion ra-tio as estimated via maximum likelihood. Con-sistency among topologies was examined withthe COMPARE TREES option in MacClade.

A phylogenetic hypothesis for the combineddataset (cyt b and 16S) was inferred using max-imum parsimony (MP). In all heuristic searchesgaps were treated as missing data, 10 random-taxon-addition replicates were conducted, andtree bisection and reconnection (TBR) was thebranch-swapping algorithm used.

For all phylogenetic analyses representatives

of two tropical bolitoglossine genera, Bolitoglossacerroensis and Dendrotriton rabbi, were used as se-quential outgroups. All analyses were rooted us-ing a nearctic bolitoglossine species Batrachosepsgabrieli.

Morphology and taxonomy.—Morphological com-parisons involved measurements of preservedspecimens with dial calipers. All measurementsare in mm. Standard length (SL) is the distancefrom the tip of the snout to the posterior mar-gin of the vent. Tooth counts are presented inthis order: left-right. Institutional abbreviationsare as listed in Leviton et al. (1985).

Criteria for recognizing species based on phy-logenetic analyses of molecular data are thoseoutlined by Good and Wake (1993), followingFrost and Hillis (1990). We name as new speciesunits that give evidence of evolutionary inde-pendence from other units. When moleculardata suggest phylogenetic independence (sub-stantially more variation between samples thanwithin accepted species), we also investigatemorphology. When molecular and morpholog-ical features both give evidence of divergence,we recognize the units as species, but in the ab-sence of morphological evidence, we are reluc-tant to name new taxa on the basis of mito-chondrial sequence data alone.

RESULTS

Sequence fragments of 385 bp and approxi-mately 520 bp were obtained for cyt b and 16Sgenes, respectively. For the 16S dataset, thealignment of the ingroup required applicationof six to 10 gaps per sequence. Insertion/dele-tion (indel) events affected between 0.6% and1.9% of the aligned sequence length, for a max-imum of 22 positions. Most indels were 1 bp inlength, and maximum indel length was 8 bp;nine indels were unique to samples or species.The aligned sequences (NEXUS files) are avail-able from the authors upon request.

Corrected sequence divergence (K2p) amongtaxa for cyt b and 16S fragments (Appendices2–6) is great in both genera, ranging as high as22.5% (cyt b) and 10.8% (16S) within Oedipina,and 30.7% (cyt b) and 17.9% (16S) within No-totriton (sensu lato). Substantial divergence isfound even within certain taxa currently rec-ognized as single species (e.g., O. poelzi, as muchas 13.7% for cyt b and 4% for 16s).

We analyzed the data separately and com-bined. We conducted NJ, MP, and ML analysesand found them to be highly concordant. Inthis section, we focus on the results of the MPanalysis. For Figures 1–4 results of bootstrap an-


Fig. 1. Strict consensus of the four most-parsimo-nious trees (688 steps, CI 0.47, RI 0.68) for 16S. Thereare 167 parsimony informative characters. Transitionsare weighted equally with transversions. Gaps aretreated as missing data. Bootstrap values (above 50%)for the MP analysis (based on 100 replicates, aboveline) and for the NJ analysis (based on 1000 repli-cates, below line) are shown. Broken lines indicatealternative positions of Nototriton richardi and Oedipinamaritima in the NJ analysis; bold lines indicate alter-native positions for N. richardi and the Oedipina gephyraclade when transversions are weighted three timestransitions.

Fig. 2. Strict consensus of 124 most-parsimonioustrees (1066 steps, CI 0.31, RI 0.70) for cyt b, showingall of the populations analyzed in this study. Bootstrapvalues for the MP analysis (above 50% based on 100replicates, above line) and for the NJ analysis (bs .70% below line, based on 1000 replicates) are shown.

alyses (bs) in both MP and NJ trees are shown;topologies are based on parsimony analysis ex-cept where noted. For 16S, we weighted transi-tions and transversions equally, and in our MPanalysis, we found two equally most-parsimoni-ous trees (Fig. 1; CI50.47, RI50.68). For cyt b(unweighted analysis), we found 124 equallymost-parsimonious trees (Fig. 2; CI 5 0.31, RI5 0.70). Analysis of the cyt b data using fullweighting (i.e., TV:TS 5 3:1 and differentialweighting of 3:5:1 for the three nucleotide po-sitions) found eight equally parsimonious trees(not shown; bs values calculated; CI 5 0.42, RI5 0.79). The MP analysis of our combined data

resulted in three equally parsimonious trees (bsvalues and decay indices are shown in Fig. 3; CI5 0.39, RI 5 0.63; Fig. 4 is a NJ phylogram withbs values indicated). The tree shown in Figure3 is fully resolved except for one polytomy (thethree trees differ with respect to close relativesin the N. picadoi group). The ML tree (notshown, likelihood score 8589.98) is highly con-cordant with the MP combined data tree in Fig-ure 3 (differing only with respect to arrange-ment of a few terminal taxa within clades).

Oedipina is well supported as a monophyleticclade. In the NJ analysis, this clade has bs valuesof 100% (16S) and 84% (cyt b), and 100% forthe combined analysis. The comparable values

46 COPEIA, 2000, NO. 1

Fig. 3. Strict consensus of the three most-parsi-monious trees (1622 steps, CI 0.39, RI 0.63) for com-bined datasets for cyt b and 16S, showing all of thepopulations for which both sets are available. Thereare 333 parsimony informative characters. Transitionsare weighted equally with transversions. Decay indexvalues are shown below lines. The symbol . 7 is usedto indicate decay indices above 7 for a given branch.Bootstrap values (above 50%) (based on 100 repli-cates, above line) are shown.

Fig. 4. Neighbor-joining tree showing branchlengths for combined datasets for cyt b and 16S, basedon all of the populations for which both sets are avail-able. Bootstrap values (above 70%) based on 1000replicates are shown.

for MP are 100% (16S) and 67% (cyt b) and100% for the combined analysis (decay indexgreater than 7). When transversions are weight-ed three times transitions, and using differentialweighting for the three positions (3:5:1) for cytb, the bs value rises to 73%. The subgenus Oed-ipina is well supported, with NJ bs values of 90%(16S), 76% (cyt b), and 100% for the combinedanalysis; comparable MP bs values are 69%(16S), less than 50% for unweighted, and 52%for fully weighted (as above; cyt b), with a bs of77% for the combined equally weighted analysis(decay index 3). There is also support for themonophyly of the subgenus Oedopinola, with NJbs values of 53% (16S) and 64% (cyt b), with78% for the combined analysis; comparable MPbs values are 60% (16S) and 56% (for cyt b un-weighted; less than 50% for fully weighted),with a bs value of 65% for the combined anal-ysis.

Within Oedipina (Oedipina) there is a well-de-fined uniformis group (maximum K2p 5 0.113for cyt b; 0.044 for 16S), including the taxa uni-formis, gracilis, and pacificensis (NJ bs of 76% for

16s and 97% for cyt b, 100% for combined anal-ysis; MP bs of 52% for 16S; bs of 93% for un-weighted and 95% for fully weighted analysis ofcyt b, with 95% for combined analysis, with adecay index greater than 7). Within this com-plex, there is good support (NJ 96% bs 16s, 97%bs cyt b, 100% combined; MP unweighted 86%bs 16S, 87% bs cyt b, 94% combined with a de-cay index of 5) for a clade of pacificensis anduniformis. There is no support (bs , 50%) fora clade comprising the 10 samples of uniformisfor our cyt b dataset (maximum K2p 5 0.046),and there is no support for pacificensis as a cladein this dataset (maximum K2p 5 0.040). Thetaxa grandis and poelzi form a clade (NJ 88% bs;MP 88%) with our 16S dataset, and althoughthis clade always appears in analyses of the cytb dataset, it is not well supported. In the com-bined analysis, there is good support for theclade (82% bs, decay index 3). Within thisclade, there is weak support for poelzi as a mono-phyletic group based on cyt b (the Monteverdesample, 57, is basal in the larger clade), andthere is support for monophyly of poelzi in the16s data (NJ 76% bs; MP 74%); however, thecombined analysis either gives support formonophyly (NJ 90% bs; MP 51%), or makespoelzi paraphyletic with respect to grandis (MP).


There is relatively great divergence within poelzi,with a K2p value of 0.093 for cyt b and 0.040 for16S between samples from Moravia de Chirripoand various populations in the Cordillera Cen-tral. There is also support for a clade includingpseudouniformis and cyclocauda (Costa Rican pop-ulation only; NJ 99% bs for 16S, 100% for cyt b,100% combined; MP 94% bs for 16S, 99% forcyt b, and 100% combined with a decay indexgreater than 7). The Honduran sample previ-ously assigned to cyclocauda (species C in thispaper) is not associated with the Costa Ricanpopulations assigned to that taxon, based onlyon our study of cyt b sequences (K2p distanceis 0.165), nor is it significantly associated withany other taxa. There is no support for any bas-al phylogenetic resolution in O. (Oedipina).

There is little basal resolution within O. (Oed-opinola), and divergences are relatively great.The recently described Honduran taxon O. ge-phyra (McCranie et al., 1993) is included in aweakly supported clade with the northern spe-cies elongata within O. (Oedopinola; NJ 52% bs16S, 55% bs cyt b, 80% bs combined; MP lessthan 50% bs 16S, cyt b). Populations associatedwith alleni cluster together (NJ 100%, MP 99%for 16S; NJ 94% bs, MP 68% for cyt b; NJ 100%,MP 98% combined, decay index 8). The varioussamples associated with complex and parvipesform a monophyletic group (NJ 94% bs for 16S,97% bs for cyt b, 100% combined; MP 84% bsfor 16S, 95% for cyt b, 98% combined with adecay index greater than 7). There is only weaksupport (NJ bs 52% cyt b) for a cluster of thealleni clade with complex and parvipes. There is acluster of gephyra and elongata in both datasetsbut with relatively weak bootstrap support.

There are two distinctive clades representingthe taxa currently assigned to Nototriton. Thefirst of these, to which the name Nototriton ap-plies, includes all of the Costa Rican taxa, as wellas three Honduran (barbouri, lignicola, limnospec-tator) and one Guatemalan (brodiei) taxa (NJ100% bs for 16S, 99% for cyt b, and 100% com-bined; MP 99% bs for 16S, 96% for cyt b, and100% combined with a decay index greaterthan 7). The second clade, which is named laterin this paper, includes Mexican (alvarezdeltoroi),Guatemalan (veraepacis and species A) and Hon-duran (nasalis) taxa and is also well supported(NJ 100% bs for 16S, 100% bs for cyt b, 100%combined; MP 100% for 16S, 97% for cyt b,100% combined with a decay index greaterthan 7).

Within Nototriton (sensu stricto), a northernclade includes the Honduran and Guatemalansamples (NJ 91% bs 16s, 57% bs cyt b, 88% com-bined; MP 87% 16S, 68% cyt b, 99% combined,

decay index 6). A group of Costa Rican speciesforms a second clade (NJ 99% bs 16S, 100% bscyt b, 100% combined; MP 90% bs 16S, 99% bscyt b, 100% bs combined with a decay indexgreater than 7). The position of the Costa Ricansamples of N. richardi is variable, being alterna-tively basal to either clade. For the MP analysisof 16S, it is basal to the northern clade, but iftransversions are weighted three times transi-tions, it is basal to the Costa Rican clade. Thisarrangement is also found in all neighbor-join-ing analyses (NJ 88% bs 16S, 58% bs cyt b, 86%combined). There is also support for this ar-rangement in the bootstrap analysis of MP trees(60% bs 16S, 61% bs cyt b, 67% bs combined).

The as yet unnamed northern clade of Noto-triton is alternatively a sister taxon of Dendrotriton(NJ 70% bs cyt b; 72% for combined data) or asister taxon of the Nototriton (sensu stricto)-Oed-ipina clade (MP 57% bs 16S); in general thisclade appears near the base of the ingroup inall trees. Within this clade, the two Guatemalanspecies are sister taxa (NJ 100% bs for 16S, 81%bs cyt b, 100% bs combined; MP 97% bs 16S,62% bs cyt b, 99% bs combined). In no instance,in any of our analyses, are the two main cladesof Nototriton (sensu lato) sister taxa.

Dendrotriton is the sister taxon of the entireassemblage (Oedipina 1 Nototriton sensu lato)considered here in the MP analysis of the 16Sdata and the combined data, but there is a po-lytomy of Dendrotriton, the newly recognizedclade within Nototriton, Bolitoglossa, and the Oed-ipina-Nototriton (sensu stricto) clade in our MPanalysis of the cyt b data. In fully weighted anal-yses of the cyt b data, the newly recognized cladeis basal to a clade (bs 65%) of Oedipina-Nototriton(sensu stricto) and Dendrotriton-Bolitoglossa (bs81% for this sister-group relationship). At thisphylogenetic depth, the 16S data are expectedto have the greater value. In the ML tree, Den-drotriton is the sister taxon of the new genus. Allanalyses (MP, ML, NJ) support a sister taxon re-lationship between Oedipina and Nototriton (sen-su stricto; MP 81% bs 16S, 55% bs cyt b, 91% bscombined, decay index 6).

DISCUSSION

Data analyses.—A high level of congruence ex-ists between the results obtained from analysisof the cyt b and 16S datasets. Accordingly, webelieve that the closest approximation to thephylogeny of the samples studied is likely tocome from analysis of the combined dataset.When the two datasets are combined, almost allof the bootstrap values in either the NJ or MPanalysis increase. At the level of the major

48 COPEIA, 2000, NO. 1

clades, our results appear to be relatively robust.There is support for a monophyletic Oedipina, amonophyletic Nototriton (sensu stricto), a mono-phyletic clade named later in this paper, and fora clade including Oedipina and Nototriton (sensustricto). The unnamed clade is not a memberof this latter clade, but it is a close, basal rela-tive, possibly the sister taxon.

Analysis of the combined data by MP resultedin only three most-parsimonious trees, andthere was almost complete resolution withmainly high bootstrap values (Fig. 3). The taxaresponsible for the alternative trees were closerelatives in the picadoi clade of Nototriton. Thetaxonomic position of the population fromMonteverde, Costa Rica (sample 16), is some-what ambiguous (e.g., Good and Wake, 1993),and our data add support to its distinctivenessand status as an unnamed species. Previous hy-potheses of relationships among the threeclades of Oedipina envisioned the Guatemalanpopulation previously called O. ignea (now O.stenopodia) as being a sister taxon of one or theother of these clades, but our analysis places itclearly outside of this cluster.

Phylogenetic relationships and species recognition.—The molecular phylogenetic hypothesis for Oed-ipina of Good and Wake (1997) is mostly con-gruent with Brame’s (1968) morphological hy-pothesis. However, Brame placed all species ofOedipina in only two groups, a uniformis group(equivalent to our subgenus Oedipina) and aparvipes group (equivalent to our subgenus Oed-opinola), whereas Good and Wake recognizedthree clades (cyclocauda, poelzi, and uniformis)within Brame’s uniformis group. Good and Wake(1997) found two most parsimonious trees. (Be-cause of a printing error only one tree appearsin the published article. A correction was pub-lished in Revista de Biologıa Tropical 45, number4). We were unable to obtain DNA from all taxasampled by them, but we investigated severalspecies unavailable to them. It is not possible todo a combined analysis of the allozyme andDNA data that would be meaningful. However,certain comparisons are appropriate. Both da-tasets identify a basal dichotomy between thesubgenera. Within the subgenus Oedipina, O. cy-clocauda, and O. pseudouniformis are sister taxain both analyses and form a well-supportedclade. There is a well-supported clade that in-cludes a restricted uniformis group (O. gracilis,O. pacificensis, O. uniformis). Both analyses rec-ognized a poelzi group and also recognized anorthern (identified as O. ignea by Good andWake, and as O. stenopodia and species B in thispaper) unit. However, basal relationships of

these four clades, each well supported individ-ually, are uncertain, and there is no support forany particular arrangement.

Our analysis is relevant to issues of speciesrecognition and species level taxonomy. Goodand Wake (1997) argued that allozymic datawere insufficient to determine whether popu-lations from Honduras (74) and Costa Rica (39,40) identified as O. cyclocauda were conspecific.Our analysis shows that they are not sister taxawith respect to mtDNA, and given the large ge-netic distance separating them, the Honduraspopulation warrants description as a new taxon;we lack adequate material for such a descrip-tion. A population from Moravia de Chirripo,Costa Rica (57), assigned by Good and Wake(1997) to O. poelzi, is well differentiated fromother members of that taxon both in allozymesand mtDNA. It is usually basal to other popu-lations of O. poelzi in phylogenetic analyses, butit is basal to a larger clade including also O.grandis in some analyses (e.g., cyt b). It probablywarrants recognition as a distinct taxon, but atpresent, we lack adequate material to prepare adescription. A population from Honduras (73)that clusters with O. stenopodia might be assign-able to O. ignea or it may represent a distinct,undescribed taxon. Our analysis also supportsthe decision of Good and Wake (1997) to sub-divide Brame’s (1968) O. uniformis into threetaxa (O. uniformis, O. gracilis, O. pacificensis).There is no support for a sister-taxon relation-ship of the Damas sample (50) of O. pacificensisand the other O. pacificensis samples in the cytb data. The K2p distances are small within theclade.

Populations (30–38, 47) form a clade that isdeeply divergent internally. Authors have vari-ously identified these populations as O. alleni,O. complex, and O. parvipes. Within this clade,populations 30–34 stand apart, but even thisgroup of five samples is deeply differentiated.Two (30–32) of the three populations we havestudied were available for allozymic work, andalthough samples were small, the populationswere very similar (Nei D approximately 0.02).In contrast, we measured large distances forboth genes studied, and furthermore a thirdpopulation (33,34) is equally distinct. This thirdpopulation occurs at much higher elevationthan the others (near Las Cruces, southwesternhighlands of Costa Rica) and differs in colora-tion (dorsal light stripe) and size (smaller, morerobust, with a short, rounded snout). It is thebasis for the inclusion of O. complex in the CostaRican fauna by Savage and Villa (1986), but itis not assignable to that taxon. Below, we de-scribe this population as a new species. The type


locality of O. parvipes is in northwestern Colom-bia, and we have no near topotypic materials forthe present study. Furthermore, the samplesthat we have had from Panama that have beenassigned to this taxon are well differentiatedfrom each other. Sample sizes are generally toosmall to permit appropriate morphologicalcomparisons, but the most divergent population(47) from western Panama is both morpholog-ically and genetically distinct from populationsin the central part of the country; we describeit below as a new species. The remaining pop-ulations (35, 36) of O. parvipes from Panama aregenetically differentiated from each other, andwe suspect that they are also differentiated fromtopotypic material; further taxonomic revisionmust await the collection of additional speci-mens. Finally, O. complex appears to be appro-priately named, because the two populationsthat we have studied that are assigned to thistaxon (near topotypic material from centralPanama, sample 38, and a population from thecentral Panamanian Cerro Campana, sample37) are not sister taxa in any of our analyses.These two populations are relatively close geo-graphically, yet they differ by 11.1% in cyt b and3.9% in 16S. More than one taxon probablyshould be recognized, but we lack sufficient to-potypic material to conduct an appropriatemorphological comparison at this time.

Within Nototriton there is good support in thecombined analysis for a monophyletic Costa Ri-can picadoi group (100% bs with a decay indexof 9) and for a northern group (98% bs with adecay index of 6), which we here identify as thebarbouri group. The Costa Rican richardi groupclusters (57% bs with a decay index of 1) withthe picadoi group. This is in accord with thefindings of Good and Wake (1993) for the CostaRican species. Within the picadoi group, a pop-ulation (16) previously assigned to N. abscondensfrom the Monteverde region displays problem-atic relationships. Given its distinctive morphol-ogy and allozyme profile (Good and Wake,1993), it merits description as a distinct species(see below). Within the barbouri group our re-sults support recognition of the recently de-scribed N. brodiei (Campbell and Smith, 1998),N. lignicola (McCranie and Wilson, 1997b), andN. limnospectator (McCranie et al., 1998). Wenote that one of the samples of N. barbouri (19)differs substantially from the other two samples(17, 18) in the cyt b tree (Fig. 1) and may rep-resent an unrecognized taxon.

Several morphological and reproductive syn-apomorphies support a clade containing Oedi-pina, Nototriton, and the new genus: the ulnareand intermedium are fused in the carpus; distal

tarsals four and five are fused; there is no oper-cular stilus; and all species examined have X-Ysex chromosomes (Sessions and Kezer, 1991;Lynch and Wake, 1978; Wake and Elias, 1983).Species of Nototriton lay eggs in moss clumps intrees, but instead of guarding the eggs, a uni-versal trait of non-bolitoglossine plethodontidsand widely present in bolitoglossines, femalesabandon the eggs (Good and Wake, 1993;McCranie and Wilson, 1992). An undescribedspecies belonging to the new genus named inthis paper and a new species of Oedipina fromPanama (see below) also lay unguarded eggs.Thus, the three genera treated in this paper dis-play a unique synapomorphy because they arethe only members of the supergenus Bolitoglossathat are known to abandon their eggs (we haveno information for some putative relatives suchas Dendrotriton and Bradytriton). Synapomorphiesunique to Oedipina among these taxa includeincreased numbers of trunk vertebrae, presenceof a spur on the squamosal bone, loss of theprefrontal bone, loss of the tibial spur, and pres-ence of mental gland clusters in a small patchat the mandibular symphysis (Wake, 1966; Wakeand Elias, 1983). We know of no synapomor-phies unique to Nototriton among these taxa, butspecies belonging to the genus as herein reviseddiffer from those belonging to the new genusin cranial structure. The prefrontal is relativelylarge in the new genus and is pierced by thenasolacrimal duct, whereas it is relatively smallor absent in Nototriton, and the nasolacrimalduct passes between the prefrontal and the na-sal (partly pierces the prefrontal in some indi-viduals of some species); the nasal is large inNototriton but relatively to very small in the newgenus; there is a distinct, tooth-bearing preor-bital process on the vomer of Nototriton (andOedipina), whereas this process is lacking in thenew genus (McCranie et al., 1998; Good andWake, 1993; Lynch and Wake, 1978). In all ofthese characters, the states in Nototriton are an-cestral, whereas the states in the new genus con-stitute synapomorphies. Within Nototriton, oste-ological differences have been noted amongtaxa, and the picadoi and barbouri groups appearto be distinct from each other osteologically,whereas richardi is poorly known and appears tohave unique osteology (McCranie et al., 1998;Good and Wake, 1993).

SYSTEMATICS

Phylogenetic analysis and taxonomy.—-We believethat taxonomy should reflect robust phyloge-netic analysis within the severe constraints ofthe Linnean system. We have identified three

50 COPEIA, 2000, NO. 1

large clades that are well supported by sequencedata. Two of these clades, Nototriton and theclade named later in this paper, are generallysimilar in overall morphology and ecology,whereas Oedipina differs dramatically from theother two in morphology and ecology. The spe-cies of Oedipina have long trunks and extraor-dinarily long tails, and most (perhaps all) aresemifossorial in habits. In our analyses, thesethree clades usually form a monophyleticgroup, but not always, and monophyly is notwell supported. For these reasons, we decline toplace the three in a single genus, for which thename Oedipina would have priority. Becausethere is no support in our data for monophylyof Nototriton and the remaining clade, we believethat the most logical and useful taxonomy is toplace the latter clade in a new genus.

The use of subgenera for clades that are mor-phologically similar but differ in molecularcharacters was advocated by Jackman et al.(1997) for Hydromantes and Batrachoseps. In bothcases, the genera are distinguished from othersby numerous synapomorphies, whereas the sub-genera are only weakly supported by morpho-logical synapomorphies but well supported inphylogenetic analyses of molecular data. This isalso true of the two major clades we have iden-tified within Oedipina, and accordingly we heretreat these as subgenera.

DESCRIPTION OF A NEW SPECIES OF OEDIPINA

FROM PANAMA

Oedipina maritima n. sp.Maritime Worm Salamander

Holotype.—-USNM 529981, an adult female fromEscudo Camp, West Point, Isla Escudo de Vera-guas, Prov. Bocas del Toro, Panama, approximate986.19N, 8184.59W, collected 28 March 1991, by R.I. Crombie.

Paratypes.—USNM 529982–529985; USNM Fieldnumber 195536 (to be deposited in a collectionin Panama), MVZ 219997, same data as holo-type; KU 116681, E end Isla Escudo de Vera-guas, Prov. Bocas del Toro, Panama, 4 m. elev.

Referred material.—USNM 313607 (two clutchesof eggs), same data as holotype.

Diagnosis.—A small (maximum size 46.2 SL),relatively slender member of Oedipina (subge-nus Oedopinola) with a narrow head and a longsnout. Distinguished from Oedipina alleni by itsnarrower hands and feet, more elongate digitsand darker coloration; from Oedipina parvipes by

its dark ventral coloration; from Oedipina savageiby having a narrower and more pointed head,and fewer maxillary teeth; from Oedipina complexby its narrower head, pointed and elongatesnout (rather than short and rounded), small,laterally oriented eyes, fewer maxillary teeth,and narrower feet with pointed digits; from Oed-ipina carablanca by its much smaller size and lessrobust habitus, small, laterally oriented eyes,narrow and pointed head, narrow hands andfeet, and pointed digits. The first four of theabove species are distinguished from O. mariti-ma by extensive sequence differences in the mi-tochondrial genes 16S rDNA and cytochrome b.

Description.—This is a relatively small, slenderspecies; adult SL for four males is 39.7–46.2, x43.3; for three females 34.6–44.3, x 39.7. Thehead is small, narrow, cylindrical and generallypointed, with an elongated, blunt-tipped snout;SL averages 10.3 times head width in 4 males(range 9.7–11.3) and 9.8 in 3 females (9.4–10.3). SL is 6.5 times head length in both males(6.2–7.1) and females (5.1–7.2). Nostrils are tinyand barely discernable. Nasolabial protuberanc-es are inconspicuous in females but slightly en-larged in males, where they extend slightly assomewhat swollen light-tipped extensions overthe underslung lower jaw. Snouts are morepointed in females than in males as a conse-quence. Eyes are small and inconspicuous, bare-ly extending beyond the lateral margins of thehead; they are directed mainly laterally, ratherthan frontally as is the situation in most bolito-glossines. The suborbital groove does not inter-sect the lip. There are 1–2 slightly enlarged pre-maxillary teeth in males, located in a far for-ward position outside the mouth, lying just infront of the small, ill-defined mental gland thatlies at the extreme anterior margin of the lowerjaw; the 1–2 premaxillary teeth in females aretiny and located well within the mouth. Maxil-lary teeth, when present, are tiny and incon-spicuous; they range in number from 0 (3males)–8 (x 2.0) in four males, and from 2–8 (x5.3) in three females. Vomerine teeth average15.5 (13–18) in four males and 17.3 (15–20) inthree females; the very small teeth are borne ina long row. There are 17 costal grooves betweenthe limbs, counting one each in the axilla andthe groin (18 trunk vertebrae). Limbs are rela-tively short; limb interval averages 8.1 in fourmales (7.5–8.5) and 8.0 in three females (7.5–9). Hands and feet are tiny, narrow and elon-gate. The digits are syndactylous, although thetips of the longest central digits are free. Thefree tips are slender and sharply pointed, andthe points are often curved toward the midline


Fig. 5. Outlines of the shape of the right foot forrepresentative samples of six species of Oedipina (Oed-opinola), drawn with camera lucida. (A) O. savagei,LACM 109558 (39.3 SL), holotype; (B) O. alleni MVZ190857 (52.4 SL), near Damas, Prov. Puntarenas, Cos-ta Rica; (C) O. maritima USNM 529981 (44.3 SL), ho-lotype; (D) O. parvipes LACM 134872 (53.9 SL), BarroColorado Island, Panama; (E) O. complex MVZ-DBW5787 (37.2 SL), Cerro Campana, Panama; (F) O. com-plex MVZ-DBW 5105 (35.0 SL), Peninsula Bohıo, Prov.Colon, Panama.

axis of the limb (Fig. 5). Fingers, in order ofdecreasing length, are 3-2-4-1; toes are 3-4-2-5-1.The tail is round, narrow in cross-section andrelatively long, tapering along the last third ofits length. Tails of most individuals may be atleast partly regenerated, but all of them greatlyexceed SL; SL averages 0.59 times tail length in4 males (0.53–0.68) and 0.57 in 2 females (0.53–0.62).

Measurements of holotype.—Head width 4.3; snoutto gular fold (head length) 6.2; head depth atposterior angle of jaw 2.3; eyelid width 0.6, eye-lid length 1.8; eye to nostril 1.6; anterior rim oforbit to snout 1.9; horizontal orbit diameter 0.9;interorbital distance 2.3; distance separatingeyelids 1.3; nostril diameter 0.1; snout projec-tion beyond mandible 0.8; distance from eye topostorbital groove 2.2; snout to posterior angleof vent (standard length) 44.3; snout to anteriorangle of vent 39.4; snout to forelimb 10.0; axillato groin 27.6; limb interval 9; shoulder width3.2; tail length 71.9; tail width at base 3.3; taildepth at base 3.3; forelimb length (to tip of lon-gest digit) 5.7; hind-limb length 7.0; hand width

1.1; foot width 1.8; free length of longest toe0.6. Numbers of teeth: premaxillary 2; maxillary0–2; vomerine 9–11.

Coloration of the holotype in alcohol.—Dorsal colordark brown, almost chocolate on the sides, ligh-ter along an indistinct dorsal band. Dorsal andlateral coloration is separated on the first quar-ter of the body by an almost indistinct light col-ored broken line, more apparent and broaderon the posterior edges of the head. A whitishband extends from the upper jaw to the inter-ocular region, covering the area between thenostrils and the eyes. The posterior edge of thislight band is not well defined and dorsal darkcoloration penetrates into it. The dorsal color-ation becomes lighter posteriorly, and along thedorsal portions of the anterior half of the tail.This light coloration does not uniformly coverthe dorsal surface of the tail. Dorsal sides of thelimbs show the same dorsal color with a lightpatch at the base of the limbs, close to its inser-tion to the trunk. Tips of digits are unpigment-ed; hands and feet are lightly colored with dark-er pigment in the area between the digits. Ven-tral coloration is lighter than dorsal, with all theventral regions from throat to vent and tail cov-ered with dense melanophores, giving a grayish-brown appearance. The throat and undersur-face of the limbs are lighter; hands and feet arealmost unpigmented ventrally. Small white dotsare spread along the flanks and the tail at lowdensity, less concentrated (scattered) on the lat-eral sides of the venter.

Color variability.—Dorsal coloration is lighter injuvenile and submature (KU 116681) speci-mens, markedly contrasted with the dark flankcoloration. They show a marked dark midver-tebral line running from the posterior edge ofthe head to the tail origin where it fades away.Tail dorsal color in juveniles is very light, but itis not sharply separated by a defined line as itis on the body. Some adult specimens (e.g.,USNM 529983) show a more contrasted lighttail pattern, always less developed than in juve-niles. The extent and definition of the anteriorlight lateral band varies from well marked andbroad at the ‘‘parotoid’’ level, to almost com-pletely absent (USNM 195536), as does the dor-sal light band. White dots along the flanks aredenser in some specimens; these specimens alsohave a few pale dots extending to the venter.Hatchlings display a bold pattern of very darkflanks and side of the head and tail with a broadlight yellowish dorsal band, extending from thesnout to the tail tip. The venter is pale yellow.The main light areas show an obscure suffusion

52 COPEIA, 2000, NO. 1

of melanic pigmentation especially evident inthe middorsal portion of the trunk and near thetail tip.

Color notes were recorded in life for USNM529985 by R. I. Crombie, as follows: dorsally me-dium brown with black mottling and silverflecks. A narrow black lateral area, less pro-nounced on tail, bordered ventrally with anarea of intense silver freckling (less concentrat-ed on lateral tail). Belly pale, unmarked graybut with some silvery flecking on throat. An in-distinct chestnut postocular stripe and indistinctdull chestnut on snout. Iris dark. Crombie re-corded the following color notes for USNM195321: dorsally medium brown with black mot-tling and silver flecks. Belly pale, unmarked graybut with some silver flecking on throat.

Osteology.—Some information on osteology hasbeen derived from radiographs of the holotypeand five paratypes. All specimens have 18 trunkvertebrae, and small ribs are borne on all in thelargest specimen and all but the last in others.Numbers of caudal vertebrae in individuals withapparently unregenerated tails range from 52 to57 (the last in the largest individual). Skulls arerelatively well articulated but little detail is evi-dent. The snout is generally well developed andprotuberant, and the nasals are curved with amedial portion that approaches an upward-growing portion of the vomer. There is no fon-tanelle between the frontals and parietals, andthe frontals have a large facial portion. Digitsare generally poorly ossified. The most com-plete phalangeal formula is 0-1-2-2 (in one ma-nus of the holotype) and 1-2-3-2-2 (in the largestindividual). However, the modal formulae are0-1-2-1 and 1-2-2-2-1. Tibial spurs are not evi-dent.

Habitat and reproduction.—Salamanders on IslaEscudo de Veraguas were collected in decayingfronds and associated moist litter near a fallenpalm in a coconut palm grove in late March,1991. The eggs were found inside a pile of co-conut trash at the base of a tree (Terminalia) onthe beach at Guayami Settlement, SE part of is-land, less than 5 m from the ocean. The eggsbelonged to two clutches, each containing sixembryos; there is one rotted egg that may be-long to either clutch. One clutch includes to-tally unpigmented embryos that are wellformed and have well-formed limbs, but thelimbs, although moderately long, have no digitsor digital rudiments. The other clutch includesnear full-term embryos that are well pigmented(see Color variability, above). These embryosbegan hatching on the way back to camp from

the field. One newly hatched embryo is 11.8mm in total length (8.7 SL) and retains gills.The gills are three lobed, one relatively largeand palmate, a second smaller but palmate, andone relatively large and slender and branchedat the base.

Etymology.—The name maritima is derived frommaritimus (L), meaning of the sea, in referenceto the type locality of this species on a low-lyingisland in the Caribbean Sea.

Comment.—Oedipina maritima does not resembleO. alleni in morphology (it is more slender andhas a narrower and more pointed head, and itgenerally lacks dorsal light coloration), and it isvery distinct in its DNA sequences. It differsfrom O. parvipes from central Panama pro-foundly in DNA sequences (4.8–4.9% for 16S,7.8–8.1% for cyt b), but appears to differ oth-erwise only in coloration (we have no allozymedata). However, we also note that evidence ofconspecificity of Panamanian and Colombianpopulations of O. parvipes is weak, and A. H.Brame Jr. and D. B. Wake have long planned toundertake a detailed analysis of this questiononce adequate material is available. According-ly, we decided to describe what is clearly an in-dependent taxon at this time and reserve acomplete analysis of Panamanian and SouthAmerican members of this genus to a later date.

This species is known only from Isla Escudode Veraguas. If it is restricted to this island, it isthe only tropical salamander that is endemic toan island. At least three species of this genusoccur on the adjacent mainland, two of whichappear to be undescribed. KU 116682, a 17.5SL juvenile from the mouth of Rio Cahuita,Prov. Bocas del Toro, Panama, 1 m. elev., cannotbe classified with certainty but may be assign-able to O. maritima.

DESCRIPTION OF A NEW SPECIES OF OEDIPINA

FROM COSTA RICA

Oedipina savagei n. sp.Savage’s Worm Salamander

Figure 6A

Holotype.—LACM 109558, an adult female fromFinca Las Cruces, 6 km S San Vito de Java, Prov.Puntarenas, Costa Rica, 88479350N, 828579300W,approximately 1200 m elev., collected 22 May1971 by R. W. McDiarmid and associates.

Paratypes.—LACM 109556–557, LACM 145447–145450 (four specimens), USNM 219122, samelocality as holotype, collected on different dates;


Fig. 6. (A) Oedipina savagei. LACM 145447, an adult male (35.7 mm SL) from Las Cruces, Prov. Puntarenas,Costa Rica, collected in August 1971. Photograph by R. W. McDiarmid. (B) Oedipina alleni. MVZ 221317, anadult female (46.4 mm SL) from 3.5 km SE Damas, Prov. Puntarenas, Costa Rica, elev. 3–5 m, collected 2August 1992. Photograph by J. Hendel.

MVZ 229360, UCR LDG 961327 (temporarilyDBW 5786), Paraguas Ridge (slopes of CerroZapote), ca. 4 km W Agua Buena, Prov. Puntar-enas, Costa Rica, approximately 88459N, 82899W,approximately 1400 m elev.

Referred material.—UCR 8210, Finca Cafrosa nearLas Mellizas, Prov. Puntarenas, Costa Rica, 1300m elev.

Diagnosis.—A small (maximum size 39.3 SL),moderately robust member of Oedipina (subge-nus Oedopinola) with a narrow head and a rela-tively short, rounded snout. Distinguished fromOedipina alleni by having more maxillary teethand a light dorsal stripe; from Oedipina maritimaby its more robust habitus, broader and morerounded head, more maxillary teeth, moreprominent and more frontally oriented eyes,and persistent light dorsal stripe; from Oedipinaparvipes by its broader head with a shorter, morerounded snout, more maxillary teeth, and lightdorsal coloration; from Oedipina complex by itsless bluntly rounded snout; from Oedipina cara-blanca by its much smaller size and less robusthabitus, and narrower hands and feet; and from

the first four of the above species by extensivesequence differences in the mitochondrialgenes 16S rDNA and cytochrome b.

Description.—This is a relatively small, moderate-ly robust species; adult standard length (SL) forthree males is 35.6–38.5 mm, x 36.8; for threefemales 34.3–39.3, x 37.5. The head is moderatein width, being broader in males. Males haveshorter, blunter snouts than females, in whichthe snout is somewhat narrow and slightly point-ed. SL averages 8.6 (range 8.3–8.8) times headwidth in four males, and 8.7 (8.4–9.0) in threefemales; SL averages 6.0 (5.6–6.4) times headlength in four males, and 6.6 in three females(6.0–7.2). Nostrils are evident and moderatelylarge for the genus. Nasolabial protuberancesare prominent in both sexes but especially inmales, in which they are swollen, knoblike struc-tures that extend below the somewhat under-slung lower jaw. Eyes are of moderate size andstand out from the head, extending laterally be-yond the limits of the head. They have a mod-erate to strongly frontal orientation. The sub-orbital groove does not intersect the lip. Thevery small premaxillary teeth (3–4 in number)

54 COPEIA, 2000, NO. 1

lie well within the mouth in both sexes, exceptfor USNM 219122, in which there is a singlesmall tooth lying outside of the mouth. This isthe only specimen that has a mental gland, asmall cluster of openings in a patch of pigmentlying just behind the mandibular symphysis.Maxillary teeth are small and moderately nu-merous, ranging between 7 and 18 (x 12.3) infour males and 7 and 15 (x 11) in three females.Vomerine teeth range between 11 and 22 (x15.8) in four males and 11 and 12 (x 11.7) inthree females; these very small teeth are bornein a long row. There are 17 (7 individuals)–18(1 individual) costal grooves between the limbs,counting one each in the axilla and the groin;accordingly we infer that there are 18 (rarely19) trunk vertebrae. Limbs are of moderatelength for this clade; limb interval averages 6.6(6–7) in four males and 6.8 in three females(6.5–7). Hands and feet are small and are mod-erately broad but short. The digits are syndac-tylous, and only the triangular tips of the lon-gest central digits are free (Fig. 5). Fingers, inorder of decreasing length, are 3-2-4-1; toes are3-2-4-5-1. The tail is round, narrow in cross sec-tion and relatively short. Tails of most individ-uals may be at least partly regenerated or werecut for extraction of DNA prior to measure-ment. Apparently complete tails exceed SL, andin the individual with the longest tail (one ofthe smaller males), SL is 0.65 tail length.

Measurements of holotype.—Head width 4.5; snoutto gular fold (head length) 5.9; head depth atposterior angle of jaw 2.9; eyelid width 0.9, eye-lid length 1.5; eye to nostril 1.1; anterior rim oforbit to snout 1.4; horizontal orbit diameter 1.1;interorbital distance 2.3; distance separatingeyelids 1.7; nostril diameter 0.2; snout projec-tion beyond mandible 0.5; distance from eye topostorbital groove 2.2; snout to posterior angleof vent (standard length) 39.3; snout to anteriorangle of vent 35.7; snout to forelimb 8.8; axillato groin 23.9; limb interval 6 1/2; shoulderwidth 3.4; tail length 39.1 (apparently at leastpartly regenerated); tail width at base 2.7; taildepth at base 2.8; forelimb length (to tip of lon-gest digit) 6.1; hind-limb length 7.1; hand width1.4; foot width 1.9; free length of longest toe0.6. Numbers of teeth: premaxillary 4; maxillary5–2; vomerine 5–6.

Coloration in alcohol.—This description is basedmainly on two recently collected specimens(MVZ 229360, UCR LDG 961327), because col-oration of the holotype and other paratypes hasbeen modified by the long period of preserva-tion. Ontogenetic change in coloration is clear-

ly evident in this species. Individuals below 30mm SL have a dorsal stripe or band that isprominent and mainly unmarked. The stripeextends from the tip of the snout to the tip ofthe tail, interrupted on the midline of the snoutby a dark incursion of pigment forming a nar-row streak, and at a few places along the backof the head or on the trunk by some obscuredark pigment. In older individuals, there iswidespread incursion of dark pigmentation intothe band from the lateral surfaces (Fig. 6A). Injuveniles the lateral surfaces are dark brownfrom the facial region and below the upper eye-lids all along the head, body and tail to its tip.This lateral stripe persists in adults. The dorsalband is always lighter than the lateral surfaces,even when extensive melanization has occurred.In all individuals the area including the back ofthe head and extending forward to and includ-ing the upper eyelids remains light and bearswhite pigment. The snout becomes increasinglymottled with increasing size as the white pig-ment is progressively restricted to the vicinity ofthe nostrils and nasolabial grooves and protu-berances. The dorsal stripe in adults is a me-lange of streaks or irregular spots, of melaninthat becomes increasingly dense with increasingsize (this observation is based in part on whathappens in close relatives). The melanic areasinterrupt and gradually replace the white pig-ment, which eventually is reduced on the backto some superficial and irregular spots. The up-per lateral surfaces are nearly unmarked darkbrown, but more ventrally on the sides of thetrunk there are increasing densities of streaksand irregular spots of extensive, prominentwhite pigmentation that continues onto the ven-tral surface. The ground color of the ventralsurface is lighter than the lateral surfaces, andthe ventrolateral streaks of white are reduced toa more regular pattern of spotting or speckling.The tail and gular areas are lighter than theventer of the trunk, and white pigment is mostevident in the gular region. The upper limb in-sertions are typically white, especially those ofthe hind limb, but they may be somewhat speck-led with white and melanic pigment. The re-mainder of the limbs including the hands andfeet are speckled with light and dark coloration.The iris is black with an overlay of delicatestrands of silvery pigment.

Osteology.—Some information on osteology hasbeen derived from radiographs of the holotypeand six paratypes. All specimens have 18 trunkvertebrae, 17 of which bear ribs in the holotype(ribs are not visible in most of the other speci-mens). Numbers of caudal vertebrae range


Fig. 7. Cross-sectional representations of thesnouts of representative members of the two subgen-era of Oedipina. Left, Oedipina (Oedipina), based on aserial-sectioned head of O. poelzi. Right, Oedipina (Oed-opinola), based on an adult, cleared-and-stained O. al-leni, and on radiographs of several species. Maximumwidth of snouts is approximately 4 mm.

from 20 (in the smallest individual) to 48(LACM 145447); the holotype has 32. Skulls arerelatively robust and well articulated with littleor no fontanelle between the frontal and pari-etal bones. Frontal bones are large with a largefacial portion that articulates firmly with the na-sals, and the snout region is well ossified, withdevelopment of at least a partial tubular ar-rangement of the nasals, vomers and maxillariesas shown in Figure 7. The premaxillary is small,and the frontal processes are fused for most oftheir length as a narrow spine, although theyseparate distally. Digits are weakly developedand in most individuals are not countable. Theholotype has phalangeal formulas of 0-1-1-1 and0-1-2-1-1. LACM 145450 has the most completedigital series with formulas of 0-1-2-1 and 1-2-2-2-1. Tibial spurs are not evident.

Etymology.—The species is named in honor ofJay Mathers Savage, who has devoted more than40 years of sustained effort to document the bi-ology of the amphibians and reptiles of CostaRica, and whose comprehensive volume on thisfauna is forthcoming. Savage’s contributions in-clude not only his superb research productivityand his role in the development and leadershipof the Organization for Tropical Studies but

also his education of several generations of pro-fessional biologists, of whom the junior authoris one of the earliest in a long list.

Comment.—We have samples from three popu-lations in southwestern Costa Rica, two of whichwere studied by Good and Wake (1997) andfound to be very similar in allozyme patterns.However, we discovered that these two popula-tions differ considerably in sequences of cyt b(8.3–9.0%) and 16S (3.4%). Because allozymicdifferentiation is so slight (Nei D 5 0.02), be-cause we find no differences in morphology(except that population 31 apparently achieveslarger body size), and because the populationsoccur in the lowlands with no apparent physicalbarriers to gene flow, we continue to considerthem to be conspecific and to represent O. al-leni. Sequence differences between these twopopulations and O. savagei are also large (pop30-31 is 6.4–6.6% and pop 32 is 10.3–11.2% forcyt b; comparable values for 16S are 3.2 and4.4%). Phylogenetic analysis consistently clus-ters O. savagei and pop 30–31. However, thefacts that O. savagei occurs at much higher ele-vations than O. alleni and that it is morpholog-ically distinct lead us to recognize it as a distincttaxon. Allozymic data suggest that reticulationwith reference to DNA haplotypes has takenplace among the lowland populations. By this,we mean that populations once were separated,and during this time sorting of haplotype line-ages took place. Upon recontact, the popula-tions interbred so that allozymic genetic dis-tances are now low. However, the existence ofwell-differentiated haplotypes indicates thatthere has been insufficient time for reductionsin haplotype diversity. Such genetic reticulationhas also been found in other plethodontid sal-amanders, such as Ensatina (Wake, 1997; Wakeand Schneider, 1998) and Batrachoseps (Wakeand Jockusch, 2000).

DESCRIPTION OF A NEW SPECIES OF NOTOTRITON

FROM COSTA RICA

Nototriton gameziMonteverde Moss Salamander

Figure 8

Holotype.—MVZ 207122, an adult female fromCarril Bosque Eterno at junction with Pantan-osa Trail, Monteverde Cloud Forest Reserve,Prov. Alajuela, Costa Rica, elev. 1600 m, approx-imately 108199N, 84847.59W, collected on 14 Au-gust 1987, by D. C. Cannatella, D. A. Good, W.Guindon, and D. B. Wake.

56 COPEIA, 2000, NO. 1

Fig. 8. (A) Nototriton gamezi, MVZ 207123, an adult male (23.6 mm SL) from Penas Blancas Trail, Monte-verde, Prov. Alejuela, Costa Rica, elev. 1540 m, collected 15 August 1987. Photograph by D. C. Cannatella. (B)Nototriton gamezi, MVZ 207121, an adult female (24.2 mm SL) from Penas Blancas Trail, Monteverde, Prov.Alejuela, Costa Rica, elev. 1535 m, collected 14 August 1987. Photograph by D. C. Cannatella.

Paratypes.—MVZ 207120–207121, 207123, PenasBlancas trail below (E) Continental Divide,Monteverde Cloud Forest Reserve, Prov. Alajue-la, Costa Rica, elev. 1530–1540 m; MVZ 207124,Pantanosa Trail, Monteverde Cloud Forest Re-serve, Prov. Alajuela, Costa Rica, elev. 1590 m;UCR 4396, La Ventana, Monteverde Cloud For-est Reserve, Prov. Alajuela, Costa Rica; UCR6200, Penas Blancas Trial, Monteverde CloudForest Reserve, Prov. Alajuela, Costa Rica.

Diagnosis.—A small (maximum size in type se-ries 26.4 SVL), slender species of Nototriton dis-tinguished from Costa Rican members of thegenus as follows: from N. abscondens by its largernostrils, shorter limbs, and more prominent

parotoid glands; from N. richardi and N. tapantiby its larger, more fully formed hands and feet(e.g., fifth toe length of adults 0.7–0.8 vs. 0–0.1),longer limbs, and more robust habitus; from N.major by its smaller size, shorter tail and largernostrils; from N. guanacaste by its shorter andnarrower head, and larger nostrils; from N. pi-cadoi by its relatively more slender habitus andshorter limbs. It differs further from N. abscon-dens, N. richardi, N. guanacaste, N. picadoi, N. bar-bouri, N. lignicola, N. limnospectator, and N. brodieiby sequence differences in the mitochondrialgenes 16S rDNA and cytochrome b.

Description.—This is a small, slender species;adult SL for the sole adult male available is 23.6;


for two females 24.2, 26.2. The head is of mod-erate length (0.19–0.21 SL in four specimens)and width (0.13–0.14 SL), and is well demarcat-ed from the trunk. Eyes are of moderate sizeand protrude slightly beyond the lateral mar-gins of the head. The small teeth are relativelynumerous (24–35, x 28.8 total number of max-illary teeth; 12–17, x 14.8 total number of vo-merine teeth). Parotoid glands are relativelyconspicuous and appear as swollen, lightly pig-mented protuberances from the posterolateralmargins of the head. The adult male has asmall, flat, round mental gland. The species hasa trunk of moderate robustness. The limbs arerelatively short (0.16–0.17 SL). The slender tailis longer than body length (x 1.1 SL, maximumlength 1.3 SL) and has a relatively strong taperto a slender, pointed tip. Limbs are relativelyshort; limb interval 5–5.5. Hands and feet bearwell-formed, slender digits that are fully inde-pendent, except for the first digit of the fore-limb which is relatively indistinct and joined tothe basal portion of the manus. The digits areslightly expanded around the terminal phalanx.Webbing is slight and is limited to part of theproximal phalanx.

Measurements of holotype.—Head width 3.5; headdepth 2.3; eyelid length 1.8; eyelid width 0.8;anterior rim orbit to snout 0.8; interorbital dis-tance 0.8; snout to forelimb 0.9; nostril diame-ter 0.2; distance between external nares 0.7;projection of snout beyond mandible 0.1; par-otoid gland 1.0 3 0.8; snout to gular fold 5.3;snout to posterior angle of vent 26.2; snout toanterior angle of vent 24.6; axilla to groin 14.3;tail length 28.3; tail depth at base 2.4; tail widthat base 2.4; forelimb length 4.0; width of hand1.2; hind-limb length 4.8; width of foot 1.6;length of longest (third) toe 0.5; length of fifthtoe 0.2. Numbers of teeth: premaxillary 5; max-illary 17–18; vomerine 8–9.

Coloration of holotype (in alcohol).—This is a gen-erally dark animal with some brightly pigment-ed highlights. The dorsal surfaces of the headand body are dark gray-brown, with some dark-er brown pigment in an obscure dorsolateralline, especially in the costal interspaces. Thelight yellow parotoid glands stand out conspic-uously at the back of the head, in the neck re-gion. The lateral surfaces of the trunk are ob-scurely mottled with cream-yellow spots, blotch-es, and streaks. The facial portion of the headand the region under the eyes and along themandibles is lightly marked with small, yellowishspots. The venter, and especially the gular re-gion, are lighter than the dorsum and are light-

ly covered with small whitish spots. The dorsalportions of the proximal limb surfaces are yel-lowish. The dorsal surfaces of the tail are brightyellow-cream, marked with some irregular smallpatches of gray-brown. The ventral surfaces ofthe tail are gray.

Coloration (in life).—Color notes were recordedin life by DBW for three specimens. MVZ207123 was black ventrally, with white spots inbands corresponding to costal segments. Therewas a distinct contrast between the black venterand the brown dorsum. The lateral surfaces ofthe trunk were streaked and mottled, and therewas an obscure dorsolateral dark line. The par-otoid region was orange-brown, as were thelimb insertions. The dorsum of the tail was or-ange-brown with some tannish streaks. The iriswas brown-bronze.

MVZ 207121 was reddish brown with a tantail, with all dorsal surfaces bearing small blackspots. The venter of the trunk was gray withhighlights of reddish brown and a rich sprink-ling of white spots. The gular region was lightergray than other ventral surfaces. Limb inser-tions were reddish brown.

MVZ 207120 was light brown dorsally withmany black spots that create a pattern of re-peated chevrons, pointing anteriorly. Dorsolat-eral lines were dark with tan streaking besidethem. The tail was tan. The ventral and lateralsurfaces of the body were black, with many ir-regularly distributed white spots. There wereblack irregular markings on the head. Limb in-sertions were orange-brown.

Habitat.—The type series was collected in theMonteverde Cloud Forest Reserve in August1987. Two specimens were collected by search-ing through heavy moss mats in openings in theforest beside a wide trail on the Caribbeancoastal slope of the continental divide. The di-vide is about 1550 m elevation at this spot, andthe salamanders were obtained within about 20m (elevation) of the divide. Air temperature was21.5 C, and two salamanders were collected attemperatures of 21.5 and 20.2 within moss mats.These animals were microsympatric with Oedi-pina poelzi and O. uniformis. The type specimenwas collected along the divide in deep forest atabout 1600 m from moss growing on a tree;temperature 20.0 C. A fourth specimen was col-lected from moss on a tree.

Etymology.—This species is named in honor ofRodrigo Gamez, distinguished Costa Rican sci-entist and public servant, and first Director ofthe Instituto Nacional de Biodiversidad de Cos-

58 COPEIA, 2000, NO. 1

ta Rica (INBio), whose superb efforts have con-tributed greatly to knowledge and preservationof Costa Rican biodiversity.

Comment.—This species has been known formany years, but it has long been considered tobe conspecific with populations from the Cor-dillera Central of Costa Rica now assigned to N.abscondens (e.g., Van Devender, 1980). The spe-cies was studied by Good and Wake (1993), whoreported details of a morphometric and allo-zyme study of the population relative to otherCosta Rican populations assigned to Nototriton.Although they found the population to be mor-phologically distinct, they included it in N. abs-condens because allozyme differences were notgreat and because the population clustered withother populations of that species (but also withN. guanacaste) in a phylogenetic analysis of theallozyme data. However, these authors notedthat the Monteverde population overlaps moreextensively in morphological traits with N. guan-acaste than with N. abscondens. Although geneticdistances are low (Nei D 5 0.05–0.06), there isone fixed allozymic difference between N. ga-mezi and N. abscondens; there are three fixed dif-ferences between N. gamezi and N. guanacaste,and seven differences between N. gamezi and N.picadoi (Good and Wake, 1993). To the charac-ters listed by Good and Wake (1993) for theMonteverde population, we note also that thespecies has distinct parotoid glands, and in thistrait, it resembles N. guanacaste. The primaryreasons for our decision to describe the speciesare that it is diagnosable on several grounds, itis well differentiated from all other Costa Ricanpopulations of Nototriton with respect to both cytb and 16S sequences (relative to N. abscondensK2p is 0.021–0.024 for cyt b, 0.018 for 16S) andit is never the sister taxon of N. abscondens inphylogenetic analyses of the sequence data. In-stead, it is usually the sister taxon of a polytomyof N. abscondens, N. picadoi (K2p is 0.016–0.019for cyt b, 0.018 for 16S), and N. guanacaste (K2pis 0.040 for cyt b; 0.014 for 16s), but in the com-bined analysis, it is a part of a polytomy of thefirst two species. Accordingly, it is unlikely thatN. gamezi is even a sister group of N. abscondens,and we believe that it merits recognition as aseparate species.

DESCRIPTION OF A NEW GENUS OF TROPICAL

SALAMANDERS

Cryptotriton new genusHidden Salamanders

Type species.—Oedipus nasalis Schmidt.

Diagnosis.—Diminutive, slender, arboreal sala-manders belonging to the supergenus Bolitoglos-sa, having moderately long tails, limbs of mod-erate length, and enlarged nostrils. The inter-medium and ulnare of the manus and distal tar-sals four and five of the pes are fused. Thegenus is distinguished from other genera withsimilar fusions as follows: from Oedipina by hav-ing only 14 rather than 18 or more trunk ver-tebrae; from Bradytriton by having more slenderhabitus, a slender, nonglandular tail, and en-larged nostrils; from Parvimolge by having fron-tal processes of the premaxillary fused togetherat their base and lacking mesopodial and hyob-ranchial mineralizations; from Thorius by havinga complete skull roof over the brain case; fromBolitoglossa by having carpal fusions, a sublingualfold, and enlarged nostrils; from Nototriton byhaving much larger nostrils and by having a pre-frontal bone that is pierced for passage of thenasolacrimal duct.

Referred species.—Nototriton adelos Papenfuss andWake; Nototriton alvarezdeltoroi Papenfuss andWake; Nototriton monzoni Campbell and Smith;Nototriton nasalis (Schmidt); Nototriton veraepacis(Lynch and Wake); Nototriton wakei Campbelland Smith.

Etymology.—From kryptos, Greek, hidden, refer-ring to the cryptic behavior of these salaman-ders, the obscurity concerning phylogenetic re-lationships that has retarded its recognition,and its close morphological resemblance to oth-er clades of tropical salamanders, and triton, Lat-in, Greek, a commonly used term for salaman-ders.

Range.—Cryptotriton ranges from the mountainsof northern Oaxaca, Mexico, through themountains of northern Chiapas, Mexico, thehighlands of Alta Verapaz, Guatemala, and theSierra de las Minas, eastern Guatemala, to theCordillera del Merendon and associate uplandsin extreme northeastern Honduras.

Comment.—As recently as 1978, this clade andother clades with diminutive, slender species(Nototriton, Dendrotriton) were included in Chi-ropterotriton, which was thought to range fromTamaulipas, northeastern Mexico, to centralCosta Rica (Lynch and Wake, 1978). Wake andElias (1983) recognized Dendrotriton as a mono-phyletic group, and restricted Chiropterotriton toMexico north and west of the Isthmus of Te-huantepec. Although they described Nototritonat the same time, they expressed doubts that itwas a monophyletic group, and those reserva-


tions are confirmed by results published in thispaper. These small, secretive, often rare sala-manders have proven to be difficult taxonomi-cally, and only with the advent of molecular ap-proaches (e.g., Good and Wake, 1997; this pa-per) are we finally able to identify the mono-phyletic lineages that formerly comprised thecomposite Chiropterotriton. Our analysis supportsthe recognition of an additional species (pop.10) from Guatemala that is being described byGood and Wake.

The monophyletic units once included in Chi-ropterotriton display homoplasy in many structur-al details as well as in overall body form andeven general ecology. They have distributionsthat overlap only slightly with each other. Chi-ropterotriton occurs west and north of the Isth-mus of Tehuantepec, and although it is diversein morphology and ecology, several species aresmall. These are either terrestrial or bromeliad-dwellers (Wake, 1987; Darda, 1994). Cryptotritonincludes small species that use either moss matsor bromeliads and occur mainly in nuclear Mid-dle America, from western Honduras intonorthern Chiapas, in areas of Atlantic drainage.We tentatively assign the rare species C. adelosto this genus, although we have been unable tostudy its DNA, mainly because of its generalmorphology (it has an arrangement of skullbones found otherwise only in Cryptotriton, Pa-penfuss and Wake, 1987) and its geographicproximity to other species in the genus. How-ever, it is the only member to occur west of theIsthmus of Tehuantepec, and it may representa phylogenetically independent clade. Dendrotri-ton includes small bromeliad-dwellers and is lim-ited to Nuclear Middle America, where it isfound in the Atlantic drainage of western Hon-duras, mainly the Pacific drainage of northwest-ern Guatemala, and on the Pacific slopes ofChiapas (Wake, 1987, 1998). Finally, Nototritonoccurs in two well-separated upland zones in theAtlantic drainage of central and western Hon-duras and eastern Guatemala (where it is mainlyterrestrial) and in mountains along the conti-nental divide but mainly of Atlantic drainage incentral and northwestern Costa Rica (where itlives in moss mats on trees or road banks, or inbromeliads, or is terrestrial). Sympatry is knownbetween Cryptotriton and Chiropterotriton innorthern Oaxaca (Wake et al., 1992), betweenNototriton and Dendrotriton on the Montana deSanta Barbara, western Honduras (McCranieand Wilson, 1997a), and between Nototriton andCryptotriton in eastern Guatemala (Campbelland Smith, 1998; McCranie et al., 1998). Wherespecies of Nototriton occur with or near eitherDendrotriton or Cryptotriton, they can be distin-

guished by being more terrestrial and by havingsmall nostrils. Nototriton, Cryptotriton, Dendrotri-ton, and several smaller clades (Bradytriton, Nyc-tanolis, Ixalotriton) all are associated with nuclearMiddle America, and in future papers we willexamine the biogeography of these taxa ingreater detail in association with further phylo-genetic analyses.

REVIEW OF OTHER GENERA IN THIS STUDY

Nototriton Wake and EliasMoss Salamanders

Type species.—Spelerpes picadoi Stejneger.

Diagnosis.—Diminutive, slender, long-tailed sal-amanders of arboreal and semiarboreal (includ-ing moss mats), terrestrial, or semifossorial hab-itats with moderately long to short legs withsmall hands and feet and small to medium-sizednostrils. The intermedium and ulnare of themanus and distal tarsals four and five of the pesare fused. The genus is distinguished from oth-er genera with similar fusions as follows: fromOedipina by having only 14 rather than 18 ormore trunk vertebrae; from Bradytriton by hav-ing more slender habitus, a slender, nonglan-dular tail, and enlarged nostrils; from Parvimolgeby having frontal processes of the premaxillaryfused together at their base and lacking meso-podial and hyobranchial mineralizations; fromThorius by having a complete skull roof over thebrain case; from Bolitoglossa by having carpal fu-sions and a sublingual fold; from Cryptotriton byhaving much smaller nostrils and by having acomplete prefrontal bone or lacking it entirely,but never being pierced for passage of the na-solacrimal duct.

Referred species.—Nototriton abscondens (Taylor),Nototriton barbouri (Schmidt), Nototriton brodieiCampbell and Smith, Nototriton gamezi Garcıa-Parıs and Wake, Nototriton guanacaste Good andWake, Nototriton lignicola McCranie and Wilson,Nototriton limnospectator McCranie, Wilson andPolisar; Nototriton major Good and Wake, Nototri-ton picadoi (Stejneger), Nototriton richardi (Tay-lor), Nototriton tapanti Good and Wake.

Range.—From central Costa Rica northward inisolated montane habitats to western Honduras.There is a large geographic gap in the rangefrom Volcan Orosı in northwestern Costa Rica(N. guanacaste) to the mountains of north cen-tral Honduras (Depto. Olancho, N. lignicola).

60 COPEIA, 2000, NO. 1

Comment.—The species originally described asNototriton sanctibarbarus (McCranie and Wilson,1997a) has recently been reassigned to the ge-nus Dendrotriton (Wake, 1998).

Oedipina Keferstein 1868

Tropical Worm Salamanders

Type species.—Oedipina uniformis Keferstein.

Diagnosis.—Elongate salamanders of moderateto large size with long to very long tails, usuallywith small to very small limbs, hands and feet;distinguished from all other tropical bolitoglos-sine salamanders by having 18 or more trunkvertebrae, rather than 14.

Content.—We treat the two clades as subgenera,Oedipina and Oedopinola. Oedipina includes gen-erally longer-bodied, longer-tailed species, usu-ally with small limbs, hands and feet. Oedopinolaincludes generally stouter, shorter-bodied,shorter-tailed species, with larger limbs, handsand feet, but some species are relatively longbodied (O. gephyra) and others have very re-duced limbs and digits (O. complex, O. maritima).In general, Oedipina has a skull that is normallyproportioned for the tribe Bolitoglossini, where-as most (possibly all) species of Oedopinola dis-play a double tubular arrangement of the nasalregion, produced by downward growth of themedial and lateral borders of the nasals and cor-responding upward growth of the vomers, form-ing a skeletal tube around the nasal capsules ofthese semifossorial to fossorial species (Fig. 7).

Oedipina Keferstein 1868

Referred species.—The content of the uniformisgroup of Brame (1968), plus additional recentlyrecognized taxa: alfaroi (Dunn), altura Brame,collaris (Stejneger), cyclocauda Taylor, gracilis Tay-lor, grandis Brame and Duellman, ignea Stuart,pacificensis Taylor, paucidentata Brame, poelzi Bra-me, pseudouniformis Brame, stenopodia Brodieand Campbell, stuarti Brame, taylori Stuart, un-iformis Keferstein.

Range.—From extreme western Guatemalasouth to central Panama along the Pacific ver-sant, and from central Honduras to central Pan-ama along the Caribbean versant; from sea levelto about 2500 m elevation; occurs on islands ofthe Golfo de Fonseca on the Pacific Coast.

Comment.—Brame (1968) concluded that Hap-toglossa pressicauda Cope, 1893, is a synonym of

Oedipina uniformis Keferstein, 1868, which wasrevised by Good and Wake (1997). If he is cor-rect in this conclusion (the type is lost), thename pressicauda is likely a senior synonym forthe recently elevated taxon pacificensis Taylor,1952. Cope reported the type specimen to havean attached tongue. Cope was an expert on sal-amander morphology and accordingly under-stood the significance of this character (nomember of the supergenus Bolitoglossa has sucha tongue, and only members of that supergenusare otherwise known from the New World trop-ics). Accordingly, we think he was unlikely tomake a mistake on such a character. We are notconvinced that the type specimen is even refer-able to the supergenus, and accordingly we arereluctant to recommend a taxonomic change inthe status of pacificensis (which was raised fromthe synonymy of uniformis by Good and Wake,1997).

Oedopinola Hilton 1946

Referred species.—–The content of the parvipesgroup of Brame (1968) and additional recentlyrecognized taxa: alleni Taylor, carablanca Brame,complex (Dunn), elongata (Schmidt), gephyraMcCranie, Wilson and Williams, maritimaGarcıa-Parıs and Wake, parvipes (Peters), savageiGarcıa-Parıs and Wake.

Range.—From Chiapas, Mexico, along the Ca-ribbean versant to northwestern Colombia, in-cluding Isla Escudo de Veraguas, Panama, andfrom Honduras to northern Ecuador along thePacific versant, including Isla Gorgona, Colom-bia; from sea level to about 1500 m.

Comment.—This is a poorly known clade com-prised of species generally known only fromvery few specimens. Isolated samples from Pan-ama, Colombia and Ecuador have been as-signed to existing taxa but several probably con-stitute undescribed species.

ACKNOWLEDGMENTS

We gratefully acknowledge the assistance ofmany friends and colleagues in assembling thematerials used in this study. Fieldwork is de-manding and difficult, and many of the specieswe have studied are rare, elusive, or both. Weespecially thank F. Bolanos, D. C. Cannatella, D.A. Good, E. J. Jockusch, E. J. Koford, P. Leon,the late J. F. Lynch, T. J. Papenfuss, S. K. Ses-sions, and N. J. Staub for help in collecting. Wealso thank individuals who have obtained criti-cal specimens for us, including J. A. Campbell,


R. Crombie, J. Denslow, L. D. Gomez, R. Ibanez,J. McCranie, M. A. Schlaepfer, and L. D. Wilson.Collecting permits have been provided by agen-cies in the following countries: Costa Rica, Gua-temala, Honduras, Mexico, and Panama. Wealso thank managers of various field sites forpermission to collect tissue samples. We thankthe curatorial staffs of the Los Angeles Museumof Natural History (LACM) and University ofKansas (KU), and J. M. Savage, for loans of spec-imens and provision of information. GabrielaParra-Olea and S. Kuchta assisted in generatingthe sequence data, and Karen Klitz preparedFigures 6 and 8, and L. Chan assisted in thepreparation of Figures 5 and 7. We thank mem-bers of the Wake Group, especially M. Mahoney,S. Kuchta and G. Parra-Olea, for discussion andreview of the manuscript, and J. McCranie andreviewers for helpful comments. This researchwas supported by National Science Foundation,MVZ, and the Gompertz Professorship of theUniversity of California.

LITERATURE CITED

ANDERSON, S., A. T. BUNKIER, B. G. BARRELL, M. H. L.DEBRUIJN, A. R. COULSON, J DROUIN, I. C. EPERON,D. P. NIERLICH, B. A. ROE, F. SANGER, P. H.SCHREIER, A. J. H. SMITH, R. STADEN AND I. G.YOUNG. 1981. Sequence and organization of thehuman mitochondrial genome. Nature 290:457–465.

BRAME JR., A. H. 1968. Systematics and evolution ofthe Mesoamerican salamander genus Oedipina. J.Herpetol. 2:1–64.

CAMPBELL, J. C., AND E. N. SMITH. 1998. New speciesof Nototriton (Caudata: Plethodontidae) from east-ern Guatemala. Sci. Pap. Nat. Hist. Mus. Univ. Kans.6:1–8.

DARDA, D. M. 1994. Allozyme variation and morpho-logical evolution among Mexican salamanders ofthe genus Chiropterotriton (Caudata: Plethodonti-dae). Herpetologica 50:164–187.

FARRIS, J. S. 1989. The retention index and the re-scaled consistency index. Cladistics 5:417–419.

FELSENSTEIN, J. 1985. Evolutionary trees from DNAsequences: a maximum likelihood approach. J.Mol. Evol. 17:368–376.

FROST, D. R., AND M. HILLIS. 1990. Species in conceptand practice: herpetological applications. Herpe-tologia 46:87–104.

GOOD, D. A., AND D. B. WAKE. 1993. Systematic stud-ies of the Costa Rican moss salamanders, genus No-totriton, with descriptions of three new species. Her-petol. Monogr. 7:131–159.

, AND . 1997. Phylogenetic and taxo-nomic implications of protein variation in the Me-soamerican salamander genus Oedipina (Caudata:Plethodontidae). Rev. Biol. Trop. 45:1185–1208.

GUTTEL, R. R., AND G. E. FOX. 1988. A compilation

of large subunit rRNA sequences presented in astructural format. Nucleic Acid Res. 16s:r175–r269.

, M. W. GRAY, AND M. N. SCHARE. 1993. A com-pilation of large subunit (23S and 23S-like) ribo-somal RNA structures. Nucleic Acid Res. 21:3055–3074.

HASEGAWA, M., H. KISHINO, AND T. YANO. 1985. Dat-ing of the human-ape splitting by a molecular clockof mitochondrial DNA. J. Mol. Evol. 22:160–174.

HILLIS, D. M., AND M. T. DIXON. 1991. RibosomalDNA: molecular evolution and phylogenetic infer-ence. Q. Rev. Biol. 66:411–453.

HILTON, W. A. 1949. Salamanders from Barro Colo-rado Island, Canal Zone. J. Entomol. Zool. 38:37–39.

JACKMAN, T. R., G. APPLEBAUM, AND D. B. WAKE. 1997.Phylogenetic relationships of bolitoglossine sala-manders: a demonstration of the effects of combin-ing morphological and molecular data sets. Mol.Biol. Ecol. 14:883–891.

KEZER, J., S. K. SESSONS, AND P. LEON. 1989. The mei-otic structure and behavior of the strongly hetero-morphic X/Y sex chromosomes of neotropicalplethodontid salamanders of the genus Oedipina.Chromosoma (Berl.) 98:433–442.

KIMURA, M. 1980. A simple method for estimatingevolutionary rate of base substitutions throughcomparative studies of nucleotide sequences. J.Mol. Evol. 2:87–90.

KLUGE, A. G., AND J. S. FARRIS. 1969. Quantitative phy-letics and the evolution of anurans. Syst. Zool. 18:1–32.

KOCHER, T. D., W. K. THOMAS, A. MEYER, S. V. ED-WARDS, S. PAABO, F. X. VILLABLANCA, AND A. C. WIL-SON. 1989. Dynamics of mitochondrial DNA evo-lution in animals. Proc. Natl. Acad. Sci. USA 86:6196–6200.

LEVITON, A. E., R. H. GIBBS JR., E. HEAL, AND C. E.DAWSON. 1985. Standards in herpetology and ich-thyology. Part I. Standard symbolic codes for insti-tutional resource collections in herpetology andichthyology. Copeia 1985:802–821.

LYNCH, J. F., AND D. B. WAKE. 1978. A new species ofChiropterotriton (Amphibia: Caudata) from Baja Ver-apaz, Guatemala, with comments on relationshipsamong Central American members of the genus.Contrib. Sci. Nat. Hist. Mus. Los Angeles Co. 294:1–22.

MCCRANIE, J. R., AND L. D. WILSON. 1992. Life historynotes. Nototriton barbouri. Reproduction. Herpetol.Rev. 23:115–116.

, AND . 1997a. A new species of sala-mander of the genus Nototriton (Caudata: Pletho-dontidae) from Montana de Santa Barbara, Hon-duras. Southwest. Nat. 41:111–115.

, AND . 1997b. Two new species of sal-amanders (Caudata: Plethodontidae) of the generaBolitoglossa and Nototriton from Parque Nacional LaMuralla, Honduras. Proc. Biol. Soc. Wash. 110:366–372.

, , AND K. L. WILLIAMS. 1993. A new spe-cies of Oedipina (Amphibia: Caudata: Plethodonti-dae) from northern Honduras. Ibid. 106:385–389.

, , AND J. POLISAR. 1998. Another new

62 COPEIA, 2000, NO. 1

montane salamander (Amphibia: Caudata: Pletho-dontidae) from Parque Nacional de Santa Barbara.Honduras. Herpetologica 54:455–461.

MILLER, S. A., D. D. DYKES, AND H. F. POLESKY. 1988.A simple salting procedure for extracting DNAfrom human nucleated cells. Nucleic Acid Res. 16:215.

MINDELL, D. P., AND R. L. HONEYCUTT. 1990. Ribo-somal RNA in vertebrates: evolution and phyloge-netic applications. Annu. Rev. Ecol. Syst. 21:541–566.

MORITZ, C., C. J. SCHNEIDER, AND D. B. WAKE. 1992.Evolutionary relationships within the Ensatina esch-scholtzii complex confirm the ring species interpre-tation. Syst. Biol. 41:273–291.

ORTI, G., AND A. MEYER. 1997. The radiation of char-aciform fishes and the limits of resolution of mi-tochondrial ribosomal DNA sequences. Ibid. 46:75–100.

, P. PETRI, J. I. R. PORTO, M. JE’GU, AND A. MEY-ER. 1996. Patterns of nucleotide change in mito-chondrial ribosomal RNA genes and the phylogenyof piranhas. J. Mol. Evol. 42:169–182.

PALUMBI, S. R., A. P. MARTIN, S. ROMANO, W. O.MCMILLAN, L. STICE, AND G. GRABOWSKI. 1991. Thesimple fool’s guide to PCR. Spec. Publ., Dept. Zo-ology, Univ. of Hawaii, Honolulu.

PAPENFUSS, T. J., AND D. B. WAKE. 1987. Two new spe-cies of plethodontid salamanders (genus Nototriton)from Mexico. Acta Zool. Mex. (n.s.) 21:1–16.

ROE, B. A., D. P. MA, R. K. WILSON, AND J. F. WONG.1985. The complete nucleotide sequence of theXenopus laevis mitochondrial DNA genome. J. Biol.Chem. 260:9759–9774.

SAITOU, N., AND M. NEI. 1987. The neighbor-joiningmethod: a new method for reconstructing phylo-genetic trees. Mol. Biol. Evol. 4:406–525.

SAKAI, R. K., D. H. DELFAND, S. STOOFFEL, S. J. SCHARF,R. HIGUCHI, G. T. HORN, K. B. MULLIS, AND H. A.ERLICH. 1988. Primer-directed enzymatic amplifi-cation of DNA with a thermostable DNA polymer-ase. Science 239:487–491.

SAVAGE, J. M., AND J. VILLA R. 1986. Herpetofauna ofCosta Rica. SSAR Contrib. Herpetol. 3:i–viii, 1–207.

SESSIONS, S. K., AND J. KEZER. 1991. Evolutionary cy-togenetics of bolitoglossine salamanders (familyPlethodontidae), p. 89–130. In: Amphibian cytoge-

netics and evolution. D. M. Green and S. K. Ses-sions (eds.). Academic Press, San Diego, CA.

TAN, A.-M., AND D. B. WAKE. 1995. MtDNA phylo-geography of the California newt, Taricha torosa(Caudata, Salamandridae). Mol. Phylo. Evol. 4:383–394.

VAN DEVENDER, R. W. 1980. Preliminary checklist ofthe herpetofauna of Monteverde, Puntarenas Prov-ince, Costa Rica and vicinity. Brenesia 17:319–326.

WAKE, D. B. 1966. Comparative osteology and evo-lution of the lungless salamanders, family Pletho-dontidae. Mem. So. Calif. Acad. Sci. 4:1–111.

. 1987. Adaptive radiation of salamanders inMiddle American cloud forests. Ann. Mo. Bot.Gard. 74:242–264.

. 1997. Incipient species formation in sala-manders of the Ensatina complex. Proc. Natl. Acad.Sci. USA 94:7761–7767.

. 1998. On the taxonomic status of Nototritonsanctibarbarus McCranie and Wilson (Amphibia:Caudata). Southwest. Nat. 43:88–91.

, AND P. ELIAS. 1983. New genera and a newspecies of Central American salamanders, with a re-view of the tropical genera (Amphibia, Caudata,Plethodontidae). Contrib. Sci. Nat. Hist. Mus. LosAngeles Co. 345:1–19.

, AND E. L. JOCKUSCH. 2000. Detecting speciesborders using diverse data sets: examples fromplethodontid salamanders in California. In: The bi-ology of Plethodontid salamanders. R. C. Bruce, R.G. Jaeger, and L. D. Houck (eds.). Plenum Press,New York.

, AND C. J. SCHNEIDER. 1998. Taxonomy of theplethodontid salamander genus Ensatina. Herpeto-logica 54:279–298.

, T. J. PAPENFUSS, AND J. F. LYNCH. 1992. Dis-tribution of salamanders along elevational transectsin Mexico and Guatemala. Tulane Stud. Zool. Bot.,Suppl. Publ. 1:303–319.

MUSEUM OF VERTEBRATE ZOOLOGY AND DEPART-MENT OF INTEGRATIVE BIOLOGY, UNIVERSITY OF

CALIFORNIA, BERKELEY, CALIFORNIA 94720-3160.E-mail: (DBW) [email protected] reprint requests to DBW. Submitted: 7Dec. 1998. Accepted: 13 July 1999. Section ed-itor: A. H. Price.


APP

EN

DIX

1.SA

MPL

ES

USE

DIN

TH

ISST

UD

YA

ND

GE

NB

AN

KA

CC

ESS

ION

NU

MB

ER

S.

Sam

ple

no.

Spec

ies

Loca

lity

Mus

eum

no.

Cyt

b16

S

1 2 3 4 5 6 7 8 9

C.

alva

rezd

eltor

oiC

.na

salis

C.

vera

epac

isC

.ve

raep

acis

C.

vera

epac

isC

.ve

raep

acis

C.

vera

epac

isC

.ve

raep

acis

C.

sp.A

Mex

ico:

Chi

apas

:21.

5m

iN

Jito

tal

Hon

dura

s:C

orte

s:M

ts.W

San

Pedr

oSu

laG

uate

mal

a:B

aja

Vera

paz:

2.5

mi

SPu

rulh

aG

uate

mal

a:B

aja

Vera

paz:

6.5

mi

ESE

Puru

lha

Gua

tem

ala:

Baj

aVe

rapa

z:6.

5m

iE

SEPu

rulh

aG

uate

mal

a:B

aja

Vera

paz:

Puru

lha

Gua

tem

ala:

Baj

aVe

rapa

z:Pu

rulh

aG

uate

mal

a:B

aja

Vera

paz:

Puru

lha

Gua

tem

ala:

Zaca

pa:S

ierr

ade

Las

Min

as

MV

Z15

8942

MV

Z12

8274

MV

Z17

2709

MV

Z21

5913

MV

Z16

7991

UT

AG

AR

59(T

iss.

).U

TA

A-5

1395

UT

AA

-513

96M

VZ

1609

01

AF1

9912

0A

F199

121

AF1

9912

2A

F199

123

AF1

9912

4A

F199

125

AF1

9912

6A

F199

127

AF1

9912

8

AF1

9919

6— —

AF1

9919

7— — — — —

10 11 12 13 14 15 16 17 18 19

C.

sp.A

N.

absc

onde

nsN

.ab

scon

dens

N.

absc

onde

nsN

.ab

scon

dens

N.

absc

onde

nsN

.ga

mez

iN

.ba

rbou

riN

.ba

rbou

riN

.ba

rbou

ri

Gua

tem

ala:

Zaca

pa:S

ierr

ade

Las

Min

asC

osta

Ric

a:A

laju

ela:

Cas

cada

deL

aPa

zC

osta

Ric

a:A

laju

ela:

Vara

Bla

nca

Cos

taR

ica:

Ala

juel

a:Va

raB

lanc

aC

osta

Ric

a:Sa

nJo

se:C

asca

jald

eL

asN

ubes

Cos

taR

ica:

San

Jose

:Cas

caja

lde

Las

Nub

esC

osta

Ric

a:A

laju

ela:

Mon

teve

rde

Hon

dura

s:A

tlant

ida:

Que

brad

ade

lO

roH

ondu

ras:

Atla

ntid

a:C

erro

Buf

alo

Hon

dura

s:Yo

ro:2

.5km

NN

EL

aFo

rtun

a

MV

Z16

0907

UC

R12

071

MV

Z20

3743

MV

Z18

1351

MV

Z19

4884

MV

Z19

4867

MV

Z20

7122

USN

M33

9712

USN

M49

7552

USN

M50

9333

AF1

9912

9A

F199

130

AF1

9913

1A

F199

132

AF1

9913

3A

F199

134

AF1

9913

5A

F199

136

AF1

9913

7A

F199

138

AF1

9919

8A

F199

199

— — — —A

F199

200

AF1

9920

1— —

20 21 22 23 24 25 26 27 28 29

N.

brod

iei

N.

guan

acas

teN

.lig

nico

laN

.lig

nico

laN

.lim

nosp

ecta

tor

N.

pica

doi

N.

pica

doi

N.

rich

ardi

N.

rich

ardi

N.

rich

ardi

Gua

tem

ala:

Izab

al:S

ierr

ade

Car

al,M

oral

esC

osta

Ric

a:G

uana

cast

e:Vo

lcan

Cac

aoH

ondu

ras:

Ola

ncho

:Cer

rode

Enm

edio

Hon

dura

s:O

lanc

ho:C

erro

deE

nmed

ioH

ondu

ras:

Sant

aB

arba

ra:E

lO

cotil

loC

osta

Ric

a:C

arta

go:T

apan

tıC

osta

Ric

a:C

arta

go:T

apan

tıC

osta

Ric

a:Sa

nJo

se:C

asca

jald

eL

asN

ubes

Cos

taR

ica:

San

Jose

:Cas

caja

lde

Las

Nub

esC

osta

Ric

a:Sa

nJo

se:C

asca

jald

eL

asN

ubes

UT

AA

-514

90M

VZ

2071

06U

SNM

4975

40U

SNM

4975

50M

VZ

2258

66M

VZ

2258

99M

VZ

2037

45U

CR

1205

7M

VZ

1948

85M

VZ

1948

87

AF1

9913

9A

F199

140

AF1

9914

1A

F199

142

AF1

9914

3A

F199

144

AF1

9914

5A

F199

146

AF1

9914

7A

F199

148

AF1

9920

2A

F199

203

AF1

9920

4— —

AF1

9920

5—

AF1

9920

6— —

30 31 32 33 34 35

O.

allen

iO

.al

leni

O.

allen

iO

.sa

vage

iO

.sa

vage

iO

.pa

rvip

es

Cos

taR

ica:

Punt

aren

as:S

iren

aC

osta

Ric

a:Pu

ntar

enas

:Sir

ena

Cos

taR

ica:

Punt

aren

as:D

amas

Cos

taR

ica:

Punt

aren

as:C

erro

Zapo

teC

osta

Ric

a:Pu

ntar

enas

:Cer

roZa

pote

Pana

ma:

San

Bla

s:N

usag

andi

MV

Z19

0857

MV

Z19

0856

MV

Z22

5903

UC

RL

DG

9613

27M

VZ

DB

W57

85M

VZ

2104

04

AF1

9914

9A

F199

150

AF1

9915

1A

F199

152

AF1

9915

3A

F199

154

AF1

9920

7—

AF1

9920

8A

F199

209

—A

F199

210

64 COPEIA, 2000, NO. 1

APP

EN

DIX

1.C

ON

TIN

UE

D.

Sam

ple

no.

Spec

ies

Loca

lity

Mus

eum

no.

Cyt

b16

S

36 37 38 39 40 41 42 43 44 44b

45 46 47 48

O.

parv

ipes

O.

com

plex

O.

com

plex

O.

cycl

ocau

daO

.cy

cloc

auda

O.

elong

ata

O.

geph

yra

O.

geph

yra

O.

grac

ilis

O.

grac

ilis

O.

gran

dis

O.

gran

dis

O.

mar

itim

aO

.pa

cific

ensi

s

Pana

ma:

Col

on:R

ıoFr

ijole

sPa

nam

a:Pa

nam

a:A

ltos

deC

erro

Cam

pana

Pana

ma:

Col

on:B

arro

Col

orad

oC

osta

Ric

a:H

ered

ia:L

aSe

lva

Cos

taR

ica:

Her

edia

:La

Selv

aG

uate

mal

a:Iz

abal

:Sie

teA

ltare

s,L

ivin

gsto

nH

ondu

ras:

Atla

ntid

a:C

erro

Buf

alo

Hon

dura

s:Yo

ro:2

.5km

NN

EL

aFo

rtun

aC

osta

Ric

a:H

ered

ia:L

aSe

lva

Cos

taR

ica:

Her

edia

:La

Selv

aC

osta

Ric

a:Pu

ntar

enas

:Cer

roPa

ndo

Cos

taR

ica:

Punt

aren

as:L

asTa

blas

Pana

ma:

Boc

asde

lTo

ro:E

scud

ode

Vera

guas

Cos

taR

ica:

Punt

aren

as:S

iren

a

MV

Z21

0405

MV

ZD

BW

5787

MV

ZD

BW

5105

MV

Z13

8916

MV

Z20

3747

UT

AA

-519

06U

SNM

3434

62U

SNM

LD

W10

502

MV

Z20

3753

MV

Z21

0398

MV

Z22

5904

MV

Z21

9593

MV

Z21

9997

MV

Z19

0859

AF1

9915

5A

F199

156

AF1

9915

7A

F199

158

AF1

9915

9A

F199

160

AF1

9916

1A

F199

162

AF1

9916

3—

AF1

9916

4A

F199

165

AF1

9916

6A

F199

167

AF1

9921

1A

F199

212

AF1

9921

3A

F199

214

AF1

9921

5A

F199

216

AF1

9921

7A

F199

218

—A

F199

219

AF1

9922

0—

AF1

9922

1—

49 50 51 52 53 54 55 56 57 58 59 60

O.

paci

ficen

sis

O.

paci

ficen

sis

O.

paci

ficen

sis

O.

poelz

iO

.po

elzi

O.

poelz

iO

.po

elzi

O.

poelz

iO

.po

elzi

O.

pseu

doun

iform

isO

.ps

eudo

unifo

rmis

O.

pseu

doun

iform

is

Cos

taR

ica:

Punt

aren

as:S

iren

aC

osta

Ric

a:Pu

ntar

enas

:Dam

asC

osta

Ric

a:Pu

ntar

enas

:Las

Cru

ces

Cos

taR

ica:

Her

edia

:Bra

ulio

Car

rillo

Cos

taR

ica:

San

Jose

:Cas

caja

lde

Las

Nub

esC

osta

Ric

a:A

laju

ela:

Vara

Bla

nca

Cos

taR

ica:

Ala

juel

a:Va

raB

lanc

aC

osta

Ric

a:A

laju

ela:

Mon

teve

rde

Cos

taR

ica:

Car

tago

:Mor

avia

deC

hirr

ipo

Cos

taR

ica:

Car

tago

:3km

EN

EJu

anV

inas

Cos

taR

ica:

Car

tago

:Los

Esp

avel

esC

osta

Ric

a:C

arta

go:M

orav

iade

Chi

rrip

o

MV

Z19

0858

UC

R12

063

UC

RE

7M

VZ

2063

98M

VZ

1812

35M

VZ

1813

48M

VZ

1637

03M

VZ

2071

28M

VZ

1948

73M

VZ

1908

52M

VZ

2037

49M

VZ

1812

29

AF1

9916

8A

F199

169

AF1

9917

0A

F199

171

AF1

9917

2A

F199

173

AF1

9917

4A

F199

175

AF1

9917

6A

F199

177

AF1

9917

8A

F199

179

—A

F199

222

—A

F199

223

— —A

F199

224

AF1

9922

5A

F199

226

—A

F199

227

—61 62 63 64

O.

sten

opod

iaO

.st

enop

odia

O.

unifo

rmis

O.

unifo

rmis

Gua

tem

ala:

S.M

arco

s:S.

Raf

ael

Gua

tem

ala:

S.M

arco

s:S.

Raf

ael

Cos

taR

ica:

San

Jose

:Cer

ros

deE

scaz

uC

osta

Ric

a:Sa

nJo

se:A

ltode

La

Palm

a

MV

Z13

8918

MV

Z16

3649

MV

Z22

1340

MV

Z22

5905

AF1

9918

0A

F199

181

AF1

9918

2A

F199

183

—A

F199

228

— —


APP

EN

DIX

1.C

ON

TIN

UE

D.

Sam

ple

no.

Spec

ies

Loca

lity

Mus

eum

no.

Cyt

b16

S

65 66 67 68 69

O.

unifo

rmis

O.

unifo

rmis

O.

unifo

rmis

O.

unifo

rmis

O.

unifo

rmis

Cos

taR

ica:

Car

tago

:Rıo

Tur

rial

baC

osta

Ric

a:C

arta

go:C

iene

gade

Col

orad

oC

osta

Ric

a:C

arta

go:V

olca

nT

urri

alba

Cos

taR

ica:

Car

tago

:Vol

can

Tur

rial

baC

osta

Ric

a:Sa

nJo

se:C

asca

jald

eL

asN

ubes

MV

Z19

4871

MV

Z19

0853

MV

Z19

4862

MV

Z19

4864

MV

Z22

5906

AF1

9918

4A

F199

185

AF1

9918

6A

F199

187

AF1

9918

8

—A

F199

229

— — —70 71 72 73 74 75 76 77

O.

unifo

rmis

O.

unifo

rmis

O.

unifo

rmis

O.

sp.B

O.

sp.C

Den

drot

rito

nra

bbi

Bolit

oglo

ssa

cerr

oens

isBa

trac

hose

psga

brie

li

Cos

taR

ica:

San

Jose

:Cas

caja

lde

Las

Nub

esC

osta

Ric

a:C

arta

go:T

apan

tıC

osta

Ric

a:A

laju

ela:

Vara

Bla

nca

Hon

dura

s:O

cote

pequ

e:G

uari

n,C

erro

El

Pita

lH

ondu

ras:

Yoro

:32

kmW

Yoro

Gua

tem

ala:

Qui

che:

Usp

anta

nC

osta

Ric

a:Sa

nJo

se:S

alsi

pued

esU

SA:C

alifo

rnia

:San

Gab

riel

Mts

.

MV

Z22

1321

MV

Z20

3751

MV

Z18

1349

USN

ML

DW

1127

0M

VZ

1677

72U

TA

A-5

1086

MV

ZS

1292

1M

VZ

2229

57

AF1

9918

9A

F199

190

AF1

9919

1A

F199

192

AF1

9919

3A

F199

194

AF1

9919

5

—A

F199

230

—A

F199

231

—A

F199

232

AF1

9923

3A

F199

234

APP

EN

DIX

2.SE

QU

EN

CE

DIV

ER

GE

NC

E(K

2p)

BE

TW

EE

NT

AX

A.D

ista

nces

with

inta

xaco

rres

pond

tom

inim

alle

vels

ofdi

verg

ence

amon

gsp

ecie

s.U

pper

half

mat

rix

isba

sed

on16

Sse

quen

ces.

Low

erha

lfm

atri

xis

base

don

cyt

bse

quen

ces.

12

34

56

78

1 2 3 4 5 6 7 8

Cry

ptot

rito

nN

otot

rito

nO

edip

ina

s.l.

Oed

ipin

as.

str.

Oed

opin

ola

Den

drot

rito

nBo

litog

loss

aBa

trac

hose

ps

***

0.24

93–0

.306

60.

2253

–0.3

255

0.22

53–0

.325

50.

2257

–0.3

104

0.24

64–0

.269

30.

2704

–0.3

040

0.23

61–0

.289

3

0.15

56–0

.179

2**

*0.

1726

–0.2

558

0.17

62–0

.255

80.

1726

–0.2

554

0.24

29–0

.304

60.

2455

–0.2

929

0.21

61–0

.287

4

0.15

91–0

.195

80.

1129

–0.1

701

***

— —0.

2308

–0.3

191

0.21

40–0

.318

00.

2416

–0.2

944

0.15

91–0

.195

80.

1129

–0.1

600

— ***

0.13

53–0

.224

70.

2445

–0.3

191

0.21

40–0

.285

50.

2416

–0.2

944

0.16

52–0

.192

40.

1148

–0.1

701

—0.

0641

–0.1

082

***

0.23

08–0

.293

40.

2567

–0.3

180

0.24

51–0

.290

1

0.17

51–0

.178

40.

1509

–0.1

639

0.17

52–0

.208

30.

1752

–0.2

027

0.17

75–0

.208

3**

*0.

2751

0.28

64

0.22

36–0

.240

10.

1838

–0.2

043

0.18

92–0

.228

10.

1948

–0.2

259

0.18

92–0

.228

10.

2075

***

0.24

42

0.21

69–0

.227

50.

1974

–0.2

188

0.22

04–0

.248

90.

2220

–0.2

489

0.22

04–0

.242

70.

2193

0.19

28**

*

66 COPEIA, 2000, NO. 1

APPENDIX 3. SEQUENCE DIVERGENCE (K2p) WITHIN Cryptotriton. Upper half matrix is based on 16S sequences.Lower half matrix is based on cyt b sequences.

1 2 3 4 5

12345

C. alvarezdeltoroiC. nasalisC. veraepacisC. veraepacisC. sp. A

(1)(2)(3, 6)(4, 5, 7, 8)(9, 10)

***0.16740.15340.14660.1469

—***

0.14870.14180.1323

——***

0.00520.0932

0.0951——***

0.0870

0.0892——

0.0448***


APP

EN

DIX

4.SE

QU

EN

CE

DIV

ER

GE

NC

E(K

2p)

WIT

HIN

Not

otri

ton.

Upp

erha

lfm

atri

xis

base

don

16S

sequ

ence

s.L

ower

half

mat

rix

isba

sed

oncy

tb

sequ

ence

s.

12

34

56

78

910

1112

13

1 2 3 4 5 6 7 8 9 10 11 12 13

N.

absc

onde

nsN

.ab

scon

dens

N.

gam

ezi

N.

barb

ouri

N.

barb

ouri

N.

brod

iei

N.

guan

acac

aste

N.

ligni

cola

N.

ligni

cola

N.

limno

spec

tato

rN

.pi

cado

iN

.pi

cado

iN

.ri

char

di

(11,

12,1

3)(1

4,15

)(1

6)(1

7,18

)(1

9)(2

0)(2

1)(2

2)(2

3)(2

4)(2

5)(2

6)(2

7,28

,29)

***

0.00

790.

0239

0.12

440.

1475

0.13

100.

0432

0.12

440.

1208

0.14

340.

0294

0.03

210.

0959

— ***

0.02

120.

1211

0.14

410.

1277

0.04

040.

1244

0.12

080.

1401

0.02

670.

0293

0.09

28

0.01

78— **

*0.

1343

0.16

490.

1343

0.04

040.

1310

0.12

740.

1468

0.01

580.

0185

0.09

28

0.05

71—

0.05

91**

*0.

0573

0.06

040.

1377

0.11

850.

1149

0.13

400.

1343

0.13

730.

1214

— — — — ***

0.06

900.

1685

0.12

800.

1250

0.13

040.

1579

0.16

100.

1277

0.05

69—

0.06

330.

0137

— ***

0.12

440.

1061

0.10

270.

1214

0.12

110.

1241

0.09

28

0.01

38—

0.01

980.

0591

—0.

0633

***

0.12

770.

1307

0.14

340.

0348

0.03

750.

0959

0.05

89—

0.06

530.

0422

—0.

0421

0.06

53**

*0.

0158

0.11

170.

1244

0.12

740.

1448

— — — — — — — — ***

0.11

460.

1208

0.12

380.

1479

— — — — — — — — — ***

0.14

680.

1499

0.13

70

0.01

78—

0.01

580.

0635

—0.

0677

0.01

180.

0653

— — ***

0.00

260.

0928

— — — — — — — — — — — ***

0.09

57

0.03

83—

0.03

620.

0508

—0.

0550

0.03

610.

0484

— —0.

0445

— ***

68 COPEIA, 2000, NO. 1

APP

EN

DIX

5.SE

QU

EN

CE

DIV

ER

GE

NC

E(K

2p)

WIT

HIN

Oed

ipin

a(O

edip

ina)

.Upp

erha

lfm

atri

xis

base

don

16S

sequ

ence

s.L

ower

half

mat

rix

isba

sed

oncy

tb

sequ

ence

s.

12

34

56

78

910

1112

1314

1 2 3 4 5 6 7 8 9

O.

cycl

ocau

daO

.cy

cloc

auda

O.

grac

ilis

O.

gran

dis

O.

paci

ficen

sis

O.

paci

ficen

sis

O.

paci

ficen

sis

O.

paci

ficen

sis

O.

poelz

i

(39)

(40)

(44)

(45,

46)

(48)

(49)

(50)

(51)

(52)

***

0.00

260.

1626

0.14

560.

1323

0.13

260.

1394

0.13

560.

1265

0.00

58**

*0.

1592

0.14

890.

1356

0.13

590.

1361

0.13

890.

1298

0.06

550.

0657

***

0.18

350.

1126

0.10

640.

1088

0.10

290.

1599

0.06

780.

0701

0.06

35**

*0.

1735

0.16

960.

1871

0.16

230.

1049

— — — — ***

0.00

520.

0402

0.02

930.

1301

— — — — — ***

0.03

470.

0239

0.12

98

0.05

320.

0512

0.04

270.

0447

— — ***

0.06

010.

1571

— — — — — — — ***

0.12

95

0.06

350.

0616

0.07

920.

0507

— —0.

0556

— ***

— — — — — — — — —

0.06

560.

0670

0.07

440.

0485

— —0.

0574

—0.

0258

0.06

340.

0678

0.07

230.

0464

— —0.

0553

—0.

0237

0.05

910.

0635

0.06

570.

0423

— —0.

0488

—0.

0402

0.03

770.

0316

0.06

780.

0674

— —0.

0553

—0.

0741

10 11 12 13 14 15 16 17 18 19

O.

poelz

iO

.po

elzi

O.

poelz

iO

.po

elzi

O.

pseu

doun

iform

isO

.ps

eudo

unifo

rmis

O.

sten

opod

iaO

.st

enop

odia

O.

unifo

rmis

O.

unifo

rmis

(53)

(54,

55)

(56)

(57)

(58,

59)

(60)

(61)

(62)

(63)

(64)

0.12

680.

1201

0.10

780.

1044

0.05

160.

0545

0.14

890.

1459

0.13

280.

1361

0.13

010.

1233

0.10

460.

1076

0.05

450.

0573

0.14

560.

1425

0.13

610.

1394

0.15

670.

1592

0.14

560.

1554

0.15

880.

1623

0.14

710.

1441

0.09

660.

0998

0.11

140.

1139

0.11

390.

1105

0.14

590.

1492

0.14

650.

1468

0.17

350.

1735

0.12

710.

1328

0.12

630.

1326

0.13

260.

1359

0.13

700.

1340

0.04

320.

0404

0.13

340.

1326

0.12

930.

1323

0.13

890.

1422

0.14

340.

1404

0.03

770.

0349

0.16

100.

1529

0.14

950.

1465

0.14

590.

1492

0.14

920.

1462

0.06

300.

0601

0.13

310.

1323

0.12

910.

1326

0.15

540.

1588

0.14

650.

1434

0.05

170.

0488

0.00

780.

0486

0.05

700.

0863

0.13

670.

1401

0.13

340.

1304

0.13

670.

1401

***

0.05

450.

0571

0.09

260.

1370

0.14

040.

1304

0.12

740.

1404

0.14

37

— ***

0.02

930.

0803

0.12

680.

1301

0.12

310.

1201

0.13

940.

1428

—0.

0019

***

0.07

170.

1274

0.13

070.

1103

0.10

740.

1361

0.13

94

—0.

0402

0.03

81**

*0.

1109

0.11

410.

1260

0.12

310.

1326

0.13

59

—0.

0783

0.07

610.

0717

***

0.00

260.

1456

0.14

280.

1359

0.13

9220 21 22 23 24 25 26 27

O.

unifo

rmis

O.

unifo

rmis

O.

unifo

rmis

O.

unifo

rmis

O.

unifo

rmis

O.

unifo

rmis

O.

sp.B

O.

sp.C

(65,

66)

(67,

68)

(69)

(70)

(71)

(72)

(73)

(74)

0.14

590.

1459

0.13

260.

1392

0.14

220.

1359

0.16

610.

1620

0.14

920.

1492

0.13

590.

1425

0.14

560.

1392

0.16

960.

1654

0.09

690.

1032

0.09

070.

0969

0.09

980.

1001

0.16

720.

1476

0.16

960.

1696

0.15

920.

1557

0.15

880.

1592

0.11

640.

1386

0.03

210.

0377

0.02

660.

0266

0.04

600.

0349

0.14

310.

1572

0.02

670.

0322

0.02

120.

0212

0.04

050.

0294

0.14

950.

1574

0.04

590.

0573

0.04

030.

0403

0.06

010.

0431

0.16

230.

1611

0.04

050.

0461

0.03

490.

0349

0.04

320.

0372

0.14

590.

1674

0.13

640.

1331

0.12

980.

1364

0.13

280.

1465

0.13

940.

1315

0.14

010.

1401

0.13

340.

1401

0.13

640.

1502

0.13

640.

1252

0.14

920.

1492

0.13

920.

1459

0.14

560.

1492

0.12

630.

1315

0.13

920.

1392

0.12

930.

1359

0.13

560.

1459

0.12

950.

1190

0.13

230.

1323

0.12

580.

1258

0.12

880.

1356

0.13

920.

1284

0.15

230.

1523

0.13

890.

1456

0.14

860.

1422

0.15

950.

1548


APP

EN

DIX

5.E

XT

EN

DE

D

1516

1718

1920

2122

2324

2526

27

1 2 3 4 5 6 7 8 9

O.

cycl

ocau

daO

.cy

cloc

auda

O.

grac

ilis

O.

gran

dis

O.

paci

ficen

sis

O.

paci

ficen

sis

O.

paci

ficen

sis

O.

paci

ficen

sis

O.

poelz

i

(39)

(40)

(44)

(45,

46)

(48)

(49)

(50)

(51)

(52)

— — — — — — — — —

— — — — — — — — —

0.05

690.

0570

0.06

550.

0570

— —0.

0490

—0.

0615

— — — — — — — — —

— — — — — — — — —

0.04

620.

0442

0.04

230.

0485

— —0.

0119

—0.

0570

— — — — — — — — —

— — — — — — — — —

— — — — — — — — —

0.04

830.

0463

0.04

440.

0465

— —0.

0119

—0.

0550

— — — — — — — — —

0.05

690.

0571

0.06

130.

0529

— —0.

0447

—0.

0747

— — — — — — — — —10 11 12 13 14 15 16 17 18 19

O.

poelz

iO

.po

elzi

O.

poelz

iO

.po

elzi

O.

pseu

doun

iform

isO

.ps

eudo

unifo

rmis

O.

sten

opod

iaO

.st

enop

odia

O.

unifo

rmis

O.

unifo

rmis

(53)

(54,

55)

(56)

(57)

(58,

59)

(60)

(61)

(62)

(63)

(64)

— — — — — ***

0.14

920.

1462

0.13

920.

1425

— — — — — — ***

0.00

260.

1337

0.13

04

—0.

0612

0.05

910.

0613

0.06

96— — **

*0.

1307

0.12

74

— — — — — — — — ***

0.00

26

— — — — — — — — — ***

—0.

0589

0.05

680.

0526

0.04

82— —

0.05

05— —

— — — — — — — — — —

— — — — — — — — — —

— — — — — — — — — —

—0.

0569

0.05

470.

0505

0.05

04— —

0.05

26— —

— — — — — — — — — —

—0.

0744

0.07

220.

0701

0.06

55— —

0.02

39— —

— — — — — — — — — —20 21 22 23 24 25 26 27

O.

unifo

rmis

O.

unifo

rmis

O.

unifo

rmis

O.

unifo

rmis

O.

unifo

rmis

O.

unifo

rmis

O.

sp.B

O.s

p.C

(65,

66)

(67,

68)

(69)

(70)

(71)

(72)

(73)

(74)

0.15

570.

1557

0.14

220.

1489

0.15

200.

1456

0.16

300.

1515

0.13

340.

1401

0.14

010.

1334

0.12

330.

1434

0.07

180.

1617

0.13

040.

1370

0.13

700.

1304

0.12

030.

1404

0.06

900.

1585

0.03

210.

0321

0.02

120.

0266

0.04

600.

0349

0.15

330.

1577

0.02

930.

0293

0.02

390.

0239

0.04

320.

0321

0.14

990.

1611

***

0.01

050.

0158

0.01

050.

0131

0.01

850.

1529

0.16

08

— ***

0.02

120.

0158

0.02

390.

0239

0.15

990.

1608

— — ***

0.00

520.

0293

0.01

320.

1461

0.14

08

— — — ***

0.02

390.

0079

0.13

940.

1474

0.00

78— — — **

*0.

0321

0.14

250.

1543

— — — — — ***

0.14

950.

1574

0.04

43— — —

0.04

42— **

*0.

1448

— — — — — — — ***

70 COPEIA, 2000, NO. 1

APP

EN

DIX

6.SE

QU

EN

CE

DIV

ER

GE

NC

E(K

2p)

WIT

HIN

Oed

ipin

a(O

edop

inol

a).U

pper

half

mat

rix

isba

sed

on16

Sse

quen

ces.

Low

erha

lfm

atri

xis

base

don

cyt

bse

quen

ces.

12

34

56

78

910

1112

13

1O

.al

leni

2O

.al

leni

3O

.al

leni

4O

.sa

vage

i5

O.

sava

gei

6O

.pa

rvip

es7

O.

parv

ipes

8O

.co

mpl

ex9

O.

com

plex

10O

.elo

ngat

a11

O.

geph

yra

12O

.ge

phyr

a13

O.

mar

itim

a

(30)

(31)

(32)

(33)

(34)

(35)

(36)

(37)

(38)

(41)

(42)

(43)

(47)

***

0.00

520.

0838

0.10

270.

1059

0.14

110.

1407

0.18

110.

1526

0.17

040.

1852

0.16

760.

1513

— ***

0.09

000.

1091

0.11

230.

1411

0.14

070.

1811

0.15

260.

1704

0.19

260.

1641

0.15

13

0.03

40— **

*0.

0664

0.06

350.

1304

0.13

670.

1422

0.16

580.

1461

0.16

650.

1461

0.13

04

0.04

43—

0.03

21**

*0.

0079

0.14

410.

1506

0.16

650.

1626

0.14

310.

1704

0.16

020.

1544

— — — — ***

0.14

070.

1471

0.16

300.

1592

0.13

970.

1668

0.15

670.

1509

0.06

99—

0.06

780.

0833

— ***

0.03

210.

0894

0.11

090.

1564

0.16

720.

1401

0.08

12

0.07

92—

0.07

040.

0954

—0.

0218

***

0.09

230.

1203

0.14

920.

1599

0.13

970.

0780

0.08

11—

0.07

890.

0903

—0.

0505

0.04

65**

*0.

1112

0.15

880.

1903

0.16

930.

1022

0.06

80—

0.05

930.

0747

—0.

0360

0.03

410.

0340

***

0.19

680.

1981

0.17

670.

1277

0.08

56—

0.08

560.

0970

—0.

0701

0.07

240.

0945

0.09

02**

*0.

1585

0.15

200.

1599

0.08

27—

0.08

480.

0961

—0.

0828

0.09

630.

1073

0.09

190.

0854

***

0.06

610.

1816

0.07

19—

0.08

090.

0945

—0.

0651

0.06

980.

0851

0.07

000.

0808

0.05

25**

*0.

1606

0.06

99—

0.07

670.

0857

—0.

0484

0.04

860.

0486

0.03

820.

0875

0.09

130.

0674

***

molecular phylogenetic analysis of relationships of the ... · phylogenetic relationships of many...

Documents