Cophixalus ateles (Boulenger, 1898)

Kraus, Fred, 2012, Papuan frogs of the genus Cophixalus (Anura: Microhylidae): new synonyms, new species, and a dichotomous key, Zootaxa 3559, pp. 1-36 : 2-4

Cophixalus ateles (Boulenger, 1898)

Boulenger (1898) described Cophixalus ateles on the basis of several specimens obtained from Moroka, a region in Central Province, Papua New Guinea, that is approximately 20–30 km E of Port Moresby in the foothills of the Owen Stanley Mts. These specimens reside in the British Museum of Natural History in London and the Museo Civico di Storia Naturale Giacomo Doria in Genoa. The most distinctive feature of this species was its very short first finger without a dilated disc, a feature that has proven diagnostic for several other Cophixalus species described since that time ( Zweifel, 1956a, 1980; Kraus and Allison, 2000, 2009; Günther, 2006). However, since Boulenger’s description, C. ateles has remained poorly known. The name had been applied to animals from the Central Highlands ( Zweifel, 1956b; Tyler, 1963 a, b) that are now referred to C. shellyi Zweifel, 1956a . From those mid-Twentieth-Century publications it appeared that C. ateles and C. shellyi were similar species: “It is not certain that the apparently diagnostic characters of C. shellyi will continue to appear so when more specimens of shellyi and ateles become available...” ( Zweifel, 1956a), and “It does appear that shellyi is even closer to ateles than originally suspected, but until additional topotypic material of the latter is obtained, the specific status of shellyi must remain suspect...” ( Tyler, 1963b). This assumption of close relationship was based primarily on the shared character of a reduced first finger, but both Zweifel (1956a) and Tyler (1963b) noted that Boulenger’s (1898) illustration of the hand of C. ateles did not agree well with available specimens. Those discrepancies, the faded color patterns and soft state of preservation of the type series ( Tyler,1963b), and the lack of additional topotypic material of C. ateles have made exact identity of that species uncertain to the present.

More recently, Kraus and Allison (2006) named three new species of closely related Cophixalus that occur in the Papuan Peninsula of New Guinea. Of these, C. sisyphus was diagnosed by its small size (SV = 12–14.1 mm), tuberculate dorsum, reduced first finger with a poorly developed disc lacking a circum-marginal groove, variegated or striped dorsal pattern, and a variety of morphometric details distinguishing it from its close relatives C. variabilis and C. timidus , also named in that same paper. Directly relevant to the present discussion is that Kraus and Allison (2006) distinguished C. sisyphus from C. ateles by the tuberculate dorsum and variegated or striped dorsal pattern of the former. This determination was based on photographs of two paralectotypes in the British Museum, which were very faded, as noted earlier ( Tyler, 1963b), and showed no evidence of dorsal tuberculation.

Since the description of Cophixalus sisyphus two additional specimens of that species (BPBM 27492, 27494) have become available from near Moroka, Central Province, PNG, the type locality of C. ateles . This unexpected discovery, and recent opportunities to directly compare paratypes of C. sisyphus (BPBM 19304–05, 19314) to the lectotype (MSNG 29116) and seven paralectotypes of C. ateles (MSNG 50182 [5 specimens under same number], BMNH 1947.2.12.6–7) allowed me to determine whether the presumptive diagnostic features of C. sisyphus were valid vis-à-vis C. ateles . They are not.

When compared side by side, there is no morphological feature to distinguish the types of the two species. Further, except for the smooth skin and some rather shrunken measurements of some of the types of C. ateles , there is no obvious difference in shape of body proportions to distinguish them from the paratypes of C. sisyphus . This is confirmed by principal components analysis of these specimens vs. type specimens of the closely related C. sisyphus , C. timidus , and C. variabilis . In both bivariate ( Fig. 1 ) and 3-dimensional (not shown) plots of wellpreserved specimens, the two specimens of C. ateles cluster with the type series of C. sisyphus to the exclusion of the similar C. timidus and C. variabilis . The specimens of C. ateles are, indeed, at one extreme of the distribution of C. sisyphus along PC1. This axis largely reflects specimen size, judging by the uniformly positive loadings of the constituent variables ( Table 1 ). The position of the C. ateles specimens at the low end of PC1 seems to be a reflection of their smaller measurements for a number of features relative to their overall body size. This apparently reflects measurement error—especially with respect to disc width—due to the soft and somewhat shrunken state of preservation of these specimens. When disc widths for these two specimens are artificially increased to be in line with those obtained from the recent C. sisyphus types of those body sizes, both C. ateles types more closely approach the distribution of the C. sisyphus types, with the C. ateles specimen lower on the graph now lying in the midst of adjacent C. sisyphus specimens (graph not shown). Thus the little difference that appears between the C. ateles types and C. sisyphus types in multivariate analysis seems to merely reflect differences in preservation quality.

Color pattern for several of the type specimens is faded, but others show the dorsal and ventral pattern elements typical of C. sisyphus , including dark dorsal blotches and supratympanic line (two specimens), a broad mid-dorsal pale stripe margined in black (one specimen), dark punctations on the chin and chest (six specimens), and paired pale pectoral spots (two specimens). Furthermore, in basing their diagnosis of C. sisyphus vis-à-vis C. ateles on the smooth skin and lack of color pattern in the latter, Kraus and Allison (2006) overlooked the fact that Boulenger (1898) mentioned that C. ateles had small warts and some pattern elements, although the latter were not specified in detail beyond noting the occasional presence of black dorsolateral lines and a light vertebral stripe. The frequent hourglass pattern seen on the dorsum and the dark flecking and blotching often seen in C. sisyphus ( Kraus and Allison, 2006) were not mentioned by Boulenger, although the black dorsolateral lines he did mention presumably represent the lateral margins of the hourglass markings described for C. sisyphus . Given the lack of structural differences between C. ateles and C. sisyphus and the fact that the few color-pattern details given by Boulenger for C. ateles are consistent with some of the patterns seen in C. sisyphus , I synonymize C. sisyphus with C. ateles . The taxonomic history of C. ateles becomes:

Figures

FIGURE 1. Bivariate plot of principal component 1 (eigenvalue = 0.125) versus principal component 2 (eigenvalue = 0.025) for two paralectotypes of Cophixalus ateles (filled diamonds), 24 types of C. sisyphus (triangles), one recent specimen of C. sisyphus from the type locality for C. ateles (open diamond), 14 types of C. timidus (squares), and 49 types of C. variabilis (circles) showing no morphometric discrimination between C. ateles and C. sisyphus. Component loadings presented in Table 1. All specimens are male.

Tables

TABLE 1. Loadings for log-transformed characters used in principal components analysis of covariance matrix for adult males of Cophixalus ateles, C. sisyphus, C. timidus, and C. variabilis. Included specimens are listed in the Appendix.

	PC1	PC2	PC3
Eigenvalue	0.12466	0.02471	0.01154
Cumulative percent variation	0.678	0.812	0.875
Variable
logSV	0.269	0.087	0.202
logTL	0.361	0.011	0.292
logEN	0.248	0.097	0.081
logIN	0.125	0.008	-0.031
logSN	0.228	0.079	0.042
logTY	0.267	-0.951	-0.115
logEY	0.208	0.064	0.258
logHW	0.240	0.017	0.149
logHL	0.240	0.014	0.194
logHandL	0.397	0.109	0.227
log3rdF	0.469	0.172	-0.455
log4thT	0.246	0.160	-0.683