Philautus, Gistel, 1848

Yu, Guohua, Rao, Dingqi, Yang, Junxing & Zhang, Mingwang, 2008, Phylogenetic relationships among Rhacophorinae (Rhacophoridae, Anura, Amphibia), with an emphasis on the Chinese species, Zoological Journal of the Linnean Society 153 (4), pp. 733-749 : 744-745

SYSTEMATICS OF PHILAUTUS , THELODERMA , KURIXALUS AND , ‘ AQUIXALUS ’

Philautus is characterized by aerial direct development of eggs into froglets, without going through an aquatic tadpole stage, and the taxonomy of this group is still in a preliminary stage ( Bossuyt & Dubois, 2001). In the present study, P. acutirostris and P. surdus , which were recognized as valid species of Philautus by most herpetologists ( Dring, 1987; Dubois, 1987; Brown & Alcala, 1994; Bossuyt & Dubois, 2001; Wilkinson et al., 2002; Frost, 2007), form a clade with high support values. However, all cases in the current study show that this clade is not related to any Chinese species of Philautus . Bossuyt & Dubois (2001) considered that P. romeri does not belong in the genus Philautus , and tentatively referred this species to the genus Chirixalus . According to Smith (1953), the tadpole of P. romeri is ‘typically ranid’; the keratodont formula of 3/3 given by this author for this tadpole is quite unusual among Chinese Rhacophorinae , and suggests that the species might belong in a new, undescribed genus. In the present study, all phylogenetic analyses recover P. romeri as the sister taxon of a group composed of other rhacophorids, except Buergeria , although the support value is weak. That is, P. romeri is neither in a monophyletic group with Philautus or Chirixalus / Chiromantis , and nor is it obviously associated closely with any other generic grouping. For these reasons, we accept that P. romeri might deserve a new generic name ( Smith, 1953), although the monophyly consisting of P. romeri , P. acutirostris , and P. surdus cannot be rejected by all statistical tests ( P> 0.05, see Table 3). In contrast to the present study, P. palpebralis and P. gracilipes were recovered as the sister taxon of a group composed of other Rhacophorinae frogs by Wilkinson et al. (2002) and Frost et al. (2006), respectively. This might be the result of sampling differences: Wilkinson et al. (2002) did not include P. gracilipes , whereas Frost et al. (2006) did not sample P. palpebralis , and neither of them included P. romeri . In the present study, all statistical tests showed no significant differences between the constrained trees based on these two hypotheses and the unconstrained trees, at a significance of 0.05 (see Table 3). Although Frost et al. (2006) erected a new genus ‘ Feihyla ’ for P. palpebralis based on the study of Wilkinson et al. (2002), the support of Wilkinson et al. (2002) for the phylogenetic position of P. palpebralis is weak. Furthermore, the placements of P. palpebralis obtained by different methods here were not uniform, and the assumed monophyly formed by P. palpebralis , P. acutirostris , and P. surdus also cannot be rejected (see Table 3). Similarly, the placements of P. jinxiuensis and P. gracilipes have not been solved; in general, the systematics of Philautus needs much additional work as most analyses so far have not recovered this genus as monophyletic.

What is surprising is that the monophyly of Theloderma , as recovered by Wilkinson et al. (2002), is not supported by the present study. Theloderma corticale is closely related to P. rhododiscus with high support values, and T. asperum is nested in P. albopunctatus (see clade II). Furthermore, the P distance between T. asperum and P. albopunctatus (1.0–1.6%), which was inferred from the 540 bp of the 16S gene without excluding the hypervariable regions, as advocated by Vences et al. (2005), is obviously lower than the distance between T. asperum and T. corticale (11.6%). Up to now, there have been no disputes about the taxonomic positions of P. albopunctatus and P. rhododiscus . Frost et al. (2006) recovered P. rhododiscus as the sister taxon to T. corticale , but the taxonomy of P. rhododiscus was not changed because, in their study, T. corticale and P. rhododiscus were the single representatives of Theloderma and Philautus , respectively, and the sister relationship between these two species that they recovered can only be treated as an indication of close relationship between Philautus and Theloderma . To correct the paraphyly as revealed here, it is suggested that P. albopunctatus should be placed into the synonymy of T. asperum , and that P. rhododiscus should be relocated in Theloderma . After these corrections, the monophyletic group formed by Theloderma and Nyctixalus (see clade II), as shown by Wilkinson et al. (2002), is supported by Bayesian analysis (100%), although the support values of the MP and ML analyses for it are not strong.

Delorme et al. (2005) included P. idiootocus along with Rhacophorus verrucosus Boulenger, 1893 in their new genus ‘ Aquixalus ’. This may have been because they do not accept phylogenetic proximity as the organizing principle in taxonomy, because this change is not supported by their dendrogram, in which the clade [( P. idiootocus , K. eiffingeri ), R. verrucosus ] was recovered. The present results confirm previous molecular studies in grouping P. idiootocus and K. eiffingeri as a monophyletic group ( Richards & Moore, 1998; Wilkinson et al., 2002; Frost et al., 2006) (see clade III), and agree with the proposal of Wilkinson et al. (2002) that P. idiootocus belongs with K. eiffingeri in the genus Kurixalus , which has been recognized by Frost et al. (2006) and Frost (2007). According to Frost et al. (2006), the morphological similarities shared by Kurixalus and ‘ Aquixalus ’ are plesiomorphic, on the basis of the tree provided by Delorme et al. (2005). However, even if R. verrucosus is also placed in Kurixalus , as modified by Frost et al. (2006), the monophyly of ‘ Aquixalus ’ still cannot be followed. In the present study, P. gracilipes and P. odontotarsus , both of which were also placed into the genus ‘ Aquixalus ’ by Delorme et al. (2005), do not form a monophyletic group. Furthermore, all phylogenetic reconstructions presented here strongly support P. odontotarsus as the sister taxon of the group associated with Kurixalus [ K. eiffingeri and Kurixalus idiootocus (Kuramoto & Wang, 1987) ] (see clade III), whereas the position of P. gracilipes is not clear, and the monophyly formed by P. gracilipes , P. acutirostris , and P. surdus cannot be rejected ( P> 0.05, see Table 3). These indicate that the new genus ‘ Aquixalus ’ needs further evaluation.