A new species of the genus Kurixalus (Anura: Rhacophoridae) is described from Yunnan, China based on morphological and molecular evidence. The new species, Kurixalus lenquanensis sp. n., is distinguished from other congeneric species by a combination of (1) smaller body size (SVL in males ranges from 25.0 to 28.9 mm), (2) obtusely pointed snout with no prominence on tip, (3) rough and brown dorsum with darker mark, (4) absence of large dark spots on ventral surface, (5) brownish clouded chin, (6) granular throat, chest, and belly, (7) presence of vomerine teeth, (8) serrated dermal fringes along outer edge of limbs, (9) slight nuptial pad, (10) golden brown iris, and (11) single internal vocal sac. The new species is known only from montane scrub vegetation at the type locality (Lenquan Village, Mengzi City, Yunnan Province) and Yangjiatian Village, Gejiu City, Yunnan Province. Genetically, the new species is nested within a clade of Taiwanese Kurixalus and recovered as the sister taxon to Kurixalus idiootocus with strong support values, indicating that the ancestor of this new species might have come from Taiwan Island or the ancestor of this new species may have been widespread in southern China and the descendent species in between Taiwan and Yunnan has become extinct.

China, Kurixalus lenquanensis sp. n., Yunnan

The genus Kurixalus Ye, Fei, & Dubois in Fei 1999 has a wide distribution including eastern India, Indochina, Sunda islands, Philippine archipelago, montane southern China, and adjacent continental islands (Hainan, Taiwan, and Ryukyu Islands); currently 14 species are recognized in this genus (Frost 2017). Owing to its morphological conservativeness, the taxonomy and systematics of this genus was once very confusing. Inger et al. (1999) considered that Chinese and Vietnamese Kurixalus odontotarsus (Ye & Fei in Ye et al. 1993) probably belong to Kurixalus verrucosus (Boulenger, 1893) or Kurixalus bisacculus (Taylor, 1962), Orlov et al. (2002) also considered Vietnamese K. odontotarsus as K. verrucosus, and Kurixalus hainanus (Zhao, Wang, & Shi in Zhao et al. 2005) was thought to be a synonym of K. odontotarsus by some authors (e.g. Fei 1999, Fei et al. 2010). However, based on evidence from molecular data, Yu et al. (2010) proposed that K. odontotarsus, K. bisacculus, and K. verrucosus should be treated as three independent species and suggested placing K. odontotarsus from Tibet and K. hainanus in K. verrucosus and K. bisacculus, respectively. Moreover, Yu et al. (2010) considered that the distributional range of K. bisacculus should be expanded to include most regions of South China. Thus, currently three species of Kurixalus (K. odontotarsus, K. bisacculus, and K. verrucosus) are recognized in mainland China and two of them (K. odontotarsus and K. bisacculus) exist in Yunnan according to Yu et al. (2010).

During recent fieldwork in Yunnan, China (Fig. 1), a small tree frog was discovered that is morphologically similar in appearance with other Kurixalus species from Yunnan. It appears to be closely related to Kurixalus idiootocus (Kuramoto & Wang, 1987) from Taiwan based on molecular phylogenetic analyses, but it is different from K. idiootocus based on the following characters: obtusely pointed snout with no prominence on tip, single internal vocal sac, loreal region oblique, absence of a pair of symmetrical large dark patches on chest, and absence of supernumerary plantar tubercles.

Figure 1. 

Known distribution of Kurixalus lenquanensis sp. n. Filled star indicates the type locality and filled circle indicates Yangjiatian Village.

Sampling. Specimens were collected during fieldwork in Honghe Autonomous Prefecture, Yunnan, China in 2015 and 2016. They were euthanized with diethyl ether anesthesia and fixed by 90% ethanol before being stored in 70% ethanol. Liver tissues were preserved in 99% ethanol. Specimens and tissue samples were deposited at Kunming Institute of Zoology, Chinese Academy of Sciences (KIZ 170175Y–170186Y) and Kunming Institute of Survey & Design, State Forestry Administration (KISD 1506203–1506204).

Morphology. The preserved specimens were examined, measured, and compared with available specimens and published descriptions of currently recognized species of Kurixalus from China and neighboring countries (Günther 1858, Boulenger 1893, Annandale 1912, Kuramoto and Wang 1987, Inger et al. 1999, Bossuyt and Dubois 2001, Matsui and Orlov 2004, Mathew and Sen 2008, Nguyen et al. 2014a, Nguyen et al. 2014b, Wu et al. 2016). Measurements were taken using digital calipers to the nearest 0.1 mm. Morphological terminology follows Fei (1999). Measurements include: snout-vent length (SVL, from tip of snout to vent); head length (HL, from tip of snout to rear of jaws); head width (HW, width of head at its widest point); snout length (SL, from tip of snout to anterior border of eye); internarial distance (IND, distance between nares); interorbital distance (IOD, minimum distance between upper eyelids); upper eyelid width (UEW, maximum width of upper eyelid); eye diameter (ED, diameter of exposed portion of eyeball); tympanum diameter (TD, the greater of tympanum vertical and horizontal diameters); distance from nostril to eye (DNE, from nostril to anterior border of eye); forelimb length (FLL, from elbow to tip of third finger); tibia length (TL, distance from knee to heel); foot length (FL, from proximal end of inner metatarsal tubercle to tip of fourth toe); and thigh length (THL, from vent to knee).

A multivariate principal component analysis (PCA) was conducted using SPSS 17.0 (SPSS Inc.) based on a correlation matrix of size-standardized measurements (all measurements divided by SVL). Scatter plots of the scores of the first two factors of PCA were used to examine the differentiation among the new species, K. idiootocus, K. bisacculus, and K. odontotarsus.

Molecular analyses. In order to support the generic placement of the new species, the phylogenetic position of the new species was investigated based on molecular data. Total genomic DNA was extracted from liver tissue stored in 99% ethanol. Tissue samples were digested using proteinase K, and subsequently purified following a standard phenol/chloroform isolation and ethanol precipitation. A fragment of the mitochondrial 16S rRNA gene was amplified and sequenced. The primers and experiment protocols used in this study are the same as with Yu et al. (2010). Homologous sequences of other Kurixalus species were obtained from GenBank and all new sequences have been deposited in GenBank under Accession Nos. KY768931–KY768944 (Table 1). Philautus abditus Inger, Orlov, & Darevsky, 1999 and Raorchestes menglaensis (Kou, 1990) were selected as outgrups based on Yu et al. (2013).

Sequences were aligned using CLUSTAL X v1.83 (Thompson et al. 1997) with the default parameters and the alignment was revised by eye. Pairwise distances between species were calculated in MEGA 5 (Tamura et al. 2011). The best substitution model was selected using the Akaike Information Criterion (AIC) in MODELTEST v3.7 (Posada and Crandall 1998). Bayesian phylogenetic inference and Maximum likelihood analysis were performed in MRBAYES 3.1.2 (Huelsenbeck and Ronquist 2001) and RAXMLGUI 1.5b1 (Silvestro and Michalak 2012), respectively, based on the selected substitution model. For the Bayesian analysis, two runs were performed simultaneously with four Markov chains starting from random tree. The chains were run for 4,000,000 generations and sampled every 100 generations. The first 25% of the sampled trees was discarded as burn-in after the standard deviation of split frequencies of the two runs was less than a value of 0.01, and then the remaining trees were used to create a 50% majority-rule consensus tree and to estimate Bayesian posterior probabilities (BPPs). For the maximum likelihood analysis, node support was estimated by 1,000 rapid bootstrap replicates.

Molecular data. The obtained sequence alignment is 870 bp in length. Both Bayesian and maximum likelihood analyses strongly supported that Kurixalus lenquanensis sp. n. is in the genus Kurixalus and is the sister taxon to K. idiootocus (Fig. 2). The average divergence (p-distance) between the new species and other congeneric species ranged from 4.72% to 16.84% (Table 2). This level of divergence in the 16S rRNA gene is indicative of differentiation at the species level in frogs (Fouquet et al. 2007) and is even higher than between other recognized congeners of Kurixalus (i.e. Kurixalus eiffingeri [Boettger, 1895]/Kurixalus berylliniris Wu, Huang, Tsai, Li, Jhang, & Wu, 2016, K. eiffingeri/Kurixalus wangi Wu, Huang, Tsai, Li, Jhang, & Wu, 2016, K. berylliniris/K. wangi, K. bisacculus/K. odontotarsus, Kurixalus baliogaster [Inger, Orlov, & Darevsky, 1999]/K. odontotarsus, and K. baliogaster/K. bisacculus; Table 2).

Figure 2. 

Bayesian phylogram of Kurixalus inferred from 870 bp of 16S rRNA gene. Numbers above and below branches are Bayesian posterior probabilities (BPP) and ML bootstrap values (only values above 50% are shown), respectively.

Morphological data. We retained the first two principal components which accounted for 43.7% of the total variance and had eigenvalues above 1.0. Loadings for PC 1 were all positive except for TD and DNE and were most heavily loaded on HW, IND, and ED (Table 3). No difference was found along the PC 1 axis between the four species. The second principal component (PC 2) accounted for 20.96% of the total variance and loaded heavily and positively on SL and DNE and negatively on FL. Differentiation was found along the PC 2 axis between the new species and K. idiootocus, K. odontotarsus, and K. bisacculus (Fig. 3). The result indicates that the new species differs from K. idiootocus, K. odontotarsus, and K. bisacculus by smaller ratio of SL divided by SVL, smaller ratio of DNE divided by SVL, and higher ratio of FL divided by SVL.

Figure 3. 

Scatterplot of principal components 1 and 2 of size-adjusted morphometric data for males of K. lenquanensis sp. n., K. idiootocus, K. bisacculus, and K. odontotarsus.

Although morphological synapomorphies of the genus Kurixalus are still not very clear (Yu et al. 2013, Nguyen et al. 2014b), intuitively K. lenquanensis sp. n. can be placed in Kurixalus because of its morphological similarity to other members of the genus (e.g. small body size, inner and outer fingers not opposable, poorly developed finger webbing, moderately developed toe webbing, and serrated dermal fringes on forearm and tarsus). Other small rhacophorid species in the genera Feihyla, Gracixalus, Chiromantis, or Philautus generally lack serrated fringes on forearm and tarsus and lack vomerine teeth (Fei 1999, Fei et al. 2010). Additionally, inner (first and second) and outer (third and fourth) fingers are opposable in all species of Feihyla and Chiromantis (Fei 1999, Fei et al. 2010). This assignment is supported by the molecular data, which indicates that K. lenquanensis sp. n. is nested in the genus Kurixalus with strong support values.

It is very interesting biogeographically that K. lenquanensis sp. n. is nested within a clade consisting of Taiwanese Kurixalus with strong support, indicating that the ancestor of the new species may have been from Taiwan Island. Another plausible scenario is that the ancestor of this new species may have been widespread in southern China and the descendent species in between Taiwan and Yunnan has become extinct. Although K. ananjevae and K. naso are not included in the present study, absence of them would have no impact on the phylogenetic position of the new species because K. ananjevae likely does not belong to the genus (Yu et al. 2013), and morphologically, K. naso is more similar to members of the K. odontotarsus species group than to K. lenquanensis sp. n. in body size and ventral color pattern.

Reproductive behavior among Taiwanese relatives of the new species varies; Kurixalus idiootocus lays pigmented eggs on land near the edge of water or in depressions where rainfall accumulates (Kuramoto and Wang 1987), whereas K. eiffingeri, K. berylliniris, and K. wangi lay eggs inside tree hollows or cut bamboos with water (Fei et al. 2010, Wu et al. 2016). Although no eggs of K. lenquanensis sp. n. were found, reproductive behavior of this new species probably is closer to that of K. idiootocus than to that of the other three Taiwanese species because 1) it is the sister taxon to K. idiootocus and 2) no tree hollows or cut bamboos were found at the type locality.

Species boundaries among members of the genus Kurixalus were previously confusing and our earlier work, based on molecular data (Yu et al. 2010), supported that there are three valid members of Kurixalus in mainland China. Therefore, with the new species described here, there are currently four Kurixalus species in mainland China: K. bisacculus, K. lenquanensis sp. n., K. odontotarsus, and K. verrucosus. However, considering the obvious geographical discontinuity in distribution between K. lenquanensis sp. n. and its congeneric relatives from Taiwan Island, additional undiscovered species of Kurixalus may exist in south China.

We are indebted to S. Lin for the donation of specimens of K. idiootocus from Taiwan and to T. Chan-ard for the loan of specimens of K. bisacculus from Thailand. Thanks also to E. J. Ely for assistance with taking measurement of K. verrucosus, to R. Min for her help with laboratory work, and to two reviewers for their valuable comments on the manuscript. This work was supported by the National Natural Science Foundation of China (No. 31301870), CAS “Light of West China” Program to G. Yu, and Natural Science Foundation of Yunnan Province (No. 2015FB176) to J. Wang.