From mainland to islands: colonization history in the tree frog Kurixalus (Anura: Rhacophoridae).

The origin and colonization history of Kurixalus, a genus of small arboreal tree frogs breeding exclusively in shallow swamps, is under disputed. On the basis of comprehensive sampling program, the evolutionary history of Kurixalus is investigated based on 3 mitochondrial genes. Our results indicate that the genus Kurixalus originated in the Asian mainland and subsequently arrived at its current distribution in Borneo, Taiwan, Ryukyu, and Hainan islands by a series of dispersal events. Moreover, the colonization of Taiwan from mainland Asia has occurred 2 times. The initial colonization of Taiwan occurred at 3.46-8.68 Mya (95% highest posterior density), which rejects the hypothesis that Kurixalus probably originated from Taiwan during the early Oligocene and favors the model of Neogene-origin rather than the model of Quaternary-origin for Taiwanese Kurixalus. Kurixalus eiffingeri has dispersed from Taiwan to the Ryukyus once or 2 times pending more data. Both transoceanic dispersal and landbridge dispersal have played a role in the colonization process; the former resulted in the colonization of Taiwan and the Ryukyus and the latter led to the colonization of Borneo and Hainan.

Taiwan and Ryukyu islands are a part of the island-arc system along the western edge of the Pacific Ocean (Vita-Finzi 2000); the former is separated from Asian mainland by the Taiwan Strait, a shallow strait about 130-km-wide at narrowest point and 60 m in average depth (Figure 1A). These continental islands emerged from the waters almost simultaneously owing to the collision between the Luzon arc and Eurasian margins during the Late Miocene (Sibuet and Hsu 2004; Huang 2017). After the emergence of Taiwan Island in the Late Miocene (ca. 6.5 Mya; Huang 2017) or early Pliocene (4–5 Mya; Ali 2018, 2020), they gradually acquired their floras and faunas mostly from the Eurasian mainland via exposed landbridge across the Taiwan Strait (Wang 1987; Hikida and Ota 1997) or via transoceanic dispersal (He et al. 2018). Based on the splitting time between the island species and their mainland relatives, 2 origin models were proposed for the colonization of Taiwan: the Quaternary-origin and Neogene-origin models (Lin et al. 2002; Su et al. 2016).

Figure 1.

Map showing the connection and isolation of Taiwan to mainland China when sea level is 0 m (A), –60 m (B), and –120 m (C). Fluctuations of sea level during the past millions years (D; Miller et al. 2005, 2011) and during the last million years (E; Bintanja et al. 2005) are presented. The map was generated using the GeoMapApp online software system (Ryan et al. 2009).

Members of the frog family Rhacophoridae are an example of a group whose distribution range extends across the entire region mentioned above. Rhacophorids are small-to-large frogs bearing several arboreal adaptations, including cartilaginous intercalary elements between the terminal and penultimate phalanges, and toe discs. This family displays a remarkable variety of reproductive modes including aquatic development, terrestrial gel-nesting, terrestrial foam-nesting, and terrestrial direct development, which facilitates less dependence on standing water bodies and may have provided new opportunities for dispersal (Meegaskumbura et al. 2015), and it is widely distributed across Asia with a disjunct occurrence in Africa (Frost 2020). Consistent with the Quaternary-origin model, rhacophorid frogs were supposed to have migrated from Chinese continent to Taiwan via landbridge during the Pleistocene glacial periods (Lue and Chen 1986), but there is no fossil record showing when the migration happened and few empirical studies have been performed to test this hypothesis (Yang et al. 1994). The genus Kurixalus, which is 1 of 2 rhacophorid genera occurring in Taiwan, currently comprises 18 recognized species (Frost 2020) and 5 unnamed lineages (Yu et al. 2018), distributed widely on the Asian continent and adjacent islands, including Taiwan, Ryukyu, Hainan, Sunda Islands, and Philippine archipelago. Contrary to the hypothesis of Lue and Chen (1986), Lv et al. (2018) considered that Kurixalus have originated in Taiwan during the Oligocene (32.8 Mya, 95% confidence interval : 25.6–39.9) and there was no exchange of Kurixalus between mainland China and Taiwan Island. However, these interpretations do not match with the geological evidence that proto-Taiwan Island did not emerged from water until the Late Miocene (6.5 Mya; Huang 2017) or early Pliocene (4–5 Mya; Ali 2018, 2020), which has also been pointed out by Ali (2020). Moreover, the ancestor of Kurixalus lenquanensis might have come from Taiwan Island (Yu et al. 2017a). Thus, origin and colonization history of Kurixalus needs further examination.

Here we reconstructed the phylogenetic relationships and ancestral biogeographic areas of Kurixalus and estimated the lineage divergence times to investigate the evolutionary history of Kurixalus and to test for the hypothesis of Quaternary-origin of Taiwanese rhacophorid frogs.

Materials and Methods

Data preparation

The classifications of Frost (2020) and Yu et al. (2018) were followed and a total of 57 individuals belonging to 15 recognized and 6 unnamed species of Kurixalus were included in this study (Appendix). Sequences encoding 3 mitochondrial genes (12S rRNA, 16S rRNA, and COI) were amplified and sequenced for K. lenquanensis and Kurixalus sp6 using the primers and experiment protocols of Yu et al. (2017b), and homologous sequences of other species were downloaded from GenBank. Eleven species representing 10 genera of Rhacophoridae were included as hierarchical outgroups according to Yu et al. (2009) and homologous sequences of them were obtained from GenBank.

Phylogenetic analysis

Sequences were aligned using the MUSCLE option in MEGA version 7.0 (Kumar et al. 2016) with the default parameters. Fragments of the 3 mitochondrial genes were treated as 1 partition for subsequent analyses because not all homologous sequences are available for some species. The Corrected Akaike Information Criterion (AICc) was used to select the best model of nucleotide substitution in jModeltest version 2.1.10 (Darriba et al. 2012). Bayesian phylogenetic inference and Maximum likelihood analysis were performed in MrBayes version 3.1.2 (Ronquist et al. 2012) and RAxML-HPC version 8.2.10 (Stamatakis 2014), respectively, based on the selected substitution model. For the Bayesian analysis, 2 runs were performed simultaneously with 4 Markov chains starting from random trees and the chains were run for 5,000,000 generations being sampled every 100 generations. Convergence and burn-in were checked using the program Tracer version 1.6 (Rambaut et al. 2014). Finally, the first 25% of the sampled trees were discarded as burn-in and the remaining trees were used to create a consensus tree and to estimate Bayesian posterior probabilities (BPPs). For the maximum likelihood analysis, node support was estimated by 1,000 rapid bootstrap replicates.

Divergence dating

Lineage divergence times were estimated using an uncorrelated lognormal relaxed molecular clock model in BEAST version 1.8.0 (Drummond et al. 2012). We chose the birth-death process as the tree prior because of the mixed inter- and intraspecies sampling in the data sets (Ritchie et al. 2017). As there is no explicit fossil record for rhacophorid frogs, we adopted an inferred range of mitochondrial sequence divergence rate to calibrate the rate of divergence. Previous empirical studies of amphibians (using a variety of calibrations, taxa, and mtDNA gene fragments) have inferred model-corrected mitochondrial sequence divergence rates between 0.8% and 1.9% (after Sanguila et al. 2011). Thus, we used a mean rate of 1.35% with a normal distribution (95% CI: 0.008–0.019) as calibration following Sanguila et al. (2011) and Gonzalez et al. (2014). The estimated temporal framework for diversification in Sanguila et al. (2011) is in line with the earlier geological events of the Mindanao Islands and is generally consistent with the postulated common mechanism that may have promoted vertebrate species diversification in the Philippine archipelago (Brown and Diesmos 2009), indicating that this choice is appropriate.

Two independent runs were conducted for 5 × 107 generations by sampling every 1,000 generations. The effective sample size (≥200) values of parameters were confirmed in Tracer version 1.6 (Rambaut et al. 2014) and then trees produced by the 2 runs were combined in LogCombiner version 1.8.0 (Drummond et al. 2012) with the first 25% of trees from each run discarded as burn-in. The maximum clade credibility tree was constructed in TreeAnnotator version 1.8.0 (Drummond et al. 2012) and was visualized in FigTree version 1.4.0 (from http://tree.bio.ed.ac.uk/software/figtree).

Reconstruction of ancestral area

We divided the present distribution of Kurixalus into 6 different biogeographic regions including South Tibet-Indian subcontinent (1), Indochina-South China (2), Taiwan Island (3), Ryukyu Islands (4), Hainan Island (5), and Sunda Islands-Philippine archipelago (6), and then assigned each species to its own region. Combined the divergence dating analysis in BEAST version 1.8.0, we used the Bayesian stochastic search variable selection (BSSVS; Lemey et al. 2009) of the discrete phylogeographic model to reconstruct ancestral ranges of lineages within Kurixalus with the specification of symmetric discrete trait substitution model.

In addition to the BSSVS analysis, ancestral ranges and biogeographical history of Kurixalus were also estimated using BioGeoBEARS package (Matzke 2013a) implemented in RASP version 4.0 (Yu et al. 2015). This method implements 3 models including dispersal–extinction–cladogenesis (DEC; Ree and Smith 2008), the likelihood version of dispersal–vicariance (Ronquist 1997; herein DIVALIKE), and the likelihood version of BayArea model (Landis et al. 2013; herein BAYAREALIKE). In each model, an additional j parameter for founder events (jump dispersal) was added; thus, a total of 6 models resulted (Matzke 2013b). We compared the fit of each model using the AIC weighted approach (Burnham and Anderson 2002) and then ancestral ranges and biogeographical events were estimated using the best-fit model. For this analysis, the ultrametric time-calibrated trees generated from BEAST analyses were used with the same range coding as BSSVS analysis. The maximum number of individual unit areas was set to 2 because no species occupied more than 2 areas in Kurixalus and the number of random trees was set to 3,000.

Results

Phylogeny

Our mitochondrial gene fragments consisted of 402 bp from 12S rRNA, 878 bp from 16S rRNA, and 807 bp from COI region. The model TIM2 + I + G was selected as the best-fit model by AICc. As this model is not available in MrBayes or RAxML, we replaced it with the GTR + I + G model in the phylogenetic analyses because the TIM2 model is a special case of the GTR model. Bayesian inference and Maximum likelihood analysis yielded similar topologies (Figure 2). Kurixalus appendiculatus was reconstructed as the sister to the clade consisted of all other congeners. The 4 species from Taiwan and the Ryukyu islands did not form a monophyly and they were grouped in 2 sibling clades with strong support values: one consisting of 3 island species (K. eiffingeri, K. berylliniris, and K. wangi) and one comprising 1 island species (K. idiootocus) and 2 mainland species (K. lenquanensis and K. sp6). Furthermore, K. lenquanensis and K. sp6 are paraphyletic with respect to K. idiootocus. The 4 specimens of K. eiffingeri from the Ryukyu Islands (A120, E43, E44, and E45) were grouped into 2 separated lineages with weak support, indicating that K. eiffingeri in the Ryukyu Islands probably have originated once or 2 times.

Figure 2.

Bayesian phylogram of Kurixalus inferred from 2087 bp of 12S rRNA, 16S rRNA, and COI sequences. Numbers above and below branches are Bayesian posterior probabilities and ML bootstrap values (only values >50% are shown), respectively. Distribution of Kurixalus species are highlighted with different color.

Divergence time estimation and ancestral range estimation

The initial divergence within Kurixalus was dated back to ca. 10.97 Mya (95% highest posterior density [HPD]: 7.03–16.54) (Figure 3B, Table 1). The time of most recent common ancestor (TMRCA) of all Taiwanese Kurixalus was estimated to be 5.60 Mya (95% HPD: 3.46–8.68) and the TMRCA of K. idiootocus and K. sp6 was estimated to be ca. 1.58 Mya (95% HPD: 0.76–2.77). The 2 lineages of K. eiffingeri from the Ryukyu Islands were grouped together in the BEAST analysis (Figure 3) and they diverged from Taiwanese lineages ca. 2.18 Mya (95% HPD: 1.31–3.43).

Figure 3.

Biogeographic division of Kurixalus (a) and divergence dating and ancestral area reconstruction (b). Color of branch indicates the BSSVS reconstruction and color of node cycle indicates the DIVALIKE + j reconstruction. Detected historical colonized events are labeled with numbers at nodes. The original map (approval number: GS(2019)1652) was downloaded from the website of Ministry of Natural Resources of the People’s Republic of China.

Table 1.

Divergence time estimates (95% HPD) and BSSVS and DIVALIKE + j ancestral range reconstructions for selected nodes as defined in Figure 3B (only area with highest probability is presented)

Node	Node age (95% HPD) Mya	Ancestral range reconstructions
		BSSVS	DIVALIKE + j
1	10.97 (7.03–16.54)	ii, 0.974	ii, 0.987
2	5.60 (3.46–8.68)	ii, 0.794	ii, 0.962
3	1.58 (0.76–2.77)	ii, 0.891	ii, 0.976
4	2.18 (1.31–3.43)	iii, 0.909	iii, 0.829
5	1.89 (1.02–3.11)	iii, 0.959	iii, 0.988
6	0.39 (0.17–0.71)	ii, 0.991	ii, 0.988

The BioGeoBEARS analyses revealed that the model of DIVALIKE + j was the best-fit biogeographical model (Table 2) and the 2 methods reconstructing ancestral range (BSSVS and DIVALIKE + j) obtained similar results (Figure 3B, Table 1). The ancestral range of Kurixalus was inferred to be Indochina-South China with single colonization of Sunda Islands-Philippine archipelago (Figure 3B). Two colonization events from mainland China to Taiwan Island were identified; 1 for the ancestor of K. eiffingeri, K. berylliniris, and K. wangi and 1 for the lineage giving rise to K. idiootocus. Kurixalus eiffingeri dispersed from Taiwan to the Ryukyu Islands. There were another 2 colonization routes including the dispersal of the ancestor of K. naso to southern Tibet-Indian subcontinent from Indochina-South China and the dispersal of K. hainanus to Hainan Island from Asian mainland.

Table 2.

Comparison of the 6 models of ancestral area estimations of Kurixalus

Model	Ln L	n	d	e	j	AICc_wt
DEC	−57.08	2	0.0060	0.0084	0	1.0 × 10−5
DEC + j	−45.84	3	1.00 × 10−12	1.00 × 10−12	0.012	0.26
DIVALIKE	−52.71	2	0.0061	1.00 × 10−12	0	0.0008
DIVALIKE + j	−44.86	3	1.00 × 10−12	1.00 × 10−12	0.012	0.70
BAYAREALIKE	−73.26	2	0.0087	0.060	0	9.8 × 10−13
BAYAREALIKE + j	−47.99	3	1.00 × 10−7	1.00 × 10−7	0.013	0.031

The model with highest value of AICc_wt was selected as the best model. n, number of parameters; d, rate of dispersal; e, rate of extinction.

Discussion

The ancestral area reconstruction indicated that originally the genus Kurixalus occurred in Asian mainland and subsequently it dispersed to the continental islands of East and Southeast Asia by multiple colonization events. The most ancient divergence in Kurixalus is the split between Bornean species and the remaining species at ca. 10.97 Mya (95% HPD: 7.03–16.54) (Node 1; Figure 3). A continuous land connection between Borneo and mainland Southeast Asia existed throughout much of the Neogene during times of lowered sea levels and would have allowed fairly unhindered migration of terrestrial biota (Moss and Wilson 1998). The ancestor of Kurixalus endemic to Borneo might have colonized the former Sundaland from mainland Southeast Asia during times of lowered sea levels. Subsequently, it might be isolated from its mainland relatives owing to the rise of the sea level during the global middle Miocene climate maximum, which resulted in extensive inundations of the former Sundaland and the loss of connection of Borneo to mainland Indochina (Hall 1998; Stelbrink 2015).

Contrary to the hypothesis that Kurixalus probably originated from Taiwan and there was no dispersal across the Taiwan Strait (Lv et al. 2018), we found that Taiwanese Kurixalus probably originated from mainland Asia via 2 colonization events (Figure 3). Initially, the ancestor of K. eiffingeri, K. berylliniris, and K. wangi dispersed to Taiwan and split from mainland species at 3.46–8.68 Mya (95% HPD; Figure 3, Table 1). This estimation coincides with that the Taiwan Island initially emerged from the waters at 4–5 Mya (Ali 2018, 2020) or 6.5 Mya (Huang et al. 1997; Lin et al. 2003; Huang 2017). Based on this result, we reject the hypothesis that Taiwanese rhacophorid frogs have originated from mainland China during the Pleistocene glacial periods (Quaternary-origin model; Lue and Chen 1986), instead favoring the model of Neogene-origin for Taiwanese Kurixalus. The late Miocene-origin or early Pliocene-origin has also been suggested for other Taiwanese animals (e.g., Creer et al. 2001; Lin et al. 2002; Shih et al. 2006; Chiang et al. 2010; Su et al. 2016). There are 2 potential mechanisms to explain the initial dispersal of Kurixalus to Taiwan: landbridge connection or transoceanic dispersal. The data of oxygen isotope demonstrated that the landbridge across the Taiwan Strait did not emerge until the sea level was frequently <‒60 m after 2.6 Mya (Figure 1B; Miller et al. 2005, 2011). Therefore, we consider that the initial colonization of Taiwan in Kurixalus might be the result of transoceanic dispersal, which has been supported in other insular amphibians (e.g., Vences et al. 2003; Komaki et al. 2017) and has been proposed as the predominant pathway for the colonization of Taiwan (He et al. 2018) and Madagascar (Crottini et al. 2012).

After the initial colonization of the Taiwan Island in Kurixalus during Pliocene to Late Miocene, the lineage giving rise to K. idiootocus also dispersed to the Taiwan Island and split from its mainland sister at 0.76–2.77 Mya (95% HPD; Figure 3, Table 1). As illustrated in Figure 1B, sea level was frequently <‒60 m during the Pleistocene, which resulted in repeated landbridge connections between the Taiwan Island and mainland Asia since 2.6 Mya. Thus, we posit that the lineage of K. idiootocus might have colonized the Taiwan Island by dispersal via landbridge, although it is also possible that the ancestor of K. idiootocus has arrived at Taiwan via transoceanic dispersal at 2.6–2.77 Mya.

Kurixalus eiffingeri might have dispersed from Taiwan to the Ryukyu Islands once or 2 times (Figures 2 and 3). More data will be needed to resolve the phylogenetic relationship between the 2 Ryukyus clades and to achieve a precise estimation of number of dispersal between Taiwan and Ryukyus (Figure 2). The Taiwan-Yonaguni Strait between the Ryukyu Islands and Taiwan has a maximum sill depth of 770 m (Nakamura 2013) with average depth of >200 m (Osozawa et al. 2012), implying that a landbridge between Taiwan and Ryukyu has never occurred during past millions years, even if the sea level was below ‒120 m (Figure 1). Thus, we consider that the colonization of Ryukyu from Taiwan in K. eiffingeri was also the result of transoceanic dispersal. Similar colonization route has also been reported for the Japanese stream treefrog (Komaki et al. 2017) and the Okinawa tree lizard (Yang et al. 2018).

Kurixalus hainanus in Hainan originally came from Asian mainland and split from mainland populations at 0.17–0.71 Mya (Node 6; Figure 3). Although Hainan is separated from southern mainland China by the Qiongzhou Strait, they were joined together frequently when the sea level was <‒40 m during the past million years (Figure 1D) because the strait is shallow with depth <40 m (Ali 2018). Thus, we consider that K. hainanus colonized Hainan via landbridge dispersal, which was also supported in bat (Mao et al. 2010) and lizard (Huang et al. 2013). Considering that the landbridge across the Qiongzhou Strait occurred multiple times during the past millions years (Shi et al. 2006), a phylogeographic study employing broad sampling will be necessary to investigate whether K. hainanus colonized Hainan once or multiple times and whether K. hainanus recolonized mainland Asia, which has been detected in other species (e.g., Mao et al. 2010).

The ancestral range of the lineage giving rise to K. yangi, K. naso, and K. sp1 was reconstructed as Indochina-South China and the split between K. naso and the clade of K. yangi and K. sp1 was estimated to be ca. 1.89 Mya (95% HPD: 1.02–3.11) (Node 5; Figure 3). This result supports the hypothesis that Kurixalus might have colonized the Indian subcontinent from northern Indochina (Yu et al. 2018). Lv et al. (2018) supposed that Kurixalus might have colonized the Indian subcontinent via a landbridge on the basis of the record of K. appendiculatus in India (Dinesh et al. 2013). This inference should be treated with caution because K. appendiculatus likely only occurs in Philippines and Sunda islands and the record of K. appendiculatus in India is likely a misidentification of K. naso or other relative according to our earlier work (Yu et al. 2017b). Geographically K. naso and the clade of K. yangi and K. sp1 are distributed on the west and east side of Patkai Range, respectively, implying that the uplift of Patkai Range during the early Pleistocene might have driven the divergence between them.

In conclusion, the genus Kurixalus originally occurred in Asian mainland and subsequently it arrived at its current distribution in Borneo, Taiwan, Ryukyu, and Hainan by a series of dispersal events. Both transoceanic dispersal and landbridge dispersal have played a role in the colonization process; the former resulted in the colonization of Taiwan and Ryukyu and the latter led to the colonization of Borneo and Hainan. In addition, Kurixalus might have colonized the Indian subcontinent from northern Indochina. Results in the present study favor the model of Neogene-origin rather than the model of quaternary-origin or Oligocene-origin for Taiwanese Kurixalus. Considering that only mitochondrial markers were used in this study owing to the absence of nuclear data for most Kurixalus species outside of China, which may lead to a biased understanding on the phylogeny if evolutionary factor such as mitochondrial capture, gene introgression, or incomplete lineage sorting occurred, further evidence based upon inclusion of more taxa and nuclear markers will be needed to achieve a complete understanding on the evolutionary history of Kurixalus.

Species	Locality	Voucher No	12S	16S	COI
Buergeria buergeri	Hiroshima, Japan	–	AB127977	AB127977	AB127977
Liuixalus ocellatus	Hainan, China	–	GQ285672	GQ285672	KP996840
Theloderma rhododiscus	Guangxi, China	–	EU215530	EU215530	KP996753
Feihyla palpebralis	Yunnan, China	–	EU215546	EU215546	DQ468683
Gracixalus gracilipes		–	GQ285668	GQ285668	KR087672
Chiromantis rufescens	Africa	–	AF458126	AF458126	KY079577
Rhacophorus schlegelii	Hiroshima, Japan	–	AB202078	AB202078	AB202078
Polypedates megacephalus	Guangxi, China	–	AY458598	AY458598	AY458598
Philautus abditus	Krong Pa, Vietnam	–	GQ285673	GQ285673	–
Raorchestes menglaensis	Yunnan, China	–	GQ285676	GQ285676	–
Raorchestes parvulus	Thailand	–	LC012865	LC012865	KR087900
Kurixalus appendiculatus	Bukit Sarang, Sarawak, Malaysia	FMNH 267896	JQ060948	JQ060937	KX554539
Kurixalus eiffingeri	Ryukyu Islands, Japan	A120	–	DQ468673	DQ468681
Kurixalus eiffingeri	Taiwan, China	11320	–	DQ468672	DQ468680
Kurixalus eiffingeri	Taiwan, China	11333	–	DQ468670	DQ468678
Kurixalus eiffingeri	Taiwan, China	E01	–	–	KT259075
Kurixalus eiffingeri	Taiwan, China	E03	–	–	KT259077
Kurixalus eiffingeri	Taiwan, China	E05	–	–	KT259079
Kurixalus eiffingeri	Taiwan, China	E22	–	–	KT259096
Kurixalus eiffingeri	Taiwan, China	E27	–	–	KT259101
Kurixalus eiffingeri	Taiwan, China	E36	–	–	KT259110
Kurixalus eiffingeri	Ryukyu Islands, Japan	E43	–	–	KT259117
Kurixalus eiffingeri	Ryukyu Islands, Japan	E44	–	–	KT259118
Kurixalus eiffingeri	Ryukyu Islands, Japan	E45	–	–	KT259119
Kurixalus eiffingeri	Taiwan, China	E46	–	–	KT259120
Kurixalus idiootocus	Taiwan, China	A127	–	DQ468674	DQ468682
Kurixalus idiootocus	Taiwan, China	–	–	AB933306	KT259131
Kurixalus berylliniris	Taiwan, China	11311	–	DQ468669	DQ468677
Kurixalus berylliniris	Taiwan, China	B02	–	–	KT259056
Kurixalus berylliniris	Taiwan, China	B05	–	–	KT259059
Kurixalus wangi	Taiwan, China	11328	–	DQ468671	DQ468679
Kurixalus wangi	Taiwan, China	W11	–	–	KT259074
Kurixalus banaensis	Krong Pa, Gia Lai, Vietnam	ROM 32986	GQ285667	GQ285667	–
Kurixalus viridescens	Hon Ba, Khanh Hoa, Vietnam	VNMN 03802	AB933284	AB933284	–
Kurixalus motokawai	Kon Tum, Vietnam	VNMN 03458	LC002888	LC002888	–
Kurixalus bisacculus	Pua, Nan, Thailand	THNHM 10051	GU227279	GU227334	KX554633
Kurixalus bisacculus	Phu Luanag, Loei, Thailand	KUHE 19333	KX554473	KX554535	KX554645
Kurixalus baliogaster	Krong Pa, Gia Lai, Vietnam	Rom 33963	KX554474	KX554536	KX554646
Kurixalus baliogaster	Tram Lap, Gia Lai, Vietnam	Rom 29860	KX554475	KX554537	KX554647
Kurixalus baliogaster	Tram Lap, Gia Lai, Vietnam	Rom 29862	KX554476	KX554538	KX554648
Kurixalus naso	Muotuo, Tibet, China	Rao 06301	KX554422	KX554484	KX554547
Kurixalus naso	Muotuo, Tibet, China	Rao 06302	KX554423	KX554485	KX554548
Kurixalus yangi	Mohynin, Kachin, Myanmar	CAS 231491	GU227278	GU227333	KX554556
Kurixalus yangi	Longdao, Ruili, Yunnan, China	Rao 14102901	KX554429	KX554491	KX554557
Kurixalus yangi	Longdao, Ruili, Yunnan, China	Rao 14102903	KX554431	KX554493	KX554559
Kurixalus odontotarsus	Mengyang, Yunnan, China	YGH 090175	GU227233	GU227282	KX554573
Kurixalus odontotarsus	Caiyanghe, Yunnan, China	YGH 090131	GU227240	GU227290	KX554576
Kurixalus hainanus	Wenshan, Yunnan, China	YGH 090044	GU227248	GU227299	KX554599
Kurixalus hainanus	Longmeng, Guangdong, China	YGH 090201	GU227265	GU227320	KX554616
Kurixalus hainanus	Diaoluo Mts, Hainan, China	Rao14111301	KX554461	KX554523	KX554620
Kurixalus hainanus	Mt. Wuzhi, Hainan, China	MVZ Herp 236722	JQ060939	JQ060928	KX554624
Kurixalus lenquanensis	Mengzi, Yunnan, China	YGH 20160036	MK348042	KY768931	MK348050
Kurixalus lenquanensis	Mengzi, Yunnan, China	YGH 20160043	MK348043	KY768938	MK348051
Kurixalus sp1	Nagmung, Kachin, Myanmar	CAS 224381	GU227274	GU227329	KX554552
Kurixalus sp1	Nagmung, Kachin, Myanmar	CAS 224563	GU227275	GU227330	KX554553
Kurixalus sp1	Nagmung, Kachin, Myanmar	CAS 225128	GU227276	GU227331	KX554554
Kurixalus sp2	Pingbian, Yunnan, China	YGH 080166	GU227244	GU227295	KX554582
Kurixalus sp2	Tuong Duong Dist, Nghe An, Vietnam	FMNH 255661	KX554450	KX554512	KX554586
Kurixalus sp2	Pakxong Dist, Champasak, Lao	FMNH 257903	KX554458	KX554520	KX554594
Kurixalus sp3	Muang Sa Kaeo, Sa Kaeo, Thailand	FMNH 265820	KX554465	KX554527	KX554635
Kurixalus sp3	Phnom Samkos, Cambodia	LSUHC 9430	KX554466	KX554528	KX554636
Kurixalus sp4	Kampot Dist, Kampot Prov, Cambodia	FMNH 261898	KX554467	KX554529	KX554637
Kurixalus sp4	Kampot Dist, Kampot Prov, Cambodia	FMNH 261899	KX554468	KX554530	KX554638
Kurixalus sp5	Pilok, Kanchanaburi, Thailand	KUHE 35069	AB933291	AB933291	KX554642
Kurixalus sp5	Nakhon Si Thammarat, Thailand	KUHE 19428	AB933290	AB933290	KX554643
Kurixalus sp6	Mainland China	YGH 1406033	MK348044	MK348047	MK348052
Kurixalus sp6	Mainland China	YGH 1406034	MK348045	MK348048	MK348053
Kurixalus sp6	Mainland China	YGH 140144	MK348046	MK348049	MK348054