One specimen of frog was collected from Mohan Town, Mengla County, Xishuangbanna Prefecture, Yunnan Province, China, which is the type locality of Rana bannanica Rao & Yang, 1997, currently considered as a junior synonym of Hylarana milleti (Smith, 1921). This specimen well agrees with the original description of R. bannanica. In regard to morphology, R. bannanica and H. milleti are, indeed, very similar, but can still be distinguished from each other. Phylogenetic analysis, based on mitochondrial gene sequences, showed that this specimen and H. milleti are different species. Hence, we remove R. bannanica from the synonymy of H. milleti.

16S, COI, cytb, Hylarana milleti, resurrection, synonym, systematics, taxonomy

The systematics and taxonomy of Golden-backed frogs of the genus Hylarana Tschudi, 1838 have long been chaotic (Dubois 1992; Oliver et al. 2015; Chan et al. 2020; Dubois et al. 2021; Reilly et al. 2022). Previously, it had been divided into 10 genera, namely Hylarana, Hydrophylax Fitzinger, 1843, Amnirana Dubois, 1992, Chalcorana Dubois, 1992, Humerana Dubois,1992, Papurana Dubois, 1992, Pulchrana Dubois, 1992, Sylvirana Dubois, 1992, Abavorana Oliver, Prendini, Kraus & Raxworthy, 2015, and Indosylvirana Oliver, Prendini, Kraus & Raxworthy, 2015 (Oliver et al. 2015; Chan et al. 2020). Later, except for Abavorana, the other nine genera were placed back into Hylarana, relegating those ranks as subgenera (Dubois et al. 2021; Reilly et al. 2022). At present, the website Amphibian Species of the World (https://amphibiansoftheworld.amnh.org/) also adopts this viewpoint (Frost 2023).

Rana bannanica Rao & Yang, 1997 is a poorly-known ranid species described in 1997 from Xishuangbanna Prefecture, Yunnan Province, China and its type locality is in Mohan Town, Mengla County, Xishuangbanna Prefecture, Yunnan Province, China, close to the border with Laos (Rao and Yang 1997). Since this species was described, there have been no formal collection reports of this species from its type locality.

Rana milleti Smith, 1921 was described from the Langbian Plateau, Lam Dong Province, central-southern Vietnam and, subsequently, this species was reported from Thailand and Cambodia (Chuaynkern et al. 2004; Stuart and Emmett 2006). Thereafter, Ohler (2007) proposed R. bannanica as a junior synonym of R. milleti by morphological data, based on several specimens collected from Phongsaly, Laos, as she considered that these specimens resemble both R. bannanica and R. milleti.

Currently, Rana milleti is regarded as Hylarana milleti and considered to be distributed in Vietnam, Thailand, Cambodia, Laos, southern China and peninsular Myanmar and R. bannanica is considered to be a synonym of H. milleti (Ohler 2007; Chan et al. 2020; Poyarkov et al. 2021; Zug 2022).

During our field surveys in Yunnan Province, China and in southern Vietnam from 2019 to 2023, we collected one frog specimen from the type locality of Rana bannanica, Xishuangbanna and six specimens of Hylarana milleti from its type locality, Langbian Plateau, Lam Dong Province (Fig. 1). We collected molecular sequence data of these specimens in order to reconstruct a molecular phylogeny and evaluate the phylogenetic position of them. Our phylogenetic analysis of molecular data demonstrated that the specimen from Xishuangbanna and the specimens from Langbian Plateau are not the same species. Thus, we resurrect R. bannanica and re-assess its taxonomic status.

Figure 1. 

Map showing the type locality (black star) of Hylarana bannanica in Xishaungbanna, Yunnan, China and the type locality (black square) of H. milleti in Langbian Plateau, Lam Dong, Vietnam and other confirmed distributions (black dots) of H. bannanica in Phongsaly, Laos, Chanthaburi, Thailand and Tanintharyi, Myanmar, respectively, based on molecular data.

The specimen from China was collected from Manzhuang Village, Mohan Town, Mengla County, Xishuangbanna Prefecture, Yunnan Province and the specimens from Vietnam were collected from Da Lat and Di Linh, Lam Dong Province. Liver or muscular tissue samples were preserved in 99% ethanol for molecular analysis. The specimen from China was deposited in Kunming Institute of Zoology, Chinese Academy of Sciences (KIZ) and the specimens from Vietnam were deposited in Zoological Museum of Moscow University (ZMMU).

Measurements were taken with a digital calliper to the nearest 0.1 mm. We measured: snout-vent length (SVL), head length (HL), head width (HW), snout length (SL), eye diameter (ED), interorbital distance (IND), tympanum diameter (TD), forearm and hand length (FHL), total arm length (ARM), tibia length (TL), tarsus and foot length (TFL), foot length (FL) and total leg length (LEG).

Total genomic DNA was extracted from liver or muscular tissues. A fragment of the 16S rRNA (16S) gene was amplified and sequenced using the primers L2188: 5’–AAAGTGGGCCTAAAAGCAGCCA–3’ and 16H1: 5’–CTCCGGTCTGAACTCAGATCACGTAGG–3’ (Hedges 1994; Matsui et al. 2006), a fragment of the cytochrome c oxidase subunit I (COI) gene was amplified and sequenced using the primers Chmf4: 5’–TYTCWACWAAYCAYAAAGAYATCGG–3’ and Chmr4: 5’–ACYTCRGGRTGRCCRAARAATCA–3’ (Che et al. 2012) and a fragment of the cytochrome b (cytb) gene was amplified and sequenced using the primers cytb-c: 5’–CTACTGGTTGTCCTCCGATTCATGT–3’ and CB-J-10933: 5’–TATGTTCTACCATGAGGACAAATATC–3’ (Bossuyt and Milinkovitch 2000). All new sequences have been deposited in GenBank. Other sequences were obtained from GenBank (Table 1).

Sequences were aligned using MAFFT 7.471 (Katoh and Standley 2013) with default parameters. Phylogenetic analyses were constructed, based on the concatenated sequences of 16S, COI and cytb genes. The best fit models were selected using the corrected Akaike Information Criterion Akaike Information Criterion (AICc) in ModelFinder which was implemented in IQ-TREE 1.6.12 (Kalyaanamoorthy et al. 2017). Bayesian Inference (BI) was performed in MrBayes 3.2.7 (Ronquist et al. 2012), based on the GTR+F+I+G4 substitution model for all partitions and the Markov chains were run for 5,000,000 generations and sampled every 100 generations. Maximum Likelihood (ML) analysis was performed in IQ-TREE 1.6.12 (Nguyen et al. 2015), based on the GTR+F+R4 substitution model for 16S, the TIM2+F+I+G4 substitution model for COI and the TVM+F+I+G4 substitution model for cytb and branch support was assessed using 1,000 ultrafast bootstrap replicates.

Morphologically, the specimen from the type locality of Rana bannanica well agrees with the original description of R. bannanica in that all measurements and the relative lengths of body parts of this specimen are within the ranges of those in the type series of R. bannanica (n = 8) (Table 2). Therefore, we consider this specimen to belong to the species described by Rao and Yang (1997). However, the ratio of head width to head length in males of R. bannanica (n = 9) is relatively smaller than that in males of H. milleti (n = 5) (HW/HL 0.31–0.35 vs. 0.35–0.36), the ratio of snout length to snout-vent length in males of R. bannanica is relatively larger than that in males of H. milleti (SL/SVL 0.17–0.18 vs. 0.15–0.17) and the ratio of total leg length to snout-vent length in males of R. bannanica is relatively smaller than that in males of H. milleti (LEG/SVL 1.53–1.69 vs. 1.70–1.78) (Table 2). In addition, R. bannanica differs from H. milleti by having relatively larger discs on tips of fingers and toes, having relatively larger outer metatarsal tubercle, having tarsal folds and having vocal sacs in males.

In phylogenetic analyses of mtDNA sequences, BI and ML methods yielded similar results. The newly-generated sequences of Hylarana milleti, as well as the sequences of the specimens of H. milleti from Lam Dong and Gia Lai provinces in Vietnam and the newly-generated sequence of the specimen from the type locality of Rana bannanica, as well as the sequences of the specimens previously identified under the name H. milleti from Phongsaly Province in Laos, Chanthaburi Province in Thailand and Tanintharyi Region in Myanmar, formed two distant, separate clades within the genus Hylarana (Fig. 2). Clearly, the specimens previously considered to be H. milleti from China, Laos, Thailand and Myanmar and the specimens of H. milleti from Vietnam belong to different species, although their phylogenetic positions within the genus have not been resolved. Therefore, we formally remove Rana bannanica Rao & Yang, 1997 from the synonym of Hylarana milleti Smith, 1921.

Figure 2. 

Bayesian phylogenetic tree of Hylarana inferred from the concatenated 16S, COI and cytb sequences. Numbers before slashes indicate Bayesian posterior probabilities (≥ 0.90) and numbers after slashes indicate ML bootstrap supports (≥ 90).

Hylarana bannanica (Rao & Yang, 1997)

Figs 3, 4, 5A

Rana bannanica Rao & Yang, 1997.

Type material

Holotype. KIZ 94001, adult male. Paratypes. KIZ 94002–KIZ 94008, seven adult males.

Type locality

Mohan Town, Mengla County, Xishuangbanna Prefecture, Yunnan Province, China.

Specimen examined

KIZ2019BN, adult male, collected by Shuo Liu from Manzhuang Village, Mohan Town, Mengla County, Xishuangbanna Prefecture, Yunnan Province, China (21°21'51"N, 101°37'17"E, 760 m elevation), on 7 May 2019.

Figure 3. 

The holotype (KIZ 9401) of Hylarana bannanica in preservative. A. Dorsal view; B. Ventral view; C. Close-up view of the right side of the head; D Close-up view of the left side of the head. Photos by Dingqi Rao.

Description of the newly-collected topotypic specimen

SVL 40.4 mm; head width less than head length (HW/HL 0.83); snout pointed, projecting beyond lower jaw; nostril slightly closer to snout than to eye; canthus rostralis distinct; loreal region vertical, slightly concave; snout length greater than eye diameter; pineal body indistinct; tympanum relatively large (TD/ED 0.82), rounded; vomerine teeth distinct; choanae close to vomerine teeth; tongue oval, posterior notch U-shaped; vocal sac openings small, on floor of each corner of mouth; vocal sacs present.

Figure 4. 

The topotypic specimen (KIZ2019BN) of Hylarana bannanica in preservative. A Dorsal view; B Ventral view; C Close-up view of the right side of the head; D Close-up view of the left side of the head; E Close-up view of the left palm; F Close-up view of the left sole. Photos by Shuo Liu.

Forelimbs relatively slender; relative lengths of fingers III > I ≥ IV > II; all finger tips expanded into small discs; webbing on fingers absent; subarticular tubercles distinct, oval and prominent; supernumerary tubercles below base of fingers II–IV present, small and rounded; thenar tubercle oval, large; median metacarpal tubercle oval, prominent; outer metacarpal tubercle elongated, indistinct; grey nuptial pad on finger I indistinct.

Hind limbs medium long; relative lengths of toes IV > V > III > II > I; all toe tips expanded into small discs, discs on toes slightly larger than on fingers; webbing on toes undeveloped; subarticular tubercles distinct, oval, prominent; inner metatarsal tubercle oval; outer metatarsal tubercle rounded; tarsal fold present.

Dorsal skin slightly rough, upper lateral skin granular, lower lateral and ventral skin smooth; dorsolateral folds distinct, supratympanic fold absent; one white gland above base of forelimb behind tympanum on each side; some tiny tubercles form indistinct skin ridges on dorsal tibia.

In preservative, dorsum light brown, lateral side greyish-brown, ventral side of head and limbs yellowish-white, belly greyish-white, iris black, pupil white.

Distribution

Currently, Hylarana bannanica is known to be distributed in Xishuangbanna Prefecture, Yunnan Province, China, as well as northern Laos (Phongsaly Province), Thailand (Chanthaburi, Ubon Ratchathani, Amnat Charoen, Sa Kaeo, Prachin Buri, Nakhon Ratchasima and Chachoengsao provinces), southwest Cambodia (Cardamon Mountains) and peninsular Myanmar (Taninthary Region), based on data from Rao and Yang (1997), Chuaynkern et al. (2004), Stuart and Emmett (2006), Yang and Rao (2008), Poyarkov et al. (2021), Zug and Mulcahy (2020), Zug (2022) and this study.

Conservation status

Further research is required to clarify the extent of the distribution, population size, trends and conservation status of the species. Hylarana bannanica is distributed over a large area including many protected areas. Across its range, the species seems not to be rare. Thus, we tentatively suggest Hylarana bannanica be considered a Least Concern (LC) species, following the IUCN’s Red List Categories (IUCN Standards and Petitions Committee 2019).

Dubois (1992) partitioned the genus Rana into many sections, one of which is the section Hylarana, which contains two subsections, namely subsection Hydrophylax and subsection Hylarana. He treated Hydrophylax as a subgenus and put it together with the newly-erected subgenera Amnirana, Humerana, Papurana, Pulchrana and Sylvirana in the subsection Hydrophylax and treated Clinotarsus, Glandirana, Hylarana, Odorrana and Pterorana as subgenera and put them together with the newly-erected subgenera Chalcorana, Eburana, Nasirana, Sanguirana and Tylerana in the subsection Hylarana and explicitly stated that these arrangements were preliminary hypotheses (Dubois 1992). Oliver et al. (2015) revised Hylarana sensu lato by utilising an inclusive molecular phylogeny, elevated Amnirana, Chalcorana, Hydrophylax, Hylarana, Humerana, Papurana, Pulchrana and Sylvirana to genus rank and erected two new genera, namely Abavorana and Indosylvirana. Chan et al. (2020) performed a comprehensive analysis of Hylarana sensu lato and also recognised the ten genera Abavorana, Amnirana, Chalcorana, Hydrophylax, Hylarana, Humerana, Indosylvirana, Papurana, Pulchrana and Sylvirana in it. Later, Dubois et al. (2021) found Chalcorana to be paraphyletic with Pulchrana and Amnirana and Sylvirana, Papurana, Hydrophylax, Hylarana and Indosylvirana to be polyphyletic; therefore, they placed all of the genera within Hylarana sensu lato, except for Abavorana, into the synonymy of Hylarana. However, we consider that these synonymous genera may be valid and need to be resurrected, at least for some of them. For example, Dubois et al. (2021) recovered Sylvirana to be polyphyletic due to Sylvirana guentheri as the sister taxon of Humerana miopus. In our phylogenetic analysis, Hylarana guentheri was also nested in the subgenus Humerana with strong support. Nevertheless, we cannot say directly that Sylvirana or Humerana is invalid, unless Hylarana guentheri is the type species of Sylvirana. Hylarana guentheri is a very common species in southern China and northern-central Vietnam and we have collected specimens of this species multiple times. This species has a large body size, which is different from most species of the subgenus Sylvirana and consistent with most species of the subgenus Humerana. Since Hylarana guentheri is not the type species of Sylvirana, therefore, we consider that Hylarana guentheri should be moved to Humerana from Sylvirana, and Sylvirana and Humerana are still valid, instead of synonymising them. Anyway, we still adopt the current arrangement (Dubois et al. 2021; Frost 2023) of Hylarana sensu lato for the time being; further morphological and molecular studies are required to elucidate the division of genera within Hylarana sensu lato.

Hylarana bannanica was described from Xishuangbanna, Yunnan, China, in 1997 and no new specimens from its type locality have been reported since then. The specimen (KIZ2019BN) we collected is the first rediscovery from the type locality of this species since it was described more than 20 years ago. Unfortunately, this specimen was misidentified as a subadult of the local common species H. cubitalis (Smith, 1917) and was not taken seriously when it was collected, so that the habitat information at the collection site and photos of this specimen in life were not recorded. In the subsequent process of specimen sorting, we discovered that this specimen is an adult male due to the presence of nuptial pad and vocal sacs, but without a large oval gland on the inner side of each elbow, which shows it is obviously not H. cubitalis, so we conducted a detailed study on this specimen. This experience reminds us that it is necessary to have sufficient taxonomic competence when doing fieldwork and that those who visit different locations should be aware of what species are distributed or what species may be distributed there.

In previous studies (e.g. Frost et al. (2006); Oliver et al. (2015); Chan et al. (2020); Dubois et al. (2021); Reilly et al. (2022)), Hylarana milleti (Fig. 5B) was once placed in the genera or subgenera Sylvirana, Indosylvirana and Papurana, respectively. In our phylogenetic analysis, it is evident that H. milleti and H. bannanica are not conspecific, but their phylogenetic positions within the genus have not been resolved. At present, we cannot determine which subgenus these two species belong to respectively. The systematic relationship within this genus needs to be further resolved.

Figure 5. 

Comparison between Hylarana bannanica (A from Chanthaburi, Thailand) and H. milleti (B from Bidoup-Nui Ba, Lam Dong, Vietnam) in life. Photos by Parinya Pawangkhanant (A) and Nikolay A. Poyarkov (B).

We thank Parinya Pawangkhanant (AUP) for providing the photo of Hylarana bannanica from Thailand. We thank George Zug (USNM) for his help in this study. Thanks to the forest rangers of Xishuangbanna National Nature Reserve for their help in the field. We also thank the editors and reviewers for their working on the manuscript. This work was supported by the Science-Technology Basic Condition Platform from the Ministry of Science and Technology of the People’s Republic of China (Grant No. 2005DKA21402), the project of the second comprehensive scientific investigation of Xishuangbanna National Nature Reserve, the project of Ministry of Ecology and Environment of China: investigation and assessment of amphibians and reptiles in Jinghong City, Menghai County and Mengla County, in part the Russian Science Foundation to N.A. Poyarkov (Grant No. 22-14-00037, molecular analyses) and in part the Rufford Foundation to T.V. Nguyen (Grant No. 39897-1; data analysis).