Consensus on the taxonomic system and phylogenetic relationships for the anuran genus Fejervarya has yet to be established. Morphological characters in this genus are generally unsuitable for species identification. To carry out molecular species identification and solve phylogenetic problems, we collected 67 Fejervarya specimens from 12 Asian countries and sequenced part of the mitochondrial (mt) Cytb gene. We also sequenced the mt 12S and 16S rRNA genes and seven nuclear genes (BDNF, CXCR4, NCX1, RAG-1, RAG-2, Rhod, and Tyr) for 25 Fejervarya taxa. These molecular markers appear to be adequate for the identification of species. We subjected the molecular data molecular to phylogenetic analyses. In the resulting trees, topotypic F. limnocharis and "F. multistriata" (from China) formed a clade. On the other hand, neither "F. limnocharis" from the Japan mainland nor "F. limnocharis" from eastern Taiwan formed a clade with the real F. limnocharis, and the genetic divergences were larger than the species threshold for frog taxa proposed in previous studies (> 3% for 16S). These results may suggest that "F. multistriata" is a junior synonym of F. limnocharis, or that only some of the populations now recognized as "F. multistriata" correspond to F. limnocharis. Our results also suggest that several cryptic species may be included among the widely distributed Fejervarya species. Finally, our datasets support paraphyly for the genus Fejervarya, although alternative phylogenetic topologies, including Fejervarya monophyly, were not rejected by KH and SH tests.

The rice frog (Fejervarya limnocharis) species complex is widely distributed, from India to Japan, and most prevalently in Southeast Asia. Conspicuous morphological variation has been reported for this species complex throughout its distribution range. In the present study, we used mtDNA gene sequence and allozyme analyses to infer evolutionary affinities within this species complex using eight populations (Sri Lanka; Bangkok and Ranong in Thailand; Taiwan; and Hiroshima, Okinawa, Ishigaki and Iriomote in Japan). We also conducted crossing experiments among four populations from Japan, Thailand, and Sri Lanka in order to find out more about the reproductive isolating mechanisms that might exist among the East, Southeast, and South Asian populations of this species complex. The crossing experiments revealed that the Sri Lanka population is reproductively isolated from the Hiroshima, Bangkok, and Ranong populations by complete hybrid inviability, and that the Bangkok population may be reproductively isolated from the Hiroshima population by partial hybrid inviability. Thus, it is not unreasonable to regard the Sri Lanka population as a species separated from F. limnocharis. The mtDNA and allozyme data showed that the Ranong population is most closely related to the Bangkok population in nuclear genome, but more similar to the Okinawa and Taiwan populations in mtDNA genome. The present, preliminary survey may raise questions about the species status of these particular populations and also about the nature of the biological species concept.

To clarify the genetic divergence in the F. limnocharis complex from Thailand and neighboring countries and to elucidate the phylogenetic problems of this taxon, we analyzed partial sequences of the mitochondrial 12S and 16S rRNA genes and the nuclear CXCR4, NCX1, RAG-1, and tyrosinase genes. The F. limnocharis complex from Thailand had three distinct haplotypes for 12S and 16S rRNA genes. Nucleotide similarities and the phylogenetic relationships indicated that the haplotype 1 group corresponded to the real "F. limnocharis", the haplotype 2 group was F. orissaensis or closely related to it, and the haplotype 3 group was possibly an undescribed species. Mitochondrial gene data also showed two major clades of the genus Fejervarya, the Southeastern and South Asian groups. Although F. orissaensis is so far known only from Orissa in India, the haplotype 2 group was observed in Thailand. This distribution pattern and the phylogeny suggested that the origin of F. orissaensis and the haplotype 2 group might lie in Southeast Asia. There was also evidence suggesting that the haplotype 3 group originated in the South Asian area and has spread to northern Thailand. The nuclear gene data did not support the monophyly of the haplotypes recognized by mitochondrial genes. This incongruence between the mitochondrial and nuclear data seems to be caused by ancestral polymorphic sites contained in nuclear genes. Although neither the mitochondrial nor the nuclear data clarified intergeneric relationships, the nuclear data rejected the monophyly of the genus Fejervarya.