All specimens were examined and stored in the collection of the College of Fisheries, Henan Normal University, Xinxiang, Henan Province, China. For morphological analysis, 10 specimens were fixed in 10% formalin. All measurements and counts were made following Kottelat (1990). Measurements were made point to point with digital calipers to the nearest 0.1 mm. X-ray films were used to count simple fin rays. Comparative morphometry was examined using PCA in IBM SPSS Statistics 22.0. The PCA analysis was processed using log-transformed morphometric data of percent of Standard Length (SL) in the variance-covariance matrix. To compare molecular characters, eighteen specimens were fixed in 95% ethanol and total genomic DNA was extracted from muscle tissue using standard phenol-chloroform extraction protocols (Sambrook et al. 1989). Two pairs of primers were used to amplify segments of COI and Cyt b by the polymerase chain reaction. The COI primers were designed with reference to other Osteichthyes COI gene sequences: Fish-CO I-F (5’-TCT CAA CCA ACC ATA AAG ACA TTGG-3’); Fish-CO I-R (5’-TAT ACT TCT GGG TGC CCA AAG AAT CA-3’). The Cyt b gene primers were designed as GluF (5’-AAC CAC CGT TGT ATT CAA CTA CAA-3’); ThrR (5’-ACC TCC GAT CTT CGG ATT ACA AGA CCG-3’) (Machordom and Doadrio 2001). PCR reactions were carried out in a 30 μl volume and amplifications proceeded for 5 min at 94°C followed by 34 cycles of 95°C for 30 s, 55°C for 30 s, 72°C for 60 s and final extension at 72°C for 10 min. The PCR products were sent to a commercial corporation for sequencing. The nucleotide sequences were assembled from independent sequence passes using the SeqMan (Swindell and Plasterer 1997) module of the DNAStar software package. Additionally, sequences for other species of Homatula were obtained from GenBank (Endruweit et al. 2018, Min et al. 2012, Xiong et al. 2017, Yue et al. 2013). Thirteen species of Homatula were studied, from which 680-bp COI sequences of 50 samples and 1,120-bp Cyt b sequences of 42 samples were used to identify the new species. A list of taxa with corresponding GenBank accession numbers is presented in Table 1. Phylogenetic analysis was performed using PhyloSuite (Zhang et al. 2020). Multiple alignments were performed with MAFFT version 7 with the default parameters (Katoh and Standley 2013). Then, ModelFinder (Kalyaanamoorthy et al. 2017) was used to ascertain the best-fit model of nucleotide substitution for the sequences using the AIC. Tree analyses were performed under Bayesian Inference (BI). All Bayesian phylogenetic analyses were conducted with MrBayes 3.2 (Ronquist et al. 2012), based on the GTR+F+G4 model estimated by ModelFinder. Four Markov chains were run for 1,000,000 generations to estimate the posterior probability distribution, sampling every 1000 generations. After discarding the first 1000 trees as a burn-in with non-stationary log likelihood values, 50% majority-rule consensus trees were estimated for the remaining trees. Finally, all sequences were grouped according to the results of the phylogenetic analysis. The genetic distances between groups were calculated using MEGA7.0 (Kumar et al. 2016), based on the K2P model.

All the sequences in this study were retrieved from GenBank and the accession numbers of the newly determined sequences in this study are MT771689-MT771705 (COI) and MT771706-MT1722 (Cyt b).

Genetic distance and phylogentic trees

Based on the COI gene, sequences of 11 studied species were used to construct a BI tree. The new species belonged to a different clade with strong high PP values (100%, Fig. 4), the genetic distance being analysed by the Kimura 2-parameter model. The distance matrix and number of sequences, determined for each species, are shown in Table 4 and Table 5. Genetic distance between H. guanheensis and other given species displayed 4.7%–10.6% variation in the COI base pairs (Table 4). The smallest distance, 2.9% (5 vs. 9) was detected between H. pycnolepis and H. anguillioides. Based on the Cyt b gene, the phylogenetic tree analysis showed that H. guanheensis is a monophyletic group with a high posterior probability value (100%; Fig. 5). Through the results of the genetic distance analysis, we found that H. guanheensis differed from the acquired congeners by 4.2%–10.3%. The genetic distance between H. guanheensis and similar species exceeded 2% variation, while the intraspecific variation of H. guanheensis was 0% (Table 5).

Figure 4.  

The phylogenetic tree of Homatula based on the COI gene; Schistura latifasciata was used as the outgroup.

Figure 5.  

The phylogenetic tree of Homatula based on the Cyt b gene; Schistura latifasciata was used as the outgroup.

New species were collected from the Guanhe River of the Hanjiang River drainage (the middle and lower reaches of the Yangtze River). In terms of geographic distribution, the known species in the Yangtze River Basin are as follows: Homatula variegata, H. wujiangensis, H. berezowskii and H. potanini. H. variegata and H. potanini are mainly distributed in the upper reaches of the Yangtze River, while H. berezowskii is mainly distributed in the Jialingjiang River and the Hanjiang River (Zhu 1989, Yang et al. 1994, Hu and Zhang 2010). Zeng studied the samples of Homatula in the Yangtze River and the Yellow River and pointed out that there were differences in interorbital width and caudal peduncle length between H. berezowskii specimens of the Hanjiang River and the Jialingjiang River (Zeng et al. 2012). The new species was initially thought to belong to H. berezowskii; however, this hypothesis was quickly abandoned due to the differences in morphology and molecular characters. H. guanheensis is distinguished from H. berezowskii in having bars at least twice as wide as the interspace (vs. equal to the interspace or slightly wider than its interspace); scales partly on anterior body (vs. scaleless); intestine flat (vs. zigzag); a higher caudal peduncle, depth 11.3% (10.1%–13.6%) in SL [vs. 10.0% (9.1%–10.9%)] (Suppl. material 1). Besides this, they have significant divergence in genetic distance (4.2%, Cyt b). The new species can be distinguished from other species in the Yangtze River Basin by the characteristics of the vertical brown bars, body scale distribution, adipose crest, intestine form and dorsal fin positions. H. guanheensis can be clearly distinguished from H. potanini and H. wujiangensis from the lateral line (complete vs. incomplete). It is distinct from H. variegata by having wider bars (vs. narrower), adipose keels not beyond the origin of anal-fin (vs. beyond), head flat, depth 45.8% (42.2%–49.9%) in HL [vs. 49.4% (43.7%–56.8%)] (Suppl. material 1). Besides, the vertical brown bars on the body of the new species are similar to H. laxiclathra. The first two Principal Components of a processed Principal Component Analysis explained 85% of the variance (Fig. 2Table 3). Meanwhile, the new species can be further distinguished from H. laxiclathra in having a flat intestine, not reaching posterior surface of the U-shaped stomach (vs. a loop anteriorly and reaching posterior surface); scales partly on anterior body (vs. scaleless); a wider interorbital width 28.5% (23.9%–31.1%) in HL [vs. 22.5% (19.9%–24.9%)]; a higher caudal peduncle, depth 11.3% (10.1%–13.6%) in SL [vs. 9.2% (8.4%–11.4%)] (Suppl. material 1). Similar to H. longidorsalis, the new species has a slender body with uniform depth, while it differs from H. longidorsalis by having scales on the predorsal body (vs. scaleless on the predorsal body); anterior nostril forming a valve (vs. forming a spool); vertical brown bars at least twice as wide as interspace (vs. bars wider than interspace on anterior portion, unclear on posterior portion); adipose keels not beyond the origin of anal-fin (vs. identical); a shorter head, length 18.7% (17.3%–20.3%) in SL [vs. 23.2% (21.5%–25.3%)]; pelvic-fin close to tip of snout, prepelvic length 46.7% (43.9%–49.33%) in SL [vs. 52.6% (50.5%–56.7%)] (Suppl. material 1).

According to the results of the phylogenetic tree, H. variegata, H. berezowskii, H. longidorsalis and the new species gather into a branch, H. longidorsalis and H. guanheensis + (H. berezowskii + H. variegata) form a sister group (supported by the Cyt b gene, Fig. 5), but the new species and H. longidorsalis + H. variegata form a sister group (supported by the COI gene, Fig. 4). The cause might vary amongst datasets and, thus, this discordance might be influenced by stochastic errors associated with a different number of species and datasets (Naylor and Brown 1998). The results of Cyt b data and the COI data show that the new species has a far genetic distance from H. berezowskii, H. variegata and H. longidorsalis.