Two new species of Tanichthys (Teleostei: Cypriniformes) from China

Abstract. Tanichthys albiventris, new species, from the River Jiangping in Dongxing City, Guangxi Province is distinguished from Tanichthys albonubes by the presence of a reddish-orange dorsal-fin margin (vs. white) and 9-10 (9 in mode) branched anal-fin rays (vs. 8 in mode). Tanichthys flavianalis, new species, from the River Jiuqu in Qionghai City, Hainan Province is distinguished from T. albiventris and T. albonubes by the presence of a golden anal-fin margin (vs. white) and 7 (rarely 6) branched dorsal-fin rays (vs. 6 in mode). In T. albiventris, T. albonubes, and T. flavianalis the black lateral stripe is located on the dorsal half of the flank, distinguishing them from Tanichthys kuehnei and Tanichthys micagemmae, in which it is mid-lateral. Tanichthys thabacensis is different from all other species of Tanichthys in the shape of the mouth and insertion of the anal fin; it is tentatively referred to as Aphyocypris.

& Šlechtová, 2019, Tanichthys micagemmae Freyhof & Herder, 2001, and Tanichthys thacbaensis Nguyen & Ngô, 2001. Unlike the other three species restricted to specific river basins with a narrow geographic distribution in northern and central Vietnam, T. albonubes has a wide distribution ranging across several different basins (the River Pearl and several small coastal rivers) in southern China (including Guangdong, Guangxi, Hainan and Hong Kong) and northern Vietnam (Quang Ninh Province) (Weitzman & Chan 1966, Pan 1991, Chen 1998, Kottelat 2001, Yi et al. 2004, Chan & Chen 2009, Li & Li 2011, Zhao et al. 2018. Recent molecular studies indicated that each of the wild populations of T. albonubes was monophyletic with significant genetic differentiation among them (Luo et al. 2015, Zhao et al. 2018. They were further considered as different cryptic species (Li et al. 2020). Although several studies made comparisons based on limited morphological characters or colour patterns among some populations from China, and reported variation among different populations (Weitzman & Chan 1966, Yi et al. 2004, Chan & Chen 2009, Li & Li 2011, Li et al. 2020, none of them provided clear diagnostic characters for different populations/ cryptic species. It is apparent that the taxonomy of the T. albonubes species group remains unresolved. Based on detailed examination of specimens of Tanichthys collected from six localities in southern China, and coupled with molecular analysis using the cytochrome b (cytb) gene, we found two new species, and describe them in the present paper (Table 1, Fig. 1).

Sampling
All specimens were collected using hand-nets. Freshly caught fish was euthanized with eugenol. Specimens used for morphological studies were initially fixed in 8% formalin for 5-7 days, then transferred to 70% ethanol for permanent preservation. A few specimens were preserved in 95% ethanol and stored in a -20°C freezer for DNA extraction. The specimens examined in this study are deposited in the Department of Oceanography, National Sun Yat-sen University, Kaohsiung (DOS) and Shanghai Natural History Museum, Shanghai (SNHM). Map of sampling localities. ① for both type locality, JP, and HZ of Tanichthys albiventris due to the short distance; ② for type locality, QH, of T. flavianalis; ③ for CH of T. albonubes; ④ for FS of T. albonubes; ⑤ for GP of T. albonubes. See Table 1 for site information.

Morphological analyses
Measurements and counts generally follow Fang (1997). Head depth was taken at the posterior margin of the orbit. Body depth was taken at the origin of the pelvic fin. The transverse scale series was counted between the dorsal and anal fin origins; both of the small scales below the origin of the dorsal fin and above the origin of the anal fin were counted as 0.5. Pelvic-and pectoral-fin rays were counted on the left side. Only the simple dorsal-and anal-fin rays above the skin were counted. The last two ray elements separated at the same base in both dorsal and anal fins were counted as one.

Molecular analyses
DNA extraction from tissues was conducted using a modified phenol-chloroform protocol (Sambrook & Russell 2001). Mitochondrial cytb of 18 specimens of Tanichthys (Table 1) was amplified by polymerase chain reaction (PCR) using the primer pair, cytbF 5'-TCG ACT AAT CAT AAA GAT ATC GGC AC-3' and cyrbR 5′-TCG ACT AAT CAT AAA GAT ATC GGC AC-3′. PCR was conducted with an initial denaturation at 95 °C for 5 min, followed by 35 cycles of denaturing at 95 °C for 40 s, annealing at 50 °C for 40 s and extension at 72 °C for 1 min. A final extension was done at 72 °C for 8 min. PCR products were purified and sequenced by Sangon Biotech Co., Ltd. (Shanghai).
A total of 1,123 bp consistent cytb sequences were obtained. For tree calculation the 18 newly amplified sequences were analysed with 21 Tanichthys sequences from GenBank (12 T. albonubes (KP893425, KP893448, KP893457,  KP893464, KP893478, KP893490, KP893505,  KP893514, MH918657, MH918658, MH918659 and MH918661), three T. kuehnei (MG952920, MH918664 and MH918663), and six T. micagemmae (HM224384, KX647170, KX647174, KX647179, KX647187 and MH918662)), and Rhodeus ocellatus (AB769519) was used as the outgroup for tree rooting. Cytb sequences were aligned using ClustalW implemented in MEGA 6 (Tamura et al. 2011) and checked by eye. Maximum likelihood tree (ML) and Bayesian inference tree (BI) were reconstructed using MEGA 6 and MrBayes 3.2 (Ronquist & Huelsenbeck 2003), respectively. PartitionFinder 2.1.1 was used to select the optimal evolutionary models for phylogenetic analysis (Lanfear et al. 2017). The best model GTR + I + G was applied for both ML and BI reconstructions. The ML tree was run with 1,000 bootstrap replications (Felsenstein 1985). Bayesian inference was run with all parameters, except topology and branch length, and allowed to vary independently using the unlink command in MrBayes v3.2 (Ronquist & Huelsenbeck 2003). Analyses were conducted with sampling for six million generations when the split frequencies were below 0.01 (two concurrent simultaneous analyses, nruns = 2; three heated chains, nchains = 4; sample frequency 1,000; burnin = 25%). Trees were visualized in FigTree v.1.3.1 (http://tree.bio.ed.ac.uk/software/figtree). Genetic distances based on cytb sequences were calculated using MEGA-X (Kumar et al. 2018, Stecher et al. 2020. Sequences generated in this study are available on GenBank and accession numbers of sequences for molecular analyses are provided in Table 1.  Diagnosis: Tanichthys albiventris is distinguished from T. albonubes by the presence of more branched anal-fin rays (9-10 vs. 8 in mode), and the colour of the dorsal-fin margin (reddish-orange vs. white); from T. kuehnei and T. micagemmae by the presence of a black lateral stripe located on the dorsal half of the flank (vs. on middle of flank), and a reddishorange dorsal-fin margin (vs. white); from T. thacbaensis by the origin of the anal fin, anterior to the base of the last dorsal-fin ray (vs. posterior), and mouth superior (vs. terminal).

Tanichthys albiventris
Description: Morphometric and meristic data of holotype and paratypes as shown in Table 2. Body moderately compressed. Mouth large and oblique; corner of mouth extending to vertical of anterior margin of orbit; lower jaw projecting beyond upper jaw. Barbels absent. Anterior and posterior narial opening confluent, not separated by a skin wall. A row of five to eight small tubercles resembling teeth on lateral margin of upper jaw posterior to premaxilla, and a projecting row of two large conical tubercles laterally on anterior lower jaw present in both adult males and females, but more developed in adult males. Four small sharp tubercles present on the lower margin of the anterior premaxilla in adult males, but not observed in females ( Fig. 2A).
Dorsal fin with 2 simple and 6 branched rays. Anal fin with 3 simple and 9-10 (mostly 9) branched rays. Pectoral-fin rays 12 (rarely 11), first and last one or two rays simple. Pelvic fin rays 7 (rarely 6), first ray simple. Principal caudal-fin rays 17 (rarely 16  Diagnosis: Tanichthys flavianalis is distinguished from T. albonubes by the presence of more branched anal-fin rays (9-10, 9 in mode vs. 8 in mode), more branched dorsal-fin rays (7 in mode vs. 6 in mode), and a reddish-orange dorsal-fin margin (vs. white); from T. albiventris by the presence of more branched dorsal-fin rays (7 in mode vs. 6 in mode), and a golden anal-fin margin (vs. white); from T. kuehnei and from T. micagemmae by the presence of a black lateral stripe located on the dorsal half of the flank (vs. on middle of the flank), and a reddish-orange dorsal-fin margin (vs. white); from T. thacbaensis by the origin of the anal fin anterior to the base of the last dorsal-fin ray (vs. posterior), and mouth superior (vs. terminal).
Description: Morphometric and meristic data of holotype and paratypes as shown in Table 2.
Body moderately compressed. Mouth large and oblique; corner of mouth extending to vertical of anterior margin of orbit; lower jaw projecting beyond upper jaw. Barbels absent. Anterior and posterior narial opening confluent, not separated by a skin wall. A row of six to nine small tubercles resembling teeth on the lateral margin of the upper jaw posterior to premaxilla, a projecting row of two or three large conical tubercles laterally on anterior lower jaw, and four to six small tubercles present on the lower margin of the anterior premaxilla in adult males, but not in females (Fig. 2B).

Colour in preservative:
Ground colour lighter towards belly. A light lateral stripe, along with a slender black stripe, running from posterior margin of operculum to caudal-fin base. Black lateral stripe less than half of width of light lateral   Table 1 for information of population GP. Branched dorsal-fin rays 6 (15) 6 (15) 6 (15) stripe, and more prominent than in life. Dorsal, anal and pelvic fins with blackish stripe near margin. Caudal fin base with a black blotch (Fig. 5).
Distribution and habitat: To date, known only from the upper reach of the River Jiuqu, in Qionghai City, Hainan Province, China (Fig. 1). The type locality at the time of sampling was a small stream (about 1.5 m wide on average) with a mixed mud and gravel substrate. The depth was mostly less than 0.4 m. There was no aquatic vegetation, but a lot of leaf litter. Coexisting fishes were O. pectoralis, Channa cf. gachua, M. hongkongensis, Rhinogobius wanchuangensis and Misgurnus anguillicaudatus.
Etymology: The specific name, flavianalis, is constructed from the Latin words flavus, meaning yellow, and analis, meaning anal, an adjective, alluding to the diagnostic light yellow anal fin with golden margin. A suggested Chinese vernacular is 黄臀唐鱼.

Molecular analyses
BI and ML trees were reconstructed based on 1,123 bps. In both trees, the two new species were monophyletic respectively. The two formed a clade together with the third Chinese species, T. albonubes, while the Vietnamese species, T. kuehnei and T. micagemmae, together formed a second clade. However, the interrelationships among species varied. In the BI tree, T. kuehnei was a monophyletic group nested within T. micagemmae while in the ML tree, reciprocal monophyly of these two species was recovered. In the Chinese clade, T. albiventris was at the basal node in both ML and BI trees. Tanichthys flavianalis was the sister group of T. albonubes in the BI tree, but formed a monophyletic group with the GP population of T. albonubes, which was sister to the remaining T. albonubes in the ML tree (Fig. 6).
In T. albonubes, high genetic distance was found not only among different populations (6.53% between the GP population and the remaining T. albonubes), but also among different individuals of the FS population (up to 4.2% between OK432540 and OK432541), at approximately the inter-specific level compared to other species in China.

Discussion
Based on the data in the present study, the two new species can easily be distinguished from all other congeneric species, excluding T. thacbaensis, by the colour of the dorsal-and anal-fin margins and the number of branched dorsal-and anal-fin rays (Table 3;  . Tanichthys thacbaensis is the only species of the genus for which colour photos are not available. It is unlikely that the opportunity to examine this species again will arise due to both the loss of all type specimens and the destruction of the type locality (Bohlen et al. 2019). However, based on the description and illustration provided in Nguyen & Ngô (2001), this species seems not to be a species of Tanichthys, but a species similar to Aphyocypris lini (Weitzman & Chan 1966) based on the position of the anal-fin origin posterior to the base of last dorsal-fin ray (vs. anterior in all other Tanichthys) and the lower jaw not projecting beyond the upper jaw (vs. projecting in all other Tanichthys).
Tanichthys thacbaensis probably has no coloured stripes across its dorsal and anal fins as found in A. lini (vs. presence in all other Tanichthys), since Nguyen & Ngô (2001) described the colour pattern of T. thacbaensis in detail, including identification of several stripes on the body and a large spot on caudal-fin base, but with no description of stripes on the dorsal and anal fins or shown on the illustration (Bohlen et al. 2019) (Fig. 7).
Colour pattern is an important character for species identification of Tanichthys (Freyhof & Herder 2001, Bohlen et al. 2019. In addition to differences in the colour of the dorsal-and anal-fin margins, the two closely related species of Tanichthys from Vietnam (T. kuehnei and T. micagemmae) can be easily distinguished from the three Tanichthys species from China (T. albiventris, T. albonubes, and T. flavianalis) by the position and width of the black lateral stripe (on the middle of the flank and of similar width as the white stripe along its upper margin vs. on the dorsal half of the flank and less than half width of the white stripe), and the colour of pigmentation below the lateral stripes (lighter than above vs. similar to or darker than above) (Lin 1932, Freyhof & Herder 2001, Bohlen et al. 2019).
The molecular phylogenetic analyses based on cytochrome b showed strong support for the monophyly of the two new species in both the BI and ML trees (Fig. 6), though the ML tree analysis was in disagreement with the BI tree analysis and morphological analysis for the GP population. In the ML tree, the GP population forms a monophyletic group with the T. flavianalis clade, and the GP + T. flavianalis clade is sister to the T. albonubes clade, including the CH and FS populations. The GP population was also considered as a cryptic species in Li et al. (2020). However, the GP population shares similar meristic characters and the abovementioned colour features with the CH and FS populations of T. albonubes (Tables 3-4; Fig. 3), and the BI tree analysis also supported the monophyly of T. albonubes containing GP, CH and FS populations (Fig. 6). Based on morphological and molecular data, the GP population is still identified as T. albonubes in the present study. It is necessary to confirm the status of the GP population by using different morphological and molecular methods in the future.
Previous studies have also provided some morphological data for different populations of Tanichthys from China, but some of the data are distinctly inaccurate based on our examination of a large number of specimens collected from the same localities to those in previous studies. Previous data may have been obtained by an inaccurate counting method. The sampling locality of the lineage "DX" (from Dongxing City, Guangxi Province) in Zhao et al. (2018) and Li et al. (2020), which is the same as the type locality of T. albiventris (coded JP in the present study), was provided by the first author of this study via Dr. Hung-Du Lin who was the corresponding author of Zhao et al. (2018). Li et al. (2020) reported eight branched anal-fin rays for this population based on the examination of 30 specimens, whereas in the present study, the number of branched anal-fin rays of the JP population is 9-10 (mostly 9), and that of the HZ population of T. albiventris from an adjacent basin is 9-11 (mostly 9). All of the 52 specimens from the two localities in the present study have more than eight branched anal-fin rays (Table 3). Furthermore, Li & Li (2011) reported data for 15 specimens from Guiping City, Guangxi Province, including anal fin ii-10, dorsal fin iii-6, pectoral fin ii-9, etc. Li et al. (2020) based on 30 specimens of the same population provided the same number of branched anal-fin rays without any variations (10) as reported in Li & Li (2011). However, the present study examined 31 specimens from the same locality reported in Li & Li (2011) (coded GP in the present study), and found 3 simple and 8-9 (8 in mode) branched analfin rays, 2 simple dorsal-fin rays above skin, and the second pectoral-fin ray branched (Tables 3-4).