Formosania tangi, a new species of suck-loach (Cypriniformes, Gastromyzontidae) from the Jiulongjiang River, southeastern China, with taxonomic notes on F. fascicauda

Abstract

For decades, populations of the suck-loach genus Formosania from the Jiulongjiang River in Fujian Province, China, have been identified as F. fascicauda, a species originally described from in a separate coastal drainage basin in Fuqing County. Employing an integrative taxonomic approach combining morphological and molecular phylogenetic analyses, we demonstrate that the Jiulongjiang River population represents a distinct species, formally described here as Formosania tangisp. nov. It is diagnosed by a combination of characters: 13 rostral barbels arranged in two rows, a distinct straight, dark, longitudinal stripe along the lateral line, and an emarginate caudal fin. Phylogenetic analyses consistently recover it as a unique evolutionary lineage showing significant genetic divergence from true F. fascicauda, with a Kimura 2-parameter distance of 4.16%. The description of F. tangi clarifies the taxonomic status of the Jiulongjiang River population and enhances our understanding of species diversity and biogeography of Formosania in the coastal drainages of southeastern China.

Introduction

The suck-loach genus Crossostoma was established by Sauvage (1878) based on specimens from the Wuyi Mountain area in Fujian Province, with C. davidi designated as the type species. Oshima (1919) independently described the genus Formosania from Taiwan, with F. gilberti as its type species. Chen (1980) later synonymized Formosania under Crossostoma, citing the absence of clear morphological distinctions. However, Novák et al. (2006) reinstated Formosania as the valid generic name after determining that Crossostoma was preoccupied by an earlier gastropod taxon. The genus currently comprises 11 recognized species (Fricke et al. 2026).

Formosania fascicauda (Nichols, 1926) was described from Fuqing County, Fujian Province, but the original description (Nichols 1926) omitted the precise collection locality within the county, which is drained by both the Longjiang (龙江) and Yuxi rivers (渔溪). Fang (1934) examined a cotype and noted the diagnostic features distinguishing it from F. lacustris (Steindachner, 1908) and F. stigmata (Nichols, 1926). Later, Chen (1980) identified specimens from the Jiulongjiang (九龙江) and Jinjiang rivers (晋江) as F. fascicauda but did not compare these specimens to topotypic material from Fuqing. Consequently, subsequent researchers have consistently assigned the Jiulongjiang River population to F. fascicauda without verification (Zhang and Zhao 2016; Sun et al. 2023).

To evaluate the taxonomic status of the Jiulongjiang River population and its relationship to F. fascicauda, we first ascertained the precise type locality of F. fascicauda. The key holotype data and photographs for this study were provided by the Department of Ichthyology, American Museum of Natural History. Comprehensive specimens were then collected from multiple coastal drainages across Fujian Province, including the Jiulongjiang River system. Integrated morphological comparisons and molecular phylogenetic analyses demonstrate that the Jiulongjiang River population is morphologically and genetically distinct from true F. fascicauda, supporting its formal recognition as a new species, which we describe herein.

Materials and methods

Specimen collection, processing, and measurement

Details of specimen collection are provided in Table 1 and Fig. 1. Fish were collected using hand nets. Following capture, specimens were immediately euthanized by immersion in a solution of methanesulfonate (MS-222). Those designated for molecular analysis were fixed directly in 95% ethanol. Specimens intended for morphological study were initially fixed in a 10% formalin solution for approximately 48 hours before being transferred to 65% ethanol for long-term preservation. All collected specimens are deposited in the Fish Museum of Shanghai Ocean University (SHOU). Comparative materials are provided at the end of this article.

Table 1.

The samples used in molecular analysis with their localities, voucher information and GenBank accession.

Species	Specimen voucher	Sampling localities	River system	GenBank accession	Source
F. tangi sp. nov.	SHOU20240801010-14	Datian County, Fujian	Jiulong-jiang	PX146867–PX146871	This study
F. fascicauda	SHOU20241122001-04	Fuqing City, Fujian	Long-jiang	PX146873–PX146876	This study
	SHOU20241019041	Fuqing City, Fujian	Yu-xi	PX146872	This study
F. davidi	SHOU202106251-53, 62-63	Qingyuan County, Zhejiang	Min-jiang	OQ605818–OQ605822	Sun et al. 2023
F. fasciolata	SHOU202107001-05	Taishun County, Zhejiang	Feiyun-jiang	OQ605808–OQ605812	Sun et al. 2023
F. galericula	SHOU202106273, 75-77, 93	Qingyuan County, Zhejiang	Ou-jiang	OQ605803–OQ605807	Sun et al. 2023
F. immaculata	SHOU202106312-16	Wuyi County, Zhejiang	Ou-jiang	OQ605813–OQ605817	Sun et al. 2023
F. stigmata	SHOU202201013, 19, 27	Yanping County, Fujian	Min-jiang	OQ605800–OQ605802	Sun et al. 2023
F. paucisquama	SHOU202110011	Puning County, Guangdong	Lian-jiang	OQ605798	Sun et al. 2023
	SHOU202110028	Jiexi County, Guangdong	Rong-jiang	OQ605799	Sun et al. 2023
F. tinkhami	SHOU202110086	Longmen County, Guangdong	Zhu-jiang	OQ605795	Sun et al. 2023
Vanmanenia stenosoma	–	Shaoxing City, Zhejiang	Qiantang-jiang	MZ853160	GenBank
F. chenyiyui	SHOU20150001	Changting County, Fujian	Han-jiang	OQ605797	Sun et al. 2023
F. lacustris	–	Taiwan	–	AY392454–AY392461, AY392463–AY392467	GenBank
	–	Taiwan	–	KX056126	GenBank
	–	Taiwan	–	NC001727	GenBank

Figure 1.

Collection sites of Formosania species for morphological comparison (data for F. lacustris and F. gilberti from Yeh et al. 2024).

Morphometric measurements followed Sun et al. (2025) and were obtained point-to-point with digital callipers (0.01 mm) for all mainland Chinese Formosania species with two rostral barbel rows and 13 rostral barbels, including F. fascicauda, F. stigmata, F. paucisquama (Zheng, 1981), F. tinkhami (Herre, 1934), and the Jiulong River population; all values are presented as percentages of standard or head lengths. Meristics were counted under a dissecting microscope, primarily on the left side. Abbreviations used in the text: standard length (SL), from tip of the snout to the last half-centrum; total length (TL), from tip of the snout to the most posterior of the caudal fin; head length (HL), from tip of the snout to the most posterior point of the operculum. For F. lacustris and F. gilberti, only meristic data were cited from Yeh et al. (2024). The meristic data from this source are recalculated and verified by us, superseding the originally published values.

DNA extraction, PCR amplification, and sequencing

To clarify the genetic relationship between the Jiulongjiang River population and the topotypic F. fascicauda population from Fuqing, 10 individuals from the Jiulongjiang, Longjiang, and Yuxi rivers were selected for amplification and sequencing of the mitochondrial cytochrome b (Cytb) gene. Genomic DNA was extracted from clips of the right pectoral fin from ethanol-preserved specimens. Samples were systematically labelled according to their collection dates.

The mitochondrial cytochrome b gene was amplified via polymerase chain reaction (PCR). The 25 μL PCR reaction mixture consisted of 9.5 μL ddH₂O, 1 μL of each primer (10 μM), 1 μL template DNA, and 12.5 μL of 2× Taq PCR Master Mix (Sangon Biotech Co., Ltd, Shanghai, China). The thermocycling protocol was as follows: initial denaturation at 95 °C for 3 min; 35 cycles of denaturation at 94 °C for 30 s, annealing at 54 °C for 45 s, and extension at 72 °C for 1 min. The primers used for both amplification and sequencing were QcytbL (5'-GACTTGAAGAACCACCGTTGTTATT-3') and QcytbH (5'-TCTTCGGATTACAAGACCGAT-GCTTT-3') (Chen et al. 2024). PCR products were purified and sequenced commercially by Sangon Biotech (Shanghai, China). Sequence chromatograms were assembled and edited using SeqMan module of DNASTAR Lasergene (Burland 2000). The newly generated sequences have been deposited in the GenBank database, with accession numbers provided in Table 1.

Phylogenetic reconstruction

A dataset for phylogenetic analysis was assembled, comprising the Cytb sequences generated in this study along with homologous sequences for other Formosania species downloaded from the NCBI database. No Cytb sequence is publicly available under the name F. gilberti, which has been historically treated as a junior synonym of F. lacustris – the name under which all previous genetic data are deposited. Although Yeh et al. (2024) revised F. gilberti as a valid species, the specific sequence data are yet not available in the NCBI database. Given that their work confirms its close relationship to F. lacustris, we conclude that its exclusion does not compromise our phylogenetic results, which focus on the distinct species of mainland China. A sequence of Vanmanenia stenosoma (Boulenger, 1901) (GenBank accession MZ853160) was downloaded and included as the outgroup. The final aligned dataset contained 53 sequences.

Phylogenetic analyses were conducted using PHYLOSUITE (Zhang et al. 2020). Sequence alignment was performed with MAFFT (Katoh and Standley 2013) under the automatic strategy and normal alignment mode. The best-fit nucleotide substitution models for maximum-likelihood (ML) and Bayesian-inference (BI) analyses were selected using ModelFinder (Kalyaanamoorthy et al. 2017) based on the Bayesian Information Criterion (BIC). The ML tree was constructed with IQ-TREE (Nguyen et al. 2015) using the TIM2+F+I+G4 model, with all other parameters set to default. The BI tree was constructed using MrBayes v. 3.2.6 (Ronquist et al. 2012) with the GTR+F+I+G4 model, running two parallel Markov Chain Monte Carlo (MCMC) analyses for 1 million generations, sampling every 100 generations. The MCMC analysis was considered to have converged, as the average standard deviation of split frequencies dropped below 0.01 (final value = 0.003854). The first 25% of trees were discarded as burn-in. Resulting phylogenetic trees were visualized and annotated using FigTree v. 1.4.4 (Rambaut 2018). Genetic distances were calculated using MEGA11 (Tamura et al. 2021) with the Kimura 2-parameter model (rates among sites: G4), computing average interspecific genetic distances.

Results

Taxonomic account

Family Gastromyzontidae Hora, 1950

Genus Formosania Oshima, 1919

Formosania tangi sp. nov.

6E9A8248-8666-5E86-A484-4A95CB0E4CD2

https://zoobank.org/39BFBF89-6178-4A97-AECB-72345DC20A10

Figs 2, 3B, 4, 7D, 8A

Figure 2.

Lateral, dorsal, and ventral views of Formosania tangi sp. nov., holotype, adult, SHOU20251010601, 63.74 mm SL.

Figure 3.

Ventral view of the head, showing the arrangement of rostral barbels. A. One-row type in Formosania davidi (SHOU20230717003, from Minjiang River, Wuyishan City, Fujian Province); B. Two-row type in F. tangi sp. nov. (SHOU20251010601, holotype); C. F. chenyiyui (SHOU20250402004, from Tingjiang River, Changting County, Fujian Province).

Figure 4.

Habitat (photographed by Yong-Sheng Lin) and live appearance of Formosania tangi sp. nov. (photographed by Hao-Jun Chen).

Figure 7.

Caudal fin morphotypes in the genus Formosania. A. Plain subtruncate (F. davidi); B. Patterned subtruncate (F. fascicauda); C, D. Patterned emarginate (C. F. stigmata; D. F. tangi sp. nov.).

Figure 8.

Pattern differences between Formosania tangi sp. nov. (A) and F. fascicauda (B), caudal fin not fully expanded.

• Crossostoma fascicauda : Chen 1980: 104 (partim: Jiulongjiang River in Longyan, Fujian).

• Crossostoma fascicauda : Chen and Tang in Le 2000: 438–567 (partim: Jiulongjiang River in Longyan, Fujian).

• Formosania fascicauda : Sun et al. 2023: 207–221. (JiulongJiang River in Nanjing, Fujian).

Type material.

Holotype • SHOU20251010601, TL 79.13 mm, SL 63.74 mm; China, Fujian Province, Sanming City, Datian County, Taoyuan Town, Jiulongjiang River; 25.843°N, 117.577°E, elevation 801 m; Yong-Sheng Lin leg.; 10 Oct. 2025 (Fig. 2). Paratypes • SHOU20251010602, SL 56.76 mm; collected from the type locality; Yong-Sheng Lin leg.; 10 Oct. 2025. • SHOU20240704603 to -605, 4 specimens, SL 56.76–73.69 mm; collected from the type locality; 4 July 2024. • SHOU202201083 to SHOU202201093, 11 specimens, SL 50.53–81.06 mm; China, Fujian Province, Zhangzhou City, Nanjing County, Jiulongjiang River; 24.630°N, 117.083°E, elevation 766 m; Yang Chen & Jia-Jun Zhou leg.; January 2022.

Diagnosis.

The new species can be distinguished from congeners by combination of following characters: rostral barbels 13, well developed, arranged in two rows (Fig. 3B) (vs one row in F. davidi, F. fasciolata, F. galericula, and F. immaculata – Fig. 3A; 12–15 rudimentary barbels in F. chenyiyui – Fig. 3C); lateral line accompanied by broad, dark, longitudinal band, margins weakly undulating or nearly straight, pale stripe between lateral line band and dark dorsal surfaces generally continuous straight (vs a thin, black, longitudinal line with alternating patches, pale stripe intermittent wavy or absent in F. fascicauda, F. stigmata, and F. paucisquama; lateral line with alternating small dark dots, flanks with interrupted, short, dark streaks or small patches along dorsal and ventral margins in F. tinkhami; lateral line with alternating small, dark dots, flanks vermicular in F. gilberti; flanks unblotched, with lateral streak in F. lacustris); fully expanded caudal-fin emarginate (Fig. 7D) (vs subtruncate in F. fascicauda (Fig. 7B), F. davidi, F. fasciolata, F. galericula, and F. immaculata (Fig. 7A)).

Description.

Dorsal iii-8, anal ii-5, pectoral i-14-15, pelvic i-8. Lateral-line canal pores and scales 72–88. Morphometric measurements are given in Table 2. See Fig. 2 for lateral, dorsal, and ventral views of the body.

Table 2.

Statistical analysis of morphometric data for two-row barbel Formosania species.

Characters	F. tangi sp. nov. (N = 16)			F. fascicauda (N = 16)		F. stigmata (N = 14)		F. paucisquama (N = 19)		F. tinkhami (N = 6)
	Holotype	Range	Mean+SD	Range	Mean+SD	Range	Mean+SD	Range	Mean+SD	Range	Mean+SD
Standard length (mm)	63.74	50.53–81.06	62.53 ± 9.16	53.98–83.30	64.08 ± 9.46	49.95–85.47	62.05 ± 11.07	49.04–65.76	56.31 ± 5.33	51.98–80.29	66.63 ± 11.38
% of standard length (SL)
Body depth (BD)	16.36	13.23–17.32	15.92 ± 1.11b	15.77–19.24	17.01 ± 0.78ab	16.10–21.93	18.28 ± 1.89a	14.14–19.40	17.10 ± 1.57ab	14.93–16.69	15.99 ± 0.67b
Head length (HL)	27.09	19.95–27.09	22.05 ± 1.95b	21.37–25.08	23.60 ± 1.05a	23.59–26.66	24.72 ± 0.88a	22.65–27.37	24.67 ± 1.39ab	23.26–24.11	23.59 ± 0.33ab
Dorsal head length (DHL)	22.26	20.34–24.06	22.64 ± 1.00b	21.63–24.07	22.61 ± 0.62b	22.32–26.24	24.08 ± 0.96a	21.36–26.41	24.53 ± 1.33a	22.70–24.72	23.69 ± 0.86ab
Head depth (HD)	11.31	10.50–13.29	11.71 ± 0.74	10.87–12.79	11.73 ± 0.52	10.65–13.51	11.78 ± 0.72	10.89–13.69	11.95 ± 0.67	11.31–12.31	11.78 ± 0.36
Head width (HW)	18.73	17.64–20.45	18.89 ± 0.77b	17.28–20.35	18.61 ± 0.80b	18.01–21.16	19.33 ± 1.07ab	18.00–22.69	20.24 ± 1.29a	18.39–20.12	19.30 ± 0.77ab
Caudal-peduncle length (CPL)	15.03	7.87–15.03	11.07 ± 2.19b	10.70–15.07	12.76 ± 0.99a	10.42–15.01	12.81 ± 1.35a	11.30–14.73	12.82 ± 1.02a	10.67–12.14	11.42 ± 0.60ab
Caudal-peduncle depth (CPD)	10.68	9.25–11.47	10.28 ± 0.59b	10.33–12.62	11.33 ± 0.70a	9.74–11.99	10.54 ± 0.65b	9.23–12.07	10.78 ± 0.86ab	9.45–10.96	10.15 ± 0.52b
Dorsal-fin length (DFL)	22.48	20.28–24.36	22.32 ± 1.05b	19.96–23.28	21.28 ± 0.99b	20.76–24.96	22.44 ± 1.38b	21.06–26.94	24.27 ± 1.58a	22.12–24.68	23.63 ± 0.90ab
Pectoral-fin length (PFL)	24.24	21.77–26.42	24.03 ± 1.15b	22.98–26.70	24.76 ± 1.16b	21.46–27.40	25.47 ± 1.67b	22.87–29.74	27.25 ± 1.74a	24.18–26.50	25.39 ± 1.10ab
Pelvic-fin length (VFL)	20.80	18.80–21.07	19.98 ± 0.86b	18.47–22.06	20.50 ± 0.93b	18.03–22.38	20.52 ± 1.44b	19.97–24.08	22.33 ± 1.24a	19.73–21.97	20.87 ± 0.94ab
Anal-fin length (AFL)	18.86	16.96–19.73	18.57 ± 0.73b	16.85–19.77	18.19 ± 0.79b	17.42–20.76	18.71 ± 0.98b	17.43–25.80	20.39 ± 1.99a	17.59–19.89	18.83 ± 0.85ab
Dorsal-fin base length (DBL)	13.07	10.85–14.15	12.73 ± 1.09b	12.32–15.75	14.23 ± 0.71a	12.27–17.56	14.57 ± 1.29a	12.17–17.50	14.28 ± 1.42a	12.59–15.18	13.69 ± 0.99ab
Pectoral-fin base length (PBL)	7.66	7.40–8.57	8.08 ± 0.39c	7.81–9.97	8.75 ± 0.53b	8.59–9.59	9.14 ± 0.28ab	8.63–10.18	9.50 ± 0.47a	8.73–10.64	9.58 ± 0.63a
Pelvic-fin base length (VBL)	5.04	3.90–5.47	4.94 ± 0.39b	5.11–7.28	5.79 ± 0.60a	5.46–6.64	5.93 ± 0.40a	4.70–6.14	5.54 ± 0.45a	5.28–5.97	5.63 ± 0.31a
Anal-fin base length (ABL)	7.56	6.43–8.43	7.49 ± 0.64	6.55–9.11	8.05 ± 0.76	6.87–10.58	8.27 ± 0.99	6.37–8.99	7.76 ± 0.74	6.88–8.98	8.04 ± 0.70
Pre-dorsal length (PDL)	50.16	49.30–53.10	50.88 ± 1.17c	48.33–53.34	49.93 ± 1.32c	50.67–57.69	53.71 ± 1.93a	49.96–56.01	52.40 ± 1.60ab	49.51–51.73	50.71 ± 0.82bc
Prepectoral length (PPL)	21.71	18.90–23.21	20.79 ± 1.00 ab	18.96–21.62	19.94 ± 0.71b	18.31–22.56	20.59 ± 1.05 ab	19.17–23.82	21.30 ± 1.23a	19.79–20.80	20.41 ± 0.41ab
Pre-pelvic length (PVL)	53.09	51.26–55.61	52.97 ± 0.93a	50.07–53.41	51.46 ± 1.06b	48.92–52.56	51.06 ± 1.18b	49.69–54.58	51.98 ± 1.26ab	51.26–53.93	52.48 ± 0.89ab
Pre-anal-pore length (PAPL)	67.95	66.47–72.42	68.52 ± 1.46	66.07–70.50	68.06 ± 1.16	64.93–69.69	67.76 ± 1.30	64.63–70.83	67.59 ± 1.45	67.51–69.61	68.31 ± 0.87
Pre-anal-fin length (PAFL)	77.71	77.50–82.66	79.78 ± 1.38c	76.44–81.06	77.87 ± 1.42b	75.64–79.95	77.92 ± 1.30b	75.13–80.84	78.32 ± 1.49ab	79.15–80.96	80.06 ± 0.69ac
Pectoral-pelvic-fin insertion (PPOL)	32.88	32.05–36.01	33.79 ± 1.08	31.10–34.56	33.15 ± 1.12	31.09–35.23	32.65 ± 1.11	29.39–36.19	32.86 ± 1.69	32.99–35.11	33.68 ± 0.81
Pelvic-anal-fin insertion (PAOL)	24.93	24.93–29.72	27.83 ± 1.44	25.99–29.31	27.40 ± 1.04	26.53–30.34	28.32 ± 1.09	24.09–29.30	27.54 ± 1.26	25.21–29.28	28.01 ± 1.52
Width between mid-pectoral base (BPW)	17.96	15.64–18.23	17.16 ± 0.68c	17.13–19.80	18.24 ± 0.84b	18.03–20.68	19.41 ± 1.06a	16.41–21.72	19.14 ± 1.37ab	17.01–19.11	18.08 ± 0.82abc
Width between mid-pelvic base (BVW)	12.86	11.81–13.73	12.75 ± 0.56d	12.67–14.37	13.48 ± 0.46c	13.74–15.97	14.57 ± 0.76a	12.14–15.80	13.82 ± 0.83bc	13.24–14.30	13.70 ± 0.41abc
Upper caudal-lobe length (UCL)	24.32	20.89–26.02	24.30 ± 1.44abc	21.15–24.89	22.82 ± 0.94b	20.93–29.34	24.36 ± 2.06abc	22.35–28.53	25.05 ± 1.74c	22.02–26.70	24.46 ± 1.71abc
Lower caudal-lobe length (LCL)	26.64	19.26–27.53	24.75 ± 2.15b	21.28–27.07	23.58 ± 1.53b	22.60–29.94	25.51 ± 1.95ab	21.72–31.24	26.70 ± 2.22a	24.19–28.53	25.82 ± 1.78ab
% HL
Snout length (SnL)	52.23	47.82–63.77	55.67 ± 4.37b	53.98–61.11	57.32 ± 1.90 ab	54.84–61.36	57.61 ± 2.08 ab	54.48–61.93	57.65 ± 2.27ab	58.70–61.66	60.20 ± 1.04a
HD	41.75	41.75–61.92	53.39 ± 4.85a	45.22–54.02	49.77 ± 2.39b	44.42–50.36	47.58 ± 1.85b	43.54–51.62	48.47 ± 2.16b	47.84–52.57	49.97 ± 1.76ab
HW	69.14	69.14–97.27	86.30 ± 8.36a	74.31–81.83	78.89 ± 2.25b	74.93–82.53	78.17 ± 2.50 b	75.72–86.97	82.06 ± 2.89b	78.23–85.34	81.80 ± 3.07ab
Eye diameter (ED)	11.93	11.93–20.38	15.96 ± 2.64a	10.77–14.97	12.62 ± 1.35c	12.95–17.02	15.03 ± 1.33ab	12.65–17.88	15.22 ± 1.29ab	11.63–15.06	13.50 ± 1.32b
Inter-orbital width (IOW)	33.29	33.29–43.75	39.04 ± 2.60a	33.83–47.21	39.42 ± 3.54a	29.42–36.45	34.15 ± 1.84b	32.00–38.38	34.80 ± 1.89b	31.03–37.59	34.80 ± 2.63b
Post-orbital length (POL)	39.26	22.64–39.26	30.54 ± 5.35b	30.29–38.25	35.31 ± 1.98a	33.75–39.93	36.20 ± 2.04a	32.25–40.53	35.78 ± 1.94a	31.35–35.52	33.61 ± 1.59ab
Mouth width (MW)	34.63	24.14–34.63	27.27 ± 3.16b	31.89–40.61	36.12 ± 2.53a	22.68–32.65	28.54 ± 2.66b	23.66–33.98	27.78 ± 2.32b	30.90–40.11	36.01 ± 2.99a
% of caudal peduncle length (CPL)
Caudal-peduncle depth (CPD)	71.09	71.09–122.51	96.11 ± 18.38a	77.12–112.02	89.40 ± 10.17ab	69.53–100.17	83.08 ± 9.44b	71.86–100.14	84.39 ± 7.20b	83.50–94.32	88.95 ± 3.99ab
% ED
Outermost pair of rostral barbels length (OBL)	128.16	78.10–167.14	109.07 ± 24.97ab	92.95–159.71	122.11 ± 22.06a	69.60–105.51	90.76 ± 10.34b	74.55–118.26	93.52 ± 12.50b	110.00–145.89	124.76 ± 12.86a
Maxillary barbel length (MBL)	97.09	49.75–118.18	82.43 ± 21.24ab	63.75–114.49	84.90 ± 16.14a	51.54–83.67	67.80 ± 9.77b	53.57–86.67	71.89 ± 9.34 ab	77.37–94.69	82.77 ± 7.04 ab
Lower lip short barbel length (LBL)	32.04	24.34–51.98	34.39 ± 6.85a	16.10–41.73	26.75 ± 6.16b	21.13–38.87	30.07 ± 5.95ab	21.59–45.13	32.28 ± 5.73ab	26.84–42.66	36.16 ± 5.63a
% POW
SnL	133.04	133.04–233.97	187.08 ± 32.22a	143.37–194.32	162.90 ± 12.41b	140.92–179.01	159.72 ± 12.37b	137.17–181.73	161.72 ± 13.07b	167.50–193.75	179.56 ± 10.63ab
Meristic counts
Dorsal-fin rays	iii,8	iii,8		iii, 8–9		iii, 8		iii, 8		iii, 8
Pectoral-fin rays	i, 14	i, 14–15		i, 14–15		i, 15		i, 15		i, 16
Pelvic-fin rays	i,8	i,8		i, 7–8		i, 8–9		i, 8		i, 8
Anal-fin rays	ii,5	ii, 5		ii, 5		ii, 5		ii, 5		ii, 5
Lateral-line scales	87	72–88a		73–86ab		76–88a		65–83b		78–90a

Note: Superscript letters indicate statistically significant differences (p < 0.05) based on one-way ANOVA with post-hoc comparisons.

Body anteriorly subcylindrical, ventral profile flat from head to abdomen; dorsal profile gently rising from snout tip to dorsal-fin origin, then sloping to caudal-fin base; posterior body laterally compressed behind pelvic-fin base. Head slightly depressed, widest posteriorly; width greater than depth; 19.9–27.1% SL. Snout rounded, blunt; length 1.33–2.34× postorbital length. Mouth small, inferior, arched. Lips fleshy; upper lip broad, papillae absent; lower lip with one pair of short pointed posterior barbels and one pair of anterior lobe-like protrusions. Maxillary barbels one pair at mouth corners, each with basal pad bearing variably distinct pointed wart-like projection. Upper and lower lips connected at mouth corners by skin flaps and maxillary barbels; upper jaw concealed by upper lip.

Rostral groove present between upper lip and snout tip, bearing four primary barbels with bases in groove, alternating with tertiary pairs. Rostral fold anterior to groove, margin bearing six tertiary barbels arranged in three pairs; inner surface with three secondary barbels between each tertiary pair, bases on fold. Anterior and posterior nostrils adjacent, small, nasal flaps well developed. Eyes round; diameter 12.0–20.4% HL. Gill openings large, extending onto ventral surface of head.

Scales minute; head scaleless. Body scaled dorsally and ventrally to posterior of pelvic-fin origin. Ventrolateral scaling from pelvic-fin origin to pectoral axilla; scales medially embedded or absent; embedded scales first visible anterior to pectoral-fin axilla. Lateral line complete; 72–88 scales.

Dorsal-fin origin above and slightly anterior to pelvic-fin origin; insertion at or slightly anterior to mid-distance between snout tip and caudal-fin base; length 21.8–26.4% SL. Pectoral fin laterally expanded, margin curved. Pelvic fin laterally expanded, with a dorsal fleshy flap at base; tip surpassing anus. Anal fin reaching or slightly surpassing caudal-fin base when adpressed; length 17.0–19.7% SL. Caudal peduncle laterally compressed; depth 0.71–1.23× length. Caudal fin forked when compressed, slightly emarginate when fully expanded; lower lobe slightly longer.

Colouration in preserved specimens.

Dorsum pale yellow, patterned brown to black; venter creamy white. Dorsal surface of head with dark vermiculate markings. Predorsal region typically with one continuous longitudinal rectangular dark blotch, margins even to weakly undulate, 2–3 small light brown saddle spots occasionally connecting to lateral pale stripes. Post-dorsal region with 2–4 saddle-shaped patches separated by pale interspaces, connected to pale lateral stripes. Dorsal surface of precaudal base with one black spot. After long-term preservation in low-concentration alcohol, yellowish areas intensified

Flank with broad, dark longitudinal band along lateral line, 2–5 scales wide, margins even to weakly undulate; a straight pale longitudinal band, 1–3 scales wide, between lateral dark band and dorsal pigmentation, usually continuous to posterior margin of head, occasionally interrupted near pectoral region. Transition zone between flank and venter variably pale or bearing narrow continuous to interrupted dark stripe. Bases of paired fins dorsally with small black spot.

Fin membranes hyaline white. Paired fins with 0–2 broad dark bars perpendicular to rays, dorsal fin with 2–3 bars, anal fin with one bar, caudal fin with 3–4 bars; lower half of caudal-fin base with a narrow, transverse, elliptical, black spot.

Colouration in life.

Bright spots on body at dorsal base of paired fins, anterior and posterior to dorsal-fin base, and on upper half of caudal-fin base (Fig. 4).

Colouration of juvenile specimens.

Predorsal region with continuous dark longitudinal stripe enclosing 1–2 pale, central, rectangular areas; postdorsal region with 2–3 fragmented saddle patches. Lateral line with straight, continuous, dark stripe extending through eye to snout. Fin markings faint, except on caudal fin.

Etymology.

The species name honours Wen-Qiao Tang, a senior Chinese ichthyology researcher, and is derived from the latinized Chinese spelling of his family name, in recognition of his contributions to the field. We propose “Wén Qiáo Yīng Kŏu Qiū” (文乔缨口鳅) as its Chinese common name.

Distribution and habitat.

This species is endemic to the Jiulong River (九龙江) system, which flows independently into the sea in southeastern China. This species inhabits streams with gravel or pebbly substrates and has a carnivorous-leaning omnivorous feeding habit.

Molecular analysis.

We conducted phylogenetic analysis using 53 sequences. After alignment and trimming, a 1140-bp sequence was obtained, containing 788 conserved sites, 352 variable sites, 110 singleton sites, and 242 parsimony-informative sites. The average nucleotide frequencies were A = 25.7%, T = 27.9%, C = 30.3%, and G = 16.1%, showing an A–T bias (53.6%).

Both maximum-likelihood and Bayesian-inference analyses produced fully congruent tree topologies. All species formed well-supported monophyletic groups (ML bootstrap ≥ 99, BI posterior probability = 1). The genus Formosania was divided into three major clades: F. chenyiyui alone constituted clade I, which was recovered as the sister group to all other species. Clade II comprised all species characterized by a single row of rostral barbels, namely F. davidi, F. fasciolata, F. galericula, and F. immaculata. The remaining species fell into clade III, which was further resolved into two distinct lineages. Lineage 1 included F. stigmata, F. paucisquama, and F. tinkhami. Lineage 2 consisted of F. lacustris, which also encompasses F. gilberti under the taxonomic framework of Yeh et al. (2024), F. fascicauda, and the new species F. tangi sp. nov., positioned at the base of this lineage. All nodes were strongly supported except for the root of clade III, which received moderate support (ML = 55%, BI = 0.66) (Fig. 5).

Figure 5.

Bayesian-inference tree based on mitochondrial cytb gene sequences from 11 species of Formosania. Maximum-likelihood and Bayesian-inference analyses yielded congruent topologies. ML bootstrap / BI posterior probability is displayed at superspecific nodes; a dash (“–”) indicates that the value is not applicable.

Genetic distances among the 11 Formosania species were calculated using the Kimura 2-parameter (K2P) model. The smallest genetic distance to F. tangi sp. nov. was observed in F. fascicauda (4.2%). Notably, this interspecific divergence exceeded that between many other congeneric species pairs, such as F. fascicauda and F. lacustris (2.93%), among F. tinkhami, F. stigmata, and F. paucisquama (2.12–3.48%), and among F. davidi, F. fasciolata, and F. galericula (2.40–3.05%) (Table 3).

Table 3.

Inter-specific mean mitochondrial cyt b genetic distances within the genus Formosania in Kimura 2-parameter genetic distance analysis (%).

	1	2	3	4	5	6	7	8	9	10
1. F. tangi sp. nov.
2. F. lacustris	4.67
3. F. chenyiyui	17.91	17.19
4. F. tinkhami	6.88	7.41	16.38
5. F. paucisquama	7.12	7.72	16.81	3.48
6. F. stigmata	7.02	7.46	16.18	3.19	2.12
7. F. fascicauda	4.16	2.93	15.97	7.00	7.84	7.33
8. F. immaculata	9.69	9.55	18.69	8.60	8.79	9.08	8.68
9. F. galericula	9.28	9.12	17.90	9.05	8.87	8.97	8.53	4.80
10. F. fasciolata	9.43	9.54	18.23	9.05	8.88	9.02	8.51	4.94	3.05
11. F. davidi	9.26	9.22	18.64	8.78	9.06	8.87	8.35	4.73	2.40	2.96

Morphometric traits.

The principal component analysis (PCA) of morphometric traits clearly separated F. tangi sp. nov. and F. fascicauda along the first two principal components. PC1 and PC2 accounted for 47.77% and 14.85% of total morphological variation, respectively, together defining the major axes of divergence (Fig. 6). The morphospace occupied by each species, represented by blue (F. tangi) and orange (F. fascicauda) polygons, showed clear separation, primarily along PC1. Specimens of F. tangi clustered in quadrants 1–3, mostly left of the y-axis, and were characterized by positive loadings on MBL, OBL, and LCL, and negative loadings on LBL and ED. In contrast, F. fascicauda grouped in quadrants 1 and 4, influenced mainly by positive loadings on POL, MW, and CPL.

Figure 6.

Principal component analysis (PCA) ordination plot of morphometric traits for Formosania tangi sp. nov. and F. fascicauda based on PC1 and PC2.

Although one-way ANOVA identified significant differences in certain morphometric ratios among the double-row barbel species, all variables exhibited broad interspecific overlap. Thus, morphometric traits offer limited diagnostic utility for distinguishing these species (Table 2).

Discussion

Caudal fin

The fully expanded caudal fin within the genus Formosania exhibits three distinct morphological patterns: a plain subtruncate type, a patterned subtruncate type, and a patterned emarginate type (Fig. 7). The plain subtruncate type, which is characterized by a nearly straight to slightly oblique posterior margin and the absence of distinct dark stripes on the caudal-fin rays, is found exclusively in the F. davidi species group (clade I in the phylogeny; Fig. 7A). The patterned subtruncate type, featuring a distinctly oblique posterior margin and 3–4 prominent dark stripes on the caudal-fin rays, is unique to F. fascicauda (Fig. 7B). All remaining species, including F. tangi sp. nov., display the patterned emarginate type, which is defined by a slightly concave posterior margin and the presence of 3–4 prominent dark stripes on the caudal fin rays (Fig. 7C, D).

We emphasize that the preceding morphological comparisons pertain specifically to the fully expanded state of the caudal fin. In an unexpanded condition, all three types exhibit a central concavity of varying extent. Moreover, while the first two types may not present a perfectly obliquely truncate posterior margin, their concavity is markedly shallower than that characteristic of the patterned emarginate type.

Flank markings

Formosania tangi sp. nov. is readily distinguished from F. fascicauda by its distinctive flank markings, with differences evident from juvenile to adult stages (earlier developmental stages were not obtained). Formosania tangi displays a straight, dark, mid-lateral stripe along the lateral line, which is consistently broader than in F. fascicauda. A dark dorsal stripe runs mid-dorsally in juveniles; the section behind the dorsal fin gradually forms regular, saddle-like blotches during growth. In contrast, F. fascicauda shows a thin, dark mid-lateral line with alternating dorsal and ventral blotches that enlarge and become irregular with age. Along the dorsum, it exhibits 5–7 saddle-shaped blotches, regular in juveniles but turning irregular and laterally extended in adults. Similar marking patterns occur in other species with two rows of barbels such as F. stigmata and F. paucisquama (Fig. 8)

Resolution of a long-term taxonomic misidentification

Chen’s (1980) report of F. fascicauda from the Mulanxi and Jiulongjiang rivers constituted the most accurate interpretation possible given the material and methods available at the time. Subsequently, its incorporation into Fauna Sinica (Chen and Tang 2000) conferred institutional authority on this identification, which shaped the scientific understanding of these populations for decades.

We now recognize this interpretation as an understandable misassignment, stemming from the absence of topotypic F. fascicauda for direct comparison and the inherently subtle morphological differentiation among many Formosania populations in the coastal basins of Fujian. In this context, F. tangi sp. nov. presents a clear exception with its distinctive pigment pattern, whereas populations in the Jinjiang and Mulanxi rivers remain morphologically similar to true F. fascicauda – a continuity that reasonably led earlier workers to presume intraspecific variation.

Our integrated analysis, based on topotypic specimens (Fig. 9) and molecular data, confirms that F. tangi is morphologically diagnosable and phylogenetically distinct. Under the current taxonomic framework that retains F. fascicauda, F. lacustris, and F. gilberti as valid species, F. tangi does not form a monophyletic group with any of these recognized taxa. These consistent lines of evidence firmly support the recognition of F. tangi as a distinct species.

Figure 9.

Dorsal and ventral views of Formosania fascicauda. (SHOU20251010101, from Longjiang River, the type locality, in Fuqing City, Fujian Province), caudal fin not fully expanded.

The present analysis conclusively establishes that F. tangi and F. fascicauda can be reliably distinguished based on external morphology. The most prominent diagnostic trait is the lateral body pattern. In F. tangi, this pattern consists of a broad, dark, longitudinal band accompanying the lateral line, with margins that are weakly undulating to nearly straight. This band is separated from the darker dorsal surfaces by a generally continuous, straight, pale stripe (Fig. 8A), a configuration that is highly distinctive in Formosania. In contrast, F. fascicauda exhibits a thin, black longitudinal line composed of alternating patches, with the pale stripe being intermittent, wavy, or entirely absent (Fig. 8B)—a pattern more similar to those observed in congeners such as F. stigmata and F. paucisquama. Furthermore, caudal fin morphology serves as a reliable secondary diagnostic character. When fully expanded, the caudal fin of F. tangi is distinctly emarginate. This condition contrasts clearly with that of F. fascicauda, in which the fully expanded fin is typically subtruncate. It is noteworthy that even in the occasional F. fascicauda specimen where the fully expanded caudal fin is not perfectly truncate and shows slight incurvation, it never exhibits the pronounced emargination characteristic of F. tangi. Collectively, these consistent and observable morphological differences robustly support the distinction between the two species.

Diagnostic key to species of Formosania

1	Tertiary rostral barbels occasionally papilliform; barbels and papillae 12–15; dorsum with broad longitudinal dark band, incurved at dorsal-fin axil	(Hanjiang River) F. chenyiyui
–	Rostral barbels well developed; barbels 13; dorsum with saddle-shaped blotches or reticulated pattern	2
2	Rostral barbels in single row, continuous with rostral fold; median caudal-fin rays unpigmented	3
–	Rostral barbels in two rows, posterior row situated along rostral groove; caudal-fin with 2–4 dark transverse stripes	6
3	Body unmarked; snout about 1.2 times postorbital length	(Oujiang River) F. immaculata
–	Body marked; snout subequal to postorbital length	4
4	Abdominal scaleless area restricted anterior to midpoint of pectoral-fin base	(Feiyunjiang, Oujiang and Aojiang rivers) F. fasciolata
–	Abdominal scaleless area extending slightly beyond pectoral-fin axilla	5
5	Longest rostral barbel about 1.5 times eye diameter; dorsum with 7–9 saddle-shaped bands	(Minjiang and Xinjiang rivers) F. davidi
–	Rostral barbels subequal to eye diameter; dorsum with reticulated pattern	(Oujiang River) F. galericula
6	Lateral line bordered by broad dark longitudinal band, margins weakly undulate to nearly straight; pale mid-dorsal stripe continuous and straight	(Jiulongjiang River) F. tangi sp. nov.
–	Lateral line with a thin black longitudinal line composed of alternating patches, or small alternating dark dots, or a continuous narrow line; pale mid-dorsal stripe intermittent, wavy, or absent	7
7	Fully expanded caudal-fin oblique subtruncate; postdorsal flank with a pale intermittent wavy stripe	(Longjiang and Yuxi rivers) F. fascicauda
–	Fully expanded caudal fin emarginate; pale stripe absent	8
8	Lateral line with a thin black longitudinal line with alternating patches	9
–	Lateral line with small alternating dark dots, or a black line	10
9	Lateral line scales 86–92	(Minjiang and Jiaoxi rivers) F. stigmata
–	Lateral line scales 76–83	(Rongjiang, Lianjiang, Hanjiang Rivers) F. paucisquama
10	Flanks unblotched	(Jhuoshui, Zhengwen and Nei Shuong rivers, Taiwan) F. lacustre
–	Flanks with interrupted dark short streaks or small patches along dorsal and ventral margins, or vermiculate	11
11	Flanks vermiculate	(Tamshui, Beihuang, Shuang and Masu Rivers, Taiwan) F. gilberti
–	Flanks with interrupted dark short streaks or small patches along dorsal and ventral margins	(Pearl River) F. tinkhami

Comparative materials

Formosania tangi sp. nov. • SHOU20240704603-605, SHOU20251010601-602, 5 specimens, 56.84–73.69 mm SL; China, Fujian Province, Sanming City, Jiulongjiang River. • SHOU202201083-093, 11 specimens, 50.53–81.06 mm SL; China, Fujian Province, Zhangzhou City, Jiulongjiang River.

Formosania fascicauda • SHOU20241019001-005, SHOU20241019007-008, SHOU20241019013-014, SHOU20241019016-019, SHOU20241019021-022, 15 specimens, 53.98–83.30 mm SL; China, Fujian Province, Fuqing City, Yuxi River. • SHOU20241122001, 1 specimen, 74.24 mm SL; China, Fujian Province, Fuqing City, Longjiang River.

Formosania stigmata • SHOU2021060126-135, 10 specimens, 50.04–75.85 mm SL; China, Zhejiang Province, Lishui City, Jiaoxi River. • SHOU2021060180-183, 4 specimens, 50.98–85.47 mm SL; China, Zhejiang Province, Lishui City, Minjiang River.

Formosania paucisquama • SHOU202311001-013, 13 specimens, 49.62–65.76 mm SL; China, Guangdong Province, Meizhou City, Hanjiang River. • SHOU202101001-005, 5 specimens, 52.26–64.03 mm SL; China, Guangdong Province, Jieyang City, Rongjiang River. • SHOU202110011, 1 specimen, 49.04 mm SL; China, Guangdong Province, Jieyang City, Lianjiang River.

Formosania tinkhami • SHOU20231108001-006, 6 specimens, 51.98–80.29 mm SL; China, Guangdong Province, Huizhou City, Pearl River.

Citation

Chen Y, Zhou J-J, Chen J-C, Yang J-Q (2026) Formosania tangi, a new species of suck-loach (Cypriniformes, Gastromyzontidae) from the Jiulongjiang River, southeastern China, with taxonomic notes on F. fascicauda. ZooKeys 1273: 147–166. https://doi.org/10.3897/zookeys.1273.184335

Funding Statement

National Natural Science Foundation of China (No. 31872207)

Additional information

Conflict of interest

The authors have declared that no competing interests exist.

Ethical statement

No ethical statement was reported.

Use of AI

No use of AI was reported.

Funding

This research was funded by a grant from the National Natural Science Foundation of China (No. 31872207).

Author contributions

Conceptualization: JQY. Methodology: JCC. Software: JCC. Validation: JCC. Formal analysis: JCC, YC. Investigation: JJZ. Resources: JJZ. Data Curation: YC. Writing – Original draft: YC. Writing – Review and Editing: JQY, JCC. Visualization: JCC. Supervision: JQY. Project administration: JQY. Funding Acquisition JQY.

Author ORCIDs

Yang Chen https://orcid.org/0009-0009-9923-917X

Jia-Jun Zhou https://orcid.org/0000-0003-1038-1540

Jing-Chen Chen https://orcid.org/0000-0001-9164-6920

Jin-Quan Yang https://orcid.org/0000-0003-0387-1824

Data availability

All of the data that support the findings of this study are available in the main text.