Laubuka tenella is a new species characterized by the colour pattern, consisting of short dark vertical bars anteriorly on the side, and a dark lateral band posteriorly on the side, combined with a relatively short pelvic fin and 29–30 lateral-line scales. It is separated from other Laubuka analysed by minimum 9 % uncorrected p-distance in the mitochondrial COI gene. The type series is composed of specimens from small streams in the Cox’s Bazar District in Bangladesh (the type locality), and the Thandwe River drainage in western Myanmar. Laubuka brahmaputraensis is strongly indicated to be a junior synonym of L. laubuca, the second known species of Laubuka in Bangladesh. Eustira ceylonensis, currently in the synonymy of Devario malabaricus, is a valid species of Laubuka.

DNA barcode, morphometry, phylogeny, South Asia, taxonomy, vertebral counts

Species of the tropical Asian cyprinid genus Laubuka Bleeker, 1860, are characterized by relatively small size, up to about 60 mm SL, a strongly compressed body and a keeled abdomen, a long and stiff falciform pectoral fin, and usually a long free distal portion of the first pelvic-fin ray (Pethiyagoda et al. 2008, Silas 1958). Laubuka resembles Devario Heckel, 1843, in presence of supraorbital and rostral neuromast-filled grooves, posteriorly situated dorsal fin, and presence of a dark blotch immediately posterior to the gill opening. Unlike Devario, most species of Laubuka show only indistinct markings, and barbels have not been recorded from Laubuka.

Species of Laubuka have been reported from southern Asia in Bangladesh, Cambodia, India, Indonesia, Laos, Malaysia, Myanmar, Pakistan, Sri Lanka, Vietnam, and Thailand (Silas 1958, Kottelat 2013). About 13 species are recognized at present (Pethiyagoda et al. 2008, Knight 2015, Lalramliana et al. 2017), but the genus has not been subject to revision since that of Silas (1958), who employed a very different concept of taxa currently in Laubuka. Banarescu (1968) examined and illustrated several species of Laubuka, but his paper is mainly concerned with the description of a new subgenus, Malayochela, monotypic with M. maassii (Weber & de Beaufort, 1912). Banarescu’s concept of the other species of Laubuka is similar to that of Silas (1958). Several of the species presently in Laubuka were previously, e.g., in Silas (1958) and Banarescu (1968), in the genus Chela (Hamilton, 1822), from which they were separated by Pethiyagoda et al. (2008). Species diversity in Chela was addressed by Knight and Rema Devi (2014), who recognized two valid species in that genus. Another genus previously included as a subgenus of Chela is Neochela Silas, 1958, containing only the miniature species N. dadiburjori (Menon, 1952). Malayochela was considered to be a distinct genus by Pethiyagoda et al. (2008).

The type species of Laubuka is Cyprinus laubuca Hamilton, 1822, which was described from “ponds of the northern parts of Bengal” (Hamilton 1822), i.e., within the present state of West Bengal in India, and Bangladesh. Hamilton (in Day, 1877) reporting on surveys 1807–1813, specifically mentioned the species from Goalpara in Rangpur (currently Goalpara District, Assam, India, Brahmaputra basin), Mahananda River in Purniah (Purnea District, Bihar, Ganga basin), and in Gorakhpur District (Uttar Pradesh, Ganga basin), suggesting that he did not observe the species in localities within present-day Bangladesh. Hamilton left India in 1815 (Day, 1877), however, and would have had opportunities to make unrecorded observations of fishes. Silas (1958), followed by Talwar and Jhingran (1991), reported L. laubuca as distributed in Pakistan, India, Bangladesh, Sri Lanka, Myanmar, Malay Peninsula, and Sumatra. Recent publications only list specimens from India (Knight 2015, Pethiyagoda et al. 2008) and Bangladesh (Rahman and Chowdhuri 2007), but there is no recent revision covering this species. Silas’s (1958) concept of L. laubuca included the southeast Asian L. siamensis Fowler, 1939, and the Sri Lankan L. lankensis (Deraniyagala, 1960), revalidated by Pethiyagoda et al. (2008), and would have accommodated also the three Sri Lankan species described by Pethiyagoda et al. (2008) as L. insularis, L. ruhuna, and L. varuna.

Despite previous records from Bangladesh, we were unable to find Laubuka in natural habitats in the Meghna, eastern Brahmaputra, and Karnafuli drainages in Bangladesh During field work 2014–2016. Laubuka laubuca was, however, present in small numbers in aquarium shops in Dhaka, and shop owners said that they had been caught locally. In 2015, we collected specimens of Laubuka from streams near Cox’s Bazar and Teknaf in extreme southeastern Bangladesh. These samples represent a species very different from L. laubuca, but morphologically indistinguishable from a sample of Laubuka from a stream on the western slope of the Rakhine Yoma in Myanmar. This paper is dedicated to the description and diagnosis of this new species.

Specimens used were available in museums, purchased from fishermen or markets; or collected in the wild using beach seine or hand net and euthanized through immersion in buffered tricaine-methanesulphonate (MS 222) until cessation of opercular movements plus an additional 30 minutes, in accordance with permits from the Swedish Environmental Protection Agency (dnr 412-7233-08 Nv) and the Stockholm Ethical Committee of the Swedish Board of Agriculture (dnr N 85/15). Collecting in Bangladesh was conducted under a permit to the University of Dhaka. Voucher specimens are deposited in the collections of The Natural History Museum, London (BMNH), University of Dhaka, Dhaka (DU), Museum of Zoology, Lund University (MZLU), and the Swedish Museum of Natural History, Stockholm (NRM).

Measurements and counts were taken as described by Fang (1997) with the exception that body depth was taken at level of the origin of the anal-fin, which is very slightly anterior to the origin of the dorsal fin. Counts of lateral-line scales do not include perforated scales on the caudal-fin base. The last two dorsal- and anal-fin rays share the same proximal pterygiophore and may appear as a single ray; these two rays are counted as 1½. Vertebral counts are given as precaudal+caudal, where the first vertebra bearing a haemal spine (anterior to the first long anal-fin pterygiophore) was recorded as the first caudal vertebra. Digital X-radiographs made with a Kevex 130kVP microfocus X-ray source and a Samsung/Rayence 17×17 inch DR panel. Statistics were calculated using SYSTAT v. 13 (Systat Software 2009).

For the genetic analysis, 655 basepairs from the 5’ end of the mitochondrial cytochrome-oxidase subunit 1 (COI, or COX1) gene were sequenced from seven morphologically identified specimens of Laubuka plus two specimens of the closely related Chela cachius (Hamilton, 1822). DNA was extracted using a Thermo Scientific KingFisher Duo (Thermo Fisher Scientific, Waltham, USA) fully automated liquid-handling instrument, with the Thermo Scientific KingFisher Cell and Tissue DNA Kit (Thermo Fisher Scientific, Waltham, USA) with the recommended protocol. PCR were performed with the puReTaq Ready-To-Go PCR kit (Amersham Biosciences AB, Uppsala, Sweden). The COI fragment was amplified using the fish barcoding primers Fish-F1 and Fish-R1 [26], with the PCR cycling: 94 °C 4 min; 35 * (94 °C 30 s; 52 °C 30 s; 72 °C 30 s); 72 °C 8 min). The PCR products were checked on agarose gel, and purified by adding 5 µL of a mix consisting of 20 % Exonuclease I (EXO) and 80% FastAP Thermosensitive Alkaline Phosphatase (Fermentas/Thermo Fischer Scientific, Gothenburg, Sweden) to each 25 µl PCR reaction, incubated at 37 °C for 30 minutes, then heated to 80 °C for 15 minutes. Sequencing of both strands of all fragments was carried out by Macrogen Europe (Amstelveen, The Netherlands). Sequences were assembled, aligned, and proofread in Geneious version 10 (Kearse et al. 2012). Geneious was used to calculate genetic distances (uncorrected pairwise p-distance, as recommended by Srivathsan and Meier (2011)), and the Geneious plug-in Species Delimitation (Masters et al. 2011) was used to calculate the probability of reciprocal monophyly under a model of random coalescence. A phylogenetic hypothesis was constructed in MrBayes version 3.2 (Huelsenbeck and Ronquist 2001; Ronquist and Huelsenbeck 2003) (2 million generations, GTR + Γ + I model), data partitioned by codon position; samples were taken every 1000 generations, and the first 25 % of samples were discarded as ‘burn-in’. Convergence was checked with Tracer version 1.6 (Rambaut et al. 2014).

The distribution map was constructed using layers from Natural Earth (http://www.naturalearthdata.com).

Comparative material. Chela cachius: DU 6116, 1, 40.9 mm SL. Bangladesh: Chittagong Division: Feni River drainage, Kohua River; M.M. Rahman, 29 May 2015. — NRM 40494. 4, 46.8–49.3 mm SL; India: Assam: about 100 km SE of Dibrugarh, small river falling into the Dibru River 3 km north of Digholtarang; F. Fang & A. Roos, 22 Jan 1998. — NRM 66988 (T10085), 1, 35.7 mm SL. Bangladesh: Dhaka Division: Padma River near Sreenagar, M.M. Rahman et al., 2 Dec 2014.

Devario cf. malabaricus.BMNH 1852.2.19.130–132, 1853.12.24.6, 1864.4–11.33; 6, 50.1– 67.2 mm SL; Ceylon.

Laubuka laubuca: Bangladesh: NRM 67315 (T10453), 1, 35.8 mm SL; NRM 67316 (T10454), 1, 31.9 mm SL; NRM 67317 (T10455), 1, 42.7 mm SL. Bangladesh, ornamental fish shops in Dhaka, said to be local fish; M.M. Rahman et al., 16 May 2015. India, Assam: NRM 40308, 1, 48.7 mm SL; NRM 40486, 1, 44.0 mm SL; NRM 40517, 55.6–58.3, Assam, about 100 km SE of Dibrugarh, rivulet falling into the Dibru River 3km N of Digholtarang, F. Fang & A. Roos, 22 Jan 1998. – NRM 44997 (T10965), 38.9 mm SL; Brahmaputra River at Tezpur, Sonitpur; H. Bleher, 28 Feb 2009. — NRM 52691, 1, 43.8 mm SL; Golaghat: Kaziranga, Deosor subdrainage; O. Åhlander et al., 24 Oct 2005. — NRM 57234, 2, 49.9–53.1 mm SL; Dibrugarh, Mr Kamal Lahkar’s home fish farm “Brahmaputra Aquaria”. F. Fang & A. Roos, 19 Jan 1998.

Laubuka sp.: NRM 12218, 59.0 mm SL; India: Kerala: Meenachil River and adjacent canals, NW of Kottayam; E. Åhlander et al., 5–6 Dec 1987.

Laubuka parafasciata Lalramliana, Valalhlimpuia & Singh, 2017: All from Myanmar, Rakhine State. BMNH 2017.8.2.2, 1, 45.6 mm SL; Kananme Chaung near Leldee village and Chaung Ma Gyi Chaung, Leldee village, 18°35.814'N, 94°22.853'E and 18°35.112'N, 94°22.182'E; R. Britz, 29 Nov 2009. — BMNH 2017.8.2.3-4, 2, 64.7–85.3 mm SL; Chaung Ma Gyi Chaung, Leldee village, N 18°35.112'N, 94°22.182’E; R. Britz, 28 Nov 2009. — BMNH 2017.8.2.21, 1, 37.8 mm SL; Kananmee Chaung; Ye Hein Htet, 3 Dec 2004. — BMNH 2017.8.2.23, 1, 46.9 mm SL; Ann Chaung near Ann; R. Britz, 24–25 Nov 2009. (Ann = 19°47’0"N, 94° 2'0"E).

Laubuka insularis: All from Sri Lanka. MZLU 962/5429, 2, 20.3–33.3 mm SL; Eastern: Rambukkan Oya River, 25 miles NE of Bibile. P. Brinck et al., 8 Mar 1962. (125). — MZLU 962/5442. 6, 32.9–44.0 mm SL; Eastern: Gal Oya River, 14 miles E of Bibile; P. Brinck et al., 8 Mar 1962. (123). — MZLU 962/5459, 1, 47.8 mm SL; North Central: Mahaweli River drainage: small stream 3 mi. S of Minneriya; P. Brinck et al., 11 Feb 1962 (67).

Laubuka cf. varuna: MZLU 962/5438, 4, 29.3–46.0 mm SL; Sri Lanka: Central: Talagala Oya stream at Pidurutalagala, 2 miles N of Nuwara-Eliya; P. Brinck et al., 4 Mar 1962. (116:I).

Laubuka siamensis: NRM 31223, 6. Viet Nam; Dong Nai River drainage, about 30 km ENE of Ho Chi MinhCity, Xa Trang Bom, road crossing small stream east of village, 10°57'1"N, 106°58'28"E; B. Björkegren, 25 Mar 1935. Information on L. ruhuna, L. varuna, L. lankensis from Pethiyagoda et al. (2008); on L. khujairokensis (Arunkumar, 2000) from Arunkumar (2000); on, L. caeruleostigmata (Smith, 1931) from Smith (1931); and on L. trevori Knight, 2015 and L. latens (Knight, 2015 from Knight (2015).

GenBank Accession numbers. New COI sequences generated for this paper are:

Chela cachius, DU 6116: MG895632; NRM 66988: MG895633.

Laubuka laubuca, NRM 67315: MG895634; NRM 67317: MG895635; NRM 44997: MG895636.

Laubuka tenella: NRM 67380: MG895640; DU 9008/NRM 67381: MG895637; NRM 67862: MG895639; DU 9006/NRM 67845: MG895638.

The first report of Laubuka laubuca from Bangladesh was by Rahman (1989), who illustrates it with a drawing copied from Day (1878: pl. 151, fig. 5), showing a specimen from Myanmar. The records by Rahman (2005) and Rahman and Chowdhury (2007), are illustrated with a photo that shows a specimen of Esomus danrica (Hamilton, 1822). No Laubuka were reported by Ahmed et al. (2013) from their survey of hillstream fishes in northern and southeastern Bangladesh. The distribution of L. laubuca within Bangladesh remains to be mapped. Based on the comparative material used here, Laubuka laubuca has a wide distribution in the Brahmaputra in Assam, whereas records from other areas need revision. Laubuka brahmaputraensis Kulabtong, Suksri & Nonpayom, 2012, is the only other species of Laubuka reported from Bangladesh. It was described on the basis of aquarium specimens without precise collecting locality, but said to be from the Brahmaputra basin in Bangladesh (Kulabtong et al. 2012), and has been considered to be a possible junior synonym of L. laubuca (Kottelat 2013). Laubuka tenella is distinct from L. laubuca in colour pattern and mitochondrial DNA, and there is no indication that it has been misidentified as L. laubuca in earlier literature.

The Cox’s Bazar and Rakhine samples of L. tenella share the same colour pattern and proportional measurements, but differ slightly in frequencies of anal-fin ray counts. Ranges overlap, however, and the anal fin count shows intraspecific variability also in other species of Laubuka. There is no variation in dorsal-fin count or circumpeduncular scale count, but these counts (ii.8½ and 12, respectively) are shared with most other species in the genus. The miniature species Laubuka fasciata ii.7½ dorsal-fin rays, and 10 circumpeduncular scales. Pethiyagoda et al. (2008) reported dorsal-fin rays ii.9½ in some Sri Lankan species and we observed this count in a specimen of an unidentified species of Laubuka from Kerala (NRM 12218) .

Laubuka khujairokensis, L. latens, L. laubuca, and L. siamensis are characterized by absence of short dark bars anteriorly on the side. However, Das (1939: fig. 1) figured a specimen identified as L. laubuca from the Damodar River near Hazaribagh (Hugli River drainage), which shows short indistinct vertical bars on the side, but no lateral stripe. Otherwise, a series of short vertical bars on the side as in L. tenella, have been reported only from Sri Lankan Laubuka, by Pethiyagoda et al. (2008). Laubuka trevori has a narrow dark stripe along the side, stated to be less distinct anteriorly (Knight 2015), but apparently no vertical bars. Consequently, L. tenella is diagnosed on the basis of colour pattern and, as verified from Bangladeshi material only, the COI sequence. Laubuka tenella is also relatively slender, and may differ in that regard from L. laubuca (Fig. 3, Tables 1–2) , and Sri Lankan species (body depth 32.8–34.6 % SL in L. ruhuna, 27.9–32.4 % in L. varuna, 27.2–30.3 % in L. lankensis according to data in Pethiyagoda et al. (2008).

Laubuka brahmaputraensis was described from three specimens collected by an aquarium fish collector and said to be from Brahmaputra but without precise locality. The published photograph (Kulabtong et al. 2012; fig. 1) of the holotype, preserved in 1995, suggests that it is in a poor state of preservation, similar to specimens kept long time in buffered formalin. The photo shows a trace of the cleithral spot but no other markings. The diagnosis and description do not include characters separating from L. laubuca reported here from Bangladesh and northeastern India. Kulabtong et al. (2012) did not have specimens of L. laubuca for comparison, but based their concept of L. laubuca on the meristic data in “Rahman (2003)” which is apparently an error for Rahman (2005). The validity of this comparison is doubtful.

Rahman’s (2005) description is not in accord with Laubuka from Bangladesh or northeastern India, but also not compatible with Esomus danrica (Hamilton, 1822), the species on the illustration of L. laubuca in Rahman (2005). Rahman’s fin counts are compatible with most species of Laubuka, but the counts of 34–36 lateral-line scales, and 20–21 predorsal scales stand out. The description is almost the same as in Rahman (1989) and the counts are identical. The high lateral-line counts and predorsal scale counts seem to be based partly or entirely on earlier literature. Silas (1958) identified L. laubuca from a wide geographical area, including India, Sri Lanka, Myanmar and adjacent south-east Asia, resulting in a lateral line count range of 31–37, as cited by Talwar and Jhingran (1991). Silas (1958) did not report specimens from Bangladesh, and his specimens from Myanmar were from east of the Rakhine Yoma. Jayaram (1981, 1999) stated the lateral line count to be “34 to 37, large, not many”, a count that may go back to the count of 34–37 in Day (1878: 598). Day’s figure, pl. 151, fig. 5, is stated to show a specimen from Burma and most likely does not show L. laubuca. Day’s description also covers material from India and Sri Lanka. Day’s Myanmar specimens, included in L. laubuca by Silas (1958), now in the BMNH, are from Mawlamyine, Sittaung River, and Mandalay (Silas 1958). Day’s image was copied by later authors to illustrate L. laubuca, e.g., in Jayaram (1981) and Rahman 1989). Pethiyagoda et al. (2008) reported elevated counts of 34–37 lateral line scales in L. lankensis and 34–36 in L. insularis, but only and 31–33 in L. varuna and L. ruhuna.

Because earlier concepts of L. laubuca were highly inclusive, including several species, and also including information copied from literature sources, published information on L. laubuca cannot be relied on for diagnosis of the species. Based on the locality information provided by Hamilton (1822; Day, 1877), it seems reasonable that the species of Laubuka present in West Bengal and Assam represents L. laubuca in the strict sense. This species, with 30–33 lateral-line scales has a colour pattern that includes a distinct vertically oriented cleithral spot, and a thin black stripe from the middle of the side which is extended caudad but not reaching to the caudal-fin base, and which is slightly expanded on the middle of the caudal peduncle. This colour pattern is illustrated by Pethiyagoda et al. (2008: fig. 13), and Lalramliana et al. (2017: fig. 2), and in Fig. 7A–B). Except for the absence of data in Kulabtong et al. (2012), L. brahmaputraensis agrees with L. laubuca in this strict sense, suggesting that L. brahmaputraensis is a junior synonym of L. laubuca.

Figure 7. 

Laubuka laubuca. ANRM 57234, 53.1 mm SL. India: Brahmaputra River drainage: vicinity of Dibrugarh BNRM 67317 mm SL, preserved in 95% ethanol. Bangladesh: vicinity of Dhaka.

Perilampus perseus M’Clelland [1839: 395, pl. 46, fig. 5 (erroneously stated to be pl. 48, fig. 5 in the text]; as P. persus in the index)] was listed as a questionable synonym of L. laubuca by Day (1878), and has not been recognized as a distinct species since. The distribution was given as “Assam, and probably Bengal”, and the description of the only specimen said to be “so much injured that I am unable to determine with accuracy the number of scales and caudal fin rays” (M’Clelland, 1839). Knight (2016) considered M’Clelland’s illustration to be reminiscent of Salmostoma, but particularly the long pelvic fin rather supports the identification as a species of Laubuka. Based on the locality information by M’Clelland (1839), P. perseus is very likely to be a junior synonym of L. laubuca. Perilampus guttatus M’Clelland, 1839, is an unneeded replacement name for Cyprinus laubuca Hamilton, 1822 (= L. laubuca).

Among the Sri Lankan species of Laubuka described by Pethiyagoda et al. (2008), L. lankensis, and L. insularis were distinguished from L. ruhuna and L. varuna by the presence of tubercles on the lower jaw. Those structures are, however, neuromast fields, as observed in L. insularis and other species of Laubuka, and should be expected from L. varuna and L. ruhuna as well. Specimens identified as possibly L. varuna (LZM 962/5438) have well developed lower jaw neuromast fields. Lower jaw neuromast sensory fields may be a unique character of Laubuka, but have not yet been studied at organ level. The diagnoses of the Sri Lankan species, based mainly on scale counts, fin lengths and body depth, make it difficult to separate them from continental Asian congenerics, but it seems unlikely that any of them is conspecific with species in the northern part of India and Bangladesh.

Günther (1868: 331) described Eustira ceylonensis as a new genus and species based on six specimens from Ceylon, purchased of Mr Cuming [Hugh Cuming, 1781–1865]. Day (1878: 599) included E. ceylonensis in Perilampus. Day’s Perilampus equates Laubuka+Chela. His description of P. ceylonensis is based on data from Günther (1868); apparently Day had not examined Günther’s specimens. Silas (1957) synonymized E. ceylonensis with Danio malabaricus (Jerdon, 1849) (now Devario malabaricus) based on his own examination of the presumed type series. Silas also provided a line drawing, supplied by Ethelwynn Trewavas, of a specimen stated to be holotype of E. ceylonensis, and which seems to show a specimen of Devario. It is noteworthy that the data from the specimens examined by Silas disagree with Günther’s description, while in the same volume Günther (1868: 282–283) lists specimens of D. micronema (Bleeker, 1863) from Ceylon (“a, b, c, d-e, f-h. Adult and half-grown”), of which a is stated to have been purchased of Mr. Cuming; and also material of other species of Danio/Devario. Silas’s assumption that Günther was unable to distinguish between Laubuka and Danio/Devario may have been somewhat precipitant. It seems rather that a sample of D. micronema (as identified by Günther) was mistaken by Silas and Trewavas for the type series of E. ceylonensis.

Günther’s description of E. ceylonensis is compatible with characters of Laubuka. Günther’s diagnosis of the genus, referring to “entire abdominal edge being trenchant”, i.e. keeled, as in Laubuka, stands in contrast to the rounded abdomen in Devario. The characters “pectorals elongate” (characteristic of Laubuka; not particularly long in Devario); “barbels none” (as in Laubuka; usually both rostral and maxillary barbels present in Devario), and “ventrals well developed” (long in Laubuka, not particularly so in Devario) also suggest a different genus than Devario. Günther’s description of the species E. ceylonensis excludes Devario by reference to absence of a symphyseal knob of the lower jaw, which is prominent in Devario and Chela, and silvery colour, not recorded from any Sri Lankan Devario. The species description, however, contains a statement that is difficult to reconcile with Laubuka or Devario: “Pectoral fin shorter than the head, extending to the ventral.” This statement also seems to be at odds with the diagnosis of Eustira, which suggests a long pectoral fin, even if “elongate” could also be understood as slender. Other parts of the description of E. ceylonensis are compatible with both Laubuka and Devario, as well as other cyprinid genera. The anal-fin count of 17 is low for Laubuka, which usually have about 20 anal-fin rays, but it may exclude the first two unbranched rays, which are difficult to discern without manipulation.

The jar with the syntypes of Eustira ceylonensis could not be located in the BMNH collection (J. Maclaine, R. Britz, pers. comm.). The specimens incorrectly considered as the type series of Eustira ceylonensis are catalogued as BMNH 1852.2.19.130–132, 1853.12.24.6, 1864.4–11.33, and are identified by us as tentatively representing Devario malabaricus sensu Pethiyagoda (1991). Eustira ceylonensis is probably a senior synonym of one of the species of Laubuka reported by Pethiyagoda et al. (2008) from Sri Lanka, but Günther’s description does not contain information that can be used to identify which of those species is affected.

Although the Myanmar and Bangladeshi samples of Laubuka tenella could not be separated by morphological characters, it may be noted that meristics and proportional measurements are very conservative in Laubuka. The colour pattern is unique to L. tenella but similar to that of other species. A longitudinal dark stripe occurs in several species, although usually narrower than in L. tenella, and short vertical bars anteriorly on the side is illustrated for L. insularis and L. varuna in Pethiyagoda et al. (2018). The combination of a posterior dark stripe and anterior series of short vertical bars is unique for L. tenella. The colour pattern is definitely different from that of L. fasciata and L. parafasciata in which there is a horizontal dark stripe along the middle of the side (Knight 2015: fig. 15; Lalramliana et al. 2017: fig. 2). Specimens referable to L. laubuca have plain sides except for a distinct dark cleithral spot, a very narrow dark stripe posteriorly on the caudal peduncle and, in juveniles only, a dark spot at the base of the caudal fin. In L. siamensis there are indistinct vertical bars anteriorly on the side, and posteriorly on the side a dark stripe ending in a triangular blotch on the caudal-fin base.

Lalramliana et al. (2017) considered 16 precaudal vertebrae to be diagnostic for Laubuka parafasciata, contrasting with 14 in other Laubuka. The count of 14 was probably based on Pethiyagoda et al. (2008) who reported 14 precaudal vertebrae in Laubuka as diagnostic from Chela cachius with 17. Our specimens of Laubuka possess 15 or 16 vertebrae anterior to that with the first distinct haemal spine (Table 3). Typically, vertebrae 5–14 bear ribs, vertebrae 15 is a short transitional vertebra without ribs or distinct haemal spine, vertebra 16 presents a haemal spine, and vertebra 17 has a haemal spine inserted posterior to the ascending first anal-fin pterygiophore (Fig. 8) There are no major differences in vertebral counts between species of Laubuka, as also suggested by preliminary data from a few species (Table 3).

Figure 8. 

X-radiograph of Laubuka tenella. Significant vertebrae indicated with sequential number. 5 = Fifth precaudal vertebra (= 5^th abdominal/trunk vertebra); 14 = 14^th precaudal vertebra, anterior of two precaudal vertebrae with haemapophyses and non-articulating ribs/ribs absent (intermediate vertebrae of Naseka 1996); 15 = last precaudal vertebra; 16 = first vertebra with haemal spine, which is first caudal vertebra of Günther (1880), Naseka (1996), Hubbs and Lagler (2007), and Bird and Mabee (2003), and the last preanal vertebra of Naseka 1996).

The terminology of vertebral elements in teleosts is diverse, with confusing synonymy. Günther (1880, 1886) and Hubbs and Lagler (2007; and editions from 1947 onwards) provided the most influential list of definitions of measurements and counts in fish taxonomy. Günther distinguished between abdominal and caudal vertebrae, and Hubbs and Lagler between precaudal and caudal vertebrae, both publications considering the first caudal vertebra to be that bearing a definite haemal spine. Günther (1880) stated more precisely: “the caudal vertebrae differ from the abdominal in having the haemapophyseal elements converted into spines similar to the neurals”.

Naseka (1996) proposed a vertebral column formula recognizing the main divisions of abdominal and caudal vertebrae. Naseka’s abdominal vertebrae include intermediate vertebrae which are vertebrae with parapophyses but no articulating ribs. His caudal vertebrae include preanal vertebrae with a haemapophysis anterior to the first (or first interhaemal) anal-fin pterygiophore, and postanal vertebrae, which are those posterior to the first anal-fin pterygiophore. Bird and Mabee (2003) mapped vertebral elements in Danio rerio, distinguishing between Weberian (within the Weberian complex), precaudal (from vertebra 5 to the one before that bearing the first haemal spine, “…vertebrae composed of centra, neural arches and spines, parapophyses, and ribs”), caudal (from that with the first haemal spine to the one before the caudal-fin vertebrae, “…vertebrae […] composed of centra, neural arches and hemal spines”), and caudal-fin (those supporting the caudal-fin rays) vertebrae.

In cypriniforms, there are always four vertebral centra within the Weberian apparatus; posterior vertebrae may vary in number and morphology. In Laubuka the Weberian vertebrae are followed by trunk vertebrae with long pleural ribs and usually three vertebrae with no ribs attached. Of these latter three, the anterior two have parapophyses, a haemal arch, and non-articulating free pleural ribs, whereas the third is slender and extends ventrad as a haemal spine immediately anterior to the long first anal-fin pterygiophore or haemapophysis (Fig. 8). On X-radiographs of fish specimens, the shape of the posterior trunk vertebrae commonly cannot be decided, but, as in Laubuka, their identity is deduced from the relative position, presence or absence of attached ribs, and the relative length of the haemapophysis/haemal spine. Roberts (1989), working with X-radiographs, suggested to count as caudal vertebrae those posterior to the first anal pterygiophore. This method was followed by e.g., Kullander and Britz (2002), and probably also Lalramliana et al. (2017).

Differences between authors in reporting vertebral counts can be attributed to different interpretations or definitions of the vertebrae in the transitional zone between trunk and tail vertebrae. The precaudal count in Pethiyagoda et al. (2008) apparently refers to the Weberian plus rib-bearing vertebrae, whereas in Lalramliana et al. (2017) it apparently refers to the Weberian plus remaining trunk vertebrae including the vertebra with the first haemal spine, i.e., all preanal vertebrae (but not abdominal or precaudal vertebrae as in Günther (1880) or Hubbs and Lagler (2007).

Vertebral numbers and the delimitation of trunk and tail vertebrae are useful taxonomic markers and may explain body proportions. Limitations in resolution of X-ray images or difficulties to determine the first haemal spine may influence the classification of trunk and tail vertebrae, however. It seems relevant in danionin cyprinids to follow Bird and Mabee (2003) in distinguishing between precaudal vertebrae (i.e. the ones anterior to the vertebra immediately anterior to the first anal-fin pterygiophore, understood as bearing the first haemal spine, but in contrast to Bird and Mabee including the four Weberian vertebrae), and caudal vertebrae (all vertebrae posterior to and including this vertebra, including caudal-fin supporting vertebrae, i.e., the the last half-centrum= compound centrum or preural+ural1) (Fig. 8). This division agrees with that of Günther (1880, 1886) and Hubbs and Lagler (2007). In groups in which the shape of transitional trunk vertebrae (Naseka’s intermediate vertebrae) is difficult to classify, some other method to count or to distinguish the vertebrae may be more practical. Because the subdivisions are based on convention only, any other method than those described above are equally valid, but the particular method of counting vertebrae needs to be explained for every instance to ensure that counts are consistent and comparisons achievable.

In the phylogenetic analysis, Laubuka tenella came out as sister group to L. laubuca, which has a complementary geographical distribution west of L. tenella. Only few sequences of Laubuka were available from GenBank to supplement our material from India and Bangladesh, and among them several lacked locality information (Fig. 4). Incorporation of additional taxa, e.g., from Sri Lanka, Myanmar and Indochina may alter significantly the pattern shown in Fig. 4. Nevertheless, the position in the tree (Fig. 4) and the 9 % p-distance difference from the most similar species, are strong indicators of species distinctness of L. tenella. One of the downloaded GenBank sequences, KT353103, from northern Bangladesh, and identified as L. laubuca in GenBank, is resolved in the L. laubuca clade (Fig. 4), but as sister to the remaining specimens. A possible explanation is that it represents one more species of Laubuka in the region, but could also reflect wide genetic variation in L. laubuca.

Laubuka tenella is known so far only from three localities, and DNA data are available from only two localities, representing two distinct haplotypes. It seems plausible to consider the localities to be part of a continuous distribution, but the differences in the COI sequence between the two Bangladeshi localities may be an indication of fragmentation. The coasts of extreme southeastern Bangladesh and southwestern Myanmar are ichthyologically underexplored. The Rakhine Yoma is a hill range separating the fish fauna in the lowlands to the west from that of the Irrawaddy basin, exemplified by several endemic species from western Rakhine Yoma (e.g., Kullander 2015, Kullander and Britz 2015). This lower costal region extends into southeastern Bangladesh, and it is not surprising to find the same species both in southwestern Myanmar and southeast, although L. tenella may be the first species reported as endemic for this particular region. Devario anomalus, in the Cox’s Bazar district, and D. xyrops Fang & Kullander, 2009, on the Western slope of the Rakhine Yoma, are two very similar sister taxa, which are considered as distinct species (Kullander et al. 2017), but with a combined distribution similar to that of D. tenella.

Specimens were collected with the assistance of Abdullah Helal, Mozammel Hossain, Suman Mandal, and Mamunur Rashid, University of Dhaka. Research on Bangladeshi freshwater fishes was supported by the project “Genetic characterization of freshwater fishes in Bangladesh using DNA barcodes” funded by the Swedish Research Council, contract D0674001 to Sven Kullander and Abdur Rob Mollah. Specimens of Laubuka parafasciata were put at our disposal by Ralf Britz, who also provided clarifying comments on morphology of vertebrae. James Maclaine and Ralf Britz searched for the types of Eustira ceylonensis.