Redescription and distributional range extension of a poorly known anchovy Stolephorus nelsoni (Actinopterygii: Clupeiformes: Engraulidae)

Background. The poorly known anchovy Stolephorus nelsoni Wongratana, 1987 (Engraulidae), previously known only from the type specimens from Australia, is redescribed and its validity confirmed, on the basis of the holotype, paratype, and 15 additional specimens from Indonesia and Australia. Similar to Stolephorus brachycephalus Wongratana, 1983 in sharing the maxilla reaching more or less to the posterior border of opercle, a predorsal-fin scute absent and the preopercle rounded, the specific distinctiveness of S. nelsoni and differences between the two species have remained unclear. This study sought to confirm the validities of the two species and the distributional range of S. nelsoni. Materials and methods. Seventeen and 9 specimens of S. nelsoni and S. brachycephalus, respectively, including the holotype of each, were examined morphologically, counts and proportional measurements following Hata and Motomura (2017). Results. Comparisons of S. nelsoni with S. brachycephalus revealed the former to have 12–14 (mode 13) branchiostegal rays [vs. 10–11 (11)], 15–17 branched anal fin rays [vs. 20–22 (20)] and relatively fewer gill rakers [37–43 (39, 40) on the first gill arch vs. 35–37 (35)]. Stolephorus nelsoni also differed from S. brachycephalus in body depth [21.7%–23.7% (mean 22.6%) of standard length vs. 18.2%–20.3% (19.1%)], anal-fin base length [16.1%–19.3% (17.6%) vs. 21.6%–24.4% (23.2%)], caudal-peduncle length [15.7%–21.7% (18.8%) vs. 12.6%– 15.2% (14.1%)] and snout length [4.5%–5.5% (5.1%) vs. 3.6%–4.3% (4.0%)]. Conclusion. The validities of both species were confirmed, Indonesian examples of S. nelsoni being the first known records of the species outside Australian waters.


INTRODUCTION
The Nelson's anchovy, Stolephorus nelsoni Wongratana, 1987, based on four specimens from Western Australia and Queensland, Australia was originally described by Wongratana (1987), and confirmed as a valid species by Whitehead et al. (1988) and Wongratana et al. (1999). Paxton et al. (2006) listed the type specimens of S. nelsoni in their review of Australian engraulid fishes, there having been no additional specimens recorded. Stolephorus nelsoni is herein redescribed as a valid species, following examination of type and additional specimens from Australia and Indonesia, the latter representing the first records of the species outside Australian waters. Morphological differences between S. nelsoni and the closely related congener Stolephorus brachycephalus Wongratana, 1983 are also provided.

MATERIALS AND METHODS
Counts and proportional measurements followed Hata and Motomura (2017). All measurements were made with digital callipers to the nearest 0.1 mm. Standard and head lengths are abbreviated as SL and HL respectively. Institutional codes follow Sabaj (2016).
Counts and measurements, expressed as percentages of SL or HL, given in Tables 1 and 2. Body compressed laterally, elongate, deepest at dorsal-fin origin. Dorsal profile of head and body slightly convex from snout tip to dorsal-fin origin, straight along dorsal-fin base. Ventral profile of head and body slightly convex from lower-jaw tip to pelvic-fin insertion, slightly convex or straight from pelvic-fin insertion to anal-fin origin, more or less straight along anal-fin base. Dorsal and ventral profiles of caudal peduncle slightly concave. Belly rounded. Anus situated just anterior to anal-fin origin. Caudal peduncle compressed; depth slightly greater than orbit diameter. Snout tip rounded; snout length less than eye diameter. Mouth large, inferior, ventral to body axis, extending backward beyond posterior margin of eye. Lower jaw slender, shorter than upper jaw. Maxilla long, its posterior tip pointed, just reaching to posterior border of opercle (beyond or just short of posterior border of opercle in some specimens). Single row of conical teeth on each jaw and palatines. Small conical teeth in patch on vomer. Small tooth  10-12 11 Pelvic-fin rays (unbranched) 1 1 1 1 1 1 1 Pelvic-fin rays (branched) 6 6 6 6 6 6 6 Caudal-fin rays (upper + lower) 10 + 9 10 + 9 10 + 9 10 + 9 10 + 9 10 + 9 10 + 9 Gill rakers on 1st gill arch ( Western Australia to Mackenzie Island, Queensland; Wongratana 1987, Whitehead et al. 1988, Wongratana et al. 1999, Paxton et al. 2006 and in Irian Jaya, Indonesia (this study) (Fig. 2).

DISCUSSION
Stolephorus nelsoni was described by Wongratana (1987) based on four specimens collected from Western Australia and Queensland, Australia, the species being known to date only from the type specimens. Examination of additional material from the northern coast of Australia and Indonesia confirmed the validity of S. nelsoni, with the unique combination of characters described in the Diagnosis (above). Although Wongratana (1987) recorded S. nelsoni as having 20 or 21 anal-fin rays, modified in Whitehead et al. (1988) and Wongratana et al. (1999) as three unbranched rays and 20 branched rays, the anal-fin ray count observed in the presently reported study was iii, 15-17, the values indicated in Whitehead et al. (1988) and Wongratana et al. (1999) being clearly erroneous.
Although S. nelsoni is similar to S. brachycephalus (Fig. 3) in sharing the maxilla well beyond the posterior margin of preopercle, the latter rounded posteriorly, and the dorsal fin without scutes, the former has more gill rakers (24 in S. nelsoni vs. 20-22 in S. brachycephalus) and branchiostegal rays (12 or 13 vs. 10 or 11), and a tooth patch on the hyoid bone upper edge (vs. teeth absent), according to Wongratana et al. 1999. The present comparisons of S. nelsoni with S. brachycephalus, based on 17 and 9 specimens, respectively, confirmed the branchiostegal ray and hyoid bone tooth conditions described by Wongratana et al. (1999). However, the lower gill raker number in S. nelsoni was 20-24 (modally 22), a greater range than that reported by Wongratana et al. (1999). The comparable range for S. brachycephalus was 20 or 21, not 20-22 as shown by Whitehead et al. (1988) and Wongratana et al. (1999). Therefore, lower gill raker numbers cannot clearly separate the two species. However, S. nelsoni differs from S. brachycephalus in having relatively more total gill rakers on the first gill arch (37-43 vs. 35-37), and fewer branched anal-fin rays (15-17 vs. 20-22) and branched pectoral-fin rays (12-14 vs. 10-12;  Fig. 4). Notwithstanding, S. brachycephalus has been reported only from specimens smaller than 41.8 mm SL (Wongratana 1983, Whitehead et al. 1988, Wongratana et al. 1999, this study), which may or may not have been adult. If larger and smaller specimens of S. brachycephalus and S. nelsoni, respectively, become available, the differences shown here between S. nelsoni and S. brachycephalus may not hold up. The collection of S. nelsoni specimens from Indonesian and northern Australian waters suggests that the species is also likely distributed off the southern coast of Papua New Guinea.