Stigmatonotus australis, Gill, 2023

Gill, Anthony C., 2023, Taxonomic identity of the enigmatic Western Australian fish species Stigmatonotus australis Peters, 1877 (Teleostei), Zootaxa 5296 (4), pp. 551-560 : 553-557

81. Stigmatonotus australis n. sp.

B. 6. D. 11, 17; A. 3, 8. V. 1, 5. L. lat. 46.

Height to length (without caudal fin) as 1: 3, head length to the same as 1: 2½. The beginning of the dorsal fin is behind the base of the pectoral fin; the spiny part, consisting of 11 rays, is lower than the soft-rayed part. The lateral line rises in a strong arc up to the dorsal fin, and is separated in the middle by only two scales, and from the beginning of the anal fin by 13 scales. Between head and caudal fin 46 scales.

Colour brown, with a light band descending from the midline of the body behind the base of the pectoral fin to the middle of the abdomen.

A single specimen, 17 millimeters long, 3 fathoms deep from Dick-Hartog (Western Australia).]

Identification and classification of Stigmatonotus australis

The following is a summary of key characters and other evidence pertinent to the identification and classification of S. australis , leading to the conclusion that it is referable to the species currently called Hypoplectrodes cardinalis Allen & Randall, 1990 . This information is presented in two sections, one of evidence that directly support this identification, and the other of characters that are apparently in conflict (and thus require explanation).

Evidence directly supporting identification to Hypoplectrodes cardinalis . Evidence in support of identification to H. cardinalis includes the following.

Additional spine below primary spine on opercle. This character has been noted as a synapomorphy of the Serranidae ( Johnson 1983) , although it has a wider distribution among acanthomorphs (e.g., Imamura 1996: fig. 1). The character is sometimes expressed numerically as presence of three opercular spines (consisting of the primary spine, a dorsal spine, and the synapomorphic ventral spine), but this definition is problematic as the upper spine may be vestigial or absent in some anthiadine species. The lower spine is consistently present in anthiadines, although sometimes additional spines or serrations may be present ventral to it (particularly in larvae and small juveniles). Peters (1877) illustrated three opercular spines in the holotype of S. australis , with a spine ventral to the primary spine.

Three supraneurals, with one in preneural space and two in first interneural space. Anthiadines have 0–3 supraneural bones. With the exception of certain species of Acanthistius Gill, 1862 (which genus is unlikely to be correctly placed in the Anthiadinae ), in anthiadines with three supraneurals, the posterior supraneural is in the first interneural space ( Pogonoski & Gill 2021). This contrasts with the plesiomorphic condition of having the third surpraneural in the second interneural space, which is characteristic of non-anthiadine serranids that have three supraneurals ( Kendall 1976). The configuration of the supraneurals is evident though difficult to see in the X-radiograph of the holotype. Aside from anthiadines, this derived arrangement of supraneurals occurs more widely among acanthomorph fishes. Johnson’s (1984) summary of osteological and other features of percoid fishes indicates the arrangement occurs in at least 17 other fish families that occur in tropical or subtropical Western Australia. However, all of these can be disregarded as candidate identifications for Stigmatonotus because they differ in other features (such as number of vertebrae or presence of the ventral opercular spine). Western Australian anthiadine genera with this arrangement include Caesioperca Castelnau, 1872 , Caprodon Temminck & Schlegel, 1843 , Epinephelides Ogilby, 1899 , Hypoplectrodes Gill, 1862 , Lepidoperca Regan, 1914 , Mirolabrichthys Herre, 1927 , Othos Castelnau, 1875 , and Selenanthias Tanaka, 1918 , and in some species of Plectranthias Bleeker, 1873 , Pseudanthias Bleeker, 1871 , and Tosanoides Kamohara, 1953 .

10+17 vertebrae. The holotype of Stigmatonotus australis has 10+17 vertebrae ( Fig. 3 ). Most serranid fishes have 10+14–15 vertebrae, whereas anthiadine species typically have 10+16. Anderson & Heemstra (1989) considered the increase in number of vertebrae to be synapomorphic for the Anthiadinae , and that species of Hypoplectrodes are further derived within the subfamily in having 10+17–18. Gill et al. (2021) provided a more extensive review of anthiadine vertebral counts, and noted that, aside from rare specimens of certain species of Plectranthias and Caprodon , a count of 10+17 vertebrae is also typical of the monotypic genera Othos and Epinephelides . Both latter genera occur in Western Australia but differ notably from the holotype of S. australis , particularly in having much smaller and more numerous scales on the body (see below).

Ornamentation of preopercle. The original description of S. australis merely states that the preopercle is serrated (“Vordeckel gezähnelt”), but the illustration shows the vertical section of the bone finely serrated, a slightly enlarged serration at the angle of the bone, and two ventrally directed serrations on the ventral section ( Fig. 1 ). These structures are visible in the photograph of the holotype, although difficult to discern ( Fig. 2 ). This configuration is typical of juveniles of certain anthiadine species ( Fig. 4 ); in adults of those species, the lower serrations and sometimes the angle serration become more prominent and are directed anteriorly (antrorse). Antrorse preopercular serrations are present in Hypoplectrodes , Epinephelides , Othos (in which only a single small serration is present), Acanthistius and some species of Plectranthias . There are three antrorse serrations in H. cardinalis , which correspond in position with the angle and lower two serrations in the holotype of S. australis .

Anal fin with eight soft rays. Most anthiadine species have seven soft rays in the anal fin.Although the presence of eight soft rays is far from unique in the subfamily, it is relatively unusual. It is characteristic of most Hypoplectrodes species, including H. cardinalis .

Number of scales in lateral series. Peters (1877) recorded 46 scales between the head and the caudal fin.Although it is not appropriate to equate this count directly with counts of lateral-line scales (in species with a complete lateral line), the two counts do approximate each other and the count of 46 allows some candidate anthiadine species to be dismissed from consideration. This includes Hypoplectrodes nigroruber (Cuvier in Cuvier & Valenciennes, 1828) (with 55–65 lateral-line scales), Epinephelides (with about 70 lateral-line scales), Othos (with 81–87), and the two local species of Acanthistius (collectively with 54–67). Hypoplectrodes cardinalis and H. wilsoni ( Allen & Moyer, 1980) have counts that more closely approach S. australis (41–45 and 44–48, respectively).

Type locality. Three species of Hypoplectrodes are known from Western Australia, but only two occur as far north as Shark Bay. Hypoplectrodes cardinalis occurs from the Recherche Archipelago in the south ( Hutchins & Swainston 1986) to 23.85°S, 113.35°E in the Gascoyne Marine Park off Lake Macleod (J.J. Pogonoski, pers. comm.), and H. nigroruber occurs around southern Australia north to Shark Bay in Western Australia (G. Moore, pers. comm.). Although H. nigroruber occurs in Shark Bay, it can be nonetheless discounted as a candidate identification for S. australis based on its smaller, more numerous body scales (see above). The third Western Australian species, H. wilsoni , is restricted to south-western Western Australia, from Lucky Bay to Houtman Abrolhos (G. Moore, pers. comm.).

Characters in apparent conflict with identification to Hypoplectrodes cardinalis . Characters that appear to conflict with identification to H. cardinalis include the following.

Number of dorsal-fin rays. Peters (1877) gave a count of XI,17 dorsal-fin rays, but his illustration shows XII,17. Unfortunately, damage to the specimen does not allow direct, conclusive evaluation of either of these counts, but the correct number of spines can nevertheless be inferred indirectly. The body is broken in two parts, with the break between the 8 th and 9 th vertebrae, and pterygiophores and their associated rays missing from the 7 th and 8 th interneural spaces. An accurate count of the soft (segmented) rays is not possible from the available radiograph, but the position of the first soft-ray-bearing pterygiophore in the 9 th interneural space is evident. This is consistent with the typical arrangement of anthiadines (and many other fishes) with ten dorsal-fin spines, with the first pterygiophore in the second interneural space bearing two supernumerary and one serial spine, two pterygiophores in the third interneural space, each bearing a single serial spine, and a single pterygiophore in the 4 th through 8 th interneural spaces, each bearing a single serial spine. Generally, in fishes with more than 10 spines, the pterygiophore serially bearing the 11 th spine is in the 9 th interneural space (and if more than 11 spines are present, such as in Acanthistius species, subsequent spine-bearing pterygiophores follow, one per interneural space in subsequent spaces), and the position of the first soft-ray-bearing pterygiophore is displaced one or more spaces posteriorly. Given the position of the first soft-ray-bearing pterygiophore in the 9 th interneural space in the holotype of S. australis , I therefore argue that it had 10 spines rather than 11 or 12. It is likely the anterior soft rays may have been confused as spines by Peters and his illustrator. If so, the total number of spines and rays in the illustration (29) is consistent with H. cardinalis (ten spines and 19 soft rays).

Six branchiostegal rays. Peters (1877) reported that the holotype has six branchiostegal rays, one fewer than in species of Hypoplectrodes . However, the anterior branchiostegal ray is small and often difficult to reveal from the branchiostegal membranes and other surrounding tissue and can be thus easily overlooked (particularly in specimens as small as the holotype). I therefore suggest that the count in the original description is erroneous. Unfortunately, damage to the specimen means this can no longer be verified.

Truncated lateral line. Peters (1877) described the lateral line as incomplete (terminating posteriorly beneath the posterior part of the dorsal fin), which contributed to classification of the species in the Pseudochromidae or Grammatidae (species of which have truncated or disjunct lateral lines). He did not provide a count of tubed scales, but there are 35 such scales evident in his illustration ( Fig. 1 ). Hypoplectrodes species have a complete lateral line, which extends to the caudal-fin base, and consists of 41 or more tubed scales. I suggest the incomplete nature of the lateral line in the holotype results from ontogenetic variation and the small size of the specimen, that the lateral line is not fully developed in the specimen. Appropriately sized specimens of H. cardinalis were not available for comparison to verify this. However, small specimens of other Hypoplectrodes species were located in the Australian Museum (AMS) collection. Notably, a 14.5 mm SL specimen of an unidentified Hypoplectrodes from New South Wales (AMS I.45405-001) has an incompletely developed lateral line composed of tubed (anterior few scales) or open-furrowed (remaining scales), which terminates beneath the posterior part of the dorsal fin. The lateral line is fully developed in an 18.5 mm SL specimen of H. maccullochi ( Whitley, 1929) (AMS I.45022-013), terminating at the hypural joint. Similarly, G. Moore (pers. comm. 2023) relayed that the lateral line is fully developed in a 17.9 mm SL specimen of H. cardinalis , the smallest specimen of the species in the Western Australian Museum collection.

Absence of transforming cteni. Peters (1877) noted that S. australis has ctenoid scales and provided an illustration of a representative scale. The illustrated scale has a single row of large cteni. This contrasts with reported squamation for Hypoplectrodes , in which scales bear additional basal rows of ctenial bases (transforming cteni; see Roberts 1993). However, this conflict can be attributed to the small size of the holotype, as transforming cteni are absent in juveniles. Moreover, transforming cteni may be restricted in distribution on a fish, and are often confined to the anterior part of the body in anthiadines ( Gill et al. 2021); therefore, the apparent absence also may be a consequence of where on the body the illustrated scale was sampled.

Number of pectoral-fin rays. Peters (1877) did not provide a count of pectoral-fin rays in the original description, but the illustration depicts 17 rays. This is one more than is known for H. cardinalis (with 15–16 rays). However, I count only 15 or 16 rays in the photograph and radiograph of the holotype.

Nomenclatural considerations

The conclusion that Stigmatonotus australis is conspecific with Hypoplectrodes cardinalis raises several nomenclatural issues, notably the need for synonymisation of the family-group name Stigmatonotidae Whitley, 1954 with Anthiadides Poey, 1861 (as already noted by van der Laan et al. 2014), the generic name Stigmatonotus Peters, 1877 with Hypoplectrodes Gill, 1862 , and the species name H. cardinalis Allen & Randall, 1990 with S. australis Peters, 1877 . In the interest of nomenclatural stability, an argument could be made for maintaining Allen & Randall’s name for the species. However, justification of this outcome (reversal of precedence) requires satisfying conditions outlined in article 23.9 of the International Code of Zoological Nomenclature ( International Commission on Zoological Nomenclature 1999), including that the senior name has not been used as the valid name for a species after 1899. Notwithstanding its absence of mention in recent checklists of Australian fishes (e.g., Allen 1985; Paxton et al. 1989; Hutchins 2001; Hoese et al. 2006), S. australis has been used as the valid name for a species in various post-1899 publications, including McCulloch (1929), Whitley (1954, 1964), Scott (1959), Böhlke (1960) and Nelson (1976, 1984). Because Stigmatonotus has been considered monotypic, that list could be extended to include publications that only mention the generic name (e.g., Springer 1982; Nelson 1994, 2006). Therefore, Hypoplectrodes australis is adopted as the valid name for the species. A synonymy for the species is given below.

Hypoplectrodes australis ( Peters, 1877)

Stigmatonotus australis Peters, 1877: 838 (type locality: Dirk Hartog Island , Western Australia).

Ellerkeldia rubra Allen, 1976: 24 , fig. 1 (type locality: off Beacon Island, Wallabi Group, Houtman Abrolhos Islands, Western Australia; secondarily preoccupied by Scopularia rubra de Buen, 1959 when both species in Hypoplectrodes ).

Hypoplectrodes cardinalis Allen & Randall, 1990: 45 (replacement name for Ellerkeldia rubra ).

Figures

FIGURE 1. Original illustration of the holotype of Stigmatonotus australis Peters (after Peters 1877; image digitally reversed).

FIGURE 2. Stigmatonotus australis Peters, holotype, ZMB 9811, ca 14 mm SL, as photographed in 2022. Photograph by E. Assel; copyright Museum für Naturkunde Berlin; used with permission.

FIGURE 3. Stigmatonotus australis Peters, holotype, ZMB 9811, ca 14 mm SL, 2022 X-radiograph. X-radiograph by E. Assel; copyright Museum für Naturkunde Berlin; used with permission. Note that the posterior portion of the specimen has been accidentally inverted.

FIGURE 4. Hypoplectrodes sp., AMS I.45405-001, 14.5 mm SL, lower part of right side of head (image digitally reversed). Arrows point to serrations on the preopercle; scale bar = 1 mm. Photograph by A.C. Gill.