Bruce: Larval development of Sillagmodes punctata, Sillago bassensis, and Sillago schomburgkn 



41 



anophores in the vicinity of myomeres 24-26. Lat- 

 eral midline pigment spreads both anteriorly and 

 posteriorly as a linear series of elongated myomeres 

 during development. By 10.1 mm, lateral midline 

 pigment consists of 18 stellate and approximately 

 evenly spaced melanophores extending from the pec- 

 toral fin to the caudal peduncle. Internal pigment 

 along the vertebrae is visible in the 10.1-mm post- 

 flexion larva but is most pronounced in post- 

 settlement larvae as clusters of melanophores located 

 over every 2-5 vertebrae. 



A single row of 16-18 melanophores is present 

 along the ventral midline of the tail in preflexion lar- 

 vae. This ventral row is gradually obscured by over- 

 lying musculature during flexion. Paired external 

 melanophores (approximately one pair per myomere) 

 subsequently develop ventrally on the tail in post- 

 flexion larvae, approximately one per myomere. Two 

 to three (most commonly three) melanophores are 

 present ventrally on the notochord tip in preflexion 

 larvae. These are retained in postflexion larvae and 

 form a band of pigment over the caudal-fin ray bases. 



Discussion 



Egg or larval development, or both, have been de- 

 scribed for only four other species of sillaginid lar- 

 vae: Sillago japonica (Kamiya, 1925; Ueno and 

 Fujita, 1954; Ueno et al., 1958; Mito, 1966 — as Sil- 

 lago japonicus; Ikeda and Mito, 1988; Kinoshita, 

 1988; Oozeki et al., 1992); Sillago sihama (Gopinath, 

 1946; Uchida et al., 1958; Ikeda and Mito, 1988; Kino- 

 shita, 1988); Sillago maculata (Miskiewicz, 1987; 

 Kinoshita, 1988); and Sillago ciliata (Munro, 1945; 

 Miskiewicz 1987; Tosh 4 ). In addition, Miskiewicz 

 (1987, p. 62) reported a series of unidentified 

 sillaginid larvae which, based on pigment on the lat- 

 eral wall of the gut below the pectoral fin base, were 

 almost certainly Sillago flindersi. 



Characters useful for the identification of tropical 

 sillaginid larvae at the family level and similarity of 

 sillaginid larvae to those from other families have 

 been considered in detail by Leis and Trnski (1989) 

 and Miskiewicz ( 1987). Although most of the charac- 

 ters discussed by these authors also apply to the tem- 

 perate species considered here, an exception was the 

 timing of gut coiling. Leis and Trnski ( 1989) reported 

 that the gut of tropical sillaginid larvae commenced 

 coiling during notochord flexion and was accompa- 

 nied by the anterior migration of the anus. In the 



South Australian species, coiling of the gut com- 

 menced prior to flexion in S. bassensis, after settle- 

 ment in Sillaginodes punctata, and had not yet com- 

 menced in the largest flexion larva available for 

 Sillago schomburgkii (although coiling of the gut was 

 present in a 10. 1-mm postflexion larva ). In all cases, 

 coiling of the gut proceeded without migration of the 

 anus and was achieved by anterior looping of the 

 midgut. The implications of these variations are un- 

 clear but suggest that, although useful on a specific 

 level, the timing of gut coiling and migration of the 

 anus are not useful characters for the identification 

 of temperate sillaginids at the family level. 



The significance of gut coiling may relate to shifts 

 in diet. Robertson (1977) reported a dietary shift in 

 postsettlement Sillaginodes punctatus (-punctata) 

 in Westernport Bay (Victoria) between November and 

 December, a shift from harpacticoid copepods, 

 gammarid amphipods, and mysids to larvae of the 

 ghost prawn Callianassa australiensis, polychaetes, 

 and juvenile crabs. Robertson correlated this dietary 

 shift with increasing body size and mouth gape as 

 well as with the availability of C. australiensis lar- 

 vae. However, from his length-frequency data, this 

 period also corresponds to the size range during 

 which postsettlement S. punctata undergo gut coil- 

 ing. Alternatively, because evacuation rates are be- 

 lieved to decrease after gut coiling (Arthur, 1976, and 

 references within; Young 5 ), perceived changes in diet 

 may be confounded by increased food retention times. 

 Stomach contents were not analyzed during this 

 study; they provide a valuable topic for further re- 

 search. 



Despite seasonal sampling over five years, larvae 

 of only three of the four sillaginid species with adult 

 distributions extending to South Australia were lo- 

 cated during this study. The lack of Sillago flindersi 

 larvae suggests either that this species does not 

 spawn in South Australian waters, that sampling 

 frequency was too course to detect the presence of 

 larvae of this species, or that S. flindersi larvae be- 

 have differently from other sillaginid species and are 

 less prone to capture (e.g. epibenthic and neustonic). 



Sillago flindersi larvae are frequently encountered 

 in similar sampling regimes in coastal waters of east- 

 ern Australia 6 and in Tasmanian waters (author's 

 pers. observ.) and thus it seems unlikely that a lack 

 of their larvae in South Australian samples repre- 

 sents an artifact of sampling or that their behavior is 

 fundamentally different from other sillaginid larvae. 



4 Tosh, J. R. 1903. Notes on the habits, development etc. of the 

 common food fishes of Moreton Bay. Queensland Marine Dep.: 

 Marine Biologist's Report. 



5 Young, J. W. CSIRO Div. Fisheries, GPO Box 1538 Hobart, Tas- 

 mania, Australia 7001. Personal commun., 1993. 



6 Miskiewicz, A. G. Sydney Water Board, PO Box A53, Sydney 

 South, NSW, Australia 2000. Personal commun., 1993. 



