Abstract.-Squirrelfish of the genus Myripristis are valued in Reproductive and trophic ecology small-scale fisheries throughout much of the tropics. The life his­ of the soldierfish Myripristis tory and species biology of most of these soldierfishes is poorly known. amaena in tropical fisheries For the brick soldierfish, M. amae­ na, in Hawaii and Johnston Atoll, we found that sexual maturity for Anderson J. Dee both sexes was reached between p.o. Box JJ54 145 and 160 mm standard length Hayward. Wisconsin 54843 at about six years ofage - a large fraction ofthe apparent maximum size and lifespan. Fecundity was James D. Parrish* relatively low and increased as the fifth power ofbody weight. Spawn­ National Biological Survey. Hawaii Cooperative Fishery Research Unit ing peaked from about early April 2538 The Mall. University of Hawaii, Honolulu, Hawaii 96822 to earlyMay, and a secondary peak occurred in September. Myripristis amaena is a nocturnal predator, feeding mostly on meroplankton, especially brachyuran crab mega­ lops, hermit crab larvae, and Soldierfish, Myripristinae, of the reported <Dee and Radtke, 1989). shrimps, but also taking a variety squirrelfish family, Holocentridae, There have been no thorough pub­ of benthic prey. In pristine fish occur widely throughout the trop­ lished studies ofthe reproduction of communities, holocentrids were ics (Greenfield, 1965, 1968, 1974). M. amaena or closely related spe­ abundant, quantitatively impor­ tant (often dominant) reef preda­ They are typically abundant and cies. Because of the wide distribu­ tors and prey. Myripristis amaena are an important component ofcom­ tion, considerable abundance, and (and probably other common and mercial, recreational, and subsis­ substantial fishery importance of important soJdierfishJ seems to be tence fisheries in much of the M. amaena, we undertook to de­ relatively long lived (at least 14 world's tropics. Throughout most of scribe more fully its food require­ years), slow growing, and late ma­ turing. The populations suffer con­ the central and western Pacific ments and trophic position in the siderable natural predation and Ocean, the brick soldierfish, Myri­ community and to quantify the re­ depend mainly on the largest and pristis amaena (Castlenau), is an productive characteristics that af­ oldest fish for reproduction. Heavy, important member of this group fect the dynamics ofits populations. unregulated fishing of these (Greenfield, 1968). It contributes The parameters we determined may soldierfish, especially at prerepro­ ductive size, may severely reduce significantly to fish communities also provide reasonable first ap­ populations. and to fishery catches in shallow proximations for similarMyripristis reef and rocky habitats. It is par­ species that are less well studied. ticularly important in the recre­ The results will contribute to an ational fishery at Johnston Atoll informed approach to management (JA), where it is typically the spe­ ofspecies that are now typically un­ cies caught in greatest abundance managed and probably overfished (Irons et a1. 1). It is also common in in most localities with even moder­ catches throughout the Hawaiian ately dense human populations. archipelago. The JA population ofM. amaena Relatively little quantitative in­ was the major focus ofthis studyfor formation hasbeen published about several reasons. Many biological the life history and biology of spe­ and ecological characteristics ofM. cies of the genus Myripristis, and amaena (e.g. size, morphology, hab­ very little is available about M. its, habitat used, fishery value) amaena in particular. Data about seem to be representative ofa num- diets are available for only a few species of Myripristis; for most of * Th whom correspondence should be sent. these, sample sizes are small (e.g. I Irons, D. K.. R. K. Kosaki, and J. D. 14 specimens for M. amaena; Parrish. 1990. JohnstonAtoll resource sur­ Hobson, 1974). Results on age and vey. Final report of Phase Six /21 July 1989-20 July 1990). Project rep. to U.S. Manuscript accepted 30 December 1993. growth from our studies of the JA Army Engineer Distsrict. Honolulu, HI, Fishery Bulletin 92:516-530 (1994). population ofM. amaena have been 150 p. 516 Dee and Parrish: Reproductive and trophic ecology of Myripristis amaena 517 ber of common soldierfish species. Johnston Atoll est 0.1 g. (Appendix A provides functions, fitted by provided a logistically good base for study, where M. regression, to convert between SL, FL, and TL.) amaena was the dominant Myripristis species, with Whole guts and gonads were excised and preserved populations not seriously depressed by fishing. The in 10% buffered formalin for further analysis. Some species was plentiful and easily collected at many specimens were frozen whole and stored for some locations throughout the year, and a good range of weeks or months before processing. sizes was readily available. The importance of M. amaena in the fishery at JA facilitated collection of Source of size-frequency data specimens and fishery data. AtJA, Puako, and the NWHI, length andweight data from all specimens collected for other purposes were Materials and methods available for size-frequency estimation. In addition, at JA, creel census data were obtained from fisher­ Specimen collection and handling men's catches on many days over the period ofstudy. Fish landed by boat and shore fishermen were ex­ Most specimens used in all analyses were taken at amined, and each specimen was measured and JohnstonAtoll, a remote, coral-rich, oceanic pinnacle weighed as above. These data provided a much larger about 1250 kIn SW ofHonolulu (Halstead and Bun­ and possibly more representative sample than our ker, 1954; Amerson and Shelton, 1976; Randall et collections alone. aI., 1985; Maragos and Jokiel, 1986). Smaller collec­ tions were made from a rich, fringing coral reeftract Feeding at Puako in South Kohala on the leeward coast of Hawaii Island (Hayes et aI., 1982). A few specimens The volume ofeach complete gut specimen was mea­ were collected from the almost uninhabited North­ sured by displacement ofwater before and after the western Hawaiian Islands (NWHI), primarily from gut was opened and all contents removed. The total shallow, coralline areas at French Frigate Shoals and volume ofcontents was determined by difference. All Midway (located about 750 and 2000 km, respectively, diet items were sorted into systematic categories and northwest ofHonolulu). identified to the lowest possible taxa. For each prey Specimens from all three locations contributed to category, the number of individuals, length, extent size-frequency analysis, visual assessment ofgonad of digestion, location in gut, and volume were re­ condition (maturity), and analysis of gut contents. corded. Whole reference specimens were used as an At JA and Puako, gonad weight was taken to deter­ aid in identifying prey items and estimatingthe origi­ mine reproductive season and size at first reproduc­ nal size ofthe prey by comparison ofthe dimensions tion; at JA, gonad samples were preserved for histo­ ofrecognizable parts. Volume ofremains ineach prey logical examination and estimation offecundity. category was estimated by displacement of water Most specimens at all locations were collected from <Wolfert and Miller, 1978). shallow waters «15 m). The major methods of col­ A measure ofoverall importance ofeach prey cat­ lection were spearing, bait casting with a line, and egory was calculated by using the Index of Relative some spot applications ofichthyocide. At JA, we col­ Importance (IRI>, as defined by Pinkas et al. (1971): lected specimens from several sites inside the lagoon andjust outside the barrier reef. Collections and ex­ IRI = frequency % x (numerical % + volume %), tensive underwater observations were made at fre­ quent intervals between February 1984 and Janu­ where frequency % = (the number ofguts contain­ ary 1986. Sampling was less intensive in some ing prey of one category di­ months because of constraints on travel to JA. At vided by the total number of Puako and in the NWHI, specimens were collected guts that contained any iden­ rather widely within coralline habitats. At Puako, tifiable prey) x 100; about half the specimens were collected as quickly numerical % = (the number ofindividuals of as feasible (May-June 1981), once the species was one prey category divided by found to be in reproductive season. Other collections the total number of prey in­ there were distributed throughout the year. In the dividuals found in all the NWHI, specimens were collected in March, April, guts) x 100; May, August, and November. and volume % = (the volume of one prey cat­ Standard (SU, fork (FL), and total (TL) lengths of egory divided by the total all captured specimens were measured to the near­ volume of all prey found in est millimeter, and weights were taken to the near- the guts) x 100. 518 Fishery Bulletin 92(31. 1994 Reproduction ties ofeggs (Khoo, 1978; Wallace and Selman, 1981). Testes from five males were examined histologically Gonads of 430 specimens from JA were visually ex­ for presence of sperm. amined macroscopically and classified (based ontheir Spawning season was estimated by plotting GSI size, color, texture, and morphology) as male, female, against month of capture for 99 sexually mature immature, or unknown. Asubsample ofthese gonads males and females (largerthan 145 mm SL) collected from specimens collected during probable reproduc­ throughout the samplingperiod. Histological sections