Dee and Parrish: Reproductive and trophic ecology of Mynpristis amaena 



527 



distribution of reproduction. Instead, small collec- 

 tions of specimens were made at regular, frequent 

 intervals only long enough to discern a time of ac- 

 tive reproduction. At only that time, a large collec- 

 tion of specimens was made quickly to permit esti- 

 mation of SFR, and seasonal collections were not 

 continued (Hayes et al., 1982). Thus, the Puako re- 

 sults serve only to establish that reproductive devel- 

 opment (e.g. GSI, Fig. 5) is high in May and June. 

 Those months probably represent a peak, but the 

 data do not well define its limits or the pattern for 

 the rest of the year. This early summer high at Puako 

 is contiguous with the late spring high at JA. 



Many tropical marine species show a collective 

 spring spawning peak and a second peak in fall 

 (Munro et al., 1973; Watson and Leis, 1974; 

 Johannes, 1978; Walsh, 1987). For Hawaiian fishes, 

 the most dominant seasonal spawning pattern, based 

 on numbers of spawning records, is a peak spawn- 

 ing period in about April and May, with a secondary 

 peak in October for some species. Based on numbers 

 of recruitment records, the dominant recruitment 

 period occurs in June and July, and a secondary peak 

 in February and March (Walsh, 1987). Many studies 

 indicate that there can be considerable variability 

 in the timing of recruitment from year to year, and 

 that the timing and intensity may vary at small spa- 

 tial scales (Victor, 1982; Williams, 1983; Sale, 1985; 

 Schroeder, 1985; Walsh, 1987, Doherty, 1991). 



Larval and newly settled M. amaena were elusive 

 throughout the present study, and few data could be 

 collected regarding recruitment. The youngest speci- 

 men for which we counted short period (apparently 

 daily) increments in otoliths (Dee and Radtke, 1989) 

 showed a discontinuity that probably represented 

 settlement from the plankton. Back calculation based 

 on the number of increments after the assumed 

 settlement mark suggests that the specimen settled 

 in early February. Although our data regarding 

 settlement are minimal, both these and our spawn- 

 ing results are consistent with the above seasonal 

 reproductive periods summarized by Walsh (1987). 

 Walsh suggested that changes in water temperature 

 or photoperiod, or both, are most likely responsible 

 for observed seasonal patterns of spawning and re- 

 cruitment in Hawaiian reef fishes. For M. amaena 

 at JA, there was no indication that water tempera- 

 ture affected the time of spawning. The full annual 

 range of temperature is very small (24.5-26.5 C). Tem- 

 peratures during the reproductive period of January 

 1986 were among the coldest recorded during the en- 

 tire study, whereas temperatures during spawning in 

 April 1984 were among the highest recorded. 



Wyatt ( 1976) reported on spawning seasons of two 

 Holocentrinae species in Jamaican waters. He re- 



corded ripe Holocentrus adscensionis collected in all 

 months except June, and only 2% ( one specimen ) ripe 

 in July. Most spawning occurred in January, Febru- 

 ary, and March, but October was also a peak month. 

 Besides ripe fish, "sexually active" gonads were com- 

 mon (14-37% of all gonads) in September through 

 May. The seasonal pattern was similar for//, rufus; 

 highest peaks of ripe gonads occurred in October 

 (44%) and February (32%). "Sexually active" gonads 

 were found in all months except July. May, June, and 

 July were the months of lowest gonad development. 

 In Bermuda ( 15° farther north in mid-ocean), Winn 

 et al. (1964) reported breeding by both these 

 Holocentrinae species in June, July, and August. 

 Variability in timing of spawning due to factors such 

 as lunar periodicity, water temperature, plankton 

 productivity, photoperiod, currents, and rainfall oc- 

 curs commonly, and spawning time can vary from 

 year to year, even at the same location ( Watson and 

 Leis, 1974; Wyrtki, 1974; Johannes, 1978; Walsh, 1987). 



The fecundity of M. amaena is relatively low com- 

 pared with many marine species. The most fecund 

 specimen examined contained fewer than 70,000 

 maturation stage eggs, and the length-fecundity func- 

 tion predicts that a specimen of length L M would con- 

 tain fewer than 100,000 such eggs. Fecundity in- 

 creases sharply with body size; it rises with the fifth 

 power of weight and more than the tenth power of 

 SL. (The sample size used for the regressions was 

 not large, but the values of r 2 indicate a reasonably 

 good fit.) These changes with size are much greater 

 than those found in many marine species. These re- 

 sults, together with the results for SFR and the old- 

 est specimen aged, indicate that the species matures 

 slowly. With a relatively long life and steeply increas- 

 ing fecundity, a very large fraction of the reproduc- 

 tive output of the population is provided by old fish. 



The number of spawnings per year is unknown. 

 There was no clear evidence in ovaries examined 

 under direct light microscopy or histologically of a 

 distinct series of distinguishable groups of ova in a 

 graded-size sequence that might represent serial 

 batches spawned within a season. However, produc- 

 tion of such serial clutches is commonplace in tropi- 

 cal, nearshore fishes. Unless individual females 

 spawn a good many batches (of the full number of 

 maturation stage eggs estimated) within each year, 

 the total or lifetime fecundity of individuals is rela- 

 tively low compared with many common marine spe- 

 cies. For example, an individual maturing at age six 

 and spawning once annually through age 14, accord- 

 ing to the body sizes indicated by our von Bertalanffy 

 expression (Dee and Radtke, 1989) and the fecun- 

 dity indicated by our length-fecundity expression, 

 would produce fewer than 300,000 eggs during such 



