496 



Fishery Bulletin 89(3), 1991 



Spawning season was estimated by visual examina- 

 tion of gonads and classification of the mature gonads 

 as either developing or gravid (Kesteven 1960). Months 

 were noted in which there was a significant increase 

 in the number of gravid fish collected relative to the 

 number of mature fish with developing gonads. Spawn- 

 ing season was also estimated by using the gonado- 

 somatic index (GSI): 



GSI = 

 (gonad wet weight/whole body wet weight) x 100. 



The GSI for each mature individual was plotted against 

 the date of capture. The spawning season was defined 

 as the period of the year during which a significant pro- 

 portion of the fish had high GSI values. 



To determine the SFR, all fish taken during the 

 spawning season were put into size-classes based on 

 standard length (50 mm SL classes for C. melampygus 

 and 100 mm classes for C. ignobilis). The number and 

 percent of individuals with gravid gonads in each size- 

 class was determined. As body size increased, at sex- 

 ual maturity the percent of fish with gravid gonads rose 

 sharply. The lower limit of the size-class in which 50% 

 gravid gonads was reached was taken as the SFR. The 

 SFR was transformed to an age at first reproduction 

 by using the length-age relation from the von Ber- 

 talanffy growth equation. 



Ovaries from gravid females captured during the 

 peak of the spawning season were used to estimate 

 fecundity, which is defined here as the estimated num- 

 ber of mature ova in the gonad of a spawning female 

 at one time (Everhart and Youngs 1981). Three 2g 

 aliquots from various parts of each ovary were com- 

 bined and placed in Gilson's fluid (Simpson 1951). This 

 solution hardens and liberates the eggs and breaks 

 down the ovarian tissue. The egg masses were left in 

 this fluid for 2 weeks, then washed with water and the 

 ovarian tissue removed. Subsamples were obtained by 

 using van Dalsen's (1977) technique as modified slightly 

 by Everson (1984). Each ovary sample was brought 

 up to 500 mL with water and mixed with a magnetic 

 stirrer. When a homogeneous mixture was obtained, 

 three 5mL aliquots were drawn. All ova more than 

 0.4 mm in diameter in each 5mL subsample were 

 counted using a binocular dissecting microscope. A 

 fecundity estimate, F, was calculated from the formula 



F = ((Nj + N 2 + N 3 )/3) x (500/5) x (G/S) 



where N,, N 2 , N 3 = the number of mature ova in 

 each subsample, 

 G = total gonad weight, 



weight of the gonad sample 

 placed in the Gilson's fluid. 



Diet 



Qontents were removed from preserved guts and all 

 prey identified to the lowest taxa possible. The number 

 of individuals and volume of each taxon were noted, 

 as well as the length, weight, and extent of digestion. 

 Many fish in advanced states of digestion were iden- 

 tified by using reference collections of Hawaiian fish 

 scales (Sylvester 1969) and skeletons. Whole prey 

 lengths and volumes were approximated in many cases 

 by comparison of the dimensions of recognizable parts 

 with reference specimens of common Hawaiian reef 

 fishes. The index of relative importance (IRI) was 

 calculated, as defined by Pinkas et al. (1971), 



IRI = (numerical % + volume %) x frequency %, 



where, for each predator species, 



numerical % = (number of individuals of one prey 

 category divided by total number of 

 prey individuals found in all the 

 guts) x 100, 



volume % = (volume of one prey category di- 

 vided by total volume of all prey 

 found in the guts) x 100, 



frequency % = (number of guts containing prey of 

 one category divided by total num- 

 ber of guts that contained any iden- 

 tifiable prey items) x 100. 



A measure of dietary overlap between C. melam- 

 pygus and C. ignobilis was provided by the index of 

 overlap, A yz (Pianka 1973). The value of the index 

 varies from 1 when diets are identical with respect to 

 proportional IRI composition to when diets are dis- 

 tinct. We calculated the index using the formula 



A yz = I 



Piy Pi 



I P.y 2 I 



1/2 



where p iy and p iz are the proportions of the total IRI 

 represented by the ith prey category for predator 

 species y and predator species z, respectively. 



The effect of individual predator size on the diet of 

 C. melampygus was assessed by comparing the occur- 

 rence of prey items found in the guts of sexually im- 

 mature and mature fish. 



