EVERSON ET AL.: MATURATION AND REPRODUCTION IN TWO HAWAIIAN SNAPPERS 



signed a maturity stage based on the criteria in 

 Table 1. Oocytes were histologically identified 

 using information provided in Crossland (1977). 

 Sectioned ovaries also were examined for the 

 presence of postovulatory follicles and oocyte 

 atresia, features used to establish criteria for the 

 estimation of spawning frequency and to sep- 

 arate juveniles from prespawning adults for de- 

 termination of size at sexual maturity (Hunter 

 and Macewicz 1985). 



The results of these four visual methods were 

 compared with gonosomatic index (GSI) values 

 [(gonad weight/body weight) x 100]. Gonoso- 

 matic indexes were calculated for both species to 

 provide a rapid but preliminary indication of 

 developmental stage, although an insufficient 

 number of uku testes were obtained. Excluded 

 from this analysis were individuals that had not 

 yet reached size at sexual maturity. Spearman's 

 coefficient of rank correlation (Snedecor and 

 Cochran 1978) was used to ascertain whether a 

 positive relationship existed between GSI and 

 maturity stage during the spawning periods for 

 females of each species. 



250 350 450 550 650 750 850 950 1050 



250 350 450 550 650 75C 850 950 



FORK LENGTH (mm) 



Figure 1. — Length-frequency distribution of male and 

 female uku and onaga. 



Size at sexual maturity (L50) was defined as 

 the smallest length category in which at least 

 50% of the individuals were mature (i.e., stage II 

 or beyond, GSI > 1.5) during the spawning 

 season. The logistic equation was fitted to the 

 percentage of mature individuals in each size 

 class (P.,) and FL (Gunderson et al. 1980; Ni and 

 Sandeman 1984); that is 



Pr = 



100 



1 + exp(aFL + b) 



where a and b are fitted parameters and L50 = 

 -b/a. We also calculated the percentage of max- 

 imum length (MAXLEN) at which sexual matur- 

 ity occurred. This ratio has been used to 

 compare proportionate size at maturity for 

 species by habitat type, zoographic province, or 

 depth range (Grimes 1987). 



Sex ratios were compiled and examined for 

 significant deviation from unity and to determine 

 whether sex ratio and size (in 50 mm FL inter- 

 vals) were independent using chi-square statis- 

 tics. Two-way contingency table analysis was 

 performed to determine whether the sex ratio 

 differed during the year when pooled into bi- 

 monthly periods. 



RESULTS 



Spawning Season 



Spawning season was determined from a wide 

 size range of uku and onaga (Fig. 1). The Spear- 

 man rank correlation coefficients (Vs) calculated 

 by comparing GSI with stages I-III were r^ = 

 0.6205 for uku (P < 0.0001), and r, = 0.8685 for 

 onaga (P < 0.0001), indicating a positive rela- 

 tionship between GSI and stage of development 

 for both species, although the correlation was 

 considerably lower for uku. In addition, the 

 range in GSI values representing stages II and 

 III ovaries was greater for uku than for onaga 

 (Fig. 2). Thus, rather than using GSI as the 

 single method for estimating spawning season- 

 ality by month or fish length, visual staging 

 methods also were used. 



Both species reached maturity in the spring 

 and summer and spawned continuously until fall 

 or early winter. Neither species reached stage II 

 of development (vitellogenesis) at any other time 

 of the year. Uku spawning began in May and 

 peaked 1 month later in June, as evidenced by 

 the sharp rise in GSI values and the presence of 

 mature and ripe fish during this time (Figs. 3, 4). 



879 



