DeMartini et al.: Body sizes at maturation and at sex change, and spawning seasonality and sex ratio of Hyporthodus quernus 
127 
Spawning seasonality 
Seasonality of spawning was gauged with a combi- 
nation of gravimetric evidence of gonadal ripening 
and histological evidence of either recent or imminent 
spawning by females. A gonadosomatic index, GSI=(GW/ 
GFBW)x 100%, where Gff=damp gonad weight and 
gonad-free body weight, GFBW=round weight minus GW, 
aggregated by month (years pooled) for each sex of adult 
fish, provided gravimetric evidence of gonadal growth. 
This simple (weight proportional) GSI was used to visu- 
alize temporal patterns. Because of the likely dynamic 
influence of reproductive stage on proportional gonad 
weight-to-body size, the calculation of a “relative gonad 
index” ( RGI ) was required in order to scale the rela- 
tion across gonad maturation stages of mature females. 
This calculation removes the confounding influence of 
maturation stage on the gonad-to-body-size relation 
(Erickson et ah, 1985). One-way ANCOVA (Zar, 1984) 
with month as a fixed factor (12 levels) and either TL 
or GFBW as covariate was used to evaluate the poten- 
tial effect of month on the RGI, controlling for fish size. 
ANCOVA was followed by a robust multiple comparison 
test (Ryan-Einot-Gabriel-Welsch [REGW] multiple-range 
test; Quinn and Keough, 2002) to quantitatively evalu- 
ate patterns among months. Plots of the proportional 
incidence for recently spawned and “imminent” ( ready - 
to-spawn) females among all mature females constituted 
histological evidence of spawning by individual fish. 
Results 
Seventy percent of the Hawaiian grouper examined 
were from the Ho’omalu Zone, west of 165°W; another 
26% were from the Mau Zone, between 165° and 161°W 
in the NWHI (Fig. 1). Fish from the early, mid-term, 
and recent series comprised about 18%, 11%, and 71%, 
respectively, of those collected. Fish were obtained in all 
months of the year (ranging from 20 in January to 108 
fish in April); and the distribution of specimens among 
months of collection was indistinguishable between the 
parent sample and a subsample used for histological 
examination. The gonads of 20 (January) to 92 (April) 
specimens per month were examined microscopically. 
The mean and median body lengths of specimens was 
661 mm TL (ranging from 241 to 1103 mm). Sizes of fish 
used for histological examination generally resembled 
those in the parent sample (mean=659 mm, median=652 
mm, range 258 to 1103 mm TL; Kolmogorov-Smirnov 
two-sample test: P>0.10). 
Comparisons of collection series 
Estimated median body size at sexual maturation varied 
by barely 3 mm between the early (576.0 ±13.4 [stan- 
dard error, SE] mm TL) and recent (579.1 ±4.0 mm TL) 
collection series. We considered this sufficient reason 
to pool samples, given the small magnitude of length 
measurement error, despite a marginally significant 
F-statistic (F 3091 =1.97, P=0.02). Estimates of adult sex 
ratios also were indistinguishable between these two col- 
lection series (maximum likelihood chi-square: j 2 =1.40, 
df=l, P=0.24). Because neither metric differed between 
series, we pooled all fish collections and present single 
estimates for each of these and related variables based 
on all combined capture series. 
Gonadal and sex allocation patterns 
Gonadal patterns Four of the 604 specimens prepared 
for histological examination were alimentary tract tissues 
collected in error and were discarded. A small minority 
of the remaining 600 valid histological specimens were 
bisexual juveniles with vestigial gonads containing chro- 
matin nucleolar oogonia, undeveloped spermatic tissue, 
or both (Fig. 2; Fig. 3, A and B). Most of these fish were 
<450 mm TL (Fig. 2). Among the remaining specimens 
whose gonads were more developed, about one-third of 
all females had immature but clearly all-female charac- 
teristics, another two-thirds were mature females, five 
were developing males with transitional gonads, and 55 
(9.2%) were fully mature males (Table 1; Fig. 2; Fig. 3, 
C and D, E and F). Developing oocytes progressed from a 
cortical alveolus (perinucleolar) and previtellogenic stage 
through vitellogenesis and ovulation, with concomitant 
increases in ovarian mass and oocyte diameter (Table 
1). The ovaries of each of three stage-5 females had mul- 
timodal size distributions of yolked or yolking oocytes 
(reject Ho: distribution unimodal normal; Kolmogorov- 
Smirnov one-sample test, all P<0.001; Fig. 30. The 
gonads of all males contained from one to dozens of 
brown bodies; in many cases, these were recognizable 
as gamma-stage (Hunter and Macewicz, 1985) atretic 
structures (Fig. 3, E and F). No mature males smaller 
than 753 mm TL were encountered. Gonads were fused 
posteriorly and the gonads of fish of all sizes and each 
sex contained a membrane-lined ovary-like lumen (e.g., 
bisexual juveniles and mature males: Fig. 3, A and E). 
Maturation scores were identical and sizes of oocytes 
were indistinguishable from the respective anterior, 
middle, and posterior sections of the ovaries of three ripe 
females from the early collection series. 
Body sizes at maturation and at sex change 
Maturation Body size at first sexual maturation was 
estimated by using the pooled sample of all immature 
(bisexual and female) fish and all recognizably mature 
females. On the basis of Equation 1, proportional body 
length at 50% maturity was best described by the logis- 
tic equation 
P 5 Q = 100 / (1 + exp ( 0.0286 - 16.574-L^) ) ; 
with SE a =0.003, SE b = 1.749, r 2 = 0.93, P<0.0001, and 
n = 540. 
Body length (L 50 ) at female maturation was 580 ±8 mm 
TL (95% confidence interval [Cl]) (Fig. 4). The respec- 
