Zhao and McGovern: Variation in sexual maturity and sex ratio of Rhomboplites aurorubens 
839 
at sexual maturity (TL 50 ) was calculated by probit 
analysis following the recommendation of Tripple and 
Harvey (1991). The likelihood ratio test at a signifi- 
cance level of 0. 10 was used to determine if the probit 
model could be fitted to the observations of matu- 
rity-at-length (SAS Institute, 1990). 
Because sex ratio may change with water depth 
and latitude of sampling sites, fish length, sampling 
year, gear, or season (Grimes and Huntsman, 1980; 
Nelson, 1988), all sexed samples, including mature 
and immature individuals between 1979 and 1993, 
were split into the same five periods as described 
above for maturity analysis. The time frame was in- 
creased to May through August to increase the 
sample size. We compared the percentages of females 
among varying water depths (midpoints: 25, 35, 45, 
and 55 m), latitudes (31°N, 32°N, and 33°N), length 
classes (TL=100 - 450 mm, with 50-mm intervals), 
periods, and gear types (traps, hook-and-line, and 
trawl). Latitudes of 31°N, 32°N, and 33°N refer to 
31°00'-31°59'N, 32°00'-32°59’N, and 33°00’-33°59’N, 
respectively. When the independence between sex 
ratio and one of above factors was tested, other fac- 
tors were kept consistent. The chi-square test and 
Fisher’s exact test were used to decide the indepen- 
dence. As a reference, Bonferroni’s method was used 
to adjust the significance level, i.e. a'=0.05/m, where 
m = the number of cases (Sokal and Rohlf, 1995). 
First, we compared the percentage of female vermil- 
ion snapper taken from varying depths with the same 
length classes (TL=2QQ-249 mm and TL=250-299 
mm respectively), same latitude (32°N), same peri- 
ods, and same gear type. Seventy-three percent of 
the vermilion snapper were collected from latitude 
32°N during 1979-93, and eighty percent of them 
were between 200 and 299 mm TL. If the hypothesis 
