838 
Fishery Bulletin 95(4), 1 997 
method of otolith sections and demonstrated that 
size-at-age decreased with time. Collins and 
Pinckney (1988) reported preliminary evidence that 
vermilion snapper caught in 1978—80 from SAB became 
reproductively mature earlier in life than those caught 
in 1972-74. Grimes and Huntsman (1980) deter- 
mined that vermilion snapper were gonochorists, fe- 
males (62.5%) significantly outnumbered males, and 
the sex ratio was dependent on fish length. However, 
Nelson (1988) reported that the sex ratio of vermil- 
ion snapper from the Gulf of Mexico differed from 
1:1 in favor of males (54.5%). He also reported that 
area and season had a significant effect on this ra- 
tio. Although what may have caused the difference 
in sex ratios of two studies was unknown, limited 
sample sizes made their comparisons among areas, 
lengths, or seasons less convincing (Grimes and 
Huntsman, 1980: n= 874; Nelson, 1988: n=881). 
In this paper, we investigated the percentage of 
mature vermilion snapper at each length class and 
age for each sex and examined the temporal change 
in maturity schedules during 1979-93. We also de- 
termined the sex ratio of vermilion snapper accord- 
ing to depth and latitude of sampling sites, fish 
length, sampling period, and types of fishing gear 
used. 
Materials and methods 
Data were collected from the SAB during 1979-93 
by reef fish surveys of the Marine Resources Moni- 
toring, Assessment, and Prediction (MARMAP) pro- 
gram. Most vermilion snapper were captured by stan- 
dardized hook-and-line and trapping gear during the 
spring and summer of 1979-93 (Collins, 1990; Collins 
and Sedberry, 1991). In addition, many vermilion 
snapper were obtained from the MARMAP trawling 
program that was terminated in 1988. In the field, 
fish were measured to the nearest mm (TL and FL) 
and weighed to the nearest gram (whole body weight). 
Water depth, latitude, and longitude of sampling sites 
were recorded. The posterior portion of the gonad was 
removed and preserved in 10% formalin buffered with 
seawater. In the laboratory, sex and maturity stages 
were determined by examining the stained gonad 
sections according to standard MARMAP histologi- 
cal criteria (Cuellar et al., 1996). All gonad samples 
taken in 1987-93 were examined with histological 
methods. During 1979-86, sex of specimens was de- 
termined and a reproductive condition was assigned 
in the field by using clearly defined visual staging 
criteria. The definitions of maturity stages by gross 
visual inspection were confirmed to be accurate by 
histological examination during 1978-80 (Collins and 
Pinckney, 1988). The codes used for defining matu- 
rity stages were consistent between histological and 
macroscopic methods. Seven reproductive stages were 
used: 1) immature, 2) developing, 3) running ripe, 4) 
spent, 5) resting, 6) developing with evidence of spawn- 
ing in the previous week, and 7) mature, but stage- 
unknown. Mature fish included those in stages 2-7. 
Percentages of mature males and females based 
on total length (TL) and age were calculated for five 
periods, each corresponding to a three-year interval, 
i.e. 1979-81, 1982-84, 1985-87, 1988-90, and 1991- 
93. Because the monthly distribution of samples was 
not necessarily similar among periods, it was essen- 
tial to define a standard month(s) if comparison of 
maturity schedules among years was to be meaning- 
ful. In addition, the reproductive condition was most 
discernible and the sizes and ages at first maturity 
could be determined during and immediately prior 
to the spawning season. Therefore, we used the data 
only from May and June that corresponded to the 
time of slow somatic growth. Age assignments of ver- 
milion snapper were based on the number of annuli 
on otolith sections (Zhao et al., 1997). We took into 
consideration knowledge of the time of annulus for- 
mation, the relative growth of the otolith margin, the 
date of sampling, and used a January 1 hatching 
date. Because only a part of the samples was aged, 
some sexed and maturity-determined samples did not 
have age information. 
Although the variation in gear type, latitude, and 
depth of sampling sites should have been included 
in the maturity analysis before the data were pooled, 
limited sample sizes of small (TL<170 mm) and young 
(age-1) vermilion snapper prevented us from com- 
paring maturity schedules by area (latitude) or by 
gear type. However, 97% of the males and females 
smaller than 170 mm were collected by trawl in con- 
sistent depth ranges during 1979-90. A majority of 
these small fish were collected from similar areas, 
with 88% males and 82% females from latitude 31°N, 
and 12% males and 18% females from 32°N. Thus, 
gear selectivity and geographical distribution can be 
disregarded as sources of bias in comparison of ma- 
turity schedules among periods. 
As recommended by Trippel and Harvey (1991), 
the G-test was used to compare maturity schedules 
of each sex among periods with the same length class 
or the same age (e.g. age 1). Maturity schedules be- 
tween sexes at age 1 were compared for the periods 
of 1979-81 and 1985-87 respectively, when sample 
sizes of both sexes were sufficient. When conditions 
of the G-test were not met, Fisher’s exact test was 
used (Zar, 1984). When the percentages of mature 
fish of each sex in 20 mm TL intervals exhibited a 
successive increase with length, the median length 
