846 
Fishery Bulletin 95(4), 1997 
tently increased during the 1980’s (Zhao and 
McGovern 2 ). The demonstration that the harvest of 
a fish stock can lead to declines in length or age at 
maturity has been reported for many fishes, includ- 
ing northeast Arctic cod (Jprgensen, 1990), Pacific 
salmon (Ricker, 1981), and California halibut (Love 
and Brooks, 1990). Changes in size or age at matu- 
rity may be the result of a density-dependent re- 
sponse to decreased stock abundance, selective re- 
moval and incomplete replacement of later-matur- 
ing fish by the fishery, or genetic change within a 
population (Nelson and Soule, 1987). Jprgensen 
(1990) attributed a decline in median age-at-matu- 
rity in northeast Arctic cod to an increase in length- 
at-age (i.e. faster growth) coincident with declining 
stock density, an idea that implicitly assumes a mini- 
mum threshold for size-at-maturity. If the scenario 
of Jprgensen ( 1990) is correct, declines in length and 
age should not occur concurrently. Furthermore, Zhao 
et al. ( 1997) indicated that the size-at-age of vermil- 
ion snapper has decreased with time. Therefore, 
changes in maturity schedules of vermilion snapper 
are not part of a density-dependent compensatory 
response to harvesting, but quite likely a result of 
the selective removal and incomplete replacement 
of faster-growing, later-maturing fish by the fishery. 
If intensive fishing pressure continues, and the early- 
maturing trait is heritable, length and age at matu- 
rity in the population will decrease with time. Life 
history theory predicts that genetic changes in life 
history characteristics will occur following increased 
mortality (Roff, 1992). Harvesting can reverse the 
relative fitness of genotypes, because an inferior 
genotype (e.g. slow-growing and early-maturing) in 
an unexploited population may be more fit under 
increased fishing pressure (Bergh and Getz, 1989). 
Early-maturing genotypes reproduce before being 
fully recruited to the fishery, whereas genotypes that 
mature at larger sizes or older ages tend to be re- 
moved before reproduction. This process would ex- 
plain the decreasing abundance of larger, immature 
fish with time and would account for declines in both 
size and age at maturity. The long-term impacts of 
size-selective fish harvests may have caused the de- 
cline in size-at-age of vermilion snapper through dis- 
proportionate harvesting of fast-growing individuals 
(Zhao et al., 1997). Similarly, it may be that late- 
maturing genotypes were removed from the vermil- 
ion snapper population in the 1980’s when fishing 
pressure was intensive. 
Maturity schedules of vermilion snapper collected 
during 1972-74, prior to heavy exploitation, were 
investigated by Grimes and Huntsman (1980). They 
used a gonadosomatic index and indicated that “most 
fish attain sexual maturity during their third or 
fourth years of life (186-256 and 256-324 mm TL), 
but a few precocious individuals may mature in their 
second year (100-186 mm TL) at about 150 mm TL.” 
It is not rigorous to compare the percent mature, 
based on age, between Grimes and Huntsman ( 1990) 
and the present study because an obvious discrep- 
ancy in size-at-age exists between Grimes (1978) and 
Zhao et al. (1997). It is meaningful, however, to com- 
pare maturity schedules based on length between 
these two studies. The maximum-likelihood esti- 
mates from the probit analysis of data from the 
present study predicted that 50% of males and 15% 
of females matured by 150 mm during 1979-81 and 
that 50% females at 150 mm matured during 1985- 
87. All males and females at 180 mm were mature in 
the present study. Differences between previous 
(Grimes and Huntsman, 1980) and present results 
could be partially due to differences in methods used 
to determine maturity (Collins and Pinckney, 1988) 
or may truly reflect the changes in maturity that 
occurred in the 1970’s. The increase in percentage of 
mature females at 150 mm was faster during the 
1980’s than during the 1970’s (i.e. an increase of 35% 
in six years from 1979-81 to 1985-87, versus an in- 
crease of less than 15% in seven years from 1972-74 
to 1979-81). The degree of exploitation may account 
for the differing rates of change in maturity while 
the fishery for vermilion snapper was initiated in the 
1970’s, but heavy exploitation did not occur until the 
1980’s. 
Sex ratios 
The chi-square analysis did not suggest significant 
differences in percentages of females among months 
(May-August) for any gear type. This information 
supported the notion that pooling data between May 
through August would not bias the comparison of sex 
ratios. Seasonal comparisons of sex ratios could not 
be done because little sampling was done in fall or 
winter. 
This study showed that the sex ratio of vermilion 
snapper was dependent on area (latitude) and gear 
type, but independent of depth of sampling sites, fish 
length, or sampling years. The reason for the signifi- 
cant differences among latitudes is unknown. How- 
ever, only 2 of the 14 cases showed a significant dif- 
ference according to Bonferroni’s method (Table 4), 
and no trend was observed between latitudes. In 
addition, relatively small sample sizes collected from 
latitudes other than 32°N may have induced errors 
in comparison. Therefore, we attribute the difference 
in sex ratio between latitudes to chance. 
Although significantly different, the percentages 
of females of vermilion snapper collected by traps 
