644 
Fishery Bulletin 95(4), 1997 
cal capacity to resist growth overfishing as those 
stocks in the northwestern Gulf of Mexico (Chit- 
tenden, 1977). The low values of t CRITIC andP a agree 
with yield-per-recruit modeling results and indicate 
that 1) for a reported maximum longevity of 8 years 
in Chesapeake Bay (Barbieri et al. , 1994a), maximum 
theoretical biomass is achieved very early in life, 
before fish reach age 2; and 2) very little potential 
for a growth span still remains when fish enter the 
exploited phase at age 2. As a precaution against 
future problems — especially considering that annual 
recruitment is reported to be highly variable and 
strongly density independent — we suggest that regu- 
latory measures for Atlantic croaker in the lower 
Chesapeake Bay be directed at maintaining the ap- 
parent current level of t (age 2; l c = 265 mm TL; 
Barbieri et al., 1994a). In addition, the magnitude 
and composition of the scrap catch for the main fish- 
eries in this area need to be estimated, and their ef- 
fect on estimates of F CUR and t c need to be assessed 
more precisely before any definite conclusion on At- 
lantic croaker yield-per-recruit can be reached. 
In contrast to what we found for the lower Chesa- 
peake Bay, results for North Carolina indicated that 
Atlantic croaker were being severely growth-over- 
fished. First, independent of the level of F or M used 
in simulations, yield-per-recruit values were consis- 
tently higher at higher levels of t c , indicating that 
age and size limits during the period 1979-81 (f =1, 
Z =177 mm TL; Ross, 1988) were unrealistically low. 
Second, estimates of F CUR for Z = 1 were not just con- 
sistently higher than F 0 x but were also well above F MAX . 
The pattern of declining yield-per-recruit values with 
increasing F at lower levels of t c agrees well with the 
high estimates of t CRITIC and P g and indicates that, con- 
trary to the pattern shown in Chesapeake Bay, maxi- 
mum cohort biomass is attained later in life (ages 5-7). 
However, differences in yield-per-recruit modeling 
results between Chesapeake Bay and North Caro- 
lina seem to reflect temporal rather than spatial dif- 
ferences in Atlantic croaker population dynamics. 
Parameters used in simulations for North Carolina 
were obtained from a study (Ross, 1988) conducted 
during a period (1979-81) that coincides with the 
occurrence of unusually large Atlantic croaker (350- 
520 mm TL; Ross, 1988) along the east coast of the 
United States (Barbieri et al., 1994a). However, since 
1982, Atlantic croaker catches in North Carolina have 
been dominated by smaller fish. Modal lengths of 
Atlantic croaker in the long haul-seine fishery dur- 
ing 1982-92 ranged from 215 to 245 mm TL; in the 
winter trawl fishery, they ranged from 215 to 240 
mm TL. In both fisheries, less than 10% of the fish 
were older than age 3 (Wilson, 1993). Therefore, 
yield-per-recruit modeling results presented here for 
North Carolina should not reflect current conditions, 
but rather be considered representative of temporal 
changes in Atlantic croaker population dynamics. 
The specific value of M used in our simulations 
had no effect on the levels of F or Z that produce 
maximum yield-per-recruit values and would not 
change conclusions for either Chesapeake Bay or North 
Carolina. However, these conclusions are still critically 
dependent on how realistic is the range of M used in 
these simulations. Methods currently used to estimate 
M have strong limitations and disadvantages (Vetter, 
1988), and the method used here is no exception. How- 
ever, we feel comfortable with the range of M used in 
this study because it agrees with values of M reported 
for other sciaenids with similar life spans, e.g. spotted 
seatrout, Cynoscion nebulosus (Rutherford et al., 1989). 
Yield-per-recruit analysis is only part of a fishery 
management strategy (Beverton and Holt, 1957; 
Gulland, 1983; Deriso, 1987). It must be applied in 
conjunction with eggs-per-recruit (Prager et al. , 1987) 
and spawning stock biomass per recruit models 
(Gabriel et al., 1989; Goodyear, 1993; Schirripa and 
Goodyear, 1994) to allow managers to examine the 
effects of different policies on both reproduction (i.e. 
egg production) and biomass yield. The pattern of 
early maturation, multiple spawning, long spawn- 
ing season, and indeterminate fecundity in Atlantic 
croaker (Barbieri et al., 1994b) suggest that repro- 
duction would be compromised only at extremely high 
levels of fishing. However, eggs-per-recruit and 
spawning stock biomass models must be applied be- 
fore this issue can be properly evaluated. 
Acknowledgments 
We would like to thank Sue Lowerre-Barbieri and 
two anonymous reviewers for helpful suggestions 
that improved the manuscript. Financial support for 
this project was provided by the College of William 
and Mary, Virginia Institute of Marine Science, and 
by a Wallop/Breaux Program Grant for Sport Fish 
Restoration from the U. S. Fish and Wildlife Service 
through the Virginia Marine Resources Commission, 
Project No. F-88-R3. Luiz R. Barbieri was partially 
supported by a scholarship from CNPq, Ministry of 
Science and Technology, Brazil (process No. 203581/ 
86-OC) and by a postdoctoral fellowship from the 
University of Georgia Marine Institute. 
Literature cited 
Alverson, D. L., and M. J. Carney. 
1975. A graphic review of the growth and decay of popula- 
tion cohorts. J. Cons. Cons. Int. Explor. Mer 36:133-143. 
