8 
Fishery Bulletin 95(1), 1997 
found that larvae that are small at age may, under 
certain circumstances, be less vulnerable to preda- 
tion than are large members of a cohort (Litvak and 
Leggett, 1992; Pepin et ah, 1992; Bertram, 1996). In 
flatfishes and in other fishes that switch habitats at 
metamorphosis, the time of transition is likely to be 
associated with high mortality. In recent laboratory 
experiments, Whiting and Able ( 1995) demonstrated 
that mortality from shrimp (Crangon septemspinosa) 
predation on settled winter flounder (10.1-14.5 mm) 
was twice that of presettled individuals (<11 mm). 
Bertram and Leggett (1994), however, could detect 
no difference in shrimp-induced mortality for winter 
flounder that differed in either length or age at meta- 
morphosis. In the present study, there was a posi- 
tive relation between length and age at metamor- 
phosis for individually reared winter flounder, but 
this trend was not evident from the larger sample of 
group-reared fish (see also Bertram et al., 1993). 
However, the results may not be strictly comparable 
because different families were used for the indi- 
vidual and group rearing. Chambers and Leggett 
(1987) reported a positive relation between length 
and age at metamorphosis for 8 of 18 families of labo- 
ratory-reared winter flounder from Newfoundland. 
The appropriate experiments have not been con- 
ducted to determine whether both large size and old 
age at metamorphosis reduce mortality due to pre- 
dation. Therefore, to date, there is no firm basis for 
interpreting the survival consequences of the indi- 
vidual variation in larval growth and development 
patterns reported here. 
The results of this study are consistent with 
Bertram et al.’s (1993) finding that size at age does 
not diverge continuously during the larval and juve- 
nile periods. Overall, the results show that juvenile 
growth rates are parallel, despite differences in lar- 
val growth rates and age at metamorphosis. The 
parallel nature of juvenile growth rates shows that 
slow-growing larvae partially compensated for their 
small size at age by increasing their juvenile growth 
rates to a greater degree than did fish that grew rap- 
idly as larvae. However, the compensatory growth 
among slow-growing larvae was not sufficient to 
cause convergence in juvenile size at age. If these 
growth rates are maintained, differences in size at 
age of juveniles that metamorphosed early and late 
would remain. 
Previous work has shown that growth rates of 
group-reared fish were maintained from weeks 1-7 
of the juvenile period (Bertram et al., 1993). In the 
present study, however, the growth rate of individu- 
ally reared juveniles was slower in weeks 5-7 than 
during weeks 1-4. There is reason to believe that 
food availability was a factor in this difference. Ju- 
veniles reared in groups in 7-L containers were ex- 
posed to concentrations of 292 prey/d per fish. Juve- 
niles reared individually in 0.4-L containers received 
100 prey/d because rations were 0.25 Artemia nau- 
pliiAmL-d) for both container sizes. Consequently, 
food availability may have limited the growth rate 
of individually-reared juveniles during weeks 5-7 
when fish were relatively large and food require- 
ments were maximal. 
An important conclusion from this study is that 
the insight into the dynamics of larval growth and 
development was gained only because the data were 
presented as individual-based observations. Although 
the CV’s of size at age would have been available if 
the weekly length measures of individuals had been 
pooled, the underlying growth dynamics and indi- 
vidual variability would have been concealed. More- 
over, a single “growth” curve fit to such size-at-age 
data would not accurately reflect the growth patterns 
of individual larvae. In this connection, we point out 
that a recent description of winter flounder larval 
“growth,” based upon reconstructions of size at age 
from otolith microstructure ( Jearld et al., 1993), bears 
little resemblance to the individual growth trajecto- 
ries shown here. 
Darwin (1859) wrote: “No one supposes that all 
individuals of the same species are cast in the very 
same mould”; but it is only recently that fishery sci- 
entists have begun to investigate the potential popu- 
lation consequences of phenotypic variability in early 
life history stages. Because mechanisms controlling 
survival and recruitment of fishes operate at the level 
of the individual (Crowder et al., 1992), baseline es- 
timates on phenotypic variability are required. This 
study clearly shows that rearing marine fish larvae 
individually in the laboratory can provide such esti- 
mates. We have shown that there is considerable 
variability in the dynamics of individual larval 
growth and development. Studies that examine the 
survival consequences of such variability represent 
a logical next step in research programs designed to 
provide a mechanistic understanding of the factors 
that affect survival during fish early life history. 
Acknowledgments 
We thank M. Litvak, F. Purton, and the staff of the 
Huntsman Marine Science Centre (HMSC) for their 
cooperation in this research. J. Farrell, S. Gutterman, 
and S. Whelan provided technical assistance. Finan- 
cial support was provided by Natural Sciences and 
Engineering Research Council (NSERC) of Canada 
operating and strategic grants to WCL. Additional 
financial support to DFB was provided by a NSERC 
