292 
Fishery Bulletin 1 14(3) 
A 
B 
Figure 1 
(A) Specific growth rates (% wet weight per d); and (B) 
RNA/DNA values of postsmolt Atlantic salmon (Salmo sa- 
lar) reared in the laboratory at 12°C. Fish were either fed ad 
libitum, fasted, or refed (fed after 11 days of fasting). Values 
are mean values (±standard deviation [SD]) for each sampling 
day. Growth rates for fish in the fed and fasted treatments were 
calculated from day 0 until the day they were sacrificed. Growth 
rates for the refed treatment were calculated from the first 
day of refeeding (day 11) until the day fish were sacrificed. 
Within a sampling day, food treatments showing a common 
superscript, or without superscripts, do not differ significantly 
(Tukey’s HSD multiple range tests). Where error bars would 
overlap, data are offset to facilitate viewing. The arrow indi- 
cates the day when food was introduced to the refed group. For 
fed and fasted fish, n-4 for each sampling day. For refed fish, 
n=22 for day 11, n=5 for days 15, 19, and 23, and re=7 for day 
27. 
ber of fish in the tanks fell below 8 individu- 
als because of sampling. This observation was 
confirmed by the decrease in average growth 
rates for fish sampled on the last day of both 
the fed (day 23) and refed (day 27) treatments 
(Fig. lA). 
After 11 days, RNA/DNA values were sig- 
nificantly greater in fed fish than in the fasted 
portion of the refed fish (Fig. IB), and from 
day 15 onward, significantly greater than val- 
ues for fish in the fasted group. Twelve days 
after refeeding, RNA/DNA values of the refed 
group were greater than the fasted fish and 
equal to the continually fed fish (day 23, Fig. 
IB). Mean RNA/DNA values in fed and refed 
fish increased in relation to their mean start 
values (Dunnett, F=0.011, P<0.0001 respec- 
tively), whereas mean RNA/DNA values for 
fasted fish decreased beginning 15 days after 
food was withheld (Dunnett, F=0.003). Re- 
peated measurements from individuals in the 
fasted group exhibited an overall significant 
decrease in RNA/DNA from their start values 
(Student’s paired t test F<0.0001, Fig. 2A). 
Repeated measurements of refed fish exhib- 
ited an overall significant increase from their 
individual start values (Student’s paired Ntest 
F<0.0001, Fig. 2B). RNA/DNA values were 
highly positively correlated with growth rates 
in the fed group, refed group, and the all-data- 
combined group, but not in the fasted group 
(Table 2). 
Beginning on day 11, RN A/pro values were 
generally greater in the fed fish than in the 
fish in the fasted treatment (Fig. 3A). In fed 
fish, mean RNA/pro values did not change from 
mean initial values (Dunnett F=0.391), where- 
as fasted fish showed modest decreases with a 
significant Dunnett value (F=0.009), primarily 
driven by 1 fish on day 19 (Fig. 3A). Beginning 
8 days after refeeding (day 19), mean RNA/pro 
values of refed fish increased from the mean 
start value (Dunnett F<0.0001). The same pat- 
tern was seen in paired data from individuals; 
repeated measurements from fasted fish exhib- 
ited an overall significant decrease from their 
start values (Student’s paired /-test F=0.002), 
whereas values of refed fish increased (Stu- 
dent’s paired /-test F<0.0001) (plots not shown). 
RNA/pro values were positively correlated with 
growth rates in the fed group, refed group, and 
the all-data-combined group, but not in the 
fasted group (Table 2). 
The DNA/pro ratio of fasted and refed fish 
was higher (smaller cells) than in fed fish, and 
the main effect of feeding treatment on DNA/ 
pro was statistically significant (Fig. 3B); but, 
because of high daily variability, coupled with a 
small sample size, most Tukey- Kramer post-hoc 
comparisons of feeding treatment within a sam- 
