296 
Fishery Bulletin 114(3) 
Figure 5 
Growth rate (% wet weight per d) versus IGFl for postsmolt At- 
lantic salmon {Salmo salar) reared in the laboratory at 12°C and 
sampled throughout a 27-day period. Fish were fed ad libitum, 
fasted, or refed (fed after 11 days of fasting). Regression lines rep- 
resent fed (— ) and refed (- -) fish. 
RNA/DNA 
Figure 6 
Specific growth rate (% wet weight per d) versus RNA/DNA for 
postsmolt Atlantic salmon {Salmo salar) reared in the laboratory 
at 12°C and sampled throughout a 27-day period. Fish were fed ad 
libitum, fasted, or refed (fed after 11 days of fasting). Regression 
lines represent fed (• — ) and refed (- -) fish. 
Atlantic salmon residing at temperatures 
near 12°C. 
Changes in RNA translation rates also 
may have influenced the relation of RNA/ 
DNA to growth rate in the refed fish. Refed 
fish accumulated more protein per unit of 
RNA than their continually fed counterparts 
as evidenced by the significantly greater re- 
gression intercept (Fig. 6). This increase in 
protein accumulation could be due to an in- 
crease in translation rate or to a decrease 
in protein degradation rate (or to both). Re- 
searchers who have observed an increase in 
protein accumulation per unit of RNA have 
attributed it to a translation increase (Mc- 
Millan and Houlihan, 1988, 1989; Miglavs 
and Jobling, 1989) although direct published 
evidence for choosing between the two path- 
ways is scarce. Because traditional estimates 
of nucleic-acid-based growth-rates do not in- 
clude a translation rate estimate, field esti- 
mates of RNA/DNA-based growth rates will 
not be exact. Nonetheless, by combining our 
refed with the fed and fasted data sets we ob- 
tained a significant regression between RNA/ 
DNA and growth rate (r2=0.692, P<0.0001) 
that can be used to coarsely estimate field 
growth rates at temperatures near 12°C. Be- 
cause temperature affects protein degrada- 
tion and translation rates (Buckley et ah, 
1999; McCarthy et al., 1999; Ciotti et ah, 
2010), a model incorporating an interaction 
term of temperature with RNA/DNA would 
need to be constructed before growth rates 
at different temperatures could be estimated. 
RNA/pro In our study, RNA/pro values were 
less correlated with growth rate than the 
RNA/DNA values, although the relation of 
both indices to growth rate was similar, i.e., 
there was a high correlation with positive 
growth rates but a range of values associ- 
ated with a constant negative growth rate. 
Unlike RNA/DNA, RNA/pro from both the 
fed and refed groups had the same relation 
to positive growth rates which could be ad- 
vantageous for estimating field growth rates 
where the past feeding history of the animal 
is unknown. Repeated measurements of in- 
dividuals indicated that RNA/pro responded 
quickly to food withdrawal (7 days) and re- 
introduction (4 days), but within any one-day 
values were variable, resulting in sampling 
days where the feeding groups could not be 
differentiated with this index. The positive 
and negative response of growth rate, RNA/ 
DNA, and RNA/pro in individual fish to the 
positive and negative changes in food avail- 
ability, respectively, and the high correlation 
