668 
Fishery Bulletin 95(4), 1 997 
C A ,= 
{.F h -F h -e~^yRjt 
1-e 
R,t 
where C At 
F. and F. 
l 2 
R. 
t 
food consumption between sampling 
periods at time t 2 and 1 7 ; 
the geometric mean gut-fullness values 
(back-transformed from log e (x+1)) at 
these time points (time points with n < 
3 fish were not used in daily ration cal- 
culation); 
the exponential gastric evacuation rate; 
and 
t 2 -t v 
Daily ration was calculated by summing estimates 
ofC Ar 
The method of Boisclair and Marchand (1993) and 
Trudel and Boisclair ( 1993) was used to estimate 95% 
confidence intervals for daily ration estimates. There 
were four steps in the analysis. First, an estimate of 
exponential evacuation rate (R g ) was made from the 
average water temperature during a given sampling 
period. Estimates of R e were calculated from the 
equation 
R e = 0.015e (0103T) , 
where T = water temperature (°C ) from Buckel and 
Conover (1996). 
Periods of declining gut fullness can be used as a 
validation of laboratory-based gastric evacuation 
rates (see Parrish and Margraf, 1990) and were used 
for seven out of ten diel beach-seine collections with 
the same data analysis techniques as those described 
in Buckel and Conover (1996). The mean field-derived 
estimate of R g for these dates was 0.24 1/h, and the labo- 
ratory-derived estimate was 0.20 1/h (± 0.038 SE). 
Second, a normal distribution of 1,000 pseudo val- 
ues of R were calculated as 
e 
R e *=R e + (SE Re xRN), 
where R = 
K = 
Re ~ 
SE 
RN = 
the pseudo value of evacuation rate; 
the estimated mean evacuation rate; 
the standard error of R e (Buckel and 
Conover 1996); and 
a random number (different for each 
calculation) from a normal distribution 
with a mean of 0 and standard devia- 
tion of 1. 
Third, a normal distribution of 1,000 pseudo values of 
gut fullness (F) were calculated for each time point as 
F t *=F t + (SE Ft x RN), 
where F f * = the pseudo value of gut fullness; 
F t = the mean log e (F+l) transformed gut 
fullness; 
SE pt = the standard error of F f ; and 
RN = a random number (different for each 
calculation) from a normal distribution 
with a mean of 0 and standard devia- 
tion of 1. 
Fourth, Monte-Carlo simulations were used to esti- 
mate C At from the above equations by randomly choos- 
ing values of R e *, F t *, and F t * from the distributions 
of 1,000 pseudo values (values of F t * and F t * were 
back-transformed before calculation of C At ). Simulated 
values of C At were generated for each of the eight time 
intervals (nine sampling points; less if a time point 
had n < 3 fish) and summed to estimate a daily ra- 
tion. This calculation was repeated 1,000 times. The 
2.5 and 97.5 percentiles of these daily ration esti- 
mates were taken as the 95% confidence intervals. 
Diet analysis 
Diets of bluefish captured with beach seines, surface 
trawls, and gill nets were quantified. In the labora- 
tory, bluefish were measured for total length (TL, ± 
1.0 mm), weighed (± 0.01 g), and their stomachs were 
extracted. Stomach contents of bluefish were identi- 
fied to the lowest possible taxon, enumerated, blot- 
ted dry, weighed (± 0.01 g), and measured (TL, ± 1.0 
mm; eye diameter, ±0.1 mm; caudal peduncle depth, 
±0.1 mm). Regressions relating prey eye-diameter 
and caudal peduncle depth to TL were used to esti- 
mate prey TL from prey pieces (see Scharf et al., 
1997). A reference collection of Hudson River fish 
species (whole fish, scales, and bones) was used to 
aid in identification of digested prey. Two indices were 
computed to describe diets (see Hyslop, 1980). The 
indices were 1 ) number of stomachs in which a taxon 
was found, expressed as a percentage of the total 
number of stomachs containing food (%F=percent 
frequency of occurrence), and 2) weight of taxon, ex- 
pressed as a percentage of the total weight of food 
items (%W=percent weight). 
Results 
Die! collections — beach seine 
A total of 1,204 spring-spawned and 64 summer- 
spawned bluefish were collected during diel beach- 
seine collections. There were five successful diel col- 
