proportionally fewer diet samples), prey mass was 
weighted on the basis of the catch rates at each station by 
using this equation: 
, 
W, x CPUE, 4 @) 
; Ns ag 
where W, = the mass of prey taxon 7; 
CPUE., , , = the catch per unit of effort for age a of species 
s at sampling station g (number of individuals 
per 250 m?); and 
aq = the number of stomachs analyzed for age a of 
species s at station q. 
S,a,q 
Ns, 
Fish age groups were based on the length—frequency 
modes provided in Hurst (2016). 
The prey-specific index of relative importance (PSIRI) 
(Brown et al., 2012) was calculated for each identified 
prey type to distinguish prey species and prey habitat 
categories that dominated the diets of northern rock sole 
and yellowfin sole. The PSIRI accounts for biases inher- 
ent in the standard index of relative importance of fre- 
quently occurring prey (Brown et al., 2012). The PSIRI 
is additive; therefore, it clarifies representation of the 
various levels of taxonomic identification inherent in diet 
studies: 
pics 2N;; a es TW, 
n n; nj 
PSIRI = 5 , (2) 
nN; 
where %N;; = the weighted number of prey category 7 in 
stomach /; 
%W;; = the weighted cumulative mass of prey cate- 
gory z in stomach /; 
n = the total number of stomachs with identifi- 
able prey; and 
n;=the total number of stomachs containing 
prey 1. 
Dietary overlap coefficients were calculated between 
flatfish species by using samples from all sampling sta- 
tions and from only those stations where the species 
co-occurred. Dietary overlap was based on the Schoener 
Index (SI), which was calculated as follows: 
SI =1-—(¥. | >) - Pal). (3) 
where p,; = the proportion by weight of prey 7 by group J; 
and 
Dix = the proportion by weight of prey z by group k. 
The general patterns of dietary overlap are based on 
analyses across the full size range (35-99 mm SL) 
(Fig. 2). We also examined the potential effect of size 
variation on dietary overlap by using analyses that 
examined overlap in 2 size classes (small: <66 mm SL; 
large: >66 mm SL) and by restricting data used in anal- 
ysis to overlapping sizes observed in the study area 
(44-90 mm SL). 
Fishery Bulletin 120(1) 
Species 
4 YFS 
e NRS 
60 80 
Standard length (mm) 
7) 
7) 
fo)) 
fe 
fe 
(3) 
i= 
rm) 
1) 
(3) 
®o 
Q 
ep) 
100 200 300 400 500 
Sample size (mm) 
Figure 2 
(A) Length-frequency distributions and (B) length— 
weight relationships of yellowfin sole (Limanda aspera) 
(YFS) (light gray area and dashed gray line) and north- 
ern rock sole (Lepidopsetta polyxystra) (NRS) (dark gray 
area and solid black line) that were collected in a coastal 
nursery in the Bering Sea in August 2012 and that had 
their stomach contents analyzed. (C) Estimated species 
richness for prey in diets of each species based on the rela- 
tionship of species richness to sample size. In panel C, for 
YFS (gray) and NRS (black), the circles indicate the total 
number of samples, the solid lines indicate interpolated 
estimates, and the dotted lines indicate extrapolated esti- 
mates; the light and dark gray bands indicate the 95% 
confidence intervals. 
