Olesiuk. Prey consumption of Phoca vitulina 



507 



lebacks, hexagrammids, skates, northern anchovy, eel- 

 pouts, smelts, crabs, pricklebacks, gunnels, an unknown 

 invertebrate, shrimp, lamprey, birds and mussels. Uni- 

 dentified prey accounted for a mean of 1.0% of the 

 overall diet within estuaries. 



Estimates of the contribution of each prey to the 

 total diet are shown in Figure 9, A-C. The estimates 

 were obtained by integrating seasonal changes in dis- 

 tribution of seals (Fig. 3) with the seasonal changes in 

 diet (Fig. 8), as per Equation 1. Although both the diet 

 and abundance of seals varied regionally outside estu- 

 aries and among individual estuaries (see Olesiuk et 

 al., 1990b for details), further stratification of the abun- 

 dance and diet data by region and by individual estu- 

 ary yielded estimates very similar to those shown in 

 Figure 9, both in absolute terms (r-=0.996 on an arith- 

 metic scale) and in relative terms (r 2 =0.879 on a log- 

 log scale). The similarity can be explained by the fact 

 that the scat samples were collected from all but one 

 of 12 inhabited estuaries, from 47 widely distributed 

 non-estuary sites and in all months of the year, and 

 were thus representative of the entire population. 



It was not possible to establish the absolute vol- 

 umes of prey consumed from scats because the samples 

 often represented only a portion of whole scats. More- 

 over, the relative volume of each prey consumed in a 

 meal could not be reconstructed from the number and 

 size of otoliths recovered and otolith length — fish weight 

 keys (McConnell et al., 1984; Prime and Hammond, 

 1987; Harvey, 1989) because otolith recovery rates var- 

 ied widely among species (Fig. 7) and intra-specifically 

 depending on the size of prey (Olesiuk et al., 1990b). 

 Furthermore, analogous analyses could not be applied 

 to the other structures used to identify prey because 

 their recovery rates also differed both among and 

 within species (i.e., structures were recovered in pro- 

 portions different from that in which they occur in 

 whole prey). However, the lower and upper limits of 

 the importance of prey in the diet, irrespective of the 

 relative volumes of each prey consumed in meals, (see 

 Table 3) indicate that estimates were fairly robust to 

 violations of the equal-volume assumption. For ex- 

 ample, the lower limits indicated that gadoids and her- 

 ring combined accounted for no less than about half 

 (47.8%) the overall diet. Indeed, subsequent calcula- 

 tions indicated that the two prey accounted for at least 

 63.2% of the total diet, because in many cases they 

 were the only two prey species in a meal such that 

 both could not have been consumed in negligible quan- 

 tities. Upper limits for the remaining prey species 

 indicated that none accounted for more than 10% of 

 overall diet, and that only five families (salmonids, 

 batrachoids, hexagrammids, embiotocids, and cephalo- 

 pods) could have accounted for more than 5% of the 

 overall diet. 



NON-ESTUARIES 



n nnnn, 



GAD HER SAL MID HEX SFP CEP FLF SCP ROK OTH UND 



SPECIES 



GAD HER SAL MID HEX SFP CEP FLF SCP ROK OTH UND 

 SPECIES 



GAD HER SAL MID HEX SFP CEP FLF SCP ROK OTH UND 

 SPECIES 



Figure 9 



Contribution of prey to the overall diet ( left scale) and esti- 

 mated annual consumption (right scale) of important prey: 

 lAl outside of estuaries; iBl within estuaries; and (Ci for the 

 entire study area. Prey codes: GAD=gadoids; HER=herring; 

 SAL=salmonids; MID=plainfin midshipman; HEX=hexa- 

 gTammids; SFP=surfperches; CEP=cephalopods; FLF=flat- 

 fishes; SCP=sculpins; ROK=rockfishes; OTH=other prey; and 

 UND=unidentified prey Dashed lines partition gadoids into 

 hake (bottom) and all other species combined (top). 



