Gudmundson et al Diet of Collorh/nus ursinus 



451 



commercial fishery trawls varied considerably 

 by sample type (Fig. 3). These findings highlight 

 the potential limitations of a single sample type, 

 such as scats, and demonstrate that a combina- 

 tion of scat and spew samples can provide a more 

 complete description of northern fur seal diet. 



The differences in prey occurrence that were 

 observed among northern fur seal scats and 

 spews in our study may have resulted from the 

 accumulation of irregularly shaped or large 

 prey parts in the stomach (Kiyota et al., 1999). 

 Large cephalopod beaks and fish bones accumu- 

 late near the pyloric sphincter and are regur- 

 gitated from the stomach, whereas small prey 

 parts pass through the digestive system and 

 are recovered in scats (Bigg and Fawcett, 1985), 

 The different rates of retention of prey parts in 

 northern fur seal stomachs would, therefore, 

 bias scat samples toward small prey species 

 or small bones of large prey species and spew 

 samples toward large prey and cephalopods. For 

 example, the F0% of Pacific salmon, a large 

 boned fish species, was higher in spews than in 

 scats. Kiyota et al. (1999) also observed a high- 

 er occurrence of Pacific salmon in spews than 

 in scats of juvenile male northern fur seals 

 from St. Paul Island. The FOVc we observed 

 for northern smoothtongue and Pacific sand 

 lance also support a prey species bias between 

 sample types; the small forage fish were among 

 primary prey species found in scat samples but 

 were absent from spew samples. 



In addition to differences observed between 

 sample types, we found that the occurrence of 

 primary prey species varied between islands. 

 Specifically, the FC/f of gonatid group I, Gb- 

 Bm, Gm-Gm, and walleye pollock varied consid- 

 erably between sample types and islands, and 

 significant interactions between sample type 

 and island were observed for a number of these 

 species. Walleye pollock was the highest occur- 

 ring prey in St. George Island scats, but the 

 F0% was significantly lower in spew samples. 

 On St. Paul Island, walleye pollock was the 

 dominant prey species, but the F09f did not 

 differ between scats and spews. Similarly, the 

 F0% of Gb-Bm at St. Paul Island did not differ 

 between sample types. However, the F0% of 

 Gb-Bm was significantly higher in spews than 

 in scats on St. George Island. The FCXf of gona- 

 tid group I differed both between sample types 

 and between islands; the occurrence of this 

 group was highest in spews, and in samples 

 from St. George Island. Overall, cephalopods 

 occurred at higher frequencies on St. George 

 Island than on St. Paul Island for both sample 

 types. The lone exception was Gm-Gm, which 

 had a lower frequency on St. George Island for 

 scats. Observed differences in cephalopod and 

 walleye pollock occurrence between St. George 



