Browne et al.: Improving pinniped diet analyses 



427 



The inclusion of all skeletal elements recovered from 

 scat increased the MNI and FO of all harbor seal prey taxa 

 (Table 4). The FO more than doubled for most taxa and 

 usually was more affected by including all prey elements 

 than was the average MNI (Table 4). We compared the 

 MNI of several common harbor seal prey estimated from 

 all structures to an estimate based on the number of re- 

 covered otoliths multiplied by a species-specific correction 

 factor for recovery rate (accounting for complete digestion 

 of the otolith; Harvey, 1989, Fig. 1 ). A value of 1.0 indicated 

 that the same estimate was derived from both methods, 

 whereas 0.5 indicated that the MNI estimated from all 

 structures was twice the estimate from otoliths. 



Seasonal variation was also apparent in estimated 

 prey mass (Table 5). In some instances, estimates were 

 based on very few otolith measurements, values were 

 taken from the literature, or mass was averaged from 

 other seasons when no intact otoliths were recovered 

 (Table 51. Because some species were difficult to discern 

 or regression relationships were unavailable, species 

 were grouped by phylogeny or size similarities (Table 5). 

 Smelts (Osmerids) were pooled by family and mass was 

 estimated from whitebait smelt (Allosmerus elongatus), 

 the most abundant species by distinguishable otoliths, 

 and eulachon (Thaleichthys pacificus), although longfin 

 smelt (Spirinchus thaleichthys) and surf smelt (Hypome- 



sus pretiosits) were occasionally identified in harbor seal 

 scats. Although smelt otoliths could be distinguished by 

 species, smelt bone could not. Smelt mass estimates were 

 based on eulachon (in their relative proportion) and white- 

 bait smelt because although less common than the other 

 three similar size species, eulachon were much larger 

 Masses of juvenile and adult salmonids were estimated 

 separately for five species identified in harbor seal scat, 

 with the exception of sockeye salmon, which was repre- 

 sented by one otolith from an adult fish. Mass for sockeye 

 was based on regressions generated for silver salmon. 

 Cutthroat (Onchorhyncus clarki) and steelhead salmon 

 (O. mykiss) were not often distinguishable by otoliths, 

 and mass estimates were based on steelhead because of 

 their numerical dominance in the lower Columbia River. 

 The sticheid-pholid group included fish from a variety of 

 families: three spine stickleback {Gasterosteus aculeatus), 

 snake prickleback (Lumpetius sagitta). high cockscomb 

 (Anoplarchus purpurescens), wattled eelpout (Lycodes pa- 

 learis). Pacific sandfish (Trichodon trichodon), and saddle- 

 back gunnel iPholis ornata). Although not taxonomically 

 related, these species were seldom represented by otoliths, 

 individuals were very small, and with the exception of 

 gunnels, were rare by number and frequency (Table 4). 

 Little has been published about relationships between 

 otolith length and fish length or fish length and mass for 



