432 



Fishery Bulletin 100(3) 



ity of detecting all prey species and of recovering remains 

 after consumption; these assumptions were violated to 

 some extent in our study because, like otoliths, passage 

 and identification of prey structures are taxon-specific 

 (Harvey, 1989; Cottrell et al., 1996; Marcus et al., 1998). 

 In captive experiments, herring (Clupea pallasi) were 

 identified by 11 structures recovered in scats (other than 

 otoliths), whereas smelt were represented only by verte- 

 brae (Cottrell et al., 1996). Smelt vertebrae cannot be used 

 to enumerate individuals, whereas several herring bones 

 commonly recovered in scats are unique or are structures 

 with definite sides (prootics, atlas and axis vertebrae), and 

 even highly eroded herring bone retain characteristics 

 identifiable to species. In contrast, bones of some taxa 

 such as pleuronectids erode rapidly, losing species traits 

 and are identified only to family. Given these factors, the 

 identification of all skeletal structures represents the rela- 

 tive consumption of some prey more accurately than oth- 

 ers. Further, MNI estimated from all skeletal structures is 

 not corrected for complete digestion of structures useful for 

 enumerating individuals. To date, there are no correction 

 factors for complete digestion of bones and this lack, doubt- 

 lessly, is a source of substantial bias. 



Behavior of both predator and prey also affects identifi- 

 cation and enumeration of prey remains in scats. Small, 

 schooling fish, such as smelts, are more likely to be con- 

 sumed in greater numbers than larger, solitary fish such 

 as hexagrammids. Smelts are most frequently identified 

 from vertebrae; therefore MNI is more severely under- 

 estimated because more than one individual is likely to 

 be consumed. Captive feeding studies also have indicated 

 that the activity of the pinniped, its meal size, size of prey, 

 and the physical structure of the prey bone all affect pas- 

 sage rate and the degree of erosion (Cottrell et al., 1996; 

 Tollit et al., 1997; Marcus et al., 1998; Bowen, 2000). 



The estimation of MNI from all prey structures recov- 

 ered in scats presents a variety of complications; however, 

 the alternative — using otolith correction factors — also 

 has problems. Otolith correction factors based on recov- 

 ery rates from feeding experiments are highly variable 

 between repeated trials of the same individuals, differ- 

 ent individuals, and different pinniped species (Harvey, 

 1989; Cottrell et al, 1996; Tolht et al., 1997; Bowen, 2000). 

 Although his results were inconclusive, Bowen (2000) sug- 

 gested that differences in activity levels may account for 

 much of the variability in digestion (and correction fac- 

 tors). Activity levels among wild harbor seals are likely to 

 be more variable than those between captive harbor seals 

 with and without access to water, and thus otolith correc- 

 tion factors may yield erroneous estimates of individuals 

 consumed by free-ranging predators. Limiting analyses of 

 diet to qualitative measures such as FO will reduce the bi- 

 ases of including bone identification; however, the overall 

 importance of frequent, small prey may be much less than 

 indicated by their relative frequency. 



Pinnipeds are considered generalist feeders and may 

 feed on large numbers of abundant, frequently encoun- 

 tered prey; however, if mass is considered a measure of im- 

 portance, they may be sustained by infrequent, large prey 

 (Tables 4—6). A few prey species were both abundant and 



frequent (herring, smelts, sculpins, flatfish), yet their esti- 

 mated masses were small. Large, infrequent species such 

 as lingcod, hake, rockfish, and salmon may contribute more 

 total mass to a hypothetical "meal." Unfortunately, these 

 prey were also poorly represented by otoliths in our study 

 and therefore our mass estimates may be inaccurate. 



All methods of examining marine mammal diets, such 

 as fecal analyses, stomach lavage, or stomach content 

 analysis, are inherently biased to some degree. Biases of 

 fecal analyses have been discussed at length in the lit- 

 erature; however, fecal analysis remains the least inva- 

 sive and least expensive technique and allows for large 

 sample sizes. Identification of all skeletal elements, rather 

 than otoliths exclusively, is an improvement on other tech- 

 niques. Although results are still subject to biases, prey 

 taxa represented by hard parts in fecal material represent 

 a minimum estimate of prey consumed. In addition, an ex- 

 amination of skeletal elements other than otoliths is man- 

 datory for assessing the impact of harbor seals on certain 

 prey species — protected salmon stocks, for example. 



Acknowledgments 



The authors would like to thank M. Gosho, B. Hanson, H. 

 Huber, K. Hughes, S. Melin, and L. Lehman for their assis- 

 tance in the field and laboratory. They thank S. Reimer 

 and W. Walker for assistance with prey identifications. This 

 manuscript was improved by comments from S. Mizroch, D. 

 Withrow, and the comments of three anonymous reviewers. 



Literature cited 



Beamish, R. J. 



1980. Adult biology of the river lamprey (Lanipetra ayresi) 

 and the Pacific lamprey (Lampetra tridentata) from the 

 Pacific coast of Canada. Can. J. Fish. Aquat. Sci. 37:1906- 

 1923. 

 Bowen, W.D. 



2000. Reconstruction of pinniped diets: accounting for com- 

 plete digestion of otoliths and cephalopod beaks. Can. J. 

 Fish. Aquat. Sci. 57:898-905. 

 Boyle. P R., G. J. Pierce, and J. S. W. Diack. 



1990. Sources of evidence for salmon in the diet of seals. 

 Fish. Res. 10:137-150. 

 Brown, R. F 



1980. Abundance, movements and feeding habits of the 

 harbor seal, Phoca vitulina, at Netarts Bay, Oregon. M.S. 

 thesis, Oregon State Univ., Corvallis, OR, 69 p. 

 Cottrell, P E., A. W. Trites, and E. H. Miller 



1996. Assessing the use of hard parts in faeces to identify 

 harbour seal prey: results of captive-feeding trials. Can J. 

 Zool. 74:875-880. 

 Dark, T. A., and M. E. Wilkins. 



1994. Distribution, abundance, and biological characteris- 

 tics of groundfish off the coast of Washington, Oregon, and 

 California, 1977-1986, 73 p. U.S. Dep. Commer., NOAA 

 Tech. Rept. NMFS-117. 

 Eschmeyer, W. N., and O. W. Herald. 



1983. A field guide to Pacific coast fishes of North America, 

 336 p. Houghton Mifflin Co., Boston, MA. 



