444 



Fishery Bulletin 100(3) 



corrected estimates of MNI for several species. The ratios 

 were quite variable but the comparison did suggest that 

 smaller prey such as smelt were more likely to be under- 

 represented by using the minimum count. 



If diet composition is based on sagittal otoliths, a fish 

 can be represented by, at most, two scats and because all 

 fish have two sagittal otoliths, fish size should not influ- 

 ence the probability that a particular fish is included, in a 

 sample of scats, except through size-specific passage rates 

 of otoliths. However, when all hard parts are included, the 

 sampling may be size-biased if the size of the prey affects 

 the number of scats in which the hard parts are deposited. 

 Larger prey contain larger hard parts that may require 

 longer passage times; therefore larger prey may be depos- 

 ited in more scats than smaller prey Also, large prey may 

 be shared among seals as a result of cooperative feeding 

 behavior and could be deposited in several scats. If either 

 situation occurs, larger prey would be more likely to be 

 included in the sample and would be over-represented. 

 Because the scat is the sampling unit, any prey-size or 

 species-specific effects on scat deposition rate may also 

 bias diet composition estimators. 



The effect of over-representing large prey depends on 

 the estimator used for diet composition. The different 

 outcomes with Equations 1-3 can be demonstrated with a 

 simple example. Consider a sample of two scats in which 

 one scat contains the remains of a 2-kg salmon and an- 

 other scat contains the remains of ten 10-g anchovy and 

 four 100-g herring. From Equation 1, the diet composition 

 would be 80% (2000/2500) salmon, 4% anchovy and 16% 

 herring based on proportions of total reconstructed bio- 

 mass. From Equation 2, we would estimate that salmon 

 represent 50% of the diet from the two samples that are 

 100% and 0% salmon, and likewise 10% anchovy and 40% 

 herring. Finally from Equation 3, we would estimate that 

 the diet was 50% salmon, 25% anchovy, and 25% herring. 

 If the small prey were undercounted in relation to the 

 large salmon, the influence of the error influences the 

 composition within the scat for Equations 2 and 3, but for 

 BR (Eq. 1) the error extends across all samples. 



As with the Columbia River harbor seal example (Fig. 3), 

 the differences in the estimators are primarily the result 

 of large prey in the weighted versus unweighted averages. 

 Some difference would be expected in the results of Equa- 

 tions 2 and 3 depending on the validity of the equal volume 

 assumption. SSFO (Eq. 3) simplifies the analysis of diet 

 composition to a measure of presence and absence by as- 

 suming that prey within the same scat were consumed in 

 equal volumes. The simplifying assumption removes the ne- 

 cessity to enumerate prey and measure mass from morpho- 

 metric relationships with prey remains. However, the equal 

 volume assumption does not seem particularly reasonable 

 and its implementation is arbitrary, depending on how the 

 prey are classified unless all prey remains can be identified 

 to species. Olesiuk ( 1993) showed that the diet composition 

 percentages for the primary prey varied by a factor of two or 

 three, depending on the assumed composition within each 

 scat. We expected that these differences would depend on 

 the diversity of the diet. How closely they represented the 

 true diet would depend on the range in prey sizes. 



From our viewpoint, we do not see a clear choice between 

 the estimators for diet composition. The use of consump- 

 tion estimates from SSFO and BR to provide a range of es- 

 timates may have limited application in cases where each 

 approach would suggest a similar conclusion. However, for 

 large prey, such as salmonids, a tenfold difference in esti- 

 mates, compounded with the uncertainty from sampling 

 and biomass estimation, may yield too little information to 

 develop a reliable conclusion about the impact of pinniped 

 predation on salmonid stocks. 



Acknowledgments 



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

 Hughes, S. Melin, and L. Lehman for their assistance in 

 the field and laboratory. The authors would also like to 

 thank P. Olesiuk for his encouragement and for succintly 

 demonstrating the problems with the minimum number 

 computation in biomass reconstruction. This manuscript 

 was greatly improved by comments from J. Jansen, P. 

 Boveng, and four anonymous reviewers. 



Literature cited 



Bigg, M. A. 



1969. The harbour seal in British Columbia. Fish. Res. 

 Board Can. Bull. 172,31 p. 

 Bigg. M. A., and M. A. Perez. 



1985. Modified volume: a frequency-volume method to 

 assess marine mammal food habits. In Marine mammals 

 and fisheries (J. R. Beddington, R. J. H. Beverton, and D. 

 M. Lavigne, eds.), p. 277-283. George Allen & Unwin, 

 London. 

 Bigg, M. A., G. Ellis. P. Cottrell, and L. Milette. 



1990. Predation by harbour seals and sea lions on adult 

 salmon in Comox Harbour and Cowichan Bay, British 

 Columbia. Can. Tech. Rep. Fish. Aquat. Sci. 1769, 31 p. 

 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. 

 Browne, P.. J. L. Laake. and R. L. DeLong. 



2002. Improving pinniped diet analyses through identifica- 

 tion of multiple skeletal structures in fecal samples. Fish. 

 Bull. 100:423-433. 

 Cochran, W. G. 



1977. Sampling techniques, 3'''' ed., 428 p. John Wiley and 

 Sons, New York, NY. 

 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. 



Forney, K. A., J. Barlow, M. M. Muto, M. Lowry, J. Baker, 

 G. Cameron. J. Mobley, C. Stinchcomb, and J. V. Carretta. 

 2000. U.S. Pacific marine mammal stock assessments: 2000. 

 U.S. Dep. Commer., NOAA Tech. Memo. NMFS-SWFSC- 

 300, 276 p. 

 Hammill, M.O., C. Lydersen, K. M. Kovacs, and B. Sjare. 



1997. Estimated fish consumption by hooded seals iCys- 

 tophora cristata), in the Gulf of St. Lawrence. Northwest 

 Atl. Fish. Sci. 22:249-257. 



