424 



Fishery Bulletin 100(3) 



skeletal elements resulted in at least two times greater 

 frequency of occurrence of some prey taxa than frequen- 

 cies exclusively derived from otoliths (Riemer and Brown, 

 1997; Boyle et al., 1990; Cottrell et al., 1996). For major 

 prey taxa, we estimated frequency of occurrence and mini- 

 mum number of individuals from fish otoliths and other 

 skeletal remains recovered from scats and average mass 

 from otoliths. 



Methods 



During 1995, 1996, and 1997, scats were collected from 

 Desdemona Sands (river km 26, 123°52'W,46°13'N), the 

 largest harbor seal haul-out site in the lower Columbia 

 River (Huber'^). Scats were collected intermittently during 

 1995. From March through August 1996 and from March 

 through October 1997, we attempted to collect 50 harbor 

 seal scats every two weeks at extreme low tides. This sam- 

 pling period coincided with Columbia River runs of spring, 

 summer, and fall chinook salmon. Scats were collected 

 from haul-outs, and upon arrival at the laboratory were 

 rinsed in nested sieves (2-mm, 1-mm, and 0.5-mm mesh 

 width). All skeletal elements were recovered, dried, and 

 stored in vials. Cephalopod remains were stored in 70% 

 isopropyl or ethyl alcohol. Other invertebrate remains 

 were relatively rare (<2'7r frequency of occurrence) and 

 their contribution to the diet was disregarded because of 

 difficulties enumerating individuals and determining pri- 

 mary from secondary (prey within large, ingested fishes) 

 prey. Otoliths were identified to lowest possible taxon, 

 their anatomical location recorded (left or right side), and 

 enumerated (number for left side and number for right 

 side). Lengths of intact left or right otoliths were mea- 

 sured parallel to the sulcus to the nearest 0.1 ram with 

 an ocular micrometer Micrometer measurements were 

 verified with hand-held calipers. Other skeletal structures 

 (such as teeth, vertebrae, and cranial bones) were identi- 

 fied to lowest possible taxon by comparing prey remains to 

 reference samples (NMMLM. 



Scat collections were divided into three seasons: spring 

 (samples collected prior to 15 May), summer (samples col- 

 lected from 15 May to 15 July), and fall (samples collect- 

 ed after 15 July). These dates distinguish runs of spring, 

 summer, and fall chinook salmon crossing the Bonneville 

 Dam (river km 235), less two weeks estimated for travel 

 from the lower Columbia River (Fryer, 1998). For each sea- 

 son, harbor seal diet was described by frequency of occur- 

 rence (FO), minimum number of individuals (MND, and 

 average prey mass estimated from otoliths of all major 

 prey taxa. Frequency of occurrence (FO) of prey taxon^' in 

 season k was defined as 



£o,M 



FO, = - 



where O,^^, = a binary variate indicating presence (1) or 

 absence (0) of taxon 7 in sample i in season k; 

 and 

 s^ = the total number of scats containing identifi- 

 able prey remains in season k. 



Rare prey taxa were grouped with similar taxa for analy- 

 ses. Unknown prey remains that were clearly distinct from 

 known taxa were considered "unidentified taxa" in sam- 

 ples containing "identified" hard parts. Scats containing 

 skeletal remains considered "unidentifiable," i.e. extremely 

 eroded bone or fragmented material, were excluded from 

 analyses. The minimum number of individuals (MND was 

 estimated from the greatest number of left or right otoliths 

 and unique or paired bone structures and expressed within 

 each season as total MNI or average MNI per scat (total 

 MNL/number of scats collected). Presence of non-unique 

 fish remains (non-unique vertebrae, gillrakers, teeth) con- 

 stituted a single individual. For example, if a scat sample 

 contained five left otoliths, three right otoliths, and six 

 atlas or axis vertebrae of a prey taxon, the MNI was six. 

 FO and MNI were calculated from otoliths and again from 

 all prey remains. Prey masses were estimated for the three 

 seasons from allometric relationships between otoliths and 

 body size (Harvey et al., 2000; Table 1). If relationships 

 were unavailable for a species, regressions generated for a 

 similar species were used. Otolith lengths were multiplied 

 by a species-specific correction factor when available or an 

 average correction factor to account for reduction in length 

 due to digestion (Harvey, 1989). All intact left or right oto- 

 liths of a prey taxa were measured, and estimated masses 

 were averaged for each season. 



Suitable morphometric regressions were not available 

 for several salmon species or did not include juvenile fish; 

 therefore, we generated regressions including subadult 

 age classes specifically for our study. In addition to pub- 

 lished regressions, relationships between salmon otoliths 

 and fish mass used in this study were calculated from Na- 

 tional Marine Mammal Laboratory reference samples, the 

 private collections of Walker* and NRC^ (Table 1 ). Because 

 of the large discrepancy in masses of adult and juvenile 

 salmonids, otoliths were identified to species and classified 

 as adult or juvenile according to species-specific lengths 

 estimated from regression equations. "Adults" described 

 all returning upriver migrants, including reproductively 

 mature individuals and jacks. "Juveniles" were seaward 

 migrants and may have included two-year-old fish of 

 some species (Groot and Margolis, 1991). Onchorhyncus 



^Huber, H.R. 1997. Unpubl.data. NationalMarine Mammal 

 Laboratory, 7600 Sand Point Way NE, Seattle, WA 98115. 



' NMML(NationalMarineMammalLaboratory). 1997. Marine 

 Mammal Prey Osteological Reference Collection. National 

 Marine Mammal Laboratorv, 7600 Sand Pomt Way NE, Seattle, 

 WA 98115. 



•• Walker. W. 1998. Unpubl. data. National Marine Mammal 

 Laboratory, 7600 Sand Pomt Way NE, Seattle, WA 98115. 



■> NRC (National Resources Consultants). 1998. Unpubl.data. 

 Natural Resources Consultants, Lie, 4055 2P' Ave. W, Suite 

 100, Seattle, WA 981 19. 



