706 



Fishery Bulletin 92(4). 1994 



Data partitioning 



A forage record was defined as the forage data spe- 

 cific to a focal animal and was used as the sample 

 unit in comparisons of prey composition, forage suc- 

 cess, and the mean number of prey captured per dive. 

 For assessing variation in prey composition and for- 

 age success, only forage records containing >10 for- 

 age dives were used; adults of unknown sex were de- 

 leted in comparisons of sex classes. Sample sizes for 

 juveniles were small and created an unbalanced sample 

 design in 2-way comparisons. Consequently, separate 

 tests were conducted to assess age-class differences. 



For comparisons of prey composition, we calculated 

 the proportion of dives resulting in the capture of 

 clams, crabs, and mussels for each forage record. 

 Differences in the proportion of prey items captured 

 by adult sea otters were tested among areas. Sample 

 sizes were insufficient to test prey composition dif- 

 ferences among areas for juveniles. Data were pooled 

 from all study areas and the proportion of prey cap- 

 tured was tested by age class. 



Forage success (the proportion of successful dives) 

 was normalized by an arcsine transformation of the 

 square root. Differences in forage success among 

 study areas and among adult sex classes (male, fe- 

 male, and females with pups) were tested. Sample sizes 

 were insufficient to test for differences among study 

 areas for juveniles. Data were pooled for all juveniles 

 and all adults to test age differences in forage success. 

 Number of prey items captured per dive was cal- 

 culated by dividing the total number of prey captured 

 by the number of forage dives per foraging record 

 and averaging these values by sex class and area. 

 Dives resulting in the capture of mussels (which 

 may be difficult to count) and dives of unknown 

 result were excluded. 



We assumed mean shell lengths of 4.0, 7.0, and 

 10.0 cm were representative of small, medium, and 

 large bivalve size classes, then estimated mean 

 wet-tissue mass of Saxidomus giganteus by using 

 the weight-length relationships generated by 

 Kvitek et al. ( 1992). We estimated caloric gain per 

 dive by using caloric values for this genus reported 

 byKenyon(1969). 



Data analysis 



Kruskal-Wallis nonparametric ( 1-way) tests were 

 used to assess differences in the proportion of 

 clams, mussels, and crabs captured among study 

 areas by adult sea otters; data were pooled for all 

 study areas and the proportion of clams, mussels, 

 and crabs were tested by age class. Analysis of 

 variance (2-way AN OVA) was used to test 1) dif- 

 ferences in forage success among study areas and 



adult sex classes, and 2) differences in the mean 

 number of prey captured per forage dive among study 

 areas and adult sex classes. A 1-way ANOVA was 

 used to test differences in the mean number of prey 

 captured per dive among study areas for juvenile sea 

 otters. A Student's t-test was used to test differences 

 in forage success between adult and juvenile sea ot- 

 ters for all study areas combined. For all compari- 

 sons, significance was set at a=0.05. 



Results 



Sea otters were observed foraging on clams (57-67%), 

 mussels (19-25%), crabs (2-A%) and green sea ur- 

 chins, Strongylocentrotus droebachiensis (0-3%) (Fig. 

 2). Clams were identified to species in 23% («=535), 

 65% (rc=957), and 63% (« = 1,060) of the observations 

 in established, intermediate, and frontal areas, re- 

 spectively. The majority of clams identified were 

 Saxidomus in established (98%), intermediate (89%), 

 and frontal (96%) areas. Other clams identified (<10% 

 per study area) were Tresus capax, Mya spp., 

 Protothaca staminea, and Entodesma macroschisma. 

 Mytilus spp. was the most common mussel observed 

 within the study areas. Crabs were primarily Tel- 

 messus spp.; however, a small number of Cancer mag- 

 ister, were recorded. Other prey which contributed 

 from <1 to 7% of the diet in each study area included 

 Clinocardium spp., Cucumaria fallax, Echiurus 

 echiurus alaskensis, Niicella spp., Octopus spp., Pisa- 

 ster spp., Pycnopodia helianthoides, barnacle (class 

 Crustacea), chiton (class Polyplacophora), tunicate 



 Established (N = 798) 



□ Intermediate (N = 1694) 



□ Frontal (N = 1852) 



M 



r~r~i 



JL 



b£dl 



Clam Mussel Crab Sea Urchin Other Unidentified 



Figure 2 



Frequency of occurrence of food items obtained by sea ot- 

 ters, Enhydra lutris, as determined by visual observation 

 along the Kodiak Archipelago during 1987 and 1988 in 

 areas of established (>25 yr), intermediate (5-15 yr), and 

 frontal (<5 yr) sea otter forage areas. 



