FISHERY BULLETIN: VOL. 82, NO. 2 



350 4O0 450 500 550 



FORK LENGTH (mm) 



 i i r 



600 650 



x 



o 



3 



300 



350 



400 



450 



500 



550 



600 



FORK LENGTH (mm) 



FIGURE 3. — Size distributions of Sebastes pinniger and S. 

 Ilavidus from summer (National Marine Fisheries Service) and 

 seasonal (Oregon Department of Fish and Wildlife) surveys. 



incomplete animals. These prey categories include 

 the two most important euphausiid species and 

 other major taxonomic groups (Table 3). Other 

 planktonic prey (e.g., copepods, chaetognaths, 

 pteropods) were occasionally present in the diet of 

 one or both species, but their contributions to the 

 overall diets were minor. Cephalopods did not 



TABLE 3. — The major prey categories used in the analysis of 

 diet variations and their respective size ranges found in the 

 stomachs of S. pinniger and S. flavidus. 



' Found in S flavidus stomachs only. 



occur in the diet of S. pinniger; thus only seven 

 prey categories were used for this species. 



We analyzed four factors that may affect the 

 diet of these two species: season, geographic area, 

 time of day, and size of fish. Each factor was 

 subdivided into four classes to elucidate the gen- 

 eral trends within each factor. Stomach content 

 data for all cruises were grouped into four sea- 

 sons, based on major periods in the hydrographic 

 regime on the continental shelf off Oregon (Huyer 

 et al. 1975; Huyer 1977): spring (March-May), 

 summer (June-August), fall (September- Novem- 

 ber), and winter (December-February). The collec- 

 tion stations for all cruises were divided into one 

 of four latitudinally defined shelf regions: North- 

 ern California-Southern Oregon (lat. 41° 00' to 

 43°50'N), Heceta Bank-Central Oregon (lat. 

 43° 50' to 45°00'N), Columbia Region (lat. 45°00' 

 to 47°00'N), and Northern Washington-Vancou- 

 ver (lat. 47° 00' to50°00'N). 



For the analysis of diel variation of feeding, the 

 local mean sampling time was adjusted to account 

 for latitudinal, longitudinal, and seasonal differ- 

 ences in daylight. Each collection time was stan- 

 dardized to an equinox day with 12 h between 

 sunrise and sunset, based on solar table values. 

 These adjusted collection times were assigned to 

 one of four time periods: morning (0800-1200 h), 

 early afternoon (1200-1600 h), late afternoon 

 (1600-1800 h), and night (1800-0700 h). Only a 

 small number of S. pinniger and S. flavidus were 

 collected at night despite extensive nighttime 

 trawling effort on several occasions during the 

 summer survey. 



Since the length distributions of the two species 

 were roughly normal (Fig. 3), dividing the length 

 range into four equal size groups would result 

 in disproportionately large sample sizes in the 

 middle size ranges. On the other hand, setting the 

 sample sizes of the four groups equal would result 

 in narrow size ranges around the mode. As neither 

 of these options seemed desirable, compromise 

 groupings were chosen. For S. pinniger, we used 

 the following size classes: <45 cm, 45-<50 cm, 

 50- < 55 cm, and s 55 cm. Similar size classes were 

 selected for S. flavidus but were offset 5 cm to 

 reflect the smaller mean size of this species. 



To test whether significant within-factor varia- 

 tion occurred in the diet of each species, contin- 

 gency tables were constructed comparing the 

 occurrence of food or a particular prey category 

 versus the absence of food or that prey category. 

 A variance test for homogeneity of binominally 

 distributed data (Snedecor and Cochran 1967) was 



274 



