walleye with either a 38.1 x 1.8 m sinking gill net 

 with multifilament, variable, stretched mesh of 

 3.81, 5.08, 6.35, 7.52, and 10.16 cm, or a 76.2 x 3.7 

 m monofilament, floating gill net with 15.25 cm 

 stretched mesh. All gill net sets were set at a 

 maximum of 2.5-h duration in order to minimize 

 regurgitation or digestion of stomach contents of 

 the walleye. In 1981 we used gill nets and a 6.15 m 

 electroshock boat with a 3,500-W generator and 

 front mounted electrodes utilizing pulsed direct 

 current at 1-4 A. Potential prey fish were periodi- 

 cally sampled with a 30.48 x 2.44 m beach seine of 

 6.35 mm stretched mesh. When we caught poten- 

 tial prey by means of gill nets, seines, or elec- 

 tioshock gear, we recorded numbers and fork 

 lengths by species. Gear selectivity prohibited re- 

 liable estimates of numerical abundance of 

 species, however, we used catch per unit of effort 

 (CPUE) to estimate change in intraspecific abun- 

 dance through time. 



For each walleye captured, we recorded fork 

 length (FL, mm), weight (g), sex, and stage of 

 maturity, took scale samples, and preserved the 

 stomach in 10% buffered Formalin 3 . Subsequently, 

 each stomach was examined and each prey item 

 was identified to the lowest possible taxon and its 

 volume was recorded. A reference bone collection 

 of potential prey species aided in the identification 

 of partially digested prey. The most useful bones 



were pharyngeal teeth, opercles, preopercles, and 

 jaw bones. Characteristics of the internal mor- 

 phology, e.g., the black peritonium of bridgelip 

 suckers, Catostomus columbianus, or the number 

 of pyloric ceca in salmonids (Scott and Crossman 

 1973), were also useful in identifying prey items. 

 We separated stomach content data into sub- 

 populations based on season and year of capture, 

 and tested for statistically significant differences 

 in numbers and volumes of individual prey items. 

 We computed statistical significance using a non- 

 parametric, multivariate test, L Nt , which has ap- 

 proximately a chi-squared distribution withp (v — 

 1) degrees of freedom, where p is the number of 

 conditions (prey taxa) and v is the number of popu- 

 lations (Koch 1969). To identify changes in the 

 importance of food items we examined changes in 

 the Index of Relative Importance (IRI), which is 

 equal to the sum of the percent by volume and the 

 percent by number, multiplied by the percent fre- 

 quency of occurrence (Pinkas et al. 1971). 



Results 



Seasonal Diet 



3 Reference to trade names does not imply endorsement by the 

 National Marine Fisheries Service, NOAA. 



The walleye size ranges were similar for both 

 years, about 200-750 mm FL (Fig. 2 ). In both years 

 fish accounted for over 99% of the total prey vol- 

 ume (Tables 2, 3). Based on IRI (Table 4), prickly 

 sculpin, Cottus asper, was the most important 

 species found in walleye stomachs. Excluding un- 



TABLE 2. — Percent by volume, percent by number, and percent frequency of occurrence 

 of foods found in the stomachs of walleye collected in the John Day pool of the Columbia 

 River April-September 1980. Sample size equals 189 walleye, with 38.1 r r empty 

 stomachs. iRaw data are in parentheses.) 



'Volumes and numbers of individual prey taxa were significantly different from those of 1981 (P - 005). 



412 



