Figure 1.— Location of Auke Bay and 

 Fritz Cove, southeastern Alaska, where 

 stomachs were collected from walleye 

 pollock in 1979-80. 



has its biases. Therefore, I chose to analyze the 

 stomach contents of walleye pollock in two ways: 1) 

 Percent by volume of major food categories to show 

 the contribution in biomass of the foods, and 2) per- 

 cent frequency of occurrence to show the diversity of 

 foods at lower taxonomic levels. 



To determine percent by volume, I either measured 

 or estimated the volume of major food categories 

 (e.g., mysids, shrimp, or fish) in each stomach, 

 depending upon the amount of food present. In 

 walleye pollock with stomach contents about >0.5 

 ml, the displacement volume of food in each major 

 category was measured in a graduated cylinder and 

 then expressed as a percentage of the total volume of 

 the stomach contents. In fish with stomach contents 

 <0.5 ml, food was divided by categories into piles of 



uniform height in a petri dish and placed over a grid. 

 The number of grid squares covered by each pile was 

 used to estimate the volume percentage of each food 

 category. I then pooled percent-volume measure- 

 ments and estimates for all fish and calculated mean 

 percentages for each food category. Walleye pollock 

 with empty stomachs or with stomachs containing 

 only a trace of food were excluded from these 

 calculations. Many stomachs contained a large per- 

 centage of flocculent, digested matter or an indis- 

 tinguishable mixture of crustacean parts; these items 

 were termed "unidentified digested matter" and 

 "unidentified crustacean fragments," respectively. 



To determine the frequency of occurrence of foods, 

 each item present in a stomach was identified to the 

 lowest practical taxon. Percent frequency of oc- 



638 



