96 



FISHERY BULLETIN OF THE FISH AND WILDLIFE SERVICE 



"frequency of occurrence," and the "volumetric." 

 Each has its shortcomings, some of which are 

 inherent in the nature of the problem, and at best 

 it can afford only a rough indication of the food 

 habits of the animal. 



The numerical system is based solely on a 

 count of organisms present, with each food ele- 

 ment evaluated as a percent of the total number of 

 all elements. This method tends to place undue 

 emphasis on food organisms with very resistant 

 parts. In summing up the food of a number of 

 individuals, instead of getting a cross section of 

 the most recently obtained food, a record is 

 obtained of the more durable elements of past and 

 recent food and a false idea of the food may result. 

 Furthermore, the numerical system does not take 

 into account the size of objects, and hence con- 

 veys little of the relative importance of the 

 separate components in terms of bulk since the 

 numerical majority may form but a small propor- 

 tion of the food. Foods that have become finely 

 broken up can be only roughly estimated by 

 number. Also, the time required for the inves- 

 tigator to make an accurate count, as for example, 

 of the thousands of crab larvae in the stomach of 

 a tuna which has gorged itself on this food, may 

 be prohibitive, necessitating the use of an esti- 

 mation based on a subsample. 



In frequency-of-occurrence analysis, each food 

 element is expressed as a percentage computed by 

 dividing the number of stomachs containing the 

 food, regardless of amount, by the total number 

 of stomachs examined. This provides a rough 

 but useful index to the overall availability, and 

 perhaps the palatability, of the food element. 



The volumetric system is based on percentage 

 by bulk. Its use reduces the overvaluation of 

 food organisms with more durable parts to a 

 minimum. A large series of stomachs yields re- 

 liable information on recent food, with old food 

 represented by traces. The size of individual 

 items is taken into account only by this system. 

 As with the other two methods of analysis, some 

 soft-bodied organisms may leave no appreciable 

 trace in the stomach and thus may be underrated 

 in importance. Other errors may result from dif- 

 ferent digestion rates. Hess and Rainwater 

 (1939), for example, demonstrated that of the 

 different kinds of immature insects fed to brook 

 trout, small soft-bodied forms were digested more 

 rapidly than large thick-skinned types. Karpe- 



vich (1941) found that Gammaridae were digested 

 more rapidly by three marine fishes than were 

 larger crustaceans and small fish. Despite its 

 defects, we believe this system to be the best of 

 the three, if conclusions are to be based on but 

 one method of analysis. 



The volumetric system can be used in several 

 ways to evaluate the amount of each kind of food 

 present. Martin, Gensch, and Brown (1946), for 

 example, describe the following two ways: 



Aggregate-total-volume method: The percent- 

 age for each kind of food is obtained by dividing 

 the total volume of all food of each kind by the 

 total volume of the stomach contents of all the 

 fish. The variation in the total volume of food 

 from each stomach influences the final result in 

 direct proportion to that volume. 



Average-percentage method: Percentage equiv- 

 alents are calculated for each food item with each 

 stomach evaluated 100 percent regardless of the 

 volume of its contents. Variation in the total 

 volume of food present, therefore, does not influ- 

 ence the results. Stomachs containing very little 

 food exert the same influence on results as do 

 well-filled stomachs. 



The aggregate-total-volume method has the 

 definite merit of reflecting truly the volumetric 

 importance of a particular food organism regard- 

 less of whether much or little of other foods is 

 present. 



Various combinations of the three basic systems 

 of analysis have been devised to present a more 

 complete picture of the food babits of the animal 

 being studied. Tester (1932) combined the volu- 

 metric and the frequency-of-occurrence methods 

 for a graphic representation of the food of the small- 

 mouth black bass, Micropterus dolomieu. The 

 relative importance of each kind of food was 

 demonstrated graphically by a rectangle in which 

 the percentage volume of a kind of food was the 

 horizontal line and the percentage frequency-of- 

 occurrence was the vertical. The vertical scale 

 was arbitrarily fixed at 40 percent of the hori- 

 zontal to give the volume of food more weight. 

 In a different method devised by Welsh (1949), 

 each kind of food was evaluated by a percentage 

 rating which was "an average of the percent of the 

 total bulk of the individual food used (indicating 

 food value), the total number of individual food- 

 animals used (indicating abundance), and the 

 total number of stomachs in which the individual 



