Stomach contents were identified to order and/or family, 

 using a dissecting microscope, and enumerated. Sample volumes 

 were determined by placing the items in a graduated cylinder 

 containing 50 ml of water and noting the displacement caused by 

 the sample. 



Since no single analysis (volumetric, numerical, or 

 frequency of occurrence) will provide an accurate analysis of 

 stomach contents, an index of all three was used. The Index of 

 Relative Importance (IRI) (George and Hadley 1979) combines all 

 three of the above types of analyses to provide a much more 

 accurate analysis of the relative importance of the food items in 

 the sample: 



RI = 100 (fl + pi + vl ) 



f (fl a + Pl a + vl a ) 



where 



fl a = frequency of occurrence 

 pl a = percent of total numbers 

 vl g = percent by volume 



The values in this index (RI a ) range from to 100, with 100 

 indicating exclusive use of a food item (Hensler 1987) . Volume 

 was substituted for weight as in Hensler (1987). 



Frequency of occurrence of a given food item (fl ) was 

 obtained by dividing the number of sample stomachs (by species) 

 that contained at lease one of that food item by the total number 



10 



