FISHERY BULLETIN: VOL. 69, NO. 2 



z 

 < 



z 

 < 



UJ 



u 



< 



3 

 K 

 u 



UJ 



o 

 a: 



UJ 

 Q. 



5 10 15 



MEAN GILL RAKER GAP (RANKS) 



15- • 



10- 



T 1 r 



5 10 15 



MEDIAN FORK LENGTH (RANKS) 



Figure 4. — Relation between percentage crustaceans by 

 volume of diet (ranked) and the (a) mean gill raker 

 gap of a fish (ranked) and (b) fork length of fish 

 (ranked). The diagonal line depicts a perfect inverse 

 relationship. 



and Duarte Bello (1961). These were not used 

 in the present analysis because sample sizes were 

 fewer than 25 fish or because the size of the 

 individual crustaceans was unavailable. Re- 

 gardless, larger A', pelamis in both studies con- 

 tained less crustaceans. However, larger T. 

 albacares in Alverson's study tended to eat more 

 crustaceans than did smaller specimens. This 



may have been because the crustaceans in their 

 diet were relatively large. 



Alverson's paper also presents a good example 

 of selectivity among crustaceans that may be 

 based on size of gill raker gaps. K. pelamis and 

 T. albacares from the same areas had markedly 

 different diets. Crustaceans contributing the 

 greatest volume to the stomachs were galatheids 

 and portunids for T. albacares but euphausids 

 for K. pelamis. Euphausids were rare in stom- 

 achs of T. albacares even when common in the 

 micronekton (Blackburn, 1968). Galatheids and 

 portunids (Longhurst, 1967; Jei'de, 1967b) are 

 typically larger in size than euphausids (Jerde, 

 1967a). The small euphausids were not impor- 

 tant in stomachs of T. albacares (i.e., 1% of 

 the volume) in any of the areas of the eastern 

 tropical Pacific studied by Alverson (1963), 

 but the larger galatheids and portunids were 

 important in the stomachs of K. pelamis from 

 certain areas. The above observations would 

 be the predictions from gill raker gaps — T. al- 

 bacares have broader gaps than K. pelamis and 

 would not be expected to capture the smaller 

 crustaceans. 



The major hypothesis under investigation in 

 the present study was that the quantity of smal- 

 ler organisms (crustaceans) eaten should be re- 

 lated to the selectivity of the gill raker apparatus. 

 The above coi'relations on central Pacific data, 

 although only crude in nature, lend support to 

 this idea. A more definitive test would require 

 extensive data on the size of food organisms 

 and the diet of scombrids over more narrow 

 length ranges than are available from the pub- 

 lished literature. 



Even though the structure of the gill raker 

 apparatus ultimately determines the smallest 

 size of prey, it is possible that actual selection 

 of fishes is made prior to ingestion (Ivlev, 1961; 

 Galbraith, 1967). Galbraith believed that the 

 gill rakers of yellow perch, Perca flavescens 

 (Mitchill), and rainbow trout, Salmo gairdrieri 

 Richardson, could have retained smaller zoo- 

 plankton than were typically found in their 

 stomachs. These species ate only larger Daph- 

 nia even though numerous smaller ones were 

 in the zooplankton. 



368 



