KOSLOW: FEEDING SELECTIVITY OF NORTHERN ANCHOVY SCHOOLS 



size at which the anchovy schoors consumption 

 is about 100%^ and the size of the largest plankters 

 enumerated (Figure 2A). The northern anchovy 

 apparently adjusts its feeding so that it continues 

 to select the largest prey over at least a hundred 

 fold range in prey size. (It should perhaps be noted 

 that the largest commonly occurring zooplankters 

 were counted in all samples, and all were found in 

 median concentrations of 172-345 fxg C/m^ (Table 

 5).) While less common, larger zooplankters may 



25 50 75 100 125 150 175 

 LARGEST PREY SIZE (XJ (pgC) 



200 



have occurred in the samples or at the sample 

 sites, relatively rare individual prey items prob- 

 ably would not affect the feeding pattern of a 

 planktivore, such as the northern anchovy.) 



Similarly, the slope of the northern anchovy 

 schools' feeding selectivity is a positive function of 

 the inverse of the largest prey sizes occurring in 

 the samples (Figure 2B), since m - Ay/ Ax = 1/xl, 

 where m = the slope of the school's feeding 

 selectivity; Ay = the difference between the frac- 

 tion consumed of the largest and smallest prey ( = 

 1); Ax = the range (w) in size of the prey organisms; 

 and where the size of the largest prey, xl»Xs, 

 the size of the smallest prey. For this limited set of 

 observations, the relation between the slope of the 

 schools' feeding selectivity im), in fact, appears 

 inversely proportional to the size of the largest 

 prey available (xl). the coefficient of the regres- 

 sion between m and 1/xl is about 1 (Figure 2B). 

 Stated another way, the northern anchovy schools' 

 feeding selectivity over a range of conditions was 

 related to the size of the largest prey available in 

 significant quantity. The largest prey may be con- 

 sidered to be entirely consumed, and the consump- 

 tion of smaller prey is approximately proportional 

 to their size in relation to the largest prey. 



DISCUSSION 



Evaluation of the Field Method 



>- 



> 



H 

 U 

 UJ 



_l 



UJ 

 CO 



z 



Q 



UJ 

 UJ 



u. 



U. 



o 



Q. 

 Q 



050 - 



40 10 5 2 1.25 



LARGEST PREY SIZE (Xl) (pgC) 



Figure 2. — (A) Predicted prey size at 100'% prey consumption 

 and (B) slope of feeding selectivity of northern anchovy schools 

 (from Figure 1) as a function of, respectively, the largest prey size 

 and the inverse of the largest prey size available in significant 

 quantity. ( See Table 4 and text for detailed explanation.) 



There are several possible biasses to the field 

 sampling method: 1) Its accuracy depends upon 

 the choice of control samples. It is clearly prefer- 

 able to sample a patch of water both before and 

 after a school has passed through it. However, bias 

 due to small-scale patchiness in selecting control 

 samples on the sides of the school's path will 

 probably lead to conservative estimates of prey 

 consumption, since the school will presumably 

 swim through the richest plankton patches. 2) The 

 method would be biased to the extent that the 

 disappearance of zooplankters in the wake of the 

 school resulted from escape from the school rather 

 than their consumption by it. However, this does 

 not appear to be a significant problem, considering 

 the large size of the fish schools sampled in this 

 study. The range of swimming speeds of copepods 

 is on the order of 5-50 body lengths/s (ca. 0.5-5.0 

 cm/s) (Enright 1977). The northern anchovy 

 schools sampled were on the order of 50-200 m in 

 length, 10-40 m in depth, and advanced at about 

 10-40 cm/s. Unless the zooplankters could detect 



139 



