276 



Fishery Bulletin 94(2), 1996 



predation rate was explained by the consumption of 

 alternative copepod prey. When one outlier point was 

 removed, the fraction of the variance explained by 

 the model increased to 46% (Fig. 8B). 



Discussion 



Selectivity for fish larvae in Atlantic mackerel 

 larvae 



Over 50% of feeding Atlantic mackerel >6 mm had cap- 

 tured at least one fish larva despite the absolute and 

 relative (to other prey) scarcity offish larvae of suit- 

 able prey size in the plankton (on average 0.14 fish 

 larva/m 3 vs. 5,920 nauplii and copepods/m 3 ). By com- 

 parison, copepods were preyed upon in proportion to 

 their relative abundance in the plankton. On the basis 

 of the small percentage of larval fish prey that could be 



