FOLKVORD and HUNTER: VULNERABILITY OF NORTHERN ANCHOVY LARVAE 



a movement directed toward the prey with the 

 mouth open. During an attack the northern anchovy- 

 predator usually increased its swimming speed, but 

 the chub mackerel increased speed only when 

 attacking larvae larger than 10 mm SL. 



Four measures of predator-prey interactions were 

 calculated: mean and maximum attack distance; fre- 

 quency of avoidance responses; frequency of 

 escapes; and predation rate (percentage of larvae 

 captured during the 5-min trials). The attack 

 distance was the distance in decimeters (dm) from 

 the prey to the point where the predator started the 

 attack. An avoidance response was a change in 

 speed or direction of a larva occurring before the 

 predator had completed the attack by closing its 

 mouth. 



An escape was defined as a larval response in 

 which the predator failed to capture the larva dur- 

 ing a single attack. Repeated attacks were scored 

 as separate events. By definition, adult Artemia 

 could not be credited with an escape since they did 

 not respond to an attack. Cases where attacked 

 Artemia were not captured were considered 

 predator errors. Predator error could only be 

 assessed for Artemia. All interactions between 

 predators and larvae in which the larvae were not 

 captured were recorded as an escape. 



Predator Performance 



The feeding success and variation in feeding rates 

 of predator groups fed live adult Artemia were 

 analyzed to estimate predator error and to deter- 

 mine if differences existed in feeding performance 

 among predator groups, or among or within ex- 

 periments. An experiment was 2-5 larval trials con- 

 ducted on a single size class of larvae on one day 

 using a single predator group. 



Predator errors were obvious when Artemia were 

 the prey because Artemia did not avoid the attack. 

 In such cases the trajectory of the attack was inac- 

 curate and the predator simply missed the prey. 

 Such errors occurred in 3.4% of the attacks on 

 Artemia; this estimate is similar to error rates esti- 

 mated for other predators (Curio 1976). We could 

 not measure the predator error when larvae were 

 the prey because we attributed any failure to cap- 

 ture a larva to larval avoidance success. Presumably 

 our estimates of larval escape probabilities include 

 an unknown number of cases where failure to cap- 

 ture a larva was the direct result of inaccuracies in 

 the predator's attack rather than being the result 

 of larval avoidance. 



Considering all northern anchovy predator 



groups, predator error in capturing Artemia was 

 higher in the first 5 trials than in the subsequent 

 trials of the experiments where Artemia trials were 

 alternated with larval trials (Fig. 1A). Predator er- 

 ror averaged 3.4% for all Artemia trials, whereas 

 it was 2.1% during the period of alternating larval 

 and Artemia trials. Similarly, adult northern an- 

 chovy took more time to capture all the Artemia in 

 the first trial than in subsequent ones (Fig. IB). No 

 decline in feeding performance on Artemia existed 

 at the end of the experiments, indicating that satia- 

 tion did not constitute a bias in the experiments. The 

 initial decline in the time required for northern an- 

 chovy predators to capture Artemia may have been 

 caused by an increase in feeding motivation, learn- 

 ing, or a decrease in fright behavior. As the decline 

 occurred during only the initial 5 Artemia trials, the 

 larval data were probably unaffected. 



Minor differences in feeding performance also ex- 

 isted among predator groups. In two experiments 

 northern anchovy predatory groups fed markedly 

 less on both Artemia and larvae (30% fewer prey 

 taken in 5 min; £-test, P < 0.05). The effect of omit- 

 ting these two experiments is indicated in the 

 results. Overall, comparisons of feeding perfor- 

 mance among groups, within trials, and among 

 experiments indicated that variation in predator per- 

 formance as measured by predation rates on 

 Artemia was not significantly biased (additional 

 details are given by Folkvord 1985) (see also Figure 

 1). 



RESULTS 



Probability of a Response to Predators 



The most striking feature of the vulnerability of 

 the youngest larval stages of northern anchovy to 

 predators was the low frequency of escape attempts. 

 Only 16% of the 6 mm larvae responded to the at- 

 tacks of northern anchovy predators (Fig. 2B) and 

 only 26% of 6.7 mm larvae responded to chub 

 mackerel predators (Fig. 3B). The probability of 

 smaller larvae (SL <20 mm) responding to either 

 chub mackerel or northern anchovy predators was 

 about the same (Fig. 4), although size, feeding 

 behavior, and body form of these two fishes were 

 distinctly different. The tendency to respond to at- 

 tacking predators steadily increased with larval size 

 until by the time northern anchovy larvae were 30 

 mm all attempted to avoid attacking northern 

 anchovy and over 80% responded to chub mackerel 

 attacks. 



861 



