HUNTER: BEHAVIOR OF LARVAL ANCHOVY 



cated that considerable flexibility exists in what 

 appeared to be a highly stereotyped feeding pat- 

 tern. Although a sequence was completed in 

 only 1 to 2 sec, the larvae depend on feedback 

 from the prey throughout the period. The lar- 

 vae respond to the prey during the sequence by 

 adjusting posture, speed, and direction of move- 

 ment and by ending or continuing the sequence. 



EXTENT OF REACTIVE PERCEPTIVE FIELD 



The size of the predator's reactive perceptive 

 field (Rolling, 1965) is an essential element in 

 the estimation of the rate of, search for prey by 

 a predator, that is, the extent of the area in 

 which a predator will react to a prey. In studies 

 on larval fishes the cross-sectional area of the 

 perceptual field is multiplied by the speed of 

 swimming to estimate the volume of water 

 searched per unit of time. Estimates of this 

 type have been made by Braum (1967), Rosen- 

 thal and Hempel (1970), and Blaxter (1966) 

 and are summarized by Blaxter (1969). 



To determine the position of prey when larvae 

 first reacted to them feeding, larvae were filmed 

 from below for horizontal measurements and 

 separately from the side for measurements in 

 the vertical plane. Seventy-one horizontal film 

 sequences of larvae, 4.0 to 24.2-mm, and 12 ver- 

 tical sequences of larvae, 7.0 to 24,7 mm, were 

 analyzed frame by frame. 



In the film analysis all measurements were 

 made in reference to the body of the larvae; X 

 signified measurements made in the axis of pro- 

 gression or swimming plane; Y those in the other 

 horizontal axis; and Z those made along the 

 vertical axis. Vertical measurements were 

 made in relation to the orientation of the larvae 

 and were not necessarily vertical in relation to 

 the water surface. Searching behavior was in- 

 dependent of body orientation. Larvae reacted 

 to prey when they swam upward, when they 

 swam downward, as well as when they swam 

 parallel to the water surface. 



In each photographic sequence the angle and 

 distance of the prey from the tip of the snout 

 of the larva were measured 15 msec before the 

 larva reacted to the prey by turning the head 

 toward it (two frames at 128 fps). To correct 

 for obvious length-dependent differences in field 



size the distance to the prey was divided by laryal 

 length and was expressed in body lengths (L). 

 Prey organisms included Brachionus, various 

 veliger larvae, Artemia nauplii, and wild cope- 

 pod nauplii of undetermined species. 



In the horizontal plane larvae reacted only to 

 prey ahead of them; prey at 90° or more from 

 the tip of the snout were not selected, and most 

 prey were less than 60° from the snout (Figure 

 7). The reactive perceptive field in horizontal 

 cross section was roughly circular. A circle of 

 radius 0.4L with the center on the axis of pro- 

 gression or X axis enclosed 90% of all prey 

 sighted (shaded area. Figure 7) , In the vertical 

 plane, larvae reacted to prey below as well as 

 above the X axis. The maximum distance above 

 and below the X axis at which prey were sighted 

 in the vertical plane was 0.3L and thus the max- 

 imum extent of Z was 0.6L (Figure 8). The 

 reactive perceptive field may be roughly tri- 

 angular in vertical cross section, because a tri- 

 angle with a central angle of 53° and altitude of 

 0.74L enclosed all but one of the 12 observed 

 values. 



Maxima seem appropriate rather than aver- 

 ages to estimate the extent of the reactive per- 

 ceptive field because only the exceptional larva 

 survives in nature and because field size may 

 change with feeding motivation. Considering 

 the two axes in the horizontal plane separately, 

 for 95% of all prey sighted the value Y for the 

 position of prey was equal to or less than 0.4L 

 from the axis of progression {X axis) , and for 

 95% of all prey the X values for prey position 

 were equal to or less than 0.74L. Ninety-five 

 percent limits could not be used to estimate Z 

 because the observations were too few to calcu- 

 late percentages. The maximum observed Z val- 

 ue above and below the X axis was 0.3L. This 

 value could be used or alternatively Z could be 

 estimated at a point on X by assuming the field 

 is triangular in cross section as illustrated in 

 Figure 8. Using the 95% limit for X, 0.74L, 

 as the point to make the cross section, we obtain 

 an estimate of Z = 0.36L. Thus, the estimate 

 of the maximum extent of Z varied from 0.30 

 to 0.36L depending on the assumptions used. 

 Assuming an elliptical cross section where Y — 

 0.40L and Z = 0.36L the area of the ellipse is 



831 



