470 
Fishery Bulletin 99(3) 
c 
CD 
O 
CD 
D. 
100 
80 
60 
40 
20 
10 
100 
80 
60 
40 
20 
10 
B 
Total locomotion 
Swim 
Turn 
Total nondirected 
Pause 
Shake 
Sink 
Standard length (mm) 
Figure 3 
Mean proportion (%) of time witch flounder larvae spent 
performing individual (A) locomotory, (B) nondirected, and 
(C) foraging MAPs over standard length (mm) during two- 
minute observation periods. Values are means (n= 70 larvae 
per length) ±SE. 
North Atlantic marine fish larvae, such as Atlantic cod 
(Puvanendran and Brown, 1999) and redfish ( Sebastes sp.; 
Laurel et al., in press) reared under similar laboratory 
conditions. Therefore, we anticipated that this range of 
prey densities would be adequate to detect differences in 
growth and survival of witch flounder in relation to prey 
density. Furthermore, as shown in experiment 2, the lunge 
frequency of witch flounder was not significantly affected 
by prey availability, which would be expected if low prey 
densities (<2000 prey/L) were to reduce consumption and 
subsequent growth and survival. 
12 
10 
>. 
o 
c 
S> 6 
cr 
CD 
4= 4 
250 p/L 
500 p/L 
1 000 p/L 
2000 p/L 
4000 p/L 
8000 p/L 
16000 p/L 
>. 
o 
c 
CD 
3 
cr 
CD 
>> 
o 
c 
CD 
CD 
CD 
c 
10 
15 
20 
25 
30 
Standard length (mm) 
Figure 4 
Frequency of (A) orient, (B) fixate, and (C) lunge MAPs 
of witch flounder larvae at different prey densities (no. of 
prey per liter) during two-minute observation periods over 
standard length (mm). Values are means (n=10) ±SE. 
Ontogeny of behavior 
Witch flounder search strategy for prey is interesting 
because it appeared to change from a saltatory to a cruise 
strategy (see O’Brien et al., 1990; Browman and O’Brien, 
1992) during the study period. When larvae were less than 
10 mm, foraging included many turns and brief periods of 
swimming that served as repositioning acts. By the time 
larvae reached an average size of 16.2 mm, most of their 
