Rabe and Brown: Behavior, growth, and survival of Glyptocephalus cynoglossus larvae in relation to prey availably 
469 
10.5 mm (Fig. 3A). Pause and sink MAPs decreased 
when larvae reached 16.2 mm, whereas a shake 
MAP stopped altogether (Fig. 3B). The durations 
of the locomotory and nondirected MAPs changed 
with size (Fig. 3, A and B) but were not significantly 
affected by prey density (Table 4). At the end of the 
study there was a slight increase in pause MAPs 
and a concomitant decrease in the duration of swim 
MAPs, the result of some larvae settling during the 
observation periods. 
Throughout the study, larvae spent between 2% 
and 10% of the total time observed performing forag- 
ing activities. The variation in total time spent for- 
aging was largely due to variation in orientation du- 
ration. The duration of the fixate and lunge MAPs 
(<2% of total time per MAP) was relatively constant 
over the observation periods, whereas the duration 
of the orientation MAP ranged from 1% to 7% of total 
time (Fig. 30. 
The frequencies of the foraging behaviors were 
highly variable and many larvae did not forage dur- 
ing the observation periods. The frequency of the ori- 
entation MAP changed throughout the study period, 
peaking between the mean sizes of 10.5 and 20.8 
mm (Fig. 4A). Within this size range, orientation fre- 
quency increased with increasing prey density. The 
effect of prey density on orientation frequency was 
significant (Table 2); the orientation frequency of lar- 
vae at 250 prey/L was significantly lower than that 
of larvae at 2000-16,000 prey/L within the 10.5-20.8 
mm size interval (Tukey test). After 20.8 mm, the 
frequency of orientation was low for all sizes and 
treatments. 
The frequencies of fixate and lunge MAPs varied 
from 0 to 4 per two-min observation. Larvae at high- 
er prey densities tended to perform more fixate and 
lunge MAPs compared with larvae at lower prey 
densities (Fig. 4, B and C), although this trend was 
not statistically significant (Table 2). The frequen- 
cies of fixate and lunge MAPs increased significantly 
with increasing larval length (Fig. 4, B and C, Table 2). 
The average lunge frequency of early- and late-stage witch 
flounder larvae was lower than that of both yellowtail 
flounder and Atlantic cod larvae (Fig. 5). 
Discussion 
Witch flounder larvae grew and survived in all treatments 
used in our study. Our experiment is the first to examine 
the early growth and behavior of witch flounder larvae 
in relation to prey availability. Larval performance can 
be influenced by many factors other than prey density, 
including temperature (Hunter, 1981), light, (Batty, 1987; 
Puvanendran and Brown, 1998), prey type (Drost, 1987), 
and turbulence (MacKenzie and Kiprboe 1995; Brownian, 
1996). We used our results 1) to describe the early growth 
and ontogeny of the foraging behavior in this species and 
2) as a preliminary step towards understanding the behav- 
ioral ecology of witch flounder larvae. 
O) 
E 
O) 
0 ) 
S 
Q 
1000 
100 
10 
0.1 
0.01 
j f 
2000 p/L 
4000 p/L 
8000 p/L 
WeekO 
10 20 30 40 50 60 
Standard length (mm) 
Figure 2 
Relationship between standard length (mm) and (A) dry weight 
(mg; note logarithmic axis) and <B> body height (mm) for individ- 
ual witch flounder larvae reared at different prey densities (no. of 
prey per liter). Symbols represent individual larvae. 
Table 3 
Percentage of witch flounder larvae reared at different 
prey densities (±SE) that survived over the experiment. 
Latvae sampled dead were not included in calculations. 
Prey density (no. of prey per liter) 
2000 
4000 
8000 
Week 2 
Week 5 
Week 12 
36.15(1.28) 
28.72 (1.54) 
14.10 (2.31) 
30.26 (6.15) 
25.13 (5.13) 
4.62 (0.51) 
38.20 (4.36) 
31.03 (3.85) 
8.97 (4.36) 
Witch flounder grew and survived equally well at each of 
the prey densities tested. Although the range of prey den- 
sities used in the rearing experiment (2000-8000 prey/L) 
was not exhaustive, these prey densities have resulted in 
informative differences in growth and survival for other 
