Rabe and Brown: Behavior, growth, and survival of Glyptocephclus cynoglossus larvae in relation to prey availably 
471 
Table 4 
Summary of ANCOVA results for locomotory and nondirected MAPs of witch flounder larvae at different prey densities ( 
per liter). Each model was run until the larval size indicated in parentheses to satisfy model assumptions. (* denotes a 
difference at a=0.05). 
no. of prey 
significant 
MAP 
Source 
df 
F 
P 
Swim frequency (20.8 mm) 
Size 
1 
32.1 
<0.001* 
Prey density 
6 
0.78 
>0.25 
Size x prey density 
6 
0.42 
>0.5 
Error 
49 
Swim duration (20.8 mm) 
Size 
1 
251.4 
<0.001* 
Prey density 
6 
1.76 
>0.25 
Size x prey density 
6 
0.17 
>0.5 
Error 
49 
Turn duration (13.8 mm) 
Size 
1 
56.5 
<0.001* 
Prey density 
6 
1.34 
>0.25 
Size x prey density 
6 
0.82 
>0.5 
Error 
21 
Pause duration (20.8 mm) 
Size 
1 
88.4 
<0.001* 
Prey density 
6 
1.82 
>0.25 
Size x prey density 
6 
0.26 
>0.5 
Error 
49 
Sink duration (18.4 mm) 
Size 
1 
79.8 
<0.001* 
Prey density 
6 
0.54 
>0.5 
Size x prey density 
6 
1.35 
>0.5 
Error 
42 
Shake duration (16.2 mm) 
Size 
1 
13.0 
<0.001 
Prey density 
6 
1.13 
>0.25 
Size x prey density 
6 
0.25 
>0.5 
Error 
35 
time was spent swimming, which was typically interrupted 
only by foraging events. During the mean size interval of 
10.5-16.2 mm, the turn and shake MAPs disappeared and 
the frequency of pause and sink MAPs decreased. These 
behavioral changes were likely related to the increase in 
larval body height, accompanied by a substantial increase 
in finfold height that occurs during this time. 
The nature of the nondirected MAPs shake, sink, and 
pause is not straightforward. Sinking has been reported 
in other species, such as the snapper (Pagrus auratus) 
and, like the pause MAP, has been interpreted as a rest- 
ing behavior. In snapper, it occurs in yolksac larvae and 
in feeding larvae during night-time periods of inactivity 
(Pankhurst et al., 1991). Sinking is typically observed only 
in the early stages of other species such as the black sea 
bream, Acanthopagrus schlegeli (Fukuhara, 1987). How- 
ever, Kawamura and Ishida (1985) noted that sinking oc- 
curs in both yolksac larvae and larger feeding larvae of 
the flounder Paralichthys olivaceus immediately after at- 
tacking a prey item. Observations of sinking in later-stage 
witch flounder larvae were not related to feeding events; 
the persistence of these behaviors in witch flounder was 
likely the result of some smaller, slower-growing individu- 
als having been included in our observations. 
Witch flounder, like other species (Holling, 1965; Houde 
and Schekter, 1980; Werner and Blaxter, 1980; Puvanen- 
dran and Brown, 1999), demonstrated increased foraging 
behavior with prey density. However, the orient MAP was 
the only foraging MAP statistically affected by prey den- 
sity. The change in orientation frequency with size is in- 
teresting because this MAP was affected only by prey den- 
sity within a limited size range. The initial low frequency of 
orient MAPs followed by an increase associated with great- 
er larval size can be explained by changes in swimming 
speeds and encounter rates (Mittelbach, 1981). However, 
the decrease in orient frequency across treatments later in 
the study period is puzzling because larger larvae are gen- 
erally competent swimmers (Rosenthal and Hempel, 1971; 
Laurence, 1972; Houde and Schekter, 1980) and are expect- 
ed to exhibit a high prey encounter rate. This decrease in 
orient frequency may be due to improved foraging ability 
associated with greater visual acuity. Miller et al. (1993) 
showed that the visual angle — the smallest angle at which 
a stimulus may subtend the eye and remain resolvable 
(Neave, 1984) — decreases during the development of three 
species of fish larvae. Thus, the eye develops such that lar- 
vae can likely detect prey items in their periphery without 
turning the head and orienting themselves toward them. 
