350 
Fishery Bulletin 115(3) 
400 
_ 350 
E 
o. 
E 300 
o 
o 
£ 250 
<5 
<D 
o 
C 200 
re 
w 
b 
150 
5 10 15 20 
Time since vessel passage (s) 
Figure 4 
Mean distance of the vermilion snapper ( Rhomboplites 
aurorubens ) off bottom, measured in the northeast 
Gulf of Mexico in 2014, as a function of the time since 
passage of the tow vessel was recorded by the benthic 
camera. The closest approach of the vehicle was at 23 
s, and the vehicle velocity was 1.5 m/s; therefore, the 
approximate distance in meters of the vehicle to the 
fish school can be calculated as (23-Z)xl.5 , where t is 
the time in seconds after vessel passage. 
between individuals within the school). The responses, 
especially swimming direction, changed abruptly twice, 
separating the period into 3 phases that can be con- 
sidered sequentially increasing states: arousal, aware- 
ness, and flight. 
During the first phase (arousal), up to 11 s after 
vessel passage, the fish were in a heightened state 
of arousal, presumably in response to the sound or 
sight of the passing tow vessel, because their group 
swimming speed and polarity were distinctly greater 
than the levels observed during the undisturbed pe- 
riod. Although this behavior was clearly a response, 
it had not progressed to active avoidance of the ves- 
sel because the individual swimming direction was 
still aligned with the path of the vessel and had a 
slight upward component of movement. Previously 
documented fish reactions to an overhead or recently 
passed vessel usually involved lateral motion or div- 
ing (Handegard and Tjpstheim, 2005; De Robertis and 
Handegard, 2013). 
In the second phase (awareness), from about 12 s to 
19 s after vessel passage, the fish displayed increasing 
signs of arousal, including an increase in group speed 
and polarity and a steady decline in nearest neighbor 
distance, all of which both theoretical and laboratory 
studies typically have shown are correlated and change 
with the arousal state of the school (Magurran and 
Pitcher, 1987; Parrish et al., 2002). During this phase, 
the fish started to display distinct avoidance behavior 
characterized by abrupt changes in horizontal direction 
from nearly parallel to nearly perpendicular to the path 
of the camera vehicle. In addition, vertical swimming 
switched from slightly upward to strongly downward. 
We interpret these behavioral changes as responses to 
the tow cable, which, by midway in the transit interval, 
would appear to the fish as a silhouetted object, ap- 
proximately 10 m above the school (assuming no hys- 
teresis in the tow cable), and approaching the fish at 
a downward velocity of ~0.9 m/s. Presumably, the tow 
cable also would produce vibrational stimuli along with 
the visual stimuli, but we were unable to distinguish 
the relative importance of the 2 stimuli in eliciting the 
observed response. Such downward and lateral avoid- 
ance behavior has been described previously for cod in 
response to the towing warps of bottom trawlers (Han- 
degard and Tjpstheim, 2005). 
During the third phase (flight), from 20 s to 22 s, in- 
dividual fish direction on the horizontal plane abruptly 
changed from perpendicular to about 45° away from 
the line of vehicle travel and swimming speed began 
to increase at an accelerating rate. During this period, 
nearest neighbor distance rapidly declined and polarity 
increased. Group speed rapidly increased, primarily 
because the individual rate of swimming increased in 
addition to the polarity (group speed equals mean indi- 
vidual speed when polarity equals 1.0). We suspect that 
the these changes were reactions to the sight or sound 
of the camera vehicle, which was approximately 8 m 
from the center of the field of view of the benthic cam- 
era when the strong change in direction occurred 19 s 
after vessel passage. Although not silhouetted against 
the sky, the camera vehicle must have been conspicu- 
ous to the fish at this distance because it has 4 contin- 
uously illuminated lights and has a looming movement 
that has been shown to elicit reactions in some Pacific 
groundfishes (Ryer et al., 2009). 
These behavioral changes in response to the pas- 
sage of the vessel, tow cable, and camera vehicle are 
consistent with the notion of threat-sensitive predator 
avoidance (Helfman, 1989; Domenici, 2010; Rieucau et 
al., 2014) where, because of energetic costs (Ydenberg 
and Dill, 1986), fish balance timing and strength of 
escape behaviors against the risk of predation. This 
tradeoff often results in a graded response, starting 
with an increase in awareness of the threat and pro- 
gressing to flight when the predatory attack is initi- 
ated. Therefore, for example, in the first phase, fish 
displayed indications of increased awareness but no 
obvious avoidance. In the second phase, fish initiated 
avoidance behavior; however, they did not greatly in- 
crease their individual swimming speed, presumably 
because the perceived threat was relatively small and 
approaching slowly. In the third phase, fish encoun- 
tered the camera vehicle, changed their swimming di- 
rection so that it was directly away from the threat, 
and rapidly accelerated their individual swimming 
speed, presumably because the camera vehicle was 
large and moving rapidly and, therefore, presented a 
more serious threat. 
