Somerton et al.: Quantifying the behavior of fish in response to a moving camera vehicle 
349 
Time since vessel passage (s) 
Figure 3 
Mean individual swimming speed and direction of vermilion snapper 
( Rhomboplites aurorubens) in the northeast Gulf of Mexico in 2014, ex- 
pressed as either (A) degrees counter clockwise relative to the camera 
vehicle tow path (toward the right) on the horizontal plane and as (B) 
degrees up and down in the vertical plane. Time is in seconds since the 
passage of the tow vessel was recorded by the benthic camera. On the 
horizontal plane, the direction of travel is from left to right; therefore, a 
left arrow is directed away from the tow vessel and toward the approach- 
ing camera vehicle. On the vertical plane, the horizontal line represents 
zero speed. The closest approach of the vehicle was at 23 s after the pas- 
sage of the vessel, and the velocity of the camera vehicle was 1.5 m/s. 
Hence, the approximate distance in meters from the vehicle to the fish 
school can be calculated as (23-/)xl.5, where t is the time in seconds after 
vessel passage. 
The nearest neighbor distance followed an almost 
inverse pattern, with a distinct, strong, decrease 
starting at about 11 s (Fig. 5D). Therefore, over the 
entire encounter with the camera vehicle and its tow 
vessel, the school increased its swimming speed ini- 
tially by the individual fish aligning with each other 
and finally by increasing the individual swimming 
rate. In addition, as the alignment of individuals was 
increasing, their spacing became progressively tighter. 
One important effect of these changes in behavior 
was a decrease in the abundance of vermilion snap- 
per within the area that was subsequently transited by 
the camera vehicle during its survey (i.e., count path; 
Fig. 6). At the time of vessel passage (t=0) all 25 mea- 
sured fish were within the count path of the camera 
vehicle, but later, in response to the various stimuli 
produced, the fish moved closer to the fixed camera 
and ultimately out of the count path. When the cam- 
era vehicle actually occupied the count path, at i=23 
s, no vermilion snapper remained in the count path 
and none were seen in the forward cameras used for 
counting (although some fish were seen by the later- 
ally pointed cameras not used for counting). Therefore, 
in this particular case, the vermilion snapper avoided 
sampling by the camera vehicle. 
Discussion 
The school of vermilion snapper observed in this study 
experienced a variety of visual and auditory stimuli 
between their undisturbed state (5 min before passage 
of the tow vessel) and the arrival of the camera ve- 
hicle. These stimuli triggered responses both at an in- 
dividual level (i.e., swimming speed and direction), as 
well as at a group level (i.e., alignment of and spacing 
