Clark et al.: Underwater video surveillance of catch saturation in lobster traps 
165 
Soak time (h) 
Figure 3 
Number of American lobsters (Homarus americanus) observed sur¬ 
rounding, entering, and escaping a ventless trap per hour and record¬ 
ed by video during a single representative trial on day 1 (hours 1-8) 
and day 2 (hours 18-24) off New Hampshire (Wallis Sands) in 2011. 
Similar data are presented for a standard trap on day 1 (hours 1-8) 
and day 2 (hours 20-24). Data are missing between days because the 
lobster-trap video system was turned off at night. 
mesh for all traps was 2 . 5x2.5 cm and the rectangu¬ 
lar trap dimensions were 90x47x35 cm. The kitchen 
of each trap contained 2 entrances, both 12.7 cm in 
diameter (Fig. 2). The parlor of each standard trap had 
a single escape vent (14.6x4.9 cm); ventless traps had 
no escape vents. 
Data analysis 
To determine factors that influenced trap saturation, 
the following parameters were quantified for each hour 
of observation of video recordings: 1) the number of 
American lobsters entering, escaping, and surrounding 
the traps and 2) the accumulated catch. When lobsters 
entered the trap, they continued to be observed to de¬ 
termine whether they would enter the parlor or leave 
the trap (see Fig. 2). In general, lobsters could not be 
identified individually; therefore, if a lobster left the 
field of view and then returned, it was treated as a 
new lobster. The rates at which lobsters accumulated 
in traps were calculated by using linear regression 
analyses. The statistical analyses presented here were 
performed with RStudio, vers. 1.0.136 (RStudio Team, 
2009). Because videos were not recorded at night, the 
beginning of saturation for ventless traps (between 16 
and 24 h; Clark et ah, 2015) was not captured; there¬ 
fore, segmented linear regression analyses were used 
to estimate rates of accumulated entries, escapes, and 
catch. Each parameter (total number of lobsters enter¬ 
ing, escaping, and surrounding traps) was compared 
by using a Mann-Whitney U test. Only the first 3 h 
of footage recorded on day 1 and on day 2 were in¬ 
corporated into the analyses because these recordings 
had the highest quality across all trials. For example, 
the total number of lobsters entering all ventless traps 
(n- 3) within the first 3 h on day 1 was compared with 
the total number of lobsters entering within the first 
3 h on day 2. To compare CPUE in standard and vent¬ 
less traps, catch values were first transformed by using 
the natural log and then analyzed by using the Mann- 
Whitney U test. All reported variations are standard 
deviations (SD) unless indicated otherwise. 
