50 
Fishery Bulletin 109(1 ) 
checked, and the presence or absence and location of 
the original test crabs were noted. The size, sex, and 
molt stage of the additional captured crabs were re- 
corded, and any unusual occurrences were documented. 
Captured crabs were numbered and placed back in 
their original locations (parlor or kitchen) in the pot. 
Pots were rebaited and set out for another 24-h period 
to assess escape rates from baited pots, to determine 
whether pots would reach some saturation point, and 
to establish the theoretical density when a pot cannot 
catch anymore crabs. After the second 24-h period, 
pots were retrieved, and similar information was re- 
corded. It was noted whether crabs from the first 24-h 
period had escaped or were still present. Size, sex, molt 
stage, and location in the pot were recorded for newly 
captured crabs. After all information was documented, 
all crabs were released. This experiment was repeated 
four times, for a total of five trials. Both the size and 
number of crabs caught were analyzed with a 2-fac- 
tor analysis of variance (ANOVA) with the factors of 
depth and test crab size. Additionally, Tukey’s honestly 
significant difference test was used for multiple mean 
comparisons. All statistical tests were conducted with 
SAS® software (SAS, vers. 9.0.0, SAS Institute, Inc., 
Cary, NC). 
Crab-trap video (CTV) system 
The crab-trap video (CTV) was modeled after the lob- 
ster-trap video presented in Jury et al. (2001). It is a 
low cost tool for observing interactions between crabs 
in and around pots in a mesocosm. Although it was not 
used for in situ observations in our experiment CTV 
could be easily modified for in situ observations. CTV 
consists of a standard commercial crab pot matching 
the specifications described above with video camera 
equipment attached for observation. A low-light, black- 
and-white, Sony time-lapse video recorder, model EVT- 
820, was used to record images every minute for 24 
hours. The camera was set 38 cm above the pot on a 
support system of four 93.9-cm long PVC pipes con- 
nected by four shorter PVC pipes 55.9 cm in length 
(Fig. 1). This configuration allowed the entire pot to 
be observed, as well as a few centimeters on each side 
of the pot. For nighttime recording, a red light, unde- 
tectable by crabs (A. Hines, personal commun. 1 * ), was 
affixed to one of the PVC legs and used to light the 
area. All images were recorded on Hi-8 tapes in an 
adjacent building connected to the mesocosm camera 
by cables. Connection cables were placed approximately 
40 cm above the surface of the water surface in the 
mesocosm experiment. The system was capable of col- 
lecting data for at least 24 hours; therefore it continu- 
ously captured all crab interactions within the field of 
view for the experimental time-frame. The system also 
1 Hines, Anson. 2003. Smithsonian Environmental Re- 
search Center, 647 Contees Wharf Road Edgewater, Mary- 
land 21037. 
Figure 1 
Diagram of the crab trap video (CTV) system. CTV is a 
time-lapse video recording system affixed to a standard 
commercial crab pot. The design allows the movements 
of crabs in and around the pot to be recorded during 
the day and night. Image not drawn to scale. Field of 
view=field of view from the camera lens. 
recorded crab approaches, entries into the kitchen and 
parlor, and escapes of crabs. 
Mesocosm experiment 
A mesocosm experiment was conducted to determine the 
influence of crab behavior on pot catch. A large above- 
ground circular mesocosm (4.8 mxl.06 m, 18.6 m 3 ) set 
on preleveled ground was used for this experiment. A 
mesocosm was used because the high turbidity of the 
Rhode River made in situ observations impractical. 
Ambient water from the Rhode River was transferred 
into the mesocosm and filtered for two days to increase 
water clarity. Water was constantly filtered when experi- 
ments were not running, and filter bags were changed 
daily. During experimental runs, filters were turned off 
and removed from the mesocosm. Fine-grain sand was 
used to cover the bottom of the mesocosm in an attempt 
to mimic the muddy-sandy substrate of the Rhode River. 
Mummichogs ( Fund-ulus heteroclitus) were placed inside 
the mesocosm to help control mosquito larvae popula- 
tions and other insects but were removed before each 
testing to limit nonsubstantial variables. As with the 
field experiment, crabs were held in deck tanks and 
were fed chopped alewife until 24 hours before use in 
an experiment. 
For each experiment, 16 male blue crabs (6 large [155 
mm CW or greater], and 10 smaller [127-150 mm CW]) 
were used. Test crabs had all appendages, and only 
male crabs of molt stage C were used to reduce any be- 
havioral variance. The number of crabs per unit of area 
