Abstract:
Red king crabs (Paralithodes camtschaticus) are caught by pots in a male-only fishery in Alaska. The objectives of this research were: (1) to examine impacts of the commercial fishery on discarded female and sublegal male crabs; (2) to examine bait efficiency; (3) to document crab behavior to pots; and (4) to develop a model describing catch versus soak time. I estimated from observer data that 64.6% of crabs in the Bering Sea fishery were females and sublegal males; I simulated commercial crab handling procedures in the laboratory to test effects on discarded crabs. Although body damage increased significantly with increased handling, there were no significant effects on righting time, feeding rate, weight gain, carapace length increment, or survival. I examined the efficiency of five potential baits (squid, herring, mussel, king crab muscle, and king crab ovary) by observing chemoreception and feeding behavior of the crabs. Chemosensory threshold varied between $10\sp{-4}$ to $10\sp{-6}$ g.L$\sp{-1}$, and feeding threshold ranged from $10\sp{-2}$ to $10\sp{-3}$ g.L$\sp{-1}.$ Crabs were most sensitive to the extract of conspecific muscle, while herring was most effective in arousing feeding behavior. Little difference existed between males and females in chemoreception and feeding behavior. Behavioral responses of the crabs to crab pots were observed by time-lapse video. Crabs approached the pot from downstream, and 78.3% of crabs searched less than 90$\sp\circ$ before leaving or entering the pot. The entry success rate was 8.1%. Only large males could begin escape from the bottom panel. Crabs had difficulties in accessing the pot and in escaping from inside the pot. The standard pot appeared inefficient in catching legal males, while it retained many non-legal crabs. I constructed a general model to describe the relationship between catch and soak time for trap fisheries. The model is expressed as $C\sb{t}=ab+a(t-b)\ e\sp{-ct},$ where $C\sb{t}$ is the catch per trap haul at soak time t, and a, b, and c are parameters to be estimated. This model is suitable for both short and long soak times.
