392 



Fishery Bulletin 98(2) 



frames from slipping off during recovery. The grap- 

 ple beam was lowered into the water with enough 

 cable to maintain a 2:1 scope (ratio of wire out to 

 depth), and towed through the suspected pot posi- 

 tion. If no pot was hooked, additional transects were 

 run parallel to the first, or the vessel was set into 

 a sharp turn and the beam was towed in a circle. 

 Hooked pots were retrieved with the ship's crane, 

 while the vessel ran at a slow speed. Recovered pots 

 were designated by the number of the most likely 

 sonar position where they were found, and informa- 

 tion was recorded as during the pilot study. The south- 

 east section of the sonar-surveyed area could not be 

 sampled owing to underwater cables and rocks. 



Nontargeted study 



The nontargeted study was conducted by the skip- 

 per and crew of the FV Big Valley without biologists 

 aboard during 19-31 August 1996 and 5-21 October 

 1996. Previous attempts to snag randomly selected 

 (i.e. targeted) pots were not particularly efficient 

 because only 43 pots were recovered during 32 days 

 of fishing. Therefore, instead of targeting specific 

 pots for recovery, the skipper was instructed to fish 

 haphazardly at his own discretion in areas where 

 commercial crab fishing had previously been con- 

 ducted, but without knowing specific pot locations. 

 This method was more efficient and productive. The 

 grapple beam was damaged, however; therefore it 

 was replaced with a 30-m long weighted chain, to 

 which were attached six short grapples, each with 

 four barbed tines. The long grapple chain was towed 

 at 2.8 km/h in a constant 10 turn until a pot was 

 hooked. Recovered pots were given sequential num- 

 bers, and data were recorded as before. 



Data analysis 



Because Tanner crabs were the target species for 

 most pots, associations between the presence of 

 Tanner crabs and other species were tested by con- 

 tingency table (chi-square) analysis. To reduce empty 

 cells. Tanner crabs were combined into logarithmic 

 categories (0, 1-9, 10-99, >99) before analysis. Asso- 

 ciation between pot condition and sex of the crabs 

 was also tested by means of chi-square. For this pur- 

 pose, pots were regrouped into two categories des- 

 ignated "best" (including pots coded as "excellent" 

 or "good") and "worst" (including "fair," "poor," or 

 "very poor" pots). Mean sizes of crabs in various con- 

 dition categories are given as mean ±SE and were 

 compared by means of the Mann-Whitney f7-test 

 (MWUl because variances were typically unequal 

 (Zar, 1984). Catch of crabs or other organisms is 



expressed as catch per pot (CPP, i.e. crab per pot). 

 CPP was not standardized by time because we could 

 not adequately determine how long each pot had 

 been in the water. The effect of holes in pot webbing 

 on crab CPP was tested with the MWU-test; addition- 

 ally, binomial confidence intervals were calculated 

 for the proportion of pots with crabs and compared 

 for overlap between pots with holes and those with- 

 out. The null hypothesis was that torn pot webbing 

 had no effect on CPP. 



Results 



Sonar survey and in-situ observations 



Submersible observations in 1995 and previous years 

 showed that seafloor substrates in the surveyed area 

 were primarily sandy silt with occasional areas of 

 compacted mud, and bottom contour was relatively 

 flat below 100 m depth (senior author, personal obs.). 

 The soft mud bottom absorbed most of the trans- 

 mitted sonar energy and reflected very little. As a 

 result, crab pots and other anthropogenic objects 

 were easily distinguishable with the sidescan sonar. 

 At least 189 objects were believed to be crab pots; 

 177 were single and 12 were connected together in 

 groups of 2 to 5. In some cases attached lines could 

 also be distinguished by sonar. Density of crab pots 

 in the 4.5-km2 area examined was 42 pots/km^. At 

 least 15 of these objects were later observed with the 

 submersible Delta and the ROV, and all were veri- 

 fied to be crab pots, including a group of five pots 

 connected together by tangled lines. Also noticeable 

 were several large linear features up to 2 km long; 

 one was later observed from the Delta to be a drag 

 scar or ditch, approximately 0.5 m deep and 3-5 m 

 wide. 



Condition of pots recovered 



A total of 147 pots of varying age, size, and type 

 were retrieved during the three studies (Table 1; Fig. 

 2). Detailed information on location, condition, and 

 contents of each pot is contained in a separate data 

 report.'* Seventy-two pots were recovered from inside 

 the sonar-surveyed portion of Chiniak Bay (these are 

 subsequently referred to as "inside" pots), including 



■• Vining, I., S. Byensdorfer, W. E. Donald.son. B. G. Stevens, and 

 G. Edward.s. 1997. Lost crab and cod pot recovery and ghost 

 fishing in Chiniak Bay and other areas in the waters around 

 Kodiak Island. Alaska. Regional information report 4K97-42. 

 Alaska Dept. of Fi.sh andGame, 211 Mission Rd., Kodiak, Alaska 

 9961.5. 93 p. 



