Boggs: Estimating capture depths of longline-caught pelagic fish 



643 



Materials and methods 



Longline fishing was conducted on board the NOAA 

 ship Townsend Cromwell in January 1989 and Janu- 

 ary-February 1990. Sets were made between lat. 14° 

 and 20°N, long. 148° and 159°W, 20-500nmi from the 

 main Hawaiian Islands, and within an area typically 

 fished by Hawaii's domestic longline fishery. Gear was 

 usually deployed in the morning and retrieved in the 

 afternoon or evening (Table 1), or occasionally at mid- 

 day to permit a second set on the same day. No sets 

 were made at night. Except for the hook timers and 

 TDRs, the fishing gear and operations were similar to 

 commercial longline fishing methods for tuna in Hawaii 

 (Kawamoto et al. 1989) prior to the advent of night 

 fishing for swordfish Xiphias gladius. Both this study 

 and the contemporary commercial longline fishery used 

 a wide variety of fishing depths. Commercial fishermen 

 used more gear (~1000 hooks), let it stay in the water 

 longer (~12h), and retrieved it faster than in this study. 

 The fishing gear consisted of 3.5mm-diameter nylon 

 monofilament main line deployed with a line thrower 

 (Kawamoto et al. 1989). The main line was supported 

 at intervals by vertical, 18m lines with floats at the 

 ends. Snap-on branch lines made of 2.1mm-diameter 

 clear-blue nylon monofilament (20 m long in 1989 and 

 11m long in 1990) were baited with thawed saury Colo- 

 labis saira on curved tuna hooks (one hook/branch line) 

 and attached to the main line between float lines. 



Hooks were size 3.6 (Japanese size is 10.9cm from eye 

 to point). Each portion of the longline between floats 

 and the attached branch lines constituted a "basket," 

 a term taken from older gear in which the number of 

 branch lines is fixed. However, this study used vary- 

 ing numbers of snap-on branch lines (12, 14, 16, or 

 20/basket), depending on the length of main line per 

 basket. 



Hook position was controlled by timing the attach- 

 ment of branch lines as the main line was thrown over- 

 board mechanically at a controlled speed. A computer 

 program was used to signal and record attachment 

 times. Deviations from the programmed instructions 

 were noted, providing a record of set times for each 

 hook. The total number of hooks in each set was 

 128-600, and the amount of main line deployed per set 

 was 9-44 km (Table 1). The amount of gear increased 

 with crew experience but also varied because of incle- 

 ment weather and equipment failures. 



Set depths 



Fishing depth was altered by varying the slack in the 

 main line and the length of line per basket (Table 1) 

 and by exogenous factors such as wind and currents. 

 Line slack was quantified as the shortening rate (Saito 

 1973), or sagging rate (Suzuki et al. 1977), equal to the 

 horizontal distance between floats divided by the length 

 of line per basket (a dimensionless ratio). At deploy- 



