644 



Fishery Bulletin 90(4), 1992 



ment, the shortening rate was the same as the ratio 

 of ship speed through the water to line-thrower speed: 

 0.40 (maximum slack) to 0.98 (no slack). The length of 

 main line per basket was 640-1427m. The predicted 

 maximum depth of the main line during each set was 

 calculated from the shortening rate and the main line 

 length per basket (Table 1), assuming a catenary shape 

 (Yoshihara 1954). 



The depth of each set was recorded with electronic 

 TDRs (Wildlife Computers, models MKII and MKIII) 

 programmed to sample depth once per minute. The 

 TDRs were attached at the deep positions, defined as 

 the attachment points for the branch line midway be- 

 tween floats (e.g., position 10 or 11 of 20 between 

 floats). In 1990, TDRs were also attached at the middle 

 positions between the deep positions and the float line 

 (e.g., at position 5 or 15 of 20 between floats). 



The time that the gear took to sink during deploy- 

 ment (0.5h) and to rise during recovery (0.5h) was 

 quantified from TDR records. Set depth was described 

 as the typical depth observed in records from the deep- 

 positioned TDRs during the period after sinking and 

 before rising. Recorded depth was examined after each 

 set and compared with predicted depth. Shortening 

 rate, the length of line per basket, or both were ad- 

 justed in the subsequent set to reach targeted depths. 



Hook depths 



The settled depth of each attachment point for the 

 branch line was estimated by interpolating between (1) 

 the TDR record for the deep and middle positions or 

 (2) the latter point and the shallowest depth of the main 

 line (assumed to equal the length of the float line). 

 Settled hook depth was calculated by adding the branch 

 line length to the interpolated depth of the branch line 

 snap. Not enough TDRs were available (2 in 1989, 10 

 in 1990) to put 1 TDR on every basket. When fish were 

 caught by baskets without TDRs, average TDR depths 

 for that set were used to interpolate settled hook 

 depths. For middle positions without TDRs in 1989, 

 depth was estimated from the mean ratio of the middle 

 position to deep-position TDR depths based on 1990 

 data. 



Hook timers 



Hook timers were made of a plastic resin cast around 

 a battery-powered microchip clock controlled by a 

 magnet (Somerton et al. 1988). They were attached to 

 the branch lines near the snap, bridging a bend in the 

 line (Fig. 1). A fish striking the hook pulled the magnet, 

 thus triggering the timer. In 1989, a rubber band held 

 the magnet in place against a test weight of about 

 l-2kg. In 1990, thread with a breaking strength of 



Main line 



Hook- 

 timer 



Snap 



Breaking 



thread 



trigger 



Holding 

 tape 



Holding y^ 

 thread ^ 



Figure I 



A hook timer and its trigger mechanism as 

 arranged in 1990, when thread triggers were 

 used. In 1989, rubber bands served as the 

 trigger. The slack loop in the branch line was 

 pulled taut when a fish struck the hook, 

 breaking the trigger and pulling the magnet 

 from its recess in the bottom of the hook 

 timer. 



4-5 kg bridged the bend in the line, and the magnet was 

 held in place by a weaker thread until the bridging 

 thread was broken (Fig. 1). Some branch lines were set 

 without timers (14% in 1989, 35.5% in 1990) to pre- 

 clude interruptions in fishing when timers were tangled 

 or otherwise unavailable. 



Hook timers indicated elapsed time in whole minutes 

 (e.g., Omin indicated 0-59 s). Timers were read as the 

 branch lines were recovered, or soon after, with cor- 

 rections made for delays. Timers were categorized as 

 being triggered (1) at recovery (< 1 min before remov- 

 ing the branch line snap), (2) while rising (>lmin- 

 0.5h before recovery), (3) while settled (>0.5-<1.0h, 

 l-<2, 2-<3h, and so on before recovery), (4) while 



