656 



Fishery Bulletin 90(4), 1992 



is caught, making the TDR equivalent to a hook timer 

 if it is close to a branch line that catches a fish (Saito 

 et al. 1970, Saito 1973, Yamaguchi 1989). The records 

 of TDRs at positions close to fish caught with hook 

 timers in the present study were checked to see 

 whether they indicated the time of capture, but the 

 depth of the monofilament longline gear was much less 

 stable (Fig. 3) than in the depth records of Saito et al. 

 (1970), Saito (1973), and Yamaguchi (1989) using TDRs 

 on rope gear. On monofilament longline gear, frequent 

 depth changes resembling fish captures occur even 

 when no fish are caught, making TDRs unreliable as 

 substitutes for hook timers. 



Viability of released fish 



Before the present study, it was believed that fish 

 would survive only a few hours after capture on 

 longline gear (Grudinin 1989, Yamaguchi 1989) despite 

 large pelagic species being known to survive capture 

 and release from other types of gear (Foreman 1987, 

 Squire 1987, Holts and Bedford 1989). Commercial 

 longline fishermen in Hawaii speculated that much of 

 their catch was made as hooks were sinking or rising, 

 because most were alive or appeared long dead (F. 

 Amtsberg, Der Fischen Co., Honolulu, HI 96822, pers. 

 commun., March 1988). Based on TDR data from fish 

 on regular longline gear (Yamaguchi 1989), vertical 

 movements stop 1.0-1.5h after capture for yellowfin 

 tuna, 1.5-4. Oh for bigeye tuna, and ~0.5h for spear- 

 fish and shark. This cessation of vertical movement has 

 been interpreted as death (Yamaguchi 1989). Grudinin 

 (1989) has reported on the diurnal periodicity of bigeye 

 and yellowfin tuna catch rates based on the proportion 

 recovered alive, assuming that tuna survive <2h on 

 longline gear. However, hook-timer results (Fig. 6) 

 show that fish survive much longer than this, suggest- 

 ing that vertical movement is not a reliable indicator 

 of survival. Alternatively, the results of the present 

 study could be specific to monofilament gear, which 

 could have less resistance to moving through the water 

 than does rope gear. 



Clearly the high proportion of live fish (Table 2) is 

 not primarily the result of capture during the 0.5 h 

 rising period. The viability of longline-caught fish is in- 

 dicated by their hooked longevity and the recovery of 

 tagged fish. As a management option, non-retention 

 of striped marlin and spearfish could reduce fishing 

 mortality due to longline fishing. The importance of the 

 reduction would depend on the length of the fishing 

 operation; but in the present study, longline fishing 

 mortality for striped marlin could have been reduced 

 by 70% (Table 2) if all live fish had been released and 

 had survived. 



Gear efficiency and selectivity 



Gear efficiency, defined as the dimensionless ratio of 

 the CPUE of one gear type (i.e., deep gear) divided by 

 the CPUE of the regular gear type, is the factor used 

 to calculate effective effort by gear fishing at different 

 depths (Suzuki et al. 1977). Total effective effort can 

 then be used to calculate indices of relative abundance 

 and to model stock production (Suzuki 1989). The most 

 thorough approach thus far has been to calculate gear 

 efficiency by area and season (Suzuki and Kume 1982). 

 A better understanding of the variables that alter 

 habitat depth would permit gear efficiency to be pre- 

 dicted as a function of environmental conditions, and 

 help account for variation in abundance indices caused 

 by environmental anomalies. 



The relative efficiency of standardized deep gear 

 (Table 5) follows the pattern observed in previous 

 studies (Suzuki et al. 1977, Yang and Gong 1988) in 

 which deep gear is more efficient at catching bigeye 

 tuna and less efficient at catching yellowfin tuna and 

 istiophorid billfish. However, the estimated efficiency 

 of the standardized deep gear for bigeye tuna in the 

 present study is greater (ratio 3.1-4.0 over the 2 years; 

 Table 5) than that reported by Suzuki et al. (1977) for 

 the central and western equatorial Pacific (1.8) or by 

 Yang and Gong (1988) for the Atlantic (1.9). Suzuki and 

 Kume (1982) have presented graphs of deep and reg- 

 ular CPUE for bigeye tuna on a quarterly basis by area 

 throughout the Pacific, and these data indicate very 

 little difference between gear types in the central 

 Pacific north of lat. 15°N. The high efficiency estimated 

 for deep gear in the present study may partly result 

 from using measured depths rather than inferred 

 depths to define deep and regular gear types. Also, a 

 high relative efficiency for deep gear may be specific 

 to the Hawaii area in the winter season. 



The relative efficiency of deep gear for yellowfin tuna 

 in the Atlantic (0.95, Yang and Gong 1988) is greater 

 than in the central and western equatorial Pacific (0.73, 

 Suzuki et al. 1977) and in the present study (0.65, Table 

 5). Relative efficiency of deep gear for striped marlin 

 in the central and western equatorial Pacific (0.28, 

 Suzuki et al. 1977) is much lower than in the central 

 Pacific north of Hawaii (0.74, Suzuki 1989), nicely 

 bracketing the estimate from the present studv (0 51 

 Table 5). 



The model estimates of gear efficiency (Table 5) are 

 not meant to supplant earlier estimates based on much 

 larger data sets (Suzuki et al. 1977, Suzuki and Kume 

 1982, Yang and Gong 1988, Suzuki 1989), but rather 

 to show how catch by hook position can be used to 

 estimate CPUE by different gear configurations, 

 especially hypothetical configurations for which no real 

 data exist. Efficiency estimates (Table 5) suggest that 



