489 



Abstract.-Swordfish (Xiphias gla- 

 diusi caught by the Hawaii-based 

 pelagic longhne fishery during March 

 1994-June 1997 were examined at sea 

 by observers of the National Marine 

 Fisheries Service, Southwest Region. 

 Observers provided unbiased size and 

 sex composition data for 4.8'J of the 

 swordfish catch and 4.9'* of the effort in 

 number of hooks of the fishery during 

 the 40-mo. period. Observers measured 

 body lengths for more than 8600 

 swordfish brought aboard participat- 

 ing vessels; sex, based on macroscopic 

 appearance of gonads, was identified 

 aboard ship for 7T7c of measured fish. 

 Sex identifications were later verified 

 (0.5% error rate) and gonadal develop- 

 mental stage described for 1336 fish 

 whose sex was identified in the field. 

 Logistic regression was used to estimate 

 sex-specific, median body size at sexual 

 maturity (L-„ ) by using microscopic mor- 

 phological evidence for gonadal develop- 

 ment. Z-jo was 102 cm ±2.5 (95'J CI) cm 

 eye-to-fork length (EFL) and 144 ±2.8 

 cm EFL for males in=506) and females 

 (f!=822), respectively. Sex ratios were 

 an increasing power function between 

 100 and 220 cm, and nearly all fish 

 >220 cm EFL were females. Sex compo- 

 sition and body size varied temporally 

 and spatially, especially the latter Rel- 

 atively more males were caught south 

 of 27'N; fem.ales dominated catches 

 north of 27°N. Small-bodied fish of both 

 sexes prevailed year-round below 22 N. 

 A greater percentage of large-bodied 

 (>156-cm [males). >172-cm [females] 

 EFL) fish were caught north of 35 'N 

 during the late summer-early winter 

 The latter observations are consistent 

 with several nonmutually exclusive 

 hypotheses of migration energetics and 

 body muscle heat conservation, both of 

 which are discussed. 



Sexual maturity, sex ratio, and size composition 

 of swordfish, Xiphias gladius, caught by the 

 Hawaii-based pelagic longline fishery 



Edward E. DeMartini 

 James H. Uchiyama 



Happy A. Williams 



Honolulu Laboratory 



Southwest Fisheries Science Center 



National Marine Fisheries Sen/ice, NOAA 



2570 Dole Street 



Honolulu, Hawaii 96822-2396 



E-mail address (for E E DeMartini) Edward Martinrffnoaa gov 



Manuscript accepted 1 December 1999. 

 Fish. Bull. 98: 489-506 (2000). 



Rapid expansion of the Hawaii-based 

 pelagic longline fishery for swordfish 

 iXiphias gladius) during the late 1980s 

 and early 1990s has generated a need 

 for an explicit management plan and 

 assessment of swordfish stocks in the 

 Pacific. Assessments for swordfish might 

 be improved by using size- or age-struc- 

 tured, rather than surplus production, 

 models for several reasons. Size- and 

 age-structured models are less depen- 

 dent on effort statistics (Gulland and 

 Rosenberg, 1992) and in the Pacific as 

 in the Atlantic, adequate effort statis- 

 tics are largely unavailable for pelagic 

 longline catches. Effort data are further 

 complicated by geographically sepa- 

 rated fisheries that differentially target 

 swordfish or other pelagic fishes and 

 by multiple jurisdictions with varying 

 data collection standards. In the Atlan- 

 tic these difficulties might be overcome 

 by using nonequilibrium production 

 models and virtual population analyses 

 (ICCAT, 1997). 



Estimates of body size and age at 

 sexual maturity are of fundamental 

 importance for the determination of 

 fishery management thresholds based 

 on size- and age-structured stock assess- 

 ments. In particular, determination of 

 the spawning potential ratio (SPR: 

 Goodyear, 1993) depends on size at 

 maturity. Sex-specific size composition 

 data are also needed to stratify catch 

 and effort statistics, thereby reducing 

 the variances of CPUE estimates. 



Information on reproductive biology is 

 inadequate for swordfish in the Pacific. 

 Nakamura et al. (1951) were the first 



to infer spawning seasonality in the 

 western Pacific based on net collections 

 of swordfish larvae. Yabe et al. (19591 

 further described swordfish eggs and 

 larvae and used plots of gonad weight 

 versus body length for fish on spawn- 

 ing grounds to approximate body size 

 at sexual maturity for female sword- 

 fish in the western Pacific. Kume and 

 Joseph (1969) estimated body size at 

 sexual maturity for female swordfish in 

 the eastern Pacific, and Sosa-Nishizaki 

 (1990) estimated body size at sexual 

 maturity for female swordfish through- 

 out the North Pacific. Uchiyama and 

 Shomura (1974) estimated total fecun- 

 dities for eight fish and provided anec- 

 dotal evidence of swordfish spawning 

 near the Hawaiian Archipelago. Adult- 

 size swordfish caught in the eastern 

 North Pacific were described by Weber 

 and Goldberg (1986) as reproductively 

 inactive during late August-November. 

 Hinton and Deriso's ( 1998) more recent 

 evaluation of swordfish gonadal index 

 data from the Japanese longline fish- 

 ery, however, has documented the sea- 

 sonal presence of reproductively active 

 swordfish near Baja California during 

 May-August. Although largely non- 

 existent for the Pacific, reproductive 

 and related growth parameters have 

 recently been estimated for swordfish 

 in the Northwest Atlantic (Arocha et 

 al., 1994; Arocha and Lee, 1995, 1996; 

 Ehrhardt et al., 1996; Arocha, 1997). 



Catch and effort data for swordfish 

 in Pacific fisheries have been summa- 

 rized by Miyabe and Bayliff ( 1987) and 

 Nakano and Bayliff ( 1992 ) and reviewed 



