490 



Fishery Bulletin 98(3) 



by Sosa-Nishizaki ( 1990) and Sosa-Nishizaki and Shimizu 

 ( 1991 ). DiNardo and Kwok ( 1998 ) have provided a prelimi- 

 nary description of catch statistics and the body size and 

 sex composition of swordfish caught by the Hawaii-based 

 longline fishery in the central North Pacific. 



Our study had several complementary objectives, and 

 the ultimate goal of providing accurate and precise infor- 

 mation for stock assessments of Pacific swordfish. In our 

 study, we estimated size at sexual maturity for swordfish 

 captured by the Hawaii-based pelagic longline fishery in 

 the central North Pacific. Because the growth rates of adult 

 male and female swordfish differ in the Pacific (Uchiyama 

 et al., 1998), as elsewhere (Ehrhardt et al, 1996), size 

 at maturity was characterized separately for males and 

 females. We classified maturity on the basis of histologi- 

 cal analyses of gonads subsampled from a larger sample 

 of field-sexed fish and used this maturity classification as 

 the basis for our estimates of size at maturity. The latter 

 are compared with a complementary characterization of 

 reproductive activity using gonad indices (Hinton et al., 

 1997). Data on body size (length), date, and location of cap- 

 ture of swordfish in the parent sample of field-sexed fish 

 were used to provide preliminary descriptions of temporal 

 and spatial patterns of size- and sex-specific catch. These 

 patterns of sex and size composition identified factors that 

 could be used for stratifying catch and effort data for the 

 Hawaii-based longline fishery, thereby reducing the vari- 

 ance of population parameter estimates based on logbook 

 records used in future stock assessments. Finally, we inter- 

 preted spatial and temporal variations in catch in terms of 

 possibly different migratory behaviors by swordfish of dif- 

 ferent body sizes and sexes. 



Methods and materials 



Shipboard collections and measurements 



About 90'^ of the swordfish analyzed were caught by com- 

 mercial vessels fishing pelagic longlines in the central 

 North Pacific from March 1994 to June 1997. Most (65^7^: 

 Ito and Machado, 1996,' 1997,2 19993) swordfish were 

 caught on trips on which swordfish were targeted. Another 

 349i were caught on "mixed sets" (both tuna and swordfish 



Ito, R. Y, and W. A. Machado. 1996. 

 Hawaii-based longline fishery for 1995. 

 Cent. Admin. Rep. H-96-12, Honolulu 

 Fish. Sci. Cent., Natl. Mar. Fish. Serv. 

 96822-2396, 45 p. 



Ito, R. Y., and W. A. Machado. 1997. 

 Hawaii-based longline fishery for 1996. 

 Cent., Admin. Rep. H-97-12.' Honolulu 

 Fish. Sci. Cent., Natl. Mar Fish. Serv. 

 96822-2396, 48 p. 



Ito, R. Y, and W. A. Machado. 1999. 

 Hawaii-based longline fishery for 1998 

 Cent. Admin. Rep. H-99-()6," Honolulu 

 Fish. Sci. Cent., Natl. Mar Fish. Sei-v. 

 96822-2396, 62 p. 



Annual report of the 



Southwest Fish. Sci. 



Laboratory, Southwest 



, NOAA, Honolulu. HI 



Annual report of the 



Southwest Fish. Sci. 



Laboratory, Southwest 



, NOAA, Honolulu. HI 



Annual report of the 



Southwest Fish. Sci. 



laboratory. Southwest 



, NOAA, Honolulu, HI 



targeted), and the remaining 1% on "tuna sets" (Boggs and 

 Ito, 1993; He et al., 1997). The remaining 10'^?^ of sword- 

 fish examined were caught on swordfish research cruises 

 of the NOAA ship Townsend Cromwell during 1992-97 in 

 the region of the commercial fishery. 



Commercial fishermen dressed all swordfish as the fish 

 were brought aboard ship; observers noted that about IC/r 

 were alive and 90'?f^ moribund or dead upon retrieval. Via- 

 bility offish when sampled is important for interpreting the 

 quality of histological specimens. Observers (NMFS SWR) 

 were assigned to participating longline vessels according 

 to a stratified random design based on vessel size (effort). 

 Observers recorded eye-to-fork length (EFL, cm) for most 

 swordfish caught. For most swordfish that were measured, 

 sex was identified on the basis of macroscopic appearance 

 of the gonads, and, for a random subsample of sexed fish, 

 a gonad tissue sample (2 cm', including gonad wall) was 

 collected from the middle of either gonad lobe and imme- 

 diately fixed in 10% buffered formalin. Swordfish gonad 

 samples were collected during most months of four consec- 

 utive years (172 trips sampled during March 1994-June 

 1997), complemented by specimens collected by research 

 cruises in April-May of 1992 and 1993, September 1996, 

 and March-April 1997. Specimens examined histologically 

 thus spanned several annual cycles; overall SlVr of the 

 fish examined were collected during March-July spawning 

 periods (Fig. 1). Fish were sampled throughout the spa- 

 tial range of the longline fishery during March 1994-June 

 1997, mostly between lat. 17-41°N and long. 141-180°W 

 (Fig. 2). 



During spring 1997, SWR observers collected whole ova- 

 ries of approximately 100 swordfish as they were being 

 dressed aboard ship. For each pair of ovaries, a fresh tissue 

 specimen was collected and fixed for later histological eval- 

 uation of developmental stage (as described below), and 

 the remainder of the ovaries were frozen for weighing 

 ashore. The latter samples were used in analyses of gonad 

 indices. 



Laboratory processing 



Field-classified sex was verified and reproductive condition 

 evaluated based on gonadal developmental stage deter- 

 mined by microscopic examination of histological prepara- 

 tions and oocyte size frequencies. Fixed gonad specimens 

 were stored for at least 60 days before oocytes were mea- 

 sured or a subsample was prepared for histological analy- 

 ses. A single series of sequential, histological sections (6 

 pm thick) (2-8: mode=3[ovaries], mode=4|testes|) was cut 

 and stained with Harris's hematoxylin, followed by eosin 

 counterstain (Hunter and Macewicz, 1985). Slide sections 

 were examined with a compound microscope at 60-300x 

 and developmental stage categorized following the stage 

 criteria of Murphy and Taylor (1990). For females, the 

 following characteristics were noted for the largest size 

 mode of oocytes present: presence and relative quantity 

 of eosinophilic yolk, partly to moderately yolked oocytes, 

 and fully yolked oocytes. If oocytes were fully yolked, 

 we further noted the presence of hydrating or hydrated 

 oocytes (HYDs), postovulatory follicles (POFs), and fully 



