252 



Fishery Bulletin 102(2) 



120 



125 



130 



135 



140 E 



Figure 1 



Fishing grounds of the gillnet (cross lines) and longline (oblique lines) fish- 

 ing boats based at Shinkang fishing port. 



Taiwan. This information could be used to determine the 

 age composition of the catch and to assess the status of 

 sailfish in these waters by using yield-per-recruit or se- 

 quential population analysis techniques. 



Materials and methods 



Materials 



Data on total length (TL), eye fork length (EFL), lower 

 jaw fork length (LJFL) (in cm), round weight (RW) (in kg) 

 and the first dorsal fins of male and female sailfish were 

 collected monthly at the fishing port of Shinkang (Fig. 1) 

 from July 1998 to August 1999. In total, 304 TLs, 1166 

 LJFLs, 1166 RWs, and 1166 dorsal fins were collected. 

 The dorsal fins were kept in cold storage before being 

 boiled to remove surrounding tissue and to separate the 

 fourth spines. Three cross sections (thickness 0.75 mm) 

 were taken successively along the length of each spine 

 with a low-speed "ISOMET" saw (model no. 11-1280) and 

 diamond wafering blades, at a location equivalent to 1/2 of 

 the maximum width of the condyle base measured above 

 the line of maximum condyle width (Fig. 2A) (Ehrhardt 

 et al., 1996; Sun et al., 2001, 2002). The sections were 



immersed in 95'/J ethanol for several minutes for cleaning, 

 placed on disposable paper to air dry, and then stored in a 

 labeled plastic case for later reading. Spine sections were 

 examined with a binocular dissecting microscope (model: 

 Leica-MZ6) under transmitted light at zoom magnifica- 

 tions of 10-20x depending on the sizes of the sections. The 

 most visible one of these three sections was read twice, 

 approximately one month apart. If the two ring counts 

 differed, the section was read again, and if the third ring 

 count differed from the previous two ring counts, the spine 

 was considered unreadable and discarded. The precision 

 of reading was evaluated by using average percent error 

 (APE) (Beamish and Fournier, 1981; Campana, 2001) and 

 coefficient of variation (CV) (Campana, 2001) statistics. 



Images of the cross sections were captured by using the 

 Image-Pro Image analysis software package (Media Cy- 

 bernetics, Silver Spring MD, 1997) in combination with a 

 dissecting microscope equipped with a charged coupled de- 

 vice (CCD) camera (model: Toshiba IK-630) and a Pentium 

 II computer equipped with a 640x480 pixel frame grab 

 card and a high-resolution (800x600 pixel) monitor. 



The distance from the center of the spine section to the 

 outer edge of each growth ring was measured in microns 

 with the Image-Pro software package after calibration 

 against an optical micrometer. The center of the spine 



