100 



Fishery Bulletin 94(1). 1996 



atresia (none, <50%, and >50%), i.e. atretic classes 

 1, 2, and 3, based on visual estimates of the propor- 

 tion of atretic oocytes and presence or absence of beta 

 (/}) atresia, atretic class 4. 



Postovulatory follicles were staged according to 

 their appearance and were assigned ages on the ba- 

 sis of the time of capture relative to the estimated 

 time of spawning. The degeneration and resorption 

 of yellowfin tuna postovulatory follicles were de- 

 scribed and classified into a series of distinct stages, 

 each with an assigned time after spawning, in 6-hour 

 intervals. An estimate of the spawning time and fre- 

 quency for females was obtained from the evalua- 

 tion of this series of yellowfin tuna ovarian tissues. 



The classification of yellowfin tuna testes was 

 based on the size of the duct, the thickness of the 

 myoid tissue (Grier et al., 1989) surrounding the duct, 

 the amount of spermatozoa within the duct, the de- 

 gree to which the duct was convoluted, whether the 



tissue adjacent to the duct appeared to be heavily 

 nucleated, and the staining characteristic of the tis- 

 sue adjacent to the duct. 



Slides were scored by stage number and specimen 

 identification was not revealed. The original vari- 

 ables were subjected to discriminant analysis oper- 

 ating on the covariance matrix by using BMDP sta- 

 tistical software (Dixon et al., 1990). This procedure 

 separates individuals into two or more groups based 

 on characteristics of variables by which the groups 

 differ (Johnson and Wichern, 1982). The two groups 

 evaluated with this technique were those individu- 

 als sampled between 0001 and 1200 h (group 1) and 

 those sampled from 1201 to 2400 h (group 2). These 

 two time intervals were chosen on the basis of the 

 approximate modal time of spawning for females. An 

 estimate of the spawning frequency for males was 

 obtained from the evaluation of this series of yellow- 

 fin tuna testicular tissues. 



Oocyte size distributions and estimation of 

 batch fecundity 



Oocytes from a 1-g tissue sample were measured to 

 the nearest 0.03 mm at 30x magnification with an 

 ocular micrometer in a dissecting microscope. For 

 each fish, mean diameter (random axis) was deter- 

 mined from 20 oocytes present in the most advanced 

 mode of oocytes. 



Following the procedures of Hunter et al. (1985), 

 batch fecundity (number of oocytes released per 

 spawning) was estimated for 12 females. For each 

 female, counts of migratory-nucleus or hydrated oo- 

 cytes in three subsamples of about 0.05 g each were 

 weighed to the nearest 0.1 mg. Migratory-nucleus 

 and hydrated oocytes can be easily distinguished 

 from other oocytes by their larger size and appear- 

 ance. Those with migratory nuclei are less opaque 

 than yolked oocytes, whereas hydrated oocytes are 

 translucent. No new postovulatory follicles were 

 present in the ovaries used for estimation of batch 

 fecundity. 



Each of the three subsamples yielded an estimate 

 of batch fecundity for each female, calculated from 

 the product of the number of migratory-nucleus or 

 hydrated oocytes per unit weight of the subsample 

 and the total weight of the ovaries. The mean of these 

 three estimates provided the spawning batch fecun- 

 dity estimate for each fish. The weight of each fish 

 used to estimate batch fecundity was estimated from 

 the length-weight relationship of Chatwin (1959) in 

 order to calculate relative fecundity ( oocytes per gram 

 of body weight). To describe the relationship between 

 batch fecundity and FL for this set of data, a linear 

 regression between the natural logarithms of the 



