640 



Fishery Bulletin 94(4). 1996 



greater than 0.445 mm. Thus, only females with a 

 MOD greater then 0.445 mm were used to estimate 

 PR Fecundity was regressed on OFWT to evaluate 

 the predictability of the relationship. 



Batch fecundity was determined by calculating the 

 number of migratory nucleus-stage oocytes (MNO's; 

 Fig 2B) or hydrated oocytes (HO's; Fig. 2C> in two 

 weighed subsamples from each female (Hunter et al., 

 1992). Ovaries that contained hydrated oocytes loose 

 in the lumen were eliminated from BF analyses. A 

 modified version of the hydrated oocyte procedure of 

 Hunter et al. (1985) was also used in BF estimates. 

 Definitions from Hunter et al. (1992) were used to 

 separate each spawning batch (Table 4). To see if 

 batch size varied with the order of spawning, a one- 

 way ANOVA was run on 49 females. To evaluate the 

 temporal effects on BF, ANCOVAwas run on monthly 

 equations of BF. 



Exact ages of POF's were not known because sam- 

 pling on MARMAP cruises occurred only from dawn 

 through dusk and because trap deployment time was 

 assigned to the catch, regardless of when the fish 

 entered the trap. Therefore, we defined POF's in 

 terms of "new" and "old," on the basis of data from 

 reports in the literature and from our observations. 

 There is limited work on other lutjanids (Davis and 

 West, 1993); thus, descriptions from Hunter and 



Macewicz ( 1985) were also used to define POF's. New 

 POF's were generally large and had an orderly ar- 

 rangement of distinct columnar granulosa cells with 

 prominent nuclei intact (Fig. 3A). New POF's re- 

 sembled the early stage POF ofLutjanus vittus ( Davis 

 and West, 1993 ). Old POF's were small, coiled, eosin- 



