334 



Fishery Bulletin 101(2) 



1999). Eggs of M. furnieri were identified on the basis of 

 the descriptions reported by Weiss (1981). 



Laboratory processing 



Ovaries were removed immediately after capture of the 

 fish and fixed in \Q'7c buffered formahn for two weeks. 

 The ovaries were weighed (GW) to the nearest 0.1 g and 

 a sample of about 2.0 g was removed from each ovary, 

 dehydrated in methanol, cleared in benzol, and embedded 

 in paraffin. Sections were cut at 4 pm and stained with 

 Harris' hematoxylin and eosin Y. Ovaries were classified 

 histologically into seven categories: developing early (1), 

 developing late (2), fully developed (3), gravid (4), partially 

 spent (5), spent (6) and resting (7). This classification is a 

 modification of that given by Mayer et al. ( 1988) and has 

 previously been employed to describe the ovarian develop- 

 ment in Cynoscion guatucupa (Macchi, 1998). The criteria 

 for identifying the different atresia stages and the atretic 

 condition of the ovary were adopted from Hunter and 

 Macewicz (1985). 



One hundred planktonic eggs collected during the main 

 spawning peak and during the end of the breeding season 

 were measured ( ±0.01 mm) by using a microscope equipped 

 with a calibrated eyepiece. Two diameters at right angles 

 to each other were measured per egg and an average was 

 taken. Egg weights were estimated for each portion of the 

 spawning period from two samples of 100 eggs each. The 

 eggs were rinsed in distilled water, dried for 20 hours at 

 60°C, and weighed (±0.1 mg). The diameter and weight of 

 the eggs collected during the spawning peak and the end of 

 the breeding season were compared by analysis of variance 

 (Draper and Smith, 1981). 



Estimation of spawning frequency 



Spawning frequency wa.s estimated by the incidence of 

 females with postovulatory follicles (POP), following the 

 method described by Hunter and Goldberg (1980). The 

 description of the stages of POP degeneration was based on 

 six females that spawned in captivity during the research 

 cruise carried out in March 1998. Two fish were sampled 

 at the time of spawning and the others were sacrificed 6, 

 12, 24, and 36 hours after spawning. The ovaries of these 



fish were preserved in 10% buffered formalin and used to 

 establish histological criteria for the aging of POPs of the 

 sea-caught females. 



Postovulatory follicle degradation in M. furnieri was 

 faster than that reported by Hunter and Goldberg ( 1980) 

 for northern anchovy (Engraulis mordax). A 24-h-old 

 POP showed advanced signs of degeneration similar to 

 those observed in E. mordax 48 h after spawning. There- 

 fore, daily percentage of spawning females (spawning 

 frequency) was estimated by taking the total of females 

 with POPs less than 24 h old (Hunter et al., 1986). Mean 

 and variance of this parameter were calculated accord- 

 ing to the equation developed by Picquelle and Stauffer 

 (1985), which allows weighing of each station according 

 to the subsample size. 



Fecundity estimation 



Batch fecundity (BP; number of oocytes released per spawn- 

 ing) was estimated gravimetrically with the hydrated 

 oocyte method on fixed ovarian samples (Hunter et al., 

 1985). Ovaries used showed no evidence of recent spawning 

 (i.e. no new POPs). Batch fecundity was determined for 87 

 females (57 from November 1995, 16 from Pebruary 1996, 

 three from November 1997, and 11 from March 1998). Three 

 pieces, approximately 0.1 g each, were removed from the 

 anterior, middle, and posterior part of each ovary, weighed 

 (±0.1 mg), and the number of hydrated oocytes were 

 counted. Batch fecundity for each female was the product 

 of the mean number of hydrated oocytes per unit of ovar- 

 ian weight and the total ovarian weight. Relative fecundity 

 (RP) was defined as the number of hydrated oocytes per 

 gram of ovary-free body weight. Pecundity values obtained 

 for the different months within the spawning season were 

 compared with a test of equality of means (Draper and 

 Smith, 1981). The relationships of batch fecundity to total 

 length and to total weight (ovary free) were described with 

 standard linear regressions. Comparisons between the dif- 

 ferent months were based on overlapping length ranges of 

 the samples, and an analysis of covariance with log-trans- 

 formed data was applied (Draper and Smith, 1981). 



Results 



Spawning season 



During November 1995, in the Rio de la Plata estuary, 93% 

 of the sampled females were found to be in a reproductive 

 state, with ovaries being either in a developing phase or 

 partially spent (maturity stages 2 to 5). Gravid females 

 (with hydrated oocytes) accounted for about 20% and spent 

 ovaries were scarce (Pig. 2). Near the end of this spawning 

 season (February 1996), 98% of the females had apparently 

 spawned at least once and the proportion of spent ovaries 

 was about 15%. During November 1997, females were 

 mainly undergoing ovarian development and only a few 

 individuals with hydrated oocytes were observed (Fig. 2). 

 In March 1998, the proportion of spent ovaries was about 

 20% and the proportion of those in the resting stage was 



