652 



Fishery Bulletin 102(4) 



40 



35 



30 



25 



20 



15 L 



80 



60 S 



40 



20 



J 



Month 



Figure 1 



Mean monthly salinities I±1SE) at the bottom of the water column 

 in the lower Swan River Estuary throughout the year and the preva- 

 lences of atresia in mature ovaries of Rhabdosargus sarba between 

 July and November, which are shown as histograms, together with 

 the number of fish examined. Closed rectangles on the horizontal 

 axis refer to summer and winter months, and the open rectangles 

 to autumn and spring months. 



cal characteristics are presented in Table 1. Because 

 stages V and VI could be distinguished macroscopi- 

 cally only during the period of oocyte hydration, the 

 macroscopic data for these two stages had to be com- 

 bined for other times. The diameters of the oocytes in 

 histological sections of an ovarian lobe from each of two 

 mature female R. sarba caught during the spawning 

 season — oocyte diameters that were typical of those 

 from mature R. sarba during this period — formed an 

 essentially continuous distribution (Fig. 2). This distri- 

 bution reflected the presence of oocytes at all stages in 

 development from chromatin nucleolar oocytes to yolk 

 granule oocytes and demonstrated that R. sarba has 

 indeterminate fecundity sensu Hunter et al. (1985). 

 Thus, the potential annual fecundity is not fixed prior 

 to the commencement of the spawning period and conse- 

 quently the potential annual fecundity of R. sarba has to 

 be estimated by using a combination of batch fecundity 

 and spawning frequency. 



Period of hydration and spawning 



The diameters of oocytes in ovaries of fish collected at 

 intervals on 1-2 September 2001 and 13 September 

 2001 and which had been retained on the 125-^im sieve, 

 produced a modal class that, for each time interval, 

 fell between 420 and 600 fim (Fig. 3). At ca. 18:30 h on 

 1 September 2001, the oocyte diameters formed a single 

 mode, and the vast majority of oocytes were less than 

 720 ;im and produced a modal class at 420-539 um 

 (Fig. 3). However, by ca. 21:30 h on the same evening. 



the maximum diameter of the oocytes had increased 

 markedly and the distribution of the oocyte diameters 

 was beginning to become bimodal. with modal classes at 

 480-539 and 780-839 ^m. By 00:30 h on 2 September, 

 the oocyte diameter distributions had become markedly 

 bimodal, and the modal diameter class of the largest 

 oocytes at this time, and also at 03:30 h, lay between 

 840 and 959 ^(m (Fig. 3). The oocyte diameter frequen- 

 cies on 13 September were essentially the same as those 

 at similar times on 1 September; the distributions were 

 unimodal at 18:30 h and bimodal at 22:30 h (Fig. 3). 

 The oocyte diameters of each fish within a given time 

 slot on 1, 2, and 13 September exhibited essentially the 

 same distribution. 



Histological sections showed that, at 18:30 h on 

 1 September 2001, most of the mature ovaries contained 

 migratory nucleus stage oocytes, i.e., oocytes in which 

 the nucleus was migrating towards the edge of the cy- 

 toplasm and a conspicuous lipid droplet was present in 

 the cytoplasm (Fig. 4A). However, it was difficult at this 

 time to distinguish migratory nucleus oocytes from yolk 

 granule oocytes under a dissecting microscope (Fig. 4B). 

 By 21:30 h, the yolk and lipid of the larger oocytes had 

 begun to coalesce and the nucleus could sometimes be 

 seen near the edge of the cytoplasm (Fig. 4Cl. Their 

 relatively larger size, translucent appearance, and one's 

 ability to detect their lipid droplet enabled these hydrat- 

 ing oocytes to be far more readily distinguished from 

 yolk granule oocytes under a dissecting microscope 

 than was the case earlier in the evening (cf. Fig. 4, B 

 and D). By 00:30 h. the largest oocytes had increased 



