64 
Fishery Bulletin 95( 1 ), 1997 
Figure 1 5 
Bottom temperature structure on the central and eastern Agulhas Bank during the October 1987 
South Coast Biomass Survey. This temperature pattern, with cold water (<12°C) at shallower 
depths and closer to the coast between Knysna and Cape St Francis, was typical of SCBS’s be- 
tween 1987 and 1995. See Figure 1 for additional localities. 
new recruits are mature in the southeastern Cape, 
but only 69% in the southern Cape. Assuming that 
size at maturity for silver kob in the southeastern 
Cape was at one time similar to that in the southern 
Cape, the removal of late-maturing fish, before they 
had spawned for the first time, could have reduced 
the sizes and ages at maturity to those recorded in 
this study. The large contribution of recruits (400- 
450 mm TL) to the southeastern Cape line catch (ca. 
50% by numberXFig. 12), supports this hypothesis. 
Investigation of sex ratios showed that there are 
substantially more females in the southeastern Cape 
(1.6x) and southwestern Cape (2. lx) stocks, but more 
males in the southern Cape (1.2x). Most natural 
populations tend to stabilize at sex ratios of 1:1 
(Conover and Van Voorhees, 1990), including those 
of other sciaenids (Shepherd and Grimes, 1984; 
Murphy and Taylor, 1989; Wilson and Nieland, 1994; 
Ross et al., 1995; Griffiths and Hecht, 1995a). The 
deviations from this ratio observed for South Afri- 
can silver kob are not easily explained. Basically, the 
reasons for an observed sex ratio that deviates from 
unity may be grouped into three categories: 1) more 
individuals of either sex are produced (e.g. environ- 
mental sex determination); 2) equal numbers of both 
sexes are produced, but those of one sex are dimin- 
ished through either emigration or mortality; and 3) 
sampling methods are biased towards one of the 
sexes. Although environmental sex determination 
can temporarily result in skewed sex ratios in some 
species (Conover and Heins, 1987), frequency-depen- 
dent selection is expected to return the ratios of such 
populations to equality through future generations 
(Conover and Van Voorhees, 1990). Higher propor- 
tions of either sex over most size classes (therefore 
spanning several age classes) in each region, with 
consistency over two periods (1978-81 and 1990-91) 
in the southeastern Cape, render environmental sex 
determination an unlikely cause of the observed sex 
ratios. Male emigration from the southeastern Cape 
and the southwestern Cape to the southern Cape is 
also unlikely. Because the smallest size class sampled 
in the southeastern Cape consisted of males without 
drumming muscles and also consisted of more males 
than females, it is tempting to suggest that male 
drumming during the protracted spawning season 
may have resulted in sex-selective predation in this 
region and in the southwestern Cape. However, there 
is no reason to believe that predation rates should 
be higher in the southeastern Cape and the south- 
