Griffiths: Life history and stock separation of silver kob, Argyrosomus inodorus 
65 
western Cape than in the southern Cape. Because 
capture methods were the same in all three regions, 
increased vulnerability of either sex to capture is not 
a plausible explanation either. Thus additional re- 
search is required before the regionally specific sex 
ratios observed for A. inodorus can be adequately 
explained. 
Silver kob use inshore (<120 m depth) sand and 
mud substrata as nursery areas. They apparently 
recruit (ca.1.5 cm TL) just seawards of the surf zone 
(5-7 m depth) but move offshore with growth. Upon 
attaining maturity they recruit to adult populations 
that are found on reefs. Distributional analyses have 
revealed that juvenile A. inodorus between Cape 
Agulhas and the Kei River comprise two disjunct 
distributional ranges, one in the southern Cape and 
the other in the southeastern Cape. Although the 
inshore areas of the southwestern Cape are not suit- 
able for trawling, analysis of commercial beach-seine 
catches (Lamberth et al., 1994) has revealed that 
juvenile A. inodorus (identified as “A. hololepidotus ”) 
are also found in False Bay. The existence of nursery 
areas and spawning grounds in each of the three 
sampling regions, and the differences in size at ma- 
turity and sex ratio, lend further credence to the sepa- 
rate stock concept. 
Conclusion 
Distributional analyses based on fishery dependant 
and fishery independent data revealed that there are 
three areas of silver kob abundance between Cape 
Point and the Kei River. The fact that each of these 
“populations” has its own spawning grounds and 
nursery areas and the fact that there are observed 
differences in growth rates, otolith-dimension and 
fish-length relationships, growth zone structure, 
sizes at maturity, and sex ratios, together indicate 
that these “populations” represent separate stocks. 
This three-stock concept is further supported by mi- 
gratory patterns indicated from catch and tagging 
data and by the oceanography between Cape Point 
and the Kei River. 
Although genetic differentiation should ideally 
form the basis of inferences concerning stock distinc- 
tion, analyses based on protein electrophoresis and 
mitochondrial DNA have generally been unsuccess- 
ful in differentiating between marine stocks (see 
Campana and Casselman, 1992; Pawson and 
Jennings, 1996), including those of sciaenids ( Ramsey 
and Wakeman, 1987; Graves et al., 1992; King and 
Pate, 1992). Although none of the data used to infer 
separate silver kob stocks necessarily reflect genetic 
differences (Ihssen et al., 1981), the identification of 
three allopatric units of fish with different popula- 
tion parameters indicates that each may respond 
differently to fishing and that the exploitation of one 
unit will not affect the size or composition of the other 
two, thereby supporting separate management of the 
three units and their stock status (Spangler et al., 
1981; Brown and Darcy, 1987; Campana and Gagne, 
1995; Edmonds et al., 1995; Pawson and Jennings, 
1996). 
Recent application of per-recruit models to South 
African silver kob (based on the results presented in 
this study) indicates that, owing to their different 
population parameters, each stock requires a differ- 
ent combination of fishing mortality and age at first 
capture for optimal exploitation (Griffiths, in press, 
c). Therefore A. inodorus should ideally be managed 
on a regional and not on a national basis. Studies of 
the life histories of two other South African sciaenids, 
Atractoscion aequidens (Griffiths and Hecht, 1995a) 
and Argyrosomus japonicus (Griffiths and Hecht, 
1995b; Griffiths, in press, b), have revealed that they 
consist of single stocks with allopatric age or size- 
determined subpopulations, even though they occur 
from Cape Point to southern Mozambique and are 
therefore more widely distributed on the eastern sea- 
board than are A. inodorus. Inferences from stock struc- 
ture, based on closely related taxa, are therefore not 
desirable because they could result in erroneous con- 
clusions and consequently in poor management. 
Acknowledgments 
The author thanks Chris Wilke, Peter Sims, and John 
Prinsloo (Sea Fisheries Research Institute) for pro- 
viding CPUE and length-frequency summaries for 
the line fishery, the inshore trawl fishery, and South 
Coast Biomass Surveys, respectively; M. Roberts, A. 
Boyd, and G. Nelson for oceanographic data and for 
helpful discussions on the oceanography of the study 
area; Malcolm Smale and Larry Hutchings for valu- 
able comments on an earlier draft of the manuscript; 
and all those commercial and recreational fishermen 
who made their catches available for sampling. The 
project was partially funded by the Sea Fisheries Fund. 
Literature cited 
Armstrong, M. J., B. A. Roel, and R. M. Prosch. 
1989. Long-term trends in patterns of maturity in the south- 
ern Benguela pilchard population: evidence for density- 
dependance? S. Afr. J. Mar. Sci. 8:91-101. 
Atkins, G. R. 
1970. Thermal structure and salinity of False Bay. Trans. 
R. Soc. S. Afr. 39:117-128. 
