484 



Fishery Bulletin 100(3) 



-5 -3.5 -2 -0 5 1 2.5 4 5.5 7 



Sand 



Factor 1 



Rocks 



Sand 



I I I i"i 

 5 8 11 14 17 20 



>^ 30 -- 



.-4 QJ 



. . o CD 



-7-5-3-11357911 13 15 17 1 



Sand 



Factor 3 



Rocks 



Figure 3 



Distribution of hake (stippled), and samples ( solid i, 

 on the principal components 1 (PCI) axis which rep- 

 resented hard and soft substrata, in the sites to the 

 north of the Orange River mouth (A). Distribution of 

 hake (B), and false jacopever (C) (both stippled) and 

 samples (solid) on the PC2 axis which represented 

 biotic richness, in the sites to the north of the Orange 

 River mouth. The variance of all species scores was sig- 

 nificantly less than that of sample scores. 



a rich infauna (Table 4), which probably reflects the strictly 

 benthic nature of their diets (Macpherson and Roel, 1987). 

 An apparent lack of strong habitat preference has been 

 observed by longfin hakes (Urophycis chesteri) off North 

 Carolina (Felley and Vecchione, 1995). Although similar re- 



sults were noted here for Merluccius capensis (e.g. Tables 4 

 and 6), this species tended to avoid the rocky extremes (as 

 in Gibbons et al., 2000). Interestingly, the species was not 

 associated with substrata that were conspicuously rich in 

 either epifauna or infauna. However, because small hake 



