Gibbons et al : Habitat use by demersal nekton on the continental shelf in the Benguela ecosystem 



483 



Table 7 



Distribution tests of individuals of different species on principal components describing environmental use by dominant demersal 

 fishes in the northern and southern study sites. Levene's test of variance equality was used to compare species' variances with 

 variance of samples on each principal component. X and XX indicate a species' variance significantly smaller than location vari- 

 ance at the P<0.10 and P<0.05 level, respectively (implying that the species uses a subset of available conditions relative to that 

 principal component). 



jacopever was very small, but it showed a weaker associa- 

 tion with high-relief hard substrata and a greater associa- 

 tion with more open areas. Helicolenus dactylopterus is a 

 conspicuous spcies in trawl catches in the region (Roel, 

 1987), unlike S. capensis, and the difference in habitat use 

 would explain this. Both species are likely to be largely 

 piscivorous as adults ( Macpherson and Roel, 1987; Meyer 

 and Smale, 1991), and therefore the association of S. cap- 

 ensis with a rich epifauna is more likely an artifact of au- 

 tocorrelation with hard substrata (per se) than real. 



Kingklip have also been correlated with rock cover (Gib- 

 bons et al., 2000). Although it is a bycatch species of the 

 commercial trawl fishery over (presumably) mixed grounds 

 (Payne, 1985), a high CPUE is achieved by using long-lines 

 over hard grounds (Badenhorst, 1988). This high CPUE 

 could imply that it is found at greater densities over hard 

 grounds and that the species shows some sort of preference 



for rocky areas (as suggested by Japp, 1990). In his study of 

 feeding patterns of kingklip, Macpherson (1983) suggested 

 that it probably inhabited holes in the mud (or caves) dur- 

 ing the daytime. Our observations indicated that this spe- 

 cies hides in holes by day (Fig. 7C), but that these holes are 

 generally at the base of rocks. Such a shelter provides them 

 with ready access to their favored prey (demersal fishes) on 

 the sandy areas surrounding (Macpherson, 1983; Macpher- 

 son and Roel, 1987) and would account for a lack of associa- 

 tion with conspicuous mobile epifauna.. 



By contrast, sole have been observed by Stein et al. 

 (1992) to favor areas of soft substrata and tend to be 

 caught in greatest numbers there (Payne, 1985). Both sole 

 and dragonets (Tables 4 and 6) certainly show morphologi- 

 cal adaptations to this substratum type and are dorsoven- 

 trally flattened to a greater (sole) or lesser (dragonet) ex- 

 tent. These species were also associated with areas having 



