It has been thought that the geographical distribution of demersal communities (as defined by Longhurst) could be tentatively delineated by using information on the position of the 20°c isotherm and bottom types associated to those assemblages. The analysis of the data presented above shows that it is possible to determine the geographical distribution of some demersal fish assemblages, especially sciaenid, sparid and lutjanid, by using such type of information. The maps produced may not be fully accurate, but they already are a valuable piece of information (this point is developed later) - our short analysis indicated the possibility that in equivalent assemblages dominant species might change from place to place. The analysis of data from trawl surveys for small areas (in spite of the general agreement with the assemblages described by Longhurst) shows the existence of apparent interruptions in the assemblages' geographical distribution, and changes in their most frequent component species. Certain characteristics of the habitat, such as the coastal rocky grounds existing off Liberia-Ivory Coast, the upwelling off Ghana-Togo and the presence of important rivers (Nigeria-Cameroon) may be obstacles to the general distribution of some assemblages, especially of those living near the coast (lutjanid and the estuarine component of the sciaenid assemblage).
The sparid assemblage living in deeper waters under more constant environment, also shows changes of the constituent species in the upwelling region (Ghana-Togo) and around Cape palmas that suggest that the sparid assemblages from Ghana-Togo and Liberia are different from those from neighbouring countries.
There are also indications that “estuarine” sciaenid and sparid groups from Liberia could be different from those of the Ivory Coast, and that species assemblages in Ghana and Togo have more affinities between each other than with those from the Ivory Coast. Furthermore, differences found in Cameroon in relation to the relative importance of the assemblages and their “dominant” species, coupled with the observed smallness of individuals support the idea that Cameroon communities might be isolated from neighbouring areas.
These facts could be interpreted in many different ways. The observed difference could reflect sampling inadequacy, actual geographical differentiation of slightly different assemblages, that could also indicate that species hierarchy structure in an assemblage is not fixed for ever but changes with time under environmental influence (including possible fishing) - the apparent geographic differentiation might then just reflect different snaphots of the evolutions of the assemblages in different areas.
Regarding temporal persistence, however, it should be noted that the most frequent species found off Ghana were the same in 1969–70 and in 1979–80. The same is also valid for the surveys carried out in 1963 and 1976 in Cameroon.
More precise information is needed on species composition, abundance, bottom temperatures, bottom types, in order to analyze the internal persistence of the assemblages specific structure in the spatial as well as temporal dimensions.
