Population substructuring in the Utaka (Teleostei, Cichlidae) from Lake Malawi (East Africa): do genetic and (eco-)morphological differentiation go hand in hand?  

Lake Malawi, in the southern part of the East African rift valley, harbours the largest and ecologically most diverse fish fauna in any comparable size area. It is home to at least 800 endemic cichlid species that presumably all originated from a common ancestor within less than 2-5 million years. Studies of morphological and genetic variation suggest that low dispersal capacities in combination with the patchy distribution of preferred habitats, differences in trophic biology, and differences in female mate choice preferences are the key mechanisms that have promoted the rapid diversification and speciation in cichlids. However, the Utaka, a poorly studied assemblage of more than 30 plankton-feeding cichlid species generally included in the non-Mbuna, appear less restricted to their preferred substrate throughout their life cycle than most other demersal taxa. All Utaka share (with some variations) the same distinctive feeding ecology of sucking plankton from the water column with their protractible mouth. Therefore, members of this group comprise interesting model species to test whether factors important in driving diversification in various rock- or sanddwelling demersal Malawi cichlids apply to Utaka in a comparable way.

The present study combines phylogenetic, population genetic, morphological as well as stable isotope analyses to examine patterns of diversification among Utaka populations.

The mitochondrial phylogeny of the Utaka shows that this assemblage contains two distant and geographically widespread mtDNA lineages. The lineage that contains individuals of the five studied Utaka species clusters with the non-Mbuna clade, which we consider to reflect the original phylogenetic position of the Utaka. The other lineage contains almost exclusively Copadichromis sp. 'virginalis kajose' individuals. We postulate that it originated through introgressive events involving a thus far unidentified (or extinct) non-Utaka species.

Examination of microsatellite variation among conspecific populations reveals that the Utaka species examined, though having different habitat preferences, show similar levels of genetic population substructuring. The observed levels of divergence among population pairs are unrelated to geographical distances between them. Therefore, and in contrast to various other demersal Malawi cichlids studied, we conclude that habitat preference and geographical barriers to dispersal have only a limited effect on the degree of population substructuring in the Utaka examined. We suggest that behavioural aspects, such as shoaling behaviour and periodic aggregating at specific "leks" for spawning, may determine the population structure in Utaka.

The detected morphological differentiation is not unambiguously associated with neutral genetic differentiation in Utaka. The species are morphologically well differentiated, but are less clearly delineated at the genetic level. Within each species, the populations show only subtle genetic and morphological differentiation. The low level of genetic differentiation observed between and within the studied species agrees with the evolutionary young age of the Utaka species and the potential occurrence of shared ancestral polymorphisms. We suggest that the distinct morphological differences observed between these genetically similar species are an indication that the origin of these species and their morphological divergence are typical for a rapid adaptive radiation, i.e. speciation and a significant (eco-)morphological differentiation that developed over a relatively short time span.

Peripatric populations of two Copadichromis species in Lake Malombe (a recent satellite lake that is connected to Lake Malawi by the Shire River) do not genetically nor morphologically differ from conspecific populations in the main lake. This may be explained by either the homogenizing effect of ongoing gene flow through the Shire River or by the short amount of time that these populations have been separated. Nonetheless, we find a subtle reduction in genetic variability in one of the two species. This is presumably the result of a founder event with a different impact on the amount of genetic variation of the Malombe populations of these two species. Furthermore, for both species the body size of specimens collected from Lake Malombe is significantly smaller than for specimens collected in Lake Malawi. This may reflect a response to the relatively higher fishing pressure in Lake Malombe.