How does the taxonomic status of allopatric populations influence species richness within African cichlid fish assemblages?

Aim  Current estimates of species richness within rapidly evolving species flocks are often highly dependent on the species status of allopatric populations that differ in phenotypic traits. These traits may be unreliable indicators of biological species status and systematists may have inconsistently assigned species among lineages or locations on the basis of these traits, thus hampering comparative studies of regional species richness and speciation rates. Our aim was to develop a method of generating standardized estimates of regional species richness suitable for comparative analysis, and to use these estimates to examine the extent and consistency of species assignment of allopatric populations within rapidly evolving cichlid fish flocks present in three east African lakes. Location  Lakes Malawi, Victoria and Tanganyika. Methods  Using published taxon co-occurrence data, a novel approach was employed to calculate standardized ‘minimum’ estimates of regional species richness for hard substrate associated complexes of cichlids within each of the lakes. Minimum estimates were based on an explicit assumption that if taxa present on equivalent habitats have disjunct distributions, then they are allopatric forms of the same species. These estimates were compared with current observed ‘high-end’ regional species richness estimates for those complexes to determine the consistency of species assignment of allopatric populations between lineages within a lake. A ‘sympatry’ index was developed to enable comparisons of levels of species assignment of allopatric populations between-lakes to be made. Results  Within each lake, the minimum and high-end estimates for species richness were significantly correlated across complexes, indicating that the complexes that contain more recognized species contain the most genuine biological species. However, comparisons of complexes among lakes revealed considerable differences. For equivalent geographical areas, substantially higher proportions of recognized species were totally allopatric within the studied Lake Malawi and Lake Victoria complexes, than those of Lake Tanganyika. Main Conclusions  Among African lakes, levels of assignment to species status of allopatric populations were found to be distinctly different. It is unclear whether the discrepancies are a consequence of differences between the lake faunas in degrees of phenotypic divergence among allopatric populations, or are simply the result of inconsistent taxonomic practices. In either case, these results have considerable wider relevance for they emphasize that quantitative measures of regional and beta diversity are critically dependent on the species status of allopatric populations, an issue usually neglected in comparative studies of species richness. The technique introduced here can be used to standardize measures of regional diversity of lineages for comparative analyses, potentially enabling more accurate identification of processes influencing rates of speciation.