Patterns of DNA sequence variation in the ribosomal DNA (rDNA) second internal transcribed spacer (ITS2) and five unlinked single-copy nuclear loci were examined for evidence of reproductive isolation among four chromosomally recognized taxa of Anopheles gambiae from West Africa: Savanna, Bamako, Mopti and Forest, as well as sibling species An. arabiensis and An. merus. Included among the single-copy loci were three sequence-tagged random amplified polymorphic DNA (RAPD) loci, two of which (R15 and R37) had been reported as discriminating between Mopti and other chromosomal forms. Each of the five single-copy sequences were highly polymorphic in most samples. However, the R15 and R37 loci had no diagnostic value, and therefore are not recommended as tools in recognition of field-collected An. gambiae chromosomal forms. Although pairwise comparisons between species generally revealed significant levels of differentiation at all five loci, variation was not partitioned by chromosomal form within An. gambiae at any single-copy locus examined. The few exceptions to these trends appear related to a location either inside or nearby chromosomal inversions. At the tryptophan oxygenase locus inside inversion 2Rb, variation was structured only by inversion orientation and not by taxonomic designation even between An. gambiae and An. arabiensis, providing the first molecular evidence that the 2Rb inversion was transferred between species by introgressive hybridization. By contrast, the rDNA showed fixed differences between species and a difference diagnostic for Mopti, consistent with effective, if not complete, reproductive isolation. The apparent disagreement between the data from this locus and multiple single-copy loci within An. gambiae may be explained by the much lower effective population size of rDNA, owing to concerted evolution, which confers increased sensitivity at much shorter divergence times. Taken together with the accompanying reports by della Torre et al. (2001), Favia et al. (2001) and Gentile et al. (2001), our data suggest that neutral molecular markers may not have the sensitivity required to detect isolation between these recently established taxa.