The high expression pattern for GL2 and TRY-1 in B. villosa leaves led us to compare the coding sequences of the five trichome regulatory genes we had isolated from B. villosa with those in B. napus, two diploid Brassica species, and A. thaliana. Extreme trichome coverage in B. villosa leaves may be due to polymorphism and evolutionary differences that impact on protein function rather than solely to specific gene transcript levels, which are dictated by promoter ����strength���� and intra-gene regulatory structures. Although analysis showed few differences Chlorthalidone between most orthologues, overall evolutionary selection was detected between GL1 proteins of hairy and glabrous genotypes by their high pairwise Ka/Ks values, potentially due to sites involved in adaptive change. Adaptive changes may occur at surprisingly few sites; consequently, the overall Ka/Ks ratio for an entire protein may remain dominated by non-adaptive changes and be substantially lower than unity as seen by the Ka/Ks scores for GL2, EGL3, TTG1, and TRY. Moreover, once a protein achieves a new advantageous function, the frequency of non-synonymous substitutions at the adapted sites will be reduced by new functional constraints. A substantial proportion of the differences that distinguished the four Brassica trichome regulatory sequences occurred outside of conserved Entacapone domains. Several of these individual amino acid differences were sufficiently dramatic to potentially change the molecular and functional properties of these proteins. Particularly, three variable positions in GL1 and GL2 translated sequences distinguished hairy B. villosa and B. rapa from glabrous B. napus and B. oleracea germplasm, and all the B. villosa genes had a minimum of one unique site compared with the other species. Bloomer reported on the effects of natural variation in GL1 from Arabidopsis and suggested that qualitative differences in trichome phenotypes might have arisen independently several times by three unique protein coding changes and a whole locus deletion. The same authors also suggested that quantitative variation might have arisen because of completely linked amino acid replacements and mutations in a known enhancer region within the AtGL1 locus.