Associations between sex-organ deployment and morph bias in related heterostylous taxa with different stylar polymorphisms

PREMISE OF THE STUDY: Populations of heterostylous species are characterized by two or three floral morphs with reciprocal positioning of stigmas and anthers. Theoretical models predict equal morph frequencies (isoplethy) when disassortative mating is prevalent in populations, but biased morph ratios may occur when variation in the expression of heterostyly causes deviations from intermorph mating. METHODS: We explore the role of sex-organ deployment in governing morph ratios in two closely related genera of Boraginaceae, exhibiting striking variation in floral traits associated with the heterostylous syndrome. We sampled 66 populations of six species of Glandora and 39 populations of three species of Lithodora across their distributional range in the Mediterranean. In each population we estimated morph ratios and measured several floral traits. We used phylogenetically corrected and noncorrected regressions to test the hypothesis that differences in sex-organ reciprocity and herkogamy are associated with deviations from isoplethy. KEY RESULTS: Biased morph ratios occurred in 24% of populations, particularly in Lithodora. Populations biased for the long-styled morph (L-morph) were more frequent than the short-styled morph (S-morph). Distylous species were less likely to exhibit biased ratios than species with stigma-height dimorphism. In Lithodora fruticosa, a species lacking reciprocity, decreased herkogamy in the S-morph was associated with increasing L-morph bias, perhaps resulting from self-interference. CONCLUSION: Striking variation in the expression of heterostyly in Glandora and Lithodora is associated with biased morph ratios, which probably result from pollinator-mediated mating asymmetries within populations.