Plant Reproduction

The most outstanding feature of flowering plants is their incredible diversity of reproductive structures. Flowering plants display more variation in their reproductive organs than the equivalent structures in any other group of organisms, and the evolution of this diversity has fascinated evolutionary biologists since Darwin. Measuring natural selection on reproductive traits is critical for understanding the evolution of floral diversification. Hermaphroditic sexual systems require measurements of selection through both female and male function, and this is rarely conducted in outcrossing plants due to difficulties of directly estimating variation in male fitness in wild populations. The application of sexual selection theory to hermaphroditic plants has led many researchers to conclude that attractive floral displays evolve mainly in response to selection through male function. Thus far, empirical support for this hypothesis is limited and there is growing evidence that the ecological and reproductive context of populations is important for determining the nature of sex specific selection on floral traits. However, few studies have used direct measurements of male and female selection gradients to critically evaluate these hypotheses. I estimated natural selection on floral traits in a wild daffodil and found contrasting patterns of selection through male and female function, revealing that estimates of selection in hermaphrodites through female function alone can be very misleading.

Stylar polymorphisms are classic examples of morphological adaptation between flowers and their pollinators. Heterostyly is one such polymorphism, where there are two (distyly) or three (tristyly) style-morphs within populations. In most heterostylous species, reciprocity of sex-organ position coupled with heteromorphic incompatibility results in symmetrical mating and the evolution of equal style-morph ratios. However, my work has demonstrated that alternatives to this symmetry in form and function are possible. Using parentage analysis and variation at microsatellite loci I demonstrated for the first time in a heterostylous species the existence of asymmetrical mating patterns among the style-morphs. In particular, I found that imperfect sex-organ reciprocity in the wild daffodil Narcissus triandrus results in asymmetrical mating and the evolution of biased style-morph ratios. These features are the consequence of the unusual association between stylar polymorphism and a self-incompatibility system that permits assortative mating in this species. These empirical results agreed with predictions of my pollen transfer models and demonstrated that small, but functionally significant variation in floral design governs fertility and mating patterns among morphs resulting in the evolution of biased morph ratios.