Increased variation in sex organ positions can increase reciprocity and pollination success in heterostylous plant populations

Aims: The reciprocal position of sexual organs in complementary floral morphs is central to our understanding of heterostyly. Reciprocity indices are used to quantify the spatial match between complementary sex organs, but previous indices fail to appropriately account for variation in sex organ positions among individuals in a population. The objective of this study was to examine how reciprocity and consequently reproductive success change with an increase in intra-population variation in sex organ heights. To this end, we formulated a reciprocity index that incorporates variation in sex organ positions among individuals in a population and asked if estimates of reciprocity can predict reproductive success in naturally occurring heterostylous populations. Methods: We developed a reciprocity index that assumed pollen transfer success equalled one for a perfectly matched stigma-anther pair, and decreased to zero with increasing mismatch. Reciprocity was quantified as the average pollen transfer success for all pair-wise combinations of complementary sex organs in the population. We examined the relationship between intra-population variation and reciprocity using simulated populations that varied in the distribution of sex organ positions, and with empirical data from natural populations. We compared previously proposed indices using the simulated and natural populations, and for a subset of natural populations we tested the ability of the indices to predict reproductive success. Important Findings: In both simulated and natural populations we observed that when differences between mean anther and stigma heights of complementary morphs are small, increasing intra-population variation in heights resulted in a monotonous decrease in reciprocity. However, when differences between mean complementary anther and stigma heights are larger, reciprocity increased, reached a peak, and then decreased with increasing variation. Previous indices failed to capture this behaviour and were largely insensitive to variation or differences in mean complementary sex organ heights. Seed set was consistently positively related to reciprocity for our index, and for two of the four previous indices. These results highlight the importance of incorporating intra-population variation in sex organ dimensions in quantifying reciprocity, and challenge the current understanding that increasing variation will always decrease reciprocity in heterostylous populations. These results may help explain why heterostylous systems exhibit, and tolerate high amounts of intra-population variation in sex organ heights.