Evolution of floral and mating system variation in Camissoniopsis cheiranthifolia (Onagraceae): An evaluation of patterns and processes

Thumbnail Image
Dart, Sara Rachel
Camissoniopsis cheiranthifolia , Mating system , Allozymes , Inbreeding depression , Pollen limitation , Geographic variation , Bud parasitism , Reproductive assurance , Floral variation
Understanding how floral traits covary with one another and with mating patterns is an important step in understanding how and why mating systems evolve. I examined the evolution of floral and mating system variation in Camissoniopsis cheiranthifolia (Onagraceae), a species that exhibits divergence in key floral traits expected to be associated with variation in the relative importance of outcrossing vs. self-fertilization. I combined geographic surveys of floral variation with genetic estimates of the proportion of seeds outcrossed (t) and confirmed that t covaried with corolla width and herkogamy in a predictable way both within and among populations. I then performed geographic surveys, manipulative experiments and genetic analyses to evaluate the potential role that; inbreeding depression (ID), interactions between flowers, pollinators and florivores, and reproductive assurance (RA) may have played in shaping and/or maintaining the geographic pattern of mating system variation in this species. The main selective factor maintaining outcrossing in large flowered (LF) populations appears to be ID, which was much stronger in LF compared to small flowered (SF) populations. These results are also consistent with purging of ID in SF populations. Increased selfing appeared to alleviate pollen limitation (PL) because it was associated with higher and less variable fruit set and reduced florivory by a microlepidopteran. However, evidence that florivores preferentially attacked larger flowers was equivocal. LF experienced stronger PL than SF populations suggesting that one condition for the evolution of selfing via RA is met in outcrossing populations. Floral emasculation experiments revealed that the timing of selfing also covaried with flower size among and within populations. SF self-pollinate before flowers open but LF do not, suggesting that selfing evolved in response to chronic outcross PL. Negative side effects of emasculation were detected which prevented a clear interpretation of the RA value of selfing. Given that much of what is known about RA comes from emasculation experiments, my results suggest that the assumptions of this approach, which are rarely verified, require more serious consideration. Taken together my results suggest that C. cheiranthifolia has evolved multiple stable mixed mating systems perhaps in response to selection for RA.
External DOI