Sharifu K. Tusuubira John Kelly ^*

• University of Kansas, Lawrence, Kansas, United States

Ecological and evolutionary changes are likely to occur rapidly when outcrossing populations experience pollinator loss. However, the number and identify of plant traits that will respond to this form of selection, as well as the overall predictability of evolutionary responses, remain unclear. We experimentally evolved 20 large replicate populations of Mimulus guttatus for 10 generations under three treatments: pure outcrossing, mixed mating (10% outcrossing) and pure selfing. These populations were founded from the same genetically diverse and outcrossing natural population. After 10 generations, all measured traits evolved with flower size, phenology, and reproductive traits diverging consistently among mating system treatments. Autogamy increased dramatically in the selfing treatment, but the magnitude of adaptation only becomes clear once inbreeding depression is factored out. Selfing treatment plants evolved reduced stigma-anther separation, and also exhibited declines in flower size and per-flower reproductive capacity. Flower size also declined in selfing populations but this was driven mainly by inbreeding depression and cannot be attributed to adaptation towards the selfing syndrome. Generally, the mixed mating populations evolved trait values intermediate to the fully selfing and outcrossing populations. Overall, our experimental treatments reiterated differences that have been documented in interspecific comparisons between selfing and outcrossing species pairs. Given that such contrasts involve species separated by thousands or even millions of generations, it is noteworthy that large evolutionary responses were obtained from genetic variation segregating within a single natural population.