The number of flowers open simultaneously on a self-fertile plant may have profound consequences for mating patterns. Self-pollination is more likely on plants with large displays since increased floral display size promotes within-plant (geitonogamous) pollinator movements. If inbreeding depression is present, this would reduce reproductive success. Self-pollination may also decrease pollen dispersal to other plants, further reducing reproductive success. We manipulated monkeyflower (Mimulus ringens) floral displays in replicate (cloned) populations consisting of individuals with unique combinations of multilocus homozygous genotypes. These marker genets facilitate unambiguous measures of paternity and individual self-fertilization rates. Plants were trimmed to experience one of an exponential series of floral display treatments: 2, 4, 8 or 16 flowers. As predicted, the fraction of bumblebee moves between flowers on the same plant increased with display size, enhancing the opportunity for self-pollination on plants with large floral displays. Self-fertilization rates closely paralleled patterns of pollinator movement. Two-flowered displays had a mean self-fertilization rate of 0.28, compared to 0.45 for 16-flowered displays. Genets differed strongly in their rates of self-fertilization and these differences were consistent across floral display treatments. The effect of floral display on selfing rate influences reproductive success since progeny produced by selfing averaged 33% lower fitness than progeny produced through outcrossing. Individual fruits in all treatments had high levels of multiple paternity (mean of 4.44 sires per fruit, including selfs), apparently as a result of both multiple visits to flowers, and deposition of mixed loads during single pollinator probes.

Key words: inbreeding depression, mating system, paternity, pollination, selfing