The W locus, a codominant locus influencing floral pigment intensity in the tall morning glory, Ipomoea purpurea, is polymorphic throughout the southeastern United States. Previous studies suggest that this polymorphism is actively maintained by balancing selection, and that increased selfing accompanied by lack of pollen discounting (''Fisher effect'') may act to protect the white allele when it is rare. Processes that act to protect the dark allele and thus stabilize the polymorphism in conjunction with the Fisher effect have not been previously detected. The goal of this study was to determine whether any of three such processes might operate in I. purpurea. Estimates of breeding system parameters in a large experimental population in which the white allele was in higher than normal frequency (0.5) provided little evidence that either dark- or light-flowered plants were more successful as pollen parents than white-flowered plants. In addition, no evidence was found for a transmission bias favoring the dark allele in the ovules produced by light heterozygotes. In contrast, a strong transmission bias favoring the dark allele in pollen of heterozygotes was observed. A simple model using parameter estimates derived from this and previous studies indicates that a balance between the Fisher effect and biased transmission in heterozygote pollen could account for many properties of the polymorphism.
