Postfertilization causes of differential success of pollen donors in Erythronium grandiflorum (Liliaceae): nonrandom ovule abortion

Seed paternity in Erythronium grandiflorum does not fully reflect the proportion of pollen on the stigma. When two types of pollen are simultaneously applied to the stigma, outbred seeds are produced over inbred, and seeds from more distant donors are produced over seeds from donors nearby. I looked for postfertilization causes of these previously reported patterns of differential success of pollen donors. I simultaneously pollinated stigmas with pollen from two donors and observed ovule development through a window sliced in the ovary. Pollen donor pairs were self and cross, donors 1 and 100 m from the recipient, and two donors each 100 m from the recipient. Since one donor was always the alternate homozygote from the recipient at the malate dehydrogenase locus, I could determine the paternity of developing seeds. When it appeared that ovules were aborting, I removed them and determined their paternity using starch gel electrophoresis. Ovules fertilized by self pollen were more likely to abort than ovules fertilized by cross pollen, and ovules fertilized by nearby donors were more likely to abort than ovules fertilized by distant donors. Ovules fertilized by donors 100 m from the recipient were equally likely to abort. There was not a significant relationship between the proportion of ovules fertilized by a pollen donor and the probability of those fertilized ovules developing into seeds. There was no relationship between ovule position within a fruit and ovule abortion. I manipulated available resources by removing leaves and by permitting only one fruit to develop per plant. Decreasing the amount of resources increased the proportion of aborted ovules. Abortion of ovules of lesser quality appears to release resources that can then be used to develop other offspring.