9807337 Marshall and Evans Understanding plant reproduction is important for both basic and applied science. We cannot understand plant ecology and evolution without knowledge of the causes of variation in plant reproductive success. From an applied perspective, plant reproduction is important because many of our agricultural products are the direct products of reproduction, so understanding of reproductive success is necessary to plant breeding programs. However, knowledge of plant reproduction has been limited because most of our information is about success of plants as seed parents, even though plants reproduce both as seed and pollen parents. If selection is occurring on the ability of plants to achieve reproductive success through pollen donation, then we may have misinterpreted or under-utilized some of the features of plants. This study focuses on testing whether selection is occurring on plants that act as pollen parents. The investigators are using an artificial selection experiment to address whether pollen-donating ability is genetically based and to test whether selection on this ability alters other characteristics of plants. Results to date show that mating characters are heritable and that selection on mating ability affects other features of the plant. The current work will replicate the experiment and learn more about how plants respond to selection. The investigators will use wild radish as a model organism because it is very easy to work with. In addition, it is closely related to other species that are grown as crops. The study will directly address a fundamental issue in plant evolutionary ecology and may have rather direct agricultural implications as well. Because the research will involve selection of seeds from different regions of the fruits, it may be possible to discern whether this form of selection could be useful in plant breeding. A study of alfalfa has already shown that seeds selected from different regions of fruits have different growth and reproductive characteristics. The current work can be used to address whether the most vigorous seeds typically come from particular regions of fruits in an unrelated plant and thus whether this might be a generally useful method of selection in plant breeding programs.