Hybridization between cultivated plants and their wild relatives has been documented for many species including sunflower, rice, sugar beet, and radish. The persistence of crop-derived traits has rarely been examined for more than two generations, and even more rarely under natural field conditions. The purpose of these studies was to determine the potential for persistence of traits from cultivated radish (Raphanus sativus) in populations of wild radish (R. raphanistrum). The first two chapters describe the results of studies that were conducted in four artificial populations created in northern Michigan; each population initially consisted of 100 F1 hybrids and 100 wild radishes planted in 1996. Three indicators of crop allele persistence were monitored over several generations: pollen viability, flower color frequency, and crop-specific allozyme allele frequency. Pollen viability is known to be low in F1 hybrids of R. raphanistrum x R. sativus due to heterozygosity for a reciprocal translocation that differs between the two parent species. The presence of plants with low viability indicates the presence of the two different chromosomal arrangements and, therefore, crop-derived genetic material. White flower color, a dominant trait, is characteristic of R. sativus; therefore, persistence of white-flowered plants in the population indicates the presence of crop alleles. Similarly, the PGM and GPI allozyme loci had alleles that could be used to determine the presence of R. sativus genetic material over generations. Pollen fertility was low (<70%) in F1 hybrids, but fertility increased significantly in subsequent years, which indicates that fertility barriers to hybridization are minimal. Increased pollen viability indicates that one of the chromosomal arrangements, probably the crop-derived form because it occurred at a lower frequency in the initial populations, was being purged from the population due to heterozygote disadvantage. White flower color frequency initially decreased, but remained relative stable over subsequent generations. Crop-specific allozyme allele frequency initially decreased, but increased in frequency or remained stable in subsequent years. Our results demonstrate that genetic differences that adversely affect fitness are selected against; however, introgression of mnay other genes from R. sativus to R. raphanistrum is likely. To further examine the relative fitness of hybrids compared to wild plants under field conditions, in 1999 I planted wild, F1, and BC1 radish seeds in a fallow field in Michigan under two growing conditions; with and without competition. Germination and mortality were significantly different among cross types (BC1>F1>wild for both measures). F1 plants bolted and flowered significantly later than wild or BC1 plants. F1 plants produced significantly fewer flowers and fruits than wild or BC1 plants, and all cross types differed significantly in seed production (wild>BC1>F1). Although measures of performance were generally lower with competition, this treatment significantly affected only seed production. There was no cross type x competition interaction. F1 fitness was markedly lower than the other two cross types; however, BC1s and wilds had similar fitness in the presence of competitors. These results indicate that even early generation hybrids are capable of ecologically significant levels of reproduction under natural conditions. This study provides further evidence that hybridization between crop and wild radish could lead to long-term introgression of crop alleles into wild radish populations.