Abstract. Long-distance colonization and rapid range expansion associated with biological invasion may have major evolutionary consequences via both stochastic processes and selection. Using large-scale population genetic surveys, we demonstrate a major shift in the relative frequency of sexually fertile diploid versus sexually sterile triploid populations associated with the invasion of North America by a clonal aquatic plant, Butomus umbellatus. Most populations across the native European range were triploid (84% of 108), whereas most introduced populations were diploid (71% of 136). We evaluated the roles of stochastic processes versus natural selection in causing this shift by surveying predominantly neutral genetic variation at 28 RAPD loci. In Europe (EU) we detected 47 distinct genotypes among 142 plants sampled from 71 populations, whereas in North America (NA) we detected only six genotypes among 138 plants from 69 populations. Of the six NA genotypes, a set of four closely related genotypes were found only in triploid populations and a pair of closely related genotypes were found only in diploid populations, and these were genetically divergent from the triploid genotypes. This result is consistent with severe founder effect. Because sex creates genotypic variation and produces offspring with greater dispersal potential than those produced clonally, we tested the hypothesis that sexual reproduction characteristic of diploids has given them a colonization advantage that accounts for their high frequency in NA. However, we found little or no evidence of sexual recruitment in introduced diploids. One very widespread heterozygous genotype occurred in 95% of 38 introduced diploid populations (i.e., 72 of 76 plants surveyed) suggesting predominant clonal reproduction. Moreover genotypic diversity was not higher within or among diploid than triploid populations in either the native or introduced range. Low genetic diversity in diploid populations was also supported by a comparison of within-population quantitative variation for plant size under a common greenhouse environment. Thus, diploids have not been favored during colonization owing to their sexual fertility. However, concurrent studies have shown that NA diploids exhibit a much higher capacity for clonal reproduction, via small vegetative bulbils, than NA triploids, which almost never produce bulbils. The same difference in clonal capacity is not a consistent feature of the native EU populations. Taken together, these results suggest that strong founder effect has set the stage for a major increase in diploid frequency due to the particular, and possibly idiosyncratic, features of the diploid and triploid lineages introduced to North America.