The introduced subspecies of the common reed (Phragmites australis (Cav.) Trin. ex Steud. subsp. australis; Poaceae) is considered one of the most invasive plants in North American wetlands. Given its relatively low seed set and its tremendous capacity to spread via stolons or rhizomes, it has generally been thought that the spread of vegetative diaspores was responsible for the establishment of new populations. To test this hypothesis, we sampled a single plant from each of 345 visually-distinct common reed stands located along the shores of Lake St. François (southern Quebec, Canada). With a set of six nuclear microsatellite markers, we distinguished 134 different genotypes. The number of individuals sharing the same genotype ranged from one to 16, and averaged 2.1. Most genotypes were encountered only once. We examined the spatial distribution of the most frequent genotypes and found little evidence of clusters along the lakeshore. These data contradict the hypothesis that a common reed invasion is initiated by the introduction of vegetative diaspores from a few clones. Rather, they clearly support the alternative hypothesis that seeds were the primary diaspores responsible for the establishment of common reed populations.