GENETIC DRIFT AND FOUNDER EFFECT IN NATIVE VERSUS INTRODUCED POPULATIONS OF AN INVADING PLANT, LYTHRUM SALICARIA (LYTHRACEAE).

There are few convincing examples of genetic drift at loci under selection in natural populations. The plant sexual polymorphism tristyly provides an opportunity to investigate genetic drift because stochastic processes interacting with frequency-dependent selection give rise to a diagnostic pattern of morph-frequency variation. A previous study of 102 Ontario populations of the introduced tristylous wetland herb Lythrum salicaria provided evidence for the role of stochastic processes during colonization. However, whether stochastic effects are greater in these recently introduced populations compared to native Eurasian populations remains unclear. The propensity of this species to invade disturbed habitats suggests that episodes of colonization and periods of small population size must also occur in the native range. A survey of 102 populations in southwestern France indicated reduced stochastic effects in native populations. Populations exhibited significantly lower morph loss than in Ontario (5% vs. 23%) and significantly higher values of morph evenness. The greater incidence of trimorphism in French populations was not associated with larger population sizes; populations were significantly smaller than those in Ontario (means: 266 vs. 487). Morph evenness was positively correlated with population size among French but not Ontario populations, providing further evidence of nonequilibrium conditions in introduced compared to native populations. The incidence of trimorphism was unexpectedly high in small native populations (N ≤ 25; 22 of 27 populations trimorphic). Computer simulations indicated that levels of gene flow on the order of m ≥ 0.05 can account for the maintenance of tristyly in small populations. The high connectivity of populations within the agricultural landscape typical of southwestern France may facilitate levels of gene flow sufficient to maintain trimorphism in small populations.