Several boll weevils, Anthonomus grandis Boheman, were captured in pheromone traps in 2004 near Tlahualilo, Durango, Mexico, an area where none had been reported for ≈10 yr. It is possible that they were from an endemic population normally too low in numbers to be detected but that increased in response to more favorable rainfall conditions in 2004. Alternatively, they may represent an influx of migrants or the immediate descendents of migrants. To identify the most likely origin of the boll weevils captured in this area, we characterized microsatellite variation of the Tlahualilo weevils and compared it with the variation from three other populations in northern Mexico and from one in southern Texas. Measures of gene flow and individual assignment tests suggest that the boll weevils captured near Tlahualilo were primarily from an endemic low-level population, but that this area also is receiving immigrants from a cotton growing region ≈200 km to the north, near Rosales, Chihuahua, which is currently under a boll weevil eradication program. Similarly, Rosales is receiving immigrants from Tlahualilo. This study shows that microsatellite markers and population assignment techniques will be practical tools for determining the most likely origins of boll weevils reintroduced to eradication zones in the United States and Mexico. Population assignment strategies based on genetic markers hold promise for replacing conventional, but spatially constrained, mark-recapture studies of insect dispersal. This relatively new and powerful analytical approach is widely used in conservation genetics and fisheries studies, but has been underused by entomologists.