Genetic erosion has been evaluated at the landrace level in the past, principally because the loss of landraces is believed to generate erosion at the allelic level; however, few studies had tested this hypothesis in the crop’s centers of diversity and domestication. Using microsatellite markers, we analyzed for genetic erosion in lima bean (Phaseolus lunatus) landraces over time in samples collected in 1979 and in 2007 in northeast Campeche, in the Yucatan peninsula, Mexico, an important diversity center and part of the putative domestication area for this crop. We found that the lima bean genetic pool from 1979 had a higher genetic diversity than the one for the 2007 pool (Nei’s diversity, H=0.18 and 0.05, respectively). Although this result could not to be explained using a bottleneck analysis, a cluster analysis showed that the alleles present in 1979 were not the same as those found in 2007, indicating an allelic displacement in the genetic pool of the lima bean landraces in the last 30 years. This displacement could be due to the introduction of improved varieties or landraces, resulting in a displacement of local varieties or to changes in the Mayan criteria for selection of germplasm or both. This study showed that the loss of landraces can generate both quantitative and qualitative changes in the genetic pool of the domesticated species. Such changes are very important to consider when planning ex situandin situ programs to conserve crop diversity in their domestication areas.