Abstract
Past climate shifts have led to major oscillations in species distributions. Hence historical contingencies and selective processes occurring during such phases may be determinants for understanding the forces that have shaped extant phenotypes. In the plant-ant Petalomyrmex phylax (Formicinae), we observed spatial variation in number of queens in mature colonies, from several queens (high polygyny) in the median part of its distribution to a moderate number of queens (weak polygyny) or even only a single queen (monogyny) in the southwestern most populations. This variation did not correlate with indicators of variation in current nest site availability and colony turnover, the supposedly determinant selective forces acting on gyny in ants. We show here that the variation in social structure correlates with a historical process corresponding to a progressive colonization of coastal southern Cameroon by the ant. Using microsatellite markers, we observed a clear pattern of isolation by distance except for the southernmost populations. Measures of genetic variability that do not take into account allele size were at equilibrium in all except the southernmost populations, suggesting recent foundation of the latter. Measures of genetic diversity taking into account allele size showed a clinal north-south decrease in variance of allele size. We propose that southern populations have yet to regain allele size variance after bottlenecks associated with the foundation of new populations, and that this variance is regained over time. Hence variation in social structure mirrors an old but still active southward colonization process or metapopulation dynamics, possibly in association with an expansion of the rain forest habitat during the late Holocene. A low number of queens in ant colonies is typically associated with strong dispersal capacity. We therefore suggest that the initial founders of new populations belong to the monogynous to weakly polygynous phenotype, and that queen number progressively increases in older populations