Abstract. Recent studies of the introduced fire ant Solenopsis invicta suggest that introduced polygyne (with multiple queens per nest) populations are strongly influenced by male-mediated gene flow from neighboring monogyne (single queen per nest) populations and selection acting on a single locus, general . This investigation formally tests this hypothesis and determines if these processes can account for the genotypic structure of polygyne S. invicta. To increase the statistical power of this test, we considered the genotypes of polygyne queens and workers at both Gp-9 and the closely linked, selectively neutral locus Pgm-3. We then constructed and analyzed a novel mathematical model to delimit the effects of monogyne male gene flow and selection on the joint genotypes at the Pgm-3/Gp-9 superlocus. Using this framework, a hierarchical maximum-likelihood method was developed to estimate the bestfitting gene flow and selection parameters based on the fit of our model to data from both the current study and an earlier one of the same population. In each case, selection on polygyne queens and workers alone, with no monogyne male gene flow, provides the most parsimonious explanation for the observed genotype frequencies. The apparent discrepancy between this result and the empirical evidence for monogyne male gene flow indicates that undocumented factors, such as other forms of selection in polygyne males or workers, are operating in introduced polygyne S. invicta.Key words. Fire ants, gene flow, isozymes, migration-selection balance, population structure, selection, Solenopsis invicta.Received March 2, 1999. Accepted September 1, 1999.Gene flow and selection are fundamental evolutionary processes that may act jointly to shape patterns of genetic variation within populations. For example, selection against deleterious alleles may reduce variation, whereas gene flow from nearby populations may act in opposition to such selection by reintroducing unfit alleles (May et al. 1975;Felsenstein 1976;Endler 1977;Barton and Hewitt 1985;Slatkin 1985;Barton and Clark 1990). Several studies have reported such a migration-selection balance within species (McNeilly 1968;Livingstone 1969;Jones et al. 1977;Turner 1977;Jones 1982;Riechert 1993;King and Lawson 1995;Chevillon et al. 1998). One well-studied case is found in introduced populations of the fire ant Solenopsis invicta in the United States, where it was thought that a recessive deleterious allele at the locus phosphoglucomutase-3 (Pgm-3) was maintained at high frequency in one social form by gene flow from neighboring populations of the alternate social form (Ross 1992;Ross and Shoemaker 1993;Ross and Keller 1995).The role of Pgm-3, however, was recently challenged by the discovery of the diallelic locus general protein-9 (Gp-9), which is tightly linked (r ϭ 0.0016) to Pgm-3 and apparently under strong selection in S. invicta. In monogyne (with a single queen per nest) populations, all females are genotype Gp-9 BB , and so the monogyne form is fixed for the Gp-9 B allele. ...