Recent advances in the statistical analysis of microsatellite data permit calculation of sex-specific dispersal rates through sex- and age-specific comparisons of genetic variation. This approach, developed for the analysis of data derived from co-dominant autosomal markers, should be applicable to a sex-specific marker such as mitochondrial DNA. To test this premise, we amplified a 449 bp control region DNA sequence from the mitochondrial genome of the collared peccary (Pecari tajacu), and estimated intra-class correlations among herds sampled from three Texas populations. Analyses on data partitioned by breeding group showed a clear signal of male-biased dispersal; sex-specific fixation indices associated with genetic variation among social groups within populations yielded values for females (F(GP)=0.91), which were significantly larger than values for males (F(GP)=0.24; P=0.0015). The same general pattern emerged when the analyses were conducted on age classes (albeit nonsignificantly), as well as categories of individuals that were predicted a posteriori to be dispersers (adult males) and philopatric (adult females and all immatures). By extending a previously published methodology based on biparentally inherited markers to matrilineally inherited haploid data, we calculated sex-specific rates of contemporary dispersal among social groups within populations (m(male symbol)=0.37). These results support the idea that mitochondrial DNA haplotype frequency data can be used to estimate sex-specific instantaneous dispersal rates in a social species.