Recombination rates vary both across the genome and between different species, but little information is available about the temporal and physical scales over which such rates change. To shed light on these questions, we performed a high-resolution analysis of a genomic region within the beta-globin gene cluster that is known to experience elevated recombination rates in humans. For this purpose, we developed new linkage disequilibrium-based methods that thoroughly search for subsets of the data with unusually high or unusually low estimated values of the population-recombination parameter (4Nr, where N is the effective population size and r is the crossover rate between adjacent base pairs). By resequencing a 15-kb segment in a human population sample, we were able to narrow the recombinational hotspot to a segment <2 kb in length that coincides with the beta-globin replication origin. In addition, we analyzed the orthologous region in samples of rhesus macaques and common chimpanzees. Whereas the analysis of the chimpanzee data is complicated by the sample structure, the macaque data imply that this region may not be a hotspot in that species. These results suggest a time scale for the evolution of hotspots in primates. Furthermore, they allow us to propose diverged sequence elements that may contribute to the differences in the recombinational landscape in the two species.