V, D, and J gene segments rearrange at different frequencies in vivo. Each rearranging gene segment is flanked by a recombination signal sequence (RSS), which is composed of a conserved heptamer and nonamer, separated by a spacer of conserved length but not conserved sequence. We summarize data from our lab and other labs showing that in many cases, but not all, the RSS can account for differences in recombination frequencies observed in vivo. Our approach is to determine the initial frequency of rearrangement of the V genes in vivo, and then place the RSSs of two V genes into a competition recombination substrate to determine the relative frequency with which the two RSSs support recombination. In one example, we have shown that a polymorphism in the heptamer of a Vkappa gene can result in a significant reduction in recombination frequency. This particular allele is prevalent in Navajos and absent in other populations. We suggest that this single change may play a major role in the high susceptibility of Navajos to Haemophilus influenzae infection, since this Vkappa gene is important in the antibody response to this bacteria. We also describe experiments showing that the sequence of the spacer of the RSS can play an important role in relative recombination frequencies.