In the neonate, Ig V-D-J junctions often occur at regions of short sequence homology, resulting in one to two predominant junctional sequences for most V-D and D-J recombinations. We have proposed that this mechanism of homology-directed recombination may play a role in the non-random usage of VH genes observed in fetal and neonatal life, since use of the short homologies at V-D junctions would preferentially make productive rearrangements for the overutilized 7183 and Q52 VH genes, and would make predominantly non-productive rearrangements for the underutilized VHJ558 gene family. Here we test this hypothesis for the 81X gene from the VH7183 family. Since pre-B cells which have rearranged the 81X gene do not appear to undergo the normal clonal proliferation before light chain rearrangement, analysis of the percentage of productive versus non-productive rearrangements for this VH gene is not skewed by the expansion of pre-B cells with productively rearranged IgH alleles. If V-D-J rearrangements were random, one would predict that only one-third of the rearrangements would be in-frame. This is close to what we observed for the 81X gene in adult bone marrow. In contrast, we show that 62% of all 81X rearrangements in fetal/newborn pre-B cells were productive. Forty-one percent of all the neonatal pre-B sequences containing DFL16 or DSP2 used homology-directed recombination to create the predominantly observed V-D junctional sequences, and 93% of those sequences were productive. This is consistent with our hypothesis that the mechanism of homology-directed recombination would result in an increased proportion of productive 81X rearrangements in the newborn. Therefore, we suggest that in fetal and neonatal life, when N regions are lacking, VH7183 and VHQ52 genes are more likely to undergo productive rearrangements than other VH families and thus are much more likely to contribute to the early B cell repertoire.