P revious studies have shown that log IgD levels in normal individuals are distributed in a nonunimodal manner. Therefore, in this study we tested whether inheritance might play a role in determination of IgD levels. IgD levels were measured in serum or plasma from 301 randomly selected children aged 6-18 yr, 245 consecutive adult blood donors, and 134 first-degree relatives of subjects with low IgD levels. Comparison of serum and plasma from five individuals revealed no difference, so the two were used interchangeably. The distributions of log IgD levels in randomly selected populations of both adults and children were nonunimodal with nadirs at 2.15 IU/ml. In both of these randomly selected populations, 13-14% of the subjects had low IgD values (<2.15 IU/ml). In addition, there was a significant sibling-sibling correlation of log IgD values (r = 0.56, n = 72, P <0.01). Because of the nonunimodality of the frequency distribution histogram for IgD values and because of the familial aggregation of these values, the study was extended to include first-degree relatives of subjects with low plasma IgD. Blood samples from 92% of living first-degree relatives, 134 individuals, were analyzed for their level of IgD, and the results of segregation and pedigree analyses of these data were compatible with autosomal recessive inheritance of an allele for low plasma IgD levels. IgD values in plasma from siblings of probands for low IgD were also non-unimodal in distribution with a nadir at ≅2.15 IU/ml. The results suggest that there is autosomal recessive inheritance of an allele for low plasma IgD.