The relationship between the biological and immunological activities of human luteinizing hormone (hLH) in plasma collected from female subjects was examined. The biological activity was measured by an in vitro bioassay and the immunological activity by an hLH radioimmunoassay (RIA), using improved reagents, such as the 1st IRP for human pituitary LH for immunoassay (code No. 68/40) as standard, a subunit-free biologically active iodinated hLH preparation as tracer and an anti-hLH serum of relatively high specificity. Similar profiles of biological (B) and immunological (I) activity were obtained in the plasma samples collected daily throughout 40 menstrual cycles (5 cycles from each of 8 subjects). However, the B/I ratios were significantly lower during the period of LH surge (P less than 0.001) than throughout the remainder of the cycle. The within- and between-assay variation in B/I ratios was investigated by the simultaneous assay of biological and immunological activities in plasma pools obtained by combining equal aliquots of plasma from each daily sample of the menstrual cycle from each of 5 cycles of each of 4 subjects. The analysis of these 20 pools revealed highly significant individual differences in B/I ratios, ranging from 0.81 to 1.33. The coefficient of variation was 20% between-subjects and 5% within-subjects. There was no seasonal variation in B/I ratios. That the individual differences in plasma B/I ratios were not attributable to the procedure of pooling was ascertained by the simultaneous assay of both activities in parellel in daily plasma samples and in the pools formed from these samples from three complete cycles. Thus the analysis of the differences in B/I ratios obtained throughout the menstrua- cycle revealed three major sources of variation. The first occurs in the form of generally elevated (higher than unity) B/I ratios, the second consists of a significant drop in B/I ratios during the midcycle LH surge, and the third source is represented by the significant between-subject differences. It is concluded that the first source is attributable to the relatively higher levels of "impurity" (i.e. biologically inactive, immunologically active material) in the standard preparation compared to those present in plasma of biologically inactive, immunologically active material of unknown composition and origin. If so, the latter source limits the quantitative significance of the RIA procedures employed. It is suggested that these three sources of variation account for most of the differences in B/I ratios for plasma lLH reported in the literature.