The ratio (B/I) of bioactive to immunoreactive LH in plasma varies during pubertal maturation. To elucidate the basis for these changes, we compared the dose-response characteristics of LH standards to those of plasma LH before and after GnRH infusion in normal males at various pubertal stages and girls with Turner's syndrome. We used a human LH (hLH) RIA modified to optimize specificity for LH bioactivity by employing a hLH tracer maximally bioactive in the rat interstitial cell testosterone production (RICT) bioassay. The pituitary hLH standards LER-907, NHPP I-1, and NHPP I-2 were not parallel to one another in the RIA, but were parallel in the RICT. Their relative slopes in the RIA were 1:1.35:1.49, respectively. The B/I for immunoreactive LH in these standards were 1 (LER-907):3 (I-1):5 (I-2). Dose-response characteristics varied greatly by patient category in the RIA. In contrast to the RICT, in which plasma from all subject groups, except the prepubertal basal group, gave parallel dose-response slopes, the groups differed in the steepness of their plasma LH RIA dose-response curves in the following order: adult post-GnRH congruent to adult basal congruent to pubertal post-GnRH greater than pubertal basal congruent to prepubertal post-GnRH congruent to prepubertal basal greater than Turner's syndrome. Prepubertal basal samples most closely resembled LER-907 in dose-response characteristics, while adult samples were most similar to NHPP I-1 and I-2. These characteristics were not related to the absolute LH concentration. Although the RIA dose-response characteristics of plasma LH changed during puberty, the B/I of basal LH did not increase through puberty using the modified hLH RIA, although B/I still rose in GnRH-stimulated samples during puberty. Our data demonstrate that the principal cause of variability in B/I is the changing dose-response characteristics of plasma LH in the RIA. We suggest that the source of this variability is a change in the molecular characteristics of LH in different states of pubertal maturation and gonadal function. The results imply that better ways of assaying LH are needed.