The levels of biologically active luteinizing hormone were determined by an in vitro bioassay method in plasma samples collected daily over a complete menstrual cycle from 12 menstruating women. These cycles were normal according to a number of criteria, including daily plasma levels of oestradiol, 17-hydroxyprogesterone and progesterone. Immunoreactive LH was estimated in the same 12 cycles by a radio-immunoassay (RIA) procedure (HCG-RIA) using an HCG antiserum and iodinated HCG. The 2nd IRP of HMG was selected as standard although significant deviations from parallelism were found with 7 out of the 12 plasma pools studied. The use of the 1st IRP of human pituitary gonadotrophins (FSH and LH (ICSH)) for bioassay (herafter HPG-1st IRP) as standard in this system resulted invariably in invalid assays, due to lack of parallelism. Immunoreactive LH was also measured in 8 of the 12 cycles by a RIA procedure (HLA-RIA) using a human LH antiserum and iodinated human LH of pituitary origin. Results are expressed in terms of the HPG-1st IRP. The plasma levels of biologically and immunologically active LH were qualitatively similar throughout the menstrual cycle. However, the LH levels measured by the bioassay invariably exceeded those estimated by the RIA procedures. The biological to immunological (B/I) ratio over the entire menstrual cycle (312 comparisons) was 5.5 with 95% confidence limits at 5.2 and 5.8 when the HCG-RIA system was employed. Using the HLH-RIA system (208 comparisons), the corresponding ratio was 6.4 (6.0:6.9). When regression lines were calculated using the bioassay results as the independent variable and the RIA results as the dependent variable, the 95% confidence limits of the regression lines did not include the origin. Furthermore,, in keeping with the high B/I ratios, the slopes of the two regression lines and their conficence limits differed markedly from unity. It is concluded that although qualitatively similar profiles were observed between the biological and immunological activities throughtout the menstrual cycle, two aspects require further attention. Firstly, the elevated B/I ratios together with the behaviour of the dose-effect lines obtained with different standards in the various RIA systems suggest that presently available reference standard preparations of pituitary and/or urinary origin are not suitable for the assay of LH in human plasms. Secondly, from the regression analyses of the biological and immunological activities it is infered that the RIA methods detect immunological activity which is not associated with biological activity. If so, the validity of these RIA procedures for specifically measuring low levels of biologically active LH in plasma may be in question.