A rapid, simple and precise luteinizing-hormone (LH) assay would appear to be the most appropriate approach to predicting ovulation in in vitro fertilization and artificial insemination programs based on natural ovulation. Therefore, we modified a standard LH radioimmunoassay into a convenient assay providing reliable results within three hours after blood collection. The antigen-antibody reaction takes place during a two-hour incubation under nonequilibrium conditions at 20 C. The immune complex is precipitated with polyethyleneglycol containing sheep anti-rabbit antibodies. The standard curve of the rapid method shows characteristics (ED-50, slope, sensitivity, range, precision) similar to those of the regular assay, and the results of both assays are comparable over a wide range of values. The interassay variation is less than 15%. The preovulatory LH surge has been evaluated in 32 cycles (20 women) with successive dosages of daily blood samples taken every three to five hours. LH surges vary in shape and duration. In most of the cycles the LH surge is relatively steep, and the maximum LH value is observed within 17 hours after the beginning of the LH rise in 87% of the cycles. Detection of the initial rise in the LH peak on the basis of successive assays of frequent blood samples seems to be reliable in predicting ovulation.