In her contribution to the international seminar on "Natural Family Planning" organized in October by the Irish Department of Health in cooperation with the World Health Organization, Professor Judith Bardwick, professor of psychology at the University of Michigan, offered an explanation for the high dropout rate from contraceptive programs based on periodic abstinence and for the difference between method failure rates of as little as 1% and user failure rates as high as 30% or more. She argued that contraceptive methods which ask women to maintain a continuous observation of the symptoms of their menstrual cycles will draw the attention of the women continuously to the act of coitus. Yet, simultaneously, the use of these symptoms as a contraceptive means requires abstinence from coitus. Even Bardwick was unable to make very definite statements regarding contraception based on periodic abstinence because of the lack of available data. A primary benefit of the Irish seminar was the presentation of the preliminary results from a 5-center international study of the Billings Method (also known as the Ovulation Method or the Cervical Mucus Method). The study, supervised by the World Health Organization in Geneva, is being conducted in Auckland, Bangalore, Dublin, Manila, and San Miguel and involves 870 couples. Over 80% of the couples are Roman Catholic. Professor H.G. Burger reported findings from the first 3 months of the study. The method (which uses observation of both the quantity and quality of vaginal mucus to determine the time of ovulation) required an average of 15-18 days abstinence in each cycle with 11-15 days available for intercourse. 90.8% of the women were thought to have acquired a good understanding of the method after the 1st cycle and 97.1% after the 3rd cycle. The couples' motivation for using the method was "religious" in 40% of the cases and "unhappiness with other methods" in about 25%. During the 2704 cycles under observation, there were 45 pregnancies in 5.2% of the women. 2 of these were regarded as user failures. Method related failures ranged from 0.48 to 3.00 and user-related failures from 0.96 to 19.25.