OBJECTIVE To study whether modulation of the GABAergic system (with sodium valproate) affects pulsatile LH secretion in the late follicular phase of normal women. METHODS Fifteen normal women volunteers were studied over an 8-hour period in the late follicular phase of two successive menstrual cycles. On each occasion, blood samples were taken every 10 minutes between 1000 and 1800 h. Nine of the volunteers--the short treatment group--were administered 400 mg of sodium valproate every 8 hours on the two days preceding their second session, and a further 400 mg at 0900 h on the day of the session. The other six--the long treatment group--were administered 400 mg of sodium valproate every 8 hours on the seven days preceding their second session and at 0900 and 1400 h on the day of the session. METHODS LH, oestradiol and progesterone were determined by radioimmunoassay, and sodium valproate by repolarization fluorescence spectrophotometry. Pulse detection was carried out both by the program ULTRA and by a method developed by the authors. RESULTS There were no significant differences in LH pulse amplitude or relative pulse amplitude between records taken in the first and second menstrual cycles, i.e. without or with prior sodium valproate treatment. Short treatment did change interpulse interval and mean secretion period, but the changes, though statistically significant, were small (about 10 minutes), so that the values for both post-treatment and control sessions were within the normal range; these parameters were unaffected by long treatment. CONCLUSIONS Activation of the GABAergic system with sodium valproate had no biologically significant effect on the late follicular phase pulsatile LH secretion of these normal women.