In a randomized study, 20 patients received temazepam 20 mg orally the night before and 20 mg in the morning of an operation performed under spinal analgesia (Group I); 20 patients received flunitrazepam I mg similarly (Group 2). Different aspects of the premedication were evaluated verbally, with the aid of a visual analogue scale, Maddox wing apparatus, the critical flicker fusion threshold test, blood pressure and heart rate measurements, serum and CSF cortisol and plasma ADH measurements, as well as CSF drug level determinations. Clinically, temazepam 20 mg proved to be comparable with flunitrazepam I mg, although the latter more effectively prevented cardiovascular changes and pre-operative hormonal stress reaction. No correlation was found between the CSF drug level (bioassayed by radioreceptor assay) and the clinical response of the two benzodiazepines, nor was there any correlation between the cortisol or ADH levels versus the CSF drug levels. On the whole, flunitrazepam proved to be marginally better than temazepam as an oral premedicant.