A quantitative method for measuring pain threshold by the use of ultrasonic stimulation in man was designed and the possibility of clinical application in assessing analgesics was investigated. Ultrasonic stimulus was given to Japanese subjects on the palmer distal part of the 2nd, 3rd and 4th fingers of both hands. The latent time between start of the stimulation and withdrawal of the hand when perceiving pain was considered the pain threshold. The ultrasonic evoked pain was a sharp pin-prick type, without sensations such as thermal and mechanical. The pain threshold lowered with increasing either stimulus intensity or water bath temperature when the hand of the subject was immersed during measurement. Normal threshold to ultrasonic stimulation measured in both 50 men and 50 women gave nearly normal distribution curves: women being more sensitive to ultrasonics than men. Analgesia with codeine phosphate (20 mg p.o.), aspirin (1.5, 1.0, 0.5 g p.o.), aminopyrine (100 mg p.o.) and mefenamic acid (500 mg p.o.) in volunteers of both sexes was demonstrated significantly using this method under double blind circumstances. Pentobarbital, diazepam, butylscopolamine, bromelain and placebo each in the usual dose used clinically failed to alter the pain threshold. Humans were at least 25 fold more sensitive than mice to the analgesics used herein.