The data presented show early results from a longitudinal survey of "therapeutic" ultrasonic instrument performance in clinical physiotherapy. Previous studies undertaken in North America and the United Kingdom have shown that these units tend not to be scientifically tested and recalibrated, and have been found to have measured outputs at variance with the metered values if tests have been undertaken. This study measured effective transducer radiating area and, concurrently, the total power output at the metered intensity used for the area estimations. Actual output, in terms of total power and space averaged intensity was then compared with expected values derived both from manufacturers' data and from measured transducer radiating area. Only four of eighteen transducers (22%) were within the specified tolerance when actual transducer area was taken into account. Range was 179-31% of expected values. False assumptions of acceptable and unacceptable output can arise from failure to measure actual transducer radiating area.