A simple device for the experimental study of the heating capabilities of diagnostic ultrasound beams is described. Some results are reported that demonstrate the manner in which the device may be used to explore the heating potential of any particular commercial transducer, operating over the full range of output conditions. The heat generated in the base of a polyethylene container, filled with water, was measured using a fine-wire thermocouple, attached externally. The majority of measurements were carried out in beams generated by a curved array operating with a modern commercial scanner (Doppler, 2.5 MHz: imaging 3 MHz). A temperature rise in excess of 30 degrees C was generated by a pulsed Doppler beam, when the water path and scanner controls were set appropriately. Comparable temperatures were measured at comparable intensities generated by Doppler beams of other scanners. Of the imaging beams studied, the greatest temperature rise observed was less than 2 degrees C, when the highest frame rate and line density were selected. The greatest temperature rise in colour Doppler mode was 7.8 degrees C. It was observed that the position of the fixed (nonelectronic) focus was significant in controlling the heating profile with depth, for scanned beams. As expected, there was a strong dependence of temperature rise on axial time-average intensity. A weak dependence on -6 dB beam area was observed over a range of beam area of about 7 to 70 mm2. A strong dependence on finite amplitude effects was observed, resulting from energy loss associated with acoustic shock propagation.