The longitudinal ultrasonic attenuation and velocity were measured in trigl-cyine sulfate (TGS) single crystal over the temperature range of 30-100 degrees C. It was found that the peak of the relative ultrasonic attenuation near the second-order phase transition is lower in the heating process than that in the cooling process, and its height decreases with the increasing frequency. Hence, the peak should be associated with the process of changing temperature and the measuring frequency. Another ultrasonic attenuation peak, which corresponds to the P-D2 peak discovered by other scientists, was also found in the experiment at 5.56 MHz. However, in experiments at 13.3 MHz and 30.8 MHz, such a peak was not observed. Considering these phenomena, it is suggested that this peak does not occur above the Curie point and is closely related to the interaction between domains. (C) 2000 Elsevier Science Ltd.