Forty pregnant women participated in a study to compare subjective with ultrasound assessments of fetal movements. A real-time ultrasound scanner was used. Movements were recorded for 45 minutes in all cases. There was a significant positive correlation between the number of movements recorded by the two methods, but the 95% confidence limits were wide and no correlation was found in those patients who recorded fewer than 20 movements in the study period. Thus "false-positive" information may be obtained from purely subjective data, and in patients reporting low "kick counts" fetal activity should be assessed from real-time ultrasound recordings.