Several indices of fetal heart rate variability, including short-term variability (STV) and long-term variability (LTV) were analyzed. All the STV indices studied were found to be almost identical statistically. The LTV indices of de Haan et al were the same in essence, while other LTV indices reflected only the fluctuation around the mean fetal heart rate. Analysis using a signal generator indicates that the value of the standard deviation of beat-to-beat difference (one STV index) of less than 7 msec, through the cardiotocomonitor is unreliable; if the standard deviation of beat-to-beat difference is to be measured with an accuracy of 1 msec, an accuracy of measurement of 0.11 msec is necessary even if the fetal heart beat data are sampled with a sampling interval of 1 msec.