In two experimental series (19 resp. 53 nights, 2 different persons in each series, test-time 10.30 p.m. to 3.00 a.m.) pulse rate after sonic booms had been recorded during sleep. In the first 3 nights the subjects slept undisturbed by noise. In the following 11 resp. 30 nights sonic booms were applied alternately 2 or 4 times. In the main series after 10 more nights without any noise 4 nights with 8 and 16 sonic booms alternately followed. The last 6 undisturbed nights in both series were used as comparison phase. The interval between two sonic booms was 40 min in nights with 2 booms, 20 min in nights with 4 sonic booms and in the nights with 8 and 16 sonic booms 8.6 resp. 4.6 min. Sound level of the sonic booms ranged from 0.48 mbar to 1.45 mbar, 1 mbar [83.5 dB (A)] in the average. The first sonic boom was applied if one of the two subjects had entered the deepest stage of sleep. Sonic booms induced a biphasic reaction in pulse rate. After an initial increase in frequency with a maximum in the 4th sec pulse rate decreased below the value before sonic boom; it was followed by a slow increase towards the baseline value. This reaction was analysed with special regard to the following factors: 1. Intensity. Due to very fast increase of noise intensity there was no significant correlation between the intensity of sonic boom and the pulse reaction. 2. Exogenic variables. There is no significant connection between postboom pulse rate and noiseless time before the sonic boom, the duration of the test series and the ambient temperature. 3. Endogenic variables. No correlation could be found between the stage of sleep and the reaction. On the contrary a very significant correlation was found between the maximum of postboom increase of pulse rate and the pulse rate before boom. With increasing pulse rate the extent of reaction becomes smaller.