Polygraphic tracings of 13 normal infants were recorded in a morning sleep at 1 and 2 weeks of age and 1, 2, 3, 4, and 6 months of age. A vibrotactile stimulus graded at 25, 50 and 100 Hz (frequency) and amplitudes of 1, 2 and 3 mm (intensity) was used, each combination being applied twice at 30 s intervals to the hand of the sleeping infant during active sleep (AS) and quiet sleep (QS). The results were analysed as percentages of failure to arouse (FTA) in relation to the number of stimulus trials, the criteria for FTA being the absence of a response in heart or respiratory rate, electroencephalogram, or chin electromyogram. The percentages of FTA from QS did not change significantly from 1 week to 6 months of age, irrespective of frequency or intensity. The percentages of FTA from AS fell sharply and significantly from 1 week to 2 months of age (P less than 0.001). At 3 months of age there was a significant increase followed by a significant decrease at 4 months of age, both changes showing a significant difference at P less than 0.05. Apart from the first week of age, the numbers of FTA from QS were greater than from AS for all stimulus trials. It is concluded that there is an arousal deficit in QS from 1 week to 6 months of age and the temporary deficit in AS at 3 months of age could explain the peak incidence of SIDS at this time.