When subjected to alternating quiet nights (32 dB) and noise nights (equivalent levels of 47 dB) a group of 14 subjects showed an average increase in the fraction of deep sleep of about 2.5% resulting from the traffic noise. Another group of 12 subjects whose noise nights were at 60 dB had an average of 4.6% increase in deep sleep during these nights. The number of wakings also increased for both groups but, as was found before, this adapted rapidly with the number of nights. The average latency of sleep onset does not appear to be affected by the traffice noise but individual differences are great and may be of opposite sign. Latency of sleep onset and waking both show appreciable ‘‘laboratory effect’’ which takes longer to disappear than the one or two nights usually assumed.