The behaviour of 95 clinically normal 6-month-old babies previously studied as neonates and as 6-week-old infants was video-recorded while they were presented with sound stimuli. The stimulus series varied in sound pressure level (SPL) (30 and 60 dB), band with (BW) and rise time (RT), and included two voice signals and a no-sound (control) trial. Video-records were made both for the whole body aspect and for a 3 1/2 X 'close-up' of the head. Later, the video-records were shown to 6 observers who were allowed to see the babies for 13 s at each trial. The first 5 s was a pre-stimulus observation period, and the second 8 s usually contained a sound stimulus. Between the trials, the observers were given 20 s in which to record pre-stimulus activity, confidence in response, and movement details. Data for movement details are not reported here. Different segments of the baby's body were observed on different occasions as detailed in our previous reports. The confidence ratings for response to tonal stimuli were analyzed using aspects of signal detection theory for differences amongst various body segments (n.s.), SPL (p less than 0.001), BW (p less than 0.05), and RT (p less than 0.01). A 60-dB broad-spectrum noise band (BSN) and the 2 voices were by far the most effective stimuli. They elicited detectablities which differed between segments. The response to different sound stimuli was affected by the baby's pre-stimulus activity state. The results are discussed in relation to our previous studies.