Of 497 single neurones recorded in the cortex in the fundus of the superior temporal sulcus (STS) of three alert rhesus monkeys, a population of at least 48 cells which were selectively responsive to faces had the following response properties: (1) The cells' responses to faces (real or projected, human or rhesus monkey) were two to ten times as large as those to gratings, simple geometrical stimuli or complex 3-D objects. (2) Neuronal responses to faces were excitatory, sustained and were time-locked to the stimulus presentation with a latency of between 80 and 160 ms. (3) The cells were unresponsive to auditory or tactile stimuli and to the sight of arousing or aversive stimuli. (4) The magnitude of the responses of 28 cells tested was relatively constant despite transformations, such as rotation, so that the face was inverted or horizontal, and alterations of colour, size or distance. (5) Rotation to profile substantially reduced the responses of 21 cells (31 tested). (6) Masking out or presenting parts of the face (i.e. eyes, mouth or hair) in isolation revealed that different cells responded to different features or subsets of features. (7) For several cells, responses to the normal organisation of cut-out or line-drawn facial features were significantly larger than to jumbled controls. These findings indicate that explanations in terms of arousal, emotional or motor reactions, simple visual feature sensitivity or receptive fields are insufficient to account for the selective responses to faces and face features observed in this population of STS neurones. It appears that these neurones are part of a system specialised to code for faces or features present in faces, and it is suggested that damage to this system is related to prosopagnosia, or difficulty in face recognition, in man and to the tameness and social disturbances which follow temporal lobe damage and are part of the Klüver-Bucy syndrome in the monkey.