The influence of stimulus form on the scalp-recorded "vertex positive peak" (VPP) evoked by images of faces (Jeffreys 1989a) was studied in seven subjects. In separate experiments, we recorded the responses to 2D images of: (1) many different depictions of human faces; (2) the heads of several different species; (3) many familiar non-face objects; and (4) stimuli where the configuration of objects were modified to produce an "illusory" or "non-contextual" subjective impression of a face. The results showed that every facial representation, including the "illusory" stimuli, and most of the non-face objects, evoked a VPP of corresponding form and scalp distribution. The object-evoked VPPs, however, were always smaller and usually later than those evoked by the faces. VPPs of longer latency but often comparable amplitude were also recorded for impoverished compared to well-defined facial representations; and for most non-human compared to human faces. Very consistent responses were recorded to repeated presentations of the same stimulus for the same subject, but there was considerable variation in latency as well as amplitude (but not form) of the VPP evoked under identical experimental conditions for different subjects. These response properties of the VPP, suggest that its underlying physiological generators are sensitive to basic configural properties of the visual stimulus; and also the face- and object-related information are processed in the same brain area(s), although not necessarily by the same physiological mechanisms.