Social psychologists have long been interested in the effects of ‘contact’ between racial groups. The conditions under which this contact can manifest have usually been experimentally manipulated in order to determine optimal combinations. A shortcoming of this approach is that it constructs contact situations that are unnatural and contrived. Some researchers have proposed an approach that examines contact as a natural phenomenon (Dixon & Durrheim, 2003). The present research adopts this approach, and reports on a naturalistic, observational study of ‘contact’ between students in university residence dining-halls. Seating patterns of students were observed for one month and analysed along dimensions of spatial variation. The results show high levels of informal segregation and that the segregation manifests as a specific spatial configuration. Such results, which occur despite the presence of apparently favourable conditions, illustrate how this approach may lead to different conclusions to those achieved through experimental manipulation.