Over recent years there have been many advances in our understanding of human face recognition. Despite this, almost all current research proceeds as though human perceivers are all equivalent. In fact, there is very large variability in people's ability to recognise faces, and in pilot work we have begun to quantify this. We will study the variation between people for two purposes. First, research on individual differences will provide a new way to examine some fundamental processes in face recognition, and particularly the dissociation between perception of familiar and unfamiliar faces. Second, variability in face processing is important for forensic purposes. Using collaborators in police and surveillance organisations, we will study recognition in realistic contexts. We aim to deliver techniques for discriminating good from poor "recognisers" and make recommendations for witness selection and officer training.