The scatter of points on an archaeological distribution map is not an end in itself, although it does provide solid information. Are we ready to concede that scientific proof can be obtained from a graphic document ? Are we able to reconstitute the organisation of a given territory from an incomplete sample of the sites it once contained ? Examples are presented from an experiment conducted on data from Iron Age and Roman Berry. Together with statistics and mathematical calculations, spatial analysis and the graphical manipulation of data can provide all the elements needed for a scientific hypothesis to be demonstrated. To identify the outcomes of spatial analyses and to sustain them as proof of particular past circumstances requires both...