Geographical Information Systems (GIS) are becoming widespread in arch
aeological investigations ranging from inventory to management and inc
luding some spatial analysis. Common among the analytical tools is the
determination of visible area, an attribute widely suggested as impor
tant in archaeological contexts although not well investigated in the
pre-GIS era. However, the existing analyses generally lack rigorous st
atistical analysis. This is partly due to the lack of spatio-statistic
al tools within GIS, and a lack of generally accepted methods among ge
ographers and spatial scientists. This payer demonstrates how clear an
d testable hypotheses can be developed within a GIS context, addressin
g questions of the archaeological significance of the visible area. Th
e hypotheses are tested using random site generation in a Monte Carlo
framework. The areas Visible from the Bronze Age cairns of North Mull
are the primary subject of analysis. It is assumed that the sites are
broadly contemporary and that there may be some similar causal influen
ce of the visible area in their location. They are compared with areas
visible from random, stratified random and proximal sites within the
general landscape. The research demonstrates that the areas visible fr
om the Bronze Age cairns are distinctive among the population of the p
ossible sites in the area, being both larger and having a greater area
of the sea visible. When the distribution of the cairns close to the
short is accommodated within the analysis by stratified sampling the s
ites are still distinctive. Furthermore, the sites seem to share a vie
w of the sea to the north of the Island. Some practical or symbolic im
portance of sea in the lives of the Bronze Age inhabitants is clearly
indicated. (C) 1997 Academic Press Limited.