Multivariate statistical techniques, such as principal component analy
sis and two-dimensional scatter diagrams, are increasingly being appli
ed to multispectral images in analytical electron and X-ray microscopy
. Combining data from different sources can often reveal more informat
ion than would be obtained if each image was processed separately. Thr
ee-dimensional data manipulation and visualisation tools are gaining p
opularity in many scientific fields. However, three-dimensional scatte
r diagrams are not yet being exploited much in analytical microscopy.
This paper demonstrates the advantages of three-dimensional scatter di
agrams and interactive correlation partitioning. An introduction to th
e methodology is included. Examples using multi-element Anger and topo
graphic images from the York multispectral analytical electron microsc
ope (MULSAM) are presented. Implementation issues are discussed briefl
y to help other workers who may wish to investigate the techniques. So
me areas of future work are suggested.