Hi! I was wondering whether anybody could help me with canonical correlation on multiple sets (3 or 4) of variates, e.g. using one of the methods suggested by Kettenring (Biometrika 58(3):433, 1971), starting from covariance or correlation matrices. If someone could point me to an efficient way to achieve this or maybe share a method, I would be forever grateful. Some additional information on the data: Submatrices of covariance matrix S: S11, S12, S13, (S14, ) S22, S23, (S24, ) S33 (, S34, S44), with dim(S) = p + q + q or p + p + q + q I would like to retrieve the canonical correlates and the canonical variates for all sets. Thanks in advance, Tsjerk