Hi, In the archive I found a function for testing contrasts after manova, based on solving 'LBM=K'. There are a few questions on my mind: 1. A contrast between levels A and B of a certain factor with levels ABC is given as (0 -1 0) because the parameters for the intercept are fixed to zero in R. A contrast between B and C would then be (0 1 -1) right? 2. How would I go about if I want to determine which variate is causing rejection of the null hypothesis? 3. Is it also possible to construct confidence intervals for these contrasts? Alternatively, Does anyone by chance know whether it is possible to do multiple comparisons on multivariate data with the multcomp package? There seems to be an indication that it is possible, but I haven't been able to get it going. Many thanks! Tsjerk