How I test for differences among time points without assuming parallelism, With a profile analysis model, Using manova R function. For example, with Potthoff and Roy data (Potthoff and Roy 1964) how I can test H: ABC=0, when A<-diag(2) and C<-rbind(diag(3),rep(-1,3)) (Davis, 2005 p.p 81 ). I have the SAS code but I want to do it with R, of course. Thank for your help.