I am using the following code to get estimates of the variances of the levels of factor 1 (factor1). lme(sp1~factor1-1,data=z.a.1,random=pdBlocked(list(pdIdent(~factro1-1),pdIdent(~factor2-1))), weights=varIdent(~1|factor1)) In this study factor 1 is crossed with factor 2. is this correct? [[alternative HTML version deleted]]