To whom it may concerned, We have a split-plot design for our experiment. We have 3 groups of pigs: piglet, 1 year old pig and 2year old pig. Each group has 5 pigs, for each pig, we take sample from brachial artery and femoral artery, i.e. two chips for each pig. So altogether we have 3*5*2 = 30 arrays. We are mostly interested in the difference between brachial and femoral in each age group and also the difference between groups for each artery type. Now our confusion is we don't know how to model the covariance between brachial and femoral within each pig in lmfit. Is there any way to model that? Or we need to use other functions like lme() to model the mixed effect? We really need this to be done as soon as possible, please give us some hint in the questions. Thank you in advance. Juan [[alternative HTML version deleted]]