Many thanks, professor Walker. I'm reading the papers that you suggest me. I think that the discrepance rise inside a variance components implemetation ... (it was my suspect) Thank You again Sorry for noise [[alternative HTML version deleted]]