On the method of Yvon in crystal statistics