. . "The last [p. 498] ground for inferring that A and B differ must be that, whilst A is an m, B is an n, and that m and n are seen to differ." . . .