We closely follow the discussion by Luttinger et al. [Phys. Rev. 118 (1960), 1417] but with a difference in taking an unperturbed system at the start of calculation. Through the consideration of graphs we are urged to treat the so-called interaction constant with an unfamiliar manner.