On a classical existence theorem for nonlinear elliptic equations