Physical reasoning suggests that aerodynamic sound from free flows may be attributed to local fluid dilatation coupled with local fluid accelerations, the former acting like a simple source field and the latter like a dipole field which reduces to a quadrupole field. A supporting mathematical development is given; it discloses the formal need of an additional surface integral. The rather poor convergence of the integrals is discussed and comparison is made with theories which rest on the dilatational effect alone.