In Part I of this paper, it was shown that the reciprocity theorem for passive linear electroacoustic transducers could be established on the basis of a pair of general linear relations connecting the voltage and current at the electrical terminals of the transducer with the pressure and velocity distribution over the transducer surface provided certain “reciprocity relations” existed among the parameters entering these equations. In Part II it is now shown that on considering a transducer to be comprised of media characterized in the usual manner by appropriate linear constitutive relations between stress, strain, electric and magnetic polarization, charge and current density, and electric and magnetic field intensity, one can establish the validity of the linear relations and the “reciprocity relations” assumed in Part I, provided that certain sufficient conditions are met. These conditions are: (l) That the coefficients in the constitutive relations satisfy certain “symmetry conditions.” (2) That no magnetostrictive media and no static magnetic field are present in the transducer (that is, that the coupling is purely electrostatic or piezoelectric or both), or that no piezoelectric media and no static charge density are present in the transducer (that is, that the coupling is purely electromagnetic or magnetostrictive or both). (3) That the transducer does not radiate electromagnetic waves from its surface.The validity of the “symmetry conditions” on the coefficients of the constitutive relations is established for frequencies sufficiently low that the harmonic changes in the variables can be regarded as adiabatic in the thermodynamic sense and under certain circumstances at higher frequencies by means of a general argument based on a principle of microscopic reversibility. One may, however, anticipate the possible breakdown of these conditions when appreciable losses due to “relaxation” phenomena in the transducer media are present. The conditions for the validity of the reciprocity theorem derived in this paper are only sufficient, but they may be considerably extended in their generality to include the presence of other phenomena (such as electrostrictive effects) by the application of the same general methods.