The possibility of relating the electric current in the ionosphere to the electric field in the frame of reference either of the neutral atmosphere (as commonly done) or of the plasma raises the question: which should be used in calculating Joule heating? The energy equations for the plasma and for the neutral medium, including collision effects, can be combined with momentum equations to separate energy transfer into work done and heating (this is not the same as separation of energy content into bulk-flow kinetic energy and thermal energy). Heating/dissipation processes can be unambiguously identified, showing that true electromagnetic (Joule) heating rate is given by J ¿ E in the frame of reference of the plasma. The major part of energy dissipation associated with J and E in the ionosphere is heating by collisions between plasma and neutrals, proportional to the square of the difference in their bulk flow velocities and distributed approximately equally between the two. The conventionally calculated ionospheric Joule heating as J ¿ E in the frame of reference of the neutral atmosphere equals the sum of total heating rate of plasma (only a small part of which is true Joule heating) plus total heating rate of neutrals, plus a term equal to work done on plasma in the neutral-atmosphere frame of reference; the latter is assumed negligible, with J ¿ B/c balanced entirely by plasma-neutral collisions and with no net work done on the plasma. This assumption implies that all the magnetic stresses exerted on the plasma at ionospheric heights are taken up by the local neutral medium; hence the dynamics of the neutral atmosphere play an important role in magnetosphere/ionosphere interactions.