IT is quite possible, though not proved, that pulsars are magnetic white dwarf stars1. On the other hand, none of the four known pulsars has been identified with a white dwarf—the star which seemed to have been identified with CP 1919 is not optically variable and, besides, is not a white dwarf. This leads to the view that if pulsars are indeed white dwarfs, they are cold (having a photospheric temperature T ≲ 104 °K) and therefore cannot be observed in the visible region. There are no other grounds for such a hypothesis, but so far no difficulties with it have arisen. Thus even with the energy losses due to radio emission and to the “stellar wind” which reach the value of P ∼ 1030 erg s−1, the oscillations are likely to become noticeably attenuated only after a period of τ ∼ 108 yr (ref. 2), and therefore may have been excited when the star was still in the “hot” state, for example, by nuclear burning on the stellar surface. Even in a relatively cold star, however, the energy P ∼ 1030 erg s−1 can, in principle, be evolved for sufficiently long periods of time from gravitational compression caused by slight variations in the equation of state. What we mean here is, for example, transition into the superconducting state, a possibility which we discuss in this letter. Furthermore, the shock waves and radio emissions originate in the atmosphere of the star, the parameters of which are rather sensitive to the temperature of the photosphere. Hence, it does not seem to be a priori impossible that powerful radio emission is true only for a cold oscillating dwarf. It is clear that all these problems should be investigated, but our purpose here is to draw attention to the fact that in the case of cold magnetic white dwarfs the superconductivity of a certain layer of the star may prove to be an essential factor.