In this paper I have set forth in detail the theory of thermal waves in inhomogeneous media; it has, I believe, independent theoretical interest. I wish also to point out the fact that the mechanism of separation of an envelope in a nova explosion has hitherto remained obscure, since it strongly depends on the nature of the energy source of the explosion. Thus, if this energy is liberated by thermonuclear reactions, it is more probable that the time of development of the phenomenon reaches hundreds or even thousands of seconds. In such a case, the ejection of an envelope of the star is the result of the total effect of an infinite series of acoustic and weak shock waves that, added together, give a powerful pressure wave [6]. But if the energy of the explosion is gravitational in nature, its liberation may be virtually instantaneous, and the mechanism that transports the energy to infinity could be either a shock wave or a thermal wave. Moreover, if the explosion is due to a rearrangement of only the outer shell of the star, a thermal wave is more probable. And although the velocities of the thermal waves themselves are high, the rate of expansion of the matter of the shell will be appreciably less, since the time (of the order of a few seconds) is too short for interaction between radiation and matter. This distribution of velocities of the matter behind the thermal front can be obtained by a numerical solution of the equations of gas dynamics with allowance for the effects of radiative thermal conductivity; I hope to find such a solution in the future.