The behaviour of radon at low temperatures presents some features which have been hitherto difficult to explain. As is well known, this gas can be condensed at the temperature of liquid air, the condensation commencing at about 120° K and being practically complete at 90° K. One would be tempted to conclude that the vapour pressure of radon is extremely small at 90° K. It is clear, however, that in usual conditions radon cannot form a solid phase extending over any area of appreciable size. In fact, the number of atoms in a monatomic layer covering an area of 1 cm
2
is of the order of 10
15
, which corresponds to 57 millicuries of radon, whereas we know that condensation occurs with very much smaller quantities than this. Of course, there is still a possibility that condensed radon exists in the form of microscopic crystals with dimensions, say, of the order of 10
-3
cm. Such crystals would contain some 1013 atoms and would have a well-determined vapour pressure not very different from the vapour pressure of microscopic crystals if the surface tension of radon is not exceptionally large. We see, therefore, that it is possible at least theoretically to treat the problem of condensation of radon from the point of view of a thermodynamical equilibrium between the gaseous and the condensed phase. Very serious objections can, however, be raised against this interpretation. Owing to the investigations of Rutherford, and of Whytlaw-Gray and Ramsay, we know fairly well the vapour pressure of radon in a region of temperatures including the boiling point (210° K) and extending over some 100°. Extrapolating these results to low temperatures, one finds that the vapour pressure of radon at 90° K should be of the order of some tenths of a bar which corresponds to a concentration of some tenths of a millicurie in 1 cm
3
, and is at least 1000 times greater than the amount of gaseous radon in equilibrium with a surface cooled in liquid air. These conclusions are supported by some measurements of the vapour pressure of radon at low temperatures to be described in this paper. In these experiments the volume occupied by radon and the area of the cooled surface were sufficiently small to allow the formation of a polyatomic layer.
Vapour pressure and condensation of radon at low temperatures
