S o ME years ago Dessler suggested a new design of a cavity for measuring the velocity of second sound in liquid helium below 1 ° K. The second sound pulses were to be produced at one focus of an ellipsoidal cavity and a carbon resistor would pick up the signal at the other focus. Because of the geometrical properties of the ellipsoid, the second sound wave reflected at the wall would always travel the same length before reaching the detector. One of the aims of the experiment was to study the shape of the received pulse at temperatures where the mean free path of the phonons was comparable with the dimensions of the cavity, this being below 0.5 ° K. Some preliminary experiments with this cavity were made above 1 ° K by Dessler and W. M. Fairbank, who also planned the cryostat for lower temperatures.
In 1960 the cryostat was redesigned by the present authors and operated down to temperatures of about 0-1 ° K. There it was discovered by chance that signals were detected when there was only gas and a helium film in the cavity. These signals were still strong at temperatures below 0-4 ° K when all the helium was condensed on the walls. Since the circumstances of this experiment were rather unusual, the original purpose was put aside and the new phenomena were investigated in detail. These will be described below. In particular, the region below 0.5 ° K, where the mean free path of the helium gas becomes larger than the cavity dimensions, has been the object of our interest and the production and properties of the "molecular beams' will be discussed. The studies were then extended to helium-3, which was available at a purity of about 99 per cent, and the results compared with those of helium-4.
Experimental
The ellipsoidal cavity consisted of two copper shells fitted together. These shells were produced by electrolytic deposition of copper on one half of a silvered lucite ellipsoid, which was finally removed from the electrode. The inside of the resulting shells was further