At first there was a rival interpretation. It was thought that the lines were produced in a cloud attached to the star--forming a kind of aureole round it. The two components travel in orbits round each other, but their orbital motion need not disturb a diffuse medium filling and surrounding the combined system. This was a very reasonable suggestion, but it could be put to the test. The test was again _velocity_. Although either component can move periodically to and fro within the surrounding cloud of calcium and sodium, it is clear that its average approach to us or recession from us taken over a long time must agree with that of the calcium and sodium if the star is not to leave its halo behind. Professor Plaskett with the 72-inch reflector at the Dominion Observatory in British Columbia carried out this test. He found that the secular or average rate of approach of the star[20] was in general quite different from the rate shown by the fixed calcium or sodium lines. Clearly the material responsible for the fixed lines could not be an appendage of the star since it was not keeping pace with it. Plaskett went farther and showed that whereas the stars themselves had all sorts of individual velocities, the material of the fixed lines had the same or nearly the same velocity in all parts of the sky, as though it were one continuous medium throughout interstellar space. I think there can be no doubt that this research demonstrates the existence of a cosmic cloud pervading the stellar system. The fullness of interstellar space becomes a fact of observation and no longer a theoretical conjecture. The system of the stars is floating in an ocean--not merely an ocean of space, not merely an ocean of ether, but an ocean that is so far material that one atom or thereabouts occurs in each cubic inch. It is a placid ocean without much relative motion; currents probably exist, but they are of a minor character and do not attain the high speeds commonly possessed by the stars.