Some of the stars are extremely rarefied. Betelgeuse, for example, has a density about a thousandth that of air. We should call it a vacuum were it not contrasted with the much greater vacuosity of surrounding space. Nowadays physicists have no difficulty in producing a better vacuum than Betelgeuse; but in earlier times this star would have been regarded as a very creditable attempt at a vacuum. The outer parts of a star, and especially the light appendages such as the solar chromosphere and corona, reach much lower densities. Also the gaseous nebulae are, as their appearance suggests, extremely tenuous. When there is space enough to put a pin’s head between adjacent atoms we can begin to talk about a ‘real vacuum.’ At the centre of the Orion nebula that degree of rarefaction is probably reached and surpassed. A nebula has no definite boundary and the density gradually fades off. There is reason to think that the fading off becomes slow at great distances. Before we pass entirely out of the sphere of one nebula we enter the sphere of another, so that there is always some residual density in interstellar space. I believe that, reasoning from the tailing off of the nebulae, we are in a position to make an estimate of the amount of matter remaining unaggregated in space. An ordinary region where there is no observable nebulosity is the highest vacuum existing--within the limits of the stellar system at least--but there still remains about _one atom in every cubic inch_. It depends on our point of view whether we regard this as an amazing fullness or an amazing emptiness of space. Perhaps it is the fullness that impresses us most. The atom can find no place of real solitude within the system of the stars; wherever it goes it can nod to a colleague not more than an inch away. Let us approach the same subject from a different angle. In the ‘Story of Algol’ I referred to the way in which we measure the velocity of rotation of the sun. We point the spectroscope first on one limb of the sun and then on the other. Taking any one of the dark lines of the spectrum, we find that it has shifted a little between the two observations. This tells us that the material which imprinted the line was moving towards or away from us with different velocities in the two observations. That is what we expected to find; the rotation of the sun makes solar material move towards us on one side of the disk and away from us on the other side. But there are a few dark lines which do not show this change. They are in just the same position whether we observe them on the east or on the west of the sun. Clearly these cannot originate on the sun. They have been imprinted on the light after it left the sun and before it reached our telescope. We have thus discovered a medium occurring somewhere between the sun and our telescope; and as some of the lines are recognized as belonging to oxygen, we can infer that it is a medium containing oxygen.