What is the shape of our universe, and what are its dimensions? This is a tremendous question to ask. It is like asking an intelligent insect, living on a single leaf in the midst of a great Brazilian forest, to say what is the shape and size of the forest. Yet man's ingenuity has proved equal to giving an answer even to this question, and by a method exactly similar to that which would be adopted by the insect. Suppose, for instance, that the forest was shaped as an elongated oval, and the insect lived on a tree near the centre of the oval. If the trees were approximately equally spaced from one another they would appear much denser along the length of the oval than across its width. This is the simple consideration that has guided astronomers in determining the shape of our stellar universe. There is one direction in the heavens along which the stars appear denser than in the directions at right angles to it. That direction is the direction in which we look towards the Milky Way. If we count the number of stars visible all over the heavens, we find they become more and more numerous as we approach the Milky Way. As we go farther and farther from the Milky Way the stars thin out until they reach a maximum sparseness in directions at right angles to the plane of the Milky Way. We may consider the Milky Way to form, as it were, the equator of our system, and the line at right angles to point to the north and south poles. Our system, in fact, is shaped something like a lens, and our sun is situated near the centre of this lens. In the remoter part of this lens, near its edge, or possibly outside it altogether, lies the great series of star clouds which make up the Milky Way. All the stars are in motion within this system, but the very remarkable discovery has been made that these motions are not entirely random. The great majority of the stars whose motions can be measured fall into two groups drifting past one another in opposite directions. The velocity of one stream relative to the other is about twenty-five miles per second. The stars forming these two groups are thoroughly well mixed; it is not a case of an inner stream going one way and an outer stream the other. But there are not quite as many stars going one way as the other. For every two stars in one stream there are three in the other. Now, as we have said, some eminent astronomers hold that the spiral nebulae are universes like our own, and if we look at the two photographs (Figs. 25 and 26) we see that these spirals present features which, in the light of what we have just said about our system, are very remarkable. The nebula in Coma Berenices is a spiral edge-on to us, and we see that it has precisely the lens-shaped middle and the general flattened shape that we have found in our own system. The nebula in Canes Venatici is a spiral facing towards us, and its shape irresistibly suggests motions along the spiral arms.