The general opinion, especially among astronomers, is that the compression of the nebulous material of the arms into globes would generate enormous heat, as in the case of the sun. On that view the various planets would begin their careers as small suns, and would pass through those stages of cooling and shrinking which we have traced in the story of the stars. A glance at the photograph of one of the spiral nebulae strongly confirms this. Great luminous knots, or nuclei, are seen at intervals in the arms. Smaller suns seem to be forming in them, each gathering into its body the neighbouring material of the arm, and rising in temperature as the mass is compressed into a globe. The spectroscope shows that these knots are condensing masses of white-hot liquid or solid matter. It therefore seems plain that each planet will first become a liquid globe of fire, coursing round the central sun, and will gradually, as its heat is dissipated and the supply begins to fail, form a solid crust. This familiar view is challenged by the new "planetesimal hypothesis," which has been adopted by many distinguished geologists (Chamberlin, Gregory, Coleman, etc.). In their view the particles in the arms of the nebula are all moving in the same direction round the sun. They therefore quietly overtake the nucleus to which they are attracted, instead of violently colliding with each other, and much less heat is generated at the surface. In that case the planets would not pass through a white-hot, or even red-hot, stage at all. They are formed by a slow ingathering of the scattered particles, which are called "planetesimals" round the larger or denser masses of stuff which were discharged by the exploding sun. Possibly these masses were prevented from falling back into the sun by the attraction of the colliding body, or the body which caused the eruption. They would revolve round the parent body, and the shoals of smaller particles would gather about them by gravitation. If there were any large region in the arm of the nebula which had no single massive nucleus, the cosmic dust would gather about a number of smaller centres. Thus might be explained the hundreds of planetoids, or minor planets, which we find between Mars and Jupiter. If these smaller bodies came within the sphere of influence of one of the larger planets, yet were travelling quickly enough to resist its attraction, they would be compelled to revolve round it, and we could thus explain the ten satellites of Saturn and the eight of Jupiter. Our moon, we shall see, had a different origin.