Various theories are reviewed. By eliminating ideas which appear to be faulty, by accepting ideas which appear to be sound and helpful, and by introducing some new ideas to fill the gaps, the outline of a complete theory is built up. The only mechanism that appears to be capable of explaining the regularities of the solar system is fluid friction, and it is inferred that the material which now forms the planets and their satellites existed at one time in the form of a vast rotating disk. Bodies of planetary mass cannot be formed by condensation in a cloud of gas, and small solid particles must have been the chief ingredient in this primitive material. The angular momentum of the disk is explained quite simply by the assumption that the material was-collected from interstellar space, and this is the only explanation that gives a result which is numerically of the right order. Condensations would form in the rotating disk, and in the first instance these would be small and numerous. Subsequently they would coalesce to-form a single large cluster in each region of interplanetary space. Each cluster would consist of a nucleus and a surrounding disk or annulus. Owing to the combined effect of viscosity and tidal friction, there would be a loss of angular momentum and the bulk of the material would be absorbed by the nucleus to form the planet. The residue of material in the annulus would condense to form the satellites. In the region which lies immediately inside Jupiter's orbit the perturbations caused by the planet would drive much of the material inwards towards the Sun. In the attenuated residue the condensations would be smaller and more numerous than usual; they would coalesce to some extent, but not sufficiently to form a single planet, and the result would be the formation of the group of small bodies we know as the asteroids. In the region outside the orbit of Neptune the material would also be highly attenuated, and here again condensations would be small and numerous, but the progress of evolution was slower, and the region is probably populated by a very large number of small clusters. Wandering clusters make their appearance from time to time as comets.