Only a fraction of one per cent of the sun's mass ejected in this fashion would be sufficient to generate the entire planetary system. Nuclei or knots in the arms of the spiral gradually grew by accretion, the four interior knots forming Mercury, Venus, the Earth, and Mars. The earth knot was a double one, which developed into the earth-moon system. The absence of a dominating nucleus beyond Mars accounts for the zone of the asteroids remaining in some sense in the original planetesimal condition. The vaster nuclei beyond Mars gradually condensed into Jupiter, Saturn, Uranus, and Neptune; and lesser nuclei related to the larger ones form the systems of moons or satellites. The orbits of the planetesimals and the planetary and satellite nuclei would be very eccentric, forming a confusion of ellipses with frequently crossing paths. Collisions would occur, and the nuclei would inevitably grow by accretion. Each planet, then, would clear up the planetesimals of its zone; and Moulton shows that this process would give rise to axial revolution of the planet in the same direction as its orbital revolution. The eccentricities would finally disappear, and the entire mass would revolve in a nearly circular orbit. Rotation twists the streams into the spiral form, and the huge amounts of wreckage from the near-collision are thrown into eddies. The fragments or particles (planetesimals) which have given the name to the theory, begin their motion round their central sun in elliptical paths as required by gravitation. The form of the spiral is preserved by the orbital motion of its particles. There is a gradual gathering together of the planetesimals at points or nodes of intersection, and these become aggregations of matter, nuclei that will perhaps become planets, though more likely other stars. The appulse or near approach is but one of the methods by which the spiral nebulæ may have come into existence. The planetesimal hypothesis would seem to account for the formation of many of these objects as we see them in the sky, though perhaps it is hardly competent to replace entirely the Laplacian hypothesis of the formation of the solar system, which would appear to be a special case by itself. It will be observed that while the Laplacian hypothesis is concerned in the main with the progressive development of the solar system, and systems of a like order surrounding other stellar centers, whose existence is highly probable, the origin and development of the stellar universe is a vaster problem which can only be undertaken and completed in its broadest bearings when the structure of the stellar universe has been ascertained.