19. This would be so, we say, if the luminous matter moved in a greatly resisting medium. But what is the measure of _great_ resistance? It is, as we have already said, that the resistance which opposes the motion shall bear a considerable proportion to the force which deflects the motion. But what is that force? Upon the theory of the universal gravitation of matter, on which theory we here proceed, the force which deflects the motions of the parts of each system into curves, is the mutual attraction of the parts of the system; leaving out of the account the action of other systems, as comparatively insignificant and insensible. The condition, then, for the production of such spiral figures as I have spoken of, amounts really to this; that the mutual attraction of the parts of the luminous matter is slight; or, in other words, that the matter itself is very thin and rare. In that case, indeed, we can easily see that such a result would follow. A cloud of dust, or of smoke, which was thin and light, would make but a little way through the air, and would soon fall downwards; while a metal bullet shot horizontally with the same velocity, might fly for miles. Just so, a loose and vaporous mass of cometic matter would be pulled rapidly inwards by the attraction to the centre; and supposing it also drawn into a long train, by the different density of its different parts, it would trace, in lines of light, a circular or elliptical spiral converging to the centre of attraction, and resembling one of the branches of the spiral nebulæ. And if several such cometic masses thus travelled towards the centre, they would exhibit the wheel-like figure with bent spokes, which is seen in the spiral nebulæ. And such a figure would all the more resemble some of these nebulæ, as seen through Lord Rosse's telescope, if the spirals were accompanied by exterior branches of thinner and fainter light, which nebulous matter of smaller density might naturally form. Perhaps too, such matter, when thin, may be supposed to cool down more rapidly from its state of incandescence; and thus to become less luminous. If this were so, a great optical power would of course be required, to make the diverging branches visible at all.