Advanced spectroscopic means reveals that the spectra of these “white” nebulæ are not simply continuous. Thus that of the Andromeda nebula shows very faint dark lines crossing it, apparently accordant with those of the solar spectrum and faint bright ones falling near and probably coincident with those of the Wolf-Rayet stars, due to hydrogen, helium, and so forth. These later observations make practically certain what earlier ones permitted us just now only to infer: that it is not composed of stars, but of something subtler still; to wit, of meteorites. The reasoning is interesting, as showing that if one have hold of a true idea, the stars in their courses fight for him. [Illustration: NEBULA M. 51 CANUM VENATICORUM—AFTER RITCHEY.] Although Lockyer has long been of opinion that the nebulæ are composed of meteorites, the present argument differs from his. The way in which their spectra establish their constitution may be outlined as follows: the white nebulæ are from their structure evidently in process of evolution, and if they are in stable motion, as we suppose them to be, their parts are moving round their common centre of gravity. As the white nebulæ resist resolution as obstinately as the green, these parts must be not only solid but comminuted (composed of small particles). Now this would be the case were they flocks of meteorites such as we have seen composed our own system once upon a time. Though all are travelling round the centre of gravity of the flock, each is pursuing its own orbit slightly different from, and intersecting those of, its neighbors. Collisions between the meteors must therefore constantly occur, and the question is, are these shocks sufficient to cause light. Let us take our own system and consider two meteorites at our distance from the Sun, travelling in the same sense, the one in an ellipse, the other in a circle, with a major axis five per cent greater and meeting the other at aphelion. This would be no improper jostle for such heavenly bodies. If we calculate the speeds of both and deduct the elliptic from the circular, we shall have the relative speed of collision. It proves to be a half a mile a second or 30 times the speed of an express train. As such a train brought up suddenly against a stone wall would certainly elicit sparks, we see that a speed 30 times as great, whose energy is 900 times greater, is quite competent to a shock sufficient to make us see stars _en masse_. But, indeed, there must be collisions much more violent than this; both because the central mass is often much greater and because the orbits differ much more, and the effect would increase as the square of the speed. The heat thus generated would cause the meteorites to glow, and at the same time raise the temperature of the gases in and about them. Furthermore, the light would come to us through other non-affected portions of gas between us and the scene of the collision. Thus all three peculiarities