temperature, till finally the main absorption of light is caused by hydrogen and iron. Toward the top of the curve we get hydrogen enormously developed. It seems that we deal with a greater and greater quantity of hydrogen as the temperature gets higher.
Side by side with this sequence in the case of stars, a similar one up to a certain point is noted in the comets. As a rule the temperature of comets is, as we should expect, very much below that reached by stars. There is, therefore, no overwhelming indication of light-absorption, and it is only in those which closely approach the sun that any indication of the absorption of light caused by the presence of iron vapor is to be seen. A comparison of the spectra observed gives a clear indication that the nature of comets and nebulæ, so far as the spectroscope can seize them, is very similar: the phenomena present themselves in the same order; a line common to both begins the story, and then bright carbon is found among the first substances indicated, and afterward absorption phenomena, produced by manganese and lead chiefly, it is supposed, are superadded.
After this cometary parenthesis I now return to consider the top of the temperature curve. I repeat that we have this sort of condition. The swarms, whether single or multiple in origin, have by collisions and gravity brought about the highest point of temperature which they can reach in consequence of these actions. Swarms of separate meteorites now give place to a globular mass of gas produced by their volatilization. It may be that this very high temperature may be produced, and this enormous globular mass of gas formed, long before all the meteorites and meteoritic dust in the parent swarm, or in that particular region of space, shall be absolutely condensed to the center; so that we see it is quite possible that this high temperature condition may last for a very long time. Hence the curve should be flat-topped—in all probability very flat—for, so far as the spectrum analysis of stars has gone at present, more than half of those which have been examined give us evidence of extremely high temperature. However that may be, it is easily to be understood that such a mass as that we are considering must be radiating with tremendous energy; for a time probably the heat which it receives by the collisions and condensation of the outer members of the parent swarm may be as great as the heat which it radiates, and under these conditions the average temperature of the gas will remain constant; but the moment the input is less than the output the mass of gas must cool, so that we have next to consider what will happen to a mass of gas cooling under these circumstances.
What will cool first? The outside. We know pretty well the chemical nature of the outside of the mass of gas we are dealing
