Before the introduction of spectrum analysis all stars were supposed to be suns, and the only difference recognized among them was one of brilliancy and the variation of brilliancy in the case of some of them. It ultimately came out that great classes might be recognized by the differences of their spectra, which were ultimately traced to differences in their chemistry and in their temperature, as determined by the extension of the spectra in the ultra-violet, the whiter stars being hotter than the red ones, as a white-hot poker is hotter than a red-hot poker. Next there was evidence to show that a large proportion of the stars were not stars at all like the sun, but swarms of meteorites; and in this way the mysterious new stars which appear from time to time in the heavens, and a large number of variable stars, were explained as arising from collisions among such swarms. The inquiry which dealt with the spectroscopic results, having thus introduced the ideas of meteor swarms and collisions to explain many stellar phenomena, went further and showed that the various chemical changes observed in passing from star to star might also be explained by supposing the whole stellar constitution to arise from cool meteoritic swarms represented by nebulæ, the changes up to a certain point being explained by a rise of temperature due to condensation towards a centre. Here the new view was opposed to that of Laplace, advanced during the last century, that the stars were produced by condensation and cooling; but Kelvin had shown, before the new view was enunciated, that Laplace’s view was contrary to thermodynamics, a branch of science which had developed since Laplace published his famous _Exposition du Système du Monde_. After all the meteorites in the parent swarm had been condensed into the central gaseous mass, that mass had to cool. So that we had in the heavens not only stars more or less meteoritic in structure, of _rising_ temperature, but stars chiefly gaseous, of falling temperature. It was obvious that representatives of both these classes of stars might have nearly the same mean effective temperature, and therefore more or less the same spectrum. A minute inquiry entirely justified these conclusions. So far had the detailed chemistry of the stars been carried in the latter years of the century that the question of stellar evolution gave rise to that of inorganic evolution generally, the sequence in the phenomena of which can only be studied in the stars, for laboratory work without stint has shown that in them we have celestial furnaces, the heat of which transcends that of our most powerful electric sparks. In this way astronomy is paying the debt she owes to chemistry. THE SUN AND HIS SYSTEM