In spectra of the solar type two great bars of violet light are stopped out by calcium; otherwise metallic arc-lines predominate, while those of hydrogen are no longer so powerfully emphasized as in white stars. Moreover, the whiteness of the unveiled Sirian photospheres has become tinged with yellow owing to the development of a shallow envelope partly impermeable to blue rays. For this reason the comparative extension of their ultra-violet spectra affords, for stars of different types, no secure criterion of relative temperature. Sound in principle, it becomes inapplicable when the unknown factor of general absorption comes into play. The energy-curve of the solar spectrum, as it is, can be determined; the energy-curve of the solar spectrum, as it would be if unaffected by general absorption, has to be constructed from inference. But only photospheres bare to space give comparable results. Hence, there are no valid grounds for asserting that Sirius is hotter than the sun, or the sun than Betelgeux. It may be so, but the evidence at present available is inconclusive. The appearances expounded in this sense may bear quite different meanings. The reasons for holding that solar mature into Antarian stars are of the same character, and of equal cogency with those tending to prove their own development from luminaries of Sirian type. There is a similar continuity of specimens. They can be ranged one after another in an unbroken series, in which, as we run down the line, primrose shades into orange, and orange into red; general absorption arrests an increasing percentage of the blue radiations, while specific absorption becomes strengthened by dusky channellings of titanium. Carbon stars are less easily located. Dr. Vogel regards them as co-ordinate with the Antarian class. The two varieties of red stars with banded spectra descend, in his opinion, from the common stock exemplified by our sun. Professor Hale also favours this view, some attendant difficulties notwithstanding. His photographs have certainly established for carbon stars links of relationship both with the Antarian and the solar families; yet the fact remains indisputable that the carbon type is, to a great extent, isolated from all the rest. Tokens of a genuine migration towards it are few and obscure.