Boeckh admitted a very slow motion of the heaven of the fixed stars, which he at first supposed to account for the precession of the equinoxes, though he afterwards abandoned that hypothesis (_Untersuchungen_, p. 93). But, as Dreyer admits (_Planetary Systems_, p. 49), it is “not ... necessary with Boeckh to suppose the motion of the starry sphere to have been an exceedingly slow one, as it might in any case escape direct observation.” It seems probable that the theory of the earth’s revolution round the central fire really originated in the account given by Empedokles of the sun’s light. The two things are brought into close connexion by Aetios, who says that Empedokles believed in two suns, while Philolaos believed in two or even in three.[825] The theory of Empedokles is unsatisfactory in so far as it gives two inconsistent explanations of night. It is, we have seen, the shadow of the earth; but at the same time Empedokles recognised a fiery diurnal hemisphere and a nocturnal hemisphere with only a little fire in it.[826] All this could be simplified by the hypothesis of a central fire which is the true source of light. Such a theory would, in fact, be the natural issue of the recent discoveries as to the moon’s light and the cause of eclipses, if that theory were extended so as to include the sun. Footnote 825: Aet. ii. 20, 13 (Chap. IV. p. 275, _n._ 609); cf. _ib._ 12 (of Philolaos), ὥστε τρόπον τινὰ διττοὺς ἡλίους γίγνεσθαι, τό τε ἐν τῷ οὐρανῷ πυρῶδες καὶ τὸ ἀπ’ αὐτοῦ πυροειδὲς κατὰ τὸ ἐσοπτροειδές· εἰ μή τις καὶ τρίτον λέξει τὴν ἀπὸ τοῦ ἐνόπτρου κατ’ ἀνάκλασιν διασπειρομένην πρὸς ἡμᾶς αὐγήν. Here τὸ ἐν τῷ οὐρανῷ πυρῶδες is the central fire, in accordance with the use of the word οὐρανός explained in another passage of Aetios, Stob. _Ecl._ i. p. 196, 18 (R. P. 81). It seems to me that these strange notices must be fragments of an attempt to show how the heliocentric hypothesis arose from the theory of Empedokles as to the sun’s light. The meaning is that the central fire really was the sun, but that Philolaos unnecessarily duplicated it by supposing the visible sun to be its reflexion. Footnote 826: Chap. VI. § 113.