Zeller’s difficulty about the meaning of τροπαί here (p. 223, n. 2) seems to be an imaginary one. The moon has certainly a movement in declination and, therefore, τροπαί (Dreyer, _Planetary Systems_, p. 17, n. 1). Footnote 130: I assume with Diels (_Dox._ p. 560) that something has fallen out in our text of Hippolytos. I have, however, with Tannery, _Science hellène_, p. 91, supplied “eighteen times” rather than “nineteen times.” Zeller (p. 224, n. 2) prefers the text of our MS. of Hippolytos to the testimony of Aetios. Footnote 131: Aetios goes on to say that the moon also is like a hollow cart-wheel full of fire with an ἐκπνοή. The difference in the figures of Hippolytos and Aetios is due to the fact that one refers to the internal and the other to the external circumferences of the rings. Cf. Tannery, _Science hellène_, p. 91; and Diels, “Ueber Anaximanders Kosmos” (_Arch._ x. pp. 231 sqq.). We saw above that the sphere of flame was broken up into rings by the expansion of the air or vapour that its own heat had drawn up from the moist, cold interior. We must remember that Anaximander knew nothing of the ring of Saturn. There are three of these rings, that of the sun, that of the moon, and, lastly, nearest to the earth, the circle of the stars. The circle of the sun was twenty-seven times, and that of the moon eighteen times as large as the earth, from which we may perhaps infer that the circle of the stars was nine times as large. The numbers nine, eighteen, twenty-seven, play a considerable part in primitive cosmogonies.[132] We do not see the rings of fire as complete circles; for the mist that formed them encloses the fire, and becomes an outer ring of opaque vapour. These outer rings, however, have openings at one point of their circumference, through which the fire escapes, and these are the heavenly bodies we actually see.[133] Footnote 132: As Diels points out (_Arch._ x. p. 229) the explanation given by Gomperz, p. 53, cannot be right. It implies the fifth century theory of μύδροι. Anaximander knew nothing of the “great mass” of the sun. Footnote 133: The true meaning of this doctrine was first explained by Diels (_Dox._ pp. 25 sqq.). The flames rush forth _per magni circum spiracula mundi_, as Lucretius has it (vi. 493). The πρηστῆρος αὐλός, to which these are compared, is simply the nozzle of a pair of bellows, a sense which the word πρηστήρ has in Apollonios Rhodios (iv. 776), and has nothing to do with the meteorological phenomenon of the same name, for which see Chap. III. § 71. It is not now necessary to refute the earlier interpretations.