that they proceed from the sun, but are only apparent, for if another cloud comes across the face of the luminary they at once vanish (29). The greater diameter of the rainbow compared with that of the sun as seen by us he accounts for by the analogy of a concave mirror, which greatly enlarges the objects reflected from it. At the conclusion of the discussion he repeats his belief that the rainbow and the corona or halo have no definite material inherent in them, but are like a mirror which reveals only a deception, the mere phantoms and empty imitations of real bodies, which certainly do not exist in the mirror, and therefore cannot come out of it (41). In Chapters XVI. and XVII. the author indulges in one of his favourite moralising episodes, suggested by the topics he has been discussing in the previous pages. He takes the existence of reflecting surfaces as his text, and from the calm surface of still water passes on to artificial mirrors, contrasting the manners and morals of early mankind, who had only pools and lakes in which to see their faces, with the luxury and vice of later ages, when the use of metals led to the invention of metallic mirrors. In this retrospect, however, he places the discovery of the use of iron before that of the other metals. The priority of bronze and the reason for it are accurately stated by Lucretius: et prior aeris erat quam ferri cognitus usus, quo facilis magis est natura et copia maior.[110] [110] _De Rer. Nat._ v. 1287. BOOK II