The establishment of the rainbow as a covenant sign of the promise that there should be no flood again, presupposes that it appeared then for the first time in the vault and clouds of heaven. From this it may be inferred, not that it did not rain before the flood, which could hardly be reconciled with Gen. 2:5, but that the atmosphere was differently constituted; a supposition in perfect harmony with the facts of natural history, which point to differences in the climate of the earth's surface before and after the flood.Clearly, there are some topics on which Delitzsch was a simple idiot. He was living in the nineteenth century, for Pete's sake, and he didn't understand how rainbows form. I've known the answer to this old chestnut since childhood; I think I might have run across a strict literalist who took this position, and realised immediately its weakness. On the other hand, to be fair to the chief proponents of literalism, AiG list this rainbow-myth as a factoid to be denied. They cite favourably exactly the same source I would cite, and I would add, lisez Chauvin, c'est notre maitre a tous.
From these words certain eminent theologians have been induced to deny that there was any rainbow before the deluge: which is frivolous. For the words of Moses do not signify, that a bow was then formed which did not previously exist; but that a mark was engraven upon it, which should give a sign of the divine favor towards men. …Hence it is not for us to contend with philosophers respecting the rainbow; for although its colors are the effect of natural causes, yet they act profanely who attempt to deprive God of the right and authority which he has over his creatures.Sometimes, it's quite sweet to read pre-17th century sources for their scientific innocence—there's a delightful passage of Augustine in which he describes, without a shade of understanding, magnetism—and then there are times when it's really quite refreshing to see how much they actually knew.