were called ταμίαι τῆς θεοῦ. Now the grotto on Cynthus certainly could not have been used for such a purpose: neither its structure nor its situation afforded the needful security.
Considering all these facts, we cannot, I think, resist the conclusion that, as early at least as 475 B.C., and almost certainly at a much earlier date, Apollo already possessed a temple in Delos distinct from the grotto. Now we know that the most ancient altars in Delos (the κεράτινος and that of Apollo Genitor) stood near the oval basin. And, as early at least as the Odyssey, the palm which saw Latona's pangs was shown near an altar. The site of Apollo's temple can scarcely be sought, then, elsewhere than on the spot where remains now exist. If all these remains were of the same age, we should have our choice between referring them to a much higher date than 400 B.C., or supposing that the temple to which they belonged had occupied the site of an older building. I have stated my difficulty. I do not propound a definite solution, for which it may be doubted whether sufficient data exist. The hypothesis, however, to which I should incline is this. The temple, of which the remains have been examined by M. Homolle, may have been partially repaired at more than one time, and these fragments, from which he estimates the date of the whole building, may be of the age which he assigns to them, i.e. about 400 B.C. But, either on these foundations, or at least on this plain, a temple of Apollo, however rude, must have stood long before
