Pausanias says ‘a sanctuary of Apollo in a cave.’ It is the fact that the sanctuary is in a cave that strikes and interests him. He does not call it a _Pythion_. But by another writer the actual word _Pythion_ is used. Philostratos[155] describes the route taken by the Panathenaic ship thus: starting from the outer Kerameikos it sailed to the Eleusinion, and, having rounded it, it was carried along past the Pelasgikon and came alongside of the Pythion, where it is now moored. The Panathenaic way has been, as will later be seen (p. 131), laid bare; for the moment all that concerns us is that the Pythion is mentioned immediately after the Pelasgikon and was therefore presumably next to it. Philostratos puts what he calls the _Pythion_ in just the place where Pausanias[156] saw his ‘sanctuary in a cave’; the two are identical. Further, any doubts as to where the ship was moored are set at rest by Pausanias himself. He saw the ship and noted its splendour. It stood ‘near the Areopagus.’ The Pythion must have stood at the North-West corner of the Acropolis (Fig. 46). * * * * * Even if we relied on literary evidence only we should be quite sure that the Pythion of which Thucydides speaks was somewhere on the Long Rocks, at the North-West end of the Acropolis. Happily however the situation is not left thus vague; the actual cave of Apollo has been found, and thoroughly cleared out, and in it there came to light numerous inscribed votive offerings to the god, which make the ascription certain. From the lower tower at the North-West corner there have always been clearly visible to any one looking up from below three caves (Fig. 21), a very shallow one immediately over the Klepsydra, and two others nearer together and somewhat deeper separated from the first by a shoulder of rock. On the plan in Fig. 22 these are marked Α, Β and Γ. The question has long been raised which of the three belonged to Apollo and which to Pan. As Pausanias[157] first mentions the sanctuary of Apollo in a cave and then passes on to tell the story of Pheidippides, manifestly _à propos_ of Pan’s cave, it has been usual to connect Α with Apollo and Β and Γ, one or both, with Pan. [Illustration: FIG. 21.] But the identification has never been felt to be quite satisfactory. The cave Α is really no cave at all; it is a very shallow _niche_. It is impossible to imagine it the scene of the story of Creousa. Moreover it bears no traces of any votive offerings having been attached to its wall, nor have any remains of such been found there.