Page:Early Greek philosophy by John Burnet, 3rd edition, 1920.djvu/56

42 as he picked up in Egypt would never have enabled him to calculate one. Yet the evidence for the prediction is too strong to be rejected off-hand. The testimony of Herodotos is said to have been confirmed by Xenophanes, and according to Theophrastos Xenophanes was a disciple of Anaximander. In any case, he must have known scores of people who were able to remember what happened. The prediction of the eclipse is therefore better attested than any other fact about Thales whatsoever.

Now it is possible to predict eclipses of the moon approximately without knowing their true cause, and there is no doubt that the Babylonians actually did so. It is generally stated, further, that they had made out a cycle of 223 lunar months, within which eclipses of the sun and moon recurred at equal intervals of time. This, however, would not have enabled them to predict eclipses of the sun for a given spot on the earth's surface; for these phenomena are not visible at all places where the sun is above the horizon at the time. We do not occupy a position at the centre of the earth, and the geocentric parallax has to be taken into account. It would only, therefore, be possible to tell by means of the cycle that an eclipse of the sun would be visible somewhere, and that it might be worth while to look out for it, though an observer at a given place