The most original theory of the winds was not any of those which attempted to account for purely local breezes but an explanation propounded by William of Conches of the circulation of the atmosphere as a whole. Unlike our modern conceptions of atmospheric circulation based on the observation of facts, William’s ingenious theory seems to have been the product of his own vivid imagination. It was founded on a persistent idea, dating back to classical times, that disturbances in the water can produce currents of air. Gervase of Tilbury, for example, states in so many words that “mountains and water cause winds” and that the swift-flowing Rhone makes the _mistral_ that blows over Provence and Dauphiny.[770] William of Conches[771] believed that there are two ocean currents trending east and west out of the equatorial ocean. Each of these was supposed to divide in two at the extremities of our _oikoumene_, making four currents which collide at the North and South Poles in the ocean perpendicular to the equatorial ring (Amphitrites). The cardinal winds are generated at four points, at the two junctions of the oceans where the currents divide and at the poles where they collide. The western division gives rise to Zephyr, the eastern to Eurus, the collision at the North Pole to Boreas, and the one at the South Pole to Auster. It may happen, however, that one of the currents will on occasion flow more strongly than its opponent and will push the point of collision beyond the pole. This displacement of the point of collision explains the blowing of the collateral winds. Absurd as it may be in itself, this theory is of interest to us mainly because it shows that William understood that a broad system of atmospheric circulation is possible and assigned to it, as well as to local breezes, a purely physical cause. Curiously enough, it is the exact reverse of our modern conception of the usual relation existing between atmospheric and ocean currents, for now we understand that the winds are more effective as the cause of the ocean currents than vice versa. William also maintained, as we shall see later,[772] that the tides are produced by the impact of ocean currents. Why then, it was asked, if the tides are of daily, periodic occurrence, do not the winds, which he tells us result from the same cause, show a similar periodicity? To this William replied[773] that the winds in fact do show such regularity but that it is not apparent to us for two reasons: in the first place, wind produced by these causes does not always reach the part of the earth where we happen to be; and, secondly, the resulting wind may blow at such a high altitude as not to be noticed by men on the ground—an observation now well known to be true. _Names of the Winds_