So far, Bradley had but come to the same point which Flamsteed had reached. Flamsteed had detected precisely the same apparent displacement of stars, and, like Hooke, had ascribed it to parallax. Cassini had shown that this could not be the case, as the displacement was in the wrong direction; and there the matter had rested. Bradley now set to follow the question up. Other stars beside Gamma Draconis were found to show a displacement of the same general character, but the amount varied with their distance from the plane of the ecliptic, the earth's orbit. The first explanation suggested was that the axis of the earth, which moves very nearly parallel to itself as the earth moves round the sun, underwent a slight regular 'wobble' in the course of a year. To check this, a star was observed on the opposite side of the pole from Gamma Draconis; then Bradley investigated as to whether refraction might explain the difficulty, but again without success. He now was most keenly interested in the problem, and he purchased a zenith telescope of his own, made, like that of Molyneux, by Graham, and mounted it in his aunt's house at Wanstead, and observed continuously with it. The solution of the problem came at last to him as he was boating on the Thames. Watching a vane at the top of the mast, he saw with surprise that it shifted its direction every time that the boat was put about. Remarking to the boatmen that it was very odd that the wind should change just at the same moment that there was a shift in the boat's course, they replied that there was no change in the wind at all, and that the apparent change of the vane was simply due to the change of direction of the motion of the boat. [Illustration: GRAHAM'S ZENITH SECTOR. (_From an old print._)] This supplied Bradley with a key to the solution of the mystery that had troubled him so long. It had been discovered long before this that light does not travel instantaneously from place to place, but takes an appreciable time to pass from one member of the solar system to another. This had been discovered by Römer from observations of the satellites of Jupiter. He had noted that the eclipses of the satellites always fell late of the computed time, when Jupiter was at his greatest distance from the earth; and Bradley's own work in the observation of those satellites had brought the fact most intimately under his own acquaintance. The result of the boating incident taught him, then, that he might look upon light as analogous to the wind blowing on the boat. As the wind, so long as it was steady, would seem to blow from one fixed quarter so long as the boat was also in rest, but as it seemed to shift its direction when the boat was moving and changed its direction, so he saw that the light coming from a particular star must seem to slightly change the direction in which it came, or, in other words, the apparent position of the star, to correspond with the movement of the earth in its orbit round the sun.