Tobias Mayer had illustrated the perspective effects which must ensue in the stellar sphere from a translation of the solar system, by comparing them to the separating in front and closing up behind of trees in a forest to the eye of an advancing spectator;[18] but the appearances which he thus correctly described he was unable to detect. By a more searching analysis of a smaller collection of proper motions, Herschel succeeded in rendering apparent the very consequences foreseen by Mayer. He showed, for example, that Arcturus and Vega did, in fact, appear to recede from, and Sirius and Aldebaran to approach, each other by very minute amounts; and, with a striking effort of divinatory genius, placed the "apex," or point of direction of the sun's motion, close to the star Lambda in the constellation Hercules,[19] within a few degrees of the spot indicated by later and indefinitely more refined methods of research. He resumed the subject in 1805,[20] but though employing a more rigorous method, was scarcely so happy in his result. In 1806,[21] he made a preliminary attempt to ascertain the speed of the sun's journey, fixing it, by doubtless much too low an estimate, at about three miles a second. Yet the validity of his general conclusion as to the line of solar travel, though long doubted, has been triumphantly confirmed. The question as to the "secular parallax" of the fixed stars was in effect answered. With their _annual_ parallax, however, the case was very different. The search for it had already led Bradley to the important discoveries of the aberration of light and the nutation of the earth's axis; it was now about to lead Herschel to a discovery of a different, but even more elevated character. Yet in neither case was the object primarily sought attained.