Thus, such powers as 3,000, 4,000, 5,000, even 6,652,[311] which Herschel now and again applied to his great telescopes, must, save on the rarest occasions, prove an impediment rather than an aid to vision. They were, however, used by him only for special purposes, experimentally, not systematically, and with the clearest discrimination of their advantages and drawbacks. It is obvious that perfectly different ends are subserved by increasing the _aperture_ and by increasing the _power_ of a telescope. In the one case, a larger quantity of light is captured and concentrated; in the other, the same amount is distributed over a wider area. A diminution of brilliancy in the image accordingly attends, _coeteris paribus_, upon each augmentation of its apparent size. For this reason, such faint objects as nebulæ are most successfully observed with moderate powers applied to instruments of a great capacity for light, the details of their structure actually disappearing when highly magnified. With stellar groups the reverse is the case. Stars cannot be magnified, simply because they are too remote to have any sensible dimensions; but the space between them can. It was thus for the purpose of dividing very close double stars that Herschel increased to such an unprecedented extent the magnifying capabilities of his instruments; and to this improvement incidentally the discovery of Uranus, March 13, 1781,[312] was due. For by the examination with strong lenses of an object which, even with a power of 227, presented a suspicious appearance, he was able at once to pronounce its disc to be real, not merely "spurious," and so to distinguish it unerringly from the crowd of stars amidst which it was moving. While the reflecting telescope was astonishing the world by its rapid development in the hands of Herschel, its unpretending rival was slowly making its way towards the position which the future had in store for it. The great obstacle which long stood in the way of the improvement of refractors was the defect known as "chromatic aberration." This is due to no other cause than that which produces the rainbow and the spectrum--the separation, or "dispersion" in their passage through a refracting medium, of the variously coloured rays composing a beam of white light. In an ordinary lens there is no common point of concentration; each colour has its own separate focus; and the resulting image, formed by the superposition of as many images as there are hues in the spectrum, is indefinitely terminated with a tinted border, eminently baffling to exactness of observation.