A fresh witness was by this time at hand. The application of the spectroscope to the direct examination of the sun's surface dates from March 4, 1866, when Sir Norman Lockyer (to give him his present title) undertook an inquiry into the cause of the darkening in spots.[461] It was made possible by the simple device of throwing upon the slit of the spectroscope an _image_ of the sun, any part of which could be subjected to special scrutiny, instead of, as had hitherto been done, admitting rays from every portion of his surface indiscriminately. The answer to the inquiry was prompt and unmistakable, and was again, in this case, adverse to the French theorist's view. The obscurations in question were found to be produced by no deficiency of emissive power, but by an increase of absorptive action. The background of variegated light remains unchanged, but more of it is stopped by the interposition of a dense mass of relatively cool vapours. The spectrum of a sun-spot is crossed by the same set of multitudinous dark lines, with some minor differences, visible in the ordinary solar spectrum. We must then conclude that the same vapours (speaking generally) which are dispersed over the unbroken solar surface are accumulated in the umbral cavity, the compression incident to such accumulation being betrayed by the thickening of certain lines of absorption. But there is also a general absorption, extending almost continuously from one end of the spot-spectrum to the other. Using, however, a spectroscope of exceptionally high dispersive power, Professor Young of Princeton, New Jersey, succeeded in 1883 in "resolving" the supposed continuous obscurity of spot-spectra into a countless multitude of fine dark lines set very close together.[462] Their structure was seen still more perfectly, about five years later, by M. Dunér,[463] Director of the Upsala Observatory, who traced besides some shadowy vestiges of the crowded doublets and triplets forming the array, from the spots on to the general solar surface. They cease to be separable in the blue part of the spectrum; and the ultra-violet radiations of spots show nothing distinctive.[464]