The spectroscope affords little information as to the constitution of Mercury. Its light is of course that of the sun reflected, and its spectrum is consequently a faint echo of the Fraunhofer spectrum. Dr. H. C. Vogel, who first examined it in April, 1871, _suspected_ traces of the action of an atmosphere like ours,[809] but, it would seem, on slight grounds. It is, however, certainly very poor in blue rays. More definite conclusions were, in 1874,[810] derived by Zöllner from photometric observations of Mercurian phases. A similar study of the waxing and waning moon had afforded him the curious discovery that light-changes dependent upon phase vary with the nature of the reflecting surface, following a totally different law on a smooth homogeneous globe and on a rugged and mountainous one. Now the phases of Mercury--so far as could be determined from only two sets of observations--correspond with the latter kind of structure. Strictly analogous to those of the moon, they seem to indicate an analogous mode of surface-formation. This conclusion was fully borne out by Müller's more extended observations at Potsdam during the years 1885-1893.[811] Practical assurance was gained from them that the innermost planet has a rough rind of dusky rock, absorbing all but 17 per cent. of the light poured upon it by the fierce adjacent sun. Its "albedo," in other words, is 0·17,[812] which is precisely that ascribed to the moon. The absence of any appreciable Mercurian atmosphere followed almost necessarily from these results. On March 26, 1800, Schröter, observing with his 13-foot reflector in a peculiarly clear sky, perceived the southern horn of Mercury's crescent to be quite distinctly blunted.[813] Interception of sunlight by a Mercurian mountain rather more than eleven English miles high explained the effect to his satisfaction. By carefully timing its recurrence, he concluded rotation on an axis in a period of 24 hours 4 minutes. The first determination of the kind rewarded twenty years of unceasing vigilance. It received ostensible confirmation from the successive appearances of a dusky streak and blotch in May and June, 1801.[814] These, however, were inferred to be no permanent markings on the body of the planet, but atmospheric formations, the streak at times drifting forwards (it was thought) under the fluctuating influence of Mercurian breezes. From a rediscussion of these somewhat doubtful observations Bessel inferred that Mercury rotates on an axis inclined 70° to the plane of its orbit in 24 hours 53 seconds.