The eastern edge of the sun is continually moving towards us with an equatorial speed of about a mile and a quarter per second, the western edge retreating at the same rate. The displacements--towards the violet on the east, towards the red on the west--corresponding to this velocity are very small; so small that it seems hardly credible that they should have been laid bare to perception. They amount to but 1/150th part of the interval between the two constituents of the D-line of sodium; and the D-line of sodium itself can be separated into a pair only by a powerful spectroscope. Nevertheless, Professor Young[632] was able to show quite satisfactorily, in 1876, not only deviations in the solar lines from their proper places indicating a velocity of rotation (1·42 miles per second) slightly in excess of that given by observations of spots, but the exemption of terrestrial lines (those produced by absorption in the earth's atmosphere) from the general push upwards or downwards. Shortly afterwards, Professor Langley, then director of the Allegheny Observatory, having devised a means of comparing with great accuracy light from different portions of the sun's disc, found that while the obscure rays in two juxtaposed spectra derived from the solar poles were absolutely continuous, no sooner was the instrument rotated through 90°, so as to bring its luminous supplies from opposite extremities of the equator, than the same rays became perceptibly "notched." The telluric lines, meanwhile, remained unaffected, so as to be "virtually mapped" by the process.[633] This rapid and unfailing mode of distinction was used by Cornu with perfect ease during his investigation of atmospheric absorption near Loiret in August and September, 1883.[634]