In putting forward these theories as to invisible stellar satellites, Bessel remarked that “light is no real property of mass,” and that the existence of countless visible stars is nothing against the existence of countless invisible and dark ones. In this he laid the foundation of the branch of science termed by Mädler the “Astronomy of the invisible.” In recent years the astronomy of the invisible has become a recognised branch of astronomical research, through the application and interpretation of Doppler’s principle in spectroscopic observations. In the course of photographing the stellar spectra for the Draper Catalogue, E. C. Pickering photographed the spectrum of Mizar (ζ Ursæ Majoris) in 1887 and again in 1889. On some of these photographs the line K was seen double, while on others it was seen under its normal aspect. This doubling of the lines indicated that the star which we see as single is in reality composed of two bodies in revolution round their centre of gravity, so close together that even the largest telescopes cannot divide them. Pickering assigned a period of 104 days, but in 1901 Vogel diminished this to 20 days. In the same year the star β Aurigæ was similarly found to be double; and in 1890 Vogel, from photographs taken at Potsdam, independently inaugurated the discovery of spectroscopic binaries. In the spectrum of Spica he discovered the spectral lines to be, not doubled, but periodically displaced, indicating the existence of a dark or nearly dark companion, both stars revolving round their centre of gravity. Spica was seen to belong to the same class as Algol, only that in the case of Algol the plane of the satellite’s orbit passes through the Earth and eclipses the star, while in the case of Spica the orbit is inclined, and the star is constant in light. The line of research commenced by Vogel and Pickering was soon followed up by these investigators, as well as by Bélopolsky at Pulkowa, Campbell at the Lick Observatory, Slipher at the Lowell Observatory, and by _Edwin Brant Frost_ (born 1866), now director of the Yerkes Observatory, and his assistant, _Walter Adams_. In 1894 Bélopolsky discovered the duplicity of several variable stars, and in 1896 that of Castor, in Gemini. Late in 1896 Campbell undertook a systematic investigation of radial motions, and has since discovered about sixty spectroscopic binaries,—among them, in 1899, the Pole Star, and in 1900 Capella. The latter discovery was made independently by _Hugh Frank Newall_ (born 1857) at Cambridge, in England. It was found by Campbell that the revolution of the stars round their centre of gravity is performed in 104 days; and it soon became apparent that, owing to the large size of the orbit, the duplicity of Capella might be observed telescopically. At Greenwich the star was seen to be elongated, but at the Lick Observatory it was seen persistently single.