panion of Procyon has a mass more than half as great as that of our sun, and is, therefore, capable of exerting appreciable gravitational influence upon a body near it. But though it is such a ponderous globe, its light is much less than that which the sun would give if placed in the same position among the stellar host. It is therefore not surprising that the gravitational influence was detected before the faint glimmer of the rays from the star was seen and understood.

Three instances have now been mentioned in which invisible celestial bodies were found by indirect evidence before they were looked for with telescopic aid-we refer to the planet Neptune, and the companions of Sirius and Procyon. Predictions fulfilled in this way should encourage confidence in conclusions based upon similar premises. That is to say, if observations show that a star is not moving through space in a straight line, they afford presumptive evidence of the existence of one or more bodies near it. The problem thus resolves itself into one of studying stellar motions.

A comparison of the exact positions of stars year by year shows that every one has a motion of its own across the blue background of infinity. The amount of movement as seen from the earth is very minute, and can only be detected by accurate determinations of position, but it is none the less real, and has to be taken into account in precise astronomy. Every movement has a cause, and when deviations from a direct course are found, it is certain that the star showing them is being disturbed by a massive body which may or may not be visible.

But what of stars which are moving straight towards the earth or away from it? Such movements cannot produce any change of position upon the background upon which they are projected. This is true enough,

but they can be detected by other effects. To an astronomer with a spectroscope, the light of a star is a gamut of colour crossed by dark or light rays comparable with musical notes. Just as the pitch of a note can be raised or lowered by rapid motion of the sounding body towards or away from the listener, so the positions of rays in the light-scale are affected by similar movements of approach or recession. If the distance between the moving star and the earth is decreasing, the rays analysed by the spectroscope are increased in colour pitch, and if the distance is increasing the rays are moved towards the lower end of the gamut of light, or, expressed in the terms of music, their notes are flat. So perfect a means does the spectroscope provide of measuring the movement, back or forth, that the velocity of a star can be determined within a quarter of a mile a second, though the star itself may be at an immeasurable distance from us. Here, then, we are provided with another means of studying stellar motions, and to it we owe the proof of the existence of many dark stars.

That such dark stars existed was first suggested by John Goodricke, who, though deaf and dumb from birth, used his sight to such good effect that his name is renowned among astronomers. He died at the early age of twenty-two, yet his observations, made in a small observatory at York, obtained for him in 1783 the Copley Medal of the Royal Society, the highest honour which the Society can confer. Goodricke was the first to make a systematic study of the variations in brightness of the star Algol-a name derived from the Persian word signifying the "demon." This star shines steadily with a brightness equal to that of the Pole Star for nearly two and a half days, and then suddenly its light is reduced. In about four and a half hours the star's

VIII. APOTHEOSIS OF THE SCIENCES.

Ceiling painting by Paul Albert Besnard (1849-...),

Salon des Sciences, Hôtel de Ville, Paris.

brightness is diminished by about two-thirds, and three and a half hours later it has regained the former intensity, which continues unaltered for another fifty-nine hours. These variations succeed one another with clock-like regularity, so that the times when the brightness of Algol will fade can be tabulated years in advance, as they are in the Nautical Almanack and similar publications. A navigator or other traveller with such a table at hand can correct his watch by observing when the star dims in brightness, and comparing the tabulated time of this occurrence with the time shown by his chronometer.

To explain the sudden reduction in light of Algol, Goodricke suggested that a dark body is revolving around the star, and periodically comes between us and it, thus causing a partial eclipse in each revolution. As a star can never be seen as anything but a point of light, whatever telescopic means are employed, it is impossible to distinguish any outlines of a dark body upon a luminous disc, such as is seen, for instance, during a partial eclipse of the sun by the moon, but Goodricke's explanation was accepted, because it accounted for the observed variations. The proof of the hypothesis was not forthcoming for more than a century after Goodricke suggested it; and then it was reached by indirect methods. Assuming that Algol has a dark companion, the two bodies must swing round their common balancing point, or centre of gravity. When the dark body is moving towards us before passing in front of the bright globe, the latter must be swinging back; and when the dark companion is receding after the eclipse, the bright star must be approaching. Algol must, therefore, alternately recede and approach in a period which coincides with that of its changes of