ALGOL, a well known variable star, also designated Per sei. Its fluctuation of brightness was noticed by G. Montanari in 1669; but the Arabic name, which signifies the Demon, seems to hint that the behaviour may have been known in ancient times. The brightness varies regularly between magnitudes 2.1 and 3.2. In 1782 John Goodricke found that the variation of light was periodic and determined the period 2d2oh49m; he suggested that the variation might be due to partial eclipses of the star by a dark body revolving round it. This hypothesis was generally accepted but it remained speculative until 1889, when H. C. Vogel demon strated its truth in a way that can scarcely be doubted.
By spectroscopic measures of the velocity in the line of sight it was shown that the bright star was travelling round an orbit in the period named. Clearly there must be an unseen mass controlling its motion. Further the reduced light occurs when the bright star is most remote from us ; that is just the time when the dark com panion passes between us and the bright star and is able to hide it. About 200 variable stars of the Algol type are now known in which the change of light is due to eclipses. It should, however, be understood that this is not the sole kind of stellar variability; and in fact the majority of variable stars have a different nature (see STAR).
Algol remains of almost constant brightness for 59 hours ; the remaining ten hours of the period are occupied by the eclipse, the light gradually decreasing to not much more than a third of its normal amount and then increasing again. The eclipse is partial. (In some variables a total eclipse occurs, totality being distin guished by a period of steady light at minimum.) The compar atively long duration of the eclipse shows that we have to do with two stars whose dimensions are by no means small in comparison with the distance between them. Closer examination shows that the light is not absolutely constant during the 59 hours. The reason is that the dark component reflects (or re-radiates) light from the bright star, and thus goes through phases like the moon. Accordingly, after the eclipse is over, the total light still continues to increase a little because the "moon" is progressing from new to full. In the middle of the S9 hour interval the light dips again because there is an "eclipse of the moon"; that is to say, the bright star is then hiding its companion. The whole phenomenon is beau tifully shown in the light-curve determined by J. Stebbins by means of selenium photometry.
We have spoken of the dark companion. In most eclipsing vari ables this only means comparatively dark, and the proper light of the eclipsing star is recognizable. But in Algol the companion is exceptionally dark, and it has not been possible to measure any light from it other than the reflected light above-mentioned. On theoretical grounds the companion is believed to give out less than a hundredth of the light of the principal star; but it must be noticed that Algol itself is an exceptionally bright star and its com panion is actually not much fainter than our sun. From a study of the light-curve, combined with the dimensions of the orbit of the principal star obtained from spectrographic measures, a great deal can be learnt about the nature of the system ; but to render the results definite an additional constant is required to be known, and through ignorance of this the dimensions formerly given have turned out to be wide of the mark.
The required datum has now been supplied by D. B. McLaugh lin, who has measured the radius of Algol by an ingenious method. Near the middle of the eclipse only a crescent of the bright star is showing; this crescent will be mainly on one side of the star just before the middle phase, and on the other side after it. Now the rotation of the star causes one side to approach us and the other to recede ; so that if we make measurements of radial velocity shortly before and after mid-eclipse we can determine the speed of rotation, very much as the sun's speed of rotation is determined by measuring first on the east and then on the west side of its disc. Having found the equatorial speed of rotation, we multiply by the period of rotation, and the result is the circumference of the star. In Algol the two components are close together, and there can be no doubt that the huge tidal forces will make them keep the same faces turned towards each other, so that the period of rotation must be the same as the period of revolution, viz., 2d21h. Knowing the circumference we can at once find the radius The following dimensions of the Algol system have been deduced: Distance between centres io,2oo,000km.
Radius of bright star 2,140,000km.
Radius of faint star 2,45o,000km.
Mass of bright star 4.3o X sun's mass Mass of faint star o•86 X sun's mass Density of bright star o• 16 (water=1) Density of faint star 0•02 Luminosity of bright star I4oXsun Distance of system 35 Parsecs (115 light years) .
Confidence in these results is strengthened by two checks. First, the mass and luminosity are in agreement with the mass-lumi nosity relation which the stars are believed to obey (see STAR). Second, the luminosity is that which is usual for a star of the spectral type B 8, to which Algol belongs. It will be noticed that the fainter star has the larger radius. The distance of the system found by this method might be checked by a trigonometrical deter mination of parallax; but very accurate measurements would be required, and it is probable that the parallax measurements hitherto published are vitiated by perturbations due to a third component of the system.
By spectroscopic measures it is found that Algol with its satel lite is describing an orbit in 1.9 years; this indicates the existence of a third member of the system. It must be at a distance of 43o million kilometres from Algol. It is intermediate in brightness be tween Algol and its close companion, giving about a I2th of the light of the former. It takes no part in the eclipses.
BIBLIOGRAPHY.-J. Stebbins, Astrophysical Journal, 32, p. 185 Bibliography.-J. Stebbins, Astrophysical Journal, 32, p. 185 (191o) ; D. B. McLaughlin, ibid.; 6o, p. 22 (1924) ; A. S. Eddington, Stars and Atoms (1927) . (A. S. E.)