ALGOL  , the Arabic name (signifying " the Demon ") of 6 Persei, a

star of the second magnitude, noticed by G . Montanan in 1669 to fluctuate in brightness . John Goodricke established in 1782 the periodicity of its change in about 2d 21h, and suggested their cause in recurring eclipses by a large dark satellite . Their intermittent character prompted the supposition . The light of Algol remains constant during close upon 56 hours; then declines in 62 hours (approximately) to nearly one-fourth its normal amount, and is restored by sensibly the same gradations . The amplitude of the phase is 1.1 magnitude; and the absence of any stationary interval at minimum proves the eclipse to be partial, not annular . Its conditions were investigated from photometric data, by Professor E . C . Pickering in 188o;' and their realization was finally demonstrated by Dr H . C . Vogel's spectroscopic measures in 1889.2 Previously to each obscuration, the star was found to be moving rapidly away from the earth; its velocity then diminished to zero pari passu the loss of light, and reversed its.direction during the process of recovery . Algol, in fact, travels at the rate of 26.3 miles a second round the centre of gravity of the system which it forms with an invisible companion, while the two together approach the sun with an unvarying speed of 2.3 miles per second .

The elements of this disparate pair, calculated by Dr Vogel on the somewhat

precarious assumption that its dark and bright members-are of equal mean density, are as follows: Diameter of Algol . . . . r,o6i,000 English miles . Satellite . . . 834,300 Distance from centre to centre . 3,230,000 „ „ Mass of Algol s solar mass . Satellite . . 9 Mean density . about 1 solar . The plane of the joint orbit, in which no deviation from circularity has yet been detected, nearly coincides with the line of sight . The period of Algol, as measured by its eclipses, is subject to complex irregularities . It shortened fitfully by eight seconds between 1790 and 1879; soon afterwards, restoration set in, and its exact length in 1903 was 2d 2oh 48'n 56°, being only two seconds short of its original value .

By an exhaustive discussion, Dr S .

Chandler ascertained in 1888 the compensatory nature of these disturbances;3 and he afterwards found the most important among several which probably conspire to produce the observed effects, to be comprised in a period of 15,000 light-cycles, equivalent to 118 years.¢ An explanatory hypothesis, propounded by him in 1892,5 is still on its trial . The system of Algol, according to this view, is triple; it includes a large, obscure primary, round which the eclipsing pair revolves in an orbit somewhat smaller than that of Uranus, very slightly elliptical, and inclined 200 to the line of sight, the periodic time being 118 years . The alternate delay and acceleration of the eclipses are then merely apparent; they represent the changes in the length of the light-journey as the stars perform their wide circuit . If these suppositions have a basis of reality, the proper motion of Algol should be disturbed by a small, but measurable undulation, corresponding to the projection of its orbit upon the sky; and although certainty on the point cannot be attained for some years to come, Lewis Boss regarded the evidence available in 1895 as tending to confirm Dr Chandler's theory.° Proceedings Amer . Acad. vol. xvi. p . 27 . z Astr . Nach . No . 2947 . Asir .

Journal, No . 165 . a Ibid . No . 509 . 5 Ibid . Nos . 255-256 . 6 Ibid . No . 343 A rival interpretation of the phenomena it dealt with was put forward by F . Tisserand in 18951 It involved the action of no third mass, but depended solely upon the progression of the line of apsides in a moderately elliptical orbit due to the spheroidal shape of the globes traversing it .

Inequalities of the required sort in the returns of the eclipses would ensue; moreover, their duration should concomitantly vary with the varying distance from periastron at the times of their occurrence . It is a

moot question whether changes of the latter kind actually occur . When they are proved to do so, Tisserand's hypothesis will hold the field . Algol gives a helium-spectrum which undergoes no alteration at minimum . Hence the light from the marginal and central portions of the disc is identical in quality, and the limb can be little, if at all, darkened by the " smoke-veil " absorption conspicuous in the sun . The rays of this star spend close upon a century in travelling hither . Dr Chase's measures with the Yale heliometer indicated for it, in 1894, a parallax of about 0" .035 ; 2 and it must, accordingly, be of nearly four times the total brightness of Sirius, while its aerial lustre exceeds seventy-fold that of the solar photosphere . Variables of the Algol class are rendered difficult to discover by the incidental character of their fluctuations . At the end of 1905, however, about 37 had been certainly recognized, besides some outlying cases of in-determinate type, in which continuous occultations by two bright stars, revolving in virtual contact, are doubtfully supposed to be in progress . (A . M .