minosity by the partial eclipse of a bright star by a dark companion.

The whole mass of the system of Algol is found to be half as great again
as that of our sun, yet the two bodies complete their orbit in the short
period of 2d 20h 48m 55s. The light remains constant during each period,
except for 9h 20m when it exhibits a considerable fall in brightness
(Clerke, "Problems of Astrophysics" page 302 and chapter XVIII.); the
curve which represents the variation in the light is shown in a figure
titled "The light-curve and system of Beta Lyrae" (Fig. 7.).

The spectroscope has enabled astronomers to prove that many stars,
although apparently single, really consist of two stars circling around
one another (If a source of light is approaching with a great velocity
the waves of light are crowded together, and conversely they are spaced
out when the source is receding. Thus motion in the line of sight
virtually produces an infinitesimal change of colour. The position
of certain dark lines in the spectrum affords an exceedingly accurate
measurement of colour. Thus displacements of these spectral lines
enables us to measure the velocity of the source of light towards or
away from the observer.); they are known as spectroscopic binaries.
Campbell of the Lick Observatory believes that about one star in six is
a binary ("Astrophysical Journ." Vol. XIII. page 89, 1901. See also A.
Roberts, "Nature", Sept. 12, 1901, page 468.); thus there must be many
thousand such stars within the reach of our spectroscopes.

The orientation of the planes of the orbits of binary stars appears to
be quite arbitrary, and in general the star does not vary in brightness.
Amongst all such orbits there must be some whose planes pass nearly
through the sun, and in these cases the eclipse of one of the stars by
the other becomes inevitable, and in each circuit there will occur two
eclipses of unequal intensities.

It is easy to see that in the majority of such cases the two components
must move very close to one another.

The coincidence between the spectroscopic and the photometric evidence
permits us to feel complete confidence in the theory of eclipses. When
then we find a star with a light-curve of perfect regularity and with
a characteristics of that of Algol, we are justified in extending the
theory of eclipses to it, although it may be too faint to permit of
adequate spectroscopic examination. This extension of the theory secures
a considerable