nitude, and when least bright
is reduced to below the third magnitude; from which it follows that its
light, when at the minimum, is only about one-third of what it is when
at the maximum. It seems definitely proved by means of the spectroscope
that variables of this kind are merely binary stars, too close to be
separated by the telescope, which, as a consequence of their orbits
chancing to be edgewise towards us, eclipse each other in turn time
after time. If, for instance, both components of such a pair are bright,
then when one of them is right behind the other, we will not, of course,
get the same amount of light as when they are side by side. If, on the
other hand, one of the components happens to be dark or less luminous
and the other bright, the manner in which the light of the bright star
will be diminished when the darker star crosses its face should easily
be understood. It is to the second of these types that Algol is supposed
to belong. The Algol system appears to be composed of a body about as
broad as our sun, which regularly eclipses a brighter body which has a
diameter about half as great again.

Since the companion of Algol is often spoken of as a _dark_ body, it
were well here to point out that we have no evidence at all that it is
entirely devoid of light. We have already found, in dealing with
spectroscopic binaries, that when one of the component stars is below a
certain magnitude[33] its spectrum will not be seen; so one is left in
the glorious uncertainty as to whether the body in question is
absolutely dark, or darkish, or faint, or indeed only just out of range
of the spectroscope.

It is thought probable by good authorities that the companion of Algol
is not quite dark, but has some inherent light of its own. It is, of
course, much too near Algol to be seen with the largest telescope. There
is in fact a distance of only from two to three millions of miles
between the bodies, from which Mr. Gore infers that they would probably
remain unseparated even in the largest telescope which could ever be
constructed by man.

The number of known variables of the Algol type is, so far, small; not
much indeed over thirty. In some of them the components are believed to
revolve touching each other, or nearly so. An extreme example of this is
found in the remarkable star V. Puppis, an Algol variable of the
southern hemisphere. Both its components are bright, and the period of
light variation is about one and