the periods before and after, no similarity can be traced between the
sun-spot curve and the wind-vane curve, and we infer that the similarity
during the period first mentioned was entirely accidental. This shows that
we must be cautious in accepting, from a limited amount of evidence, a
connection between two phenomena as real and established; for it may be
purely fortuitous. We may particularly remark that it is desirable to have
repetitions through several complete periods instead of one alone. It is
possible to reduce to mathematical laws the rules for caution in this
matter; and much useful work has already been done in this direction by
Professor Schuster of Manchester and others, though as yet too little
attention has been paid to their rules by investigators naturally eager to
discover some hitherto unthought-of connection between phenomena.

[Sidenote: Faculae follow spots and the chromosphere.]

With this example of the need for caution, we may return to phenomena of
which we can certainly say that they vary sympathetically with the
sun-spots. Roughly speaking, the whole history of the sun seems to be
bound up with them. Besides these dark patches which we call spots (which,
by the way, are not really dark but only less bright than the surrounding
part of the disc), there are patches brighter than the rest which have
been called faculae. With ordinary telescopes, either visual or
photographic, these can generally only be detected near the edge of the
sun's disc; but even with this limitation it can easily be established
that the faculae vary in number and size from year to year much in the same
way as the spots, and this conclusion is amply confirmed by the beautiful
method of observing the faculae with the new instrument designed by
Professor Hale of the Yerkes Observatory. With this instrument, called a
spectroheliograph, it is possible to photograph the faculae in all parts of
the sun's disc, and thus to obtain a much more complete history of them,
and there is no doubt whatever of their variation sympathetically with the
spots. Nor is there any doubt about similar variations in other parts of
the sun which we cannot see _at all_ with ordinary telescopes, except on
the occasions when the sun is totally eclipsed. Roughly speaking, these
outlying portions of the sun consist of two kinds, the chromosphere and
the corona, the former looking like an irregular close coating of the
ordinary sun, and the latter lik