e a pearly halo of light extending to
many diameters of the sun's disc, but not with any very regular form.

[Illustration: PLATE XV. SMOOTHED SUNSPOT CURVE (WOLF) COMPARED WITH THE
NUMBER OF TURNS MADE IN EACH YEAR BY THE OSLER ANEMOMETER VANE OF THE
ROYAL OBSERVATORY, GREENWICH (THE EXCESS OF THE DIRECT TURNS (D) OVER THE
RETROGRADE TURNS (R) OR _VICE VERSA_.)

THE UPPER CURVE IS IN EACH CASE THE SUNSPOT CURVE, THE LOWER THE VANE
CURVE. THE BREAK IN 1882 IN THE VANE CURVE IS DUE TO THE OMISSION OF
EVIDENTLY ACCIDENTAL TURNS FROM THAT DATE.]

The chromosphere, from which shoot out the prominences or "red flames,"
can now be observed without an eclipse if we employ the beautiful
instrument above-mentioned, the spectroheliograph; and Professor Hale has
succeeded in photographing spots, faculae, and prominences all on the same
plate. But although many have made the attempt (and Professor Hale,
perhaps, a more determined attempt than any man living), no one has yet
succeeded in obtaining any picture or evidence of the existence of the
corona excepting on the occasion of a total solar eclipse.

[Sidenote: Eclipses of sun.]

[Sidenote: Total eclipses rare.]

Now these occasions are very rare. There are two or three eclipses of the
sun every year, but they are generally of the kind known as partial; when
the moon does indeed come between us and the sun to some extent, but only
cuts off a portion of his light--a clean-cut black disc is seen to
encroach more or less on the surface of the sun. Most of us have had an
opportunity of seeing a partial eclipse, probably more than once; but few
have seen a total eclipse. For this the moon must come with great
exactness centrally between us and the sun; and the spot where this
condition is fulfilled completely only covers a few hundred miles of the
earth's surface at one moment. As the earth turns round, and as the moon
revolves in its orbit, this patch from which the sun is totally eclipsed
travels over the earth's surface, marking out a track some thousands of
miles in length possibly, but still not more than 200 miles wide; and in
order to see the sun totally eclipsed even on the rare occasions when it
is possible at all (for, as already remarked, in the majority of cases the
eclipse is only partial), we must occupy some station in this narrow belt
or track, which often tantalisingly passes over either the ocean or some
regions not easily accessible to civilised man. Moreover,