oft from the photosphere with speeds so terrific, or
the conditions which bring about the downrush of such gigantic masses of
vapour from above. In the spectra of the prominences on the sun's limb
also we often see the bright lines bent or shifted to one side. In such
cases what we witness is evidently caused by movements along the surface
of the chromosphere, conveying materials towards us or away from us.

An interesting application of this beautiful method of measuring the
speed of moving bodies has been made in various attempts to determine
the period of rotation of the sun spectroscopically. As the sun turns
round on its axis, a point on the eastern limb is moving towards the
observer and a point on the western limb is moving away from him. In
each case the velocity is a little over a mile per second. At the
eastern limb the lines in the solar spectrum are very slightly shifted
towards the violet end of the spectrum, while the lines in the spectrum
of the western limb are equally shifted towards the red end. By an
ingenious optical contrivance it is possible to place the spectra from
the two limbs side by side, which doubles the apparent displacement, and
thus makes it much more easy to measure. Even with this contrivance the
visual quantities to be measured remain exceedingly minute. All the
parts of the instrument have to be most accurately adjusted, and the
observations are correspondingly delicate. They have been attempted by
various observers. Among the most successful investigations of this kind
we may mention that of the Swedish astronomer, Duner, who, by pointing
his instrument to a number of places on the limb, found values in good
agreement with the peculiar law of rotation which has been deduced from
the motion of sun-spots. This result is specially interesting, as it
shows that the atmospheric layers, in which that absorption takes place
which produces the dark lines in the spectrum, shares in the motion of
the photosphere at the same latitude.

[Illustration: Fig. 20.--View of the Corona (and a Comet) in a Total
Eclipse.]

[Illustration: PLATE V.

TOTAL SOLAR ECLIPSE, JULY 29TH, 1878.

THE CORONA FROM THE PHOTOGRAPHS.

(HARKNESS.)]

We have yet to mention one other striking phenomenon which is among the
chief attractions to observers of total eclipses, and which it has
hitherto not been found possible to see in full daylight. This is the
corona or aureole of light which is suddenly seen to surr