ing region. Again, photographs in polarised
light attested the radiance of the outer corona to be in large measure
reflected, while that of the inner ring was original; and the inference
was confirmed by spectrographs, recording many Fraunhofer lines when the
slit lay far from the sun's limb, but none in its immediate vicinity. On
plates exposed by Mr. Dyson and Dr. Humphrys with special apparatus, the
coronal spectrum, continuous and linear, impressed itself more
extensively in the ultra-violet than on any previous occasion; and Dr.
Mitchell succeeded in photographing the reversing layer by means of a
grating spectroscope. Finally, Mrs. Maunder, at Mauritius, despite
mischievous atmospheric tremors, obtained with the Newbegin telescope an
excellent series of coronal pictures.[585]

The principles of explanation applied to the corona may be briefly
described as eruptive and electrical. The first was adopted by Professor
Schaeberle in his "Mechanical Theory," advanced in 1890.[586] According
to this view, the eclipse-halo consists of streams of matter shot out
with great velocity from the spot-zones by forces acting perpendicularly
to the sun's surface. The component particles return to the sun after
describing sections of extremely elongated ellipses, unless their
initial speed happen to equal or exceed the critical rate of 383 miles a
second, in which case they are finally driven off into space. The
perspective overlapping and interlacing of these incandescent outflows
was supposed to occasion the intricacies of texture visible in the
corona; and it should be recorded that a virtually identical conclusion
was reached by Mr. Perrine in 1901,[587] by a different train of
reasoning, based upon a distinct set of facts. A theory on very much the
same lines was, moreover, worked out by M. Belopolsky in 1897.[588]
Schaeberle, however, had the merit of making the first adequate effort
to deduce the real shape of the corona, as it exists in three
dimensions, from its projection upon the surface of the sphere. He
failed, indeed, to account for the variation in coronal types by the
changes in our situation with regard to the sun's equator. It is only
necessary to remark that, if this were so, they should be subject to an
annual periodicity, of which no trace can be discerned.

Electro-magnetic theories have the charm, and the drawback, of dealing
largely with the unknown. But they are gradually losing the vague and
intangible ch