ble.
In the case of a Solar eclipse it may be total, perhaps, in Africa, may
be of six digits only in Spain, and of two only in England. Under the
most favourable circumstances the breadth of the track of totality
across the Earth cannot be more than 170 miles, and it may be anything
less than that down to zero where the eclipse will cease to be total at
all, and will become annular. The question whether a given eclipse shall
exhibit itself on its central line as a total or an annular one depends,
as has been already explained, on the varying distances of the Earth and
the Moon from the Sun in different parts of their respective orbits.
Hence it follows that not only may an eclipse show itself for several
Saros appearances as total and afterwards become annular, and _vice
versa_, but on rare occasions one and the same eclipse may be annular in
one part of its track across the Earth and total in another part, a
short time earlier or later. This last-named condition might arise
because the Moon's distance from the Earth or the Sun had varied
sufficiently during the progress of the eclipse to bring about such a
result; or because the shadow just reaching the Earth and no more the
eclipse would be total only for the moment when a view perpendicular
upwards could be had of it, and would be annular for the minutes
preceding and the minutes following the perpendicular glimpse obtained
by observers actually on the central line. The eclipse of December 12,
1890, was an instance of this.

If the paths of several central eclipses of the Sun are compared by
placing side by side a series of charts, such as those given in the
_Nautical Almanac_ or in Oppolzer's _Canon_, it will be noticed that the
direction of the central line varies with the season of the year. In the
month of March the line runs from S.W. to N.E., and in September from
N.W. to S.E. In June the line is a curve, going first to the N.E. and
then to the S.E. In December the state of things is reversed, the curve
going first to the S.E. and then to the N.E. At all places within about
2000 miles of the central line the eclipse will be visible, and the
nearer a place is to the central line, so much the larger will be the
portion of the Sun's disc concealed from observers there by the Moon.
If the central line runs but a little to the N. of the Equator in Winter
or of 25 deg. of N. latitude in Summer, the eclipse will be visible all over
the Northern Hemisphere, and the co