ach recurrence are reproduced _almost exactly_. Here, then, lies the
_crux_ of the situation. For it is quite evident that were the
conditions _exactly_ reproduced, the recurrences of each eclipse would
go on for an indefinite period. For instance, if the lapse of a saros
period found the sun, moon, and earth again in the precise relative
situations which they had previously occupied, the recurrences of a
solar eclipse would tend to duplicate its forerunner with regard to the
position of the shadow upon the terrestrial surface. But the conditions
_not_ being exactly reproduced, the shadow-track does not pass across
the earth in quite the same regions. It is shifted a little, so to
speak; and each time the eclipse comes round it is found to be shifted a
little farther. Every solar eclipse has therefore a definite "life" of
its own upon the earth, lasting about 1150 years, or 64 saros returns,
and working its way little by little across our globe from north to
south, or from south to north, as the case may be. Let us take an
imaginary example. A _partial_ eclipse occurs, say, somewhere near the
North Pole, the edge of the "partial" shadow just grazing the earth, and
the "track of totality" being as yet cast into space. Here we have the
beginning of a series. At each saros recurrence the partial shadow
encroaches upon a greater extent of earth-surface. At length, in its
turn, the track of totality begins to impinge upon the earth. This track
streaks across our globe at each return of the eclipse, repeating itself
every time in a slightly more southerly latitude. South and south it
moves, passing in turn the Tropic of Cancer, the Equator, the Tropic of
Capricorn, until it reaches the South Pole; after which it touches the
earth no longer, but is cast into space. The rear portion of the partial
shadow, in its turn, grows less and less in extent; and it too in time
finally passes off. Our imaginary eclipse series is now no more--its
"life" has ended.

We have taken, as an example, an eclipse series moving from north to
south. We might have taken one moving from south to north, for they
progress in either direction.

From the description just given the reader might suppose that, if the
tracks of totality of an eclipse series were plotted upon a chart of the
world, they would lie one beneath another like a set of steps. This is,
however, _not_ the case, and the reason is easily found. It depends upon
the fact that the saros do