s extend is included in the time through which calculations of
eclipses have been made in the French work entitled _L'Art de verifier
les Dates_. I have several times had occasion to recalculate with great
accuracy eclipses which are noted in that work (edition of 1820), and I
have found that, to the limits of accuracy to which it pretends, and
which are abundantly sufficient for the present purpose, it is perfectly
trustworthy. I have therefore made a comparison of the _Chun-Tsew_
eclipses with those of _L'Art de verifier les Dates_. The result is
interesting. Of the 36 eclipses, 32 agree with those of the _Art de
verifier les Dates_, not only in the day, but also in the general track
of the eclipse as given in the _Art de verifier_, which appears to show
sufficiently that the eclipse would be visible in that province of China
to which the _Chun-Tsew_ is referred." Airy then proceeds to point out
that, with regard to the four eclipses which he could not confirm, there
cannot have been eclipses in April 645 B.C. or in June 592 B.C. It
appears, however, from a note by Williams, that the date attached to the
eclipse of 645 B.C. is, in reality, an erroneous repetition (in the
Chinese mode of expressing it) of that attached to the next following
one, and in the absence of correct date it must be rejected. In the
record of 592 B.C., June 16, no clerical error is found, and there must
be an error of a different class. The eclipses of 552 B.C., September
19, and 549 B.C., July 18, to which there is nothing corresponding in
the _Art de verifier_, are in a different category. These occur in the
lunations immediately succeeding 552 B.C., August 20, and 549 B.C., June
19, respectively, and there is no doubt that those which agree with the
_Art de verifier_ were real eclipses. Now there cannot be eclipses
visible at the same place in successive lunations, because the
difference of the Moon's longitudes is about 29 deg., and the difference of
latitudes is therefore nearly 3 deg., which is greater than the sum of the
diameters of the Sun and Moon increased by any possible change of
parallax for the same place. These, therefore, were not real eclipses.
It seems probable that the nominal days were set down by the observer in
his memorandum book as days on which eclipses were to be looked for.
Airy conjectured that the eclipses of 552 B.C., August 20, and 549 B.C.,
June 19, were observed by one and the same person, and that he possessed
s