ior
to 1750. The results, published in 1878,[961] proved somewhat
perplexing. They tend, in general, to reduce the amount of acceleration
left unaccounted for by Laplace's gravitational theory, and
proportionately to diminish the importance of the part played by tidal
friction. But, in order to bring about this diminution, and at the same
time conciliate Alexandrian and Arabian observations, it is necessary to
reject _as total_ the ancient solar eclipses known as those of Thales
and Larissa. This may be a necessary, but it must be admitted to be a
hazardous expedient. Its upshot was to indicate a possibility that the
observed and calculated values of the moon's acceleration might after
all prove to be identical; and the small outstanding discrepancy was
still further diminished by Tisserand's investigation, differently
conducted, of the same Arabian eclipses discussed by Newcomb.[962] The
necessity of having recourse to a lengthening day is then less pressing
than it seemed some time ago; and the effect, if perceptible in the
moon's motion, should, M. Tisserand remarked, be proportionately so in
the motions of all the other heavenly bodies. The presence of the
apparent general acceleration that should ensue can be tested with most
promise of success, according to the same authority, by delicate
comparisons of past and future transits of Mercury.

Newcomb further showed that small residual irregularities are still
found in the movements of our satellite, inexplicable either by any
known gravitational influence, or by any _uniform_ value that could be
assigned to secular acceleration.[963] If set down to the account of
imperfections in the "time-keeping" of the earth, it could only be on
the arbitrary supposition of fluctuations in its rate of going
themselves needing explanation. This, it is true, might be found in very
slight changes of figure,[964] not altogether unlikely to occur. But
into this cloudy and speculative region astronomers for the present
decline to penetrate. They prefer, if possible, to deal only with
calculable causes, and thus to preserve for their "most perfect of
sciences" its special prerogative of assured prediction.

FOOTNOTES:

[Footnote 796: _Neueste Beytraege zur Erweiterung der Sternkunde_, Bd.
iii., p. 14 (1800).]

[Footnote 797: _Ibid._, p. 24.]

[Footnote 798: _Phil. Trans._, vol. xciii., p. 215.]

[Footnote 799: _Mem. Roy. Astr. Soc._, vol. vi., p. 116.]

[Footnote 800: _Month.