ed at the Sun on a surface of water and saw
it distinctly. At the end when we found no part of the Sun was any
longer eclipsed, and that its disc appeared in the water as a complete
circle, its altitude was 12 deg. in the E., less the one-third of a division
of the instrument, which itself was divided to thirds of a degree. One
must therefore reduce the stated altitude by one-ninth of a degree,
leaving, therefore, the true altitude as 11 deg. 53' 20"." The skill and
care shown in this record shows that the Arab who observed this eclipse
nearly a thousand years ago must have been a man of a different type
from an ordinary resident at Bagdad in the year 1899. No description is
given of the instrument used, but presumably it was some kind of a
quadrant. It does not appear why some of the observations were made at
Bagdad and some at Cairo. The Bagdad observations commence with an
eclipse of the Sun on November 30, 829, and end with an eclipse of the
Moon on November 5, 933. The Cairo observations begin with an eclipse of
the Sun on December 12, 977, and end with an eclipse of the Sun on
January 24, 1004. These statements apply to the 25 observations which
Newcomb considered were trustworthy enough to be employed in his
researches, but he rejected three as imperfect.

I have broken away from the strict thread of chronological sequence in
order to keep together the notes respecting Arabian observations of
eclipses. Let us now revert to the European eclipses.

Under the date of A.D. 733, the _Anglo-Saxon Chronicle_ tells us that,
"In this year AEthelbald captured Somerton; and the Sun was eclipsed, and
all the Sun's disc was like a black shield; and Acca was driven from his
bishopric." Johnston suggests that the reference is to an annular
eclipse which he finds occurred on August 14, at about 81/4 h. in the
morning. In Schnurrer's _Chronik der Seuchen_ (pt. i., Sec. 113, p. 164),
it is stated that, "One year after the Arabs had been driven back across
the Pyrenees after the battle of Tours, the Sun was so much darkened on
the 19th of August as to excite universal terror." It may be that the
English eclipse is here referred to, and a date wrong by five days
assigned to it by Schnurrer. Humboldt (_Cosmos_, vol. iv. p. 384, Bohn's
ed.) reports this eclipse in an enumeration he gives of instances of the
Sun having been darkened.

On May 5, A.D. 840, there happened an eclipse of the Sun which, amongst
other effects, is said to ha