and on the refractive Powers of different Gases. 183 
his instruments and his method still exist, and the authors 
of the memoir availed themselves of them; adding thereto 
such improvements as were suggested by their own genius 
and the progress of science. 
They employed the same prism as that of Borda. It is 
a brass tube, terminated by two planes of glass carefully 
luted to the tube, oblique to its axis, and making between 
them an angle of 143° 7’ 28” (sexagesimal degrees), deter- 
mined by processes which ascertained it within a few se- 
conds. This tube is sustained by a support with a stop- 
cock, by means of which the vacuum is made in its interior 
by an air-pump. It is furnished with a barometer commu- 
Nicating with the interior, and acting the part of a proof to 
ascertain the degree of vacuum made in it. In the results 
an account was kept of a very small error in the parallelisms - 
in the two surfaces of each of the glasses at the ends of the 
tube; an error which was discovered by observations, and 
which only rose to 16:6”; this quantity is constantly added 
to the angles observed ; it is but a very trifling portion of 
the total deviation, which rises to 362°6”, 
In order to determine this deviation the following process 
was resorted to:—The apparatus, being placed on the top of 
the Palace of the Senate (formerly the Luxembourg), imme- 
diately behind it was placed a repeating circle, the lower eye- 
glass af which was constantly pointed to the thunder-rod of 
the observatory, 1400 metres distant. The lower eye-glass 
was directed upon the same object, but traversing the prism ; 
and the visual ray was affeeted with all the deviation due to 
the refraction. The angular quantity of it was very exactly 
observed; then, after having half turned the prism upon 
its vertical axis (being on a horizontal plane), which this 
turning had doubled, it was again observed, and was found 
so great, that the bent ray passed from one extremity to the 
other of the facade of the observatory. This whole angle 
was observed a great number of times, in order to take ad- 
vantage of the characteristic property of the repeating circle; 
and thus a perfect exactitude was obtained. An account 
was always kept of the variations of the barometer and of 
the hygrometer during the operations. 
The 
