122 Dulong and Petit on the Measure of Temperatures, [Fus. 
air is ascertained, whence, by a very simple calculation, the 
temperature of the air thermometer is deduced. It is searcely 
necessary to add, that the tube had been carefully dried, and 
that for each measurement the correction arising from the 
capillary depression was made. ‘ 
The indications of this thermometer add nothing to the preci- 
sion of those furnished by the mercurial thermometer. But we 
took that opportunity of again comparing the two thermometers. 
The results deduced from this comparison entered into the 
determination of the means inserted in Table I. 
It remains now to describe the kind of micrometer which 
we employed to measure the height of the columns. This mstru- 
ment (fig. 4) is composed of a thick copper rule, A B, along 
which moves stiffly, but smoothly, a piece of copper, MN.PRS, 
carrying at its two ends, M and 8, two rings, in which a micro- 
meter telescope, O O’, turns, furnished at its focus with a 
horizontal wire. From the telescope is suspended a very 
sensible level, the graduated tube of which serves to regulate 
the optical axis. This piece of copper, M N P RS, is susceptible 
of two movements, one very rapid, by unscrewing the lateral 
screw, C; the other very gentle, by turning the adjusting screw, 
D. The whole instrument turns round a vertical axis, which 
rests upon a thick triangular plane of copper, furnished with a 
screw at each of its summits, 
The construction of this instrument enables us, as is evident, 
to measure the difference between the height of two columns, 
which are not situated in the same vertical. it is necessary for 
this, after having directed the glass to one of the points, to cause 
the axis to turn, in order to bring it im the azimuth of the other 
point. It is then raised or depressed the requisite quantity, 
which is measured on a scale engraven on the opposite face of - 
the rule, A B, by means of a vernier moved by the piece, 
MNPRS. A micrometer serew would probably have been 
preferable had it not been for the rapidity which our experiments 
required. The vernier enabled us to appreciate the 50th of a 
millimetre, a degree of precision which we thought sufficient. 
To give to this instrument all the requisite exactness, it was 
necessary that the smallest differences between the two heights 
should be appreciable ; and that in the passage from one obser- 
vation to another, the glass should preserve its horizontality, or 
at least that we should be able to appreciate the derangement. 
The first of these conditions was satisfied by giving the telescope 
a sufficiently high magnifying power ; and as for the second, the 
particular care with which the micrometer was made, the solidity 
of the support on which it rested, and which was independent 
of the rest of the apparatus, might have led us to consider it as 
satisfied. However, we measured beforehand for the distance 
at which the telescope pointed, to what difference of height 
would correspond a change of inclination equal to one degree 
