and on the refractive Powers of different Gases. 155 
gramme. The mercury was weighed with great precaution, 
in order to compare its weight with that of water; and set- 
ting out with that of the air, a relation upon which several 
useful results rest, and particularly the measure of heights by 
the barometer, the metal was introduced into a matrass with 
a straight beak, aud from which all air was carefully expelled 
by the ebullition of the mercury, and by afterwards putting 
it under the receiver of the air-pump. The same precautions 
were taken with the water; all the corrections were made in 
respect to the temperature and the density of the atmospheric 
air, as well as for the dilatation of the vessels ; and there was 
found for the final relation of the weight of the mercury and 
of the air, 2. e. for the constant coefficient to employ in the 
calculations of the measurement of heights by the barometer 
expressed in metres, the number 18332, at the freezing point 
and at the pressure of 0°76 of a metre. 
We know that this same coefficient may be obtained by 
a very different method, 7. e. by comparing the barometric 
heights observed in the open air in stations differently ele« 
vated, with the real differences of height of the same:stations 
determined by levelling or by trigonometrical measures.. De 
Lucs, Schuckburgh, Trembiey, and other men of science, 
have proceeded in this manner; but no person has followed 
up this part of the research, or fathomed the different ele- 
ments of it with more sagacity than M. Ramond, whose 
conclusions we shall shortly state. 
The coefficient indicated by M. Ramond as the most con= 
venient to apply to observations made in the open air is = 
18336, and differs only by four unities of the fifth order of 
decimals, from that which the authors concluded from their 
experiments in close vessels; an agreement which forms a 
presumption very favourable to the exactitude of this funda- 
mental determination. M.Ramond has also shown that 
this coefficient, introduced into the formula of M. de la 
Place, would give the heights of mountains in a manner so 
nearly corresponding, that the uncertainty which remains is 
within the limits of those errors of which the minutest and 
most accurate observations are susceptible. 
The specific gravity of each of the gases employed may 
6 be 
