of Barometrical Mensuration. 81 



The most simple way of considering the subject, in a general 

 point of view, appears to me to be that which was, I believe, first 

 suggested by Sir George Shuckburgh, {Phil. Trans. Vol. Ixvii. 

 p. 55Q.) namely, to make a comparison of the specific gravities 

 of mercury and air at a fixed temperature, and under a given 

 pressure, the foundation of the operation. In this manner we 

 calculate the height of a column of air, compared with any 

 given column of mercury of equal base, supposing it of equal 

 density throughout. The calculation of the gradual diminution 

 of density, which takes place, for equal ascents in the atmos- 

 phere, according to a geometrical progression, is made in the 

 usual manner, by means of logarithms. This latter calculation 

 may be deemed invariable under all circumstances ; the former 

 includes all the adventitious circumstances, and all the effects 

 of disturbing causes. 



The well-known accuracy of MM. Biot and Arago, assisted 

 by the nicety of modern instruments, has determined the relative 

 specific gravities of dry air and mercury at a temperature of 32°, 

 and under a pressure of 30.00 inches to be as 1 to 10.435. 

 The height of a column of air, therefore, of equal density through- 

 out, which would balance a column of mercury of 30 inches under 

 these conditions, would be very nearly 26.090 feet. 



Now these proportions may be disturbed in two ways by the 

 operation of heat. In the first place, its expansive power 

 acting upon the mercury may dilate or contract its particles, 

 so that a column of 30 inches, being more or less dense, will 

 require an equipoise of greater or less length, according as its 

 temperature is below or above the standard at 32°. This effect 

 has been most minutely appreciated, and its correction is applied 

 with the utmost ease and precision. In the second place, the 

 power of heat acting upon the air occasions a much more im- 

 portant dilatation or contraction of its parts, and gives rise to 

 much greater differences in the height of the equiponderant 

 column. The expansion of air has been determined with pre- 

 cision by the experiments of M. Gay Lussac, and from them we 

 infer, with confidence, that it increases or diminishes ^|^^ part 

 for every addition or subtraction of 1° of heat. 



Vol.. XIII. G 



