IN THE HALL OF THE ROYAL SOCIETY. 
547 
inches in the tube. The whole instrument has been inclosed in an exterior 
ornamental case resembling an architectural column. The pedestal (A, B) 
conceals the boiler with its brick-work, and upon the capital (C, D) stands 
a glass-case including that part of the tube to which the oscillations are con¬ 
fined, and the apparatus for measuring them. 
As much interest will attach to the accurate comparison of the water-baro¬ 
meter with the mercurial barometer, it is of great importance that several cor¬ 
rections should be attended to in the first reading of their respective heights, 
to reduce the columns to the same invariable circumstances under which alone 
such comparison can be properly made; for this purpose the variations of the 
density of the liquids, and the expansion of the scales, from variations of tem¬ 
perature, together with the capillary action of the tubes, must be taken into 
account. To facilitate this object, I have constructed the two following 
Tables of double entry; by which the observations may be reduced to the 
temperature of 40° (39°*38) or that of the maximum density of water, in which 
the expansion of the brass scales is also allowed for; which is a correction of 
considerable amount in the long scale of the water-barometer. 
The data upon which these Tables have been calculated are as follows: 
1st, The specific gravity of water at different temperature, as determined by 
the experiments of Hallstrom, taken from Dr. Thomson’s late work upon Heat 
and Electricity, p. 28. 
2nd, The determination of the linear expansion of brass at *0000104 per 
degree of Fahrenheit. 
The height of the column is assumed to be in inverse proportion to the spe¬ 
cific gravity; and the correction to the maximum density at 40° (or more cor¬ 
rectly 39°*38) is calculated accordingly. From this correction is deducted, or 
to it is added, the expansion or contraction of the brass scale on either side of 
60°, calculated on the preceding datum. 
