ON THE MECHANICAL EQUIVALENT OF HEAT. 
407 
Oi:e end of a piece of strong rubber tube was then fastened on the glass tube 
protruding from the flask, w'hile its other end was fixed to the vessel shown at A, 
which was open to the atmosphere. 
Fig. 11. 
Mercury was poured into the glass funnel at A, and it was raised till there was a 
solid column of mercury from the bottom of the flask to the surface in A. The water 
in the beaker was then heated by a Bunsen flame till it boiled. This boiling was 
continued during a whole day, the water in the beaker being replenished as required. 
By adjusting the level of the free surface of the mercury at A, any required 
pressure could be put on the vapour column which formed over the water in the flask 
neck and displaced some of the mercury from the bottom. Also, by suddenly raising 
the pressure, the vapour was compressed and cold mercury flow’ed down into the 
flask, condensing the vapour in the neck as it descended. By this means the water 
in the flask could be made to boil briskly for a few moments now and then, so as to 
facilitate the escape of the air. At the close of the day the levels of mercury and 
water wmre adjusted so as to give the requisite pressure on the vapour column. The 
length of this column was then measured, and knownng the diameters of the flask 
neck and tube, it was easy to calculate the volume of vapour. 
This was 2‘2 cubic inches. 
If this be reduced to a temperature of 32° and atmospheric pressure, the proportion 
of air by volume appears to be 1’6 per cent. 
This number is considerably less than the 2*5 per cent, already mentioned, but as 
