CHANGE IN WEIGHT DURING CHEMICAL REACTION. 
255 
whilst the contained solutions were being mixed a fine crack was developed across 
the lower portion of a sealed reaction vessel; this crack rendered the vessel useless 
for the purpose for which it was designed. Following a habit we have of retaining for 
the time being anything connected with our work, the vessel was placed inside a 
beaker and set aside. At the end of about eight weeks it was discovered that the 
whole of the liquid had left the vessel and passed out into the beaker; seeing that 
the vessel was still intact, it was obvious that a considerable pressure had been 
generated within; in no other conceivable way was it possible to account for the 
ejection of the liquid. 
An experiment was now devised to enable us to measure the growth in the 
pressure which follows the mixing of the two solutions. A flask was prepared and 
charged after the manner already described, but the capillary tube was made as 
uniform as possible and longer, and then bent so as to 
assume the form shown in fig. 4. A spot at the base 
of the capillary tube was softened before the blowpipe, 
and a short fine capillary tube, t, drawn out and left 
open ; it was now possible to introduce a convenient 
quantity of mercury into the U-shaped limb; this done, 
the vessel was placed for some hours in a thermostat, 
maintained at a mean temperature of 29°‘5 C., and 
then hermetically sealed by fusing the tips of the 
capillary tubes t and h ; the atmospheric pressure at 
the time of sealing was observed and recorded. Next, 
the solutions were mixed, the vessel replaced in the 
thermostat, and the sealed mercury manometer, m, 
read from time to time. During the first day the 
manometer remained unaffected, but at the expiration 
of the second day its indications showed that the pressure within had commenced 
to fall; the diminution in the pressure became more and more marked until the sixth 
day during which the change reached a maximum; from this time onward the 
pressure began to slowly return towards its original value, the increase during the 
four succeeding days being equal to 20 mm. of mercury. 
We next tried the effect of light upon the mixture, and in order to carry out this 
experiment the flask was removed from the thermostat, allowed to cool, and then 
placed together with a 25 candle-power “ tantalum” lamp inside a bright tin box, the 
top of which was then covered with tin-foil. Switching on the lamp for short inter¬ 
vals only, the pressure was now observed to increase so rapidly that it was deemed 
wise to terminate the experiment at the end of an hour. The flask was then replaced 
in the thermostat and the new pressure measured ; this pressure remained unaltered 
during the succeeding day, but 24 hours later it had still further increased. The 
experiment, which had by this time been proceeding for a fortnight, was now brought 
