CHANGE IN A^HIIGHT DURING CHEMICAL REACTION. 
253 
out the cooling experiment, the vessel, loosely corked and fitted with a thermometer 
placed centrally within, was first warmed and the temperature of its contents raised 
about 6° C. above that of the surrounding air ; it was next placed upon a light three- 
point support of fire clay and surrounded with one of the copper cylinders from the 
balance case ; this done, the top of the cylinder was covered with a tin-foil lid through 
which the stem of the thermometer projected ; finally, one of the tubes leading from 
the air-purifying apparatus was passed downwards and just through the tin-foil lid, 
and the whole covered with a large glass dome. The time required for the vessel and 
its contents to cool from 21° C. to 16°‘2 C. was then determined, a chronometer being 
for this purpose read for each fall of 0°‘5 C. until the temperature had sunk to 16°'5 C. ; 
as the rate of cooling had then become very slow, chronometer readings were 
recorded for three succeeding temperature steps each of 0°1 C. 
For this experiment we purposely chose a day when the weather was dull and the 
air fairly still ; under those conditions the temperature of our balance room is 
remarkably constant; in this present instance the variations kept within the limits 
15°'2 C. and 15°’4 C. during the 6 hours the experiment lasted. The results derived 
from this experiment were now set forth in the form of a curve (No. 3) of dimensions 
corresponding to those chosen for Curve No. 2 ; we were thus enabled to institute a 
direct comparison between the two. 
Superposing the cooling and decrease in weight curves, it is found that the former 
very approximately coincides with the axis of the undulations of the latter; this very 
approximate coincidence appears to persist and improve until finally the two curves 
merge into one and so become indistinguishably the same. The portion a/3 of the 
change in weight curve was obtained by extrapolating, and the point marked y 
indicates the result obtained during our final weighing in this portion of our 
experiment. 
The results we have just given and briefly discussed aflbrd strong confirmatory 
evidence of the correctness of the theory already advanced, and which indicates that 
for a body of some considerable volume the apparent mass, as determined within a 
very restricted air space free from convection currents, must appreciably increase 
when the temperature of the body is raised, and decrease when the temperature is 
allowed to fall. 
With regard to the undulations wliich appear in the first portion of the Curve 
No. 2, it must be confessed that we are not yet in possession of sufficient data to be 
able to definitely assign a cause from which they may arise; we may, however, state 
that the results of several experiments led to the conclusion that the undulations 
will be inappreciable unless the temperature of the charged vessel which is being 
weighed exceeds that of the air by 1° C. or more. The more or less harmonic form 
of the undulations suggest two possibilities : (l) for a given temperature diflerence. 
periodic oscillations may be set up in the air shell which encloses the warm vessel, (2) 
the transference of the heat energy from the interior to the bounding surface of the 
