CHANGE IN WEIGHT DUEING CHEMICAL REACTION. 
247 
Let the temperature of a sphere having V = 150 c.c. be 0°'l C. greater than the 
mean temperature of the air within the balance case; we shall commit no serious 
error if we assume that the enveloping shell of air which is in close contact with and 
therefore readily warmed by the walls of the vessel is likewise 0°‘l C. greater than 
the mean temperature of the general body of air ; for theory shows that such a shell 
of warmer air, though thin, will be appreciable and fairly well defined. Taking the 
weights of 1 c.c. of dry air under ^ = 760 mm., and at the respective temperatures 
of 15°‘0 C. and 15°T C. as coj and 0)2 mgr., we find the dilference — 002 — '0005 mgr. ; 
the corresponding difference for 150 c.c. (the volume of the sphere) is therefore 
= "075 mgr. Under the conditions which existed during our final experiments it was 
estimated that the smallest change in weight that could be detected with certainty 
equalled '002 mgr.; and therefore '075 mgr. would, comparatively speaking, produce 
a very marked effect upon the R.P. value. 
In order to discover how far the temperature might be affected during the process 
of wiping, we introduced into a reaction flask, containing 150 c.c. of water, a 
thermometer graduated in C. ; holding the flask Ijy the upper end of tlie neck 
the contents were shaken and the temperature noted. A doubled silk or linen 
handkerchief was then employed in the usual manner and the flask wiped for the 
space of half a minute preparatory, as it were, to weighing, and the temperature again 
noted ; this experiment was repeated a number of times both with the silk and the 
linen ; in each case the results were surprisingly concordant. With silk the mean 
rise in the temperature was 0°'15 C. and with linen 0°'25 C. These experiments, 
therefore, show the correctness of the more or less prevailing idea that for the purpose 
of wiping glass vessels silk is superior to linen, the superiority being with reference 
to the resultant temperature changes and, according to our measurements, in the ratio 
of 25/15. 
If we assume an otherwise strictly constant temperature within the balance case 
and apply the Stefan-Boltzman law of cooling, it becomes evident that a very 
considerable period must elapse before it will be possible to weigh correctly the 
slightly warmed vessel and its contents. If the above line of argument followed for 
the assumed difference of 0°'l C. be extended so as to include the results obtained in 
the experiments upon the effects produced during wiping, it is found that the 
apparent increase in the weight of the vessel due to this cause alone is respectively 
0'113 and 0'188 mgr. ; in practice, however, other efiects such as those arising from 
the action of convection currents would modify and lessen these values. So far as I 
am aware, little or no attention has hitherto been given to the possible efiects 
following very slight differences in the temperature of a vessel and the general body 
of air surrounding it ; for our present purpose the influence exercised l)y even minute 
differences in the temperature of the air and that of the vessel may become highly 
important and must not be neglected. 
Having reviewed and indicated the sources and approximate magnitudes of at least 
