CHANGE IN WEIGHT DURING CHEMICAL REACTION. 
238 
reduced for a standard temperature of 18° C. We now possessed all the information 
necessary for a first determination of the difference in the weights of vessels A 
and B. 
Leaving the positions of the vessels unchanged and the beam free to swing, we, 
somewhat later, effected, by the methods of reversal and vibration described al)ove, a 
second determination of the apparent difference in the weights of A and B. Following 
this plan, we usually obtained during the course of a day from 2 to 3 complete and 
perfectly independent comparisons of the relative masses of A and B. From these 
completed determinations, a mean difference value for the whole day was deduced. 
With the object of removing any possible ambiguity, the actual data together with 
the final results secured during one day are set out in Table I. on p. 232. 
Similarly, by using every possible combination of the two sets of reduced B.Ps., we 
obtain the differences tabulated below :— 
Table II. 
From experiments (1) and 
mgr. 
(2), A = B + 0-264 
mgr. 
- -004 
)) 
(1) 
yy 
(4) „ = „ 
•260 
- -008 
J) 
}» 
(1) 
yy 
(6) „ = „ 
•263 
- •OOS 
Differences 
yy 
(3) 
yy 
(2) „ = „ 
•274 
+ -004 
from the 
>5 
yy 
(3) 
yy 
D) ” = )> 
•270 
+ -002 
final mean 
yy 
(3) 
yy 
(6) „ = „ 
•273 
+ -005 
value. 
>> 
yy 
(5) 
yy 
(2) „ = „ 
•270 
+ -002 
jy 
yy 
(5) 
yy 
(A „ = „ 
•266 
- -002 
yy 
yy 
(5) 
yy 
(6) „ = „ 
•269 
+-001 J 
Final mean 
differences found 
on July 18 = 
•268 
± -004 
This table shows the maximum variations from the final mean value for the day to 
be approximately = ±'007 mgr.; and as the vessel and its contents weighed 127 gr. 
we observe that in this case, which is one of the most concordant of the whole series, 
the mean value was liable to a fluctuation the order of which was about 1 in 17,000,000 
parts. 
Proceeding according to the plan just explained and indicated in Tables I. and II., 
we obtained the 9 daily final mean differences in the weights of A and B set forth in 
Table III.; the maximum variations to which the several mean values were subject, 
and also the corresponding degrees of accuracy attained are likewise stated. There 
were in all 169 weighings of A against B before reversal, and 173 after. 
At 10.30 a.m. on July 25, the contents of A were mixed; 7 hours later the first of 
the new series of weighings was commenced and continued at intervals during the 
succeeding 5 days, after which the observations ceased. The daily final mean values 
for the difference in the weights of A and B were the following :— 
July 25, 26, 27, 28, 29, 30. 
Differences in milligrammes = 0’37, ‘28, ‘22, '35, ’37, '36. Mean = '325. 
VOL. CCXII.-A. 2 H 
