224 PROCEEDINGS OF THE AMERICAN ACADEMY 



was easy, however, to determine the exact loss, and as in a like inves- 

 tigation, even when using the greatest care, one must ever be liable to 

 accidents of this sort, it is important to dv?ell on this point. 



Ill losing the tare we had apparently lost the thread of our inves- 

 tigation, but we were readily able to recover this thread without inter- 

 ruptiug the general course of the work. It must be borne in mind 

 that our immediate object was to obtain by the chemical method a 

 series of results for the tare of the empty globe which could be com- 

 pared with the mean value 2.5573 obtained by the method of Regnault. 

 Here had come a change of tare, and we could obviously find a new 

 standard of comparison by redetermining the new value by the old 

 method. But this was not necessary. In the first determination by 

 the chemical method, we had observed with great precision the pres- 

 sure and temperature at which the balloon was filled. We do the 

 same in the second determination, to which we next proceed, although 

 these values are not needed in this determination, by itself considered, 

 any more than they were in the first. If now from the tare of bal- 

 loon and gas in the first determination (11.5907) we deduct 2.5573, 

 we have the weight of carbonic acid which the balloon contains at a 

 known temperature and pressure as found by the method of Regnault. 

 If we deduct the same quantity from the tare found in the second de- 

 termination, we have the weight of gas which the balloon holds at 

 another temperature and pressure by the same method. If now there 

 has been no change of tare, these two weights ought to agree exactly 

 when one is reduced to what it would have been under the conditions 

 at wliich the otlier was taken ; and if there has been a change of tare, 

 the difference of the weights thus reduced will give the exact amount 

 of the loss, on the basis solely, let it be noticed, of results obtained 

 after the method of Regnault. Thus we have, — 



Grams. 



By No. 1, weight of CO^ at 763.85 mm. and 25°.08 9 0273 



" No. 2, " " 756.73 " " 22°.40 9.0334 



Weight No. 1 reduced at " " " " 9.0241 



Loss of tare 9.3 m. g. 



Obviously the accuracy of this result depends upon the accuracy 

 with which the observations of temperature and pressure were made, 

 and upon the correctness of the data on which the reductions are based. 

 That the accuracy is extreme will appear from a comparison of the 

 values of the specific gravity of carbonic acid, given on page 229, which 

 were deduced from the same elements. And although in such work 

 as this there is an obvious liability to error from just such a change of 



