OF ARTS AND SCIENCES. 225 



tare as we have been discussing, yet the loss or gain thus resulting is 

 usually of such a magnitude that it cannot be overlooked. 



Besides the one of which the details have been given, we made by 

 the chemical method two other determinations of the tare of the empty 

 globe in precisely the same manner as before described, and obtained 

 the values 2.5481 and 2.5475 respectively. In order to bring these 

 results into comparison with that obtained by the first determination, 

 or those obtained by the method of Reguault, we must add 9.3 m.g. 

 to each value. We then have for 



Tare of Empty Balloon. 



Regnault's Method. Chemical Method. 



No. 1 2.5573 2.5569 



No. 2 2.5572 2.5574 



No. 3 2.5574 2.55G8 



Average 2.55730 2.55703 



o 



These results are essentially identical, and the two methods confirm 

 each other. But having reached this important conclusion, which was 

 the main object of our investigation, we freely concede the preference 

 to Regnault's method. The difference between the results of the 

 two methods, as here exhibited, is insignificant ; but so long as the 

 validity and value of the correction for the contraction of the balloon 

 on exhaustion has been thus established, Regnault's method will be 

 preferred because it is practically more simple, and hence the results 

 are more accordant than those obtained by the chemical method first 

 described in this paper. • 



During the course of our investigation there were two marked 

 changes in the tare of our balloon. One of these has been de- 

 scribed. The other arose from the circumstance that the thin paraf- 

 fine (vaseline) which was most suitable for lubricating the stop-cocks 

 during the winter no longer kept the joints tight when the heat of 

 approaching summer rendered the material more liquid ; and it became 

 necessary to remove the stoppers and relubricate them, and it was the 

 excess of cerate then used which melted and ran into the tubulaturea 

 subsequently. 



There are then three separate values of the tare of the empty bal- 

 loon to be used in our further calculations, all determined by or de- 

 pendent upon Regnault's method, and all known with equal accuracy. 

 These values we shall distinguish by the letters A, B, and C, and we 

 shall designate by the same letters the experimental data or calculated 

 results into which these values enter as elements. 

 VOL. xxiv. (n. s. xvi.) 15 



