633 TRANS. ST. LOUIS ACAD. SCIENCE. 



My calorimeter was cylindrical, of the very thinnest brass, and 

 could contain about 30 grms. of water. It was supported on three 

 sharp metal pins inside of a water mantle, to avoid interference 

 from foreign radiation and conduction. Its water value together 

 with that of all accessories was 0.8 g. The water value of the 

 test tube used to contain the liquids examined was 0.2124 g. 

 The average quantity of substance used in the determinations was 

 5.6 g. The glass tube was held firmly clamped immersed in 

 an oil bath, which was itself in turn immersed in a water bath. 

 By this means the temperature was capable of being maintained 

 very constant, so that the fluctuations of the temperature of the 

 mercury amounted at most to several tenths of a degree during 

 an hour. The temperature of the mercury bath was always about 

 50° C. The usual time of the immersions of the glass tube was 

 15 minutes. The temperature of both the mercury bath as well 

 as the calorimeter could be determined to 0.01° C. with certainty. 

 My thermometers were exceptionally concordant in their regis- 

 trations. The corrections for influence of radiation, ^A T. were 

 made by Rcgnaidfs method. To make them as small as pos- 

 sible the method of Rumford was adopted ; the temperature 

 of the calorimeter was regulated so that at the end of an experi- 

 ment it stood as much above the temperature of the mantle as it 

 was below it at the beginning. The temperatures of the calorim- 

 eter were recorded from 20 to 20 seconds. 



The following is an estimate that I made of the maximum of 

 error that could possibly occur in these determinations : 



Special heat of substance 0.5229. 



Error in reading calorimeter 0.01° sp. H. 0.5208 diflF. = 21 



Error in ^ A T 0.005° " 0.5219 " = 10 



Error in reading mercury bath. . 0.1° " 0.5248 " =19 

 Error in water value of calorim- 

 eter 0.1° •' 0.5210 " =19 



Total error 69=1. i^/o 



This, it must be remembered, is the maximum error. 



In the following I give my results. The column headed n 

 shows the number of equivalents of water added consecutively to 

 one equivalent of the acid ; the column under ;/' shows the total 

 number of equivalents of water added ; that headed P shows the 

 percentage of acid in these various mixtures ; the column headed 



