398 



Dr. A. Scott, 



[June 16, 



These ends were more or less satisfactorily attained by the use of 

 the apparatus employed, which was entirely of glass with the excep- 

 tion of the junctions at H and 0. The gas generators contained only 

 small vol umes of gas, and could easily be exhausted by means of the 

 apparatus itself. It is evident that by filling A and B with mercury, 

 completely closing all the stopcocks, and then lowering M and open- 

 ing e, the air in the oxygen generator would be in great part drawn 

 into A ; on now closing e and raising M this air could be expelled by 

 opening /. By repeating this several times an almost perfect vacuum 

 could be produced. Before collecting the gas for the experiments, 

 after exhausting the air gas was evolved and exhausted, and this 

 again repeated. The gases were measured saturated with moisture, 

 and after measurement were expelled into G; from this they were 

 drawn into E and exploded, and the residue measured in a small tube 

 and analysed by explosion with either hydrogen or oxygen as re- 

 quired. The results of every experiment made are given in the 

 following table, from which it will be seen that in no case, even when 

 the maximum value is given to it, does the ratio exceed 2 vols, of 

 hydrogen to 1 vol. of oxygen, although in four cases it is exactly 

 2:1. 



The mean of the twenty-one experiments gives the ratio — 



1-9857 : 1 from Column E. 

 1-9941 : 1 „ F. 



Excluding experiments IV and VI, in which the gases contained much 

 impurity, we get the ratio — 



1-9897 : 1 from Column E. 

 1-9959:1 „ F. 



Taking experiments I, III, XV, and XVIII, in which the purest 

 gases were used, we get — 



1-9938 : 1 from Column E. 

 1-9964 : 1 „ F. 



or taking experiments I, III, XIV, XV, XVIII, and XX we get — 



1-9938 : 1 from Column E. 

 1-9967 : 1 „ F. 



Taking as the most probable ratio 1*994 : 1, and the density of 

 oxygen referred to hydrogen as 15*9627, we get the atomic weight of 

 oxygen as 16*01. 



The oxygen was in each of the first twenty experiments prepared 

 from potassium chlorate, and in the twenty-first from mercuric oxide 



