8 



DR. R. ANGUS SMITH ON THE 



It would appear as if he went in both cases slightly to 

 excess ; but it is not well to attempt to judge on this point. 



Dumas and Boussingault together and Brunner obtain 

 also less oxygen ; they used weights ; and they also are 

 men standing, like Regnault and Bunsen, in the foremost 

 rank. 





Oxygen 

 per cent. 





Paris 



20*8l0 



20-856 

 20784 

 20757 

 20773 

 20793 



20-8lI 



Dumas & Boussingault. 



M. Stas. 



M. Marignac. 



M. Brunner. 



Verver. 



Brussels 



Greneve 



Bern 



Taulhorn 



Groningen 



Copenhagen 





In looking over the analyses already presented, some 

 of them means of hundreds, and the whole representing 

 many years of labour, we see at once how many give the 

 amount of oxygen to be above ao'g. As a rule those num- 

 bers which fall below 20*9 represent air from cities and less 

 pure places or from high mountains, or they have been ob- 

 tained by weighing the oxygen, a method which seems 

 always to give lower results. Take the conclusion of 

 Cavendish, wonderful at the time, that the average was 

 20*833, we are surprised at the accuracy of the man who 

 used a method by which no one now seems able to obtain 

 any reliable results. It cannot be supposed to take from 

 the honour of Cavendish, if we add one-tenth of a per cent, 

 to his figures after sixty years of scientific activity has made 

 that apparatus the plaything of ])oys which in his time it 

 required a philosopher to handle. He used acid liquids ; 

 and how easy it is to lose a fraction of a per cent, every 

 man who has worked with gas must know and feel most 

 keenly. If, however, any one shall object to this reasoning, 

 we must put up against him Saussure, a man using the 

 same method, and we find his average much higher, viz. 



