138 Mr. Harold B. Dixon on 



question, on which there is Httle experimental evidence. 

 If, for instance, in the combustion of a hydrocarbon, or of 

 cyanogen, the carbon is first burnt to carbonic oxide, which 

 subsequently is burnt to carbonic acid, the rate of the 

 explosion-wave should correspond with the carbonic oxide 

 reaction, in this case the primary reaction ; whereas, if the 

 carbon of these gases burns to carbonic acid directly, in one 

 stage, then the rate of the explosion-wave should correspond 

 with the complete reaction. 



Now, if we adopt Berthelot's formula as a working 

 hypothesis, we can calculate the theoretical rates of explo- 

 sion of marsh gas, ethylene, or cyanogen : (i) on the 

 supposition that the carbon burns directly to CO2, and (2) 

 on the supposition that the carbon burns first to CO, and 

 the further oxidation is a subsequent or secondary reaction. 

 On the first supposition, if 100 represents the rate of explo- 

 sion of these three gases burning to carbonic oxide, the 

 addition of the oxygen required to burn the gases to 

 carbonic acid should increase the rate of explosion : — 



Marsh Gas. Ethylene. Cyanogen. 



Calculated rate of explosion 



when burnt to CO- 



ion I 



104 103 107 



Whereas if these gases always burn first to carbonic oxide, 

 and the extra oxygen is inert in propagating the explosion- 

 wave, then the addition of this inert oxygen would diminish 

 the rate of explosion : — 



Marsh Gas. Ethylene. Cyanogen. 



Calculated rate of explosion when burnt | qq o_ 



to LU with inert oxygen present ... J ^ 



The experiments show that if 100 be taken as the rate of 

 explosion when the oxygen is only sufficient to burn the 

 carbon to carbonic oxide, the following are the rates found 

 when oxygen is added sufficient to burn the carbon to 

 carbonic acid : — 



