368 Mr. R. H. M. Bosanquet on the Uluminating-poicer 



Dis. Gas and Benzene-Dis. Gas Mixture. 



Benzene. Dis. Gas. x. y. 



100 -1415 7-33 



3-19 96-30 -2013 23*54 



100 1 240 (calc.) 



If we estimate the error to be expecte J , by analogy from 

 the ethylene, at 1/7, this would give 274. But there is a 

 probability that the benzene molecule may be but little 

 attacked by absorption, in which case the straight line would 

 hold more nearly than in the case of ethylene. So that the 

 value may probably be between 240 and 274, and the number 

 in the last paper (255, p. 125) may not be far wrong. But I 

 shall use the number obtained from the new scheme of 

 illuminating values, 262*4. Whence we have : — 



Heavy Hydrocarbon mixture. Benzene and Ethylene. 



x. y. 



Benzene 1 262*4 



Ethylene *5 68*5 



Hydrocarbons ^ 



H = 4*32 V *690 142*4 (calc.) 



C=2*98 ) 



The number obtained in the former paper was 177. The 

 rectilinear method seems likely to hold here, as there cannot 

 be any absorption, and there seems to be no change of solvent 

 capacity. The next step is to draw the line from the dis. gas 

 point to the point which represents the hydrocarbon mixture. 



Total Gas (omitting the oxygen) . 



Hydrocarbons *690 142*4 



Dis. Gas -1415 7*33 



Total Gas -1782* 16*6 (calc.) 



Here there is likely to be some absorption, and the calculated 

 value ought to be reduced, for it lies where the true locus 

 would fall below the straight line. But it is not probable that 

 the correction would be so great as in the case of the ethylene 

 mixtures ; we have already observed on the probability of the 

 smaller liability of the benzene molecule to absorption. The 

 lower ethylene mixture falls below the corresponding straight 

 line in this manner by 2 candles. If we suppose the excess 

 of 16*6 over 16 to represent the correction required in the 

 case of the gas, it would be 1/3 of the correction for the 

 corresponding ethylene mixture, which seems likely enough. 



To complete the second approximation properly, we need 



* Calculated from the analysis. 



