1820.] Dr. Henry's Experiments on the Gas from Coal. 39 



The sums of the numbers thus theoretically obtained do not, it 

 is true, exactly correspond with the experimental ones ; but they 

 approach as nearly as, from the nature of the subject, can be 

 expected, the greatest disagreement (that in the diminution by 

 firing) not much exceeding -rrVth of the observed amount. 



In a similar manner we may explain the composition of the 

 hghter and less combustible products obtained at advanced 

 periods of the distillation. For example, a portion of the last 

 product of gas from cannel, distilled in a glass retort, gave the 

 foUowins; results : 



o 



Measures of gas 20 



Mixed with oxygen 30 = 28 pure oxygen + 2 azote 



Total 50 



Fired 22 



Washed with potash 18 = 14*7 oxygen + 3-3 azote 



13-3 oxygen spent. 



In this experiment, 1'3 more azote were found in the residuum 

 than can be traced to the oxygen employed. The combustible 

 gas was, therefore, only 18|- measures; the carbonic acid pro- 

 duced 4 5 the oxygen spent 13'3 ; and the diminution by firing, 

 28. The following supposed mixture will explain these facts : 



Measures of Take oxygen. Give carb. acid. Dim. by firinp. 



2 carb. hydr 4 2 4 



2 carb. oxide 2 2 2 



15 hydrogen 7^ — 22i- 



19 13i 4 28^ 



In this instance, the hypothetical constitution coincides even 

 more nearly with the facts than in the former case. It must, 

 indeed, be acknowledged that the explanation rests on hypothesis 

 only ; but it is on an hypothesis which is perfectly consistent 

 with a copious and increasing induction of facts, all tending to 

 establish a limitation to the proportions in which bodies combine ; 

 while the opposite explanation is at variance with this general 

 law of chemical union. 



«ary appanitus, which was fnund, from long disuse, (o have b;X(>.ne unfit for the 

 purpose. So far as respects thepractieal objects of this paper, the oinissii n is ef 

 no consequtnce. 



