ELEMENTARY COMPOSITION OF NITROGLYCERINE. 



89 



broken, and a stream of oxygen was passed through to ensure the complete 

 combustion of the carbon. In our first experiment the nitroglycerine was 

 weighed out in a short glass tube, and this was dropped into the combustion 

 tube ; but an explosion occurred at an early stage of the ignition, which, 

 although damaging the furnace and slightly injuring one of us who happened 

 at the moment to be within a few inches of the tube, satisfied us that the 

 explosive force of the quantity of material employed (0*23 grm.) was not so 

 great as to prevent our continuing the experiments with the adoption of very 

 ordinary precautions. It was next attempted to burn the nitroglycerine in the 

 form of dynamite, using pure and previously ignited Kieselguhr as the 

 absorbent ; but this also gave rise to an explosion, though of greatly diminished 

 violence. Ultimately it was found that the combustion could be performed 

 without any risk of explosion by adopting the following method. A quantity 

 of the liquid (from "2 to "4 grm.) was weighed out in a porcelain boat con- 

 taining finely-divided copper oxide, and was then covered with another layer of 

 the oxide. The boat was dropped into the combustion tube, and its contents 

 were scraped out and well mixed up with the granulated copper oxide by 

 means of a long bent copper wire. The tube was then filled up in the 

 customary manner. The chief difficulty attending this method is to avoid the 

 introduction of moisture by the copper oxide and consequent raising of the 

 hydrogen percentage. The precautions taken against this were increased in 

 each experiment, so that the last hydrogen determination is probably the most 

 reliable, while the first is considerably too high. All the combustions were 

 conducted with unusual slowness. The same means was employed for filling 

 the tube in the nitrogen determinations by Dumas's method. 





Calculated. 



Found. 







I. 



II. 



in. 



IV. 



v. • 



L 3 , 



15-86 



16-05 



15-89 



15-80 





• . • 



H 5 , . 



2-20 



2-99 



2-40 



2-08 







N„ 



18-50 









17-93 



17-97 



9) . . . 



63-44 



... 



... 



... 



... 



... 



Two nitrogen determinations were also made with the same nitroglycerine 

 before it had been placed in vacuo, and the results, though slightly higher, 

 hardly warrant the belief that there was any difference in the composition. 

 They were (1) 18-25 and (2) 18-06. 



We consider that the above figures prove nitroglycerine to be glyceryl 

 tri-nitrate, the slight deficiency of nitrogen being possibly due to the presence 



