1819.] the different Species of Pit-Coal. 89 



3. Constituents of each Species of Coal. 



The method which I employed to analyze the different species 

 of coal was this. I reduced one grain of the coal to a fine pow- 

 der, and mixed it intimately with 140 gr. of peroxide of copper 

 in a fine powder, which had been subjected to a red heat in a 

 crucible, and had been kept in a well-stopped glass phial. This 

 mixture was put into a copper tube, and occupied a portion of it 

 amounting to about four inches in length at the upper end of the 

 tube. The copper tube was then filled with peroxide of copper. 

 I first luted to the end of the copper tube a glass tube of the 

 shape and size described in the Annals of Philosophy, xii. 109, 

 having previously filled it with dry muriate of lime in powder. 

 That part of the copper tube containing the mixture of coal and 

 peroxide of copper was then heated to redness, and kept in that 

 state till all extrication of gas was at an end. The gas was 

 collected over mercury, its quantity was estimated, and its 

 nature ascertained. The glass tube, by the increase of its 

 weight, indicated the quantity of water which had been evolved 

 during the process. 



The inconveniences attending this apparatus were the follow- 

 ing : 1 . The muriate of lime was apt to melt at the upper extre- 

 mity of the tube, owing partly to the water which it absorbed, 

 and partly to the heat to which it was exposed. When this 

 happened, it effectually prevented the gaseous product from 

 making its way through the tube ; so that the process for that 

 time failed. 2. The increase of weight which the tube sustained 

 Varied considerably in different trials with the same coal ; though 

 the process was conducted in every respect in the same way. 

 As the glass tube was luted to the copper tube by means of fat 

 lute, and as this lute during the process was unavoidably raised 

 to a temperature higher than that of boiling water, I could not 

 be certain that some part of the increase of weight of the tube 

 did not proceed from the lute. 



To get rid of this inconveniency, I got a brass tube with a 

 small bore (not larger than a moderate sized wire), about four 

 inches long, bent into the shape of the upper extremity of the 

 glass tube. One end of this brass tube was ground so as to fit 

 exactly the extremity of the copper tube ; and as brass is more 

 expanded by heat than copper, there was no risk of its loosening 

 during the continuance of the experiment. The other end of the 

 brass tube was shaped so as to fit the extremity of a glass tube 

 similar to the one formerly employed ; only it was deprived of its 

 upper end ; its diameter was rather greater, and it was stouter. 

 The end of the glass tube was cemented to the brass tube in one 

 case (for I have several different tubes) with plaster of Paris, in 

 another by shell lac ; and I found both methods to answer 

 equally well. The glass tube was filled with pounded muriate of 

 lime, which was prevented from making its way into the brass 



