Prapofthns of Carbon in Minera! Coal. 489 



«ny air from nitre, nor confequently contribute to its decompofition, »s the fubfequent 

 experiments fuflicicntly evince; but is confumed partly by the pure air fpontaneouDy 

 emitted by nitre during its ignition, and partly by the ambient atmofpheric air. 



Nay, when mineral coal is employed in the decompofition of nitre, the (hare which the 

 mere carbonaceous part of the bitumen contained in it contributes to the decompofition 

 ■will be found fo fmall that it merits no confideration in the general account. 



The firft ftep towards carrying this analytic plan into execution mud therefore be, to 

 determine the quantity of pure carbon neceflary to decompofe a given quantity of pure 

 nitre. But here many practical difficulties occur, which {hall prefently be mentioned. The 

 mod perfeifl method of obviating them was that employed by the ever memorable Lavoifier : 

 he mixed the pureft nitre with charcoal alfo purged of the inflammable as well as other 

 airs and water which it ufually abforbs. In the proportion which, after fevcral trials, he 

 found requlfite for the entire decompofition of that fait, rammed them into a copper tube, 

 fired them, and continued the inflammation under water ; by which means the charcoal was 

 »£led on folely by the air educed from the nitre to the entire exclufion of the external air, 

 and this air was educed folely by the ignited charcoal to the entire exclufion of external 

 heat ; advantages that cannot be procured by the ufual mode of e<Fe(fl;ing this decompo- 

 fition. Thus he found the proportion of charcoal neceflary for the entire decompofition of 

 nitre, to be as I to 7,57, or, in other words, that 13,21 parts charcoal decompofe 100 of 

 nitre * : and yet even In this experiment I find a fmall inaccuracy, as he did not take the 

 water employed in mixing the nitre and charcoal into the account ; and hence, and for 

 feme other reafons, the detail of which would lead me too far, I think the proportion fliould 

 be as I to 7,868 nearly, or that 12,709 charcoal decompofe lOo of nitre ( but the difference 

 is of little importance. 



This mode of experimenting, however, is Inapplicable on the prefent occafion ; the dif- 

 ferent fpecies of mineral coal being not fo readily Inflammable as to carry on the combuftion 

 in this manner. Hence I contented myfelf with the common manner only, ufing fuch pre- 

 cautions as to render Its refults tolerably uniform, and repeating each experiment feveral 

 times. 



I examined the purity of the nitre I employed by nitrated filver, and found by the quan- 

 tity of fallted filver produced, that 480 grains of the nitre contained 3,5 grains of common 

 fait (13 J grains of muriated filver. Indicating too of common fait): hence the conftant quan- 

 tity of nitre I ufed was 483^ grains, except in the experiments on bitumens, as I had not 

 enough of them to expend In fo large a quantity of nitre. 



The nitre was heated barely to rednefs, before any coal was projefted on It, In a wind fur- 

 nace and a very large crucible ; upon this uniform degree of heat much of the uniformity 

 of dilTerent experiments on the fame fpecies of coal depends. 



In my firft experiments the coals were reduced to a very fine powder, and then projected 

 on the ignited nitre ; but I obfervcd that by this method much more of each fpecies of coal 

 was requlfite to alkalize the llandard quantity of nitre, than when It was reduced to a coarfe 

 powder about the fize of a pin's licad, or fomewhat larger ; and the reafon is, that by the 

 force of the cxplofion much of the finer powder is carried ofl^, without having been in 



■' II Mem. Sqav. F.tnng. 6jC. 



Vol. I.— February 1798. 3 1^ contaa 



