Memoir upon Coffee. 29 



of sulphate and oxalate of lime. The prusslate of ammonia 

 indicated the presence of an atom of iron; the g-allated al- 

 cohol acted in the same manner, by commLmicating to the 

 solution of this salt a purple colour. 



BcinT desirous of knowing the products which coffee not 

 boiled in water would furnish by its decomposition, I re- 

 duced 200 parts of it into a rough powder, which I intro- 

 duced into a glass retort, to which I joined a tubulated ma- 

 trass, communicating by means of a tube with the hydro- 

 pneumatic tub. The first portions of liquid which passed 

 into the matrass after the apparatus was heated were only 

 a white limpid liquid, which became brown in proportion 

 as I augmented the heat : to this first product succeeded a 

 citrine oil slightly fetid, which becair.e black and thicker 

 on the temperature being raised. Towards the end of the 

 operation there was sublimed into the' neck of the retort a 

 saline matter, confusedly crystallized, mixed with black em- 

 pyreumatic oil : I then augmented the fire until the bottom 

 of the retort was reddened, in the hope of liberating some 

 gaseous substances ; but I did 'not succeed : so that from 

 this I was led to believe, that the principles of which coffee 

 i^ composed, and which are disengaged by fire, are combined 

 in a state different from that in which they exist in the 

 o-rain, and that the proportions of each of them are exactly 

 those which ought to form the ammonia, carbonic acid, oil, 

 water, and pyromucous acid, which arc the products of this 

 operation. The retort was broken : it contained a charcoal 

 occupying all the capacity of the vessel ; it appeared deco- 

 rated with the most brilliant colours, and of various hues; 

 the lead gray with the metallic aspect was very distinct to- 

 wards that part which touched the bottom of the retort; its 

 weight was 50 parts. 



The liquid contained in the matrass was composed, as 

 wc have said, of thick black oil, very much charred, and of 

 a brownish liquid, fetid and astringent : its total weight, 

 isolated from the oil, was 88 parts ; that of the oil was 33. 

 Tl>e neck of the retort still contained a sublimed saline mat- 

 ter, of a fetid ammoniacal smell, weighing 22 parts, and 

 which was only carbonate of ammonia: the loss was seven. 



The 



