- PRODUCTS FEOM COAL. 39 



coal*, and this may be done in the manner shown in the annexed woodcuts, 

 Figs. 6 and 7. 



A is the cast-iron retort, 4 feet long, 12 inches wide, and 4 inches high, 

 set in a furnace exposed directly to the flame, the flue passing beneath and 

 over it, without guard of any kind, and secured into the side walls by the 

 snugs, bb. P P is a wrought-iron welded tube, about f inside diameter, in- 

 serted into the retort at B, and running the entire length from B outwards ; 

 it is bent in the form of a gauge, and must not be less than two feet from B 

 to C : at the socket D, one foot from the bottom bend, a glass tube about 14 

 inches long is firmly fixed by a little plaster of Paris : the whole tube from 

 the retort to the commencement of the glass should be thickly wrapped with 

 flannel or woollen listing (the thicker the better) , to ensure an unvarying tem- 

 perature. Pour mercury into the glass tube until it is just visible at D : heat 

 the retort to a bright red by daylight for a few hours, to decompose all vapour 

 in the gauge ; mark the glass gauge as the mercury rises (as at 1, 2, 3, etc.), 

 fixing a point (say for instance at 6) for the heat at which the distillation is 

 to be carried on (say 27 Wedgewood, the degree of heat at which copper 

 melts), and the apparatus is then ready for the reception of the coalf. The 

 coal must be prepared as follows : sift a quantity through a sieve with meshes 

 about f ths of an inch apart, and then through another much finer, so that the 

 smallest pieces of coal used will be about the size of coffee-berries. Weigh 

 correctly 3 pounds, and spread it over a surface of 360 square inches ; that 

 is, on a tray of sheet-iron, 10 inches wide and 36 inches long, turned up at 

 the sides. Have the lid of the retort ready luted, and when the gauge marks 

 6, introduce the coal on the tray and immediately secure the lid. 



Before I describe the remaining portion of the apparatus, it will be neces- 

 sary to make a few remarks upon the necessity of having the barometric-gauge 

 attached to the retort. In the first place, then, if the heat at which the dis- 

 tillation is carried on were not uniform in all the experiments, the results, 

 even from the same piece of coal, would vary from 50 to 60 per cent, both in 

 quantity and quality. If the retort is too cold, nitrogen and hydrogen are libe- 



* Of course the coal supplied from a known main need not be subjected to this test ; but that 

 procured from new districts I should certainly advise to be tried by the method described. 



t It is always well to have a certain fixed degree of heat, such as the melting-point of tin, 

 marked as 1 on the glass gauge lead, marked 2, and so on, both because errors are then less likely 

 to occur, and memoranda, when referred to, more readily understood. The mercury will always 

 indicate the same for the same temperature. 



