. Examination of various Prociffes. .' J7 



•o^d water, miH. deprived of the marir^e fait V(rhich \x. contains. . .2. The waOiing^ of falt- 

 petre, when it is performed after tlie'feiuHon and cryftaliizatioH^requires 40 of'soper ccitit! 

 of water inftead of 35. '' '' " 



3. Trial has been made to diffolye faltpetre in 20 or 25 per cent, of boiling water; to 

 take out the fea-falt in proportion as it falls by the ebullition of the liquor j to dilute the 

 fluid with 30 per cent, of additional water, and then to convey it into the veflel of cryftal- 

 lization. It was expedted that the wafliings with cold water might be avoided, or confi- 

 derably diminiftied by thefe means 5 but, not to mention that ebullition maintained for four 

 or five hours to feparatc the fea-falt fuppofes very great confumption of time, fuel, and falt- 

 petre, the wafliings are flill indifperifable to remove the colouring principle, and to carry 

 off the laft portions of muriate of foda. 



4. It may be thought that the quantity of waiter in the wafliings might probably be di- 

 miniflied ; but there is reafon to fear that when the faltpetre is loaded with fea-falt, it 

 cannot be perfectly refined by the ufc of a lefs quantity of water than we have prefcribed. 



5. The operator may alfo be tempted to diminifli the proportion of water employed in 

 the folution; but we are convinced by numerous experiments, that this proportion is the 

 mofl; fuitable. If it be increafed, the faltpetre remains diflblved in the fluid ; if it be dimi- 

 niflied, it fettles or falls down in a folid mafs. Obfervation has proved, that the degree of 

 faturation bed adapted to this work, is between the 66th and 68th degrees of the hydro- 

 meter (fp. gr. 1.848, and i.8d>6.) 1 ' 



6. It may likewife be thought more Ample and economical to treat the folution of crude 

 faltpetre with potafli. But it is to be feared in this cafe that part of the alkali might be em- 

 ployed in decompofing the muriate of foda, to convert it into muriate of potafli ; and it mull 

 be obferved, that this laft fait is not at all proper to decompofe the earthy nitrates, notwith- 

 ftanding the aflertions of flcilful chemifts to that effeft. 



It appears, therefore, more convenient to defer the treatment of the mother waters, and 

 not to ufe potafli till after the fea-falt has been feparated by evaporation. . ' : 

 This procefs, therefore, unites a number of advantages. 



1. Itconfumes much lefs fuel : for, inftead of two long folutionsand ebullitions, nothing 

 more is fequired than to give the water a boiling heat in order to dilTolve the faltpetre. 



2. It requires lefs time. Three days are fuflicient to purify the faltpetre to the degree 

 fuitable for making gunpowder. 



3. It difpofes the faltpetre to dry more readily. As the cryftals are no larger than fTnall 

 needles, a few hours' expofure to the air are fufEcient for its complete deficcation. This 

 advantage is inefl;imable, particularly in a feafon wherein feveral months would be required 

 to drain the large loaves of nitre ; and in which, confequently, the fabrication of gunpowder 

 would be either retarded or fufpended, and the drying-houfes encumbered with quantities 

 of humid nitre. ' ~ 



4. It requires lefs fpace. A boiler five feet in diameter and four in depth, a veflel for 

 cryftallization of a few feet dimenfions, and thirty tubs, are perfedly fufficient to refine fif- 

 teen thoufand weight daily. " 



5. It occafions lefs lofs. Very accurate experiments have fliown that the folutions ufed 

 in the ancient procefs occafioned a lofs of faltpetre, by mere evaporation, which amounted 

 to 7 per cent, of the original quantity. By this new procefs, the water which holds the falt- 



E 2 ^ petre 



