Purificatkn of Fixed Alkalis. 1<^5 



and evaporated in a glafs retort, till it begins to depofi: regular cryftals. If the mafs fliould 

 confolidate ever fo little by cooling, a fmall quantity of water is to be added, and it muft 

 be heated again to render it fluid. After the formation of a fufficient quantity of regular 

 cryltals, the fluid, which is very brown, is to be decanted, and the fait, after being fuflfered 

 to drain, muft be re-diflbived in the fame quantity of water. The decanted liquor muft be 

 ];ept in a well clofed bottle, and fuflfered to become clear by fubfidence for feveral days. 

 It muft then be decanted for a fecond evaporation and cryftallization. The procefs muft 

 be repeated as long as the cryftals aflxird, with the leaft poflible quantity of water, folutions 

 perfeclly limpid. Thefe folutions are to be preferved in well dofed bottles, to defend them 

 from the accefs of air. 



7. The greateft difiiculty of this procefs arifes from the facility with which the lixivium 

 aflumes the folid form. To obviate tliis inconvenience, a fmall portion of the lixivium 

 may be concentrated to the point at which it becorhes converted into a folid mafs by 

 cooling. The facuration of a lixivium confiderably evaporated may be afcertained by 

 throwing fmall pieces of this mafs into it during its cooling. When thcfe are no longer 

 diflblved, it is a proof that the lixivium is at the required point. I have (hewn, in a par- 

 ticular article on the cryftallization of falts, the principles on which this pra£lice is 

 founded *. 



8. This method of proceeding is abfolutely indifpenfable in the cryftallization of foda, 

 which in other refpecls is made in the fame manner as that of pot-afh. The alkali would 

 not fail, without this management, to form a folid mafs during its cooling. 



9. With regard to the foreign falts which are mixed with the pot-afli, the greateft por- 

 tion feparates by cryftallization after the firft evaporation of tlic lixivium. The reft is 

 feparated during the fecond concentration by the continual fliimming of the pellicle. The 

 little which may remain with the pot-afti muft precipitate, for want of water of folution, in 

 a lixivium wherein the alkali itfelf is no longer diflTolved but by its own water of cryftal- 

 lization. 



10. The property of cauftic alkalis to diflblve in highly recSlified alcohol, with the ex- 

 clufion of every foreign falt,-would afford an excellent means of obtaining this fait very 

 pure, if their mutual adlion did not afford a new fource of impurity. For when an alkali 

 abfolutely pure and cryftallized is diflblved in fpirit of wine, even without heat, the fluid af- 

 fumes a very brown colour, which becomes ftill deeper after decantation from the faline mafs. 



11. The cryftallization of pot-afti is very different, accordingly as the cryftals are formed 

 with cold or heat. In the firft cafe, the cryftals obtained are odahedrons in groups, 

 which contain 0.43 water of cryftallization, and excite by their folution in water, even in 

 the fummer, a degree of cold very near the point of aqueous congelation. In the fecond 

 cafe, cryitallinc tranfparent very thin blades, of extraordinary magnitude, are formed, 

 which, by an affemblage of lines, in direftions that crofs each other to infinity, prefent 

 an aggregate of cells or cavities, moft commonly fo perfedly clofed that the veffel may 

 be inverted without the cfcape of the fmalleft drop of the lixivium, though fometimcs 

 included to the quantity of an ounce or two. For this reafon it is neceffary to break this 

 fine cryftallization, that the fluid may run off. The cryftals prefent in their regular 

 formation rcdlangular tetragonal blades, which, as they contain little water of cryftalliza- 

 tion, produce a confuicrable degree of heat when diflblved in water. 



* Chcm, Anual. auf 1795, b. i. f. C. 



12. By 



