DETERMINATION OE LITHIA IN MINERAL WATERS.rate can happen with a slag, considering that its type as silicate is determined beforehand in the charges of a blast furnace, by furnishing to the silica of ores, stone, and ash of coal, perfectly definite basic elements, in certain proportions primarily calculated, which necessarily determine a composition which the ultimate analysis of the slag, as run from the furnace, corroborates very closely in the great majority of cases. When it does not happen, it has also to be attributed to disturbing factors, and they can be found, for instance, in this fact that calcium, combined with sulphur, which the slags often contain in very serious quantities, has been calculated as oxide of calcium or lime. The same can be said of magnesium and manganese. i>fc of sulphur in a slag is by no means an exceptional circumstance ; ¡S# of manganese sulphide is frequently met with in Scotch slags. In other cases alumina, owing to certain condition of the furnace, or the proportions of fluxes, will play the part of an acid, and, in these conditions, ought not to figure in the determinations of the type as saturating silica.(To be continued.)DETERMINATION OF LITHIA IN MINERAL WATERS.By E. Waller. Ph. 1).Practically, three methods are now available. 1. The phosphate method (Mayer's modification) (Ann. Chem. v. Pharm. 98, 193). 2. The amylalcohol method (Gooch, Am. Chem. Jour., 9, 33). 3. The fluoride method (Carnot, Bull. Soc. Chim. [3] 1, 280).Rammelsberg's method (Pogg. Ann., 66, 79) somewhat similar in principle to that of Gooch, in that it depends upon the comparatively greater solubility of lithium chloride in an organic solvent, has been comparatively little used, on account of the difficulty and expense involved in obtaining the pure anhydrous alcohol and ether necessary for the process. Moreover the experiments of J. L. Smith (Am. Jour. Sci. [2] 16, 56), rearranged in convenient form for reference by Gooch (loc. cit.) do not indicate that it is
