In some cases, especially when the lyes are very dirty and the total residue in the crude glycerine runs high, for which there is a penalty usually attached, a double filtration of the lye is advisable. This is carried out by first making the lye slightly acid in reaction by the addition of alum and acid, then filtering. This filtered lye is then neutralized to the proper point with caustic, as already described, and passed through the filter press again. While in the method of treating the lyes as given sulfuric acid is used for neutralizing, some operators prefer to use hydrochloric acid, as this forms sodium chloride or common salt, whereas sulfuric acid forms sodium sulfate, having 3/5 the graining power of salt, which eventually renders the salt useless for graining the soap, as the percentage of sodium sulfate increases in the salt. When the salt contains 25 per cent. sodium sulfate it is advisable to throw it away. Sulfuric acid, however, is considerably cheaper than hydrochloric and this more than compensates the necessity of having to eventually reject the recovered salt. It may here also be mentioned that recovered salt contains 5-7 per cent. glycerine which should be washed out in the evaporator before it is thrown away. The following tables give the approximate theoretical amounts of acids of various strengths required to neutralize one pound of caustic soda: For 1 pound of caustic soda-- 3.25 lbs. 18deg. B. hydrochloric (muriatic) acid are required. 2.92 " 20deg. B. " " " " " 2.58 " 22deg. B. " " " " " For 1 pound of caustic soda-- 1.93 lbs. 50deg. B. sulphuric acid are required. 1.54 " 60deg. B. " " " " 1.28 " 66deg. B. " " " " It is, of course, feasible to neutralize the spent lye without first determining the causticity by titrating a sample and this is often the case. The operator under such conditions first adds the sulfate of alumina, then the acid, using litmus paper as his indicator. Comparatively, this method of treatment is much slower and not as positive, as the amount of acid or alkali to be added is at all times uncertain, for in the foaming of the lyes their action on litmus is misleading.