The phosphoric acid is determined in one-half of the filtrate and in the remaining half the iron is reduced with zinc and determined with potassium permanganate in the usual way. The phosphoric acid and iron having been thus determined the alumina is estimated by difference. The chief objection to this process is in the excessive quantity of acetic acid used and the danger of solution of the precipitated phosphates caused thereby. =43. Estimation of the Lime and Magnesia.=—The filtrate and washings from the first precipitation, (paragraph =41=,) of iron and alumina in the method of Krug and McElroy, above described, are collected and sufficient ammonium oxalate is added to precipitate the calcium. The precipitated calcium is very fine and should be collected on a gooch, under pressure. The filtrate and washings from the calcium precipitate are again collected, and a solution of sodium phosphate added to precipitate the magnesia. The solution must be kept cool and slightly alkaline with ammonia during the above operations in order to prevent the separation of molybdic anhydrid. =44. Estimation of Sulfuric Acid.=—As a rule, sulfates are not abundant in mineral phosphates. In case the samples are pyritiferous, however, considerable quantities of sulfuric acid may be found after treatment with aqua regia. The acid is precipitated with barium chlorid, in the usual way, in an aliquot portion of the filtrate first obtained. The precipitate of barium sulfate is washed with hot water until clean, dried, ignited, and weighed. If the portion of the filtrate taken represent half a gram of the original material then the weight of barium sulfate obtained multiplied by 0.6858 will give the quantity of sulfur trioxid in one gram. =45. Estimation of Fluorin by the Method of Berzelius as Modified by Chatard.=—The method of estimating fluorin as proposed by Berzelius, has been found quite satisfactory in the laboratory of the Geological Survey, with the modifications given below.[34]