48 THE TESTING OF EOAD MATERIALS. 



silica filtered, washed, and burned white over an ordinary Bunsen flame. The 

 error resulting from a single evaporation on the water bath is in most cases very 

 nearly counterbalanced by the error introduced by omitting the blasting, while the 

 saving in time is very considerable. The slight amount of silica which remains in 

 solution is carried down later with the iron and alumina, and serves to counterbal- 

 ance the small quantity of aluminum phosphate and alumina almost always found 

 in the silica. 



Whichever of the two procedures as given above has been followed, the liltrate 

 from the silica is transferred to a 500 cc graduated flask, diluted to the mark and 

 very thoroughly shaken. This is solution A. A 100 cc pipette which has been 

 carefully calibrated to divide the contents of the flask into exact fifths under stand- 

 anl conditions is used for dividing the solution into aliquot parts. Manganese 

 usually shows itself in the green color which it imparts to the fusion, and a slight 

 pinkish color which can be observed when the fusion is first treated with hydro- 

 chloric acid. A 100 cc portion having been transferred to a beaker, its treatment 

 will depend upon whether or not manganese is present. 



(1) When manganese is present. Add a solution of sodic carbonate until one or 

 two drops in excess have produced a permanent precipitate. Clear with one drop 

 of concentrated hydrochloric acid. Dilute to 200 cc, add ten drops of glacial acetic 

 acid and 2 grams of crystallized sodic acetate. Heat to boiling and boil three min- 

 utes. Allow to settle and filter hot. Wash a little with 2 per cent sodium acetate 

 solution, dissolve the precipitate in an excess of hydrochloric acid, and reprecipftate 

 with ammonia filter, wash, ignite, and weigh. The two filtrates are combined and 

 boiled for some time with additions of bromin water until the precipitation of man- 

 ganese dioxid is complete. The precipitate is ignited over the ordinary Bunsen 

 flame and weighed as Mn 3 O 4 . The filtrate from the manganese dioxid precipitate 

 which contains calcium and magnesium is put on the steam bath for concentration 

 while other portions of the analysis are proceeded with. 



(2) When manganese is absent. Ammonia is added in slight excess, the solution 

 bnili-d for two minutes and filtered hot, using the pump. Unless the precipitate is 

 a small one, another precipitation is necessary. The precipitate is dissolved in hydro- 

 chloric acid in the same beaker, reprecipitated with ammonia, filtered, washed with 

 hot water, blasted, and weighed. If necessary the combined filtrates are placed on 

 the steam bath for concentration. The paper containing the precipitate is transferred 

 wet to a weighed crucible. If the accurate method has been used, there will be a 

 weighed crucible ready to receive it. The paper and contents are rapidly dried by 

 keeping the crucible moving over a free flame, the paper is burned off and the pre- 

 cipitate finally blasted and weighed. This weight represents ferric ox id. alumina, 

 titanium oxid, ami phosphoric anhydrid. 



TOTAL IKON. 



< >ne hundred cubic centimeters of solution A is poured directly through a reductor 

 and followed by I")0 ccof distilled water. The clear, colorless solution is transferred 

 to a beaker, and to it is added 50 cc of dilute sulphuric acid and finally 5 cc of a 

 I'D per cent solution of nijinganoiis sulphate. The contents of the beaker are imme- 

 diately titrated with standard permanganate solution. The titration with bichrom- 

 ate is not nearly so rapid as the titratioti with permanganate and with very small 

 amounts of iron the end point can not be accurately distinguished. A large number 

 of trial< has shown that this direct method of titration in the presence of hydro- 

 chloric acid is quite as accurate as those which call for an evaporation with sulphuric 

 acid or a preliminary precipitation with ammonia, filtration, and re-solution in sul- 

 phuric acid. A permanganate solution is made by dissolving 1.975 grams in a liter. 

 One cubic centimeter of this solution equals 0.005 gram ferric oxid (Fe 2 O 3 ). The 





