THE ANALYST. 109 INORGANIC ANALYSIS. On ths Estimation of Water in Silicates by the Borax Method. P. Jannasch and P. Weingarten. (Zeit. A ~ ~ o r g . CILewz., xi., 37.)-This is R modification of the process already described (ibid., viii., 352), which was designed to replace the method of water estimation where the mineral is opened out with plumbic oxide, and which was found to yield irregular results in the case of silicates rich in calcium. The new process permits of the portion of material operated upon being used for other subsequent determinations. 1-5 grainmes of borax, dried as described below, are placed in a platinum boat 7.5 c.ni. long by 1.5 deep, which is put into the middle of a thin-walled combustion tube 75 c.m. in length, followed by a plug of glass wool. To free the tube from the last traces of moisture it is heated gently in a combustion furnace for half an hour while a current of dry air is passed through, great care being taken to avoid fusing the borax.After this the tube is allowed to cool, the air current is stopped, and the weighed quantity of the substance under examination is mixed as rapidly and as iiztimntely as possible with the borax in the boat by means of a spiral of platinum. wire. The boat is pushed back to its former position, the glass-wool plug replaced, and the tube again connected to the air supply. The unweighed calcium chloride tube at the exit end of the combustion tube is changed for a weighed one, to the outer end of which the first tube is attached. The part of the tube containing the boat is then very gradually heated up, the remainder being kept as cool as possible by not covering it with tiles.When the borax is fused to a clear homogeneous glass in which bubbles are no longer being formed, the air-current is started once more, and the whole length of the tube is heated with a Bunsen burner to drive the moisture into the calcium chloride tube, which, after cooling to the temperature of the balance-room, is reweighed. Should fluorine be present in the substance analysed, a layer of granulated lead chromate 5 c.m. long, with a glass wool-plug at either end, is placed between the platinum boat and the exit end of the combustion tube, to absorb any volatile fluorides.110 THE ANALYST. The powdered borax employed in the process is prepared (Zcit.Anorg. Chenz.. viii., 352) by heating pure crystallized borax to incipient fusion in a platinum dish or crucible, allowing to cool, reducing to fine powder and finally heating to a dull red heat while the powder is stirred with a thick platinum wire. C. H. C. Volumetric Determination of Copper. M. Rupeau. (Bzdl. SOC. Pharnz. dc Bordeaux, October, 1895 ; through A m . Chiwz. AmZyt., i., 32.)-The author proposes the employment of a solution of picric acid, which gives with copper a precipitate insoluble in ammoniacal liquids. The standard picric acid solution is made by dis- solving rather more than 7.5 grammes of the acid in hot water, and, after adding ammonia (30 to 40 c.c.) till a decided odour is manifested, making up to 1 litre. After standing for some time, the solution, is filtered and standardized against a solution of 1 grainme of pure red copper, dissolved in nitric acid and made up to 100 C.C.with water (1 C.C. = 10 mgms. Cu). The picric acid is then run in, with constant agitation. So long as a green tinge is observed there is no danger of over- stepping the end point; but when this is approached the precipitate is allowed to settle, and the addition is proceeded with cautiously until a decided yellow tinge without any tendency to green is obtained in the clear supernatant layer of fluid. The titration of the solution under examination is performed in an exactly similar manner. When silver is present, 10 C.C. of ammonia must be added to the nitric solution to keep this metal in solution while titrating, and the same holds good for zinc as well.The oxide thrown down by the ammonia obscures the final reaction, but this may be obviated by adding a little tartaric acid. Iron is removed by simply boiling the solution, adding ammonia to precipitate the ferric oxide. I n the inost complex case likely to occur--i.e., one in which the copper is associated with all the four metals mentioned above-the method is as simple as Lead behaves differently. when iron is the only admixture. c. s. A Critical Examination of the Processes for the Estimation of Phosphoric Acid. C. Meineke. (Chem. Zeit., 1896, xx., 108.)-The author has already stated, some ten years ago, that it is perfectly possible by gentle ignition to convert the yellow ammonium phospho-rnolybdate into a form in which it can safely be weighed with the assurance of its containing a definite percentage of phosphorus.He has now reinvestigated his process, checking it on specinlens of phosphorite, on pure disodium phosphate, and on silver phosphate, and he finds that the phosphoinolybdic anhydride so obtained contains constantly 3.949 (&0.008) per cent. of P,O,; corre- sponding, therefore, to the formula 24Mo0, + P,O,, which (taking Mo = 95-9, P = 30.96, and 0 = 15-96) demands 3.944 per cent. The molybdate solution should be prepared by dissolving 150 grammes of the ammonium salt in 150 C.C. of ammonia (specific gravity 0.91) and 850 C.C. of water, pouring the whole into 1 litre of nitric acid (specific gravity 1.2). The liquid is heated to 90" C. for ten minutes, filtered from the molybdic acid, and preserved in the dark.THE ANALYST, 111 I n testing solutions poor in iron, the precipitation is so arranged that at least 5 per cent.of ammonium nitrate is present. To ensure this, ammonia (specific gravity 0.91) and nitric acid (specific gravity 1-4) are prepared, which, when mixed together in equal volumes, yield a moderately acid solution of the nitrate of about 53 per cent. con- centration. These are added in the necessary amounts in succession to the liquid under examination, and when the temperature has fallen to 50” C., the molybdate is added, the whole thoroughly stirred, and set aside till clear. If the solution, on the other hand, contains much iron, at least 10 per cent. of ammonium nitrate must be present, and enough free nitric acid to represent 5 to 10 per cent.of the 1-4 strength. The solution is heated to the boil, the inolybdate introduced, stirred up, and set aside. Further heating after the addition of the precipitant rnust be avoided. The precipitate should be pale yellow, and the liquid must also be light in colour. It is washed with a liquid consisting of 100 C.C. each of the same acid and ammonia diluted to 1 litre with water. il final rinse in pure cold water is given, followed, if desired, by alcohol and ether, and the filter is then dried. I t is ignited in a porce- lain basin or crucible over an argand, or in a mume, at a temperature which is not visibly red by daylight. When hot, the residue has a somewhat grayish colour, but when cold it is blue-black.I t must be free from yellow or green particles, and also froin free iiiolybdic acid ; and it should be cooled in the desiccator. The author has also examined the statement that the presence of ammonium chloride in solutions containing much iron, which are to be precipitated by molybdate solution, is to be avoided. A series of careful experiments show that this substance, even when present in abnormally large amounts, is entirely without action on the results. The remainder of the paper contains a long and elaborate examination of the two processes-Maercker’s and Wagner’s--for the estimation of phosphoric acid by magnesia mixture when following the molybdate precipitation. The former neutralises the ammoniacal solution of the yellow precipitate, introduces the magnesia mixture, a d finally adds ammonia in 24 or 3 per cent. excess.Wagner prefers to add the reagent direct to the alkaline solution, which should contain the same amount of free ammonia. Both processes are liable to several errors : the pre- cipitate is somewhat soluble, it is apt to vary in composition, and, according to the temperature of ignition, greater or less loss of phosphoric acid iiiay occur. Neubauer has published a table of corrections for this last source of error, but it cannot be considered exact. The present author, when using Wagner’s process, prefers to ignite the precipitate gently, weigh it, ignite again over the blow-pipe to constant weight (10 to 15 minutes) and weigh once more, The original precipitate, consisting of an indefinite mixture of pyro- and meta-phosphate, suffers a loss of P,O, corre- sponding to the Mg(PO,), it contains ; and if this loss be multiplied by 2.56, and the weight so obtained be subtracted from the original, the remainder expresses the true amount of pyrophosphate in the compound.I n this manner the quantity of phosphoric acid in the precipitate can be determined exactly : the yield usually being a trifle above the theoretical. It may be observed that without correction, Wagner’s process gives sufficiently exact results for many practical purposes, such as the analysis of manures : for when 1 gramme of substance is taken, the phosphoric acid comes out only about 0.2 per cent. too low.112 THE ANALYST. With Maercker’s modification, the original precipitate i s apt to vary in composi- tion still more ; but good results are to be obtained by preventing the precipitation of the magnesium phosphate from occurring too early during the addition of the reagent. Experiments show that it is better to effect this by means of ammonium citrate than by a large excess of hydrochloric acid ; and the process is best arranged 60 that the liquid remains clear until half of the necessary amount of magnesia mixture has been run in. I n all cases the precipitate must be ignited strongly to constant weight, in order to remove the niolybdic acid which is always present. F. H. L.