this one the nitro-mercurous phosphate was placed in a mass in the center of the crucible and completely covered with cupric oxide. When the crucible was heated the cupric phosphate formed coated the outside of the pellet and prevented the further escape of the mercury. This proves the necessity of mixing thoroughly. Care should be taken that the mercurous nitrate used is not basic. The salt used in this investigation was prepared in the following way. Pure mercury was dissolved, by aid of gentle heat, in pure nitric acid in a flat, open vessel. More mercury was now added and the whole boiled until every trace of nitrous acid was driven off. The solution was then allowed to crystallize. To the solution formed from these crystals metallic mercury is added to prevent the formation of mercuric nitrate. By this method the salt is obtained perfectly free from mercuric nitrate and the nitrites. Ammonio-sodic phosphate.-As the ammonia present decomposes the mercurous nitrate, it is well to boil with a little caustic soda until it is driven off. In the first analysis magnesic oxide was substituted for the cupric oxide, but in every other respect the process was the same. 2 Found. Required. (1) 1:3500 grms. gave 3991 grms. P2O5 = 29.56 29.36 66 66 Ammonio-magnesic phosphate.--The solution was freed from ammonia as in the previous case. By ignition the following results were obtained. (1) 5936 grms. gave 3664 grms. Mg,P20, (2) 6135 66 4123 7 43.00 pr. ct. P2O 2 = 42.99 66 By the mercurous nitrate process I obtained the following results: (1) 10785 grms. gave 4627 grms. 42.92 pr. ct. P2O, 66 •4084 66 Calcic phosphate.-This salt was obtained by precipitating calcic chloride with an excess of disodic phosphate. The precipitate was then carefully washed and dried. This was dissolved in as little nitric acid as possible. If the solution is too acid the precipitate is not formed until sodic hydrate is added. The nitro-mercurous phosphate was ignited with magnesic oxide. By this means the following results were obtained: (1) 6190 grms. gave 1606 grms. P2O, 25.94 pr. ct. Mr. Waldo Lincoln employed this process in making an analysis of bone earth, and permits me to cite his results. The nitro-mercurous phosphate was ignited with cupric oxide. The first analysis presents another instance of failure on account of the precipitate not being properly mixed with the cupric oxide. Aluminic phosphate.-This was prepared by adding an excess of disodic phosphate to a solution of potash alum. The precipitate thus obtained was carefully washed and dried. The phosphate was dissolved in the least possible quantity of nitric acid and then treated as in the previous cases. An excess of nitric acid must be carefully avoided. The precipitate was ignited with cupric oxide. (1) 5685 grms. gave 1634 grms. P2 5 28.77 pr. ct. (2) 5827 In order to test the process thoroughly it was thought best to determine the composition of this salt and then to compare the found with the calculated percentage. The water was estimated by simple ignition. (1) 8689 grms. gave 2624 grms. H2O 30.18 pr. ct. (2) 8370 2 The aluminic oxide was taken by difference yielding 41·12 pr. ct. These results lead to the formula or (Al2O3)2 (P2O5)+8H2O Al2O3+Al2P20,+8H2O showing it to be a basic aluminic phosphate. I attempted to determine ferric and uranic phosphates in the same way, but on the addition of the sodic hydrate the metallic oxides were precipitated, rendering this method worthless for them. The iron solution was then very largely diluted and precipitated without the addition of sodic hydrate, but all of the phosphoric acid could not be thrown down. Next the ferric was reduced to the ferrous salt by the addition of metallic iron, by passing a stream of sulphydric acid through the solution, and by the addition of sulphurous acid, but without better results. The phosphate was dissolved in potassic cyanide and the ammonio-magnesian solution added to precipitate the phosphoric acid. I hoped that potassic ferrocyanide would be formed which would remain in solution and allow the phosphoric acid to combine with the potassium. The experiment, however, proved unsuccessful, and with these results the further investigation of ferric phosphate was suspended. Although the ignition with cupric oxide leaves nothing to be desired in point of accuracy, and is a much quicker method than any now in use, yet it was thought that some other agent might be employed which would hasten the process still more. Stannic oxide was tried for this purpose. To a weighed quantity of pure tin* the nitro-mercurous phosphate was added, then sufficient nitric acid, sp. gr. 1170, was poured upon the mass to oxidize the tin completely. The whole was evaporated to dryness on a radiator, ignited and weighed. The plan seemed very promising, but the results, from some unknown cause, were far from being satisfactory. Another idea was to fuse the nitro-mercurous phosphate with some substance which had a low fusing point, so that the mercury might be volatilized, the phosphoric oxide remaining. Accordingly it was fused with free sulphur but some of the phosphoric oxide was always volatilized. Potassic dichromate and plumbic chromate were both subjected to trial, but as the results were very unsatisfactory, further investigation in this direction was abandoned. These results prove conclusively the value of this process for the estimation of phosphoric acid in all cases except those of ferrict and uranic phosphates. * Chemically pure tin can be obtained in considerable quantities, and with very little trouble, in the following manner. Ammonio-stannic chloride is first produced by adding to one molecule of stannic chloride, two of sal ammoniac. Dissolve in the least possible quantity of water and add an excess of chlorhydric acid, which causes the double chloride to crystallize out immediately. Purify completely by recrystallization from acid solutions. Dry the salt at 120°. Then fuse with an equal weight of a mixture of one part of potassic cyanide and one of potassic carbonate in a porcelain crucible On cooling a button of chemically pure tin will be found. Rose proposes in the case of ferric phosphate, to fuse the mercurous phosphate obtained with the mixed carbonates; then dissolve in dilute chlorhydric acid, and precipitate with the ammonio-magnesian solution. This will not effect a complete separation. Otto's process of precipitating the phosphoric acid as ammonio-magnesic phosphate in the presence of tartaric acid was modified and tested as follows: Ferric phosphate was prepared by precipitating ferric chloride with disodic phosphate. The precipitate was thoroughly washed by decantation and then evaporated to dryBy this means the phosphate was obtained as a light yellow powder. A weighed quantity of this was dissolved in dilute chlorhydric acid, sufficient tartaric ness. 3. On the use of Porous Cones in Filtration. The sulphides of arsenic and antimony being quite easily obtained have long been regarded as furnishing the readiest and best means for the estimation of these two elements. A serious objection has however arisen in the fact, that they must be de termined upon weighed filters, and that paper filters cannot be dried above 100° without danger of loss. Hence a filter which could be weighed easily and withstand a high temperature became a desideratum. Taylor accomplished this with sand filters, and obtained some excellent results. This method, however, requires experience and careful manipulation, reasons which will probably prevent its being generally used. Under these circumstances the idea of using a cone made of very por ous earthenware, and as a substitute for the paper filter, presented itself to me, and has been carried out in the following manner. acid added to keep the whole of the iron in solution, next ammonia in excess, the solution was then heated till it boiled briskly, and the ammonio-magnesian solution was added. The precipitate was ignited and weighed as magnesic pyrophosphate. This method gave the following results. (1) 6084 grms. gave 3437 grms. Mg,P20, 36 13 per cent. P2Os (2) 7411 I would call especial attention to the fact, as it shortens the process considerably. that the phosphate solution was boiling when the ammonio-magnesian solution was added. Dr. Gibbs found that when the ammonio-magnesic and ammoniomanganic phosphates were precipitated from boiling solutions, that on cooling they came down as beautiful, highly crystalline precipitates, which, instead of requiring twenty-four hours for precipitation, were ready to be filtered within an hour. This method has been in use in this laboratory for the past six months, an has met with uniform success. It applies as well to the estimation of phosphoric acid as to that of magnesium and manganese. With this additional treatment for the separation of iron, it is believed that this process will apply to the majority of cases met with in practice. *This Jour. II, vol. xliv, p. 215. The cones* are made of very light, porous earthenware, and have an angle of about sixty degrees. They are used in the following way: A section of a seamless rubber tube a, is stretched around the mouth of a funnel b, preferably a Bunsen funnel, allowing a portion of the tube to project above the top. This part will immediately arrange itself at right angles to the top of the funnel; into the circle thus formed the cone c, is put. It is then connected with the Bunsen pump. When the cone is moistened and the pressure applied, the rubber band forms an air-tight joint and the liquid runs through with great rapidity. Before the cones are applied to quantitative work they must be carefully washed, first with concentrated chlorhydric acid, then with distilled water, dried, and weighed. A small porcelain crucible was always kept at the balance in which to weigh them. this apparatus the following results have been obtained. A sample of ordinary crystallized potassio-antimonylic tartrate gave (1) 11069 grms. gave 5570 grms. (2) 1.6985 66 66 66 •8550 With Sb3 Sa 3 =35.94 pr. ct. Sb. =35.96 66 66 A specimen of the same salt carefully recrystallized, gave upon analysis the following data. (1)7755 grms. gave 3955 grms. Sb2S=36.40 pr. et. Sb. (2) •6050 66 66 ⚫3085 Mean 36.41. =36.42 66 66 Mr. W. Lincoln also made an analysis of the same salt by means of the cones, and kindly permits me to use his results. (1) 4683 grms. gave 2388 grms. Sb,S,=36-41 pr. ct. Sb. (2) 8135 In all of these analyses Sb=120. The Sb,S, was precipitated in the manner recommended by Sharples. It is necessary that the boiling should proceed for some time, a brisk current of sulphydric acid being passed through the liquid. The precipitate is then filtered upon the weighed cone and the whole dried in an air bath at 300°. The antimonous sulphide is at this temperature converted into the gray crystalline modification. * Much credit is due to the firm of A. W. & H. C. Robertson of Chelsea, Mass., who, by their superior mechanical skill, have aided me much in accomplishing the desired result. Taylor obtained as the mean of three analyses 36.08. This Jour., II, vol. 1, page 248. |