goodly number (twenty) elements are operated. The parts of the apparatus are (Fig. 1):-A vessel of sheet copper 160 cm. long, 11.5 cm. deep, and 5 cm. wide (inside). This vessel is provided with a removable lid (which, however, needs rarely to be removed); the object of the lid is to keep out dust. The lid is provided with a hopper-like opening, c, which itself has a cap or cover. The vessel is divided into twenty water-tight compartments by partitions, each of exactly the same height (6.5 cm.); each compartment accordingly holds about 260 cc. of water. In the centre of the bottom of each compartment (Fig. 3, section) is a small hole, a, about 5 mm. in diameter; over this hole is soldered a galvanised iron tee, b. In one branch of the tee is inserted, by means of a cork, a glass tube, c, which serves as a water-gauge; into the other branch is screwed an ordinary brass gas-cock, d (male, small size). To the nozzle is attached, by means of a cork provided with an air vent, the syphon. The syphon is, preferably, what is known as the engineers' syphon. It consists of a glass tube, e, 10 cm. long and 18 cm. in diameter inside. The lower end of this tube is fitted with a cork through which passes a small glass tube, f, about 4 mm. in diameter inside. This small tube is about 9 cm. long; its upper end is V-shaped (ground sloping on two sides), and extends about 5 cm. into the larger tube. Over the V-shaped tube. and supported by it, is a cap or glass tube, g, closed at its upper end, about 7 mm. in diameter inside, and of such length (about 4'5 cm.) that the syphon delivers 10 to 15 cm. of water at each overflow. The flow of water, drop by drop, into the syphon is so regulated by the cock that it (the syphon) delivers water to the funnel underneath more slowly than the water runs out of the funnel. Intermittent washing of the contents of the funnel is thus effected. The inverted small funnel in the larger funnel throws the water around over the edge of the filter-paper. The wash-water is caught by a trough which conveys it into a sink.

The machine is charged by pouring water (distilled) through the opening, c; the excess of water, after all the compartments are full, runs out through the overflow, I, which is on a level with the tops of the compartments. Laboratory of the North Carolina Department of Agriculture, Raleigh, N.C.

THE

5

It is well recognised by everyone conversant with this matter that as many different methods as possible for determining each atomic weight should be employed. F. W. Clarke has often emphasised this point. The difficulty in carrying out this recommendation is that so few substances lend themselves to accurate analysis. Crystalline precipitates always carry down some of the mother liquor from which they were prepared, and hence many usual analytical methods are not permissible when great accu racy is required. For this reason, and so as to get consistent and comparable results with different metals, the analysis of the chlorides and bromides has been one of the chief methods employed at Harvard, because the spongy nature of the silver halides makes it possible to wash them with unusual thoroughness. But some means of verifying these results by comparison with an entirely different standard would be highly welcome. For this purpose a new process, involving cross-reference, has recently been carried out in the Wolcott Gibbs Memorial Laboratory, and the present paper describes this new method of attack.

The plan was simply this :-The purest sodium carbonate was to be weighed with scrupulous care, dissolved in water, and exactly neutralised with hydrobromic acid. The amount of silver necessary to precipitate this hydrobromic acid was then to be determined by the usual process. In this way the chemical equivalence between the weights of sodium carbonate and silver could be simply determined, and the silver halide standard referred directly to an entirely new substance, namely, sodium carbonate. The accuracy of the process turned of course upon obvious criteria: first, the purity of the sodium carbonate and hydrobromic acid; secondly, the complete and exact neutralisation without loss of material; and thirdly, the purity of the silver. The sodium carbonate, silver, and silver bromide alone were weighed, but the hydrobromic acid obviously had to be prepared in a state free from every trace of other acids, as well as from bases. As will be seen, the outcome was very satisfactory.

Preparation of Materials.

In an investigation of this kind the details of experimentation are of the utmost importance, therefore the necessary minutiæ are given below.

Three different specimens of sodium carbonate were prepared; the first simply by many recrystallisations from water in platinum vessels, the second from recrystallised

AND THE ATOMIC WEIGHT OF CARBON REFERRED to silvER AND BROMINE.*

By THEODORE W. RICHARDS and CHARLES R. HOOVER. INVESTIGATIONS concerning radio-activity have increased rather than diminished the interest in atomic weights. The possibility indicated by recent research that variations in these figures may exist in radio-active elements stimulates further research in all directions concerning these quantities, for such variations cannot be due to chance, but must have fundamental cause (Richards and Lembert, Journ. Am. Chem. Soc., 1914, xxxvi., 1329; also Zeit. Anorg. Chem., 1914, lxxxviii., 429, and Science, June 5, 1914, 831. This conclusion has been supported also by Soddy, Proc. Chem. Soc., 1914, xxx., 134; by Maurice Curie, Comptes Rendus, 1914, clviii., 1676; and by Hönig schmid and Mdlle. S. Horovitz, Ibid., p. 1786; also Zeit. Electrochem., 1914, xx., 319). To be sure, no valid evidence of variation in any value, except that for lead, has as yet been found, but whether or not the atomic weights vary in a few cases, they still probably remain the most fundamental figures which come within our scientific ken; therefore no amount of trouble is too great to secure complete and satisfactory knowledge of them.

Contribution from the Wolcott Gibbs Memorial Laboratory of
From the Journal of the American Chemical

Harvard College. Society, xxxvii., No. 1,

silver oxide, and the third from sodium amalgam electrolytically produced from hydroxide, resulting from the reaction of barium hydroxide on sodium sulphate. In making the first sample, 3 kilogrms. of "C.P." sodium carbonate were dissolved in pure water, filtered while cold through hardened filter-paper, and at first recrystallised four times from pure water in porcelain dishes with centrifugal washing and draining. The solution of the crystals was again filtered as before through washed filter-paper, and recrystallised five times in platinum vessels, again with centrifugal draining and washing. The solution of this ninth crop of crystals was finally filtered through the platinum mat of a carefully prepared Gooch-Munroe crucible and recrystallised, making ten crystallisations in all. The last crop of crystals, contained in a platinum dish, was dehydrated in a vacuum desiccator over fused sodium hydroxide, and was called Sample A.

As briefly suggested above, the second sample, B, was prepared in a very different way. Sodium hydroxide (previously twice recrystallised) was added in excess to twice recrystallised silver nitrate. The washed precipitate of oxide was treated with a solution of sodium bromide, which had been five times recrystallised as hydrate. The resulting sodium hydroxide solution was then once recrystallised from concentrated solution by strong cooling, and was transformed into sodium carbonate by passing pure carbon dioxide into the solution through an inverted platinum funnel. The solution was filtered through a

platinum mat and the salt was thrice crystallised. All these operations were carried out in platinum except the crystallisation of sodium hydroxide, which was carried out in silver.

For the sake of especial care, the preparation of the third sample, C, involved yet different processes. Sodium sulphate was added in slight excess to barium hydroxide, each having been previously four times recrystallised. The solution was decanted from the precipitate through a platinum filter and the resulting sodium hydroxide crystallised by cooling. A solution of this sodium hydroxide was electrolysed in an amalgamated iron dish. The solid amalgam after thorough washing was partly decomposed by water, and the solution was evaporated to crystallisation. The pure caustic alkali was transformed into carbonate, as in the case of Sample 2, and the salt thus obtained was thrice recrystallised. All operations except the electrolysis in the iron dish were carried out in platinum or silver.

The hydrobromic acid was prepared from an especially pure sample of potassium bromate, which we owed to the kindness of Edward Mallinckrodt, jun., of St. Louis. Three kilogrms. of this material, which showed only a trace of chlorine and no iodine, were three times recrystallised in porcelain; even the first crystals gave no test for chloride. Portions of the very pure product were decomposed to bromide by heating in platinum, and the bromide was fused. Proper quantities of bromate, bromide, and very pure redistilled diluted sulphuric acid were gently heated in a glass-stoppered distilling flask. The pure bromine which distilled off was twice redistilled from the similar potassium bromide and passed with hydrogen over hot platinised asbestos. The hydrobromic acid thus formed was dissolved in pure water, concentrated, and three times redistilled through a quartz condenser.

Three samples of silver were used. The first was made from nure silver nitrate by a well tested method (Richards and Wells, Carnegie Inst. of Washington, 1905, Publication No. 28, 19; Journ. Am. Chem. Soc., 1905, xxvii., 475; Zeit. Anorg. Chem., 1905, xlvi.. 74). The salt, five times recrystallised, was reduced with ammonium formate prepared from redistilled ammonia and formic acid. The metal was thoroughly washed, dried, and fused in hydrogen on a boat of pure lime made from six times recrystallised calcium nitrate. The buttons were then etched with nitric acid, washed with ammonia and much water, and dried at 400° in a vacuum.

The second sample was prepared with twice as much care, but appeared to be no purer. Silver nitrate, ten times recrystallised, was reduced in a silver dish with ammonium formate, made by passing ammonia from redistilled ammonium hydroxide, contained in a platinum retort, into formic acid, twice redistilled in quartz, contained in a gold flask. The silver powder was washed many times with purest water, then fused and dried, as in Sample I.

and in other cases where carbon dioxide must be excluded, was drawn direct from the vent supplying the filtered air to the laboratory before its entry into the room, and was washed and driven by a water blast through a long tube of copper oxide and through Emmerling towers, containing in succession solution of copper sulphate, potassium permanganate dissolved in dilute very pure phosphoric acid, pure permanganate made alkaline by potassium hydroxide which had been fused to drive out volatile substance, and concentrated potassium hydroxide, then over solid fused potassium hydroxide, and finally, where it was necessary to have especial purity and absence of moisture, over fused potassium hydroxide and resublimed phosphorus pentoxide, and over hot copper oxide, being filtered at last through a porous cup to retain dust particles (Richards and Cox, Fourn. Am. Chem. Soc., 1914, xxxvi., 819). The apparatus was of course constructed entirely of glass with fused or ground joints.

Carbon dioxide was generated in an Ostwald apparatus by the action of diluted redistilled hydrochloric acid on selected pieces of pure marble. The gas was first passed through three towers containing sodium bicarbonate solution, and afterwards through three towers of boiled concentrated sulphuric acid, to which a trace of potassium dichromate had been added, and finally through a long U-tube of resublimed phosphorus pentoxide.

In one preliminary experiment nitrogen was used. This was prepared by the Wanklyn method of passing ammonia and air over heated copper. The apparatus for this purpose was kindly loaned us by Prof. Baxter and Dr. C. J. Moore. Glass stopcocks were lubricated with a very small amount of a mixture of viscous paraffins, containing a little pure melted rubber, as suggested by Ramsay. They were so arranged that either pure air or carbon dioxide could be passed, separately or mixed, through the familiar Harvard bettling apparatus (Richards, Faraday Lect., Journ. Am. Chem. Soc., 1911, xcix., 1203; see Richards and Parker, roc. Acad. Arts and Sei., 1897, xiii., 86), and also when needed through vessels in which the solution of the sodium carbonate and the evaporation of the bromide were carried out.

Preparation of Sodium Car1onate for Analysis. The almost anhydrous carbonate, after long drying in a desiccator over fused caustic alkali, was placed in a weighed platinum boat, covered with a piece of platinum foil, and heated gently in a quartz tube attached to the bottling apparatus. A removable mica sleeve wound with a thin ribbon of a commercial high-melting alloy furnished the necessary heat, with the help of a moderate electric current. After completing the dehydration the platinum foil was removed and the boat and contents were heated for several hours at a gradually increasing temperature until the salt just fused, pure carbon dioxide being passed through the tube. As soon as fusion had taken place the application of heat ceased and pure air was admitted, gradually decreasing the amount of carbon dioxide until when the boat was cold pure air was passing over it. The trials, which was not the case when the salt was fused in nitrogen or air, a slight gain in weight being usually noted in the latter case. In the presence of air at about 850° fused sodium carbonate attacks platinum to a slight extent, especially if the salt is kept fused for a considerable time. Hoping to avoid this difficulty we used a gold boat in preliminary work, but this steadily gained in weight even after sixty hours' treatment with nitric acid and prolonged ignition in air. Returning to the platinum boat we found that in a stream of pure carbon dioxide free from air it was possible, by very gradually increasing the temperature until the salt was just fused, to prevent any significant action on the platinum boat. Thirteen fusions decreased the weight of the boat only 0.27 mgrm. or o'02 mgrm. during each fusion.

The quantitative identity of these three samples is evidence that all were as pure as possible.

In a research of this kind, involving exact neutralisation, especial pains must be taken not only with the solid substances concerned but also especially with the water and gases which are employed. Water was thrice distilled, alkaline permanganate and a few drops of sulphuric acid being used in successive distillations as usual. A thoroughly washed condenser of block tin was employed. Especial care was taken to exclude the products of combustion of the burners from the product. For much of the work, where carbon dioxide must be rigorously excluded, the water from the final distillation was condensed in such a way that it was collected cold in a stream of air wholly freed from carbon dioxide. The air used for this purpose,

There is reason to believe that sodium carbonate thus fused in a stream of pure carbon dioxide is purer than

naturally takes much longer to arrive at its maximum availability and optimum mechanical condition. It is usually kept for at least one month before shipping, but adverse weather conditions may delay shipment considerably longer, and even seriously affect the quality of the final product. The production of acid phosphate by the "open-dump" method is impracticable in the vicinity of towns or in a rich farming country unless phosphate rock very low in fluorine compounds is used, for the fumes given off during the process are not only so obnoxious as to constitute a nuisance, but are quite injurious to both animal and vegetable life. At points where these fumes do no harm, however, and where the climate is not too cold, an excellent product is obtained by this method at less cost and with less danger than by the "den " system.

The Drying of Acid Phosphate.

Acid phosphate which is carefully made, especially that produced by the "den" system, seldom requires any sub sequent drying. It is customary abroad, however, to dry superphosphates artificially, particularly when they contain an excess of phosphoric acid or are in a poor mechanical condition due to improper mixing. There are two general methods employed in drying acid phosphate. The first consists of the application of artificial heat, and the second of adding some material to take up or combine with the water or free phosphoric acid present.

In Europe a number of machines for artificially drying acid phosphate have been patented. Among the most efficient of these are the dryers of Lutjens, and of Moller and Pfeiffer (Fritsch, "Manufacture of Chemical Manures," 1911, p. 123). In both of these machines the disintegrated acid phosphate is submitted to the action of a current of hot air under pressure. No direct heat can be used in drying acid phosphate because of the tendency of the material to revert at high temperatures.

The second method of drying acid phosphate is often practised in this country when the material is too sticky or wet (due to faulty manipulation) to be uniformly spread on the field or mixed with other fertiliser ingredients. Such a condition when due to an excess of phosphoric acid can be frequently remedied by mixing the sticky mass with small percentages of phosphate rock or limestone. If the condition is due to the presence of a large proportion of iron and aluminium, the addition of finely ground peat or calcined gypsum will dry the material. In expelling the water from acid phosphate by artificial heating the value of the fuel consumed must be added to the cost of production, but no matter how the drying is done it entails additional handling, which is always expensive and should be avoided.

Storing the Acid Phosphate.

In order that the acid phosphate produced may contain a maximum quantity of soluble phosphoric acid when ready for shipment, it is usually stored in well-ventilated buildings for at least two weeks. During this time the quantity of the so-called available phosphoric acid should steadily increase. This is especially true of properly mixed acid phosphate made by the "open-dump" method where the heat is not sufficiently great to bring about rapid chemical reactions.

On the other hand, the storing of acid phosphate for protracted periods in large piles often seriously impairs its mechanical condition and sometimes its chemical composition. The pressure exerted on the material in the lower part of the heap, coupled with its contraction as the mass cools, tends to pack it. When the formation of gypsum is still in progress the superphosphate often becomes so closely cemented that it is difficult to break up. Again, improperly mixed acid phosphate or that high in compounds of iron and aluminium when closely packed is very apt to become gummy or to revert. Porter states that in making acid phosphate by the "open-dump" method the material should not be discharged on the pile until it is stiff enough to "set up" (Fourn. Ind. Eng. Chem., 1911, iii., 108).

The storing of acid phosphate in medium-sized piles, however, should cause no trouble, provided it is not allowed to stand too long and the climatic conditions are not unfavourable. Even when the material improves by storing it is hardly economical to keep it over a few months, as the interest on the money invested more than counterbalances the added value of the product due to the increase in available phosphoric acid.

In Table V. the figures of which are taken from Fritsch ("Manufacture of Chemical Manures," 1911, p. 137), the changes taking place in stored acid phosphate made from three different samples of Tennessee phosphate (A, B, and C) are shown.

TABLE V.-Changes taking place in Acid Phosphates made from Tennessee Rock on Storing from Two and a-half to Four and a-half Months. Fe2O8+Al2OS.

P205. Insoluble. Per cent.

Insoluble. Per cent.

Soluble. Per cent.

1.16

I'34

1'57

It is not stated whether the acid phosphate was made by the "open-dump" or "den" method, but an inspection of the table will show that little change has taken place in the material after keeping several months. In Table VI. are given the analyses of two piles of acid phosphate sampled after standing certain definite periods of time. The acid phosphate in both cases was made by the "open-dump" system.

TABLE VI.-Analyses of Acid Phosphate from Two Piles after Standing for certain Periods.

Although the percentage of available phosphoric acid continued to increase after storing the material for several months, this increased availability was largely offset by a corresponding rise in the moisture content of the product.

Disintegrating the Acid Phosphate.

Before acid phosphate can be bagged and shipped it must be broken up and put through coarse sieves. In the case of superphosphate which has been carefully made it often suffices to throw the material by means of shovels upon inclined screens, the force of the impact being great enough to disintegrate the lumps. When dealing with acid phosphate, however, which has been improperly made or stored for a long time, it is often necessary to use a machine for breaking up the material. The ordinary crushing devices do not answer for this purpose, owing to the tendency of the acid phosphate to pack or become sticky when pressure is applied, so disintegrators of a special type must be employed.

CHEMICAL NEWS, July 2, 1915

In a machine (of which an illustration is given in the original) complete pulverisation is brought about by submitting the material to innumerable shocks, but in such manner that no opportunity is given the acid phosphate to pack or gum together.

The disintegrator consists of a number of concentric cages made up of steel bars, all of which are enclosed in a casing. The cages are usually four in number, the first and third attached to a shaft which revolves in one direction, and the second and fourth attached to another shaft having the same axis but revolving in the opposite direction. The casing can be readily opened and the cages slid apart and cleaned.

The acid phosphate is fed through a hopper into the inner or smallest revolving cage, and is thrown by centrifugal force against the bars and into the second cage, which is revolving in the opposite direction. From the second it is thrown into the third, and then into the fourth, finally being discharged from the machine thoroughly disintegrated by the numerous impacts it has received. Two scrapers fitted to the outside cage prevent the material from adhering to the casing and clogging the machine.

After disintegration the acid phosphate is ready to be bagged or mixed with other ingredients to make a complete fertiliser.

COST OF PRODUCTION.

The cost of producing acid phosphate depends on a number of factors, which vary widely. These are the size, location, and equipment of the plant, and the cost of the sulphuric acid employed in the process.

The use of rock mills which grind the largest quantity of rock with the least expenditure of time and power and the employment of mixers having a capacity of 2 tons instead of i ton tend to reduce the cost of acid phosphate per ton. Plants located at seaports, where the cost of manufacturing sulphuric acid is less and the price of Florida rock usually lower, can often produce acid phosphate cheaper than those located at inland points. On the other hand, factories located at inland points which are within easy access of the phosphate fields can obtain their phosphate rock cheaper than those more distant from the source of supply. Again, those plants which have their own acid factories can manufacture sulphuric acid cheaper than it can be bought by companies which do not make their own acid.

The initial cost of producing acid phosphate by the den system is greater than by the open-dump method, but since the material can be shipped much sooner when made by the former method, the greater cost is compensated somewhat by the more active capital.

At inland points, such as Atlanta, Augusta, and Birmingham, the cost of producing acid phosphate (16 per cent citrate soluble), exclusive of office expenses, varies from 6.75 to 8 dols. per ton. At seaports, such as Charleston, Savannah, Baltimore, and Norfolk, the cost ranges from 6.20 to 7.50 dols. per ton. In Table VII. is given the cost of producing acid phosphate at a plant running under good conditions located at a seaport and using Florida phosphates.

The figures given in Table VII. were compiled from data obtained through personal inspection of the principal fertiliser factories of the south and east. The costs do not include overhead charges, which vary greatly according to the size and number of the plants run by a company

DISPOSAL OF PRODUCT.

Acid phosphate is sold on the basis of its so-called available phosphoric acid content, and is worth f.o.b. the factory from 40 to 56 cents per unit, depending on the location of the plant and grade of the product. (The unit is I per cent of a ton, or 20 pounds. One ton of 16 per cent acid phosphate contains 320 pounds P2O5).

The availability of phosphoric acid is determined by its solubility in a solution of ammonium citrate. There seems

9

The phosphates of South Carolina (27 per cent P2O5) and those of Northern Africa (26 to 30 per cent P205) yield, as a rule, acid phosphate containing 14 per cent available phosphoric acid.

Florida land pebble phosphate gives an acid phosphate containing 16 per cent of available phosphoric acid, and the highest grade rock from Florida, Tennessee, and certain islands in the Pacific Ocean (containing from 35 to 38 per cent P2O5) yield a product containing from 18 to 21 per cent of available phosphoric acid.

Acid phosphate is usually put up in 200 pound sacks and shipped in closed box cars. The sacks are treated with a solution of silicate of soda, paraffin, or some other substance to prevent the acid phosphate from acting upon them.

The latest official figures on the output of acid phosphate are those for 1909, which show a total production of 3,062,834 tons. It is needless to say that the production has increased enormously since these figures were compiled.

DOUBLE ACID PHOSPHATE.

The double acid phosphate now marketed contains from two to three times as much soluble phosphoric acid as ordinary superphosphate, and is therefore very valuable in the manufacture of concentrated fertilisers. Before the discovery of extensive deposits of high-grade phosphate rock, both in this country and abroad, the making of double superphosphate was widely practised, since it afforded a ready means of utilising low-grade phosphates. Now, however, most of the commercial rock is so high in phosphoric acid that it is unnecessary to resort to schemes for enriching the soluble product obtained therefrom, but in Germany, France, and several other foreign countries, as well as in the State of South Carolina, where a comparatively low grade of phosphate is mined, this process is still used with considerable success.

The two main chemical reactions involved in the manufacture of double acid phosphate are:-First, sufficient dilute sulphuric acid is added to phosphate rock to convert the hypothetical tricalcium phosphate into phosphoric acid and gypsum, and, second, the phosphoric acid thus obtained is used to convert the tricalcium phosphate of a fresh supply of rock into monocalcium phosphate. The reactions in their simplest form may be represented thus:

1. Ca3(PO4)2 + 3H2SO4 = 2H3PO4 + 3CaSO4.
2. 4H3PO4 + Ca3(PO4)2 = 3Ca H4(PO4)2.

The process, however, is by no means as simple as it at first appears, for there are several distinct operations which not only require the watchfulness of a competent superintendent but the control of a skilful chemist.

The phosphate rock and dilute sulphuric acid (16° B.) are run into a vat simultaneously, and stirred thoroughly