ethereal oxygen. This is equivalent to 3 x 9.8, and thus A NEW AND SENSITIVE INDICATOR FOR ACIDIMETRY AND the volumes of these two classes of oxygen, under the con ditions, are equal to three times the unit. volumes are The relative O" Methyl ester, H.CO2CH. 171'1 Since acetic acid is to some extent associated at the critical point, association as such exerts but little or no effect on the volume. THE ESTERS. (From van 't Hoff's text book "Lectures on Theoretical or Physical Chemistry "). HCOOCH3.. H.COOC2H5 CH3COOC2H5.. CH, COOCH,.. H.COOCHII ALKALIMETRY, AND FOR THE DETERMINATION OF BETWEEN THE LIMITS OF 6 AND 8 ON THE SORENSEN SCALE. By MARSTON TAYLOR BOGERT and GEORGE SCATCHARD. (Concluded from p. 141). Experimental. Dinitrobenzoylene Urea, CH4O6N4.- 20 grms. anthranilic acid were dissolved in 700 cc. water and 15 cc. concentrated hydrochloric acid (calculated, 12 cc.) by warming. The solution was filtered, cooled, a solution of 15 grms. KNCO (calculated, 12 grms.) in 50 cc. water added slowly with mechanical stirring, this stirring being continued for twenty minutes after all the KNCO solution had been added. The uraminobenzoic acid precipitated as a pasty white mass of microscopic needles. 300 grms. NaOH were added with cooling. This dissolved the uramino acid, and the sodium salt of benzoylene urea soon After standing for four separated in crystalline form. 228 hours the sodium salt was filtered out, dissolved in a litre of boiling water, precipitated with acetic acid, the free benzoylene urea filtered out, washed with water, and dried at 120°. Yield, 21 8 grms., or 92 per cent. product formed colourless needles, m. 353-4° (corr.), which m. could not be raised by further crystallisation. The volumes of the isomeric esters are nearly the same as those given above. Since V(CH2) is 57 = 6x9'5, V(O2) = 57 in the acids and esters. This shows that, under the conditions, O2 occupies the same volume as CH2. It follows that O" 3H, >0 = 3H. The volume of O2 in combination, 57, is very similar to its volume in the free state, 53. It is concluded that, at the critical point, oxygen possesses a relative volume of 3, except alcoholic oxygen, which occupies 2 unit volumes or stereo. And so other conditions-boiling, melting-points, solid state-oxygen tends to assume its normal volume, 2H. THE HALOGENS. (CI) 58.7 The Ten grms. of this benzoylene urea were heated on the water bath with 100 cc. concentrated sulphuric acid (which did not quite dissolve it all), and 12 cc. (calculated, 8 cc.) concentrated nitric acid (grms., 142) added. Heat was evolved, and the mixture turned bright red, but soon changed to bright yellow. After heating for an hour at 100 the mixture was cooled and poured into a litre of ice and water, the precipitate filtered out, washed with water, and recrystallised from a litre of 50 per cent acetic acid. The crystals were removed, washed, and dried at 120°. Yield, 144 grms., or 92 per cent. Subs., 0'2026, 0'2786; H2O, 00290, 00419; CO2, Calculated for C8H4O6N4-C, 38.08; H, 160. Found Subs., o 1858, o'1655; 36 60 cc. N at 17° and 765'4 22.30. As thus prepared the compound formed pale greenish yellow prisms, decomposing at 274-5° (corr.), which 4X59 3 decomposition point could not be altered by further 2 X 598 recrystallisation. 100 cc. of its aqueous solution saturated at 23° gave 00164 grm. residue at 110°, and 100 cc. of the water gave From the Journal of the American Chemical Society, xxxviii., No 8. 0 0007 grm. residue. This solubility of 00157 grm. per | each solution were made up, and each of these solutions 100 cc. is equivalent to o 00062 mols. per li re, and corresponds to 14-15 drops of oor molar indicator solution to 10 cc. of liquid to be tested. The substance is very difficultly soluble in cold alcohol, or in ether, benzene, toluene, ligroin, chloroform, carbon tetrachloride, or carbon disulphide; slightly soluble in | acetone, ethyl acetate, cold acetic acid, or boiling alcohol; moderately soluble in boiling water; readily soluble in boiling glacial acetic acid. It can be recrystallised from water or acetic acid. In solution of sodium hydroxide or carbonate it dissolves with a yellow colour, but is reprecipitated from such solutions by saturation with CO2. For the preparation of the sodium salt 25 grms. of the dinitrobenzoylene urea were dissolved in 115 cc. molar NaOH and 500 cc. boiling water, the solution filtered and cooled. A mass of long bright yellow needles crystallised out. These were removed, pressed as dry as possible, and then left over concentrated sulphuric acid in an evacuated desiccator. The rest of the dinitrobenzoylene urea was recovered by acidifying the filtrate. The so lium sat thus dried in vacuo was then heated to constant weight at 140-150°. 03195 grm. subs. lost o 019+ grm. H2O. Calculated for Found-H2O, C8H3O6N,Na.H2O-H2O, 6·17. 6.07. Subs. (dried at 140-50°), 0'3001; NaSO4, 0'0777. Calculated for CsH3O6N,Na-Na, 8.39. Found-Na, 8.38. The solubility of the salt at 20° and at 2° was determined by taking a measured volume saturated at the desired temperature, heated to boiling. acidifying with hydrochloric acid, cooling, filtering out the precipitated dinitro benzoylene urea, and drying to constant weight at 120° in a Gooch crucible. This gives the difference in solubility between the sodium salt and the free dinitrobenzoylene urea. 1. 50 cc. solution saturated at 20° gave 0.4898 grm. dinitrobenzoylene urea, indicating a solubility of 00389 mols. per litre, ог 11359 grms. CsH3O6N4Na. H2O per 100 cc. 2. 50 cc. solution saturated at 2° gave o 1294 grm. dinitrobenzoylene urea, indicating a solubility for the sodium salt of oo103 mols. per litre, or o 3008 grm. per 100 cc. In both the above calculations the solubility of free di nitrobenzoylene urea at the temperature used has been assumed as zero. This of course is not strictly accurate, as the free dinitrobenzoylene crea is itself slightly soluble but it is sufficient for all practical purposes, and those who wish closer figures can readily obtain them from the solubility results recorded above for the free dinitrobenzoylene urea. The indicator solution was prepared by dissolving 0.292 grm. of the salt in 100 cc. water, and was therefore o or molar. The dropper used delivered 22-23 drops per cc. Fifteen drops of this indicator solution in 10 cc. O'OI molar HCI gave no precipitate; 20 drops gave consider able. The solubility is thus about four times the amount ordinarily used in practice. Determination of Hydrogen-ion Concentration.-Phos phate and borate solutions were prepared as described by Sörensen. To 10 cc. of each four drops of indicator solu tion were added. That with an index of 6 gave a colour less solution, while that with an index of 8 gave a distinct greenish yellow, the colour developing evenly in the inter mediate solutions. Borate solutions with an index of 9 and 10 gave the same colour as those with 8; 11 gave a slightly deeper colour. o'i molar NaOH (index 13) gave a much deeper greenish yellow, one drop of the indicator giving as much colour as four drops in a solution with index of 8. Effect of Toluene and of Chloroform.-Phosphate solutions were prepared of index 6'4, 70, and 7.6. 30 cc. of then divided into three lots of 10 cc. each. These separate lots were then grouped into three sets of three lots each, one of each index. One set was shaken with excess of toluene and filtered. Another set was given a similar treatment with chloroform. The third set was kept as a blank. Four drops of indicator were then added to each solution. No difference could be detected between the solutions of the same concentration in the three sets. Effect of Protein.- 25 cc. of egg white were diluted to 100 cc. with water, filtered, and 2 cc. of this solution added to a mixture of 14 cc. primary and 6 cc. secondary phos. phate solution. This was then divided into two 10 cc. portions, to one of which p-nitrophenol was added, and to the other the new indicator. Both showed an index of 6.60. Salt Effect. -160 cc. secondary phosphate solution and 40 cc. primary phosphate solution were mixed, and 12 grms. NaCl (approximately molar) added. Electromotive force measurements of this solution showed an index of 6 74, while the new indicator gave 6'90. Such a solution is more concentrated than any generally measured. The effect is apparently due to the large concentration of Na+ ion, which is responsible for the pre sence of unionised salt molecules which are coloured like the ion. Fading Effect.-Three sets each were made up of phosphate solutions of index 6.6 and of 7 6. Four drops of indicator were added to one set upon a certain day, to the second set on the second day, and to the third set on the third day. On this third day no difference could be detected between the three sets of the same concentration. After standing for a week the first set was compared with freshly prepared standards with the following results:That with index of 7 6 had faded to 7'45-750, and that with index of 6 6 to 6 55-6.60. Titrations.- Titrations of o or molar NaOH and HC', using 10 drops indicator, gave 1'0011, 1'0004, 10009, and 10016 for the ratio of base to acid. Titrations of o'i molar NH4OH and HCl, using 10 drops indicator, gave 1 1257, 11262, and 1'1261 for the ration of acid to base. Titrations of o 1 molar acetic acid and NaOH, using 5 drops of indicator, gave 10129 and 10138 for the ratio of base to acid, but it was necessary to titrate by comparison with a colour standard instead of to the first appearance of a yellow colour. samples of o 005 mol. NaCl (the approximate concentra- Effect of Nitrous Acid.-30 cc. molar HCl were diluted Summary. A dinitrobenzoylene urea has been discovered whose mono-sodium salt is a very sensitive indicator fer bydrogerion concentrations between the limits of 6 and 8 on the Sörensen scale, changing from colourless to greenish yellow. Structurally, and in its behaviour as an indicator, it resembles p-nitrophenol more closely than any of the other well-known indicators. Like the latter, its chief disadvantage is its yellow colour, which renders it unsuitable for work in artificial light. It is but slightly affected by neutral salts, not at all by chloroform or toluene, proteins (egg albumen) have ng CHEMICAL NEWS, March 30, 1917 Determination of Fluorine in So.uble Fluorides more influence upon it than upon p-nitrophenol; its colour For the preparation of neutral ammonium citrate solutions, for fertiliser or soil analysis, it should prove superior to rosolic acid (commercial c ralline). It can be prepared easily from anthranilic acid by the method described. The experimental work upon which this paper is based was carried out by Mr. Geo ge Scatchard in partial fulfilment of the requirements for the degree of Doctor of Philosophy under the Faculty of Pure Science of Columbia University. Acknowledgments are also due to Dr. H. A. Fales, of th's University, for much valuable assistance and advice. 149 and the pure radio-lead treated exactly as in the case of the other samples previously described (loc. cit., p. 224). The amount of substance at hand being rather small, the work could not be done quite as accurately as before. The density determinations were made in the same pycnometer as before, by the second method described on p. 223 (loc. cit.)-the volume of the pycnometer having been redetermined because its tip had been broken in the meantime. Four identical determinations gave 5.7200 as the weight of water in the pycnometer at 19'94°, weighed in air. Therefore, the volume of the pycnometer was 5'7361 cc. Density of Lead from Cleveite. Wt. in Obs. wt. vac (W). 4'4252 44250 4'4252 4 4250 Obs. wt. 5:3287 53427 5'7361 0.3924 II 277 5.3286 53436 5 7361 0·3925 11:274 4'4252 4 4250 DENSITY OF RADIO-LEAD FROM PURE By T. W. RICHARDS and C. WADSWORTH. THROUGH the kindness of Prof. Ellen Gleditsch of the University of Kristiania, we have been so fortunate as to receive a specimen of lead sulphide from carefully selected Norwegian cleveite. According to Dr. Gleditsch, The Norwegian uraninites are very old and very unaltered. They are found in well developed crystals and occur in connection with the pegmatite dykes in south-eastern Norway." The sample in question occurred in cubic crystals near Langesund. As Hönigschmid has already pointed out (Sitz. Wien. Akad., 1914, lxxii., II.a, 20), the properties of radio lead obtained from pure minerals of this scrt are far more interesting and significant than those of lead obtained from ordinary uranium ores, which doubtless contain some ad. mixture of ordinary lead. (The name radio lead is used provisionally to designate lead which appears to be the result of radioactive transformation). Hönigschmid has shown that the lead from pure cleveite has an atomic weight as low as 206'06, and our own experience with the sample referred to above essentially confirms this result, as will be shown in another communication. So far as we know, however, the density of lead of this kind has not yet been determined, and accordingly the present paper recounts such a determination, which forms an interesting sequel to the recently published results on the density of Australian radio lead (Journ. Am. Chem. Soc., 1916, xxxviii., 221). The par fication of the sulphide, which doubtless contailed traces of sulphides of other metals, was carried out as follows: -The specimen was dissolved in nitric acid and crystallised three times with centrifuging as nitrate,a process which Baxter's experience has shown to be an excellent one for the purification of lead from other metals. (Baxter and Grover, Journ. Am. Chem. Soc., 1915, xxxvii., 5). From this purified nitrate the chloride was precipitated by pure hydrochloric acid, and this salt was crystallised three times. The final crystals, after draining on the centrifuge, were stored in a vacuum desiccator over caustic soda. The chloride thus prepared was used for the determination of the atomic weight, the density being determined in the material saved from the filtrates from that determination. These filtrates contained excess of silver, therefore enough hydrochloric acid was added to bring the concentration of the dissolved chloride ion to o'orn rnal, because at this concentration silver chloride is most nearly insoluble (G. S. Forbes, Journ. Am. Chem. Soc., 1911, xxxii., 1937). When the precipitated silver chloride had settled and had been removed by filtration through a Gooch-Munroe crucible with platinum mat, the resulting solution was concentrated and crystallised once more as nitrate. The pure crystals were electrolysed, Average, 11.273 The density of this sample, presumably a nearly pure is almost identical with that of pure lead, as indicated by IN determining fluorine gravimetrically there are several methods in use. The method of Rose (Liebig's Ann., 1849, Ixxii., 343) consists in precipitating calcium carbonate together with calcium fluoride so that the precipitate of calcium fluoride may, with some degree of satisfaction, be filtered and washed. After being ignited the calcium carbonate is dissolved out by means of 15 N acetic acid, and the residue of calcium fluoride is washed, ignited, and weighed. This procedure is open to two objections, that the calcium fluoride is very appreciably soluble in the dilute acetic acid, and that two filtration are required. Starck and Thorin (Zeit. Anal. Chem., 1912, li.) precipitate calcium fluoride along with a known weight of calcium oxalate and determine the fluorine by difference. These authors claim that the precipitate so formed is granular and easy to wash, but Adolph found it very hard to handle (Fourn. Am. Chem. Soc., 1915, xxxvii., 2500). Starck makes use of the mixed chloride and fluoride of lead (Zeit. Anorg. Chem., 1911, lxx., 173). This method is said to give good results provided that care is taken to use very little wash-water. In the attempt to precipitate fluorine so that it could be separited from fluosilicic acid by filtration, an excess of powdered calcium sulphate, CaSO4 2H2O, was used. This gave a precipitate of calcium fluoride and sulphate which was easier to filter and wash and had almost no tendency to run through the pores of the filter. It was thought that it might be possible to adapt this to the determination of fluorine in soluble fluorides. Calcium fluoride when treated with sulphuric acid is converted to sulphate, the fluorine being expelled as hydrofluoric acid. Now, if a mixture of fluoride and sulphate of calcium be similarly treated with sulphuric acid, the only change will be the conversion of the fluoride to sulphats. A grm, molecule of calcium fluoride, 78, when changed to sulphate will weigh, 136, so that an increase of 58 parts by weight will mean that the precipitate contained 78 parts of calcium fluoride. In perfecting this method there were several difficulties which arcse in connection with the filtering, ignition, &c., of the precipitates. These will be enumerated, and then will be described the expedients which were used to avoid these difficulties. 1. Since the fluoride has to be treated with acid, it cannot be filtered on asbestos because the hydrofluoric aci1 set free will attack the mineral of the filter. 2. If an ordinary filter is used particles of the precipitate adhere to it, and, when burnt along with the paper, the calcium sulphate is partly turned to sulphide and leads to incorrect results. 3. If the mixed precipitate is heated to redness in order to obtain a constant weight the mass fuses together, and it is almost impossible to completely decompose the solid mass with sulphuric acid so as to convert all of the fluoride to sulphate. 4. When the excess of sulphuric acid is being driven off so that the residue of calcium sulphate may be weighed, great care has to be exercised to prevent spattering if the heit is supplied by placing a Bunsen burner beneath the crucible. The detailed directions for the determination of fluorine by the use of powdered calcium sulphate as mentioned above will now be given, and it will be made clear how each of the above difficulties was surmounted. The solution of the fluoride, which should occupy as small a volume as practicable, say, about 30 or 40 cc., and should be neutral, is heated to boiling, and powdered calcium sulphate is added. After standing from thirty minutes to one hour, with frequent stirring, the precipitate of fluoride and sulphate of calcium is washed by decantation several times, and then put into the filter for final washing. The filter consis's of a perforated platinum crucible in the bottom of which is a small disc of ashless filter-paper, cut so as to fit exactly in the bottom without being bent up around the sides. By keeping gentle suction upon the crucible the disc is held in place, and the filtrate comes through without the least turbidity. As soon as the precipitate has been sufficiently washed it is transferred with the aid of a fine jet of water from a wash bottle to an ordinary platinum crucible; the disc of paper is washed free of the precipitate, and is ignited on the lid of the crucible, while the precipitate in the crucible is evaporated on the steam bath to dryness. If now this residue is heated to redness to obtain a constant weight it melts, and becomes very difficult to decompose with sulphuric acid. By experimenting it was found that at a temperature around 300° C. the calcium sulphate loses all of its crystal I water, and a constant weight is obtained. The proper temperature is obtained by heating the platinum crucible within an ordinary iron crucible of diameter about 3 inches at the top, used as a radiator. In order to equalise the heat a thin piece of asbestos was placed in the bottom, and upon this was placed a small triangle for the platinum crucible to rest upon. By heating the bottom of the iron crucible with a Bunsen burner to a low red the contents of the platinum crucible reach a constant weight within one hour or less. When a constant weight has thus been obtained the residue is mixed with a little water and several cc. of pure sulphuric acid. This mixture is now evaporated on the steam-bath as far as it will go at this temperature, and then, by heating further, the sulphuric acid is driven off, the last traces requiring the application of a red heat for a few moments. As was stated above there is great danger of spattering when the excess of acid is being driven off. In order to avert this danger and to permit the acid to be driven off very quickly the following method was adopted :-The lid is placed on the crucible, which is resting on a triangle. A Meker burner is fastened above, and slightly to one side of the crucible by means of an adjustable clamp. The flame of the burner is allowed to impinge from above at an angle of about 45° on the farther side of the lid of the crucible. In this way the heat can easily be regulated so that the sulphuric acid volatilis:s rapidly, and since the heat radiates from above there is almost no tendency to spatter. The residue obtained by igniting at 300° consists of a mixture of calcium fluoride and sulphate, while the residue remaining after volatilisation of the sulphuric acid consists entirely of calcium sulphate. The increase in weight of the contents of the crucible is due to the replacement of two atoms of fluorine by the sulphuric acid radical. The changes which have taken place may be represented by 2NaF CaF2 CaSO4. Thus, 84 of sodium fluoride gives 78.0 of calcium fluoride and 136 07 of calcium sulphate. Therefore 58.07 of increase corresponds to 84 of sodium fluoride and 78 of calcium fluoride, and to estimate the calcium fluoride present in the precipitate multiply the increase in weight by 78/58 07 13431, and for the sodium fluoride multiply by 84/58.07 = 1·4465. In order to test out the method which has been described in detail a solution of pure sodium fluoride was used. Commercial sodium fluoride, even that marked C.P., contains silica, and so, for a standard solution, pure hydrofluoric acid was neutralised to phenolphthalein with pure sodium hydroxide obtained by allowing moist air to act on metallic sodium in absence of carbon dioxide. This solution was diluted so that it contained about 3 grms. of sodium fluoride per 100 cc. To ascertain the exact concentration of this solution, first the method of Rose, of precipitating the fluoride along with the carbonate of calcium, was tried. This precipitate was so hard to handle in that it ran through into the filtrate, and clogged up the pores of the paper as well. that the attempt was abandoned. Since the solution of which the standard was desired contained no other compound besides sodium fluoride, its concentration was finally determined by evaporating measured portions to dryness in a platinum crucible, igniting to about 300° C., and weighing. The following results were obtained on several portions:-Ten cc. portions of the sɔdium fluoride solution gave o‘2724, 0 2721, 0.2723, 0.2725, 0.2721 grm. of sodium fluoride. One determination with 20 cc. gave 0'5450 grm. The average of all gave for the standard of the solution that 10 cc. contained 0.2724 grm. of sodium fluoride. Table VI. shows the results obtained by carrying out the determinations as outlined above. On account of the solubility of calcium fluoride there will be a tendency for the results to run low, and, unless the filtrate and washings are kept to a low volume, large negative errors are liable to occur. In several determinations the filtrate was about 200 cc., and the results here were about 14 per cent low. On account of this danger a solution saturated with pure calcium fluoride and calcium sulphate was used for wash-water in order to eliminate the solubility error, and results very close to the calculated CMICAL NEWS, Efficiency of Calcium Chloride, &c., as Drying Agents March 30, 1917 value were obtained even when the filtrate was allowed to get quite large. These results are shown in h and i of Table VI. TABLE VI. Error. Sol NaF used. Cc. Diff. +0.62 +0.25 +0.0003 -0'00017 0'27237 -0'00003 05448 03752 05427 02724 01881 0 2721 10 0 2724 01886 0.2728 -0.0010 -0'0021 The weighed precipitate of calcium fluoride and sulphate might be converted, according to Loczke (Zeit. Anal. Chem., 1910, xlix., 329), into chloride and sulphate by evaporating with hydrochloric acid, and the calcium sulphate be weighed, the fluoride being determined by difference. However, this would require another filtration and would be less accurate. 151 to test two other common drying agents, fused sodium and potassium hydroxides. The air current method was employed, the details of the apparatus and procedure being essentially identical with those described by Baxter and Warren. Calcium chloride was prepared in an anhydrous condition by fusion in an open platinum dish. (The calcium chloride was undoubtedly basic after fusion; it is, however, unlikely that the basic impurity was present in quantity sufficient to affect the results even if of a greater efficiency than the -0062 neutral salt). While still warm it was crushed to pieces -0015 the size of a small pea and packed in a glass stoppered - 0'37 U-tube about 15 mm. in diameter, the column of salt being -0.385 about 30 c.m. long. (In the paper by Baxter and Warren -0082 it is erroneously stated that the column of salt was 30 mm. +0'11 long, as in reality it was ten times this length). In most of the experiments the air was passed over a 20 per cent solution of sodium hydroxide to ensure an excess of moisture before being conducted through the calcium chloride tube. After a considerable amount of moisture had been absorbed by the calcium chloride, in a few experiments the air was first dried by means of concentrated sulphuric acid in order that the equilibrium might be approached from the reverse side. The results with calcium chloride at o° and 25° were very consistent and satisfactory, but the first experiments at 50 showed gradually higher values for succeeding experiments. Fearing that the chloride bad by this time absorbed so much water that its efficiency had been impaired, the salt was again fused. While the results then became more nearly concordant, irregularities still persisted, until the layer of active salt was doubled to about 70 cm. by inserting a second U-tube containing the fused salt. The results then showed the most satisfactory agreement. In order to find out whether the layer of chloride had been sufficiently long at the lower temperatures, additional experiments- Ncs. 4, 5, 13, 14-were carried out at the lower temperatures with the longer layer, with results identical with the earlier ones (see Table I.). The attempt was made to weigh the phosphorus pentoxide tube in which the water was absorbed to hundredths of a mgrm., by using a No. 10 Trömner balance and gold plated weights carefully standardised by the Richards method. No claim is made, however, that the accuracy is greater than a tenth of a mgrm. It was repeatedly shown that the tube remained constant in weight within this amount for periods of fourteen to sixteen hours. Since, during the precipitation of the fluoride of calcium by the sulphate, an equivalent amount of sulphate ion is set free, it is obvious that this method is adopted without any modifications to the separation and determination of fluorides in the presence of sulphates. It may also be used for the separation of fluorine from other radicals which do not form insoluble compounds with calcium. The method of Bunsen for determining fluoride in the presence of phosphoric acid would be applicable here provided that it were accurate. This consists in weighing a mixed precipitate of orthophosphate and fluoride of calcium, and then converting it to phosphate and sulphate by ignition with sulphuric acid. He claims that the final residue after ignition consists entirely of calcium ortho phosphate and calcium sulphate, but Treadwell and Koch (Zeit. Anal. Chem., xliii., 469), in experimenting to decide upon the best method for this sepiration in wines and beers, heated a weighed amount of calcium orthophosphate with sulphuric acid, as directed by Bunsen, until a constant wight was obtained. Instead of obtaining the original weight they obtained a much greater weight. Upon testing the residue they found that the precipitate contained a large amount of calcium sulphate, and that metaphosphoric acid had been formed and condensed partly on the lid of the crucible. American Journal of Science, xlii., p. 464. THE EFFICIENCY OF CALCIUM CHLORIDE, By GREGORY P. BAXTER HOWARD W. STARKWEATHER. In a recent paper upon the efficiency of certain drying agents, Baxter and Warren (Journ. Am. Chem. Soc., 1911, xxxiii., 340) make the statement that Dibbits (Zeit. Anal. Chem., 1876, xv., 159) found the aqueous vapour pressures of the lowest hydrate of calcium chloride to be 0'29 mm., 2'17 mm., and 3'50 mm. at o°, 24°, and 30° respectively. Our attention has been called by Dr. W. F. Hillebrand to the fact that these figures refer to a hydrate containing 26 per cent of water. As we have been unable to discover reliable data upon the drying efficiency of an hydrous calcium chloride, the matter has been investigated experimentally (see, however, Marden and Elliott, Journ. Ind. and Eng. Chem., 1915, vii., 320). We have found the efficiency of this salt to be far greater than the above yalues indicate, In addition, we seized the opportunity | No blank experiments were considered necessary, since in the runs with potassium hydroxide the gain in weight of the tube was scarcely perceptible. While three points are obviously insufficient to define a curve exactly, since the plot of the logarithm of the vapour pressure against the reciprocal of the absolute temperature is very nearly a straight line it seems worth while to point out that the preceding averages can be represented by the Inodification of the Antoine formula (Comptes Rendus, 1890, cx., 632; 1891, cxii., 284; 1892, cxii, 328; 1893, cxvi., 170) |