ii. 20 ATJSTrLACTS OF CIIEMICAL PAPERS. Inorganie Chemistry. The [Suggested] Formation of Helium and Neon in Geissler Tubes containing Hydrogen. ARNALDO PIUTTI (2. Elelitro- chem. 1922 28 452-453).-The author has varied the experi- inental conditions operating in Geissler tubes with a view to the production of helium and neon from hydrogen. In all 70 different experiments were carried out and variations were made in the size of the electrodes their difference apart the nature of the current etc. In no case was found any trace of the D line of helium or of the D line of ncon or any evidence of the conversion of hydrogen into neon or helium (cf. A. 1920 ii 311). Hydrogen as a Halogen in Metallic Hydrides. DWIGHT C. BARDWELL ( J . Amer. Chem. Xoc. 1932 44 2499-2504).-A new method is described for the preparation of pure sodium and potassium hydrides. Pure sodium or potassium was placed in a steel test-tube and heated to 400" by immersing in a eutectic bath of sodium and potassium nitrates.Pure hyclrogen made by passing electrolytic hydrogen through a tube containing platinised asbestos at 400" two tubes containing phosphoric oxide and finally bubbling through sodium potassium eutectic alloy was passed into the molten metal by a steel tube when the hydride formed in fine white crystals higher up the test tube where i t was scraped into glass side tubes and sealed up. Calcium hydride free from metallic calcium was similarly prepared in the form of white crystals. The absence of free metal in the above-named compounds was demon- strated by leaching with liquid ammonia which remained quite colourless.Solutions of the hydrides in a molten eutectic mixture of potassium and lithium chloride were electrolysed when it was found chiefly in the case of calcium hydride that hydrogen was liberated a t the anode in quantities strictly in keeping with Faraday's law whilst no gas was liberated at the cathode. J. F. S. CHARLES A. KRAUS and WILBUR B. DEXTER ( J . Amer. Chem. Xoc. 1922 44 2468-2 i?l).-Water having a specific conductivity of 0.12 x 10-6 W. E. G. Improved Still €or Producing Pure Water.ENORGANIC CHEMISTRY. ii. 21 0hl;zs-l may be obfainsd in large quantities by the use of a still constructed as follows. A copper boiler of 90 litres capacity is connected a t the top with a block tin tube which leads to a tinned copper chamber containing a filter of asbestos fibre or glass wool between two pieces of nickel gauze.The steam passes through the filter and a portion condenses and runs off through a drip tube. The residue of the steam passes into a settling chamber also of tinned copper and thence to a block tin worm in which the water is condensed. ohms1 a t 18" may aklo be prepared in quantity. The principle on which the preparation depends consists in allowing about 20% of the steam to pass away uncondensed. This carries with i t practically all the carbon dioxide and other volatile impurities present in the water. The still used in this case is a copper boiler as before connected directly by a block tin pipe to the settling chamber. From this the steam passes through a block tin pipe which at its lower end is surrounded by a condenser jacket to a separating chamber wiiere a large fraction about 70-80~o of the water condenses and xns away; the residual 20% of the steam carrying all the volatile impurities escapes from the top of the separating chamber.In both the above cases ordinary distilled water is placed in the boiler and this is made alkaline with sodium hydroxide and a little potassium permanganate is added. When not in use the water i n the boiler is kept near to the boiling point so that air is not dissolved in it to any large extent. Structure of Mcdecules of Water. IRVING LANGMUIR (Science 1921 53 580).-1n view of Dennison's demonstration (Physical Rev. 1921 !17 20) that ice consists of molecules of the formula H,O a strUci,ure is proposed in which the four hydrogen nuclei bind the two oxygen atoms. The duplet held by each hydrogen nucleus has one of its electrons in each of the oxygen octets instead of the more usual arrangement in which both electrons of a duplet form part of the same octet.Properties of Pure Hydrogen Peroxide. 111. 0. MAASS and W. H. HATCHER I J . Amer. Chem. Soc. 1922 44 2472-2480; cf. A. 1921 ii 106).-A continuation of previous work in which the properties of pure hydrogen peroxide have been investigated. It is shown that hydrDgen peroxide is diamagnetic and has a sus- ceptibility of 8-8 x 10-7 a value which is larger than that of water. This may be taken as evidence in favour of a structure for hydrogen peroside in which the oxygen atoms are not connected in the same way as in the oxygm molecule.The solubility of anhydrous hydrogen peroxide has been measured in a number of organic solvents with the following results a t 0". Benzene is quite im- miscible alcohol completely miscible and ether partly miscible. The solubility curves of sodium chloride sodium nitrate sodium sulphate and sucrose in anhydrous hydrogen peroxide have been determined and it is shown that the degrec of dissociation is of the same order as in water. The tendency toward the formation of molecular compounds is less than in the case of water there Water having a specific conductivity 0.05 x J. F. S. A. A. E.ii. 22 ABSTRACTS OF CHEMICAL PAPERS. being evidence of only one compound namely Na,S04,2H202. The action of the halogen hydrides on pure hydrogen peroxide and its aqueous solutions has been examined and the conditions have been determined under which the oxidation of the haloid occurs.These are to form the subject of a future paper. The halogen8 are found to be less soluble in hydrogen peroxide than in water. Am- monia dissolves in pure hydrogen peroxide with the formation of a crystalline compound which melts at 24.5" and is stable in the absence of water. This compound is slightly soluble in ether and may be formed by passing ammonia into an anhydrous solution of the peroxide in ether. A compound of the formula 2NH3,H20 was not obtained. The stability of the compound NH3,H20 shows that i t is the hydroxyl-ion which causes the decomposition of hydrogen peroxide.Water dissociates or possibly hydrolyses it and with the consequent formation of hydroxyl-ions the decom- position is cumulative. Whether or no this compound is an ammonium salt NH,O,H is not certain. The vapour pressure dielectric constant and conductivity of salts in hydrogen peroxide are under investigation and the results will be published shortly. J. F. S. Activation of Chlorine. GERSLD 1,. WENDT ROBERT 8. LANDAUER and W. W. EWING ( J . Amer. Chem. Xoc. 1922 44 2377-2382).-A number of methods have been used in an attempt to activate chlorine but neither the light from a 3000-watt Bovie mercury lamp nor a high potential electric discharge was found capable of converting clilorinc into an active form such that it would react with hydrogen in the dark. The experiments described by Draper (PhiE.Mag. 1843 [iii] 23 431; 1848 25 9 1845; 26 473; 27 327) have been repeated and confirmed. It is shown that the fact that the latent period in the hydrogen-chlorine reaction can be destroyed by previous insolation of the chlorine must be attributed to the destruction of substances which inhibit the reaction rather than to any activation of the chlorine itself. Preparation of Pare Ozone and Determination of its Mole- cular Weight. SEBASTIAN KARRER and OLIVER R. WULF ( J . Amer. Chem. SOC. 1922 44 2391-2397).-Pure ozone may be prepared by leading the gas issuing from an ozoniser into a bulb condenser cooled with liquid air until the bulb in which the liquid mixture is collected is about one-half to three parts filled with the deep blue liquid mixture of ozone and oxygen.I n this process a considerable concentration may be obtained by periodically lifting the condenser from the liquid air and so effecting a partial con- densation only. The mixture is then distilled a t -182" when for a long period oxygen containing only 0.8% of ozone passes off that is the distillation proceeds as though liquid oxygen and liquid Ozone were immiscible. At the temperature stated liquid ozone has a vapour pressure of G mm. After distillation has been carried on for some while and most of the liquid has passed off and the residue is very small in vol~imc the temperature rises slowly to -11go and then a blue gas practically pure ozone commences J. F. S.INORGANIC CEEMSTRY. ii. 23 to distil over. At this st,age of the operations extreme care is necessary if explosions are to be avoided.After the blue gas has distilled over for some time it becomes much darker in colour and this gas which is pure ozone was collected in a globe of known volume. The pressure of the gas in the bulb the weight and the pressure after decomposing the ozone by heating to ordinary temperature were determined and the molecular weight was cal- culated. As a mean of many experiments the value 47.3 was obtained. It is shown that there is no evidence to confirm the existence of tetratomic and hexatomic allotropes of oxy en. f. F. S. The Polythionic Acids. F. FOERSTER and A. HORNIG (2. nnorg. Chem. 1922 1.25 86-146).-The reaction between sul- phurous acid and hydrogen sulphide can be represented by the equilibrium H2S+H2S03 =s= H,S,O,+H,O.Further reaction with hydrogen sulphide produces sulphur whereas excess of sulphurous acid leads to the formation of pentathionic acid. Of the poly- thionic acids and their alkali salts the tetrathionates are the moat stable and the trithioriates the most unstable ; the pentathionates occupy a mean position. None of these compounds is completely stable in aqueous solution the following decompositions taking place S,O,” =+= S406:’+S. S,O,” e. S3?61’+S. S,O,”+HzO= SO;’+S2O,”+2H’. 5,0 +He + HSO +S. Themsolubihty of sulphur removes it from the equilibrium and a t boiling temperature the sulphur dioxide is removed and SO,” is alone left in solution; side reactions were also found to take place. The more sulphur dioxide and sulphur there remains in the solution the greater is the tendency of thiosulphuric acid to polymerise e.g. S,O,”+ A high concentration of hydrogen-ion retards the decomposition of tetra- and penta-thionic acids but not of trithionic acid. S,O(’H’ S40,”+ Hi30,’. S40 g’+ S20,”+ H’ += S50 6/’+ HSO,’ . W. T. Specific Volume of Liquid Ammonia. C. S. CRAGOE and D. XI. HARPER (U.S. Bur. Standards Sci. Papers 1921,420 287- 315) .-The apparatus used in the determination of the specific volume of saturated liquid ammonia is described and the results obtained for every degIee from -70“ to 100” are tabulated. These may be expressed by the equation p=[4-2830+0.s130552/133-8 where p is expressed in C.C. pey g. and 8 in “C. A. R. P. OLIVER R. WULF FARRINGTON DANIELS and SEBASTIAN KARRER (J.Amcr. Chem. Soc. 1922 4.4 239&2401).-The oxidation of nitrogen tetroxide to nitrogen pentoxide by ozone has been investigated a t 25”. The oxidation occurs almost instantaneously and in consequence a piece 0-i apparatus has been constructed whereby the oxidation can be followed titrimetricadly using the disappear- ance of brown fumes as an indication of the end-point thus carrying out a true titration in tlic gaseous phase. The results show that - 0.0082861(133 - e)]/[i + 0.424805dECCie + o.oit~38(133 - e)] Oxidation of Nitmogen Tetroxide by Ozone.ii. 24 ABSTRAC!CS OF CHEMICAL PAPERS. one molecule of ozone is required to oxidise one molecule of nitrogen tetroxide at 25" thus only one atom of the ozone molecule is avail- able in the oxidation at this temperature. This result is in striking contrast to the oxidation of sulphur dioxide to the trioxide by ozone a t temperatures below 40" in which one molecule of ozone oxidises three molecules of sulphur dioxide all three atoms of the ozone molecule being available for oxidation.J. F. S. Decomposition of Nitrogen Pentoxide in the Presence of KARRER ( J . Amer. Chem. SOC. l922,44,2402-2404).-According to the experiments of Daniels and Johnston (A. 1921 ii 249) nitrogen pentoxide decomposes rapidly a t the ordinary temperature. It is shown that in the experiments on the oxidation of nitrogen tetroxide by ozone (preceding abstract) in which a trace of ozone in excess was present a very long period of time elapsed before any visible decomposition of the pentoxide was observed. Calculations from Daniels and Johnston's results show that the brown colour ought to have returned to the gas in 15.3 minutes but actually it was found that this did not happen until the gas had been kept at 80" for 195 minutes.It follows thc?refore that a trace of ozone retards the decomposition of nitrogen pentoxide. If traces of the decom- position products are added to the system the decomposition takes place with its normal velocity. Cryoscopy of Boron Trifluoride Solutions. V. Systems with Methyl Ether and with Methyl Chloride. A. F. 0. GERMANN and MARION CLEAVELAND (Science 1921 53 582).- The melting-point curve of boron trifluoride and methyl ether shows a eutectic a t 3 mol. yo of boron trifluoride and a maximum at 50 mol. yo corresponding with the formation of the compound BF ( CH,),O previously prepared by Gasselin.The melting-point curve of boron trifluoride and methyl chloride shows two maxima at 15 mol. yo and 33 mol. % respectively of boron trifluoride and an angular point in the curve at 50%. The form of the maximum a t 150/ is identical with that a t 50% for methyl ether and the mixture yields a compound having the same f. p. and b. p. as the compound BF3,(CH3),0. It appears that methyl et,her may be present as an impurity in methyl chloride as a constant boiling mixture. The use of boron trifluoride for its detection is suggested. Ozone. FARRINGTON DANIELS OLIVER R. WULF and SEBASTIAN J. F. S. A. A. E. Catalysis in the Interaction of Carbon with Steam and Carbon Dioxide. H. A.NEVILLE and H. S. TAYLOR (Science 1921 53 677).-Potassium carbonate and other salts accelerate the interaction of carbon and steam and carbon and carbon dioxide. Reduced nickel also markedly promotes interaction of carbon and carbon dioxide. It has been shown that adsorption of carbon dioxide by carbon at 445" is increased by addition to the carbon of such accelcrating agents although the latter themselves show no adsorptive property for the gas. A. A. E.Ib'ORGANIC CIIEJIISTRT. ii. 25 Vapour Density of Technical Phosgene. A. F. 0. GERMANN and VERNON JERSEY (Science 1921 53 581).-Pure carbonyl chloride has a vapour 1;ension of about 562 mm. a t 0"; technical phosgene which contains impurities probably consisting of carbon dioxide and hydrogen chloride gives after repeated fractionation a value for the vapour density of 4-5263 g.per litre (uncorr.). A. A. E. Cryoscopy of Carl~onyl Chloride Solutions. I. System with Chlorine. A. F. 0. GERMANN and VERNON JERSEY (Science 1921 53 582).-The melting-point curve of solutions of carbonyl chloride and chlorine exhibits a eutectic point at 25 mol. yo of chlorine and angular points at Go/ 11% SO% 63% 76% and 91 yo indicating the existence of the following compounds which dissociate at the melting point 16COC1,,C12 ; 8COc1,,C1 ; COCl,,C1 ; 3COC1,,5Cl2 ; C0Cl2,3C'l ; and C0C1,,1OCl2. The mechanism of the catalytic preparation of carbonyl chloride from carbon monoxide and chlorine is briefly discussed. The Identity of Amorphous and Crystalline Silicon. WILHELM MAWCHOT (2. anorg.Chem. 1922 124 333-334).-The identity of amorphous and crystalline silicon has been proved by an X-ray examination carried out by Debye and Frauenfelder. The preparation and properties of the two forms have been given in previous papers (cf. A. 1922 ii 144 286 764). They can also be distinguished by rubbing them with lead peroxide; nothing happens in the case of the crystalline variety but the pseudo- amorphous forms ignite the brown form exploding. The greater reactivity of the latter is due to its'very fine state of division. Magnetic Analysis of Silicate$ and Silicic Acids. PAUL PASCAL (C'ompt. rend 1922 175 814--816).-Three typss of hydrated silicon dioxide were investigated magnetically all of which behaved as a mixture of anhydrous oxide and water. The conclusion is drawn that definite silicic acids do not exist (cf.van Bemmelen A. 1908 ii 838; 1909 ii 234; and Le Chatelier A. 1903 ii 1033). The Reduction of Potassium Chlorate by Ferrous Sulphate. &I. HERSCHKOWITSCH (2. anorg. Chem. 1922 125 147-154).-A dry mixture of potassium chlorate and ferrous sulphate remains unchanged up to 100"; a trace of water causes the evolution of chlorine dioxide and chlorine. In very dilute solutions the chlorate is completely reduced to the chloride. Crystd Structures of Potassium Chloroplatinite and of Potassium and Ammonium Chloropalladite. ROSCOE G . DICKINSON ( J . Amer. Chem. SOC. 1922 44 2404-2411).-Tetra- gonal crystals of po1;assium chloroplatinite and potassium and ammonium chloropalladite have been examined by X-ray methods using photographs of spectre and symmetrical and unsymmetrical Lznc photographs The simplest structure which will account for A.A. E. W. T. H. J. E. W. T.ii. 26 ABSTRACTS OF CHEMICAL PAPERS. the X-ray data has been described. In this structure each platinum or palladium atom is surrounded by four equidistant and equivalent chlorine atoms lying in one plane. The distance between the platinum atom and a chlorine atom in potassium chloroplatinitc is 2*33&0-05 A.U. It is further shown that the structure of the chloroplatinites is very closely connected with that of the cubic crystals of the chloroplatinates. The tetragonal structure is only slightly larger than the cubic one in its horizontal dimensions but considerably shorter vertically. A table is given in which the dimensions of the salts under examination are compared with those of the chloroplatinates and chloropalladates.The Hydrates of Sodium Sulphide. A. SANFOURCHE and (MLLE) A. M. LIERAUT (Bull. SOC. chim. 1922 [vl 31 966-972; cf. Sabatier A. 1879,865 and 866 ; Gottig A. 1886,980 ; Parravano and Fornaini A. 1907 ii 951).-8 thermal study of the system sodium sulphide-water reveals or confirms the existence of three hydrates Na2S,9H,O Na,S,5-5H20 Na2S,H20. The evidence for the existence of the last is less trustworthy than for the former two hydrates the reason being that experimental difficulties hinder the investigation of the system under those conditions which render the hydrate stable. The monohydrate and nonahydrate both decompose at temperatures below their respective melting points.The hydrates Na,S,H,O and Na,8,5*5Hz0 appear to undergo trans- formation into each other in the solid state. Other hydrates which have been described are not detected in this method of investigation. H. J. E. Sodium Silicate Crystals ALFEED HEINRICH ERDEN- BRECHER (2. anorg. Chem. 1922,124,339-354).-This investigation of the hydrates of sodium silicate was carried out by plotting the cooling curve of the fused salt. This on melting gives a viscous paste; it was therefore mixed with a little paraffin oil the latter having no effect on the cooling curves of sodium carbonate and sodium sulphate. The following hydrates were obtained. Na,Si03,9H,0 rhombic m. p. 47.0" ; Na,SiO3,6H,O monoclinic m. p. 62.5"; and Na2Si03,4H20 hexagonal m.p. about 85.0". The hydrate Na2Si0,,14H,0 probably exists also. The concsentration of the alkali determines which hydrate crystallises from an alkaline solution. Crystal Structure [of Csesiwn Chloride and Thallous Chloride]. WHEELER P. DAVEY and FRANCES G. WICK (Physical Rew. 1921 17 403404).-C~sium chloride is considered to be a simple cube of czsium-ions with a chlorine-ion at the centre of each cube of side 4.12 A. Thallous chloride is similarly constituted the side of the cube being 3*S5 A. These results are inconsistent with those of Bragg. The Crystal Structure of Ammonium Chloride. RALPH W. G. WYCKOPF (Amer. J . Sci. 1922 [v] 4 469475; cf. A. 1922 ii 290).-The structure of the low temperature form of ammonium chloride as determined from X-ray powder measure- J.F. S. W. T. A. A. E.INORGANIC CIIEMISTRP. ii. 27 ments is at variance with that deduccd from it study of the face development and etch figure formation of this substance. Lauc photographic data liavc been obtained by the author which are in agreement with the powder data in assigning one chemical molecule within the unit cubc. The symmetry of the crystal as a whole must be hemimorphic hemihedral (tetrahedral) or possibly holohedral but in no case can the arrangement of the atoms of the crystal agree with the enantiomorphic hemihedry which studies of face development and etch figures assign to it. The latter " chemical " means of deducing crystal structure supplies information with regard to surface of the crystal and the internal structure of the crystal is only one of a nunber of factors which Spontaneous Decomposition of Ammonium Chlorate. FRED FAIRBROTHER ( J .Arner. Chem. SOC. 1922 44 2419-2422).- About 30 g . of ammonium chlorate were kept in contact with 100 C.C. of its saturated solution a t ordinary temperature; for about three weeka the salt remained white whilst a small quantity of a colour- less gas was evolved. Later the gas became tinged with greenish- yellow and a slow but continuous evolution of gas occurred. By the beginning of the fifth week the solution had become quite yellow and the amount of solid had decreased and the rate of decomposition appeared to be increasing. The liquid gradually became darker and considerable quantities of " euchlorine " wcre evolved. The temperature rose to 3 0 4 0 " and within a few hours the system exploded.Solut'ions of ammonium chlorate in the absence of the solid may be kept indefinitely without any decom- position. The solid residue obtained from the decomposition is ammonium nitrate. It is shown that the rate of decomposition of ammonium chlorate is autocatalytically accelerated when the products of decomposition are not quickly removed. The reaction is regarded as taking place as follows. In the initial stage a very slow autoxidation of the molecule is taking place with the pro- duction of nitric acid nitrogen chlorine and water. The nitric acid reacts with more chlorate producing chloric acid and chlorine &aside. These then oxidise more of the ammonium radicle with the formation of more nitric acid and the liberation of chlorine. This takes place more rapidly than the initial stage and conse- quently the decomposition tends to increase in speed unless the acid gases are removed quickly The more stable perchlorate does not appear to be formed.WHEELER P. DAVEY and E. 0. HOFFMAN (Physical Be.. 1920 15 333).-Calcium magnesium cadmium and nickelous oxides form a face-centred cube of the metal interpenetrating a similar face-centred cube of oxygen so that the combined lattices form a simple cube of which the side is half that of the face-centred cube; the lengths of t h e side of this simple cube are 2.37 8. 2.09 h. 2.36 A. and 2.10 pi. respectively. Alumina forms a rhombohedron of which the bear on the nature of these surface phenomena. w. E. G. J. F. S. Crystal Analysis of Metallic Oxides.ii. 26 ABSTRACTS OF CHEMICAL PAPERS.hexagonal axes each have a length of 4.86 8. and the vertical axis 6-62 8. The Ammines of the Strontium Haloids. GUSTAV F. HUTTIG (2. aizo~g. Chem. 1922 124 322-332).-Uetammine strontium chloride was degraded isothermally constancy of pres- sure indicating the existence of compounds. The formation of the following ammines were thus proved [Sr(NH,),]CI (3.5") [Sr(NH3),]12 (134") [Sr(NH,)]I (204"). I n brackets are given for each complex the temperature a t which its dissociation pressure is ALFRED C . EGERTON (Nature 1922 110 773).-It has been found possible to separate zinc by fractional distillation into a distillate and a residue the densities of which bear to that of the initial metal the ratios 0.99971 1 and 1.00026 1 respectively the differences being greater than the probable experimental error.The change in the atomic weight implied by this separation is about 0.035 which is considerably less than might be expected if the metal mas com- posed of equal quantities of an isotope of atomic weight 64 on the one hand and of isotopes of atomic weight 66 68 and 70 on the other. A. A. E. W. J. BENTLY and PAUL L. STERN (Science 1921 53 143).-Ammonia free from oxygen and moisture is passed over zinc dust (previously washed with a solu- tion of ammonia and ammonium chloride alcohol and ether and dried in a vacuum) for thirts minutes a t 650" and the product cooled to at least 200" before exposure to the air. The highest yield obtained was 36.8% of nitride.A. A. E. HUBERT THONAS STANLEY BRITTON ( J . SOC. Chem. Ind. 1922 41 349- 352T) .-Various methods for the separation of glucina and alumina have been reinvestigated with the object of developing if possible a method which could be worked for the isolation of glucina from beryl. Although this object has not been completely attained it was found that more than 90% of the alumina can be easily separated from the glucina by crystallising out the former in the form of potassium alum under suitable conditions. The ore is ground to a flour and fused with potassium hydroxide. The ground product is treated with sulphuric acid and the precipitated silica separated. The filtrate containing the sulphates of glucinum aluminium and potassium with excess of sulphuric acid is adjusted to about SN-acidity by addition of potassium hydroxide as it was found that whereas the solubility of glucinum and aluminium sulphates and of alum are somewhat reduced in 5N-sulphuric acid that of potassium sulphate is considerably increased and an increase in the liquid phase of the proportion of potassium to the other sulphates caiises an increased separation of alum.The A. A. E. [Sr(NH,)]C& (45'5") [Sr(NH3) ~]Br2 (30") [Sr(NH3)2]Br2 (6s0) [Sr(NH3)]Br2 (146") [Sr(NHJ,II (31") [Sr(NH3),]& (74*5") equal to 100 rnm. w. T. Separation of tho Isotopes of Zinc. Preparation of Zinc Nitride. The Extraction of Glucina (Beryllia) from Beryl.IN ORcfUIC CHEMISTRY. ii. 29 acid liquid is therefore saturated with potassium sulphate a t the boiling temperature and set aside to crystallise at 0".The solid phase obtained under these conditions always consisted entirely of alum and represented 92.9% of the alumina present in one of the experiments quoted where the molecular proportions of glucina to alumina were 3 1 as in beryl and where 4 mols. of potassium sulphate were present in the solution for crystallisation. At 25" 87.6% of the alumina was separated. The glucina was separated from the mother-liquor by adding sodium hydroxide sufficient to redissolve the prccipitate diluting and boiling when glucinum hydroxide was precipitated. The paper contains full data and curves representing the solubility of potassium sulphate in sulphuric acid of various strengths. Constitution of Aqueous Solutions of Thallium Salts.C. DRUCKER (2. Elektrochem. 1922 28,463467).-From measure- ments of the depression of freezing point and concentration potentials of solutions of thallous nitrate it is shown that this salt gives rise on ionisation to appreciable quantities of T1,"-ions. Below C=0*06 molar the solutions follow the simple Ostwald dilution law and the association of T1' to Tl," may be neglected. At higher concen- trations all the ionic equilibria obey the law of mass action and for 2T1' e Tl," K=2-1. For TINO e Tl'+NO,' K=0*27 for all of the concentrations investigated. The mobility of the complex thallium-ion is 110. These data for the bivalent thallium-ion are in accord with the previous work on thallium sulphate solutions (cf. A. 1921 ii 161). The interpretation of the conductivities of thallous salts is satis- f actorily explained on the classical dissociation theory.The Ghosh theory of complete ionisation of electrolytes leads to erroneous results when applied to solutions of thallous nitrate. G. F. M. W. E. G. The S-re and Chemical Activity of Copper Films and the &lour Changes accompanying their Oxidation. C. N. HINSHELWOOD (Proc. Roy. SOC. 1922 [ A ] 102 318-328).- The diffraction colours produced on it copper surface by oxidation become more intense as the surface increascs in chemical activity. ,4fter repeated oxidation and reduction the permanent colour sequence is purple blue green light green (almost yellow) puryle blue black and this order of colours may be traversed time after time. The film of copper which is formed in this manner is com- posed according to Beilby of small discrete units giving an " oFen formation " or granular structure.These granules are shown by the author to possess an order of magnitude less than 1 p. The extent to which these granules are converted into oxide determines the colour of the diffracted light the colour phenomenon being independent of the thickness of the film. Thus when the granulcs are one-third copper oxide and two-thirds copper the colour of the scattered light is bright blue. No diffraction effects occur during the reduction of the oxide. The reduction phenomena arc in agreement with the view that fresh nuclei of copper grow a t theexpense of the oxide nuclei. After repeated treatment the copper film becomes incapable of further sub-division and a limiting velocity of oxidation is attained.W. E. G. Phenomena of Diflhsion in Metals in the Solid State and Cementation of Non-ferrous Metals. 111. Influence of the Vapour Pressure of the Migratory Elements in the Form- ation of the Superficial Cementated Layer. G. SIROVICH and A. CARTOCETI (Gnxzetta 1 932 52 ii 233-245).-The cementation of copper by inems of ferro-manganese or chromo-manganese (cf. A. 1922 ii 68 671) consists first of the formation at the surface of the bar of a layer of alloy containing the migratory element in definite concentration mid secondly of the flow of this element into the interior of the bar. The authors explain the first stage of the process on the assumption that a t the experimental tem- perature employed the cementated material exhibits a certain vapour or dissociation pressure in the various elements constituting it so that the medium in which the material to be cementated may after some time be regarded as saturated as regards the vapours of these elements. This idea is develoDed and the results of further experiments are brought into conformity with it.T. H. P. Phenomena of Diffusion in Metals in the Solid State and Cementation of Non-ferrous Metals. IV. Cementation of Copper by means of Ferro-aluminium. G. SIROVICH and A. CARTOCETI (Gazzetta 1922 52 g 245-249; cf. preceding abstract) .-Cementation of copper by means of aluminium like that by means of manganese may result in the penetration of the metal to considerable depths and by relatively high proportions of the migrating metal.T. H. P. Relation between Colour and Texture of Alloys. MASUMI CHIKASHIGE (2. anorg. Chem. 1922 124 335-338).-Copper reflects red to orange-yellow light the addition of tin gives a white alloy and the whiteness reaches a maximum when the composition corresponds with Cu,Sn. The colour then gradually assumes the uolour of tin which only reflects light from red to the beginning of blue. W. T. Catalysis in the Reduction of Oxides and the Catalytic Combination of Hydrogen and Oxygen. R. N. PEASE and H. S. TAYLOR (Science 1921 53 677).-Oxygen and water vapour when present in hydrogen used for the reduction of copper oxide markedly inhibit Che reaction whilst the addition of reduced copper to the oxide appears to accelerate the reaction. The induction period in the reaction is attributed (a) to initial drying of the oxide ( b ) to slow initial formation of copper which then acts as a catalyst.The mechanism of the catalytic combustion of hydrogen and oxygen in presence of copper is briefly discussed. A. A. E.MOROANIC CHEMISTRY. ii. 31 Separation of Isotopes by Distillation and Analogous Processes. ROBERT S. MULLIKEN ( J . Amer. Chem. Xoc. 1922 44 2387-2390).-A continuation of previous work on the separ- ation of the isotopes of mercury. It is shown that the separation coefficient of mercur is 0.0063 and not 0-0057 as previously stated (A. 1922 ii 295). $he progress of the separation has been followed by means of density determinations. The effect of various factors such as pressure rate of distillation and form of the apparatus on the efficiency of separation of isotopes by the method of non- equilibrium distillation has been studied and it is shown that an efficiency of 50% is readily attained. A film of dirt on the mercury increases the efficiency a t the higher rates of distillation.It is pointed out that irreversible distillation may be applied it6 a new factor in the separation of non-isotopic liquid mixtures. The failure of attempts to secure a separation of the isotopes of mercury by electrolytic solution or by reaction with sulphur is ascribcd t o the great difficulty of obtaining non-equilibrium conditioiis a t the boundary between two condensed phases. Potassium Ammonioaluminate and Ammonionianganite. FRANCIS W. BERGSTROM (Science 1921 53 $78).-Po&ssiUrrz ammonioaluminate Al(NH,),*NHK has been prepared by the action of potassamide in liquid ammonia solution on amalgamated aluminium and potassium ammoniomunganite Mn(NHK),,2NH3 by the action of an excess of potassamide on manganese thiocyanate.The Growth of the Grains in Unworked Iron and the Appearance of a Striated Structure after Heating it at more than l l O O o KARL DAEVES (2. anorg. Chem. 1922 125 167- 172).-1ron which contains but a small amount of carbon m~umes a new structure on being heated at 1125" ; this is due to the separ- ation of ferrite and pearlite in certain directions which are orientated crystallographically. Rapid cooling results in the incomplete separation of pearlite. This structure is closely associated with the development of coarse grains and the iron is deteriorated; it can be overcome by heating a t 900".W. T. L. DEDE (Z. anorg. Chem. 1922 125 2846).-The addition of phosphoric acid solution to a ferric chloride solution results in a considerable imrease in the specific conductivity of the solution; the latter also increases rapidly with the further addition of the acid solution and reaches a constant value when equivalent amounts of salt and acid are mixed. This is assunled to be due t o the formation of the complex t I ichlorophosphatoferric acid [CI,Fe(PO,)]H,. This complex acid could not be isolated; the same applies to its salts. Further addition of phosphoric acid solution whilst having no appreciable effect on the conductivity resulted in the replace- ment of the three chlorine atoms of the complex by another phos- phato-group i.e. cliphosphutoferric acid; this wm isolated and analysed ; it corresponds with [Fe(P04),]H,+S$H,0.The com- plex is very stable and the author pomts out its application to J. F. S. A. A. E. Complex Phosphato- and Arsenato-metal Acids.ii. 32 ABSTRAC!L'S OR' CHEIIIICAL PAPERS. analysis. Arsenic acid gave analogous compounds. Aluminium and glucinum salts also show a considerable increase in specific conductivity on being mixed with a solution of phosphoric acid. These are being further investigated. Behaviour of the Stannic Acids towards Solutions of Alkaline Hydroxides. GEORGE ERNEST COLLINS and JOHN I~EEFOOT WOOD (T. 1922 121 2760-2765). Crystal Structure of Titanium Zirconium Cerium Thorium and Osmium.ALBERT W. HULL (Physical Rev. 1921 18 88-89).-Tifanium crystallises in the hexagonal system holohedra.1 class with axi@ ratio 1.59&0'01. The side of the unit triangular prism is 2.97 A. and its height 4-72 8. The lattice is made up of two sets of these prisms the atoms of one set being in the centre of the prisms of the other set. Zirconium has a similar structure,owith axial rat'io also 1.!S. The side of the unit triangle is 3.23 A. and its height 5.14 A. Cerium has a similar structure with axial ratio 1-62. T,he side of the elementary tri- angle is 3.65 A. and height 5.96 A. A face-centred cubic form with sido of cube 5-12 A. is also present but may possibly be due to the presence of impurity. Thorium has a face-centred cubic lattice with side of cube 5.04 a. Osmium crystallises in a face- centred cubic lattice with axia'l ratio 1.59. The side of the unit triangle is 2.714 A. and its height 4.32 8. The Isotopes of Antimony. F. W. ASTON (Nature 1922 110 732).-The mass-spectrum of antimony obtained by the use of antimony trimethyl gives two lines at 121 and 123 respectively the former being the more intense by 10-20~o. The masses of these isotopes of antimony are most probably less than whole numbers by one to two parts per thousand. The results are in excellent agreement with the atomic weight (121 *77) obtained by Willard and McAlpine (A 1921 ii 405) but not with the accepted figure (120.2). A. A. E. The Alleged Variable Composition of Triple Chlorides containing Silver and Gold. HORACE L. WELLS (Amer. J . Sci. 1922 [v] 4 476-482; cf. A. 1922 ii 449 514).-From Bayer's analyses of cssium silver gold chloride (A. 1920 ii 688) the conclusion is drawn that the triple salt Cs2AgAuC1 is a definite invariable compound and that Bayer analysed mixtures of this salt with Cs,Au,Cl, and CsAuC1,. The rubidium salt Rb,Ag,Au,CI,,~ is also invariable. It is probable also that Suschnig (A. 1922 11 014) analysed mixtures in his investigation of the triple bromides of rubidium silver and gold. Hydrolysis of Platinum Salts. 11. Potassium Platini- bromide. EBEN HENRY ARCHIBALD and WILLIAM A. GALE (T. W. T. A. A. E. W. E. G. 1922 121 2849-2857).