137 General and Physical Chemistry. Photographic Investigations of the Ultra-Violet Spark Spectra emitted by Metallic Elements and their Combina- tions under Varying Conditions. By W. N. HARTLEY (Chem. News, 48, 195-196).-It has been shown (Brit. Assoc. Jour.? 1882) that t,he spectra of metallic solutions are the same as those from metallic electrodes, the principal difference being that short lines in the spectra from the metals become long in the spectra from solutions, whilst very short lines sometimes disappear, as for example in the case of zinc. This is probably due to the solution not being able to contain a sufficient quantity of metal to yield an image of them : thus the very short lines of the aluminium spectrum are not repro- duced in solutions of the chloride unless the solutions ai-e extremely concentrated.With regard to the short lines being lengthened by moistening iridium electrodes with calcium chloride, it has now been shown that moistening with water has the same effect : hence the sup- position that a chloride of the metal was formed is untenable. The very short lines in the zinc spectrum are also lengthened by moisten- ing the electrodes with water. This variation in the spectra appears to be due t o the cooling action of the water on the negative electrode, since heating the electrodes produces a reverse result. Carbon gives two spectra in air when dry, and a third when moistened with water ; the three have been photographed, but cannot be exactly described without maps. Numerous experiments have been tried to determine which non-metallic.elements are capable of yielding spark spectra when they are combined with metals. Chlorides, bromides, iodides, snlphides, nitrates, sulphates, selenates, phosphates, carbonates, and cyanides yield none. Hydrochloric acid solutions of arsenites, areenates, and antirnonates yield spectra of arsenic and antimony respectively, afid solutions of borates and silicates yield characteristic spectra (see below) of the non-metallic constituents ; even if sodium salts are employed no metallic lines appear in the case of borates, and with silicates only the strongest sodium line ( h = 3301) is visible, even in concentrated solutions. Spark. r--" (wave-lengths) . (wave-lengths) . Boron ----Kz 3450.1 2881-0 2497.0 2631.4 2496.2 2.541.0 2528.1 2523-5 2518.5 2513.7 25013.3 2435.5 251 5.5 VOL.XLVI. Carbon spectra lines (Liveing and Dewar). r----- 7 Spark. Arc. - 2881.0 2541.0 - 2528.2 2528.1 2523-6 2523.9 2518.7 2518.8 2515.8 2515.8 2514.0 2514.1 2506.3 2506.6 2478-3 2434.8 1138 ABSTRACTS OF CHEMICAL PAPERS. It will be observed that these silicon lines are identical with those (annexed table) attributed by Liveing and Dewar (Proc. Roy. Soc., 33, 403) to carbon, and from many hundred spectra taken between graphite poles it is apparent that in the arc spectrum carbon yields but one line (2478.3 wave-length) in the ultra-violet. The ultra-violet spectrum of beryllium has been obtained from the solution of its chloride, and the following lines were observed :- Wave-lengths. 3320.1 3129.9 2649.4 2493.2 24’77.7 Description. Strong, sharp.Very strong, extended. Strong, sharp. Strong, sharp. Strong, sharp. From these observations and the general grouping of the lines, the author feels inclined to regard beryllium as the first member of the dyad series to which barium, calcium, and strontium belong. Reasons are given for not classing beryllium with other metals. D. A. L. Production of Electricity by Condensation of Aqueous Vapour. By S. K~LISCHER (Ann. Phys. Chew%. [2], 20, 614-620). -The production of electricity by condensation of aqueous vapour presents a problem of considerable meteorological importance as regards the origin of atmospheric electricity. It is, however, pro- bable that the production of electricity observed is in most cases due to the friction between the water particles and the condensing surface.Tn this paper the problem is examined experimentally by means of an apparatus which permitted the condensation of aqueous vapour by cooling. It consisted essentially of a series of beakers filled with ice, and covered externally with tinfoil; the beakers were placed on a plate of galvanised iron connected with a quadrant electrometer, and the whole combination was enclosed in a metallic box. Although deviations of the needles of the electrometer were observed, yet they were of the same magnitude and direction whether the beakers were filled with ice or not; aud secondly, they were sometimes in one, and sometimes in the other direction. Other experiments are described in which air was compressed in, and then allowed to expand from a vessel resembling the electric egg, the metallic stopcock of which was in connection with a quadrant electrometer.But in this case, although a pressure of 25 atmospheres was used, and the aqueous vapour fell in the form of fine dew on releasing the pressure, yet there was no development of electricity. V. H. V. Measurement of the Quantity of Electricity produced by a Zamboni’s Pile. By E. RIECKE (Ann. Phys. Chem. [2], 20, 512- 524) .-This paper contains a series of determinations in absolute measure of the quantities of electricity produced by three Zamboni’s piles containing a large number of platinum plates iiiterposed between strips of silk. A long series of tables of those quantities obtained on days of different relative hcmidity are given, and formulaeGENERAL AND PHYSICAL CHEMISTRY.139 for their calculation as well as for diffeibences of potential are also quoted. V. H. V. Influence of Galvanic Polarisation on Friction. B-y K . WAITZ (Ann. Phys. Chem. [2], 20, 285--303).-In 1874 Edison noticed that the friction between a metallic and a porous plate moistened with some conducting liquid, was diminished when an electric current was sent through this combination from the porous to the metallic plate. Further changes in the friction are produced by variations in the intensity of the current. This fact has been prac- tically applied in the construction of telephones and electromoto- graphs. In this memoir the phenomenon is more completely investigated. The apparatus consists in the main of a clay cylinder filled with acidulated water, and enclosed within a glass vessel filled with water of t>he same concentration. A platinum foil is introduced into the inner, and a strip of glass in the outer vessel, on which a small platinum foil is stretched ; this latter is connected with a mechaiiical arrangement whereby the platinum foil on the glass strip can be pressed against the clay cylinder with various degrees of pressure. The whole aryangement is enclosed in circuit with two Daniell's cells, a metallic arrangement to measure the degree of pressure, and a, rheostat to vary the intensity of the current.In many experiments it was found that the friction between the platinum and the porous cell is materially diminished when the intensity of the current is sufficient to decompose the acidulated water. As the contact of the platinum and the clay was not found to be sufficiently perfect, a polished glass cylinder was substituted.The alteration of friction between glass and various metals, platinum, palladium, gold, and nickel, introduced into such solutions as sulphuric acid, potash, and soda, and potassium ferrocyanide was carefully examined : in the original memoir extensive tables are given of the results obtained in the course of the investigation. As a general result it may be stated that there is a diminution of resistance when the metallic plate is the anode, but an increase when the plate is the kathode, and the intensity insufficient to decompose the electrolytic liquid. This latter fact is contrary to the experience of Koch, who found no alteration in the case of the kathode.As the diminut'ion of resistance appears only when bubbles of gas appear at the surface of the plate, it appears probable that the occlusion of the gas by the electrode is the cause of the phenomenon. Another hypothesis is that the formation of an electric double stratum, called forth by the polari- sation on the surface of the electrodes, alters the external friction between the metallic surface and the glass on the one hand, and the glass and the liquid on the other. Further experiments will decide bet ween these two hypotheses, but the writer inclines t o the latter. V. H. V. Relations between Coefficients of Friction and Galvanic Conduction. By E. WIEDEXANN (Ann. Plqs. Chenz. [ d ] , 20, 537- 538).-In connection with a remark of G.Wiedemann ou certain rela- 1 2140 ABSTRACTS OF CHEMICAL PAPERS. tions existing between coefficients of friction and galvanic conduction, and the theoretical deductions drawn therefrom, a series of investiga- tions has been made on this point. However, in order to prove how far these relations hold good, the author examined solutions of crys- talline zinc sulphate in water and aqueous glycerol of varioue concen- trations. The coefficients of friction were measured by Spring’s apparatus, those of conduction by a Wheatstone’s bridge. Ratio of coefficient Ratio of coefficient of friction of conduction Strength of solution r--h-- 7 r----- 7 1 per cent ......... 1 68.7 1 12.1 2 Y, ........ 1 29.8 1 9.52 5 9 ) ........1 6.5 1 3.68 These numbers show that the ratios existing between the co- efficients of friction and conduction are not simple, and further that the nature of the solvent and also its concentration exert a most marked influence on them. V. H. V. ZnS047H20. I n water. I n glycerol. In water. I n glycerol. Galvanic Temperature CoefRcient. By V. STOUHAL and C. BAROS (Ann. Phys. Chew. [2], 20,525-536).-Previous reseawhes have not established any differences of specific resistance as factors of the temperature corresponding to known differences of composition of various forms of iron and steel. It appeared, however, probable to the writer that iron, whose coefficient of resistance varies according to its degree of hardness and temper, should also possess various temperature coefficients.In this connection it has been shown by Matthiessen and Vogt that, the temperature coefficients of platinum- silver alloys decrease with the proportion of platinum ; like dif- ferences have been observed by the authors in the case of German silver. Similarly it is to be expected that iron, containing various proportions of carbon, which may funct.ion as the second metal, should display similar varia,tions. Experiments were accordingly made on steel tempered a t various “heats,” and specimens of bar and pig-iron. I n the case of steel the temperature coefficient of the steel varies as its coefficient of resistance continuously with the hardness of the steel, and decreases with increase of the temperature a t which it is tempered. Tlie experiments of Matthiessen and Vogt on bar-iron, and those quoted in this memoir on pig-iron, show that the temperature coeffi- cient varies with the proportions of carbon, and further, that the specific resistance of the latter is far greater than that of the hardest 5: teel.The above observations point to a marked anarlogy existing between alloys and steel in their galvanic relations, and to a general law that in all cases of homogeneous substances differences of specific resist- ance cause similar differences of temperature coefficients, increments of the former corresponding to diminutions of the latter. In alloys it is the small proportion of added metal which raises the specific re-GEXERAL AND PHYSICAL CHEMIfjTRY. 141 Boiling point. sistance, in steel the carbon produces the same effect.It is also most probable that the magnetic, as the galvanic, temperature-coefficients vary with the degree of hardness in the same way. Other experiments are also promised on the analogy existing between alloys and steel as regards their physical properties. v. H. v. Dependence of the Boiling Point on Pressure. By G. W. A. KAHLBAUM (Ber., 16, 2476--2484).--It has usually been assumed that a dinhution of pressure of 1 cm. lowers the boiling point 1". Experiments made in an apparatus in which the pressure could be kept constant whilst a considerable amount of liquid distilled, showed thak this assumption is quite erroneous. The experiments were mainly for pressures of 5-100 mm. The nature of the variations is best shown by the results obtained with some of the fatty acids.P* R~. Formic acid. Boiling point. mm. Pressure. --- 24 '84 27 *66 32 -58 41 *40 49 *66 74.54 760 -00 P. R* Boiling point. I-- -- P* R* 56.5' 57.6 63-5 68'8 69.2 7 0 4 139.4 Propionic acid. 0'112 0'111 0*1U4 0'098 0.098 0.097 - mm. Pressure. ---- 21'31 22 '46 31 -34 41 -70 44 '20 47 *30 '760 .OO 63.5" 75.2 81.4 87.5 89-8 161.5 - 0.131 0-171 0.110 0.103 0.101 - - 21.8' 22.6 24.6 2'7.9 30.5 37-6 100.6 Butyric acid. 0.1071 0.1065 0*104 0.101 0.099 0.092 - mm. Pressure. --- 10 -06 21 -48 31 *94 43 -12 48 *go 760 *OO - From these results it will be Been that the variations are peculiar to each suhstance and pressure. The figures in the third column of the tables give the ratio of the diminution of boiling point to diminu- tion of pressure (from the boiling point at 760 mm.), termed by the author the specijic remission.On obtaining by means of curves the specific remission for 0 mm. pressure, it appears that for the snb- stances experimented with (lower fatty acids, alcohols, and anhydrides) n diference in the composition of CH, corresponds with a diference 91' 0.01 in the spec@ remission f o r 0 mm. pressure. A. J. G. Heat of Combination of Carbon and Oxygen. By A. BOILLOT (Conzpt. rend., 97,41)0--491).-Thermochemical investigations should determine the quantity of heat due to each individual constituent entering into combination or being liberated by decomposition, but, up to the present the determinations made measure only the sum or the difference of the effects due to the different constituents.In the combination of carbon and oxygen, how much heat is developed or absorbed by the carbon and oxygen respectively? Let A be the heat developed by the combination of 2 vols. of oxygen (weighing 2.666) with 2 vols." of carbon vapour (weighing 1) to form 2 vols. of car- * The volume ratios and weights are given as in the original paper.142 ABSTRACTS OF CHEMICAL PAPERS bonic anhydride (weighing 3.666), B the heat developed by 1 vol. of oxygen combining with 2 vols. of carbonic oxide to form 2 vols. of carbonic anhydride, then A - B will be the heat developed by the union of 1 vol. of oxygen wit.h 2 vols. of carbon vapour with forma- tion of 2 vols. of carbonic oxide. Let x be the heat absorbed by the unit weight of carbon passing into the state of vapour occupving 2 vols., then A - 13 + a: will he the heat furnished by 1 vol.of oxygen combining with 2 vols. of carbon vapour to form 2 vols. of carbonic oxide, and A - B + 2 = B, whence 2 = 2B - A. Using 6 grams of diamond, the results will be A = 47 cal., A - €3 = 12.9 cal., B = 34.1 cal., and x = 21.2 cal. A + x = 68.2 cal. will be the total heat furnished by the combination of 2 vols. of oxygen (weighing 16 grams) with 2 vols. of carbon vapour (weighing ti grams] to form 2 vols. of carbonic anhzdride (weighing 22 grams). Of these 68.2 cal., 21.2 cal. are absorbed by the carbon. C. H. B. Sodium Alcoholates. By DE FORCRAND (Compt. rend., 97, 108- Ill).-The heats of solution of the three known alcoholates in water a t 20" are as follows :- C2H5Na0, solid + 13.47 cal.C,H5Na0,2C2H60, solid + 10.46 ,, CzH5Na0,3C2H,0, solid + 12.34 ,, and the heats of formation are therefore- 'LC2H60, liquid, + Na,O, solid, = C2H,Na0, solid, + H,O liquid .................... develops + 34-70 cal. C2H60, liquid + NaHO, solid, = C2H5Na0, solid, + H20 liquid.. ........................ 7, + 0.25 99 It would appear, therefore, that the heat of formation of sodium alcoholate from sodium hydroxide, like that of sodium glycollate, is practically nil. C,H,NaO, solid + 2C2H,0, liquid, = C,H,NaO, solid + 3c2H60, liquid, = C:H5Na0,2C2H,0, solid + C2H60, liquid, = C2H,Na0,2C2H,0, solid.. ................ develops + 8.06 cal. C2H,Na0,3CzH60, solid. ................. 9 , + 8-64! 9 1 C2H,Na0,3C2H,G, solid. ................. 9 , + 0.58 7 , A considerable proportion of this development of heat is due to the solidification of the alcohol.C,H,NaO, solid + H20 liquid = C2H60 liquid, + NaHO C2HsNa0,2C,H,0, solid, + H,O liquid = 3C,H,O liquid C2H,Na0,3C2H,0, solid + H,O liquid = 4CzH60 liquid The inverse reactions give- solid.. ........................................ + 1.19 cal. + NaHO, solid ................................ - 6.82 ,, + NaHO, solid ................................ - 7.44 ,, The decomposition of the alcoholates by water in excess takes placeGENERAL AND PHYSICAL CHEMISTRY. 143 by reason of the development of heat which accompanies the hydration of the two products of the reaction, a development which amounts to + 17.78 or + 19.78 cal., according as the amount of alcohol is 3C2H60 or 4c2H60.From these data it follows that C2H,0 iiquid + Na solid = C,H5Na0 solid + H gas develops + 32.13 cal., zt number very similar to that developed by the action of sodium on wat'er. Direct determination of the heat developed by the action of sodium in a large excess of alcohol gives the heat of solution of the anhydrous alcoholate in excess of alcohol as 12.65 cal., a value closely approaching the heat of solution of sodium hydroxide by water. These results show that water and alcohol have almost equivalent functions with respect t o sodium and sodium oxide. This fact and the dissociation of the sec0ndar.y hydrates and alcoholates explain the equilibrium which is established in liquids containing alcohol, water, and sodium oxide. C. H. B. Aqueous Solution. By J.A. GROSHANS (Ann. Phys. Cl~enz. [2], 20, 492--512).--The sp. gr. of an aqueous solution of 1 mol. of a substance in 4 mols. of water can be expressed by the interpolation formula a = 1 + in which a is the molecular weight of the substance, 18 that of water, a is a constant, and X = __ where /3 is another constant. I n this communication, certain relations exist- ing between aa for the hnlo'id salts and the nitrates of certain metals me brought out. Thus, for example, the differences of the values for zn for the iodide and bromide of strontium is equal to the difference between the same values for the iodide and bromide of mdtgnesium. The same difference exists between the values for the bromides and chlorides of magnesium and cadmium. A few exampIes are quoted below :- i 8 ( ~ + xj' a Value for an.Value for aa. SrI, .......... 283.85 MgT,.. ...... 226.93 SrBr, ........ 211.37 MgBr2.. .... 152.38 72.48 74.55 MgBr, ........ 152.38 CdBr, ...... 226.93 MgC1, ........ 78.47 CdCl, ...... 152.38 73.91 74-53 Hence it follows, for these salts, that R(12 - Br,) = R(Brll, - Cl,, = I, - Br2 = Br, - C1, = 73.80. Similarly these differences for au are the same, not only for the non-metallic but also for the metallic radicles. Thus Pb - Ba (aa) = Cd - Mg (au) = 72.48 to 74-35: so also Ba - Sr = Sr - Ca = 43.15. A series of analogous differences is quoted in this memoir, derived not only from the above formula, b u t also from modifications of it in which other constants144 ABSTRACTS OF CHEMICAL PAPERS. are introduced. The compounds more especially investigated are the halogen acids and their salts together with the nitrates of the alkaline earth and the magnesium-zinc-cadmium-group. V.H. V. Specific Gravity of Normal Salt Solutions. By C . BENDER (Ann. Plys. Chena. [ 21, 20, 560-578).-Valson has noticed (Compt. rend., 73 and 77) certain differences existing between the sp. grs. of metallic salts containing 1 gram of salt in 1 litre. For example, there is a constant difference between the sp. grs. of the potassium and ammonium salts, whatever be the non-metallic radicle associated with them, and, conversely, there is a constant difference between the sp. grs. of the chlorides and nitrates, whatever be the metallic radicle. So that to each metallic and non-metallic radicle there can be assigned a certain value or modulus with which it may be said to enter into solution. The following table explains the above statement :- c1.Br. I. I(.. .......... 1.0444 1.0800 1.1135 NH4.. ........ 1.0157 1.0520 1.0847 0.0287 0*0280 0.0288 K. Na. NH,. NOj.. ........ 1.0591 1.0540 1.0307 C1 ............ 1-0444 1.0396 1.0157 0.0147 0.0144 0.0150 If, then, the sp. gr. of one salt, preferably that of ammonium chlo- ride as the lowest, be taken as a standard, then the sp. gr. of other salts can be deduced from it, by means of the equation d = d, + mb + ms, in which d, is the sp. gr. of ammonium chloride, and nq,.rns the moduli of the metallic and non-metallic radicle respectively. Valson has also observed that similar relations exist between the refraction equivalent of the metallic salts.From these facts, the following general law may be deduced:- Elements or radicles, entering into conabination, are endowed with certain plzysizal constants which are independent of the chentical nature o f the resaiIta<nt compounds. It is, however, pointed out that these observations were restricted to salt solutions of the same strength, and presupposes the same coefficient of expansion for each of the salt solutions. I n this paper these observations are extended to solutions contain- ing one or more molecules of salt; it is shown that the modulus (vide g/,pra) divided by the number of molecules in solution is a constant. This statement is explained by the table below, in which /A. represents the number of molecules and A the modulus :- A P - A NaCl.A. - p. NH4C1. KC1. A. 1.. 1-0157 1.0444 287 2 i 7 1.0401 244 244 2 . . 1.0308 1.0887 579 289 1.0788 450 240 3.. 1.0451 1.1317 866 289 1.1164 713 238GENERAL AND PHYSICAL CHEMISTRY. 145 A series of analogous examples of a large number of metallic salts is given in the original memoir ; some of the values for the modulus in =Am units at O", deduced from various experiments, are given below :- Metal. A. Element or Radicle. A. 0 Cl.. .......... 0 NH, .......... K ............ 289 Br .......... 373 NO 163 Na .......... 238 L i . . (SO,), ........ 206 .......... 78 +Ba C2H302.. ...... -15 .......... 735 +Sr .......... 500 +Mg .......... 210 +M.n .......... 356 t Z n .......... 410 +Pb .......... 1087 From these data the following will be the general equation for calculating the sp.gr. of a salt solution containing p molecules in solution : dp = d(p.)o +p(mb + m), in which cia, mb, and mS represent the same values as in the equation above. Attention is drawn to the general interest attached to these relations, and to their wider appli- cation to the more complicated organic combinations. A New Liquid of High Specific Gravity, Refraction Equiva- lent, &c. By C. ROHRBACH (Ann. Phys. Chem. [2], 20, 169-174).- Schaffgotsch, Church, and Klein have proposed various liquids of high sp. gr. for the practical separation of minerals. In the present memoir, a solution of the double salt of barium and mercuric iodide is recommended for this purpose, which forms a highly refractive golden-coloured liquid of sp.gr. 3.575 to 3.588. It boils at about 145", evolving with the steam red vapours of mercuric iodide. It has the advantage that it neither decompose carbonates, nor absorbs carbonic anhydride from the air, but readily takes up water. Tables are given of various minerals which may be separated by means of this liquid. The following refraction values for Praunhofer's lilies were obtained :- V. H. V. F. nF-nC - C. D. E. nC. iodide sp. gr. = 1.7754: 1.7930 1.8065 1.8488 0.409 Barium mercuric 3.564 ........ 1 G and the following Fraunhofer's lines could not be measured, owing to strong absorption in the violet. The high dispersive power of the liquid is also shown by the well-defined separation of the two D lines nD1 and nD2. This liquid may possibly be available for other mineralogical and physical investigations. V.H. V. Compressibility of Gases. By E. H. AMAGAT (Ann. Chim. P h p [ 5 ] , 28, 456-464).-The author contradicts Cailletet's statements (Ann. Chim. Phys. [5], 19), (1) t,hat mercury absorbs oxygen at the146 ABSTRACTS OF CHEMICAL PAPERS. Pressure em. 50". cm. ------- 74 1-0037 72 291 1.0143 282 3 47 1.0075 143 100". cm. 200". cm. 1 300". 1-002'7 '71 1*0009 72 1-0003 1.0085 250 1*0041 287 1.0017 -__----- 1.0051 141 1.0025 143 1.0015GENERAL AND PHYSICAL CHEMISTRY. 147 carbonic anhydride was observable ; (3) within temperature o€ + 23" to - 8" a rise of temperature produces an acceleration, a fall a corre- sponding retardation of the absorption. In the three years 5.135 C.C. of carbonic anhydride under standard conditions were absorbed by 1 square metre of the glass wool.The results of these experiments are in direct contradiction to others on the same phenomenon ; this discrepancy is, however, due to the fact that in former experiments the phenomenon was supposed to be finite and not continuous. As the chemical affinity of the silicic anhydride for the basic sub- stances in t'he glass is greater than that of the carbonic anhydride, the supposition of a combination of the latter with the glass is pre- cluded ; thus it appears thak the carbonic anhydride is condensed as such. The long continuation of the phenomenon can be explained by an imperfect interpenetration of the glass by the molecules of the liquid carbonic anhydride. Further experiments will be required to prove whether after a sufficiently long interval of time this interpene- tration becomes a sufficiently diminishing quantit,y ; when this point is reached, the conditions remaining the same, neither condensation nor evaporation of the liquid carbonic anhydride would take place, although either one or the other might be caused by a change of the conditions.It would thus be not altogether impossible that for every series of temperatures there would he a corresponding series of levels of the layer of the carbonic anhydride. It was further found by experiment that atmospheric air behaves towards a glass surface exactly as carbonic anhydride. V. H. V. Similarity of the Behaviour of Ultramarine in a very fine State of Division to that OF Metallic Sulphides in the Colloidal State.By P. EBELL (Ber., 16, 2429--2432).-Ultramarine in a very finely divided state, as it is obtained by grinding and elutriation in the course of its manufacture, shows great similarity in behaviour to the colloidal metallic sulphides described by Spring (Abstr., 1883, 904). The finer parts will remain siispended in pure water for months ; the liquid passes unaltered through several thicknesses of Swedish filter paper, does not show any sign of turbidity when examined in a layer 2 cm. thick, and on evaporation leaves the ultramarine in a lustrous layer on the walls of the vessel. Microscopic examination ( x 1200) shows only points appearing partly colourless, partly pale blue by transmitted light, deep blue by reflected light. The addition of small quantities of salts, &c., to the liquid caiises the separation of the ultramarine : on washing with pure water or very great dilution, the ultramarine passes again into suspension.The author regards it as doubtful if the copper sulphide in the so-called colloidal state is really in solution in water, or whether it is not rather in suspension in a state of division far finer than that in which ultramarine ca,n be obtained by mechanical means. A. J. G . Specific Volumes of Liquid Substances. By H. KOPP (Bw., 16, 2458--2460).-The author thinks it advisable to call attention to148 ABSTRACTS OF CHEMICAL PAPERS. the fact that in his original paper (AnmaZen, 96, 155) the relations between composition and specific volume as derived from the observa- tions then accessible were not advanced as being the correct ones, but only as useful approximations. In comparing the calculated and observed volumes, it is necessary to bear well in mind the real nature of any difference : for instance, taking methyl alcohol (cal.vol. 40.8, obs. 42.1, diff. 1.3) and amyl benzoate (cal. 240 0, obs. 247.7, diff. 7*7), there is not, as a t first sight appears, a much greater divergence in the latter case, but in each case the difference is the same, = 3.2 per cent. of the calculated specific volume. The results obtained by the later experiments with bodies whose relations to one another are now more clearly known, certainly show less agreement with a law of atomic volume than was expected from the earlier experiments ; but the agreement of individual substances with many so-called laws is often more general than absolute.A. J. G.137General and Physical Chemistry.Photographic Investigations of the Ultra-Violet SparkSpectra emitted by Metallic Elements and their Combina-tions under Varying Conditions. By W. N. HARTLEY (Chem.News, 48, 195-196).-It has been shown (Brit. Assoc. Jour.? 1882)that t,he spectra of metallic solutions are the same as those frommetallic electrodes, the principal difference being that short lines inthe spectra from the metals become long in the spectra from solutions,whilst very short lines sometimes disappear, as for example in thecase of zinc. This is probably due to the solution not being able tocontain a sufficient quantity of metal to yield an image of them :thus the very short lines of the aluminium spectrum are not repro-duced in solutions of the chloride unless the solutions ai-e extremelyconcentrated. With regard to the short lines being lengthened bymoistening iridium electrodes with calcium chloride, it has now beenshown that moistening with water has the same effect : hence the sup-position that a chloride of the metal was formed is untenable.Thevery short lines in the zinc spectrum are also lengthened by moisten-ing the electrodes with water. This variation in the spectra appearsto be due t o the cooling action of the water on the negative electrode,since heating the electrodes produces a reverse result. Carbon givestwo spectra in air when dry, and a third when moistened with water ;the three have been photographed, but cannot be exactly describedwithout maps.Numerous experiments have been tried to determinewhich non-metallic. elements are capable of yielding spark spectrawhen they are combined with metals. Chlorides, bromides, iodides,snlphides, nitrates, sulphates, selenates, phosphates, carbonates, andcyanides yield none. Hydrochloric acid solutions of arsenites,areenates, and antirnonates yield spectra of arsenic and antimonyrespectively, afid solutions of borates and silicates yield characteristicspectra (see below) of the non-metallic constituents ; even if sodiumsalts are employed no metallic lines appear in the case of borates, andwith silicates only the strongest sodium line ( h = 3301) is visible,even in concentrated solutions.Spark.r--"(wave-lengths) .(wave-lengths) . Boron ----Kz3450.1 2881-02497.0 2631.42496.2 2.541.02528.12523-52518.52513.725013.32435.5251 5.5VOL. XLVI.Carbon spectra lines (Liveing andDewar).r----- 7Spark. Arc.- 2881.02541.0 -2528.2 2528.12523-6 2523.92518.7 2518.82515.8 2515.82514.0 2514.12506.3 2506.62478-32434.8138 ABSTRACTS OF CHEMICAL PAPERS.It will be observed that these silicon lines are identical with those(annexed table) attributed by Liveing and Dewar (Proc. Roy. Soc.,33, 403) to carbon, and from many hundred spectra taken betweengraphite poles it is apparent that in the arc spectrum carbon yieldsbut one line (2478.3 wave-length) in the ultra-violet.The ultra-violet spectrum of beryllium has been obtained from thesolution of its chloride, and the following lines were observed :-Wave-lengths.3320.13129.92649.42493.224’77.7Description.Strong, sharp.Very strong, extended.Strong, sharp.Strong, sharp.Strong, sharp.From these observations and the general grouping of the lines, theauthor feels inclined to regard beryllium as the first member of thedyad series to which barium, calcium, and strontium belong.Reasonsare given for not classing beryllium with other metals.D. A. L.Production of Electricity by Condensation of AqueousVapour. By S. K~LISCHER (Ann. Phys. Chew%. [2], 20, 614-620).-The production of electricity by condensation of aqueous vapourpresents a problem of considerable meteorological importance asregards the origin of atmospheric electricity.It is, however, pro-bable that the production of electricity observed is in most cases dueto the friction between the water particles and the condensing surface.Tn this paper the problem is examined experimentally by means of anapparatus which permitted the condensation of aqueous vapour bycooling. It consisted essentially of a series of beakers filled with ice,and covered externally with tinfoil; the beakers were placed on aplate of galvanised iron connected with a quadrant electrometer, andthe whole combination was enclosed in a metallic box. Althoughdeviations of the needles of the electrometer were observed, yetthey were of the same magnitude and direction whether the beakerswere filled with ice or not; aud secondly, they were sometimes in one,and sometimes in the other direction.Other experiments are described in which air was compressed in,and then allowed to expand from a vessel resembling the electric egg,the metallic stopcock of which was in connection with a quadrantelectrometer.But in this case, although a pressure of 25 atmosphereswas used, and the aqueous vapour fell in the form of fine dew onreleasing the pressure, yet there was no development of electricity.V. H. V.Measurement of the Quantity of Electricity produced by aZamboni’s Pile. By E. RIECKE (Ann. Phys. Chem. [2], 20, 512-524) .-This paper contains a series of determinations in absolutemeasure of the quantities of electricity produced by three Zamboni’spiles containing a large number of platinum plates iiiterposedbetween strips of silk.A long series of tables of those quantitiesobtained on days of different relative hcmidity are given, and formulaGENERAL AND PHYSICAL CHEMISTRY. 139for their calculation as well as for diffeibences of potential are alsoquoted. V. H. V.Influence of Galvanic Polarisation on Friction. B-y K .WAITZ (Ann. Phys. Chem. [2], 20, 285--303).-In 1874 Edisonnoticed that the friction between a metallic and a porous platemoistened with some conducting liquid, was diminished when anelectric current was sent through this combination from the porous tothe metallic plate. Further changes in the friction are produced byvariations in the intensity of the current.This fact has been prac-tically applied in the construction of telephones and electromoto-graphs.In this memoir the phenomenon is more completely investigated.The apparatus consists in the main of a clay cylinder filled withacidulated water, and enclosed within a glass vessel filled with waterof t>he same concentration. A platinum foil is introduced into theinner, and a strip of glass in the outer vessel, on which a smallplatinum foil is stretched ; this latter is connected with a mechaiiicalarrangement whereby the platinum foil on the glass strip can bepressed against the clay cylinder with various degrees of pressure.The whole aryangement is enclosed in circuit with two Daniell's cells,a metallic arrangement to measure the degree of pressure, and a,rheostat to vary the intensity of the current.In many experimentsit was found that the friction between the platinum and the porouscell is materially diminished when the intensity of the current issufficient to decompose the acidulated water.As the contact of the platinum and the clay was not found to besufficiently perfect, a polished glass cylinder was substituted. Thealteration of friction between glass and various metals, platinum,palladium, gold, and nickel, introduced into such solutions as sulphuricacid, potash, and soda, and potassium ferrocyanide was carefullyexamined : in the original memoir extensive tables are given of theresults obtained in the course of the investigation. As a general resultit may be stated that there is a diminution of resistance when themetallic plate is the anode, but an increase when the plate is thekathode, and the intensity insufficient to decompose the electrolyticliquid.This latter fact is contrary to the experience of Koch, whofound no alteration in the case of the kathode. As the diminut'ion ofresistance appears only when bubbles of gas appear at the surface ofthe plate, it appears probable that the occlusion of the gas by theelectrode is the cause of the phenomenon. Another hypothesis is thatthe formation of an electric double stratum, called forth by the polari-sation on the surface of the electrodes, alters the external frictionbetween the metallic surface and the glass on the one hand, and theglass and the liquid on the other.Further experiments will decide bet ween these two hypotheses,but the writer inclines t o the latter.V. H. V.Relations between Coefficients of Friction and GalvanicConduction. By E. WIEDEXANN (Ann. Plqs. Chenz. [ d ] , 20, 537-538).-In connection with a remark of G. Wiedemann ou certain rela-1 140 ABSTRACTS OF CHEMICAL PAPERS.tions existing between coefficients of friction and galvanic conduction,and the theoretical deductions drawn therefrom, a series of investiga-tions has been made on this point. However, in order to prove howfar these relations hold good, the author examined solutions of crys-talline zinc sulphate in water and aqueous glycerol of varioue concen-trations. The coefficients of friction were measured by Spring’sapparatus, those of conduction by a Wheatstone’s bridge.Ratio of coefficient Ratio of coefficientof friction of conductionStrength of solution r--h-- 7 r----- 71 per cent .........1 68.7 1 12.12 Y, ........ 1 29.8 1 9.525 9 ) ........ 1 6.5 1 3.68These numbers show that the ratios existing between the co-efficients of friction and conduction are not simple, and further thatthe nature of the solvent and also its concentration exert a mostmarked influence on them. V. H. V.ZnS047H20. I n water. I n glycerol. In water. I n glycerol.Galvanic Temperature CoefRcient. By V. STOUHAL andC. BAROS (Ann. Phys. Chew. [2], 20,525-536).-Previous reseawheshave not established any differences of specific resistance as factors ofthe temperature corresponding to known differences of compositionof various forms of iron and steel.It appeared, however, probable tothe writer that iron, whose coefficient of resistance varies accordingto its degree of hardness and temper, should also possess varioustemperature coefficients. In this connection it has been shown byMatthiessen and Vogt that, the temperature coefficients of platinum-silver alloys decrease with the proportion of platinum ; like dif-ferences have been observed by the authors in the case of Germansilver. Similarly it is to be expected that iron, containing variousproportions of carbon, which may funct.ion as the second metal,should display similar varia,tions. Experiments were accordinglymade on steel tempered a t various “heats,” and specimens of bar andpig-iron.I n the case of steel the temperature coefficient of the steel varies asits coefficient of resistance continuously with the hardness of thesteel, and decreases with increase of the temperature a t which it istempered.Tlie experiments of Matthiessen and Vogt on bar-iron, and thosequoted in this memoir on pig-iron, show that the temperature coeffi-cient varies with the proportions of carbon, and further, that thespecific resistance of the latter is far greater than that of the hardest5: teel.The above observations point to a marked anarlogy existing betweenalloys and steel in their galvanic relations, and to a general law thatin all cases of homogeneous substances differences of specific resist-ance cause similar differences of temperature coefficients, incrementsof the former corresponding to diminutions of the latter.In alloysit is the small proportion of added metal which raises the specific reGEXERAL AND PHYSICAL CHEMIfjTRY. 141Boilingpoint.sistance, in steel the carbon produces the same effect. It is also mostprobable that the magnetic, as the galvanic, temperature-coefficientsvary with the degree of hardness in the same way.Other experiments are also promised on the analogy existingbetween alloys and steel as regards their physical properties. v. H. v.Dependence of the Boiling Point on Pressure. By G. W.A. KAHLBAUM (Ber., 16, 2476--2484).--It has usually been assumedthat a dinhution of pressure of 1 cm.lowers the boiling point 1".Experiments made in an apparatus in which the pressure could bekept constant whilst a considerable amount of liquid distilled, showedthak this assumption is quite erroneous. The experiments weremainly for pressures of 5-100 mm. The nature of the variations isbest shown by the results obtained with some of the fatty acids.P* R~.Formic acid.Boilingpoint.mm.Pressure.---24 '8427 *6632 -5841 *4049 *6674.54760 -00P. R*Boilingpoint.I-- --P* R*56.5'57.663-568'869.27 0 4139.4Propionic acid.0'1120'1110*1U40'0980.0980.097 -mm.Pressure.----21'3122 '4631 -3441 -7044 '2047 *30'760 .OO63.5"75.281.487.589-8161.5-0.1310-1710.1100.1030.101 - -21.8'22.624.62'7.930.537-6100.6Butyric acid.0.10710.10650*1040.1010.0990.092 -mm.Pressure.---10 -0621 -4831 *9443 -1248 *go760 *OO -From these results it will be Been that the variations are peculiarto each suhstance and pressure.The figures in the third column ofthe tables give the ratio of the diminution of boiling point to diminu-tion of pressure (from the boiling point at 760 mm.), termed by theauthor the specijic remission. On obtaining by means of curves thespecific remission for 0 mm. pressure, it appears that for the snb-stances experimented with (lower fatty acids, alcohols, and anhydrides)n diference in the composition of CH, corresponds with a diference 91'0.01 in the spec@ remission f o r 0 mm.pressure. A. J. G.Heat of Combination of Carbon and Oxygen. By A. BOILLOT(Conzpt. rend., 97,41)0--491).-Thermochemical investigations shoulddetermine the quantity of heat due to each individual constituententering into combination or being liberated by decomposition, but,up to the present the determinations made measure only the sum orthe difference of the effects due to the different constituents. In thecombination of carbon and oxygen, how much heat is developed orabsorbed by the carbon and oxygen respectively? Let A be the heatdeveloped by the combination of 2 vols. of oxygen (weighing 2.666)with 2 vols." of carbon vapour (weighing 1) to form 2 vols. of car-* The volume ratios and weights are given as in the original paper142 ABSTRACTS OF CHEMICAL PAPERSbonic anhydride (weighing 3.666), B the heat developed by 1 vol.ofoxygen combining with 2 vols. of carbonic oxide to form 2 vols. ofcarbonic anhydride, then A - B will be the heat developed by theunion of 1 vol. of oxygen wit.h 2 vols. of carbon vapour with forma-tion of 2 vols. of carbonic oxide. Let x be the heat absorbed by theunit weight of carbon passing into the state of vapour occupving2 vols., then A - 13 + a: will he the heat furnished by 1 vol. ofoxygen combining with 2 vols. of carbon vapour to form 2 vols. ofcarbonic oxide, and A - B + 2 = B, whence 2 = 2B - A.Using 6 grams of diamond, the results will be A = 47 cal., A - €3= 12.9 cal., B = 34.1 cal., and x = 21.2 cal. A + x = 68.2 cal.will be the total heat furnished by the combination of 2 vols.ofoxygen (weighing 16 grams) with 2 vols. of carbon vapour (weighingti grams] to form 2 vols. of carbonic anhzdride (weighing 22 grams).Of these 68.2 cal., 21.2 cal. are absorbed by the carbon.C. H. B.Sodium Alcoholates. By DE FORCRAND (Compt. rend., 97, 108-Ill).-The heats of solution of the three known alcoholates in watera t 20" are as follows :-C2H5Na0, solid + 13.47 cal.C,H5Na0,2C2H60, solid + 10.46 ,,CzH5Na0,3C2H,0, solid + 12.34 ,,and the heats of formation are therefore-'LC2H60, liquid, + Na,O, solid, = C2H,Na0,solid, + H,O liquid .................... develops + 34-70 cal.C2H60, liquid + NaHO, solid, = C2H5Na0, solid, + H20 liquid.. ........................ 7, + 0.25 99It would appear, therefore, that the heat of formation of sodiumalcoholate from sodium hydroxide, like that of sodium glycollate, ispractically nil.C,H,NaO, solid + 2C2H,0, liquid, =C,H,NaO, solid + 3c2H60, liquid, =C:H5Na0,2C2H,0, solid + C2H60, liquid, =C2H,Na0,2C2H,0, solid.................. develops + 8.06 cal.C2H,Na0,3CzH60, solid. ................. 9 , + 8-64! 9 1C2H,Na0,3C2H,G, solid. ................. 9 , + 0.58 7 ,A considerable proportion of this development of heat is due to thesolidification of the alcohol.C,H,NaO, solid + H20 liquid = C2H60 liquid, + NaHOC2HsNa0,2C,H,0, solid, + H,O liquid = 3C,H,O liquidC2H,Na0,3C2H,0, solid + H,O liquid = 4CzH60 liquidThe inverse reactions give-solid.. ........................................ + 1.19 cal.+ NaHO, solid ................................- 6.82 ,,+ NaHO, solid ................................ - 7.44 ,,The decomposition of the alcoholates by water in excess takes placGENERAL AND PHYSICAL CHEMISTRY. 143by reason of the development of heat which accompanies the hydrationof the two products of the reaction, a development which amounts to + 17.78 or + 19.78 cal., according as the amount of alcohol is3C2H60 or 4c2H60. From these data it follows that C2H,0 iiquid + Na solid = C,H5Na0 solid + H gas develops + 32.13 cal., ztnumber very similar to that developed by the action of sodium onwat'er. Direct determination of the heat developed by the action ofsodium in a large excess of alcohol gives the heat of solution ofthe anhydrous alcoholate in excess of alcohol as 12.65 cal., a valueclosely approaching the heat of solution of sodium hydroxide bywater.These results show that water and alcohol have almost equivalentfunctions with respect t o sodium and sodium oxide.This fact and thedissociation of the sec0ndar.y hydrates and alcoholates explain theequilibrium which is established in liquids containing alcohol, water,and sodium oxide. C. H. B.Aqueous Solution. By J. A. GROSHANS (Ann. Phys. Cl~enz. [2],20, 492--512).--The sp. gr. of an aqueous solution of 1 mol. of asubstance in 4 mols. of water can be expressed by the interpolationformula a = 1 + in which a is the molecular weight ofthe substance, 18 that of water, a is a constant, and X = __ where/3 is another constant. I n this communication, certain relations exist-ing between aa for the hnlo'id salts and the nitrates of certain metalsme brought out.Thus, for example, the differences of the values forzn for the iodide and bromide of strontium is equal to the differencebetween the same values for the iodide and bromide of mdtgnesium.The same difference exists between the values for the bromides andchlorides of magnesium and cadmium. A few exampIes are quotedbelow :-i 8 ( ~ + xj'aValue for an. Value for aa.SrI, .......... 283.85 MgT,.. ...... 226.93SrBr, ........ 211.37 MgBr2.. .... 152.3872.48 74.55MgBr, ........ 152.38 CdBr, ...... 226.93MgC1, ........ 78.47 CdCl, ...... 152.3873.91 74-53Hence it follows, for these salts, that R(12 - Br,) = R(Brll, - Cl,,= I, - Br2 = Br, - C1, = 73.80.Similarly these differencesfor au are the same, not only for the non-metallic but also for themetallic radicles. Thus Pb - Ba (aa) = Cd - Mg (au) = 72.48 to74-35: so also Ba - Sr = Sr - Ca = 43.15. A series of analogousdifferences is quoted in this memoir, derived not only from the aboveformula, b u t also from modifications of it in which other constant144 ABSTRACTS OF CHEMICAL PAPERS.are introduced. The compounds more especially investigated are thehalogen acids and their salts together with the nitrates of the alkalineearth and the magnesium-zinc-cadmium-group. V. H. V.Specific Gravity of Normal Salt Solutions. By C . BENDER(Ann. Plys. Chena. [ 21, 20, 560-578).-Valson has noticed (Compt.rend., 73 and 77) certain differences existing between the sp.grs. ofmetallic salts containing 1 gram of salt in 1 litre. For example, thereis a constant difference between the sp. grs. of the potassium andammonium salts, whatever be the non-metallic radicle associated withthem, and, conversely, there is a constant difference between thesp. grs. of the chlorides and nitrates, whatever be the metallic radicle.So that to each metallic and non-metallic radicle there can be assigneda certain value or modulus with which it may be said to enter intosolution. The following table explains the above statement :-c1. Br. I.I(.. .......... 1.0444 1.0800 1.1135NH4.. ........ 1.0157 1.0520 1.08470.0287 0*0280 0.0288K. Na. NH,.NOj.......... 1.0591 1.0540 1.0307C1 ............ 1-0444 1.0396 1.01570.0147 0.0144 0.0150If, then, the sp. gr. of one salt, preferably that of ammonium chlo-ride as the lowest, be taken as a standard, then the sp. gr. of othersalts can be deduced from it, by means of the equation d = d, + mb + ms, in which d, is the sp. gr. of ammonium chloride, and nq,.rns themoduli of the metallic and non-metallic radicle respectively. Valsonhas also observed that similar relations exist between the refractionequivalent of the metallic salts.From these facts, the following general law may be deduced:-Elements or radicles, entering into conabination, are endowed withcertain plzysizal constants which are independent of the chentical natureo f the resaiIta<nt compounds.It is, however, pointed out that these observations were restrictedto salt solutions of the same strength, and presupposes the samecoefficient of expansion for each of the salt solutions.I n this paper these observations are extended to solutions contain-ing one or more molecules of salt; it is shown that the modulus (videg/,pra) divided by the number of molecules in solution is a constant.This statement is explained by the table below, in which /A.representsthe number of molecules and A the modulus :-AP- ANaCl. A. - p. NH4C1. KC1. A.1.. 1-0157 1.0444 287 2 i 7 1.0401 244 2442 . . 1.0308 1.0887 579 289 1.0788 450 2403.. 1.0451 1.1317 866 289 1.1164 713 23GENERAL AND PHYSICAL CHEMISTRY. 145A series of analogous examples of a large number of metallic saltsis given in the original memoir ; some of the values for the modulusin =Am units at O", deduced from various experiments, are givenbelow :-Metal.A. Element or Radicle. A.0 Cl.. .......... 0 NH, ..........K ............ 289 Br .......... 373NO 163 Na .......... 238L i . . (SO,), ........ 206 .......... 78+Ba C2H302.. ...... -15 .......... 735+Sr .......... 500+Mg .......... 210+M.n .......... 356t Z n .......... 410+Pb .......... 1087From these data the following will be the general equation forcalculating the sp. gr. of a salt solution containing p molecules insolution : dp = d(p.)o +p(mb + m), in which cia, mb, and mS representthe same values as in the equation above. Attention is drawn to thegeneral interest attached to these relations, and to their wider appli-cation to the more complicated organic combinations.A New Liquid of High Specific Gravity, Refraction Equiva-lent, &c.By C. ROHRBACH (Ann. Phys. Chem. [2], 20, 169-174).-Schaffgotsch, Church, and Klein have proposed various liquids ofhigh sp. gr. for the practical separation of minerals. In the presentmemoir, a solution of the double salt of barium and mercuric iodideis recommended for this purpose, which forms a highly refractivegolden-coloured liquid of sp. gr. 3.575 to 3.588. It boils at about 145",evolving with the steam red vapours of mercuric iodide. It has theadvantage that it neither decompose carbonates, nor absorbs carbonicanhydride from the air, but readily takes up water.Tables aregiven of various minerals which may be separated by means of thisliquid. The following refraction values for Praunhofer's lilies wereobtained :-V. H. V.F. nF-nC - C. D. E. nC.iodide sp. gr. = 1.7754: 1.7930 1.8065 1.8488 0.409Barium mercuric3.564 ........ 1 G and the following Fraunhofer's lines could not be measured, owingto strong absorption in the violet. The high dispersive power of theliquid is also shown by the well-defined separation of the two D linesnD1 and nD2. This liquid may possibly be available for othermineralogical and physical investigations. V. H. V.Compressibility of Gases. By E. H. AMAGAT (Ann. Chim. P h p[ 5 ] , 28, 456-464).-The author contradicts Cailletet's statements(Ann. Chim.Phys. [5], 19), (1) t,hat mercury absorbs oxygen at th146 ABSTRACTS OF CHEMICAL PAPERS.Pressure em. 50". cm. -------74 1-0037 72291 1.0143 2823 47 1.0075 143100". cm. 200". cm. 1 300".1-002'7 '71 1*0009 72 1-00031.0085 250 1*0041 287 1.0017-__-----1.0051 141 1.0025 143 1.001GENERAL AND PHYSICAL CHEMISTRY. 147carbonic anhydride was observable ; (3) within temperature o€ + 23"to - 8" a rise of temperature produces an acceleration, a fall a corre-sponding retardation of the absorption. In the three years 5.135 C.C.of carbonic anhydride under standard conditions were absorbed by1 square metre of the glass wool.The results of these experiments are in direct contradiction toothers on the same phenomenon ; this discrepancy is, however, due tothe fact that in former experiments the phenomenon was supposedto be finite and not continuous.As the chemical affinity of the silicic anhydride for the basic sub-stances in t'he glass is greater than that of the carbonic anhydride,the supposition of a combination of the latter with the glass is pre-cluded ; thus it appears thak the carbonic anhydride is condensed assuch.The long continuation of the phenomenon can be explainedby an imperfect interpenetration of the glass by the molecules of theliquid carbonic anhydride. Further experiments will be required toprove whether after a sufficiently long interval of time this interpene-tration becomes a sufficiently diminishing quantit,y ; when this pointis reached, the conditions remaining the same, neither condensationnor evaporation of the liquid carbonic anhydride would take place,although either one or the other might be caused by a change of theconditions.It would thus be not altogether impossible that for everyseries of temperatures there would he a corresponding series of levelsof the layer of the carbonic anhydride.It was further found by experiment that atmospheric air behavestowards a glass surface exactly as carbonic anhydride.V. H. V.Similarity of the Behaviour of Ultramarine in a very fineState of Division to that OF Metallic Sulphides in the ColloidalState. By P. EBELL (Ber., 16, 2429--2432).-Ultramarine in a veryfinely divided state, as it is obtained by grinding and elutriation in thecourse of its manufacture, shows great similarity in behaviour to thecolloidal metallic sulphides described by Spring (Abstr., 1883, 904).The finer parts will remain siispended in pure water for months ; theliquid passes unaltered through several thicknesses of Swedish filterpaper, does not show any sign of turbidity when examined in a layer2 cm. thick, and on evaporation leaves the ultramarine in a lustrouslayer on the walls of the vessel. Microscopic examination ( x 1200)shows only points appearing partly colourless, partly pale blue bytransmitted light, deep blue by reflected light. The addition of smallquantities of salts, &c., to the liquid caiises the separation of theultramarine : on washing with pure water or very great dilution,the ultramarine passes again into suspension. The author regards itas doubtful if the copper sulphide in the so-called colloidal state isreally in solution in water, or whether it is not rather in suspensionin a state of division far finer than that in which ultramarine ca,n beobtained by mechanical means. A. J. G .Specific Volumes of Liquid Substances. By H. KOPP (Bw.,16, 2458--2460).-The author thinks it advisable to call attention t148 ABSTRACTS OF CHEMICAL PAPERS.the fact that in his original paper (AnmaZen, 96, 155) the relationsbetween composition and specific volume as derived from the observa-tions then accessible were not advanced as being the correct ones, butonly as useful approximations.In comparing the calculated and observed volumes, it is necessaryto bear well in mind the real nature of any difference : for instance,taking methyl alcohol (cal. vol. 40.8, obs. 42.1, diff. 1.3) and amylbenzoate (cal. 240 0, obs. 247.7, diff. 7*7), there is not, as a t first sightappears, a much greater divergence in the latter case, but in eachcase the difference is the same, = 3.2 per cent. of the calculatedspecific volume.The results obtained by the later experiments with bodies whoserelations to one another are now more clearly known, certainly showless agreement with a law of atomic volume than was expected fromthe earlier experiments ; but the agreement of individual substanceswith many so-called laws is often more general than absolute.A. J. G