The expression T is equivalent to the first term of Clausius's and ᎡᎢ Van der Waals' formulæ, but it is by no means so simple. It (v-B)' would appear that b is not any simple function of the volume of the gaseous or liquid matter; and we are still in ignorance of the true relations between volume on the one hand, and pressure and temperature on the other. With constant volume, however, temperature and pressure have the simple relation already mentioned.

The behaviour of acetic acid was also investigated, and the data furnished by E. and L. Natanson (Abstr., 1886, 657) for nitric peroxide were also made use of. These two substances exhibit a behaviour markedly different from that of stable bodies. For whereas the isochoric lines (or lines of equi-volume) are straight with stable substances, when mapped against temperature and pressure as ordinates and abscissæ, those of acetic acid and nitric peroxide are curves of double flexure, tending at higher temperatures and pressures to become tangential to isochoric lines, calculated on the theoretical assumption that acetic acid and nitric peroxide have respectively the simpler formulæ C2H,O2 and NO,; and at lower temperatures and pressures approaching, as if they would ultimately touch, the theoretical isochoric lines of the more complex molecules C.H2O, and N2O1. The gradual dissociation of these molecules is thus traced, and these substances show and exhibit no analogy with stable substances. It cannot therefore now be asserted that the abnormal vapour-density of these compounds is ascribable to the same cause as the high vapour-density of stable substances at high temperatures and correspondingly high pressures.

The paper also shows the deviation from Boyle's and Gay-Lussac's laws exhibited by ether. It is to be noted that at any given volume smaller than 300 c.c. per gram, the pressure of vapour is below what is calculable on the assumption that ether gas is "perfect." At greater volumes, this divergence is barely noticeable by experiment. Keeping the volume constant, if the temperature and consequently the pressure be raised, there is a continually decreasing difference between the found and what may be called the "theoretical" pressure. At a certain sufficiently high temperature, the found and "theoretical" pressures are identical. At a still higher temperature, the "theoretical" pressure is below the found pressure. The physical meaning of this fact appears to be as follows:-The pressure of a gas depends on the number of molecules present in unit volume, on the average velocity of each molecule, and on the number of impacts on unit area of the surface of the containing vessel in unit time. With constant volume, since the mean distance between the molecules remains constant (on the assumption that the individual molecules are incompressible), the cohesion of the molecules is assumed to be constant. But the rise of pressure produced by rise of temperature of a theoretical gas is based on the assumption that each impact takes place at the centre of each molecule; that is, that the actual volume of the molecules themselves is nil. But as this is not the case, as impacts must take place at some distance from the centres of the molecules, they must necessarily be more frequent. The effect of cohesion is to reduce the pressure of the gas, by reducing the average velocity of the molecules, and this, for

any given volume, by a constant amount. Hence below a certain temperature the pressure will be less than that of a normal gas, and if the temperature be reduced sufficiently, will become negative. With rise of temperature the average velocity of each molecule will increase at the same rate as in the case of a perfect gas, but the number of impacts, owing to the increased chances of collision, and consequently the pressure, will increase at a greater rate than if the gas were perfect. Hence a temperature will ultimately be reached when the pressure will be as much decreased by cohesion as it is increased by the more frequent encounters of the molecules; and at that temperature the density of the gas will be normal. At still higher temperatures, the pressure, and therefore the value of the expression pt/v, will be greater, and the vapour-density less than that of a perfect gas. That the numerical value of the expression pt/v would ultimately exceed unity has indeed been experimentally proved by Natterer, in his experiments on the compression of the so-called permanent gases, at temperatures far above their critical points.

In conclusion, it is pointed out that the equations of Clausius, van der Waals, and Sarrau do not express the true relation between volume, temperature, and pressure. As before stated, the relation of temperature and pressure to volume is by no means so simple as they represent it to be.

W. R.

Compressibility of Solutions of Gases. By F. ISAMBERT (Compt. rend., 105, 375–377).—All known facts indicate that the solution of a gas in a liquid is a very complex phenomenon, and these experiments were undertaken with a view to throw some light on the problem of the nature of solution. The coefficients of compressibility obtained were as follows:

Water: 0.0000443 at 20°.

Ammonia (330 litres of gas per litre of liquid): −0·0000381 at 21 -22.5°.

Ammonia (140 litres of gas per litre of liquid): −0·0000389 at 20-4 -22.2°.

Hydrochloric acid (at 22°): -0.0000366 at 19-6-20·5°.

Alcohol: -0·0001076 at 19°.

Ammonia in alcohol (87 litres per litre): -0.0001071 at 18.0— 19.7°.

Sulphurous anhydride in alcohol (60 litres of gas per litre of solution): -0.0001031 at 18-9-20·3°.

Ether: −0·000183 at 21·5°.

Ammonia in ether: -0.000185 at 21-22°.

The solution of hydrogen chloride, which is usually regarded as containing definite hydrates, has a compressibility less than that of water, and in this respect resembles saline solutions. An aqueous solution of ammonia behaves in the same manner. The volume of aimonia dissolved by ether is too small for any conclusion to be drawn from this case, whilst the compressibility of the solution of ammonia in alcohol is practically identical with that of alcohol, and that of sulphurous anhydride in alcohol is slightly lower. The mere dissolution of a gas in a liquid has very little effect on the compressibility of

the latter. An aqueous solution of ammonia, however, behaves like a true compound. C. H. B.

Solution. By W. DURHAM (Chem. News, 56, 152-153).-In this paper the author draws attention to the support the theory, promulgated by him in 1878, receives from the thermochemical results accumulated of late years. According to this theory, dissolution of a substance in a liquid is due to the chemical affinity of the elements of the substance dissolved for the elements of the solvent, hence common salt dissolves in water, on account of the affinity of sodium for oxygen and chlorine for hydrogen; further chemical affinity is not exhausted in all cases of chemical combination, but sometimes sufficient affinity remains to form "solution compounds."

In support of these statements, he points out that as regards chlorides, bromides, iodides, sulphates and nitrates, the available data indicate that the heat of dissolution varies directly; firstly, as the heat of combination of the positive element of the salt with oxygen in water varies; secondly, as the heat of combination of the negative element with the hydrogen varies; but that it varies inversely as the heat of combination of the positive and negative elements of the salt varies. Moreover, when salts differing in heats of combination are dissolved, the heats of dissolution differ to a similar extent, the greatest heat of dissolution being associated with the lowest heat of combination.

The author shows that the absolute amount of heat of dissolution appears to arise from a balancing of affinities among the constituent elements, for instance, when M,Cl2- [M,OAq + Neutr.] = H2, Cl2, Aq — H2O there is no heat of dissolution, and the salt is not soluble. It also would appear that, in neutralising an oxide by an acid solution, the operation is incomplete as long as the salt remains in solution; positive heat of dissolution indicating that the oxide and acid are not completely decomposed, whilst negative heat would indicate that the salt and water, resulting from the double decomposition, are not completely formed. When both parts are complete there is insolubility. In the case of sulphates, when the heat of combination of the oxide with the sulphuric anhydride is equal to the heat of combination of the metal with sulphur, there is insolubility, but when the former is less than the latter solubility appears. Examples are given in the original paper illustrating all the above statements and deductions. The author suggests that dissolution is probably a periodic function of the elements. D. A. L.

Solution. By S. U. PICKERING (Chem. News, 56, 181-182).— The author, commenting on Durham's paper (preceding Abstract), complains of his (Durham's) reasoning in a circle when drawing his conclusions from the thermochemical data. Moreover, the author warns those in search of laws in thermochemistry as to the uncertainty attached to experimental numbers which agree absolutely with or come even within 100 cals. of the theoretical.

D. A. L.

Nature of Solution. By S. U. PICKERING (Chem. News, 56, 191 -192).-Referring to a recent communication by Nicol (ibid., p. 162), the author states that the hypothesis on which Nicol bases his argument of the vapour-pressures being a measure of the heat of dissolution is a pure assumption, and one for which we have no particle of evidence.

In a few subsequent remarks, he endeavours to clear up some of the points in dispute between himself and Nicol. D. A. L.

Salt Solutions. By C. BENDER (Ann. Phys. Chem. [2], 31, 872888). The author describes further experiments on "corresponding" solutions (Abstr., 1885, 12). Tables are given showing the volumes, densities, and coefficients of expansion of solutions containing from 1 to 6 gram-molecules, or half-molecules, of lithium, ammonium, and barium chlorides per litre. By the aid of these and previous tables, the author has determined by trial the strengths of solutions which, when mixed in equal volumes, furnish solutions of which the densities and coefficients of expansion are the means of those of the constituents. The relative strengths of corresponding solutions are indicated by the values of the coefficient u. n = number of grammolecules or half-molecules per litre

NaClμ = n; KClμ = n; (BaCl2)μ = n; NH ̧Clμ = n; LiClμ = zn.

Solutions correspond electrically when the means of their specific resistances and conductivities most nearly approach the resistances and conductivities of their mixtures in equal volumes. Electrical measurements were made by Kohlrausch's method, and are given in tables for each salt. These solutions contain

NaClμ = n; LiCl = n; †(BaCl2)μ = 2; KClμ = {n; NH ̧C1⁄4μ = ‡n•

The resistance of a mixture of two solutions is almost always less than the arithmetic mean of the resistances of the separate solutions. For corresponding solutions, the difference is a minimum.

A similarly simple law does not hold for the conductivities of mixtures. Thus solutions of NaCl and NH,Cl, and of KCl and BaCl2, correspond as regards conductivity in two very different relative states of concentration. The idea of correspondence is therefore based primarily on resistance.

The author answers the objection of Arrhenius (Abstr., 1887, 415), that correspondence is based on a purely arithmetical relation, and points out that this relation is more general, and not more arbitrary, than that on which Arrhenius founds his idea of "isohydric solutions.

Ch. B.

Compressibility of dilute Salt Solutions and of solid Sodium Chloride. By W. C. RÖNTGEN and J. SCHNEIDER (Ann. Phys. Chem. [2], 31, 1000-1005).-Schumann (Abstr., 1887, 696) states that weak solutions of potassium and calcium chlorides at 15°, and of ammonium and strontium chlorides at 0°, are more compressible than pure water. In their previous work (Ann. Phys. Chem.

[2], 29, 165), the authors have not observed such an anomaly. They have therefore repeated Schumann's experiments as nearly as possible, but with entirely opposite results. No details are given.

The authors also state that they have found no perceptible difference between the compressibilities of air-free water and of water saturated with air at ordinary temperature and pressure. The difference, at any rate, is less than 0.2 per cent.

By methods already described, they have also found the relative apparent compressibility of solid sodium chloride 0049. This value is not very different from the number 0044, which is obtained when the interpolation formula for the relative apparent compressibility of sodium chloride solution (loc. cit.) is extended beyond the limits of solubility of the chloride, and n put = co. It may perhaps be possible in this way from the known compressibility of a solution to calculate, at least approximately, that of the dissolved solid. The above direct determination gives for the true compressibility 5.0 × 10-6. By calculation from the interpolation formula, the number 48 x 10-6, and from the formula for the relative molecular compressibility the number 4.7 x 10-6, are obtained.

These numbers differ greatly from that obtained by Braun (Abstr., 1887, 436), namely, 14 x 10-6. The authors therefore give details of their own experiments. They have further made determinations, using comparatively low pressures, in order to eliminate errors due to inclusion of air in crevices in the crystals. These led to the above number 0049. Ch. B.

Water of Crystallisation of dissolved Cobalt Salts. By J. KALLIR (Ann. Phys. Chem. [2], 31, 1015-1028).-It is known that the dehydration by heat of cobalt chloride in solution is effected at a lower temperature when the solution is saturated with sodium chloride. The author has studied this reaction quantitatively by observing the accompanying changes in the absorption-spectrum in the neighbourhood of the D line. Let A be the absorption coefficient for a known solution of the hydrated salt of thickness d, E that of the same solution when dehydration is complete, and M the coefficient when at any temperature a fraction 1/ has been converted. In the latter case, the absorption will be equivalent to that produced by a thickness da of dehydrated solution, and a thickness d-d/x of unchanged solution, whence

A beam of parallel rays from a petroleum lamp, or from the sun, was directed perpendicularly upon a glass trough with parallel sides containing the solutions under experiment (0-0746 gram of CoCl, per litre). The coefficients were measured by Glan's photometer (Ann. Phys. Chem. [2], 1, 351; 12, 481; 14, 177; and 15,337). The conversion was assumed to be complete when rise of temperature caused no further change in the absorption. This occurred at 84° for a saturated salt solution, at 94° when the solution contained 1.12 grams NaCl to 5.22 grams cobalt solution. In any particular solution, the