small crystals of the pure salt were used, the relative tension was 0:32 to 0.39 at 14.9-24-3°; the value of t1 t2, which the author assumes to be a measure of the attraction between the salt and its water of crystallisation (compare Abstr., 1887, 628), decreasing with the rise of temperature. A partially dehydrated specimen of the salt containing about mol. H2O, when placed in a moist atmosphere until water equal in amount to 3 mols. had been absorbed, showed a higher and less constant relative tension, the highest value being 0.485 at 21.5°.

With sodium thiosulphate, Na2S2O3 + 5H2O, two series of experiments were also made, and in the first series 3 mols. H2O were lost with a relative tension = 0·33 to 0:36 at 29·1-32.9°, and a further 14 mol. with a relative tension = 0.17 to 0.22 at 31·7-34.2°, the remainder of the water being given off with a relative tension = 0.047 to 0.053 at 50·6-52·7°. The second series of experiments showed that 3 mols. were lost with a relative tension =0·26 to 0.28 at 18.3-21.6°, that a further 14 mol. was lost with a relative tension = 0.08 to 0.12 at 15.9-22.1°, and that the tension became imperceptible when the salt contained 0.6 mol. H2O. The partially dehydrated salt, containing 2 mols. H2O, exposed in a moist atmosphere until it contained 4.75 mols. H2O, lost 2:55 mols. with a relative tension = 0·29 to 0:31 at 23·6-25.9°, and the salt with 2 mols. H2O, obtained in like manner from the salt + 0.49 mol. H2O, lost approximately 1.5 mol. with a relative tension = 0·10 to 0·19 at 5·3—16·1°. It was found, however, that when sodium thiosulphate was heated to drive off the water of crystallisation rapidly and then exposed to moist air until 4.9 mols. H2O had been taken up, it lost the whole of the water with a relative tension = 0·21 to 0·32 at 15·8-30.6. W. P. W.

Relation between the Compressibility of a Solution and the Compressibilities of its Constituent Parts. By F. BRAUN (Ann. Phys. Chem. [2], 32, 504—508).-Röntgen and Schneider (this vol., p. 22) have given for the compressibility of rock-salt the value 5 x 10-6, which does not agree with the author's previous determination (Abstr., 1887, 436), which gave 14 x 10-6, This result was obtained by comparatively rough methods, but the author points out that it agrees very closely with the value 16 x 10-6, deduced from Voigt's determination (Ann. Phys. Chem. Ergbd., 7, 214) of its two coefficients of elasticity.

Röntgen and Schneider assume that if y be the compressibility of a solution containing per unit volume a volume V' of water, and V" of salt, and if be the compressibility of water and y" that of the salt, then (1) = 'V' + '"V". Now if n be the number of molecules in a solution containing a percentage p of salt of molecular weight m, then (2) n = 106p/m(100-p). They then take a constant a such that (3) n = aV"/V', and eliminating V' and V" from equations (1) and (3), they obtain the equation (7 — y') (n + a) = Y') (n + a) = (7' — 7′′)a, which they write (y b)(n+ a) = (1 − b)a, where y is the apparent compressibility of the solution, and b that of the salt. If be the cubic compressibility of glass, (5) y = (1 − x)/(7' — «), and

b = (y" — x)/(y). They determined n and y for five different concentrated solutions of sodium chloride. Two of these results they used to determine a and b by means of (4). They then found that by substitution in (4) the relations obtained between n and y agreed well with the experimental determinations. Assuming the formula to hold good when n becomes infinite, y will in that case be equal to b, and determining from (5), they obtained values varying from 4.7 x 10-6 to 48 x 10-6, which agreed well with the value 5 × 10−6 determined directly for the solid salt.

In a previous memoir, the author found that (1) does not agree with experiment, if for example were given, would have to be negative to give the correct value of y. If, therefore, the compressibility of a solution is equal to the sum of the compressibilities of its constituents, the compressibility of the water of the solution must be diminished by the presence of the salt, and this is shown to be the case by Röntgen and Schneider's recent results. The other discrepancy the author traces to these observers having determined the value of a from two equations of the form (4), the observations then show that a and b are constants, so that if the value of V" calculated from (3) be substituted in (1), this equation will naturally agree with experiment, since it is inerely another form of (4). The value of a can, however, be obtained directly from (2) and (3), and the value thus obtained is about five times as great as before, and (1) will only hold good for this value of a, for then only does the ratio V'/V' express the actual conditions. With this value of a, b is no longer found to be a constant. The author gives a numerical example, in which Röntgen and Schneider's value of a leads to an impossible result. He observes that as this correction does not alter the form of (4), all conclusions drawn simply from the form of this equation will still hold good. G. W. T.

Dilatation and Compressibility of Liquids. By E. H. AMAGAT (Compt. rend., 105, 1120-1122).-The author has determined the expansion and compressibility between 0° and 50°, and 1 atmosphere and 3000 atmospheres, of water, of ether, of methyl, ethyl, propyl and allyl alcohols, of ethyl chloride, bromide and iodide, of carbon bisulphide and of phosphorous chloride.

In all cases, except that of water, the coefficient of expansion diminishes with increased pressure, but the rate of diminution decreases as the pressure rises, although it remains distinct even at 3000 atmospheres. The increase in the coefficient of expansion due to a rise of temperature, likewise diminishes under increased pressure. The coefficient of expansion of ether under a pressure of 3000 atmospheres is only onethird of its value at normal pressure, and when this liquid is compared with carbon bisulphide, it is found that whilst under normal pressure the ether is much the more expansible of the two, under 2500 atmospheres the coefficients are identical, and under 3000 atmospheres the coefficient for carbon bisulphide is higher than that for ether.

Under very high pressures, the perturbations which are observed in the case of water, and which are due to the existence of a point of maximum density, gradually disappear, and under 3000 atmospheres

the law of expansion of water agrees with that of other liquids (compare Abstr., 1887, 695). C. H. B.

Compressibility of an Aqueous Solution of Ethylamine. By F. ISAMBERT (Compt. rend., 105, 1173-1175).—The coefficient of compressibility of ethylamine between 5° and 7° and from 8 to 45 atmos. is 0·000120, a value which approximates to that for ether at 0°, and is much higher than that for water. The compressibility of an aqueous solution of ethylamine is much lower than the value calculated on the assumption that the two liquids are simply mixed, and it decreases as the proportion of water increases. Ethylamine in fact behaves like ammonia, and this result confirms the author's conclusion that solutions of the ammoniacal bases behave like true chemical compounds more or less dissociated in presence of excess of water. The view that a compound is formed is supported by the contraction which takes place, and the development of heat, which as in the case of ammonia is equal to the heat of volatilisation of the amine.

C. H. B.

Oxygen Carriers. By L. MEYER (Ber., 20, 3058-3061).— Experiments were made with various metallic salts with a view to determine their power of expediting the oxidation of sulphurous acid. The oxygen and sulphurous acid were passed, for four hours, at as uniform a rate as possible, through the solution of known concentration contained in a flask heated on a water-bath. The amount of sulphuric acid was then determined.

The most active salt was manganous sulphate, of which 2:404 grams (MnSO. + 5H2O) was dissolved in 200 c.c. of water; this solution in four hours gave six times as much sulphuric acid as the salt contained. Manganous chloride is also very active; copper sulphate gives less sulphuric acid. After copper salts, the salts of iron and cobalt are the most active, the chlorides being more so than the sulphates. The salts of nickel, zinc, cadmium, and magnesium produce less sulphuric acid, whilst dilute solutions of thallium and potassium sulphates and free sulphuric acid produce none at all (compare Kessler, Ann. Phys. Chem., 1863, 119, 218).

N. H. M.

Explosive Decomposition of Picric Acid and other Nitroderivatives. By BERTHELOT (Compt. rend., 105, 1159-1162).— When picric acid is heated in a capsule, it melts and gives off inflammable vapours, which burn with a smoky flame, but it does not explode. A small quantity when heated in a closed tube can be sublimed without decomposition. If, however, the picric acid is thrown into a vessel previously strongly heated, the quantity of the solid being so small that the temperature is not materially reduced, then it decomposes with detonation accompanied by a vivid flash of light. When the quantity of picric acid introduced is sufficient to somewhat cool the bottom of the vessel, detonation does not take place at once, but the acid is partially volatilised, and a somewhat less violent explosion follows. If the quantity is still larger, the acid decomposes, but there is no explosion. Similar results were obtained with mono- and di

nitrobenzene, and mono-, di-, and tri-nitronaphthalene. The tendency to detonate increases with the number of NO2 groups.

Nitro-derivatives may decompose in several different ways. If the heat developed by combustion is carried away with sufficient rapidity, there is no deflagration or detonation, but if the heat accumulates the temperature may rise sufficiently high to produce detonation. The heating of only a small part of the containing vessel to a high temperature may produce an explosive wave, which may then propagate itself through the whole mass, and thus produce an explosion. C. H. B.

Relative Size of the Molecules, calculated from the Electric Conductivity of Salt Solutions. By G. JÄGER (Monatsh., 8, 498-507).—In this paper, an attempt is made to determine the relative diameters of some of the elementary molecules and atomic groupings, adopting as a basis for the calculations the results obtained by Kohlrausch in his investigations on the electric conductivity in aqueous solutions of certain metallic hydroxides and salts.

If, in such an aqueous solution, a cylindrical section is taken of unit dimensions, there are contained therein m electrochemical molecules; taking, then, the electromotive force in the direction of the axis as unity, the kathion will be propelled in the one direction with velocity U, and the anion with velocity V in the other. If e is the quantity of positive or negative electricity with which each molecule is endowed, then the coefficient of conductivity = (u + v)m where €U = u_and_eV v, and the specific molecular conductivity λ = u + v. But according to Hittorf v/u + v = n, in which n is the number of molecules passing in unit time through unit space. Hence u = (1 n)λ and v = nλ. If the molecules of the ions pass with a certain velocity, they meet in unit time a certain number of other molecules of a different kind passing in the opposite direction; therefore they require a certain amount of energy to overcome the necessary resistance which is proportional to their rate of passage. If for the sake of simplicity the molecules are assumed to be spheres, and the solution is so dilute that there is no interaction between the molecules of the dissolved substance; also if a molecule of radius r passes in a certain direction whilst the environment consists of molecules of radius p, and the number of molecules in unit volume is a, then r + p = R if the forces are resolved in two directions. The result is the same if the radius of the moving molecule is R, whilst the molecules in the environment are reduced to a mathematical point. Now (1) v C/R2 = C/(r + p)2, in which C is a constant obtained by integration, while for another molecule (2) v' = C/(r′ + p)2; dividing (1) by (2) v/v′ = (r' + p)2¡(r + p)2, from which r= (r' + p)√v/v — p. To solve this equation the values for the relative velocities have been determined by Kohlrausch, whilst to find r′ and p the diameters of the molecules of water and chlorine calculated by O. Meyer are used. The values are for water d = 96 × 10-9, for chlorine & = 44 × 10-9 centimetres, while U for water = 49. Hence the diameter of a given molecule is—

=

expressed in 10-9 centimetres. The following results are given for the relative diameters of the elementary molecules and of certain atomic groupings arranged in order of magnitude.

H.
H2. I.
15 32 91

Br. (CN). Cl. K. (NH). (NO3). (ClO3).

91

95 96 97

99

100

111