acid in both cases. The salt which, for a given weight of the solution, can furnish the most chromic acid is the superior. The solubilities of the two salts are as follows:-At 15°, 100 parts of water dissolve 12.5 parts of potassium dichromate and 83-16 parts of anhydrous sodium dichromate. The available oxygen, therefore, in a saturated solution is 0-0852 per cent. for the sodium salt, as compared with 0.0180 per cent. for the potassium salt. B. H. B. Ferric Chloride as an Exciting Agent for Voltaic Batteries. By H. N. WARREN (Chem. News, 55, 49).—The power of a current from an ordinary dichromate cell may be very greatly increased by employing a slightly acidified, strong solution of ferric chloride mixed. with bromine, in place of the potassium dichromate. The bromine serves to reoxidise the ferrous chloride formed, and when consumed may be again set free by the addition of bleaching powder. D. A. L. Specific Inductive Power of Liquids. By NEGREANO (Compt. rend., 104, 423-425). The author has determined the dielectric constants K and the refractive indices of benzene (both pure and mixed with thiophen), toluene, xylene (mixture of isomerides), metaxylene, pseudocumene, cymene, and terebent hene. The square roots of the dielectric constants differ from the refractive indices in the second decimal place only. The dielectric constant seems to decrease with a rise of temperature within certain limits. the series examined, the value of the dielectric constant increases as the molecule becomes more complicated. The ratios K − 1 and d In increase irregularly with an increase in molecular weight. K 1 The ratio (K + 2)d' however, is practically constant, and for the same liquid is the relation which connects the dielectric constant with the density. In this particular series, the value of this constant is about 0:34. Electromotive Force of some Thermo-elements consisting of Metals and Solutions of their Salts. By A. EBELING (Ann. Phys. Chem. [2], 30, 530-543).-According to Gore (Proc. Roy. Soc. 36, 50, and 37, 251), the electromotive force of a thermo-couple formed of a metal and a salt solution changes with the concentration of the solution; according to Bouty (Abstr., 1881, 336), who employed the same metal in the electrodes and the solution, it is independent of the concentration. The author has, therefore, reiuvestigated the question, using pure copper in contact with copper sulphate or nitrate, and amalgamated zinc free from arsenic in contact with zinc sulphate, nitrate, or chloride. An element consisted of two glass cylinders containing the air-free solution, and communicating through a narrow glass syphon tube. In each mass of liquid was immersed an accurately graduated thermometer, round the bulb of which the metal was wound in the form of wire. One cylinder was surrounded by melting ice, and the other heated to various temperatures. The electromotive force was measured by Du Bois Raymond's modification of Poggendorff's compensation method, using two Daniell cells which were from time to time compared with Helmholtz's constant calomel cell. It was impossible to find two wires of the same metal so perfectly homogeneous as to give no current in the above cell, even with both contacts at the same temperature. The strength of this chemical current was therefore measured at constant temperature before and after each set of experiments, the rate of its change with time calculated, and a correction thus made. Since observations with different elements could not all be made at precisely the same temperature, this source of error was allowed for on the assumption that the electromotive force is proportional to the difference in temperature of the contacts. The temperature intervals were 20°, 35°, and 47°. Tabular statements of the results in each case are given. The author's conclusions are: The electromotive force is not proportional to the difference of temperature of the contacts, but increases with it in some greater ratio. With any constant temperature difference, the electromotive force diminishes slightly with the time, probably as the result of chemical actions. The electromotive force does not increase continuously with the concentration, but exhibits maxima and minima. For the same acid, these maxima correspond with the same state of concentration. Some idea of the results may be gathered from the following table of electromotive forces due to a difference of temperature P signifies the percentage of anhydrous salt in solution. = 20°. The results are expressed in terms of a calomel cell as unit. To reduce them to electromagnetic units the figures must be multiplied by 10542-104 cm.lg.sec.-2. The maxima for strong solutions are indicated by brackets, for weak solutions by asterisks. Certain other conclusions are also stated, but with some reserve. According to Neumann's law, substances of similar constitution and properties have equal atomic heats; but according to the fourth statement above the maximum electromotive force for different metals, but the same acid, occurs for solutions of different concentration. Also, the strength of solution for which the electrical conductivity is a maximum is also that for which the electromotive force is a maximum. This is shown by a table of electrical conductivities taken from Wiedemann's Electricity. Finally it appears that here, as in the case of ordinary thermo-couples, the worst conductors make the most effective combinations. Сн. В. Galvanic Polarisation of Aluminium. By F. STREINZ (Phil. Mag., [5], 23, 304).-The author finds that if aluminium plates be polarised, the difference of potential between the oxygen plate and the zinc of the polarising cell increases within very wide limits with the electromotive force of the cell, whilst the hydrogen plate shows very little polarisation for small electromotive forces, but with large electromotive forces acquires a difference of potential opposite in direction to that usually obtained. The small amount of gases usually obtained from aluminium electrodes can be explained by the great opposing force of oxygen polarisation. H. K. T. Conductivity of Mixtures of Aqueous Solutions of Acids. By S. ARRHENIUS (Ann. Phys. Chem. [2], 30, 51-76).-Previous observations by Bouchotte, Paalzow, Bender, and Klein, have led to no general conclusions. In the following experiments, the author has used Kohlrausch's method, employing a telephone as indicator. The fact that the molecular conductivity of a solution is not proportional to the amount of dissolved electrolyte, proves that the solvent liquid is not always uniformly distributed amongst the dissolved molecules, even when the electrolyte is a single substance. Ostwald, however, by some experiments, which are here described for the first time, has shown that the molecular conductivity of a mixture of butyric and acetic acid solutions in any proportions, is always the sum of the molecular conductivities of the constituent solutions. Each electrolyte behaves as if the other were absent. Such solutions are termed by the author "isohydric." In the foregoing case, the two acids are very similar; but when they are of different character, their solutions can be isohydric only under special conditions. Now, the specific conductivity of any solution is directly proportional to the contained mass of electrolyte, and inversely proportional to the frictional resistance to the transference of the ions. For dilute solutions, the latter may be regarded as not changing with dilution. Hence, when two weak solutions are mixed, and the distribution of the water is not thereby altered, the conductivity is the arithmetical mean of the separate conductivities; or if lv, and v2 are the specific conductivities and volumes of the original solutions, and L and (v1 + v2) those of the mixture (Lv1 + v2) = lv1 + h1⁄2v2 = N1μ1 + N1⁄2μ2, Νιμι Ν2μ2, where N1 and N2 are the numbers of dissolved gram-molecules, and 1, 2 the molecular conductivities of the electrolytes. When the distribution of the water is altered by mixture, let v1 and v2 become v1 + dV and v2 dV; then L, 1, and με will increase by dL, du1, and du2. Therefore The ratios du/dV and dμ/μ2dV can be calculated from Ostwald's data (J. pr. Chem. [2], 32, 300). For a small increase of v1, du/μdV =σ, where the values of σ are different for different acids, and also diminish gradually with dilution. Finally, (v v2)dL = (401 + : 402)dV. The author has calculated for solutions of equal conductivity of six different acids, and finds that its value is greater the more feeble the acid (strength = molecular conductivity). When, therefore, two solutions of different acids, of not very different conductivity, are mixed, and when the specific conductivity of the mixture is found to be greater than the arithmetical mean of the specific conductivities of the original solutions, then the more feeble acid has taken part of the solvent water from the stronger; and vice versa. When the conductivity of the mixture is equal to the above arithmetical mean, the solutions are isohydric. In practice, it is desirable to compare feeble acids with a stronger acid, rather than with each other, so that σ, and σ, may be as different as possible; and to use equal volumes, since dV is then a maximum. The solutions should also be dilute, and of such strengths that their specific conductivities do not greatly differ. From two observations, then, in which one solution is varied, the isohydric strengths can be found by linear interpolation. The following propositions are experimentally proved. If two solutions are isohydric when mixed in equal volumes, they are also isohydric when mixed in any other ratio. If a solution A is isohydric with solutions B and C, B and C are also mutually isohydric. In this table, isohydric solutions are arranged in horizontal series. The upper of each pair of numbers gives the corresponding strength of the solution in gram-molecules per litre; the lower gives the conductivity in mercury-units x 108. The latter is uncertain, sometimes to the extent of 5 per cent. From the table, it may be seen that the conductivities of isohydric solutions are usually about equal. It is also shown that solutions of equal conductivity are approximately isohydric. The difference in the latter case is usually a slight increase of conductivity. By a method of approximations, the conductivity of any mixture of these acids may be calculated from Ostwald's data. No simple relation exists between the molecular strengths of isohydric solutions, such as Bender (Ann. Phys. Chem. [2], 22, 197) claims to have proved. Сн. В. Theory of Voltaic Action. By J. BROWN (Proc. Roy. Soc., 41, 294-315). A number of experiments are described in this paper to establish the two following propositions: (1) that the difference of potential near two metals in contact, as observed by Volta's bimetallic condenser method or Thomson's quadrant method, is due to the chemical action of a film of condensed vapour or gas on the surfaces of the metals; (2) that these two metals, with their liquid film, form a cell similar to one composed of the same metals as elements, and a liquid of the same kind as electrolyte, which in contact experiments is divided by the intervening insulating diaphragm of air or other gas. From this it follows that in these experiments it is the difference of potential at the outer surfaces of the two metals which is measured. Thus, in the case of a single metal covered by its chemically active electrolytic film, at the surface of the film and metal there is an electromotive force corresponding with the chemical action between |