the contact-theory would have amounted to from 10 to 15 scale-divisions, and would therefore be easily observed.

(5.) To eliminate the possibility that the absence of a deflection might be due to unknown external effects, a Daniell's cell was introduced into the insulated system, and the deflection obtained was found to agree closely with the value predicted theoretically.

(6.) The author claims that he has proved that no change is produced when the grating is brought into contact with a conductor consisting of a different metal. Such a change is required by the contacttheory, but the author has not attempted to determine whether it would be required by the chemical theory, nor, as far as he is aware, has the question been considered by others.

G. W. T.

Resolution of the Electromotive Forces of Galvanic Elements into their Differences of Potential. By J. MOSER (Monatsh., 8, 508-509.)--Examples are given in this paper of the resolution of the electromotive forces of certain galvanic combinations, and determination of each of the components at the electrolytic surface. The sum

of these several components is shown to be equal to the total electromotive force of the cell. The method of "drop electrodes," as applied by the author to exclude the electromotive force of contact metals, is used, a method based upon an observation of Helmholtz, that if an insulated mass of mercury, dropping quickly, is in contact with an electrolyte at the dropping point, there is no difference of potential between the mercury and the electrolyte.

Thus the total electromotive force of the combination Zn | dilute ZnCl2 concentrated ZnCl, | Zn = 0·15 volt. The values of the electromotive force at the three surfaces v, f, and c, were found by the drop electrode method to be

v = 1.1 volt

c = 0·98 and ƒ = 0·27

ƒ + c = 0·95.

0.15, or the value found for the total electro

a value agreeing with that obtained by direct observation. A similar agreement between the sum of the differences of potential and the total electromotive force in the Latimer-Clark's cell is also noticed. V. H. V.

Property of the Alkalis of increasing the E.M.F. of Zinc. By J. H. KOOSEN (Ann. Phys. Chem. [2], 32, 508-515).-Grove, Joule, Poggendorff, and others have observed the increased E.M.F. obtained by substituting aqueous potash or soda for the dilute acid in a cell. The author states that according to his researches the alkali, say potassium hydroxide, has a four-fold action: (1) it breaks up

into potassium and oxygen; (2) potassium replaces zinc in its electrical action, for instance, in a Daniell's cell potassium sulphate is formed in place of zinc sulphate, as the copper is deposited; (3) the zinc is oxidised by the oxygen of the potash; (4) the resulting zinc oxide is dissolved by the aqueous potash. (1) Diminishes the E.M.F., whilst (2) and (3) increase it; (4) may possibly slightly increase it, but it would be desirable to have positive evidence on this point.

If the external resistance of a cell is very great, almost the whole heat of combination will be developed in the external circuit. Let the E.M.F. of a Daniell's cell be taken as the unit of E.M.F., and let the unit of heat be the amount generated in a Daniell's cell by the decomposition of n grams of zinc, where n is the atomic weight of that metal. Then from Thomsen's determinations, the E.M.F. of a Daniell's cell is given by the equation ZnSO4(248) CuSO,(198) = 50 = 1 Daniell; and with potash substituted for dilute sulphuric weid, K,SO,(337) + 2ZnOH,Ô(166) – 2KHO(232) – CuSO (198) = 73 1.46 Daniell. For the author's zinc-bromine-platinum cell, ZnBr2(91) = 1.82 Daniell.

=

The results thus thoretically obtained in the case of these and a few other simple cells considered by the author agree closely with experimental determinations. The heat of combustion of zinc oxide and potassium is not taken into account, as it is certainly not more than 2 per cent. of that due to the decomposition of zinc in a Daniell's cell, and according to Favre and Silbermann it would seem that thi ́s chemical action is merely local, and does not contribute to the electrical energy. Practically the author finds that a Daniell's cell with a solution of potassium hydroxide is very satisfactory for giving continuous currents through a high resistance, and sodium hydroxide gives still better results, as sodium sulphate, being more soluble, does not so readily crystallise on the porous cell.

It is very important to prevent interdiffusion of the fluids as far as possible, and for this purpose the author uses a double porous cell, having the intermediate space filled with a solution of potassium and sodium sulphates respectively, with a small admixture of sulphuric acid to diminish the resistance. Cells of this kind have been kept joined up through a resistance of 20 or 30 ohms for weeks together without any perceptible diminution in the E.M.F. The alkaline solution must not be too dilute or the zinc will soon become coated with oxide.

The author states that his zinc-bromine-platinum cell is still better for giving a continuous current through a high resistance. It requires no porous cell, has a high E.M.F., and has been kept in continuous action for months without any change in the E.M.F., in fact until one or other of the elements was completely decomposed. The only precaution necessary is to cover the bromine with a layer of petroleum, which completely prevents all evaporation and smell. G. W. T.

Electrolytic Formation of Hydrogen Peroxide at the Anode. By M. TRAUBE (Ber., 20, 3345–3351).—From the results of previous

experiments (Abstr., 1885, 1108), it was concluded that hydrogen peroxide is formed by the union of molecular oxygen with hydrogen. Richarz, on the other hand (this vol., p. 11), maintains that that bydrogen peroxide is produced by the oxidation of water.

When I per cent. sulphuric acid is electrolysed in presence of alcohol or hydrogen peroxide at the anode, these are rapidly oxidised. In the electrolysis of a 1 per cent. solution of chrome alum, not a trace of ozone is formed, and chromic acid appears at the anode. When lead is used as anode in the electrolysis of 1 per cent. sulphuric acid, it becomes covered with lead peroxide. These experiments point to the presence of free oxygen-atoms which only unite to passive molecules when there is nothing to oxidise, and also show that water is not oxidised by oxygen-atoms. In the electrolysis of sulphuric acid, it is suggested that persulphuric acid is formed by the action of nascent oxygen, and that this decomposes into sulphuric acid (2 mols.) and hydrogen peroxide (1 mol.).

As further proof in favour of the constitution previously ascribed to hydrogen peroxide (Abstr., 1886, 660), it is mentioned that the peroxides of hydrogen, of the alkalis and alkaline earths, of zinc, cadmium, and copper, have quite different chemical properties from those of lead, silver, manganese, and nickel, &c. Only those of the first group yield hydrogen peroxide when treated with acids. The peroxides of the second group can all be prepared by oxidising the oxides or hydroxides in alkaline solution. The peroxides of the first group possess powerful reducing properties, whilst those of the second group are indifferent to oxidising agents. Hence it is concluded that the peroxides of the two groups are differently constituted, and that hydrogen peroxides cannot have the formula HO.OH. The constitution represented by the formula HO OH is considered to be the only one possible.

N. H. M.

Electrolytic Conductivity of Halogen-compounds. By W. HAMPE (Chem. Zeit., 11, 816; 846–847; 904-905; 934-935; 1109— 1110; 1158).—Pure dry salts fused in hard glass tubes or porcelain crucibles, or dissolved in dry ether or absolute alcohol, were submitted to a current from eight large Bunsen chromic acid cells, platinum electrodes being used when the conditions of experiment permitted. The following salts proved good electrolytes in a state of fusion:All the haloïd salts of lithium, sodium, potassium, rubidium, and casium, beryllium chloride, magnesium chloride and bromide, strontium and calcium chlorides and bromides (part of the liberated metals combining with the silica of the glass or porcelain), barium chloride, lanthanum, didymium, and cerium chlorides (this confirms Hillebrand and Norton's experience), indium chloride, thallous and cuprous chlorides (thallic and cupric chlorides decompose when fused), tantalum chloride, and thorium chloride, although the latter is not suitable for electrolysis, as the melting and boiling points are near together. In concentrated alcoholic solutions, cupric chloride behaves as an electrolyte. Gold chloride in carbon bisulphide is a non-conductor; when mixed with ether a pasty mass forms, and the supernatant

liquid gradually deposits gold; this liquid conducts slightly, and after 15 minutes the negative pole becomes gilded. Aqueous solutions of gold chloride owe their electrolytic conductivity to the presence of hydrochloric acid. Zinc and cadmium iodides, bromides, and chlorides are good electrolytes when fused, but their conductivity ceases on solidification. In concentrated solutions, these salts conduct well, in weak solutions badly; in absolute alcohol all, but only zinc chloride and bromide dissolve in ether and conduct well, whilst the other salts dissolve in ether and conduct only sparingly or not at all. Fused mercaric chloride, bromide, and iodide are feeble electrolytes, especially the first; in solution in ether, none conduct, owing probably to sparing solubility; in absolute alcohol, they conduct slightly; better when the solution is hot than when it is cold. Mercarous chloride fused in closed glass vessels proved electrolytic. Aqueous solutions of mercuric chloride conduct badly at first, with deposition of mercury and mercurous chloride, and evolution of oxygen and chlorine; subsequently mercury only is deposited, and the electrolysis proceeds at a quicker rate; in this case the current in the first instance produces electrolytes (hydrochloric acid for instance), and when these are present in abundance the process goes well.

Fused gallous chloride is a very good conductor, globules of metal separate at the negative pole, but the chlorine at the positive pole combines with the gallous chloride to form gallic chloride; the latter is not such a good conductor as gallous chloride, no metal separates, evidently owing to its entering into combination with the gallic chloride to form gallous chloride. Stannic chloride does not conduct; when electrolysed in aqueous solution the hydrochloric acid only conducts, the tin being set free by the hydrogen (Hittorf has stated the contrary). Fused stannous chloride on the other hand is an excellent electrolytic conductor, the current continues to pass even on cooling as long as the mass remains soft, but ceases to flow when the chloride is solid. Fused lead chloride, bromide, and iodide conduct very well with vigorous decomposition, and, unlike the other salts examined, they conduct when solid; Buff characterised this conduction as metallic, Wiedermann regarded it as electrolytic; the author confirms the latter view. At ordinary temperatures, no current passes through the cold fused lead salts; it is first observed to flow at 110° through the chloride, at 115° through the bromide, and at 150° through the iodide, the conductivity then increases as the temperature rises, in the manner usual with electrolytes. The trihaloid compounds of antimony in a fused state behave as feeble electrolytes, the tri-iodide being the best; but in carbon bisulphide solutions they do not conduct electricity. Vanadium tetra- and tri-chlorides do not conduct, but when decomposed with water the solutions electrolyse, but in neither case is the metal deposited. The following are non-conductors-Fused aluminium chloride or bromide (Buff's statement to the contrary is probably based on experiments with impure chloride), boron chloride, titanium and silicon tetrachlorides and bromides, vanadyl chloride, niobium pentachloride, all the haloïd compounds of phosphorus, arsenic trichloride, and antimony pentachloride. Yttrium and zirconium chlorides sublime without melting, and when

the former is dissolved in water it conducts on account of the hydrochloric acid, but no metal is deposited; the latter is a non-conductor; titanium hexa- and di-chlorides are infusible solids, so do not come under consideration.

D. A. L.

Molecular Heats of Gases. By H. LE CHATELIER (Bull. Soc. Chim., 48, 122-124). -The fact that the curves of the specific heats of carbonic anhydride and water vapour converge to a point below zero, leads to the supposition that molecular heats of all gases tend to the same limit as the temperature approaches absolute zero. The author has compared the specific heats, calculated on this assumption, with the actual determinations of Wiedemann. The specific heats can be represented by the expression C = 68+a[273 +t], a being a constant which depends on the nature of the gas and has a higher value the more complex the molecule. The numbers calculated by this equation differ from the actual numbers by quantities less than the experimental errors. A similar agreement is found between the numbers for carbonic anhydride at high temperatures deduced from this equation and the actual numbers obtained by the author and Mallard.

α

C. H. B.

Influence of Small Amounts of Impurities on the Vapourtension of Liquids. By G. TAMMAN (Ann. Phys. Chem. [2], 32, 683-699).—The increase in the vapour-tension of a liquid observed by Wüllner and Grotrian (Ann. Phys. Chem. [2], 11, 545), when the space occupied by its vapour in the manometer has been decreased by compression, and attributed by them to a dependence of the tension on the amount of liquid in contact with the vapour, is shown to be caused and brought about by traces of impurity in the liquid. These traces are often so exceedingly small as not to be recognisable by the ordinary tests. But the vapour, if containing this impurity, consists of a mixture of two substances one of which is more volatile than the other, so that compression causing a greater condensation of the less volatile, causes a rise in the vapour-tension, and expansion has just the opposite effect. A difference is therefore observed between the tension measured after compression and after expansion. This difference is reduced in magnitude by purification of the liquid used, but could not in any case be entirely eliminated, although an approximation to a constant tension was reached in the case of water. measurement of the vapour-tension, once after compression and once after expansion, will give a rough measure of the purity of the substance used. Experiments are given with water, ether, carbon bisulphide, benzene, methyl and ethyl alcohols, chloroform and acetic acid.

H. C.

Α

Dissociation of Crystallised Lead Acetate and Sodium Thiosulphate. By W. MÜLLER-ERZBACH (Ber., 20, 2974-2981).Two series of experiments were made with lead acetate. In the first series, large crystals of the commercial salt were employed in the state of powder, and with it the relative tension was 0.27 to 0.38 at temperatures varying between 12.5° and 21.8°; whilst in the second series, in which