ELECTROLYSIS 107 and the danger of confusing electrolytic effects with effects due to disruptive discharge by convection. Gases have, however, been decomposed by the silent discharge, as C0. 2 into CO + O. From Faraday s time attempts have continually been made to classify strictly, according to their chemical com position or constitution, the liquids capable of electrolytic conduction, but hitherto without very much success. It must be remembered that, as the resistance of a liquid increases, the tests of electrolytic conduction become less and less sensitive. We can consider a body an electrolyte if we can (1) collect the products of decomposition, or (2) demonstrate their presence on the electrodes by means of the return current due to polarization. If the resistance be very great the former method becomes evidently very difficult, and in the latter complications are introduced which cannot here be discussed (see ELECTRICITY). On the other hand, we might easily be misled into consider ing a body an electrolyte from the presence of mere traces of a foreign substance. Thus at one time water was regarded as the only electrolyte, but it is found that the purer the water is the less does it conduct electricity, and now Kohlrausch and Nippoldt have shown that the presence of one 10-millionth of H 2 SO 4 would be sufficient to account for its observed conducting power, so that the weight of evidence goes to show that water itself is not an electrolyte at all It is not, then, surprising that views on the question of what constitutes an electrolyte have changed considerably. Davy and the older chemists, as mentioned above, considered water to be the only electrolyte ; Faraday, by electrolysing fused chlorides, &c., dissipated these notions, but still re garded water as the electrolyte which was decomposed when acids were subjected to the electric current, and his general conclusion was that an electrolyte must be a compound con sisting of an equal number of chemical equivalents of its elements, that is, in modern notation, must be of the type M y R ^ where x and y are the atomicities or valencies of the elements whose atomic weights are represented by M and 11, and thus that two elements would by uniting form only one electrolyte (Ejcp. Res., 679-701, 830). The oxygen salts for which Faraday assigned no law were in cluded by Daniell in the same formula as binary compounds, of which the part 11 acting as anion was no longer an element but a compound; thus ZnS0 4 was shown to be split up by electrolysis into Zn and S0 4 ; in that case y would represent the basicity of the acid forming the salt. This hypothesis lacks definiteness, on account of the varia tion of the atomicity of the elements, and falls through altogether in the case of copper and iron, which form each two chlorides, (CuCl 2 ,Cu 2 Cl 2 ), (FeCl 2 ,Fe 2 Ci c ), both electrolytes, and in consequence Wiedemauu (Galv., Bd. i. 295, 346a, 418 (5)) modifies the statement of the hypothesis, and considers that for a body to be an electrolyte it must be capable of formation by double decomposition from one of the simple binary electrolytes, the exchanging atoms or groups of atoms forming the ions of the new compound. Thus silver acetate gives, by double decomposition with sodium chloride, silver chloride and sodium acetate. Sodium acetate arid silver chloride are therefore electrolytes of which Ag, Cl, Na, C g H 3 O 2 are the respective ions. This hypo thesis may be illustrated by a great number of instances : the case of the decomposition of uranium compounds, as UOC1 into UO and Cl, is a very good example. But Wiedemann s view would indicate that a body, in order to be an electrolyte, need but be one of a " series of salts," and we then see no reasan for excluding the hydrogen salts from the class ; thus 1LO and II Cl can be easily formed by double decomposition, yet the former is, when pure, one of the worst liquid conductors, while tho latter as liquefied gas is apparently uot decomposed even by 5640 cells of De la Rue s chloride of silver battery, but gives vibrations indicating very high resistance. 1 Bleekrode has also shown that, of all the pure liquefied hydrogen acids, only HCN is an electrolyte. On the other hand, liquefied NH 3 , which is not formed, so far as we are aware, by double decomposi tion, is electrolysable by only a moderate battery of Bunseu s cells, giving a blue liquid at the cathode. Moreover, Buff (Ann. d. Chem. und Pharm., Bd. ex.) has electro lysed molybdic and vanadic anhydrides after the manner Mo0 3 = MoO 2 + O, but these bodies are not obtainable by double decomposition with a simple electrolyte. Miller (Elements Chem., i. 282 (v) ) considers that an electrolyte must be a combination of a conductor and a non conductor, and so the majority of electrolytes are. But alloys behave to a certain extent as electrolytes when fused (see VVied., Galv., Bd. i. 328), and SnCl 4 , though consist ing of a conductor and a non-conductor, is not an electro lyte ; so that this classification is not exclusive. It would therefore appear that the condition does not lie in the chemical constitution of the body, but rather in its molecular state, and to this points the fact that two non conductors, as II 2 O and HC1, on being mixed form a very good conductor. In addition to this, quantitative measure ments of the resistance of electrolytes show that, in the case of many salt and acid solutions, there is a point of con centration below saturation, for which the conductivity is a maximum. This would scarcely be the case if one alone of the bodies were the conductor. The liquids which do not conduct are very various, including, besides oils and resins and other organic bodies, benzine, iodide of sulphur, carbon disulphide, glacial acetic acid, fused boracic anhydride, antimonic oxide and oxy- chloride, the higher halogen salts of tin, liquid sulphurous anhydride, pure water, and pure halogen acids. For others see article ELECTRICITY, p. 51. In the description of the phenomena, in the typical case Fara- of electrolysis given above, it was stated that the amount of day slaw chemical decomposition in any time is proportional to the ^ e whole quantity of electricity which passes through the electro- liquid in that time ; this is true in all cases of electrolysis, lyte. and was established by Faraday (Exp. Res., . 505, and ser. vii.). It forms part of the general law to which his name is attached, but we prefer to consider it separately for reasons that will appear when we discuss th i statement of that law. We may put it thus : If W be he mass of an electrolyte," 1 decomposed by the passage of a qt antity E of elec tricity, then, as long as the ions remain of tl z same nature, U), where K is a constant dependent only on the nature of the electrolyte, and therefore independent of the nature or size of the electrodes and of any secondary actions which may take place. It is evident that if we can prove the truth of this lawVolta- for one electrolyte, with ions which do not vary with varia- tions of electromotive force, we shall have a very con venient means of measuring the total amount of electricity which passes through any circuit in a given time by in troducing such an electrolyte into the circuit, and measur ing the amount of decomposition in the given time. Fara- 1 Bleekrode and Dela Rue, Proc. Roy. Soc., xxv. p. 323. In fact, dis ruptive discharge occurs by convection currents, or, if the electrodes be sufficiently near, by spark. Similar phenomena may be observed l>y immersing the poles of a Iloltz machine in paraftin oil. 2 In what follows, the term electrolyte is used in its most general sense, to signify any liquid or mixture of liquids through which the current passes," and not necessarily ore definite chemical compound. Hence the necessity for the rendition that the ions shall not vary, as in mixed electrolytes ions for high electromotive forces are different
from those for low (vid. inf.).