C H E M I S T R Y from the ratio in which the potassium chlorate becomes distributed between the solid solvent (T1C10.,) and the liquid, it appears that the molecular weight of potassium chlorate is the same in both the liquid and the crystalline condition. One of the most brilliant and fruitful generalizations of recent times has been van’t Hoff’s extension to dilute soluExtension ^0T1S °f the well-known laws connecting presofgaseous sure, volume, and molecular weight in gases. laws to Although a liquid exerts no pressure other than solutions, thah due to its mass upon the surface of a containing vessel, the existence in solutions of a state comparable with that which obtains in gases becomes evident on studying the phenomena of liquid diffusion. If, for example, a solution of sugar be enclosed in a vessel having semi-permeable walls—'b e., permeable by the water, but not by the dissolved substance,—and then the vessel be placed in water, diffusion of water into the solution sets in and continues until a certain pressure is attained : this is termed the osmotic pressure, and is found to be proportional to the number of molecules dissolved in the unit volume of liquid. Variations in temperature affect osmotic pressure in the same way that they affect gaseous pressure, consequently Avogadro’s theorem is directly applicable to dilute solutions; in other words, equal volumes of dilute solutions (of any one solvent) under like conditions of temperature and osmotic pressure contain equal numbers of molecules. If osmotic pressure could be easily measured, a simple method of determining the molecular weight of dissolved substances would be to determine the osmotic pressures they develop ; but, unfortunately, this is not the case, and the determinations are affected by considerable experimental errors. But there are certain properties of dilute solutions which can be easily measured which vary jm^'i passu with the osmotic pressure. It has long been known that aqueous solutions of equimolecular concentration of salts such as sodium chloride, potassium chloride, &c., deposit ice when cooled to the same temperature, and van’t Hoff has shown that the depression of the freezing-point of a solvent by addition of small proportions of a soluble substance is proportional to the variation in the osmotic pressure. The law that the freezing-point of a solvent is depressed to an extent proportional to the number of molecules of dissolved salt present in unit volume forms the basis of the now well-known freezingpoint method of determining the molecular weight of dissolved substances. In a similar manner the vapour pressure of a solvent is lowered by the presence of dissolved material to an extent which is proportional to the osmotic pressure, and in weak solutions the decrease in vapour pressure is directly proportional to the rise in boiling-point of the solvent. On this is based the now familiar method of determining molecular weights by observing the rise in boiling-point of a solvent produced when known small amounts of soluble substances are dissolved in it. Both methods afford anomalous results when applied to aqueous solutions of electrolytes; but these meet with an explanation when the diminution in average molecular weight, i.e., the increase in the number of active molecules, consequent on the occurrence of ionic dissociation, is allowed for. The ionic dissociation hypothesis is undoubtedly a conception of great beauty, and the success that has attended ionic its application in the case of dilute solutions is hypothesis of the most striking character. But it would insuffibe unscientific to yield to its charms without c,ent having fully considered the probability and sufficiency of the explanation which it affords, and the difficulties which stand in the way of its acceptance. Chemistry is so wide a subject, chemical phenomena are so complex and so hard to analyse and interpret, the difficulty of giving a quantitative expression to them is so
739
great, that chemists are but rarely prepared to undertake such a discussion ; and it is easy to accept guidance and to yield allegiance to high authority, especially when it is exercised in the most autocratic and dogmatic manner possible hitherto. Unfortunately the advocates of the dissociation hypothesis have declined even to consider the objections which may be raised against it from the chemist’s side. One of the most fundamental objections to the hypothesis is that the number of compounds which are per se electrolytes being limited—to only a few fused salts, in fact ■—the explanation it affords of the occurrence of chemical change is at best only a very partial one. Water and all other liquids appear to be non-conductors in the pure state, and the arguments on which the assumption is based that water is dissociated to a very slight extent all involve reasoning in a circle. Pure water—a pure substance—is and must ever be an ideal conception ; but as the resistance which water opposes to a current of electricity gradually increases as it is purified, and at last becomes enormously high, the facts warrant no other assumption than that it would be a non-conductor in the simple and pure state. The arguments based on the observation that it is slightly active chemically are of no independent value, since an impurity which would condition conductivity would equally condition chemical activity. The assumption which is made, in order to explain the constancy in the heat of neutralization of strong acids generally by alkalies, that water is only very slightly dissociated, involves a sharp distinction between water and other compounds, and consequently between oxygen and other elements, which is in no way warranted and is in opposition to the periodic generalization. It is generally recognized that oxygen and the halogens present many analogies, so that water should— and does—in many Avays resemble the halhydrides; to assume, therefore, that water resists “ionic” dissociation whilst hydrogen chloride suffers almost complete disruption into “ ions ” is altogether unjustifiable and illogical. When Avater is mixed, say with hydrogen chloride, a non-conductor per se, a solution is obtained Avhich is readily electrolysed. But solvents generally do not act in this manner—in fact, only a limited number have the poAver of forming conducting solutions, and none is so active as Avater. To account for the influence of the solvent, F. Kohlrausch, in 1875, put forward the view that Peculiar of the ions which Clausius assumed to be present in a liquid are thereby prevented from coming too frequently into contact and merely recombining, and that consequently they are free during a sufficient length of time to be amenable to the guiding influence of the electromotive force. Parenthetically it may be pointed out that the dissociation postulated by Clausius Avas supposed to affect only a relatively small number of molecules at any one moment, whereas the dissociation postulated by the ionic hypothesis, in the case of strong acids and metallic salts, affects the major proportion of the molecules. Moreover, the process supposed to attend dissolution is one of a most extraordinary character; it may be likened to the case of a number of individuals jumping into a swimmingbath, and nearly all of them instantly parting with their arms and legs. If the solvent act as a mere screen, any solvent should suffice to render a potential electrolyte a conductor, but this is not the case. Moreover, as a number of substances AArhen fused (silver chloride, stannous chloride, &c.) are conductors in the absence of any solvent, the screen hypothesis is not only an unsatisfactory and incomplete explanation of the facts, but its introduction appears to be uncalled for. Again, if the Clausius hypothesis be accepted, and it be assumed that some of the molecules in a liquid collide Avith sufficient violence to bring about their decomposition, it is to be expected
