woolly-looking crystals after a slight rubbing, and although the wire was rapidly withdrawn as soon as the crystals began to appear, it left behind it a curious kind of trail, consisting of an axis of the form of the wire, repeating its bends and irregularities, while small acicular crystals started out at right angles to this line, and thickly studded it soon after the whole of the solution became solid. Zincic acetate (1 to 1) did not display crystals as soon as the side was rubbed, but slowly, in the course of half-an-hour or so, the part rubbed was occupied by a dense crop of beautiful crystals, the points of which passed beyond the axis of the tube. Sodic acetate solution does not readily yield to the effort of rubbing, except in certain states of the weather when the crystallizing force is strong; but a case occurred to me some years ago which exactly suits my present purpose. Several flasks containing a highly supersaturated solution of sodic acetate, to which oil had been added without any nuclear effect, were emptied into a small, stout cylindrical glass, where it remained during some weeks without change, the layer of oil on its surface preventing it from absorbing moisture. It was several times. stirred with the finger without effect, for this, on being introduced, became coated with oil and so prevented contact. The finger was then pressed against the strong side of the vessel with a considerable amount of force and drawn slowly from the bottom upwards. The solution immediately became solid. I attributed this effect at the time to nuclear action, but I now regard it as an excellent illustration of the point which I am seeking to establish. A supersaturated solution of sodic carbonate is very sensitive to any interference with its adhesion. A speck of carbon or of ferrous oxide, coming to the surface within the tube, is sufficient, on cooling down the solution, to cause the salt to start from the speck in crystalline lines in all directions, producing immediate solidification.* In addition to amber and resin some tubes were coated with other substances, such as sulphur, shellac, stearine, and paraffin; but as some of these, in addition to the objection of being opaque, melted when the boiling solution came into contact with them, I need not refer to them any further, except perhaps to paraffin, with which an interesting result was obtained. A large tube was lined with paraffin and passed, telescope fashion, over a smaller tube, containing a nearly boiling solution of sodic sulphate. When cold, the tubes were slowly reversed, so as to pour the solution into the coated tube; but it solidified as soon as it came into contact with the paraffin. If, however, it were poured in while still warm, the solution accommo * In the experiments which I had the honour of exhibiting to the Society after the reading of my first paper (28th May, 1868), a case of this kind occurred and was pointed out by my late lamented friend, William Allen Miller, as "a very pretty effect." dated itself to the new circumstances and remained liquid when cold. The results of these experiments seem to point to the conclusion that adhesion to the side of the vessel is one of the conditions under which the state of supersaturation is maintained; and that whatever interferes with this adhesion, must prevent supersaturation or destroy it. This conclusion was tested by another mode of experiment, starting from the idea that if the solution were made to expose an amount of free surface, about equal to the surface of contact with the vessel, the adhesive force would scarcely be sufficient to maintain the state of supersaturation, and the solution in cooling would crystallise as in an open evaporating dish. This is what really did happen in the case of ammonia alum (5 to 3). The solution in three covered shallow vessels deposited massive crystals in cooling, and when cold, the mother liquor was no longer supersaturated. A solution of the same degree of strength was boiled in a flask and the opening plugged with cotton-wool: this deposited no salt in cooling, because the area of adhesion was greatly in excess of that of the free surface. The solution under such conditions will remain unchanged for months. In all cases the re-boiling of the solution in the flask into which it is filtered greatly promotes adhesion, and, consequently, the duration of the solution in the state of supersaturation. But to return. A similar crystallisation took place in the case of sodic sulphate; but that arose chiefly from the kind of vessel used and the mode of covering it, as will be noticed presently. I must first refer to supersaturated solutions of sodic sulphate (3 to 1) in shallow foot glasses, four inches in diameter and one inch deep in the middle. The boiling solution was poured into these glasses, which were immediately covered with shallow glass vessels, inverted over them so as to fit nicely and rest on their edges, or to fit with friction to the sides. The result was curious and interesting. Instead of crystallising in cooling, the solution relieved itself by throwing down large quantities of the seven-watered salt in finely shaped masses, while the liquid portion remained supersaturated.* Next day these solutions were taken into the open air and uncovered. From a point at the extreme edge of each, crystallisation set in and spread like a fan over the surface and through the solution. This was not a case of nuclear * A three-ounce flask was coated with resin, and into it was poured a boiling solution of magnesia sulphate (3 to 1). The flask was covered with a small beaker and set on the window ledge to cool. A fine display of crystals of the modified salt sprang from the bottom nearly to the surface of the solution, which remained supersatuated. Here again, as in the case of sodic sulphate, the solution relieved itself by this deposit and there was sufficient adhesion on the part of the remaining liquid to maintain a moderate degree of supersaturation. action derived from the atmosphere (rain having fallen during some hours and having scarcely ceased) but was due to another force. Last summer, as noticed in my former paper,* drops of sodic acetate exposed on glass always started the crystallisation from the edge. So, in these wide shallow glasses in the open air, crystallisation set in from the edge, or capillary curve, where the solution is most drawn out and attenuated. Now, supposing evaporation to be active all over the surface, this thin portion, becoming thinner by the process, would be the first to detach a molecule of the salt, and this, once set free, would act as a powerful nucleus to the whole. Lines, or rather thin planes of crystals, diverge from the point, and this so quickly as not to afford time for the molecules to arrange themselves into crystalline forms. On one occasion the mass of seven-watered salt rose so near the surface, that the solution resting on it was scarcely more than a film. This, by evaporation, also disengaged a minute portion of the normal salt, which acted as a centre of crystallisation. Solutions containing two parts of salt to one of water behaved much in the same way as the stronger ones. A weaker solution, containing only one of salt to one of water, accommodated itself to the conditions under which it was placed. It did not deposit any modified salt, and on being uncovered in the open air, crystallised from the edge. In the case before referred to, where the solution crystallised in cooling, as in an open evaporating dish, namely, in the form of large prisms of the normal salt, the vessel was a shallow dessert dish with a scolloped edge, so that the protecting cover did not confine the air over the surface of the solution, but allowed it to circulate; hence, in cooling down, the molecules were also in a condition to circulate and gradually to arrange themselves into groups, ready to assume the crystalline form when the temperature had sufficiently declined. Whereas, in closely covered vessels, the strong adhesion to the side. prevents this circulation and re-arrangement, the adhesion holding the molecules in a forced state, bearing some kind of analogy to the glass in Prince Rupert's drops and the Bologna phial. It must also be noted, that in the last case also, there was no nuclear action; for, had there been, the solution would have cooled down to a certain point, and then have suddenly crystallised in closely packed planes; whereas, in the case now referred to, the crystals were large and of the usual shape, while the mother liquor was no longer supersaturated. Solutions of zincic sulphate and of magnesic sulphate (3 to 1 each) in these covered shallow vessels, also threw down a good deal of the modified salt, in the case of the zinc solution the monohydrated salt *Proc. Roy. Soc., xxvi, 528. in long clustered cylinders, but being left undisturbed for a day or so, the remaining supersaturated liquor in both cases became solid, the normal salt being formed. Sodic acetate solution (6 to 1) poured boiling into these large shallow vessels and covered over did not change during twenty-four hours. On being taken into the open air and uncovered, the solutions remained some time in the liquid state, but on touching the bottom of the vessel with the finger, they immediately became solid. The alum solution, before referred to, deposited the normal salt in large crystals, while the solution was still warm. It could not relieve itself in any other way, because by its constitution it cannot form a salt of a lower degree of hydration than the normal, that is, with 12 aq., and the throwing down of that in considerable masses, so as to get rid of the state of supersaturation altogether, forcibly illustrates the dependence of that state on the adhesion of the solution to the sides of the vessel. The weather was mild while these experiments were in progress. At lower temperatures the solutions would doubtless have been even more sensitive. The result of these experiments satisfies me that the state of supersaturation is dependent on the adhesion of the solution to the sides of the vessel, coupled with the tension of the surface. This tension may be lowered, so as to throw more work upon the other force, which it may be able to bear; but anything that effectually detaches a portion of the solution from contact with the sides of the vessel, produces the sudden crystallisation of the solution. If this view be correct, many of the phenomena of supersaturation are accounted for, and the whole subject is far advanced towards that obedience to law, which can alone invest it with dignity. March 21, 1878. Sir JOSEPH HOOKER, K.C.S.I., President in the Chair. The Presents received were laid on the table and thanks ordered for them. The following Papers were read: I. "Contact Theory of Voltaic Action." Parts I and II. By W. E. AYRTON and JOHN PERRY, Professors in the Imperial College of Engineering, Tokio, Japan. Communicated by Professor Sir W. THOMSON, F.R.S. Received October 2, 1877. PART I. The contact theory of voltaic action seems to have undergone no development since the date of Sir W. Thomson's experiment, which consisted in connecting a plate of zinc and a plate of copper by means of a drop of water, when it was found that the metals were brought to the same electric potential, although when metallically connected they were at different potentials. He believed that any electrolyte would behave in exactly the same way as the water of his experiment, equalizing the potentials of any two metals connected by it. The electromotive force of a simple cell, ought, in accordance with the theory, to be equal to the difference of potentials between zinc and copper in contact. A test founded on this deduction was very difficult to apply, because there was no exact determination of the difference of potential of zinc and copper in contact, Sir W. Thomson, in his experiment, having really measured the difference of potential between air at the surface of a zinc plate, and air at the surface of a copper plate. In the absence of this test, the equality of the electromotive forces of simple cells in which zinc and copper are the metals (the liquids being water, dilute sulphuric acid, and sulphate of zinc) was held as a proof of the theory. Now it is known that when two pieces of the same metal are dipped into any two liquids, which are diffusing into one another, a difference of potentials is established between the metals, and the electromotive force of a cell of this kind can in no way depend on a difference of potentials due to metallic contact. So that although in such a cell there is an action which is somewhat the same as the action in a simple voltaic cell, the theory took no account of it whatever. In fact, the explanation of voltaic action given in the latest |