of the distinguished Genevese sufficiently comprehensive to include all the phenomena, while he may regard it as a merit in himself to endeavour to confirm and support it. Nevertheless I beg to remind him that when the theory was started in 1772 it met with no favour, nor when it was restated in 1786; but when again brought forward in 1803 it was violently opposed in the very journal* in which M. Gernez now advocates it. Nor was it until M. Donny published in 1846, in the same journal†, his remarkable researches that the theory rose at all into favour; but at that time, and even long after, the merit was given to Donny and not to De Luc. The details given in my "Historical Notes" established, I believe for the first time, the claim of De Luc, and the identity of Donny's results with his.

It is remarkable that although De Luc published his theory several times with very full experimental details, it attracted no attention, although many researches were being made about the same period, and the Committee of the Royal Society were considering their Report on Thermometers. De Luc was much in England and on intimate terms with many of the Fellows of the Royal Society, was admitted a Fellow himself in 1773, and as Reader to Queen Charlotte resided at Windsor and published several of his works, including the theory in question, in London. And yet his theory attracted no attention, and researches on the phenomena of boiling were made during upwards of half a century entirely without reference to this theory. Thus in 1784-85 Achard‡ studied the subject of boiling with reference to the material of the vessel and the influence of solid substances on the boiling-point. In 1812, and again in 1817, Gay-Lussac worked in the same direction as Achard, and also attached importance to the cohesion of the liquid, its adhesion to the sides of the vessel, as well as to pressure in influencing the boilingpoint. In 1825 Bostock published his experiments on the boiling of ether and the influence of solids thereon; and still no reference is made to De Luc. Bostock suggests that the presence of air may have something to do with the results, but he has difficulty in explaining them on this ground. In 1835 Le Grand published some experiments on boiling, and endeavoured to show that the process becomes difficult in proportion as the air dissolved in the liquid is expelled. But I do not remember that he makes any reference to De Luc, any more than Magnus, who in 1836 recurred to the old idea of cohesion, or Schönbein,

* Ann. de Chim. et de Phys. Ser. 1. vol. xlix. p. 235.

↑ Ibid. S. 3. vol. vi. p. 167.

The references to the authorities here briefly enumerated are given in my "Historical Notes."

Phil. Mag. S. 4. Vol. 49. No. 327. June 1875.

2 H

who in 1837 attributed the action of bits of wood, wire, &c. in liberating vapour from boiling liquids to the air contained in their pores or adhering to their surfaces, their action ceasing when the air is driven off. M. Gernez quotes this conclusion with applause; but he does not quote Schönbein's doubts, forgetting perhaps that the doubts of a man of genius are often of more value than the conclusions of an ordinary man. I pointed to the fact some years ago* that Schönbein was by no means satisfied with the theory which attributed the action of solids in liberating gases or vapours from liquids to their carrying down air, a film of which was supposed to adhere to all bodies exposed to it; and he expressed his opinion that any one would perform an important service both to physics and to chemistry who could satisfactorily account for the varied phenomena connected with the subject of nuclei.

In 1842 Marcet attributed to differences of adhesion of the liquid to the sides of vessels of different material those differences in the boiling-point which had been so often noticed. In 1844 Magnus recurs to the same idea, and also to that of cohesion among the particles of the liquid; but they make no reference to the presence of gas in the liquid; and it was not until 1846, or three quarters of a century after the promulgation of De Luc's theory, that we have a De Luc redivivus in the person of M. Donny. According to him, boiling is not an inherent property of liquids; they only boil when they contain air-that is, when they are not pure. Heat liberates bubbles of air nearest to the source of heat; each air-bubble presents to the liquid molecules surrounding it a surface which promotes the vaporization of these molecules; and when the tension of the vapour is sufficient to counterbalance the pressure to which these bubbles are submitted, nothing further opposes the development of this vapour, which then forms currents that traverse the liquid and give rise to ebullition. Hence, according to this view, ebullition is a kind of evaporation, extremely rapid, which operates upon those interior surfaces of the liquid which limit a bubble of some aëriform fluid. If the quantity of air in the liquid be small, the boiling-point may rise; the boiling-point is constant only when the liquid contains air. It is difficult, if not impossible, to get rid of the dissolved air. The boiling-point is also singularly influenced by the forces of cohesion and adhesion, the cohesion of the molecules of water being superior to a pressure of three atmospheres, or a column of water of 30 metres. This conclusion was arrived at by heating water in a chemically clean tube in a bath of chloride of calcium as high as 138° C. (280°.4 F.). Donny's experiments excited considerable attention among * Phil. Mag. for September 1869.

physicists, but, so far as I know, did not have the effect of reviving De Luc's theory until I published an account of it in my "Historical Notes." From 1843 to 1863 it was considered a settled point that, in proportion as water is deprived of air, the character of its ebullition changes, becoming, as it does, more and more abrupt, and boiling, like sulphuric acid, with frequent soubresauts, and the water becomes superheated until relieved by a burst of vapour. In 1863 Mr. Grove stated that if water be boiled in an open vessel it continually reabsorbs air and boils in the ordinary way-although, in a tube with a narrow orifice and with a layer of oil on the water, however long the boiling be continued, a minute bead of nitrogen gas is always to be detected. I have already made some remarks on this form of experiment*. About the same time the experiments of M. Dufour excited considerable attention. In order to get rid of the adhesion of the liquid to the sides of the vessel, he heated a mixture of oil of cloves and linseed-oil and dropped into it water heated to 80° or 90° C., and then gradually raised the temperature of the bath, in one case until it reached 178° C. (352°-4 F.); and he supposed the water to be at that high temperature without boiling. Many years ago I showed that drops of water, ether, alcohol, &c. could be deposited on oil heated to 450° or 500° F. when they assumed the spheroidal state and rolled about on the surface for a long time. In some cases when a drop slipped beneath the surface it exploded and scattered the oil about; but in other cases it was shot up again to the surface, where it continued to roll about as before. Hence I ventured to suggest in my "Historical Notes" that the globules in Dufour's experiment were in the spheroidal state, especially as he admits that when they touched the side of the vessel or were touched with solids introduced into the bath they exploded. But what strengthens my suspicion is the admission on the part of M. Dufour that globules of water, although still retaining their dissolved air, admit of being superheated in his oil-bath. This is exactly what I found in my experiments on the spheroidal condition of liquids on hot oil; the globules that sank without exploding contained air, but were protected from the superior heat of the oil by a coating of vapour and by slow vaporization from the surface.

As to the action of solid nuclei, M. Dufour admits that his results were not always concordant, but were even sometimes contradictory; and he attributes these irregularities to differences in the surface of bodies, to greater or less degrees of roughness, to the absence or presence of foreign corpuscles adhering to them, to the presence or absence of an adhering film of air;

Phil. Mag. for March 1872.

and he attributes the sudden change of state in the globules from liquid to vapour to the capillary action of porous bodies.

The various effects obtained by me from solid nuclei in what I term the chemically unclean or non-catharized state, are admitted by M. Gernez (p. 356); but, while refusing to admit my explanation of them, he invokes the aid of M. Verdet (p. 352), who says: "The following cause acts doubtless in many cases, perhaps in all. If in certain parts [of the solids] the surface be not wetted by the liquid, and if in these parts there exist very small and very fine asperities, the capillary forces compel the liquid to withdraw from the sides to the neighbourhood of these asperities (as happens when we dip a very fine steel needle into mercury); and in this way is formed a veritable free surface where evaporation is a constant and necessary phenomenon. When the vapour formed between this free surface and the side has acquired a sufficient elastic force, it becomes disengaged and ebullition sets in. In this way it may be explained how sulphur and shellac, which water does not wet, and metals which it wets less completely than glass, are useful in accelerating ebullition. In M. Donny's experiment it was necessary that the tube be first disembarrassed from fatty matter by washing with sulphuric acid; the prolonged boiling of the water had probably for its final object, not only the getting rid of the dissolved air, but also the bringing about a certain chemical action of the water on the glass, the result of which was to produce a more intimate contact and a stronger adhesion."

All this is put forth with great ingenuity and ability. At the risk of being tedious, I must here repeat my theoretical views as to the function of solid surfaces introduced into a gaseous solution, or into a liquid at or near the boiling-point. The term "gaseous supersaturated solution" refers to such liquids as soda-water, Seltzer water, champagne, &c.; and seeing that, in a large number of cases in which solid nuclei separated gas from them, there was a precisely similar action of nuclei in separating vapour from liquids at or near the boiling-point, it seemed not unreasonable to suppose that these last-named liquids are constituted like the former. (Herr Schröder subsequently adopted the same view*.) Moreover in both cases the received opinion is that nuclei act by carrying down air, into which the gas or the vapour is said to expand, and so escape; whereas, according to my theory, I endeavour to show that as gas or vapour will adhere to an oily, fatty, or greasy body, or to a body that has been handled, while water will not so adhere, it is only necessary to introduce such a body into the solution to see that it becomes

• Poggendorff's Annalen, 1869.

immediately covered with bubbles of gas or of vapour. These bubbles escape from its surface so long as it continues to be covered, more or less, with a film of a body that can be touched by gas or vapour and not by water. A flint pebble that has been exposed to the air of a room or handled and put into a solution of gas or of vapour, immediately becomes covered with bubbles; but if broken in half and returned to the solution, not a single bubble is to be seen on the fractured parts; for these are specimens of nature's clean surfaces*. If air has any action to perform in the matter, why should the unclean and not the clean surfaces carry it down?

Some liquids contain their own nuclei, as in the case of milk. When this is heated over the fire it becomes more and more charged with vapour; and at a certain point the particles of butter disseminated through it, assisting the expansive force of the heat, produce such a sudden burst of vapour as to cause the liquid to boil over.

If a body, such as a glass rod, be made chemically clean and then be plunged into a supersaturated solution of gas or of vapour, not a single bubble will be seen upon it, because the solution, whether of water and gas, or water and vapour, adheres to it perfectly. If the clean but wet glass rod be left to dry in the dusty air of the room, and when dry be plunged into the solution, it will be active; but if left to dry in the pure outer air of the country, and when dry be plunged into the solution, it is inactive, because it is still in a clean or catharized state.

But it is contended by M. Gernez, and before him by Herr Schröder, that the effect of washing the glass rod or other solid surface in sulphuric acid, or in a caustic alkaline solution, is to get rid of the film of air that adheres to it; while Herr Schröder admits that unclean bodies act because they are covered more or less with a film of fatty organic matter, but according to him it is this film which enables the air to adhere to the solid. This is the critical point of the inquiry. I say that when a clean glass rod or wire &c. is drawn through the hand and immediately immersed in a supersaturated solution, the gas or the vapour is in a condition to leave the solution and adhere to the unclean surface. If an unclean glass rod be heated and introduced below the surface of the solution, and by moving it against the clean side of the vessel leave a greasy line, such line becomes instantly covered with gas- or vapour-bubbles. A rod or wire heated in the flame of a spirit-lamp not sufficient to clean it,

M. Gernez broke a Prince Rupert's drop below the surface of a superheated liquid; the fragments did not act because, according to him, they contained no air. I say they were inactive because they were chemically clean.