*

illuminated. In fact many such bodies are evanescently phosphorescent, and the survival of the phosphorescence when the stimulating light has been withdrawn can be observed, and its duration measured, by the phosphoroscope. Although in most cases its duration is but a small fraction of a second, nevertheless in every case that has been observed, indeed in every case that can be observed with the phosphoroscope, it is a duration of immense length compared with the almost inconceivable rapidity of molecular events, in comparison with which even the thousandth part of one second is a vastly long period of time.†

In all such bodies therefore there are events of the class Bb.

Solids only seem to have been examined by the phosphoroscope. But we may feel assured that the same dynamical conditions prevail, certainly in liquids and probably in gases.

Let us next consider what bearing this has on the interpretation which is to be put on a high ratio of the two specific heats.

An event of the Bb class, which subsides so rapidly as to require the phosphoroscope to detect it, will behave, in any protracted experiment for determining the ratio of the two specific heats, in the same way as events of the Ba class. But this ceases to be the case where the ratio of the two specific heats is determined by experiments on sound; and in all the experiments which have been made use of it ceases to be the case when Bb events are as slow in subsiding as some of them are in conspicuously phosphorescent bodies. Now the method by which the ratio of the two specific beats has been determined for argon and helium has been by experiments on sound; and as the value furnished by this method depends upon Ba events, it is competent to supply information about them only. It gives no information as to the energy involved in events of the Bb class. Accordingly it remains quite possible that Bb events may be easily evoked by stimulation of argon and helium, and that while in existence they may engross a considerable share of the total energy in the gas. That this is the case would seem to be evidenced by the vivid spectra which these gases exhibit under the influence of electricity.

The colours of objects, when not mere interference phenomena, are due to the excitation of Ba or Bb events within the molecules by certain rays of the incident light. In both cases the acting rays yield up their energy; but when Ba events take their place, the body is simply warmed: when it is Bb events that come into existence, the body for a short time subsequently radiates light. In the one case the colour of the body is that of incident light which is not absorbed; in the other case it is in general the colour of incident light which is absorbed. Of course if both causes are in operation they produce their conjoint effect. A few outlying cases, such as that of fluorescence, require a slightly modified treatment.

The thousandth of a second bears about the same relation to molecular events that 10,000 years does to the motions of the limbs of animals.

In this connection, it is well to call to mind that phosphorescent events can be excited with even greater splendour by electricity than by exposure to light,* as has been abundantly shown by experiment, especially by many experiments of exceptional brilliancy that have been made by Mr. Crookes.

There appears, therefore, to be no ground for the supposition which has been sometimes entertained, that there is incompatibility between the two facts that have been observed-the fact that in these two gases the ratio of the two specific heats is near its maximum value, and the fact that these gases, when stimulated by electricity, furnish brilliant spectra.

Events of the Bb class in phosphorescent bodies may be made to reach the intensity which enables them to emit visible radiations in any one of three ways, either—

1. By exposing the phosphorescent body to light of suitable wavelength and sufficient intensity; or

2. By exciting certain other electrical events in its neighbourhood;

or

3. By raising the temperature of the whole phosphorescent body to a white heat.

If the phosphorescence is excited by either of the first two of these methods, the phosphorescent substance remains at a temperature as tested by the thermometer, immensely short of that which would enable an incandescent body to emit light of equally high refrangibility.

Accordingly the luminous effects within Geissler tubes do not prove that the temperature of the luminous gas is very high: an inference which is often erroneously drawn. In fact, Bb motions, when once excited within molecules, may continue for a considerable time to be more active [or, it may be, less active] than the Ba motions simultaneously going on, since there is but feeble interaction between them.

If the molecule consist of but one chemical atom, there may be both Ba and Bb events going on within that so-called atom. If the molecule consist of two or more chemical atoms, a part of the Ba events may be motions of the centres of mass of these atoms relatively to one another. But this is not always the case the bonding between the atoms that form the molecule may be such that there is but little of this relative motion. Accordingly, the ratio of the two specific heats being large does not necessarily imply that the molecule is monatomic. The inference involves the erroneous supposition that there are no events going on within the molecule, and few

* Or, rather, by electrical events other than light; since light is itself a manifestation of electricity.

degrees of freedom in its motion :* a supposition which is an example of how ready we are to think that Nature must work simply when she works on a very small scale, and of the further error of imagining that a little rigid body is something exceptionally simple-the fact being that a rigid body is only a figment of the imagination, and that in Nature it is physically impossible. Moreover, it so happens that the bodies in Nature which most nearly resemble vigid bodies, namely, elastic solids, are amongst those bodies whose internal constitution is most complex.

Beside the Bb motions, there may be other internal events more or less isolated from both the Ba and the Bb events. By two events being isolated is to be understood their being unable to interchange energy. We may call these in succession Bc, Bd, &c., events. When they exist, the body will usually emit two or more spectra under variations of the external stimulus, whether luminous or electrical. And we must bear in mind that Ba events may also be the source of a spectrum.

The simplest supposition as to the interaction between the ether and the molecules of matter is that which is based on Faraday's law of electrolysis, which, as von Helmholtz pointed out, and as the present writer had previously shown, implies that there is a certain electrical charge, of the same amount in all cases, associated with each chemical bond (see Philosophical Magazine' for October, 1894, p. 418). The approximate amount of this charge, which Helmholtz designated the atom of electricity, and which the author has called the electron, can be computed. According to a determination made by the author in 1874, it appears to be about three-eleventhets

That there are few degrees of freedom in the molecule is sometimes supposed to follow from the dynamical investigation; but this appears to be a mistake. The Maxwell Law of the partition of kinetic energy is only known to prevail 1° where the kinetic energy is expressible as a sum of squares; 2° where certain initial conditions of the motions of the system of bodies have been complied with; and 3° where the subsequent events are due exclusively to the interaction of the bodies of the system.

No one of these is known to be true of any gas; and the second of them if fulfilled initially will in general cease to prevail so soon as any agency other than the dynamical action of the molecules intervenes. Radiant heat, light, electricity, and many (probably all) chemical reactions, are agencies of this kind.

That the dynamical investigation, based on data simpler than those that prevail in nature, offers in a striking way an explanation of the numerical values for the ratio of the two specific heats as determined by experiment in several gases, in no degree proves that those simpler data are what exist in nature. Many and very various phenomena of light are explainable in a very striking way by the simple hypothesis that light is an undulation of transverse motions; but it would be a rash inference to conclude from this that electromagnetic waves are mere transverse motions. The data of Nature have always to be simplified before they can be used as the data of mathematical investigations.

(3 × 10-") of the C.G.S. electrostatic unit of quantity. Other estimates of the amount of this remarkable unit of electricity have since been made, and do not materially differ from the above. Here, then, we have all the machinery required. The motions which go on actively within chemical atoms can scarcely fail to wave about these electrons which are so closely associated with them; and we know that the waving about of such of these charges of electricity as for the time happen to be undisguised, must generate in the ether exactly such electro-magnetic waves as those revealed to us in the spectra of gases (see 'Transactions of the Royal Dublin Society,' vol. 4 (1891), p. 583).

V. "On the Velocities of the Ions." By W. C. DAMPIER WHETHAM, M.A., Fellow of Trinity College, Cambridge. Communicated by Professor J. J. THOMSON, F.R.S. Received May 2, 1895.

(Abstract.)

In a previous communication to the Royal Society (Phil. Trans.,' 184, (1893), A. p. 337), I have described a method of experimentally determining the velocities of the ions during electrolysis, by observations on the phenomena which occur when a current of electricity is passed across the junction of two salt solutions, one at least of which is coloured.

The results obtained agreed, within the limits of experimental error, with the numbers deduced by Professor F. Kohlrausch from measurements of the conductivities. The method is, however, seriously restricted by the conditions necessary for its success. two solutions must be of different densities, of different colours, and of nearly equal conductivities at equivalent concentrations.

The

In order to extend the method, I have used solid solutions in agaragar jelly, tracing the motion of the ion by the formation of a precipitate, Jelly solutions were employed by Dr. Oliver Lodge, but, although he got a good result for hydrogen with phenol-phthallein as an indicator, when he tried to measure the velocity of barium and strontium by watching the formation of precipitates, the experiments were not very successful. This was probably due to the fact that, when a precipitate forms, it removes some of the electrolyte from solution, and so increases the specific resistance and the local potential-gradient. In order to eliminate this disturbing cause, instead of setting up two solutions which completely precipitate each other, like barium chloride and sodium sulphate, use was made of barium chloride and sodium chloride, just enough sodium sulphate being added to the latter to enable the motion of the barium ions to be traced by the

formation of a slight precipitate of barium sulphate. That this did not seriously affect the result was shown by making two determinations of the velocity of the barium ion, much more sodium sulphate being present in the first case than in the second. The numbers obtained for the specific ionic velocity were 0.000385 and 0.000390 cm. per second respectively.

The influence of the jelly was examined by determining the velocity of the bichromate group (Cr2O,) with solid solutions of potassium bichromate and potassium chloride-the motion being indicated by the change in colour. The result was 0.00044 cm. per second. In the former paper the velocity of the same ion, measured by means of aqueous solutions of these salts, came out 0.00047 cm. per second. The influence of the jelly thus appears to be small, as is also indicated by Arrhenius' experiments on the conductivity of such solutions ('B. A. Report,' 1886, p. 344).

The apparatus used and the method of measurement were the same as in the former investigation. Two vertical glass tubes, about 2 cm. in diameter, were joined by a third, considerably narrower, which was bent parallel to the others for the greater part of its length. In this tube the jelly solutions came in contact with each other, and formed a slight precipitate at the junction. When a current was passed from one solution to the other, this precipitate spread, and the rate at which its advancing surface moved, was measured on a glass scale placed behind it, the observations being made through a telescope.