is, in fact, the sender, the ether the wire, and the eye the receiving instrument, in this new telegraphy.

As before, let us endeavour to see if our ideas may be rendered more clear by any of the more familiar phenomena in sound.

Let us call the motions, of whatever nature they may be, set up in a molecule of matter, vibrations. Then, to get the most concrete notion of such a light source, let us compare it with the most simple sound source, a tuning-fork.

The same tuning-fork will always give us the same sound, but the sound is more complex than might at first sight be imagined. In addition to the prevailing note, which depends upon the number of vibrations per second, as we have seen, there are other tones which have a definite relationship to the prevailing, or, as it is termed, the fundamental note. The loudness of all these, as we have also seen, depends upon the amplitude of the vibration of the tuning-fork.

But there are more complicated cases of vibration than these.

In the violin we have a convenient example which shows us that the quality of the sound produced, that is the quality of the vibration of the string set up, depends upon the manner in which the bow is drawn over the string.

Similarly, the same vibrating plate, when damped at different points, vibrates in quite a different manner.

Now does the spectroscope throw any light upon molecular questions? is there any hope that the

In sodium we may say that the longest line is double; I refer to D' and D". All the lines are double.

In magnesium the longest line is a triple combination. This is repeated exactly in the violet.

In manganese we may almost say that the same thing happens, but the phenomenon is much more absolute in the case of those particles such as sodium and magnesium, which, on other grounds, I suspect to be of the simplest structure.

21. Our knowledge of the vibrations of particles will be incomplete until the vibration is known from the extreme violet (invisible) to the extreme red (invisible).

In the meantime great help may be got from inferences, and, in the case of metalloids at low temperatures, from the position of their continuous absorption: and it is a question whether light may not be thus thrown upon the opacity of some solid substances and the transparency of others.

I think it not too much to say that already, in the case of some gases and vapours which are apparently transparent, it is as certain in some cases that their absorption is in the ultra red, as it is certain that in the case of others the absorption is in the ultra violet. And further, it can scarcely be that this absorption is not of the continuous or fluted kind-in other words, that no gas is "atomic" in the chemist's sense, except when subjected to the action of electricity, or, in the case of hydrogen, to a high temperature.

no breaks in the series of wave-lengths. We may also drive a platinum wire to incandescence in the same way by means of electricity. Analyse the light by means of the spectroscope, the spectrum is the same. as that of the poker. Further, we can go to the sun, and divest it in imagination of the atmosphere which absorbs so much of its light, and we know that, with a small exception, we shall get a perfectly continuous spectrum similar to that in the case of the poker or platinum wire.

In this continuous spectrum we have a spectroscopic fact connected with that kind of molecular motion which physicists attribute to particles so long as they are closely packed together in the solid state, and so long as they have but a small free path, as in the fluid state.

2. When particles are in a state of gas or vapour, and are rendered incandescent by high tension electricity, linespectra are produced in the case of all the chemical

elements.

These line-spectra are only to be obtained from gases and vapours, and, with few exceptions, only when we employ high-tension electricity.

We get a spectroscopic result perfectly distinct from the one we had before, precisely in the case where according to the physicists we have an enormous motion and agitation of particles.

3. The characteristic vibration of a particle is independent of length of free path.

In the Kinetic theory, as generally enunciated, there is nothing to show that the same particle may not be in question in the solid, liquid, and gaseous states, the only change of condition being in the amount of free path.

Now as the spectra of solids and gases present a complete difference in kind (see 1 and 2), if the particle were always the same, it would be necessary to assume that, under different conditions of free path, the same particle can be thrown into different states of vibration and give us different spectra.

The question is, have we at the present time any facts at our disposal? I think we have, although some may not consider them sufficient in number or cogency; but it must always be remembered that it is precisely in such questions as these that experiments on an extended scale become almost impossible.

All the facts we have, however, tend to show that a known change of molecular condition is always accompanied by a change of spectrum, e.g., sulphur vapour above and below 1,000° C.-at which point its vapour density changes-has distinct spectra.

Salts have spectra of their own, in which no lines, either of the constituent metal or metalloid, are to be found.

Another line of argument. In some cases we can mix vapours with liquids and the spectrum of the vapour remains unchanged in character; that is, the circumambient particles of the liquid behave in one case in exactly the same manner as the circumambient particles of the air do in another.

4

Iodine in bisulphide of carbon, N, O, in water, and didymium salts in water are illustrations.

Nay, we may even enclose, or appear to enclose, some substances in glass (salts of didymium, erbium, nickel, cobalt), and we get a spectrum so special in each case that we know that the particles are still going through their motions, are still vibrating, in spite of the absence of free path, and in spite of the "solid" state of their surroundings.

I shall, elsewhere, use other lines of argument to show that the reason that we so rarely see these characteristic spectra in connection with the solid state lies in the fact that the solid state is one reached not only by reduction of free path, which enables the molecules to lie nearer together, but by a reduction of molecular agitation, which in all probability enables them to combine inter se.

4. In some cases particles in a state of gas or vapour can be set swinging by heat waves.

Salts of sodium and strontium, subjected to the heat of a Bunsen burner, are at once dissociated, and the particles of the metals are set swinging by the heat waves, and we get the lines in the spectra of their vapours. Now that is not only true for salts of strontium and sodium, but for some of the elements themselves. But if salts of iron, or of the other heavy metals are placed in the flame, we do not get bright lines. Or again, in some other vapours, such as sulphur, we only get a spectrum, not