I will now abandon the Nicols, and send through the piece of Iceland spar (B, fig. 53), already employed (in Lecture III.) to ill.u.s.trate the double refraction of light, our sifted beam of invisible heat. To determine the positions of the two images, let us first operate upon the luminous beam. Marking the places of the light-images, we introduce between N and L our ray-filter (not in the figure) and quench the light. Causing the pile to approach one of the marked places, the needle remains unmoved until the place has been attained; here the pile at once detects the heat. Pushing the pile across the interval separating the two marks, the needle first falls to 0, and then rises again to 90 in the second position. This proves the double refraction of the heat.

[Ill.u.s.tration: Fig. 53.]

I now turn the Iceland spar: the needle remains fixed; there is no alteration of the deflection. Pa.s.sing the pile rapidly across to the other mark, the deflection is maintained. Once more I turn the spar, but now the needle falls to 0, rising, however, again to 90 after a rotation of 360. We know that in the case of light the extraordinary beam rotates round the ordinary one; and we have here been operating on the extraordinary heat-beam, which, as regards double refraction, behaves exactly like a beam of light.

-- 10. _Magnetization of Heat_.

To render our series of comparisons complete, we must demonstrate the magnetization of heat. But here a slight modification of our arrangement will be necessary. In repeating Faraday's experiment on the magnetization of light, we had, in the first instance, our Nicols crossed and the field rendered dark, a flash of light appearing upon the screen when the magnet was excited. Now the quant.i.ty of light transmitted in this case is really very small, its effect being rendered striking through contrast with the preceding darkness. When we so place the Nicols that their princ.i.p.al sections enclose an angle of 45, the excitement of the magnet causes a far greater positive augmentation of the light, though the augmentation is not so well _seen_ through lack of contrast, because here, at starting, the field is illuminated.

In trying to magnetize our beam of heat, we will adopt this arrangement. Here, however, at the outset, a considerable amount of heat falls upon one face of the pile. This it is necessary to neutralize, by permitting rays from another source to fall upon the opposite face of the pile. The needle is thus brought to zero. Cutting off the light by our ray-filter, and exciting the magnet, the needle is instantly deflected, proving that the magnet has opened a door for the heat, exactly as in Faraday's experiment it opened a door for the light. Thus, in every case brought under our notice, the substantial ident.i.ty of light and radiant heat has been demonstrated.

By the refined experiments of k.n.o.blauch, who worked long and successfully at this question, the double refraction of heat, by Iceland spar, was first demonstrated; but, though he employed the luminous heat of the sun, the observed deflections were exceedingly small. So, likewise, those eminent investigators De la Povostaye and Desains succeeded in magnetizing a beam of heat; but though, in their case also, the luminous solar heat was employed, the deflection obtained did not amount to more than two or three degrees. With _obscure_ radiant heat the effect, prior to the experiments now brought before you, had not been obtained; but, with the arrangement here described, we obtain deflections from purely invisible heat, equal to 150 of the lower degrees of the galvanometer.

-- 11. _Distribution of Heat in the Electric Spectrum_.

We have finally to determine the position and magnitude of the invisible radiation which produces these results. For this purpose we employ a particular form of the thermo-pile. Its face is a rectangle, which by movable side-pieces can be rendered as narrow as desirable.

Throwing a small and concentrated spectrum upon a screen, by means of an endless screw we move the rectangular pile through the entire spectrum, and determine in succession the thermal power of all its colours.

[Ill.u.s.tration: SPECTRUM OF ELECTRIC LIGHT.]

When this instrument is brought to the violet end of the spectrum, the heat is found to be almost insensible. As the pile gradually moves from the violet towards the red, it encounters a gradually augmenting heat. The red itself possesses the highest heating power of all the colours of the spectrum. Pushing the pile into the dark s.p.a.ce beyond the red, the heat rises suddenly in intensity, and at some distance beyond the red it attains a maximum. From this point the heat falls somewhat more rapidly than it rose, and afterwards gradually fades away.

Drawing a horizontal line to represent the length of the spectrum, and erecting along it, at various points, perpendiculars proportional in length to the heat existing at those points, we obtain a curve which exhibits the distribution of heat in the prismatic spectrum. It is represented in the adjacent figure. Beginning at the blue, the curve rises, at first very gradually; towards the red it rises more rapidly, the line C D (fig. 54, opposite page) representing the strength of the extreme red radiation. Beyond the red it shoots upwards in a steep and ma.s.sive peak to B; whence it falls, rapidly for a time, and afterwards gradually fades from the perception of the pile. This figure is the result of more than twelve careful series of measurements, from each of which the curve was constructed. On superposing all these curves, a satisfactory agreement was found to exist between them. So that it may safely be concluded that the areas of the dark and white s.p.a.ces, respectively, represent the relative energies of the visible and invisible radiation. The one is 7.7 times the other.

But in verification, as already stated, consists the strength of science. Determining in the first place the total emission from the electric lamp, and then, by means of the iodine filter, determining the ultra-red emission; the difference between both gives the luminous emission. In this way, it is found that the energy of the invisible emission is eight times that of the visible. No two methods could be more opposed to each other, and hardly any two results could better harmonize. I think, therefore, you may rely upon the accuracy of the distribution of heat here a.s.signed to the prismatic spectrum of the electric light. There is nothing vague in the mode of investigation, or doubtful in its conclusions. Spectra are, however, formed by _diffraction_, wherein the distribution of both heat and light is different from that produced by the prism. These diffractive spectra have been examined with great skill by Draper and Langley. In the prismatic spectrum the less refrangible rays are compressed into a much smaller s.p.a.ce than in the diffraction spectrum.

LECTURE VI.

PRINCIPLES OF SPECTRUM a.n.a.lYSIS PRISMATIC a.n.a.lYSIS OF THE LIGHT OF INCANDESCENT VAPOURS DISCONTINUOUS SPECTRA SPECTRUM BANDS PROVED BY BUNSEN AND KIRCHHOFF TO BE CHARACTERISTIC OF THE VAPOUR DISCOVERY OF RUBIDIUM, CaeSIUM, AND THALLIUM RELATION OF EMISSION TO ABSORPTION THE LINES OF FRAUNHOFER THEIR EXPLANATION BY KIRCHHOFF SOLAR CHEMISTRY INVOLVED IN THIS EXPLANATION FOUCAULT'S EXPERIMENT PRINCIPLES OF ABSORPTION a.n.a.lOGY OF SOUND AND LIGHT EXPERIMENTAL DEMONSTRATION OF THIS a.n.a.lOGY RECENT APPLICATIONS OF THE SPECTROSCOPE SUMMARY AND CONCLUSION.

We have employed as our source of light in these lectures the ends of two rods of c.o.ke rendered incandescent by electricity. c.o.ke is particularly suitable for this purpose, because it can bear intense heat without fusion or vaporization. It is also black, which helps the light; for, other circ.u.mstances being equal, as shown experimentally by Professor Balfour Stewart, the blacker the body the brighter will be its light when incandescent. Still, refractory as carbon is, if we closely examined our voltaic arc, or stream of light between the carbon-points, we should find there incandescent carbon-vapour. And if we could detach the light of this vapour from the more dazzling light of the solid points, we should find its spectrum not only less brilliant, but of a totally different character from the spectra that we have already seen. Instead of being an unbroken succession of colours from red to violet, the carbon-vapour would yield a few bands of colour with s.p.a.ces of darkness between them.

What is true of the carbon is true in a still more striking degree of the metals, the most refractory of which can be fused, boiled, and reduced to vapour by the electric current. From the incandescent vapour the light, as a general rule, flashes in groups of rays of definite degrees of refrangibility, s.p.a.ces existing between group and group, which are unfilled by rays of any kind. But the contemplation of the facts will render this subject more intelligible than words can make it. Within the camera is now placed a cylinder of carbon hollowed out at the top; in the hollow is placed a fragment of the metal thallium. Down upon this we bring the upper carbon-point, and then separate the one from the other. A stream of incandescent thallium-vapour pa.s.ses between them, the magnified image of which is now seen upon the screen. It is of a beautiful green colour. What is the meaning of that green? We answer the question by subjecting the light to prismatic a.n.a.lysis. Sent through the prism, its spectrum is seen to consist of a single refracted band. Light of one degree of refrangibility--that corresponding to this particular green--is emitted by the thallium-vapour.

We will now remove the thallium and put a bit of silver in its place.

The are of silver is not to be distinguished from that of thallium; it is not only green, but the same shade of green. Are they then alike?

Prismatic a.n.a.lysis enables us to answer the question. However impossible it is to distinguish the one _colour_ from the other, it is equally impossible to confound the _spectrum_ of incandescent silver-vapour with that of thallium. In the case of silver, we have two green bands instead of one.

If we add to the silver in our camera a bit of thallium, we shall obtain the light of both metals. After waiting a little, we see that the green of the thallium lies midway between the two greens of the silver. Hence this similarity of colour.

But why have we to 'wait a little' before we see this effect? The thallium band at first almost masks the silver bands by its superior brightness. Indeed, the silver bands have wonderfully degenerated since the bit of thallium was put in, and for a reason worth knowing.

It is the _resistance_ offered to the pa.s.sage of the electric current from carbon to carbon, that calls forth the power of the current to produce heat. If the resistance were materially lessened, the heat would be materially lessened; and if all resistance were abolished, there would be no heat at all. Now, thallium is a much more fusible and vaporizable metal than silver; and its vapour facilitates the pa.s.sage of the electricity to such a degree, as to render the current almost incompetent to vaporize the more refractory silver. But the thallium is gradually consumed; its vapour diminishes, the resistance rises, until finally you see the two silver bands as brilliant as they were at first.[24]

We have in these bands a perfectly unalterable characteristic of the two metals. You never get other bands than these two green ones from the silver, never other than the single green band from the thallium, never other than the three green bands from the mixture of both metals. Every known metal has its own particular bands, and in no known case are the bands of two different metals alike in refrangibility. It follows, therefore, that these spectra may be made a sure test for the presence or absence of any particular metal. If we pa.s.s from the metals to their alloys, we find no confusion. Copper gives green bands; zinc gives blue and red bands; bra.s.s--an alloy of copper and zinc--gives the bands of both metals, perfectly unaltered in position or character.

But we are not confined to the metals themselves; the _salts_ of these metals yield the bands of the metals. Chemical union is ruptured by a sufficiently high heat; the vapour of the metal is set free, and it yields its characteristic bands. The chlorides of the metals are particularly suitable for experiments of this character. Common salt, for example, is a compound of chlorine and sodium; in the electric lamp it yields the spectrum of the metal sodium. The chlorides of copper, lithium, and strontium yield, in like manner, the bands of these metals.

When, therefore, Bunsen and Kirchhoff, the ill.u.s.trious founders of _spectrum a.n.a.lysis_, after having established by an exhaustive examination the spectra of all known substances, discovered a spectrum containing bands different from any known bands, they immediately inferred the existence of a new metal. They were operating at the time upon a residue, obtained by evaporating one of the mineral waters of Germany. In that water they knew the unknown metal was concealed, but vast quant.i.ties of it had to be evaporated before a residue could be obtained sufficiently large to enable ordinary chemistry to grapple with the metal. They, however, hunted it down, and it now stands among chemical substances as the metal _Rubidium_. They subsequently discovered a second metal, which they called _Caesium_. Thus, having first placed spectrum a.n.a.lysis on a sure foundation, they demonstrated its capacity as an agent of discovery. Soon afterwards Mr. Crookes, pursuing the same method, discovered the bright green band of _Thallium_, and obtained the salts of the metal which yielded it. The metal itself was first isolated in ingots by M. Lamy, a French chemist.

All this relates to chemical discovery upon earth, where the materials are in our own hands. But it was soon shown how spectrum a.n.a.lysis might be applied to the investigation of the sun and stars; and this result was reached through the solution of a problem which had been long an enigma to natural philosophers. The scope and conquest of this problem we must now endeavour to comprehend. A spectrum is _pure_ in which the colours do not overlap each other. We purify the spectrum by making our beam narrow, and by augmenting the number of our prisms.

When a pure spectrum of the sun has been obtained in this way, it is found to be furrowed by innumerable dark lines. Four of them were first seen by Dr. Wollaston, but they were afterwards multiplied and measured by Fraunhofer with such masterly skill, that they are now universally known as Fraunhofer's lines. To give an explanation of these lines was, as I have said, a problem which long challenged the attention of philosophers, and to Professor Kirchhoff belongs the honour of having first conquered this problem.

(The positions of the princ.i.p.al lines, lettered according to Fraunhofer, are shown in the annexed sketch (fig. 55) of the solar spectrum. A is supposed to stand near the extreme red, and J near the extreme violet.)

[Ill.u.s.tration: Fig. 55.]

The brief memoir of two pages, in which this immortal discovery is recorded, was communicated to the Berlin Academy on October 27, 1859.

Fraunhofer had remarked in the spectrum of a candle flame two bright lines, which coincide accurately, as to position, with the double dark line D of the solar spectrum. These bright lines are produced with particular intensity by the yellow flame derived from a mixture of salt and alcohol. They are in fact the lines of sodium vapour.

Kirchhoff produced a spectrum by permitting the sunlight to enter his telescope by a slit and prism, and in front of the slit he placed the yellow sodium flame. As long as the spectrum remained feeble, there always appeared two bright lines, derived from the flame, in the place of the two dark lines D of the spectrum. In this case, such absorption as the flame exerted upon the sunlight was more than atoned for by the radiation from the flame. When, however, the solar spectrum was rendered sufficiently intense, the bright bands vanished, and the two dark Fraunhofer lines appeared with much greater sharpness and distinctness than when the flame was not employed.

This result, be it noted, was not due to any real quenching of the bright lines of the flame, but to the augmentation of the intensity of the adjacent spectrum. The experiment proved to demonstration, that when the white light sent through the flame was sufficiently intense, the quant.i.ty which the flame absorbed was far in excess of that which it radiated.

Here then is a result of the utmost significance. Kirchhoff immediately inferred from it that the salt flame, which could intensify so remarkably the dark lines of Fraunhofer, ought also to be able to _produce_ them. The spectrum of the Drummond light is known to exhibit the two bright lines of sodium, which, however, gradually disappear as the modic.u.m of sodium, contained as an impurity in the incandescent lime, is exhausted. Kirchhoff formed a spectrum of the limelight, and after the two bright lines had vanished, he placed his salt flame in front of the slit. The two dark lines immediately started forth. Thus, in the continuous spectrum of the lime-light, he evoked, artificially, the lines D of Fraunhofer.

Kirchhoff knew that this was an action not peculiar to the sodium flame, and he immediately extended his generalisation to all coloured flames which yield sharply defined bright bands in their spectra.

White light, with all its const.i.tuents complete, sent through such flames, would, he inferred, have those precise const.i.tuents absorbed, whose refrangibilities are the same as those of the bright bands; so that after pa.s.sing through such flames, the white light, if sufficiently intense, would have its spectrum furrowed by bands of darkness. On the occasion here referred to Kirchhoff also succeeded in reversing a bright band of lithium.

The long-standing difficulty of Fraunhofer's lines fell to pieces in the presence of facts and reflections like these, which also carried with them an immeasurable extension of the chemist's power. Kirchhoff saw that from the agreement of the lines in the spectra of terrestrial substances with Fraunhofer's lines, the presence of these substances in the sun and fixed stars might be immediately inferred. Thus the dark lines D in the solar spectrum proved the existence of sodium in the solar atmosphere; while the bright lines discovered by Brewster in a nitre flame, which had been proved to coincide exactly with certain dark lines between A and B in the solar spectrum, proved the existence of pota.s.sium in the sun.

All subsequent research verified the accuracy of these first daring conclusions. In his second paper, communicated to the Berlin Academy before the close of 1859, Kirchhoff proved the existence of iron in the sun. The bright lines of the spectrum of iron vapour are exceedingly numerous, and 65 of them were subsequently proved by Kirchhoff to be absolutely identical in position with 65 dark Fraunhofer's lines. ngstrom and Thalen pushed the coincidences to 450 for iron, while, according to the same excellent investigators, the following numbers express the coincidences, in the case of the respective metals to which they are attached:--

Calcium 75 Barium 11 Magnesium 4 Manganese 57 t.i.tanium 118 Chromium 18 Nickel 33 Cobalt 19 Hydrogen 4 Aluminium 2 Zinc 2 Copper 7

The probability is overwhelming that all these substances exist in the atmosphere of the sun.

Kirchhoff's discovery profoundly modified the conceptions previously entertained regarding the const.i.tution of the sun, leading him to views which, though they may be modified in detail, will, I believe, remain substantially valid to the end of time. The sun, according to Kirchhoff, consists of a molten nucleus which is surrounded by a flaming atmosphere of lower temperature. The nucleus may, in part, be _clouds_, mixed with, or underlying true vapour. The light of the nucleus would give us a continuous spectrum, like that of the Drummond light; but having to pa.s.s through the photosphere, as Kirchhoff's beam pa.s.sed through the sodium flame, those rays of the nucleus which the photosphere emit are absorbed, and shaded lines, corresponding to the rays absorbed, occur in the spectrum. Abolish the solar nucleus, and we should have a spectrum showing a bright line in the place of every dark line of Fraunhofer, just as, in the case of Kirchhoff's second experiment, we should have the bright sodium lines of the flame if the lime-light were withdrawn. These lines of Fraunhofer are therefore not absolutely dark, but dark by an amount corresponding to the difference between the light intercepted and the light emitted by the photosphere.

Almost every great scientific discovery is approached contemporaneously by many minds, the fact that one mind usually confers upon it the distinctness of demonstration being an ill.u.s.tration, not of genius isolated, but of genius in advance. Thus Foucault, in 1849, came to the verge of Kirchhoff's discovery. By converging an image of the sun upon a voltaic arc, and thus obtaining the spectra of both sun and arc superposed, he found that the two bright lines which, owing to the presence of a little sodium in the carbons or in the air, are seen in the spectrum of the arc, coincide with the dark lines D of the solar spectrum. The lines D he found to he considerably strengthened by the pa.s.sage of the solar light through the voltaic arc.

Instead of the image of the sun, Foucault then projected upon the arc the image of one of the solid incandescent carbon points, which of itself would give a continuous spectrum; and he found that the lines D were thus _generated_ in that spectrum. Foucault's conclusion from this admirable experiment was 'that the arc is a medium which emits the rays D on its own account, and at the same time absorbs them when they come from another quarter.' Here he stopped. He did not extend his observations beyond the voltaic arc; he did not offer any explanation of the lines of Fraunhofer; he did not arrive at any conception of solar chemistry, or of the const.i.tution of the sun. His beautiful experiment remained a germ without fruit, until the discernment, ten years subsequently, of the whole cla.s.s of phenomena to which it belongs, enabled Kirchhoff to solve these great problems.

Soon after the publication of Kirchhoff's discovery, Professor Stokes, who also, ten years prior to the discovery, had nearly antic.i.p.ated it, borrowed an ill.u.s.tration from sound, to explain the reciprocity of radiation and absorption. A stretched string responds to aerial vibrations which synchronize with its own. A great number of such strings stretched in s.p.a.ce would roughly represent a medium; and if the note common to them all were sounded at a distance they would take up or absorb its vibrations.

When a violin-bow is drawn across this tuning-fork, the room is immediately filled with a musical sound, which may be regarded as the _radiation_ or _emission_ of sound from the fork. A few days ago, on sounding this fork, I noticed that when its vibrations were quenched, the sound seemed to be continued, though more feebly. It appeared, moreover, to come from under a distant table, where stood a number of tuning-forks of different sizes and rates of vibration. One of these, and one only, had been started by the sounding fork, and it was the one whose rate of vibration was the same as that of the fork which started it. This is an instance of the _absorption_ of the sound of one fork by another. Placing two unisonant forks near each other, sweeping the bow over one of them, and then quenching the agitated fork, the other continues to sound; this other can re-excite the former, and several transfers of sound between the two forks can be thus effected. Placing a cent-piece on each p.r.o.ng of one of the forks, we destroy its perfect synchronism with the other, and no such communication of sound from the one to the other is then possible.

I have now to bring before you, on a suitable scale, the demonstration that we can do with _light_ what has been here done with sound. For several days in 1861 I endeavoured to accomplish this, with only partial success. In iron dishes a mixture of dilute alcohol and salt was placed, and warmed so as to promote vaporization. The vapour was ignited, and through the yellow flame thus produced the beam from the electric lamp was sent; but a faint darkening only of the yellow band of a projected spectrum could be obtained. A trough was then made which, when fed with the salt and alcohol, yielded a flame ten feet thick; but the result of sending the light through this depth of flame was still unsatisfactory. Remembering that the direct combustion of sodium in a Bunsen's flame produces a yellow far more intense than that of the salt flame, and inferring that the intensity of the colour indicated the copiousness of the incandescent vapour, I sent through the flame from metallic sodium the beam of the electric lamp. The success was complete; and this experiment I wish now to repeat in your presence.[25]

Firstly then you notice, when a fragment of sodium is placed in a platinum spoon and introduced into a Bunsen's flame, an intensely yellow light is produced. It corresponds in refrangibility with the yellow band of the spectrum. Like our tuning-fork, it emits waves of a special period. When the white light from the electric lamp is sent through that flame, you will have ocular proof that the yellow flame intercepts the yellow of the spectrum; in other words, that it absorbs waves of the same period as its own, thus producing, to all intents and purposes, a dark Fraunhofer's band in the place of the yellow.

In front of the slit (at L, fig. 56) through which the beam issues is placed a Bunsen's burner (_b_) protected by a chimney (C). This beam, after pa.s.sing through a lens, traverses the prism (P) (in the real experiment there was a pair of prisms), is there decomposed, and forms a vivid continuous spectrum (S S) upon the screen. Introducing a platinum spoon with its pellet of sodium into the Bunsen's flame, the pellet first fuses, colours the flame intensely yellow, and at length bursts into violent combustion. At the same moment the spectrum is furrowed by an intensely dark band (D), two inches wide and two feet long. Introducing and withdrawing the sodium flame in rapid succession, the sudden appearance and disappearance of the band of darkness is shown in a most striking manner. In contrast with the adjacent brightness this band appears absolutely black, so vigorous is the absorption. The blackness, however, is but relative, for upon the dark s.p.a.ce falls a portion of the light of the sodium flame.

[Ill.u.s.tration: Fig. 56.]