another jet, from which burns the blue flame of carbonic oxide gas, the body which produces the blue lambent flame frequently seen in coal fires: the temperature of this flame is somewhat higher, and has been calculated at 3,000°. If I bring a little common salt (sodium chloride) into these flames, you observe that in all cases we get them coloured yellow. I have here a hydrogen flame, of which the temperature is 3,259° C., and you see, when we bring the sodium salt into it, we have the same yellow colour produced; in other words, we cannot get sodium vapour either red-hot or blue-hot, it always remains yellowhot; that is to say, the first moment that the sodium vapour becomes luminous, it gives off this particular and peculiar yellow light, and if we heat it more, the effect is not to alter the refrangibility of the rays, but merely to increase their intensity.

As a further illustration, we have this oxyhydrogen flame, of which the temperature is said to be about 6,000° C. If I bring a piece of soda into it, the effect is intense ignition; but still there is only the yellow light, no blue light. This indicates to us that when a body becomes gaseous, the light which it gives off is of a particular kind, and does not alter when we increase the temperature. One other experiment will indicate this to you still more fully, and this I can make by means of the electric spark, which I have here the means of producing. The temperature of this electric spark is so high that it has never been measured, but it is certainly infinitely higher even than the temperature of the oxyhydrogen flame. Still, if I bring this piece of sodium salt into the electric spark, I find that the same thing occurs-I get the same yellow-coloured light; and if I take some other substance, such as lithium, the per

manent red colour which lithium vapour gives off will be clearly seen.

Now the methods by means of which we can obtain bodies in the state of luminous gas vary with the nature of the substance, but I would beg you to understand that the property which we have noticed with regard to sodium and the other alkalies is not confined to those bodies which have the power of being volatilized in such a flame as I have burning before me. This property belongs to matter in general; it belongs to every chemical element; and if we can by any method get the vapour of a chemical element so hot as to become luminous, we find that the light emitted by it is peculiar to itself, and is distinctive of that special body, whether under the ordinary circumstances the element be gascous, solid, or liquid. Hence you see that we have at last reached the principles upon which the science of spectrum analysis is based, by means of which we can detect the presence of any of the elementary bodies when they can be obtained in this state of glowing gas.

We must now pass on to the consideration of the various methods by which the elements can be obtained as luminous gases.

I purpose to confine our attention in this lecture to the method by which we can detect the presence of the metals of the alkalies and alkaline earths. Let me, however, first point out to you the kind of spectrum which we obtain when we look at any one of these variouslycoloured flames through a prism or spectroscope, the construction of which we will now briefly consider.

The simplest form of spectroscope which Bunsen first. adopted is represented in Fig. 22. It consists of a common hollow prism (F) placed in a box; a telescope (c) is fixed

at one side of the box, and a slit is placed at one end of a tube having a lens at the other end, in order to obtain a pure spectrum, and to render the rays parallel; and this

collimator (B) is fixed at the other side of the box. The substance to be examined is placed in the non-luminous Bunsen's flame, and the light passing through the slit falls upon the prism, and having been split up into its

constituent parts, the differently-coloured rays pass through this telescope, are magnified, and then fall upon the retina. In Fig. 23 we have the more perfect form of

the instrument represented, as made by Steinheil of Munich. With this we are enabled to use two flames, and the apparatus is so arranged that we can see the two spectra placed one above the other. The object of this superposition of the spectra is evident: it is to enable us to see whether the substance under examination really is the body which it is supposed to be. For instance, putting a small quantity of the substance we know to contain sodium in this flame, we place a substance sup

FIG. 24.

posed to contain sodium in the other flame, and then by means of a small reflecting prism placed on the end of the slit we have the spectra of these two flames sent into the telescope one above the other, so that we see at the same time the spectrum of the pure sodium and the spectrum supposed to be that of sodium; and we can readily observe whether the lines coincide. If they coincide, and the two spectra have these lines exactly continuous one below the other, then we are quite certain that sodium, or any other substance which we may have been investigating, is

1 This instrument consists of a prism (a) fixed upon a firm iron stand, and a tube (6) carrying the slit (d), seen on an enlarged scale in Fig. 24, through which the rays from the coloured flames (e and é) fall upon the prism, being rendered parallel by passing through a lens. The light having been refracted, is received by the telescope (ƒ), and the image magnified before reaching the eye. The rays from each flame are made to pass into the telescope (f); one set through the upper uncovered half of the slit, the other by reflection from the sides of the small prism (c), Fig. 24, through the lower half; thus bringing the two spectra into the field of view at once, so as to be able to make any wished-for comparison of the lines. The small luminous gas flame (h), Fig. 23, is placed so as to illuminate a fixed scale contained inside the tube (g): this is reflected from the surface of the prism (a) into the telescope, and serves as a means of measuring the position of the lines.

present. Another arrangement for facilitating the comparison of spectra consists in the illuminatel millimetre scale contained in the tube g (Fig. 23), a magnified reflection of which is thrown into the telescope from the surface of the prism. The illuminated scale is thus seen between the two superimposed spectra, and the position of any line or lines can be accurately determined. The further arrangements--mechanical and optical-of these inrtruments I need hardly trouble you with in detail. I have here a variety of spectroscopes kindly lent to me by the maker, Mr. Browning; one with one, one with two, one with three, and one with four prisms. The more prisms we employ, of course the greater dispersion we get, the more is the light drawn out into its special varieties; and the greater also is the intensity of the light which it is necessary to employ in order to get the rays to pass through this greater number of prisms.

I will next show you a drawing of the actual arrangement used by Kirchhoff (Fig. 25). There you see the prisms employed, four in number, placed one behind another on a horizontal table of cast iron. The light passes through the slit at the end of this tube. Here (top of Fig. 25) is an enlarged representation of the slit, the breadth of which can be altered at pleasure by means of the screw; on this slit is placed a small reflecting prism to enable us to get two superposed spectra. The light passes through the fine vertical slit, the rays are rendered parallel by the lens fixed at the end of the tube (A); it then passes through these four prisms, and the rays thus split up into constituent parts fall on to the telescope (B), at the end of which the eye is placed. This, then, gives you the simplest, and at the same time the most delicate and complete, form of spectroscope.