36 



NATURE 



[Nov. 13, 1879 



it is particles of carbon, probably in the solid state, that 

 in the electric light afford us every variety of ray so as to 

 enable us to get from them a continuous spectrum. 

 When, however, we go from solid and liquid particles to 

 those of a gaseous nature, we find the various molecules 

 so far .ipart that each one is unconstrained by its neigh- 

 bour ; it is thus like a bell left to itself, in which case it 

 gives out its own peculiar kind of light just as a 

 bell, left to itself, will give out its own peculiar note. I 

 will now show you on the screen the various rays or 

 luminous notes given out by particles of incandescent 

 vapour of silver. 



We tints see what is the spectroscopic difference 

 between solids or liquids, and gases, the former when 

 sufficiently heated giving out a continuous spectrum con- 

 sisting of ail different rays of light, the latter a discon- 

 tinuous spectrum consisting of only a few different rays. 



The next point to which I will call your attention is a 

 very important one. A particle when cold or compara- 

 tively cold absorbs those very rays which it gives out when 

 hot. Now it is known that incandescent vapour of the 

 metal sodium, gives out under certain conditions a peculiar 

 monochromatic yellow light, which we call the double line 

 D. This light is so strictly monochromatic that all bodies 

 under its illumination appear either yellow or black, as 

 you will see by the following experiment. 



Now suppose we take the electric lamp, the carbon points 

 of which, as you already know, give out all kinds of iight, 

 and suppose we place between these points a piece of 

 metallic sodium ; while this sodium is in the act of being 

 volatilised, and its vapour comparatively cold, you will see 

 that it will stop one particular kind of light, and will thus 

 cause a black line. When, however, the vapour is suffi- 

 ciently hot, this black line will be changed into a bright 

 yellow one. You thus see that when we have an incan- 

 descent body which gives us all rays, and when between 

 it and the eye we insinuate some comparatively cold 

 sodium vapour, we get a certain definite black absorption 

 line. 



Now the curious point is that the sun's light gives us 

 this black line, so that if I could replace the electric light 

 by the sun, I should have a black line thrown upon the 

 screen in the very position where you saw it when the 

 sodium was introduced. 



This means that between the source of the sun-light 

 and the eye, we have sodium vapour in a comparatively, 

 remember only comparatively, cold state, and as this 

 vapour is certainly not in the earth's atmosphere, it can 

 only be in the atmosphere of the sun. I need not tell you 

 that although colder than the particles beneath it which 

 give us sun iight, it must be in reality very hot. The dis- 

 covery that there was vapour of sodium in the atmosphere 

 of the sua was due to Stokes, and it has since been found 

 out by Kirchhoff that we have black lines in sun light cor- 

 responding in position with the bright lines of iron vapour, 

 the bright lines of hydrogen, the bright lines of mag- 

 nesium vapour, and the bright lines of many other 

 elements, and we may therefore assume as Kirchhoff 

 assumed, as a first and approximative hypothesis, that 

 the vapour-, of iron, magnesium, hydrogen, &c, as well as 

 that of sodium exist in a comparatively cold state in the 

 atmosj here of our luminary ; — more recent work by Hug- 

 ging-, an 1 o hers has shown that the same remark applies 

 to the at no>pheres of many other stars. 



You thus see that there are two ways by means of 

 which the chemical composition, or rather perhaps the 

 atomic stru tare of bodies may be indicated by the spec- 

 trum. At a comparatively low temperature this structure 

 will be indicated through the lines that are absorbed or 

 rendered bla k, while at a comparatively high temperature 

 it will be indicated by the bright lines that are given out. 



Thus at a comparatively low temperature a solution 

 which contains blood will indicate the presence of this 

 subsian e by certain very peculiar black lines. Blood, 



however, is easily decomposed by a high temperature, and 

 accordingly when such is applied we no longer get the 

 bright equivalents of these black lines, but something 

 very different, namely, the bright lines of iron, and of 

 those other elements into which blood is decomposed as 

 the temperature is raised. In short when raising the 

 temperature of a substance its black lines will be con- 

 verted into bright ones only in those cases where no 

 molecular change has taken place between the two tem- 

 peratures. Even in the case of elements like sodium 

 Roscoe and Schuster have shown that the absorption 

 spectrum at a low temperature is different from, and more 

 complex than, the radiant spectrum at a high temperature, 

 and other elements have been tried in this way by Lockyer 

 and others with similar results. We may imagine with 

 much propriety that the molecule of sodium vapour at a 

 low temperature is a larger and more complex structure 

 than it is at a high temperature, where the splitting up or 

 dissociating agency of heat has been freely employed. 



We come at last to the important question which it is 

 my object to discuss. Has a study of the spectrum 

 thrown any light on the ultimate constitution of matter, 

 or is it likely to do so ? 



You are aware that chemists and physicists have begun 

 to speculate as to the possibility that the so-called 

 elements may be in reality nothing more than combina- 

 tions differing in numbers and in tactical arrangement, of 

 some one kind of primordial atoms. 



This idea was first entertained by Dr. Prout, the well- 

 known physician and chemist. He pointed out that the 

 atomic weights of the so-called elements are very nearly- 

 all multiples of the half of that of hydrogen, so that the 

 various elements may possibly be looked upon as formed 

 by a grouping together of certain atoms of half the mass 

 of the hydrogen atom. 



M. Stas, the distinguished Belgian chemist, instituted 

 a laborious series of experiments with the view of testing 

 this doctrine. He came to the conclusion that the atomic 

 weights of the various elements were not precisely multiples 

 of the half of that of hydrogen, there being greater 

 differences than could possibly be accounted for by errors 

 of experiment. His researches, however, seemed to show- 

 that in many cases there was a very near approach to 

 Prout's imagined law. But here we must bear in mind 

 the great difficulty, or indeed impossibility, of obtaining 

 substances absolutely free from all impurities (indeed 

 Dumas showed that oxygen forms part of the silver with 

 which Stas worked), so that we may be excused from 

 imagining that Stas has settled the point in the negative. 

 We are thus driven to look to the spectrum as a likely 

 means of throwing some light on this very interesting and 

 important speculation. 



Let us now, therefore, endeavour to realise what would 

 be the behaviour of the spectrum if the so-called elements 

 were not capable of simplification, and also what would 

 be its behaviour if they were, and then find with which of 

 these two hypotheses the true behaviour of the spectrum 

 agrees best. Now if the elements were absolutely simple 

 bodies, we might still expect that the molecule of vapour 

 of an element would be at a low temperature more com- 

 plex than at a high one, and would therefore give out a 

 more complex spectrum. As, however, the temperature 

 was made to rise we might expect ultimately to obtain a 

 certain spectrum which would represent the simplest 

 mode of vibration of that element, and which would 

 thenceforward remain, however much higher the tempera- 

 ture should be made to mount. Lockyer has written 

 much on this point and given many facts in support of 

 this view. 



And again we should have no reason for supposing that 

 the lines of the ultimate spectrum of one element should 

 coincide in position with those of the ultimate spectrum 

 of another element. If therefore we had a mixture of all 

 the elements, and subjected this mixture to a very high 



