394 



NATURE 



\Augnst 25, i! 



undoubtedly due to Prof. Young, and although he has lately 

 seen cause to withdraw somewhat from this first view of his, he 

 is still driven to think that two of the most important lines in 

 the solar spectrum, H and K, are really due to two substances — 

 hydrogen and calcium. There is one more thing which must be 

 said with special reference to this : the work is not to be nega- 

 tived by a mere assertion that the line is double ; it must in a great 

 many cases be shown that it is a double-double or a treble-treble 

 line. For instance, take the case of 1^4, which may very well be 

 double for aught we know ; that line is coincident, so far as we 

 can make out with the means at our command, with magnesium, 

 iron, nickel, uranium, chromium, and cerium. It will not do 

 to limit oneself to the statement that these lines are double — 

 they must really be perfect famihes of lines in order to prevent 

 this explanation being, as I think it is, the more probable. 



Another objection, again, a very important objection, is of a 

 somewhat different nature, and may shortly be called the bell 

 hypothesis. It is to the effect that these molecules, or atoms, 

 are very extraordinary things indeed, as we can well imagine 

 them to be, and that the spectrum which they produce depends 

 entirely upon the manner in which they are struck ; so that in 

 fact it would seem at first useless to construct a map of any 

 spectrum at all, for fear the substances we wish to observe in our 

 laboratories should be struck in different ways and should render 

 the map perfectly useless. I say the idea is really that the mole- 

 cules struck differently would give us different spectra. Now 

 if the difference were only slight, that would not much matter ; 

 it would be very difficult to withstand that hypothesis. But it 

 must be remembered that in this work we are dealing with this 

 extraordinary fact, that over the region which we have been spe- 

 cially studying there are no lines of iron common to spots and 



flames ; so that if we had not any iron at all to experiment with 

 we should be perfectly justified in asserting the iron lines in flames 

 were produced by one substance and those in spots by another, 

 because no two lines agree in these two spectra. The spectra 

 are as distinct as the spectrum of magnesium and the spectrum of 

 iodine, or any other two bodies. Now if the bell hypothesis is 

 to explain that, it explains too much, because if it is true of any 

 two bodies, it must be true of all bodies, and therefore all 

 spectra are the result of the ;ame thing struck differently, and 

 spectrum analysis would then cease to be spectrum analysis, for 

 it wuuld simply record changes rung on the same molecule by 

 the various methods of striking. Then again there is another 

 thing to be said, that no statement of this bell hypothesis which 

 I have heard gets us out of the difficulty that we are sinning 

 against the law of continuity in advancing it, for the reason that 

 if you begin with a known compound body, let us say, a salt of 

 calcium, the change from a salt of calcium to calcium is the same 

 in kind and about the same in degree as the change from one 

 form of calcium to the other, if we can talk of different forms 

 of calcium on the meie strength of spectroscopic work. I mean 

 that there are more important changes to be got out of the ob- 

 servations of the metal calcium than there are to be got by 

 passing from a salt of calcium to calcium ; so that if the bell- 

 hypothesis proves anything it proves that a compound body is a 

 simple one. 



It will be seen that the special import of these considerations 

 lies in the question of the short lines ; leaving all considerations 

 dependent on long lines, by which the presence of impurities may 

 be recognised, out of consideration altogether. But it may be 

 further said that a method of purifying spectra and eliniinating 

 any spectroscopic defects which were due to the presence of 



FiG. 46. — Diagram showing how the cvulution of 



impurities has been before the scientific world for some years, 

 and so far as I know its validity has never been called in ques- 

 tion. Where that method has been employed I believe it has 

 been employed honestly, so that until that method is called in 

 question by better work — and the work I know will be severe — 

 I think we are boujid to accept the results which have been 

 obtained by it. 



Then there are some theories which I might be permitted to 

 say a few words about, but in reference to them I need only call 

 attention to Dr. Schuster's admirable report on the progress of 

 spectrum analysis, given to the British Association at the 

 meeting at Swansea. It will be seen from that report that none 

 of the theories which have been put forward really account for 

 all the facts observed. It is shown that phenomena have been 

 recorded which are not to be explained away by the bell theory 

 or any other. Such a phenomenon, and a perfectly distinct 

 one, is the change due to the thickness in the vapour. Changes 

 also due to varying temperature and other causes have been 

 seen, for which the theories in question do not account. Now 

 changes in temperature may probably affect large reaches of the 

 spectrum, but in the cases we have studied we got most diverse 

 effects in lines so close together in the spectrum that it requires 

 a considerable amount of dispersion to find out that they are 

 really not single lines. 



The New TVuory of Chemical Evolution 

 What then is the view of the evolution of chemical species 

 to which wefare led by om- study of the sun and stars? I 

 think that after all it is but a slight expansion of the pre- 



hemical forms may be indicated by their spectra. 



sent chemical view. Chemists regard matter as composed of 

 atoms and molecules, about which more presently. The view 

 now brought forward simply expands the series into a larger 

 number of terms, and suggests that the molecular grouping of a 

 chemical subst.ance may be simplified almost without limit if the 

 temperature be increased. A diagram (Fig. 46) will show ex- 

 actly what I mean. If we assume a very great difference in 

 the temperature which can be brought to bear upon a substance 

 we may nssume that at the highest temperature we have, for 

 simplicity's sake say, a certain line represented by a single circle ; 

 let us imagine the temperature reduced, we shall then get another 

 spectrum, which we can represent by a double circle, if we like 

 to assume that the evolution is one which proceeds by constant 

 additions of the original unit. Coming lower down, we get 

 another substance formed with a more complex spectrum repre- 

 sented by three circles ; lower down still we have one repre- 

 sented by four circles, another by five, another by six, and so 

 on. We might take another sujiposition, easier perhaps to some 

 minds, and suppose that evolution proceeded, not by the addi- 

 tion of the initial unit, but by the constant doubling of the sub- 

 stance of the molecule itself. Instead, therefore, of our circles 

 increasing by one, we shall have one, two, four, eight, sixteen, 

 thirty-two, and it will be readily understood that if there are a 

 considerable number of stages of temperature, both within our 

 ken and beyond our ken, and if some substances form themselves 

 perpetually by doubling, then the unit with which we can expe- 

 riment at low temperature, call it the chemical atom or the 

 chemical molecule, or what you will, must be a very complex 

 thing indeed. If the lower spectrum represents that of a complex 



