324 



NATURE 



\Augnst 4, 1 88 1 



whatever is visiljle. Therefore as the spectra of prominences and 

 of storms may be stated to be the spectra of the hottest regions 

 of the sun that we can get at in our inquiries. The hues in the 

 solar spectrum afifected neither in spots nor flames give us an 

 approach to the cool spectrum we are in search of. We might 

 expect if differences were observable that we should get some- 

 thinT like this — 



Lines special to prominences = hottest. 



Lines special to spots = medium. 



Lines affected neither in spots nor storms.. = coolest. 

 How have these views been tested. The first attempt made 

 to get light out of this inquiry was one which simply dealt 

 with a long catalogue of lines observed by Prof. Young in the 

 memorable expedition of his to Mount Sherman, AAhere, at the 

 height of between 8000 and gooo feet, with perfect weather and 

 admirable instrumental appliances, about a month was employed 

 ia getting such a catalogue of lines as had never been got before. 

 But it was found that, although the result of this inquiry was 

 absolutely m harmony with these views, still after all one wanted 

 more facts. Therefore we have endeavoured to get some of the 

 facts here. And the way in which they have been collected is 

 as follows : — During the last two years the spectra of 100 sun- 

 spots have been observed in the observatory here — observed in 

 a new fashion, and for a good reason I think. In this changeable 

 climate it does not do to do as we began by doing — to attempt 

 to ob erve all the lines acted upon in a solar spot. The exces- 

 sive complication, and the intense variation of a spot-spectrum 

 from the ordinary solar spectrum, cannot be better shown than 

 by throwing on the screen the spectrum of one of the sun-spots 

 lately observed at Greenwich. 



The figure (Fig. 34) shows a limitel part of the solar spec- 

 trum, and the lines thickened in the spot-spectrum. It will be 

 seen therefore that to tabulate the existence and thickness and 

 ictensities of these lines over the whole of the solar spectrum 

 would be a work which it would be difficult to accompli6h in a 

 single day, even if the day were absolutely fine. So that was 

 given up in favour of a limited inquiry over a small part of the 

 solar spectrum ; limited further by this, that we only get the 

 twelve lines most affected in each spot on each day. In this way 

 we insure a considerable number of ab_=olutely comparaljle 

 observations, and we can more easily compare the spot results wiih 

 those which had been obtained in the obervation of the brightest 

 lines in prominences, because when we begin to observe lines in 

 the solar prominences one nalurally begins by observing the 

 brightest lines first. So that by observing the darkest lines first 

 in the case of spots, one has a fairer comparison. 



A diagram (Fig. 35) will show the result of our observations of 

 100 spots over a very limited pai-t of the solar spectrum. We will 

 begin by the individual observations. We have at the top lluj 

 iron lines recorded among the Fraunhofer lines ; below we have 

 the iron lines recorded as iron lines by Angstrom, who used an 

 electric arc. Lower down we have the iron lines recorded by 

 Thalen, who u^ed the electric spark. It will be seen that there 

 is a very con iderable difference in the spectrum of iron as viewed 

 by means of ihe spark and by means of the arc, and that there is 

 an equal difference between the jpectnim of iron in the sun, 

 that is to say, in the whole «un, determined by the Fraunhofer 

 lines, and the spectrum of either the arc or the spark. It is 

 also to be n-ited that the solar spectrum is more like the spec- 

 trum of the arc than the spectrum of the spark. 



Since the relative iritensities in all these cases are represented by 

 the length of the lines, we have here an opportunity of observing 

 and discu sing the accuracy of KirchhofT's statement that 

 the iron lines in the sun correspond absolutely in intensity 

 with the lines of iron seen in a light soui'ce here. It is 

 necessary first of all to see which light source he fixes on, 

 whether the arc or the spark. When this has been done it is 

 found that the statement is really true with regard to neither. 



That however is a digression ; to proceed v\dth the diagram, 

 descending from this general spectrum of iron which we get by 

 the absorption of the whole atmosphere of the sun independently 

 of the hottest region and the coldest region — descending from 

 the general to the particular — and taking that particular part of 

 the solar atmosphere v\here the spots produce their phenomena, 

 let us see what are the results in the case of the spots? We 

 have in the vertical lines a record of the lines which are afifected 

 in each spot, and each of the spaces included between the 

 horizontal lines repreents a particular spot, the date being given 

 on the right hand side ; and these 100 lines which we have here 

 represent the phenomena produced by lOO spots. The diagram 



is a small portim of the larger map. Now the wonderful thing 

 that one is at once struck with is the absolute and complete irregu- 

 larity of the whole result. There is no continuity among any 

 of these lines. A careful inspection of the diagram shows us 

 that, speaking in a general way, each of these lines is seen in 

 one spot or another absolutely without the other. We have an 

 inversion in the intensities of the lines when piassing from spot 

 to spot. Whenever we get a line intensified by Thalen, we miss 

 it in the spots, and, as a rule, what happens is that the spectrum 

 of the spot is not only simpler than the spectrum of the arc, but 

 simpler than the spectrum of the spark. 



Now the importance of these statements depends on other 

 statements which we can bring to confront with them. The next 

 diagram shows the observations of 100 prominences observed 

 between the years 1872 and 1876. (The diagram was thrown on 

 the screen.) Prominences exist in a region of the solar atmo- 

 sphere not very far from that occupied by the s]iols, but we have 

 already seen that whereas the spots are produced by a downrush 

 of cool material, prominences are produced by an uprush of hot 

 material. Let us see therefore if any change is produced in the 

 phenomena ; whether we shall have exactly the same lines from 

 the flames, or the prominences, as we have from the spots ; 

 whether we shall get the same information or no. 



Here ai-e the facts with respect to Tacchini's observations : — 

 We begin as before with the whole absorption of the sun. Ang- 

 strom's map, and Thalen's map. I think you will see a very 

 considerable change ; the iron lines (for we are only dealing 

 with iron) most prominent in the prominences are vastly different 

 from the iron lines most thickened in the spots. The difference 

 is shown in the annexed diagram (Fig. 36), which represents those 

 individual observations both of spots and flames treated in a 

 certain way with refei'ence to the discu sio 1. I will at once 

 explain to you what that certain way is. We have, as before, 

 the three data to begin with, and we have treited the sun-.spot 

 observations so that the lengths of the lines will represent the 

 number of times they have been seen in 100 sun-spots ; the line 

 at wave-length 4919 S, for instance, has been seen seventy-two 

 times ; that line, in fact, has been seen more than any other ; 

 the one at 5005 'o some forty times, and so on; very many 

 lines having been seen less than ten times. In another part of 

 the same diagram we have summarised the individual results 

 obtained from Tacchini's observation of prominences in exactly 

 the same way. The line 501 7 '5 was seen in 66 prominences 

 out of 100. But why I am particularly anxious to show this 

 diagram is this, that it brings out the perfectly natural fact — 

 for it is the natural fact— that over this region of the spectrum, 

 at all events, no iron lines affected in the spots are visible in the 

 prominences. If we asmme that the region occupied by pro- 

 minences is hotter than the region occupied by spots, that higher 

 region ought to do this work, and it ought to be a work of 

 simphfication. Therefore I say it is a perfectly natural result, 

 and not one to be wondered at, that in the spectra of the flames 

 there is no line coincident with any of the lines seen frequently 

 widened in the spots. 



Now we have these three solar spectra here which we can 

 compare one with the other. Fii-st of all we have the iron spec- 

 trum of the sun taken as a whole. Then we have next the 

 spectrum of spots, which we know to be hotter than the sun 

 taken as a whole. Then we have the spectrum of flames, which 

 we know to be hotter than the spot-. It will be seen that the 

 story, as it runs from the top of the diagram downwards, is a 

 story of greater simplicity, as it ought to be, and it was explained 

 in the diagram which I exhibited befoi-e I began to show these 

 results of absolute hard facts. It will be seen that the sim- 

 plicity brought about by the reduction of lines actually seen 

 as to number, is accompanied by the appearance of new lines 

 (produced by the transcendental temperatures) in these regions. 

 This first discussion of a large number of spectra and of spots, 

 as compared with storms, is, I submit, in absolute harmony with 

 the view of the dissociation of the elementary bodies by the 

 solar temperature suggested by Sir Benjamin Brodie in 1867, 

 and therefore I may further add that to me, at all events, il 

 is absolutely inexplicable on any other view. 



J. NORMA-N LOCKYER 

 ( To he continued. ) 



INTERNATIONAL MEDICAL CONGRESS 



THIS Congress, which opened by an informal reception 

 at the College of Physicians on Tuesday, has so far 

 been a real success. It has brought together something 



