368 



NATURE 



^August 1 8, 1 88 1 



one must take into account ; it is too coarse a statement — I 

 do not use the word in any offensive sense — to say lliat the 

 iron lines in the sun correspond with the iron lines seen on 

 the earth. Which iron lines — which of these horizons — are 

 to be taken ? It will be seen in a moment, if there are differ- 

 ences between these horizons, that if we take any one, v\ e throw 

 all the others out of court, and we have no right to do that ; 

 so that statement about the coincidence in the intensity could 

 not be made with the facts now at our disposal. Any one wish- 

 ing to make that statement would have to go over thai work, 

 and he would, following it lionestly, I believe, find tliat the 



statement was true in no instance whatever. Fig. 40, which is 

 an engraving from a photograph, will show the kind of differ 

 ence (jne gets, even when one deals With the electric arc, which 

 undoubtedly gives an iron spectrum which is the nearest approxi- 

 mation to the Fraunhoferic spectrum. The lines at wave-lengths 

 4325 'o, 43007, 427 1 'o are three of the strons^est iron lines in the 

 arc spectrum, and those at 407 fo, 4063'o, 4045'0 are also strong 

 iron lines, though less strong than the others. Now it will be 

 seen tliat in the solar spectrum the last three are much more im- 

 portant, much thicker, and much darker than the first, so that 

 here is an absolute inversion in the thickness of the lines. I 



appeal to the photograph because there is no partiality about it ; 

 it has no view, no anxiety therefore to intensify one particular 

 part of it at the expense of the other. This photograph is re- 

 ferred to only as the exemplar of many similar reversals which 

 we see whenever such observations are made. 



Let ns now take some iron lines wliich have been studied in 

 spots and storms, and consider the differences in their intensity 

 among the Fravtnhofer lines. We may also note the changes 

 brought about in our laboratories. 



The diagram (Fig. 41) gives the main results in a con- 

 venient manner. It does not profess to go over the whole 

 ground, but I think it will enable me to point out the way in 

 which the phenomena observed on the sun are re-echoed and 

 endorsed by the work which has been done in the laboratory, 



Fig. 40. — Anomalous reversals (iron) from a photograph. 



and how severe the tests applied have been, and how well the 

 view has borne the strain. 



The diagram refers to three lines visible in the first map — three 

 lines that in an instrument of ordinary dispersion might easily 

 be mistaken for a single line in the sun, We have, as before, 

 the intensities among the Fraunhofer lines recorded in the upper 

 part of the diagram ; we then go to our photographs of the arc, 

 and find that the line at 4923^2 is entirely absent. We then pass 

 on to the quantity coil, which gives us the three lines ; but there 

 is a difference between the intcn^ties of the lines as seen in the 

 quantity coil with a jar, and the lines seen in the sun, 491S 

 being thinner than in the sun. If we take the jar out of circuit 

 4923*2 almost disappears, and we get very nearly the same result 

 as we get from the arc. We then try the intensity coil, wliich 

 is supposed to give us an equivalent or higher temperature than 

 the quantity coil does. What do we find there ? That 4923-2 

 is^enormously expanded and developed, apparently at the ex- 



pense of 4918, which becomes thin. Taking the jar out, we 

 come back to a result which is very much like the solar 

 spectrum, with the difference, however, that 491S is somewhat 

 less intense than in the sun. Then c >me the facts which have 

 already been brought forward throughout with special reference 

 to these particular lines, that the two lines which are seen alone 

 in the arc are seen alone in the spots, or at all events in 73 

 spots out of 100 ; and the other line which is so enormously 



Fic. 41 — .^'i.i.,ia ., ^u.wiu^ ilie beuav.._ur or l]ir^:e iron lioes under different 

 conditions, solar and lerrestrial. i, solar spectrum ; 2, arc ;_ 3, quantity 

 coil with jar; 4, quantity coil without jar; 5, intensity coil with jar; 

 6. intensity coil without jar ; 7, spots observed at Kensington : 8, Pro- 

 minences observed by Tacchini ; 9, prominences observed by Young ; 

 10, reversed in penumbra of spot observed on August 5, 1872, by Young ; 

 It, motion indicated by change of refrangibility. 



expanded when we use the highest temperature is seen alone in 

 52 out of 100 prominences by Tacchini. Again, further con- 

 necting this diagram with the last one, we have found in several 

 cases when a ctiange of refrangibility has been observed in the 

 iron lines in the spots visible on the sun that the two lines 4918 and 



