502 



NATURE 



\April 2<^, 1878 



to an eruption of calcium. This, of course, is only a 

 suggestion, but the fact that it is a suggestion merely 

 shows how important it is that this point should engage 

 attention next July. If the prominences are then con- 

 stituted as they were in '75, this violet line will doubtless 

 turn up again, and that is why I have been most anxious 

 to point out not only the conclusions to which we have 

 been led, but the extreme difficulty of arriving at any 

 conclusion whatever, unless by one method or another 

 we have an absolute comparison of the spectrum of the 

 prominences with that of the sun itself. 



I have before referred to the fact of the registration 

 on the plates of a continuous spectrum. If we were to 

 suppose the whole light of the corona to be due to 1474 

 light, for instance, we should expect to get just as definite 

 an image of the corona in the prismatic camera as in an 

 ordinary one. And if everything outside the moon gave 

 us nothing but a line spectrum, the moon's limb would 

 have a perfectly defined edge. Now as a matter of fact, 

 only one such edge is seen in the photographs. We have 

 only one complete ring with a thoroughly defined hard 

 outline, such as that to which reference has been made. 

 This hard ring corresponds to the second spectral image of 

 the prominences, and is a continuation of it. Supposing 

 we were right about the prominences, the ring would be 

 due to the high temperature h line of hydrogen (supposing 

 us wrong it might be a companion line to 1474); as the 

 observations of Respighi, Janssen, and others, in the 

 Indian eclipse of '71 endorsed the American observations 

 of '69 that the hydrogen lines are the strongest in the 

 photographic parts of the corona, we may very possibly 

 be really dealing with hydrogen. 



Now the edge of the corona, or the upper part of it 

 considering it as the sun's atmosphere, as seen on our 

 photographs, is precisely such as would be given by 

 homogeneous light ; that is, there is a distinct image, 

 and there is one image and not three or any other 

 number. Have we any means of determining the 

 Avave-length of the light by which this image has 

 been produced? Let me give an idea of one method 

 which we employed : — A circle of the same size as 

 the image of the moon on a photographic enlargement of 

 the original negative was cut in paper and placed over 

 the enlargement until the corona was symmetrical round 

 it, as we know it to have been symmetrical round the 

 moon's body, or nearly so, at that phase of the eclipse. 



We found as a considerable endorsement of the assump- 

 tion which we made regarding the hydrogenic origin of 

 the chromospheric images, that the paper circle in this 

 position had its circumference coincident with the hard 

 ring to which I have referred as being a continuation of 

 the middle spectral image of the prominences. Next, 

 one of the ordinary photographs of the corona was en- 

 larged to the same size as that of the one produced in the 

 prismatic camera. When these were superposed so that 

 the outlines of both coincided as much as possible, it was 

 again found that the edge of the moon lay along the ring. 



Now then for the continuous spectrum. The general 

 wooUiness of the photographs which at first sight gives rise 

 to the idea that they were out of focus, and that there is 

 nothing to be got out of them, is of course only in one 

 direction, that at right angles to the edge of the prism 



employed. There is a well-defined structure running 

 parallel to this direction, which of course is the line of 

 dispersion; this structure is doubtless due to irregularities 

 in the corona, drawn out by the prism into bands; it is 

 easy to determine the limits of this continuous spectrum. 

 Examining the centre of the photographs we find that 

 on one side the structure stops short at F, on the other it 

 extends to a considerable distance beyond the prominence 

 image in the ultra-violet, spaces of light being visible 

 beyond 3530. 



From these data we concluded that the continuous 

 spectrum-giving region extends at least to a distance of 

 3' of arc from the sun' s limb. This continuous spectrum 

 is well shown on photographs taken at the beginning and 

 end of the eclipse. One of the plates of the prismatic 

 camera was exposed, until the signal for the end of 

 totality was given. Dr. Schuster states that all the 

 observers agreed that the signal was given rather too 

 late, and the fog on the plate indicates an intense illu- 

 mination ; nevertheless, the edge of the sun is not drawn 

 out into a continuous band but rather into three distinct 

 bands. It is probable, therefore, that when the plate was 

 exposed, only the lower part of the chromosphere had 

 appeared, and that it gave out light of such intensity that 

 everybody imagined that the sun itself had come out of 

 eclipse. I observed this myself in 1871, and a very 

 striking fact it is. 



So much then for the results obtained by the prismatic 

 camera in '75. When the report is issued— and its issue 

 cannot be much longer delayed — it will be seen that the 

 hasty sketch I have now given can be followed in greater 

 detail. 



One of the most remarkable points about the expe- 

 dition to Siam was the failure to obtain even spectra of 

 the sun with the ordinary telespectroscopic cameras 

 employed. No doubt the unforeseen delays which left 

 very little time for the adjustment of instruments, hare a 

 great deal to answer for. I have little doubt that if the 

 attempt is made next July, when any quantity of skilled 

 help will be at hand, and any amount of rehearsal will 

 be possible, that a full measure of success will be 

 obtained, at all events for the most photographic part 

 of the spectrum. An ordinary photograph of the corona 

 was obtained by Dr. Schuster in two seconds ; and my 

 experience with photographic spectra enables me to say 

 that this photograph was taken by means of an almost 

 monochromatic light — that near G. Now as the coming 

 eclipse will enable an exposure of almost 100 times 

 longer than this to be employed, I do not think that 

 the undoubted feebleness of the object need be feared. 

 Besides, this method would enable us to pick up the 

 light of those lower reaches of the chromosphere which, 

 as has been already stated, are of such extreme bril- 

 liancy as to have been mistaken, on many occasions, for 

 the sun itself. 



Up to the present time no attempt has been made to 

 obtain a photographic record of the polarisation of the 

 corona. The difference of colours indicating radial polari- 

 sation observed by me when I used the biquartz in 1871, 

 certainly have left the impression on my mind that it 

 would be quite easy to obtain a permanent record of 

 them. This would be a very valuable result, and one 



