392 



NATURE 



[August 26, 1897 



THE APPROACHING TOTAL ECLIPSE OF 



THE SUN.^ 



V. 



The Spectrum of the Corona. 



I HAVE already stated that one great advantage pos- 

 sessed by the prismatic camera is that it has enabled 

 us to distinctly separate the spectrum of the corona from 

 that of the chromosphere. It has done this because 

 the coronal radiations are hot most intense near the 

 prominences. 



The following reproduction (Fig. 19) of a photograph, 

 recently published by the Royal Society, will show how 

 clean cut the distinction is. A nearly complete ring is 

 shown in 1474 light, while the H and K light only provides 

 us with dots showing the positions of the brightest 

 prominences. 



It is seen at a glance that while the greatest intensities 

 of the H and K lig^ht occupy similar positions, the loci 

 of greatest intensity in the case of the 1474 light are widely 

 different in position ; that is, they do not occupy the same 

 positions as the rings assumed perfect. Careful study 

 shows that the 1474 ring marks out the brightest 

 parts of the corona, and neglects the chromosphere and 

 prominences altogether. 



Except in the case of the ring in the green at 1474 of 

 KirchhofPs scale the coronal rings are very feeble, and 



hydrogen, have not previously been recorded by the 

 prismatic camera, though some of the lines correspond- 

 ing to them appear to have been photographed with slit 

 spectroscopes. The rings photographed in 1893 are 

 compared in the following table with the results obtained 

 by the slit spectroscopes in the years 1882, 1883, 1886,1 

 and 1893,'"^ those lines only which are possibly common 

 being included. 



We see then that all the coronal radiations above 

 referred to probably correspond with lines photographed 

 by Dr. Schuster in 1886. The intensities, however, are 

 not the same. 



The number of coronal rings recorded with the pris- 



FiG. 19. — The 1474 light ring compared with the prominences as depicted in H and K light during the eclipse of 1896. 



their wave-lengths can, perhaps, not be depended upon 

 to within one-tenth metre. They were read off from 

 the African negatives by direct comparison with the 

 spectrum of Arcturus. Possibly, owing to the great 

 intensity of the continuous spectrum, the Brazilian 

 negatives show only the green ring. 



The wave-lengths of the coronal radiations are as under, 

 the numbers following them indicating the number of 

 African photographs in which they have been detected. 



Wave-length, 

 Rowland. 



3987 

 4086 

 4217 

 4231 

 4240 

 4280 

 4486 



/53I6-9 



1.1474 K 



(6) 

 (4) 

 (2) 

 (5) 

 (I) 

 (2) 

 (3) 

 (5) 



It is almost impossible to form any trustworthy 

 estimate of the relative intensities of the rings, but it may 

 be noted that the one at wave-length 3987 comes next to 

 1474 K in order of brightness. There are indications of 

 other extremely faint rings, the positions of which cannot 

 be determined with the necessary degree of accuracy to 

 enable a useful statement to be made touching their 

 vave -lengths. 



Coronal rings, other than those due to 1474 K, or 



1 Continued from page 368. 



NO. 1452, VOL. 56] 



matic camera is very much smaller than the number of 

 lines attributed to the corona photographed with the slit 

 spectroscopes in this and previous eclipses. This is, no 

 doubt, partly due to the rings being submerged in con- 

 tinuous spectrum, which is relatively more intense in the 

 case of the prismatic camera. Further, as already pointed 

 out, it is not yet established that many of the lines 

 recorded as being due to the corona by the slit spectro- 

 scopes are not due to glare. 



By a comparison of the results obtained with slit 

 spectroscopes and prismatic cameras, it would seem to 

 be possible to determine which of the lines recorded by 

 the former instruments really belong to the coronal 

 spectrum. The most intense light will give the strongest 

 glare, and therefore the brightest lines of the chromo- 

 sphere and prominences will become superposed on those 

 due to the corona. As the results obtained with the 

 prismatic cameras are so very definite with regard to the 

 spectrum of the prominences, it seems only necessary to 

 subtract the common lines of the spectra recorded by 

 the two instruments from the total number recorded by 

 the slit spectroscopes in order to determine those which 

 certainly belong to the corona. 



An attempt has been made to investigate the coronal 

 spectrum in this way by reference to the slit spectra of 

 1886 and 1893, but no satisfactory results can be obtained 

 in this way until slit spectra taken with greater dispersion 

 become available. 



1 Phil. Trans., 1889, A, p. 335. 2 Roy. Sac. Prac.,\o\. 56, p. 20. 

 3 Intensity on a scale, where 6 = brightest line. 



