446 



NATURE 



[Sei'TEmber 9, 1897 



which are present in the Fraunhofer spectrum, it shows 

 many bright lines which are not represented among the 

 dark ones. This again indicates that the chromosphere 

 is not the origin of the Fraunhofer spectrum 



It is all the more important to call attention to the 

 advantage we now possess in being able to directly com- 

 pare photographs of the chromosphere obtained during 

 eclipses with others of the spectra of stars resembling 

 the sun, since, as I have already stated, if all goes well 

 next year, double the dispersion utilised in 1893 will be 

 employed. This is certain not only to enable us to 



prominence with the lower parts gradually cut off by 

 the moon's edge. In the case of a prominence at the 

 opposite limb, similar sections will be represented in 

 successive photographs, and the last photograph taken 

 during totality will show the spectrum of the greatest 

 part of the prominence. 



Some of the 1893 prominences (Nos 3 and 19) have been 

 investigated in this way, and particulars of their spectra 

 at various heights recorded. The height above the 

 photosphere, reckoned in seconds of arc and in miles, 

 at which each spectrum is given, has been calculated. 



(0 



(2) 



Fig. 24. — -The spectrum of A 



(i) compareil with that of the base of the chromosphere photographed during the eclipse of 1893 (2). 



obtain more accurate wave-lengths, but the number of 

 Arcturus- and chromospheric-lines obtained by the same 

 instrument will be very greatly increased. 



"In the meantime we must take the above as one of the 

 most positive results secured in the eclipse of 1893. 



The Spectra of Prominences at Different Heights. 

 There is another matter of almost equal importance in 

 which the increased dispersion designed to be employed 

 in 1898 will in all probability prove of the utmost value. 



The relative intensities of the l-nes at different heights 

 have been tabulated. In this way it has been found 

 that some of the lines remain of the same relative in- 

 tensity throughout all parts of the same prominence ; 

 others again dim rapidly in passing towards the upper 

 parts, while some, but not so many, brighten. 



The prominences are also seen to behave differently in 

 respect to some of the lines; thus the line at X 38s6'5 

 disappears before a height of 2000 miles is reached in 

 prominence No. 3, but remains visible at a height of 



Fig. 25. — The spectra, a, <5, c, of Prominence No. 3, photographed during the eclipse of 1893, compared with the spectrum of the base of chromosphere (d). 



(a) 22"-26" above photosphere. 

 W 7"- 9" „ 



I refer to the differences in the spectra as we work out- 

 wards from the photosphere. 



The questions touching the spectra of prominences at 

 different heights, which the prismatic camera enables 

 us to study with minuteness, must really lead in time to 

 a much better knowledge of the loci of absorption m the 

 solar atmosphere. If we consider a prominence on that 

 part of the sun's limb where the second contact takes 

 place, the first photograph taken during totality will 

 show the spectrum of the whole prominence, and succeed- 

 ing photographs will give the spectrum of the same 

 NO. 1454. VOL. 56] 



over 4000 miles in prominence No. 19. Lines also occur 

 in one prominence which do not appear in the other, e.g. 

 A 43i3'2. Other differences are also revealed, but it may 

 be remarked that too much stress should not be laid on 

 the presence or absence of the very faintest lines in some 

 of the photographs, as variations may be partially attri- 

 buted to differences in the quality of the photographs, and 

 also to the time of exposure and degree of development. 

 The changes in the spectrum of a prominence in pass- 

 ing from the top towards the base are illustrated in Fig. 

 25. Spectra a, b, and c represent the spectrum of Pro- 



