August 30, 1877] 



NA TURE 



365 



limited area upon the disc or margin of the sun, but the 

 spectrum of light from the whole disc. I have not used 

 an image of the sun upon the slit of the spectroscope, but 

 have employed the beam reflected from the flat mirror of 

 the heliostat without any condenser. 



In support of the above assertions the accompanying 

 photograph of the solar spectrum with a comparison 

 spectrum of air, and also with some of the lines of iron 

 and aluminium, is mtroduced. The photograph itself is 

 absolutely free from handwork or retouching. It is 

 difficult to bring out in a single photograph the best points 

 of these various substances, and I have therefore selected 

 from the collection of original negatives that one which 

 shows the o.\ygen coincidences most plainly. There are 

 so many variables among the conditions which conspire 

 for the production of a spectrum that many photographs 

 must be tiken to exhaust the best combinations. The 

 pressure of the gas, the strength of the original current, 

 the number of Leyden jars, the separation and nature of 

 the terminals, the number of sparks per minute, and the 

 duration of the interruption in each spark, are examples 

 of these variables. 



In the photograph the upper spectrum is that of the 

 sun, and above it are the wave-lengths of some of the 

 lines to serve as reference numbers. The wave-lengths 

 used in this paper have been taken partly from Angstrom 

 and partly from my photograph of the diffraction-spectrum 

 published in 17S2. The lower spectrum is that of the 

 open air Leyden spark, the terminals being one of iron 

 and the other of aluminium. I have photographed 

 oxygen, nitrogen, hydrogen, and carbonic acid, as well as 

 other gases in Pliicker's tubes and also in an apparatus 

 in -vhich the pressure could be varied, but for the present 

 illustration, the open air spark was, all things considered, 

 best. By other arrangements the nitrogen lines can 

 readily be made as sharp as the oxygen are here, and the 

 iron lir.es may be increased in number and distinctness. 

 For the metals the electric arc gives the best photo- 

 graphic results, as Lockyer has so well shown, but as my 

 object was only to prove by the iron lines that the spectra 

 had not shifted laterally past one another, those that are 

 here shown at 4325, 4307, 4271 4063, 4045, suffice. In 

 the original collodion negative many more can be seen. 

 Below the lower spectrum are the symbols for oxygen, 

 nitrogen, iron, and aluminium. 



No close observ.ation ir. needed to demonstrate to even 

 the most casual observer that the oxygen lines are found 

 in the sun as bright lines, while the iron lines have dark 

 representatives. The bright iron line at G (4107), on 

 account of the intentional overlapping of the two spectra, 

 can be seen passing up into the dark absorption Ime in 

 the sun. At the same time the quadruple oxygen line 

 between 4345 and 4350 coincides exactly with the bright 

 group in the solar spectrum above. This oxygen group 

 alone is almost sufficient to prove the presence of oxygen 

 in the sun, for not only does each of the four components 

 have a representative in the solar spectrum, but the relative 

 strength and the general aspect of the lines in each case is 

 similar. I do not think that in comparisons of the spectra of 

 the elements and sun, enough stress has been laid on the 

 general appearance of lines apart from their mere posi- 

 tion ; in photographic representations this pomt is very 

 prominent. The fine double line at 4319, 4317, is plainly 

 represented in the sun. Again there is a remarkable 

 coincidence in the double line at 4190, 41S4. The line at 

 4r33 is very distinctly marked. The strongest oxygen 

 line is the triple one at 4076, 4072, 4069, and here again a 

 fine coincidence is seen, though the air spectrum seems 

 proportionately stronger than the solar. But it must be 

 remembered that the solar spectrum has suffered from 

 the transmission through our atmosphere, and this effect 

 is plainest in the absorption at the ultra-violet and 

 violet regions of the spectrum. From some experiments 

 I made in the summer of 1873 it appeared that this 



local absorption is so great, when a maximum thickness 

 of air intervenes, that the exposure necessary to obtain 

 the ultra-violet spectrum at sunset was two hundred times 

 as long as at mid-day. I was at that time seeking for 

 atmospheric lines above n like those at the red end of 

 the spectrum, but it turned out that the absorptive action 

 at the more refrangible end is a progressive enfeebling, as 

 if a wedge of neutral tinted glass were being drawn 

 lengthwise along the spectrum towards the less refrangible 

 end. 



I shall not attempt at this time to give a complete list 

 of the oxygen lines with their wave-lengths accurately 

 determined, and it will be noticed that some lines in the 

 air spectiuni which have bright analogues in the sun are 

 not marked with the symbol of oxygen. This is because 

 there has not yet been an opportunity to make the neces- 

 sary detailed comparisons. In order to be certain that a 

 line belongs to oxygen, I have compared, under various 

 pressures, the spectra of air, oxygen, nitrogen, carbonic 

 acid, carburetted hydrogen, hydrogen, and cyanogen. 

 Where these gases were in Plucker's tubes a double 

 series of photographs has been needed, one set taken 

 with and the other without Leyden jars. 



As to the spectrum of nitrogen and the existence of this 

 element in the sun there is not yet certainty. Never- 

 theless, even by comparing the diffused nitrogen lines of 

 this particular photograph, in which nitrogen has been 

 sacrificed to get the best effect for oxygen, the character 

 of the evidence appears. The triple band between 4240, 

 4227, if traced upward into the sun, has approximate 

 representatives. Again at 4041 the same thing is seen, 

 the solar bright line being especially marked. In another 

 photograph the heavy line at 3995, which in this picture 

 is opposite an insufficiently exposed part of the solar 

 spectrum, shows a comparison band in the sun. 



The reason I did not use air in an exhausted Pliicker's 

 tube for the production of a photograph to illustrate this 

 paper and thus get both oxygen and nitrogen lines well 

 defined at the same time, was partly because a brighter 

 light can be obtained with the open air spark on account 

 of the stronger current that can be used. This permits 

 the slit to be more closed and of course gives a sharper 

 picture. Besides the open air spark enabled me to employ 

 an iron terminal and thus avoid any error arising from 

 accidental displacement of the reference spectrum. In 

 Plucker's tubes with a Leyden spark the nitrogen lines 

 are as plain as those of oxygen here. As far as I have 

 seen, oxygen does not exhibit the change in the character 

 of its lines that is so remarkable in hydrogen under 

 the influence of pressure as shown by Frankland and 

 Lockyer. 



The bright lines of oxygen in the spectrum of the solar 

 disc have not been hitherto perceived, probably fiom the 

 fact that in eye observation bright lines on a less bright 

 background do not make the impression on the mind that 

 dark lines do. When attention is called to their presence 

 they are readily enough seen, even without the aid of a 

 relerence spectrum. The photograph, however, brings 

 them into a greater prominence. From purely theoretical 

 considerations derived from terrestrial chemistry and the 

 nebular hypothesis, the presence of oxygen in the sun 

 might have been strongly suspected, for this element is 

 currently stated to form eight-ninths of the water of the 

 globe, one-third of the crust of the earth, and one-fifth of 

 the air, and should therefore probably be a large con- 

 stituent of every member of the solar system. On the 

 other hand the discovery of oxygen and probably other 

 non-metals in the sun gives increased strength to the 

 nebular hypothesis, because to many persons the absence 

 of this important group has presented a considerable 

 difficulty. 



At first sight it seems rather difficult to believe that an 

 ignited gas in the solar envelope should not be indicated 

 by dark lines in the solar spectrum, and should appear 



