228 



NATURE 



{Jan. 9, 1879 



the sun which is strongly reversed, so far as our present 

 knowledge extends. The various forms of lithium, 

 therefore, may be shown in the following manner. 



^ 



FEEBLE SPARK 



Fig. 5. — The Molecular Groupings of Lithium. 



It is remarkable that in the case of this body which at 

 relatively low temperature goes through its changes, its 

 compounds are broken up at the temperature of the 

 Bunsen burner. The spectrum, e.g. of the chloride, so 

 far as I know, has never been seen. 



Hydfooen 



All the phenomena of variability and inversion in the 

 order of intensity presented to us in the case of calcium 

 can be paralleled by reference to the knowledge already 

 acquired regarding the spectrum of hydrogen. 



Dr. Frankland and myself were working together on the 

 subject in 1869. In that year {Proc, No. 112) we pointed 

 out that the behaviour of the h line was hors ligne, and 

 that the whole spectrum could be reduced to one line, F. 



*' i.oThe Fraunhofer line on the solar spectrum, named 

 Ji by Angstrom, which is due to the absorption of hydro- 

 gen, is not visible in the tubes we employ with low battery 

 and Leyden-jar power ; it may be looked upon, therefore, 

 as an indication of relatively high temperature. As the 

 line in question has been reversed by one of us in the 

 spectrum of the chromosphere, it follows that the chromo- 

 sphere, when cool enough to absorb, is still of a relatively 

 high temperature. 



" 2. Under certain conditions of temperature and pres- 

 sure, the very complicated spectrum of hydrogen is 

 reduced in our instrument to one lijie in the green corre- 

 sponding to F in the solar spectrum." 



As in the case of calcium also, solar observation affords 

 OS most precious knowledge. The h line was missing 

 from the protuberances in 1875, as will be shown from the 

 accompanying extract from the Report of the Eclipse 

 Expedition of that year : — 



" During the first part of the eclipse two strong pro- 

 tuberances close together are noticed; on the limb to- 

 wards the end these are partially covered, while a series 

 of protuberances came out at the other edge. The 

 strongest of these protuberances are repeated three times, 

 an effect of course of the prism, and we shall have to 

 decide if possible the wave-lengths corresponding to the 

 images. We expect a priori to find the hydrogen lines 

 represented. We know three photographic hydrogen 

 lines : F, a line near G, and h. F is just at the limit of 

 the photographic part of the spectrum, and we find indeed 

 images of protuberances towards the less refrangible part 

 at the limit of photographic effect. For, as we shall 

 show, a continuous spectrum in the lower parts of the 

 corona has been recorded, and the extent of this con- 

 tinuous spectrum gives us an idea of the part of the 

 spectrum in which each protuberance line is placed. We 

 are justified in assuming, therefore, as a preliminary 

 hypothesis, that the least refrangible line in the pro- 

 tuberance shown on the photograph is due to F, and we 

 shall find support of this view in the other lines. In 

 order to determine the position of the next fine the dis- 

 persive power of the prism was investigated. The prism 

 was placed on a goniometer table in minimum deviation 

 for F, and the angular distance between F and the hydro- 

 gen line near G, i.e., Hy, was found, as a mean of several 

 measurements, to be 3'. The goniometer was graduated to 

 15", and owing to the small dispersive power, and therefore 



relatively great breadth of the slit, the measurement can 

 only be regarded as a first approximation. Turning now 

 again to our photographs, and calculating the angular 

 distance between the first and second ring of protuber- 

 ances, we find that distance to be 3' 15". We conclude, 

 therefore, that this second ring is due to hydrogen. We, 

 therefore, naturally looked for the third photographic 

 hydrogen line, which is generally called It, but we found no 

 protuberance on our photographs corresponding to that 

 wave-length. Although this line is always weaker than 

 Hy, its absence on the photograph is rather surprising, 

 if it be not due to the fact that the line is one which only 

 comes out at a high temperature. This is rendered likely 

 by the researches of Frankland and Lockyer {Proc. 

 Roy. Soc, vol. xvii. p. 453). 



" We now turn to the last and strongest series of pro- 

 tuberances shown on our photographs. The distance 

 between this series and the one we have found reason 

 for identifying with Hy is very httle greater than that 

 between H/3 and Hy. Assuming the distances equal, we 

 conclude that the squares of the inverse wave-lengths of 

 the three series are in arithmetical progression. This is 

 true as a first approximation. We then calculated the 

 wave-length of this unknown line, and found it to be 

 approximately somewhat smaller than 3,957 tenth-metres. 

 No great reliance can be placed, of course, on the number, 

 but it appears that the line must be close to the end of 

 the visible spectrum. 



" In order to decide if possible what this line is due to, 

 we endeavoured to find out both by photography and 

 fluorescence whether hydrogen possesses a line in that 

 part of the spectruta. We have not at present come to 

 any definite conclusion. In vacuum tubes prepared by 

 Geissler containing hydrogen, a strong line more re- 

 frangible than H is seen, but these same tubes show 

 between Hy and H5, other lines known not to belong to 

 hydrogen, and the origin of the ultra-violet line is there- 

 fore difficult to make out. We have taken the spark in 

 hydrogen at atmospheric pressures, as impurities are 

 easier to eliminate, but a continuous spectrum extends 

 over the violet and part of the ultra-violet, and prevents 

 any observation as to lines. We are going on with 

 experiments to settle this point. 



" Should it turn out that the line is not due to hydrogen, 

 the question will arise what substance it is due to. It is 

 a remarkable fact that the calculated wave-length comes 

 very close to H. Young has found that these calcium 

 lines are always reversed in the penumbra and imme- 

 diate neighbourhood of every important sun-spot, and 

 calcium must therefore go up high into the chromo- 

 sphere. We draw attention to this coincidence, but our 

 photographs do not allow us to draw any certain con- 

 clusions. 



"At any rate, it seems made out by our photographs 

 that the photographic light of the protuberances is in 

 great part due to an ultra-violet line which does not 

 certainly belong to hydrogen. The protuberances as 

 photographed by this ultra-violet ray seem to go up 

 higher than the hydrogen protuberances, but this may be 

 due to the relative greater length of the line." 



In my remarks upon calcium I have already referred to 

 the fact that the line which our observation led us to believe 

 was due to calcium in 1875, was traced to that element in 

 this year's eclipse. The observations also show the curious 

 connection that, at the time when the hydrogen lines were 

 most brilliant in the corona, the calcium lines were not 

 detected ; next, when the hydrogen lines, being still 

 brilliant, the h line was not present (a condition of things 

 which, in all probability, indicated a reduction of tem- 

 perature), calcium began to make itself unmistakably 

 visible ; and finally, when the hydrogen lines are absent, 

 H and K become striking objects in the spectrum of the 

 corona. 



To come back to 7i, then, I have shown that Dr. 



