V. " Notes on the Atmospheric Lines of the Solar Spectrum , and on certain Spectra of Gases .
" By Dr. JOHtN HALL GLADSTONE , F.R.S. Received May 30 , 1861 .
In the paper of Sir David Brewster and myself on the lines of the solar spectrum * , attention was drawn to the following among other phenomena:1st .
" When the sun descends towards the horizon and shines through a rapidly increasing depth of air , certain lines which before were little if at all visible , became black and well defined , and dark bands appear even in what were formerly the most luminous parts of spectrum .
" These we termed '"atmospheric lines .
" We did not wish to express by that term anything beyond the fact above mentioned ; yet we threw out the idea that these lines may have their origin " in the air that encircles our globe .
" ' 2nd .
In the case of those artificial flames whose spectra " consist of a series of luminous bands separated by dark spaces ... these luminous bands sometimes coincide with the dark lines of the solar spectrum .
" About the same time Kirchhofft published his theory that this remarkable coincidence is due to the presence in the atmosphere of the sun of the substances which emit these luminous bands , and their appearance as dark lines is because " the spectrum of an incandescent gas becomes reversed , when a source of light of sufficient intensity , giving a continuous spectrum , is placed behind the luminous gas .
" It also appears that ' sodium vapour at a temperature much below that at which it becomes luminous , exerts its absorptive power at exactly the same point of the spectrum as it does at the highest temperatures which we can produce .
" It is , however , as yet an undetermined question at what temperature below that at which it becomes luminous , a gas ceases to absorb ; though , from the close connexion which has been shown to exist between emission and absorption , there can be little doubt that such a limit would be arrived at .
This theory of the origin of Fraunhofer 's lines , and the uncertainty just alluded to , suggested the desirableness of comparing the " atmospheric lines " of absorption with the luminous balnds which make their appearance in the linear spectra of the light emitted by the different constituents of the atmosphere when sufficiently heated .
If all the rays emitted by these different constituents should be found to coincide with the " atmospheric lines , " it would show at once the origin of the lines , and prove that gases at the ordinary temperatire will absorb rays of the same refrangibility as those they emit when themselves glowing .
If , on the contrary , they should be found not to coincide , it would prove by an extreme case ( for we operate thus upon many miles of oxygen and nitrogen gases at least ) that this connexion between absorption and emission either does not extend to these elements , or is confined within those narrower limits of temperature which theory seems to require .
It was also conceivable that the atmospheric lines might coincide with the rays emitted , not by all , but by a portion of the constituents of the atmosphere .
The following data exist for this comparison .
Angstrom * has delineated the lumiinous bands due to the gas when the electric spark is sent through nitrogen , oxygen , carbonic acid , hydrogen , &amp;c. Pliicker t has described and measured those bands which appear in the spectra of the same and other gases when they serve as the residuary gas in Geissler 's tubes .
Under the influence of the electric discharge , the vapour of water , carbonic acid , ammonia , and other compounds are decomposed ; but it occurred to me that the oxyhydrogen flame must consist partly , if not entirely , of glowing steam , since it is only when the two elements combine that the heat is evolved ; and that similarly the flame of carbonic oxide burning in air must contain the light emitted from glowing carbonic acid .
I therefore subjected these two flames to prismatic examination .
The result of the comparison has been that not one of these observations shows any accordance between the luminous bands due to the gas , and the dark lines that make their appearance in the solar spectrum when the sun is shining through a great depth of air .
Hydrogen alone is inconclusive .
Neither is there any accordance between these luminous bands and the more prominent lines of the ordinary solar spectrum .
This shows that oxygen and nitrogen , and perhaps other gases , though in enormous quantity , do not absorb at the ordinary temperature rays of the same refrangibility as those they emit when heated by the means specified .
It would not be legitimate to infer from this that the atmospheric lines have not their origin in the absorbent power of one or more of the common constituents of the atmosphere .
It is well known that some gases when placed before a continuous spectrum produce lines of absorption ; among these are bromine and iodine vapour ; yet the dark lines caused by these two halogens * do not coincide with the bright lines into which Plucker found the light of Geissler 's tubes containing bromine and iodine to be resolved by the prism ; nor have I succeeded in reversing them by bringing these substances into a very hot but little luminous flame .
Bright lines were discerned , but in other positions .
In connexion with this subject , it may be worth noting that a prismatic examination of the sun 's rays passing through three inches of mercury vapour at above 300 ?
C. , did not afford the least indication of the reversal of the bright rays that appear when that metal is rendered incandescent .
From the fact that the atmospheric lines do not always present the same appearance when the sun is on the horizon , and that the band 6 has been observed during a shower , and the most prominent lines during a fog , it has been sometimes supposed that the aqueous vapour in the atmosphere is the cause of them .
Yet this can scarcely be .
They seem not to be due to little vesicles of condensed vapour ; for the sun 's rays when passing through the edges of a cloud do not exhibit them , unless , of course , near the horizon .
And they seem not to be due to gaseous water ; for they appear near sunset when the aqueous vapour in the atmosphere is reduced to a minimum by frosty weather , though they are not seen when the sun is higher up in the heavens on a damp warm day .
From Sir David Brewster 's notes , it appears that February 10 , 11,12 , and 13 of the year 1838 were frosty days ; yet the lines were well seen : , , , , C 6 , C 15 , C 16 , and D are specially mentioned .
On the 13th , when the thermometer stood at 23 ?
F. , 4 is said to have been the most prominent .
In the paper referred to at the commencement , we , in common with most of those who worked on the subject before the appearance of Kirchhoff and Bunsen 's paper * , ascribed a bright line coincident with D to other spectra than that of soda .
This was no doubt owing to the almost universal trace of that subtance .
Theelectric lights produced between charcoal points by Professor IIolmes 's magneto-apparatus , and by a galvanic battery with M. Serin 's lamp , were found to be identical , when subjected to prismatic analysis .
Each exhibited a continuous spectrum , and not those variations of bright and dark which other observers , as well as myself , have noticed in electric lights from charcoal points of but inferior quality .
The spectrum extended at both ends beyond that of ordinary diiect sunlight , and the only lines which I could discern were bright ones in the violet or lavender region .
The following refractive indices were determined for the magneto-electric light ; and they are compared with the refractive indices of the nearest principal dark lines of the solar spectrum as determined with the same prism .
The relative position of these bright rays and the dark lines in the lavender part of the spectrum , which are only visible in the sun 's light under the most favourable circuamstances , must not be relied on as accurate , since it was determined by measurement , and not by direct comparison .
Possible inaccuracy of adjustment will render the Pogg .
Ann. , cx .
p. 161 , fourth place of decimals quite uncertain .
The remarkable extension of the spectrum of this electric light , both at the red and violet ends , is , however , indisputable .
Electric light .
Sunlight .
Limit of red ... ... ... ... . .
1'6025 ( about ) Line X ... ... ... ... 1-6038 - .
, , A ... ... ... ... 1-6069 ~~ ~ _~^ G , , G ... ... ... 1-6404 Bundle of violet rays ... 1'6436Faint violet ray ... ... ... 1'6516 , , K ... ... ... 1-6513 2nd , , , ... ... . .
16531 ( about ) Bright violet ray ... ... ... 16550 , , I ... ... .
1.6548 Faint ray ... ... ... ... ... ... 16562 ( about ) - , , ... ... .
, ... ... . .
1-6574 ( about ) , , L ... ... . .
1-6567 Limit of violet ... ... ... ... 1-6664 ( about ) N st ) ... ... 1-6642 Limit 'f v:o e1 --666`(about )

