On Spectra of Electric Light , as modified by the Nature of the Electrodes and the Media of Discharge .
By the Rev. T. R. ROBINSON , D.D. , F.B.S. , &amp;c. Received June 19* , -Read June 19 , 1862 .
THE important discovery of WHEATSTONE , that the spectra of electric sparks contain brilliant lines whose character depends on the nature of the electrodes , after being almost neglected for several years , has lately become an object of great interest ; and much additional light has been thrown on it by several physicists , of whom MASSON , ANGSTROM , 0 PLUCKER , and KIRCHHOFF are the most conspicuous .
ANGSTROM has announced as a general law that the lines in question are produced by the electric current igniting the medium in which the discharge takes place , or molecules of the electrodes which are torn off by its passage , that each of these actions produces its own spectrum , and that those spectra are simply superposed without any modification .
The gases with which he worked were at the ordinary pressure .
PLUCKER , on the other hand , used the wellknown Geissler tubes , which contain minute proportions of highly rarefied gas or vapour .
He attaches very little importance to the lines due to the electrodes ( metallic lines ) , which he thinks are confined to those portions of the spark near the electrodes ; and he maintains that in the centre of an exhausted tube of sone length only gaseous lines are seen .
He was embarrassed in several instances by the decomposition or absorption of the gaseous media ; and there must always be some doubt as to the precise nature of these media , as the tubes are hermetically sealed .
On the other hand , KIRCHHOFF seems to attach most importance to the metallic lines , whose influence he has exhibited to a wonderful extent by a spectrum-apparatus probably unrivalled .
All hold the doctrine of an essential connexion between the character of the spectral lines and the chemical nature of the substances which are present in the track of the discharge ; and the last , in conjunction with BUNSEN , has based on this principle the new system of spectral analysis , which is rapidly becoming popular , and has applied it to explain the dark lines of the solar spectrum in a way which , if not absolutely certain , is singularly elegant .
Yet it is impossible to overlook the fact that , in all this , much is assumed , not proved .
Has it been established that these lines depend so absolutely on chemical character that none of them can be common to two or more different bodies 1 Has it been ascertained that , while the chemical nature of the bodies present remains unchanged , the lines never vary if the circumstances of mass , density , &amp;c. are changed And what evidence have we that spectra are superposed , so that we observe the full sum of the spectra which the electrodes and medium would produce separately .
Lastly , is it certain that electricity produces light merely by its heating power ?
and may not the same action which produces thermic vibrations , also ( and independently ) produce luminous !
Some of these questions I have endeavoured ; to investigate and the attempt cah scarcely fail to be of use , wereit only to direct attentionot tohe subject.:CMy attention was originally turned to it by observing , some years ago , that the discharge in carbonic oxide , which is white at common pressures , becomes bright green when the gas is highly rarefied .
The spectra in these cases differed so much that I determined to examine them in various gases and metals ; and I procured the apparatus which seemed necessary .
That of STEINHEIL had not been contrived then ; and my want of experience in such researches caused at first some mistakes , but by degrees I corrected them ; and though my instrument is of comparatively limited power , the results are , , I hope , not without importance ; and I have given details sufficient to estimate the errors which may affect them .
The source of electricity in my experiments was an induction-machine .
Till July 1860 I used one by HEARDER , having three miles of secondary wire ; after that , one which I have described* , with the substitution of copper wire for iron .
It has 9'5 miles , of which three are not lapped , merely varnished ; and when excited by three Grove 's , each of which has 49 square inches of platinum acting , it gives abundantly sparks 6'8 inches long .
By a simple device its two coils can be in an instant made to act collaterally instead of consecutively , thus reducing the intensity but doubling the quantity .
Then the sparks are only 2-5 inches , but very dense and luminous , and possessing a far higher power of ignition .
A Leyden jar , each coating 1'25 foot , was normally connected with the terminals , as the spectrum of the simple spark is far more faint and unsteady : I have given one as a specimen in Table III .
With this jar there is a continuous stream of discharge at 0'6 inch .
The discharges were made in the open air , or in tubes about 0'2 inch diameter and 6 inches long .
With rarefied gases they would be more luminous if the tubes were capillary ; but in that case it would be very difficult to clean them from the coatings of metal or oxide which are deposited , most densely near the negative electrode , but often on the whole interval .
In some metals , of which the most notable are lead , cadmium , bismuth , antimony , arsenic , and above all tellurium , this deposit was so thick that at the close of an experiment it was difficult to see the fainter lines , and I found it necessary to employ for common density a tube 1 inch diameter .
Even this was coated , but not so thickly as to give much trouble .
The upper electrode ( which was , in all cases but one , positive ) was attached to platinum wire fused in the tube .
It was sometimes soldered to it , at other times twisted with it ; and when I could not procure the metal as wire or foil , a globule of it was fused by a gas-blowpipe , a platinum wire inserted in this , and the heat withdrawn .
For nickel and cobalt , the blowpipe was oxyhydrogen .
As the electric light does not spread over the positive electrode , it was unnecessary to 'insulate the platinum wire ; but with the negative it is otherwise ; and when I had not enough of Ry .
T. ; 3 .
EROBI : PSON ON S::'PECTRA QOF BLEIOTWEI .
DiiHT .
the body to give alength of some inches it ' was ; inserted in a piece of quill tube , wh-ose lo^er extremity was fused on platinum wire reaching up to it ; and this was inserted i the idischarge-tube .
: ' .
In tubes the distance of the electrodes was about 0-75 , in open air 0-100 .
The discharge-tube was cemented in a cap screwed to the air-pump , or rather to a transfer piece.:This is a cylinder of brass with a concave screw above , a stopcock below iscrewing to the air-pump , and a lateral one connected with a desiccator .
The desiccator isa bottle 8 inches deep and 2 inches diameter , fitted with a ground stopper , in which are:two apertures .
In one of these is cemented a tube which is connected with the transfer ; in the other , one descending to the bottom of the bottle , where it is drawn into a ; capillary point , its upper end being connected with either another desiccator or a gasometer .
-The desiccator is filled nearly with sulphuric acid .
: Supposing the transfer and the tube exhausted , shut off the air-pump cock and cautiously open that which connects the desiccator ; gas bubbles up slowly through the acid and fills the transfer and tube .
Shut off the desiccator , connect the pump , and exhaust .
By repeating this process ad libitum all traces of air or any gas previously used are removed , and nothing is present but the subject of experiment .
It is almost needless to say that every part of this apparatus , including the air-pump , must be absolutely* air-tight .
To ascertain whether this means of desiccation is sufficient , I used Kater 's oat-beard hygrometer , In the air of the room it read at 60 ? ?
3-358 R. It was then placed under a receiver containing a capsule exposing 20 square inches of sulphuric acid for five hours at a pressure of 0'3 inch , when it read 0'455 R. The receiver was then filled with air drawn through the desiccator as rapidly as could be done without drawing acid into the pump .
In 101 it read 0*285 R. Admitting air and introducing a slip of bibulous paper moistened with water , in 40m it was 8'480 R ; and when left in the air of the room some hours it was again 3'258 .
It is therefore evident that in this respect nothing more can be desired .
Sulphuric acid also absorbs sulphurous acid and nitric oxide ; but when other acids might be present , a second desiccator was used filled with concentrated solution of potassa .
The gasometer was made of a Woulfe 's bottle holding 75 cubic inches .
In its central neck is cemented a siphon , one , branch of which reaches to its bottom , the other to the bottom of a bottle of rather larger capacity ; in the other two necks , stopcocks are cemented , one of which ( A ) is connected by elastic tube with the desiccator , the other ( B ) with any apparatus for generating gas .
Suppose the Wolf filled with water ( or on a smaller scale with mercury ) .
Let gas be supplied to B , the water will be displaced by it through the siphon into the bottle ; then closing B and opening A , the transfer and its tube are filled .
If the water has been boiled for a few hours and cooled without exposure to the air , I find that gas continues in this gasometer sensibly pure for a much longer period than the duration of a day 's observations : For operating on vapours free from any mixture of permanent gas , I use a mercurial apparatus , consisting of a strong Woulfe 's bottle in whose three necks are ground ( 1 ) a glass stopcock , ( 2 ) a tube in which 'a platinum wire is sealed long enough to dip in the mercury and serve as the negative6 electrode , ( 3 ) a tube also dipping in the mercury , and with a platinum electrode at top .
If the bottle be filled one-third with pure mercury , and the stopcock be connected by a caoutchouc tube* with the transfer and the apparatus exhausted , the tube can be filled with mercury , which on erecting it falls , leaving a vacuum , through which discharges can be passed by the electrodes .
If the tube has beenperfectly cleaned by filling it with nitric acid and washing it with distilled water , dried by sulphuric acid in a vacuum , and finally wiped with a morsel of linen which has been boiled in distilled water , after being fastened to a flexible wire , on inclining the tube the mercury will fill it without leaving the least speck of air , and will often adhere with considerable force .
It , however , always falls at the first discharge .
In this case the space is filled with mercurial vapour alone .
If a few drops of any volatile fluid be introduced into the tube , by filling it with mercury the excess is expelled , and the vacuum contains only its vapour highly rarefied .
With phosphorus , the tube filled with water was warmed till it fused and adhered to the upper part ; the water was removed , and the tube put in its place .
The apparatus was then repeatedly filled with dry nitrogen till all traces of moisture disappeared .
If a platinum wire of sufficient length be introduced below , both electrodes are platinum ; and if the upper part be bent so that the descending branch remains full of mercury , both are of that metal .
Prisms.-For three-fourths of the observations I used a prism by MERZ , with an angle 45 ? ?
35'A4 .
By sets of from eight to twelve it gives , for the deviations of FRAUNItOFER'S lines , 0 A ... 32 20-00 ... 1-6230 B ... 32 32-04 ... .
1-6259 C ... 32 38-88 ... .
1-6285 D ... 32 56-50 ... .
1-6336 E ... 33 20-26 ... 16.405 F ... 33 41-67 ... .
1'6467 G ... 34 2306 ... .
16587 H ... 35 0-06 ... .
16692 It is therefore nearly identical with FRAUNHOFER'S flint No. 2 , but in dispersive power it is far inferior to those used by MASSON and PLUCKER .
After working with it a long time , I found that several bright but cloudy bands which it showed were resolved into two , three , or more by a prism of bisulphuret of carbon having an angle of 60 ? .
The indices of this fluid change so much with temperature , that I did not venture to employ it ; and I obtained from Mr. DUBOSCQ a prism of 60 ? ?
3'-54 angle , which , though nearly of the same density as the Merz , is more effective , in the proportion of 3 to 2 .
Determining with it FRAUNHOPER'S lines , and comparing with their deviations in Merz , Ix This , with which I was supplied by Messrs. SILvER , is far superior to vulcanized tube , which always leaks .
The glass apparatus was made with great precision and intelligence by Mr. CAsELeA .
942 formed by interpolation a Table of reduction to the latter , by means of which all are given on the same scale .
The two glasses are so nearly similar , that it was not necessary to go beyond second differences .
Even this prism leaves several bands unresolved , though often giving a suspicion of their compound character , which in some recent instances I have verified by combining two 60 ?
-prisms of the bisulphuret .
It is an object of great interest to ascertain whether there be any special relations between the wave-lengths of these luminous bands .
In aid of this I subjoin a Table giving the value of x for every five minutes of MERZ'S deviations within the range of my observations .
It has been computed by a very simple form of interpolation given by BI Professor STOKES* .
Assuming p-=A+-B , and taking b and for any two of FRAUNHOPER 's lines , we get , for any intermediate io , the simply by proportional parts .
This is so accurate , that it gives correctly one of the intermediate lines by taking the double interval , as D from C and E ; even H from F and G. p -X .
aX .
a2X ... X. AX .
A 2X. .
X. AX .
A2 .
32 35 2493 33 25 1909 +2 34 15 1620 -1 +1 40 2405 +9 30 1872 3+ 20 1599 -19 +2 45 2326 +10 35 1837 +3 25 1580 _1 --69 --32 -18 50 2257 63 +6 40 1805 -3 1 30 1562 -18 + 55 2194 5+5 45 1774 +2 35 1544 8+0 33 0 2136 5+5 50 1745 +21 40 1526 +1 --53 27 15 -17 5 2083 49 +4 55 1718 + 45 1509 + 55 1118 --16 +1 10 2034 +4 34 0 1692 +1 50 1493 1+1 -5 25 --15 15 1989 +45 1667 24 + 55 1478 14 +1 20 1948 +2 10 1643 -+ 1 35 0 1464 25 1909 39 215 1620 25 1909 +2 15 1620 +1 , ... .K Theodolite.-For the use of this instrument I am indebted to Mr. GRUBB , who made it many years since to determine the p0s of the glasses for his object-glasses .
It is of simple and very firm construction .
A strong brass disk , supported by three screws , has on its upper surface a circle 95 inches diameter graduated to half degrees , and carries laterally the supports of a collimating telescope 9-5 inches focus and 1 inch aperture , which is provided with an adjustable slit .
Above the disk turns a brass plate bearing two verniers in the plane of the divisions ( which read to minutes ) , and supporting the telescope , with a triple object-glass 7 inches focus and 0'9 inch aperture .
Below the disk turns another circle , similarly divided on its cylindric surface ; its axis rises through that of the upper plate , and carries a table 3 inches diameter , which bears the prism .
The axis of the prism is adjusted by observing the images of the slit reflected from its surfaces , which also give its angle by means of the lower circle .
Some of these matters require a few remarks .
1 .
The telescope ( and the collimator also ) , though sensibly achromatic on a day-object or the moon ( and very sharp ) , is over-corrected for G and the rays beyond it * as no provision was made for changing the distance of its object-glass from the system of wires , I at first : had some difficulty from parallax at that end of the spectrum , till such an adjustment was applied .
The eyepiece ( positive , magnifying nine times ) was also not achromatic , and had to be constantly focused .
It was replaced by a microscope 's objective about 0,4 inch focus , magnifying 14 with the prism .
A higher power , and even a larger object-glass , avails little in comparison of an increase of the prism 's dispersion .
Thirdly , I found the cross of spider 's lines useless , except for the brightest lines .
In the spectra of rarefied gases , which are very faint , it is difficult to see them , and I substituted the point of a fine needle , carefully ground to be a sharp wedge .
One can estimate very nicely the equality of the tongues of light on each side .
These changes were not made till about fifty spectra had been measured , which , therefore , are not as well determined as the rest .
2 .
The collimator certainly possesses advantages over the simple slit which FRAUN : HOFER and his predecessors used in studying the spectrum .
It requires no correction for parallax , secures from any accidental shift of the theodolite , and brings the observer close to his work ; but it has the great defect of diminishing the light .
Unless the slit be very narrow , it is impossible to distinguish close and fine lines .
I find that with the instrument which I am describing , and the Merz prism , I cannot see D double if the slit subtend more than 72 " , which corresponds to a width of --inch .
If it be 3 ' ( the opening used by PLtC : KER ) I cannot see any of FRAUNHOFER'S lines , and the finer parts of electric spectra are lost .
Obviously the quantity of light must diminish with the slit , and the evil is made greater by the necessity of keeping the latter at some distance from the discharge .
If it be-neater than 1-5 inch , or at most 1 inch , the inductive action of the spark charges the theodolite , and the observer , on applying his eye to the telescope , gets a stream of pungent sparks anything but pleasant .
But in order that the whole object-glass may be illuminated , we must have this distance less than A , t 7 ' AA cosecant 72"f and f being the aperture and focus .
This limit in the present instance =0 03 inch , from which it is obvious that a comparatively small portion of the light can reach the objectglass .
I was led to this discussion by a fact which at first startled me a little .
When obtaining the deviations of FRAUNHOFER'S lines , I was surprised to find that H was not visible to me , nor any line beyond h ; though in 1838 , with the same prism , but with a slit in the shutter 15 feet distant , I saw several beyond K. I concluded that my eyes had become insensible to rays of short wave-length ( in analogy to WOLLASTON'S inaudible sounds ) , or that their humours had undergone some change by which they absorbed that part of the spectrum ; but never suspected the collimator .
However , last year , while examining a very fine Munich grating belonging to Mr. STOKES , I was surprised to find that it showed me the missing lines perfectly : here the aperture was a slit 31 at 18 feet .
ie was so kind as to entrust me with the grating ; and on my return home I found that with the collimator it behaved no better than the prisms had done , but that with 4-41 the plain slit it recovered its power : the same was the case with my prisms .
Thatfthb j6cns , ianid placing it at that distance from the slit ( --j ) .
This arrangement gave vision that was first-rate , and showed lines in an electric spectrum 14 ' beyond the last of those -that were visible in the usual mode of observing .
It 'follows from these facts , that the.collimator 's focus should be as ' long as possible consistent with convenience ; --and I will suggest that it should be a cassegrain instead of an achromatic .
The equivalent focus 'of this is from six to seven times that of the large mirror , and I find by a rough trial that the image is very sharp .
If the mirrors of A were glass silvered on FOUCAULT'S lamn , it would have as much light as the achromatic , and be free from all chromatic error .
I have recently found that the brightness of faint lines is much improved by 'usinina ndric lens before the slit .
The slit was generally one minute wide .
3 .
The precision of the angles measured depends , on the determination of the indexcorrection , on the precision of the bisection , and on the reading of the circle .
The zero was obtained at the beginning , and also often at the end of each set , by bisecting the posing a slip of red glass ; latterly , in preference , by using the flame of a Bunsen burner in which chloride of sodium was present .
The bisection is sometimes doubtful from the flicker of the discharges , more frequently from faintness of the lines , whether intrinsic or relative to the ground on which they are seen , and occasionally from Want of sufficient light to see the point .
The verniers.read only to minutes ; but the half minute is easily estimated , and so 0 ' 25 may be considered the uncertainty of reading .
It is desirable to form some estimate of the probable amount 'of error due to the combined action of the three .
This is done by observing twice over the same spectra , and comparing the differences of the observed lines ; from which may be obtained the probable error , and the probabilities of given observed differences being errors of observation , or evidences of the lines not being identical .
Unless the graduation were much finer than it is , this process would not be of any real value ; and a much simpler one may serve .
In this work 325 such differences were observed , of which number there were , Comp. Equal .
to 0 ' ... .
86 ... ... .
86. .
239 From 0 ' to + ?
0'5 ... 1.31 Not exceeding 0'-5 .
217. .
108 From +0'"5 to + 0. .
89 Not exceeding 10. .
306. .
19 From +l'0 to +'-5. .
15 Not above 1'5 .
321. .
4 Above 1'"5 ... 4 We may reason thus : if the difference of the places of a certain line in two spectra exceed 0 ' 5 , either the lines are identical and the difference is error , or they are distinct ; 108 217 the probability of the first happening =25 and therefore that of the other =325 or it is 2 to 1 they are not the same .
If the difference exceed 1'"0 , it is 16 to 1 , if 1'5 , : 80 to 1 .
Two things , however , must be remembered : these errors may be + ; and therefore , in comparing a series , the limit is nearly doubled ; and further , these probabilities may be much modified by other circumstances .
For instance , a line or band may be identified by some peculiarity , even if the difference be greater than the probable limit .
In general I would fix the limit to +-1'0 from the mean .
I at first read both verniers ; but found this consumed so much time , that I determined the excentricity of the upper circle and allowed for it .
This correction for the vernier A =0'-75 X sin ( 160 ?
+p- ) .
The gases which I selected for experiment are-i1 , air ; 2 , nitrogen obtained by heating nitrite of potassa with saturated solution of sal-ammoniac ; 3 , oxygen ; 4 , hydrogen : these last two were obtained by electrolysis of pure oil of vitriol diluted with eight volumes of distilled water , in a voltameter of peculiar construction .
A porous cell has a cover cemented on it with three tubulures ; one for admitting the dilute acid ; one for a strong platinum wire to which is soldered with gold a platinum sheet , exposing 19 square inches ; the third carries away the evolved gas .
Round this cell a larger platinum is rolled , and it is immersed in a jar filled with the same dilute acid .
When connected with the three Grove 's already mentioned , it gives 8 inches of hydrogen per minute .
Both the gases I believe to be quite pure , except as to ozone , when thus obtained .
5 .
The carbonic oxide was got by heating sulphuric acid with ferrocyanide of potassium .
As to the metals , the platinum and silver were obtained from Messrs. JOHNSON and MATTHEY , the aluminium from Paris .
I am indebted to the kindness of Dr. MATTHIESSEN for calcium , tellurium , and gold .
The palladium was given to me by Dr. WOLLASTON ; tin , lead , and bismuth reduced from oxides carefully prepared ; zinc , iron , and antimony , deposited by electrolysis on platinum wires .
Cadmium , nickel , cobalt , magnesium , sodium , and potassium were got from Messrs. JACKSON and TowNsoN .
The spectra at common pressure ( C.P. ) are in general magnificent objects .
Their ground seems to be a continuous spectrum , of which , however , the brightness varies very much with different substances and at different parts .
Sometimes , especially at the violet end , this ground is so faint that its presence might be questioned ; but I believe it exists even there .
On this are , as it were , superposed luminous lines of every degree of brightness , from a splendour almost insupportable , to a faintness such that ( at least with my optical means ) the least glimmer of diffused light in the telescope totally effaces them .
Most of these lines are ( as might be expected ) as broad as the image of the slit : a few are much broader , even to six or seven times .
Such are , I think , always cloudy and ill defined , giving the impression that they are groups of finer linles , which the optical power is insufficient to separate .
In several instances this is shown to be the fact by the combination of two fluid prisms of 60 ?
( 2BS .
C ) , of which the two most remarkable are those which I call h , in the green , and x " , at the beginning of the violet .
The first was seen in the Merz prism as a broad bright band , but it is a crowd of very fine lines , of which the central one is much the brightest ; and some of the others are 946 ' developed in different spectra : the second ( not resolved in Duboscq ) consists of six bright and sharp lines .
Even this power fails to decompose the remarkable blue-green and violet bands which characterize the hydrogen spectra C.P. Perhaps a third prism might , but there is no room for it on the theodolite .
A third class of lines is narrower than the slit , down to the finest hairbreadth ; they are mostly sharp and well defined , sometimes very bright .
Their occurrence is not easily explained ; for in the ordinary conditions of these observations each ray illuminates the whole slit , and the image of the line due to it ought to be as broad as that slit .
The only explanation which has yet been proposed ( by PLUCKER ) is that they are caused by .
the overlapping of two images the distance of whose centres is less than the slit .
It is , however , liable to two objections-that the brightness of such an overlap cannot exceed twice that of the ground on which it is seen , and that it would be often resolved by a prism of higher dispersion : the combination 2BS .
C disperses seven times as much as the Merz , and ought surely to break up some of them .
This is not , as far as I have examined , the case ; and some of these lines are as intense as any in the spectrum. .
These narrow lines are sometimes very thickly crowded , as in the green and blue of iron spectra , and in those of carbonic oxide ; and possibly they may compose the bright ground when so close as to be unresolvable .
There is a seeming tendency in them to be grouped in two , three , or higher numbers , which , however , might disappear with more powerful prisms ; but with those I use one can scarcely avoid thinking that there is some special connexion between the components of such groups ; such are enclosed in brackets .
Some of the most brilliant are double , as the orange one y , supposed to correspond with D ( though its components seem to me more separated than those of that line ) , the splendid yellow 6 , and the greens ?
and 8 .
At that boundary of the spectrum which corresponds to th e negave electrode ( and in a much less degree at the positive ) extremely intense lines are seen , especially in the green , which however e are short : bismuth , zinc , lead , and arsenic are the best examples of this .
These are generally supposed to be metallic lines , and to proceed from the intense dispersion of metal near the electrodes , especially the negative .
It has even been proposed to consider this not crossing the entire breadth of the spectrum as a test of metallic origin , and to regard the others as gaseous .
I , however , find that these very lines can be traced entirely across though sometimes very faint , except when the entire part is covered with a sort of bright haze , through or on which nothing can be seen which is not very bright .
It seems to me that the existence of a line and its brightness depend on different causes ; and I shall give instances when the same line assumes very different aspects as the media of discharge are changed .
I may add that the sort of haze just mentioned does not occur at the negative electrode. .
The spark discharge without a jar is so much fainter that none but the brightest lines can be seen .
If the surface of the jar be increased ( for the converse reason ) more lines are visible , but I think no new ones are produced .
One or two examples will be given , If the secondary coils be arranged for quantity , the red end is not changed , the violet is much brighter .
On rarefying the gas in which the discharges are made ( R ) there is at first no change , except perhaps a little diminution of brightness , till at a certain pressure ( varying with the media and diameter of the tube ) the spectrum fades away .
Sometimes , as with CuO ( i. e. copper electrodes in oxygen ) , all vanishes except a trace of the lines at 0'75 inch ; in SbN all but a trace of ' ; in NiN , at 2'8 inches , all but a suspicion of violet light , and sometimes perfectly dark bands take the place of bright lines .
In others the change is not so striking .
In Al , Air , at 1*2 inch , the chief violet lines remain tolerably conspicuous ; ~ can be seen , but a and X vanish : in Pt , Air , tube'l inch diameter , at 6 inches the first half fades to a neutral-tinted haze with faint alternations of brightness , and the rest has eleven definite but faint bands .
In general , however , it may be considered the rule that from 3 inches to 1 inch these spectra almost vanish .
I will only mention another , PbHI .
In hydrogen spectra , at C.P. the most distinctive characters are a very bright red line ( found in all spectra , but not so bright ) ; a very bright and broad one at the confines of green and blue , and a similar one in the violet .
Of these the last diappears at 53 inches ; the red at 311 inches ; the blue not totally till 0'15 inch ; but this is doubtful , as it may have been confounded when waning with another which was near its place .
It deserves notice that this disappearance of the peculiar hydrogen bands is also produced when the hydrogen is diluted with air .
The electrodes were cobalt ; 20 cubic inches of hydrogen were mixed with 10 of air , and the spectrum of the mixture examined ; its volume was reduced to 20 , 10 more of air added , and so on ; the degree of dilution was easily found from the number of additions ; and it was found that the violet band vanished when H is 02 of the mixture , the blue when 0'09 .
These correspond to hydrogen under the pressures 6 inches and 2-7 inches : the red could not be determined , as air has the same band .
If in any case we rarefy beyond the limits of these transition spectra , bright lines reappear , but not all in the same places or with the same characters .
Ordinarily the brilliant lines of the C.P. spectra are wanting , though sometimes lines which are faint in them assume this type in the others .
The red , yellow , and green seem diffused in cloudy light ; and the violet system is replaced by a set of broad cloudy bands nearly equidistant , and more conspicuous than the less refrangible ones .
The red band , which almost always begins the C.P. spectra , is often wanting , and when it does occur is insulated in darkness , but at the other end both spectra are nearly of the same length .
These R spectra are in general much less luminous , and show little distinction of colours unless the metal of the electrodes be easily vaporized .
That of tellurium is very bright , containing thirteen brilliant lines ; and under every circumstance of pressure or discharge , those of potassium and sodium show the dazzling orange bands .
As might be expected from their greater faintness , they contain fewer lines , or at least fewer are visible ; but it is remarkable that of this number a considerable proportion is not found in the C.P. spectra .
The percentage deduced from fifteen metals is 948 REV , T.t. .
ROBINSON ON SPECTRA OF ELECTRIC LIGHT .
Air ... ... ... .
039 Nitrogen ... ... . .
032 Oxygen ... ... . .
039 Hydrogen ... . .
066 Carbonic oxide ... .
0'37 Of this , two explanations may be given .
It may be said that these lines are not seen at C.P. because they are overpowered by the brightness of the ground on which they are seen .
In some cases ( of which examples are given hereafter ) , especially in hydrogen , this cause does act in some degree .
In that gas the centre of the C.P. spectrum was at first observed , where , as I have stated , the finer lines are not easily seen with some metals ; and then the percentage was very high ; but at the negative boundary this bright haze does not interfere , and when observed there , the proportion is almost exactly that of the other gases .
Besides , it occurs frequently that in the C.P. spectrum , on each side of the place where the missing line should be , faint and narrow lines are seen with perfect distinctness .
The other solution is this , that these lines do not depend entirely on the chemical character of the media and electrodes , but also on their molecular condition .
On any other supposition it seems hard to conceive the passage from the C.P. to the transition spectrum , and the increase of brightness from that to the R one , the same chemical elements being present in the three .
In presenting the measures of these spectra , there is a difficulty arising from the impossibility of giving , within any practicable limits , the distinctive characters of the phenomena , though they are of considerable importance .
For instance , the orange band y , which in the sodium spectrum is " intensely bright , " is in Pt , CO " faint , scarcely visible .
" To tabulate them I must restrict myself to a few distinctive symbols .
First , the lines which are far transcendent in brilliancy , and are not less broad than the image of the slit , I denote by a * .
This implies merely that they surpass the others greatly ; one in red or violet may fully deserve this symbol , though it would seem dull beside 6 or d. Not very frequently the same line has this mark in C.P. and R , but , except with sodium and potassium , less bright ; and sometimes one which is faint in C.P. is a except H , C.P. ( which was taken with Duboscq , and at the negative boundary ) were with Merz , but with the improved telescope .
As first observed , and at the centre , this last showed only eight lines .
It will be remarked that all the lines in the C.P. of nitrogen , and all but one in that of oxygen , are found in air , as might be expected ; often a line is common to the three , and then that in air is of intermediate character .
But the same line is also often found in the other two gases .
It is generally supposed that this indicates the metallic origin of that line ; but it will be found that many occur , not only in all these gases , but with all or nearly all the electrodes which I have tried .
I shall return to this at another time , now making a few remarks on those before us .
The pair a , c ' are of almost universal occurrence at the origin of the spectrum ; a has close before it a narrow but bright companion , whose place has sometimes been taken ; but outside of that there are merely shadowy traces scarcely ever bright enough to be bisected .
In a very few cases , however , one of them becomes predominant , e.g. silver and iron .
In H ( hydrogen ) spectra , one of this pair acquires an intense brilliancy , which is peculiar to this gas , while the other fades away .
It is to be noted that a is exactly in the place of the solar line C. These and 3 , ' are in the red .
No. 7 , y is an orange band , nearly but , I think , not exactly in the place of D. When viewed with the combination 2BS .
C , it is double in all cases which I have examined ; but the distance of the centres of its components is fully twice that of the components of D. No. 11 , S is yellow , generally extremely brilliant , except in one ; it is seen in 2BS .
C double , the second one being about half the breadth of the preceding .
Before it is a narrow orange line , and before that No. 10 , which seems to attend it constantly .
Nos. 14 s , the origin of green , 15 s ' and 16 s " are of constant recurrence .
With the higher prism-power the first and third are each double , and some between g ' and i " , of which one ( as here in air , C.P. ) has been occasionally observed .
No. 18 , ' was at first observed as a bright cloudy band ; 2BS .
C shows that the whole of that region is covered with close lines , of which one is more conspicuous than the rest .
In CO many more are conspicuous .
No. 21 , ; is very brilliant ( except in oxygen and CO , when its following companion is the brightest ) .
With 2BS .
C it is close double of two equal* .
This is also the case with No. 24 , 0 , at the boundary between green and blue , which is , as a rule , the most intensely bright in all the C.P. spectra , in many of which its light is almost blinding .
In R it is always faint though present .
It is preceded by two narrow ones of unequal brightness , which seem to form with it a system .
Another system seems to be found from 32 to x " 36 , remarkable for its beauty and peculiar character ; x ' is vivid blue , close double of equals , x " , broad and cloudy , begins the violet .
With 2BS .
C x is composed of two further apart , and c " of six .
It is less developed in nitrogen and H ; x 37'40 was at first observed as a band ; in CO the whole of its vicinity is covered with narrow violet lines .
The pair 28 , 29 , and the triplet p , pi ' , pi " , are also of very constant occurrence .
Some of these are peculiar .
In air , C.P. , No. 12 is the brightest of all , ~ is very brilliant , so is No. 46 .
The H , C.P. , is also greatly developed .
It was taken at the negative electrode only .
Nos. 2 , 29 , and 48 reach quite across the spectrum ; the others more or less , but on an average about one-third of the way from each electrode , being brightest at the negative .
The deposit in the tubes was very considerable , but it does not seem to have cut off much of the spectra .
To examine the effect of the diameter of tube , I took ( with Duboscq ) the air R. in a compound tube , of which the wider parts were 0'4 inch bore , and the narrower 0-04 and 1*5 long .
The distance between the electrodes was 5 inches .
The discharge was pretty bright in the narrow parts , and red , except on the negative wire , where it was blue .
They are as follows : Narrow .
Blue negative .
Red positive .
I I , II 1 .
3 .
3 .
4 .
5 , 6 .
7 .
8 .
9 .
10 .
11 , 12. .
13 .
14 .
15 .
16 .
17 .
18 .
19 .
20 .
21 .
22 .
23 .
24 .
25 .
26 .
27 .
28 .
29 .
30 .
31 , 32 .
33 .
32 32 33 E 33 34 38'1 f. 49-1 vf .
58 ) vf .
many .
0o6 J vf .
5-3 n. 7-3 n. 94 vf .
11 b. others , 14 1 f. some .
16-4 f. 20 f. 22-4 w. 25-7 J w. 28*4 nf .
one .
31-4 nf .
34-4 c. 37-5 c. 40-8 nb. 44-1 w. 50 b. 55-3 b. 0-4 b. 7-5 b. 13-4 nb. 18 c. 193J c. 26-5 b. 33-1 c. 39*6 c. 49'5 vf .
57-8 vf .
58 vf .
6 f. 107 c. 20-7 } c. 23-4 c. 257J c. 29'4 f. 32-1 nc .
35'1 vf .
37-8 f. 40-8 nc .
44-1 vf .
46-7 vf .
51'3 c. 54-7 f. 0-1 f. 6*9 f. 12.7 vft 18.6 f. 25-8 c..ee:eee ? ? ? ?
58 vf .
7 vf .
10'5 nb. 23 f ... ... .
30-8 vf .
37'8 nf .
44'1 f. 49-3 vf .
54-3 f. 0-4 f. 6'9 f. 13-4 vf .
o ... . .
18'6 w. 26-5 vf .
33-4 v-f . ?
... The three evidently represent the same system of lines , for those missing in the second and third have been lost by their faintness .
Nos. 12 , 19 and 46 , Table IX .
, are resolved by Duboscq into pairs .
It also shows 17 which were not visible in the other .
Nos. 5 and 11 of that again are not found in these : the first of these is almost peculiar to R. spectra ; and its not being visible in the narrow tube here seems to argue that condensed discharge cannot show it .
I I3 I &lt; I ? !
i I ?
i. 'L lREV .
T. R. ROBINSON ON SPECTRA OP ELECTRIC LIGHT .
These spectra are in some respects peculiar , those of tellurium most so .
In N , C.P. , No. 36 is so like the distinctive band of hydrogen that at first I suspected the presence of that gas .
The only difference I can find between them is that at the negative electrode it separates into two sharp and bright lines .
Still one can scarcely see it , without thinking that there is some close connexion between Te and H. I have already noticed the extreme brilliancy of its R. spectrum .
It is very difficult to work with from its rapid dissipation at the negative electrode , which is so great that a wire 0-01 in diameter disappears in a few minutes ; and the deposit is of extreme opacity .
In H few lines were seen ; for I was obliged to take them at a distance from the electrode , as it became invisible in a few seconds .
With arsenic , H , C.P. , I was more successful ; the centre of the spectrum was a bright cloud , in which only the three bright bands and two or three others were visible .
But at the negative boundary I got the largest number of lines , and the brightest that I ever met in H. The arsenic of course soon lined the tube ( 0'5 inch diameter ) with a dark mirror , through which .
the flash seemed crimson ; but an ellipse opposite the plates which form the slit remained quite clear .
I suppose the induction of the brass on the tube caused this .
The potassium ( air ) is of great beauty ; the double orange * is of as great intensity as with sodium .
No. 7 is an odd-looking olive band .
In H , C.P. , only the second orange * remains , but it is very bright ; the first one is entirely wanting,.but reappears in H , R. On exhausting , the second retains its brightness when the rest have vanished .
TABLE XIX.-Sodium .
Air .
lNitrogen .
Air .
Nitrogen .
C.P. R. C.P. R. C.P. R. C.P. R. 1 .
32 ... ... ... 36'5 vf ... ... .
26 .
3 51'5 nb. 51-3c .
51-6 vf .
51-3 c. 2 ... ... .
371 f ... ... .
double .
3 .
40'5 * ... ... ... ... . .
27 ... ... ... ... ... ... 52 4 .
41-5 n ... ... ... . .
double .
5 .
495 ... . .
48-1c ... ... .
28 ... 54 f ... . .
6 ... ... .
; ... ... ... 534 29 56-5 * 559 566f. .
7 .
55.5 n. 56'3 * 55'3 vb ... ... .
30 .
33 57-5 * ... ... 572 b ... ... .
8 .
32 57'5 vb ... ... ... ... . .
31 .
34 0-5* 11 f. 59-1 b ... ... .
9 .
33. .
26 b ... 32 ... ... ... ... .
291 n ... ... .
10 .
35 5n .
3 vf .
4 33c .
33 ... ... ... ... .
4.6 nf ... ... 11 .
55 ... ... ... ... ... ... 34 ... ... . .
5-2 nc .
12 .
8 n ... ... ... ... ... 35 ... ... .
82f ... 78 C. 13 .
10 n. .
0 ... .
-7 c ... ... .
361 l3n ... ... .
14 .
11-5n .
117 c. 12-4 f f. 12-4nb .
37 .
' 12-5nf ... ... one here .
15 , 16-5 w ... ... 17-7 15-7 f. 38 .
" 135 c. 13-4 vf .
13-1 nb. 14-4 c. 16 ... ... .
21:7 vf ... ... ... ... many here .
17 .
245 * 254 f. 254 The flash of the discharge is white , with a dense yellow atmosphere , to which I think the sodium line No. 8 is due .
very faint .
The Air , R. spectrum is , with the exception of this line , TABLE XX.--Graphite .
Air .
'Nitrogen .
Oxygen .
Carb. oxide .
Air .
Nitrogen .
Oxygen .
Carb. oxide .
C.P. C.P. C.P. C.P. R. C.P. C.P. C.P. C.P. R. I 1.32 ... ... ... ... ... 33-9o ... ... .
34 .
33 45 * 45-2 1 b. 45-4 b ... . .
2 , ... . .
36-3n ... ... .
: 358 fo .
35 .
46-2 7 b. 46:7 b. 467 n. 46-7 3 .
39 * 398 f. 37'7 c. 384 * 38-4 fo .
36 .
49-1 f. 493f ... ... .
479 nc .
4 .
418 ... .
40-3f ... ... .
37 ... ... 51-6 } o. 526 n. 51-3'\ * 51-3nc .
5 .
s 44 ... ... 43'3 f. 429 f 44-2 fo .
38 .
53-5 f. 54-2 C 5453 n. 536 n. 52 b. 6 .
i ... ... 45.6 vf .
46-2 f. 455 f ... .
39 .
56 * 559 56 55 563 55 ... ... 7 ... ... 481 * 48'5 4-'1 f ... ... .
double .
8. .
50 ... ... 501c .
51c .
40 ... ... .
572 582 57 c. 57-9 9 ... ... .
525n 527n ... ... 57 ... ... 41 .
33 585 * 59-1 vb. 598* 59-1 ... ..:10 .
' 56 * 542* 55-3* 56 b. 56 c. 42 .
34 2 nc ... ... . .
11 .
32 s ... ... ... ... ... 58-3nc .
59.3c .
43 .
3-5 38 no ... ... ... ... . .
12 .
33 ... ... 8 un .
2-3 bvn .
2-6 b. 26c .
corn13 .
5 , ' 4-9* 4* 3-6b ... ... pound .
14. .
61 f. 6-7c .
6 nf .
6 n. 44 .
517 nc ... ... ... ... .
6'5 15 ... ... 7-3* 45 .
9 9 ... .
5 un .
91 c ... ... 16 ... ... .
8'7vb .
9 vb. 8-3 f. 8-7 1 f. 46 , ' ... \ ... 10'nvb .
11'8vn .
10-1 c , ... ... 17 .
10 b. 102b .
10 ' 4 c. 10.4 f , 10-7 f. 47 .
tc " 13 12'4nb .
134 vb. 131 ... ... 18 1 b ... ... ... . .
11-4 f. 11'-7nc .
48 ... ... .
seen ... ... 147 f. 15 f. 19 .
13 b 12 b. 12 c. 124 J..49 1-7 f. 49 .
: ... ... 17 b 176 f3 f. 20 ... ... .
146 vf .
15 15 n. 14:7 f. 50. .
f ... ... 19'6 18 n. 19 b. 18-6 Jf .
' 21 .
, ... ... 16-4 ?
vf .
51 .
19'9vb .
206 n ... . .
22 .
18-5 '18-1 vf .
184 18. .
52 ... ... .
22-3f .
21-9n .
21'9 b. 23-2 f. ' ' 23 .
19 ... ... 9-9vf .
20'7 20 no. 21 f. 53 ... ... .
243 f , ... ... 24 ... ... .
22-8 vf .
23-4 b. 22'7 nco .
54 ... ... .
27 f. 258 272* ... ... 25 .
246 failed totally in getting the 0 spectra ; for at the very first discharge the mercury was so rapidly oxidated that not a glimmer could pass the thick deposit on the tube .
This singular energy is probably due to the presence of ozone in the electrolytic oxygen ( though it was passed through a tube filled with silver-leaf ) .
I shall repeat the experiment , passing it through a red-hot tube .
This fact may explain the absorption of oxygen in Geissler tubes , which was observed by PLuCKER ; for Hg vapour must be present in them , as they are exhausted by a mercurial air pump .
The deposit in CO was black below , grey above : the first was dissolved by nitric acid , the other not ; it was probably some form of carbon from the decomposition of the gas ; of that , however , there was a large residue present .
The spectrum of Hg vapour has a few very brilliant lines , but the rest are narrow and faint on a very dark ground .
I had expected by comparing it with Pt , Hg vapour , Table II .
, to obtain at once the lines of platinum , which I supposed would be superadded to those of Hg ; but , to my great surprise , that spectrum has fewer lines ( 28 , while the other has 48 ) , and only three occur in it which the other wants .
A similar deficiency occurs in the gas spectra of Table XXI .
: N , C.P. , wants 22 mercurial lines ; N , R. , 25 ; H , C.P. , 20 ; CO , C.P. , 12 ; and CO , R. , 30 .
These observations , it must be remarked , are strictly comparable ; they were all made with the Duboscq prism , nearly at the same time , and with the apparatus in precisely the same condition .
From this may be inferred , either that mercury gives different lines from its vapour , or that the gaseous media have lines which interfere with those of the electrodes so as to destroy each other .
I must reserve for a second communication the complete discussion of these and the other spectra , as it would , I fear , swell this one beyond all reasonable limits .
[ P.S. ( See page 954 .
) A more probable explanation of this peculiar form of the spectrum was suggested to me by Mr. STOKES , which is as follows : The aperture of the collimator was 1 inch , and its focal length 9'5 inches .
The distance of the spark from the slit was probably less than 1'5 inch , say 1'25 .
Hence if we assume for simplicity that all the light emerging from the collimator entered the object-glass of the observing telescope , which was nearly true , each element of the slit 1'25 1 would be capable of being illuminated by rays from a length 1= 95 X1=76 inch of the spark , provided the spark reached so far , the centre of that line being in a prolongation of the line joining the centre of the object-glass with the element in question .
If the length of the spark were very great compared with 1 , the central parts and those near the electrodes would illuminate widely separated elements of the slit , and the variations of brightness in a vertical direction would nearly correspond with the real variations of brightness of the different parts of the spark-though of course there would be a " ragged edge " above and below , where the spectrum was formed by partial pencils .
If , on the other hand , the length of the spark were infinitely small compared with l. 6u2 973 each illuminated element of the slit would receive rays from the whole spark , and the spectrum would be a band of uniform brightness in a vertical direction , and terminating abruptly above and below .
In the actual experiment the length of the spark was -6 inch , a quantity comparable with I , being a little less .
Hence the spectrum would consist of a central stripe of uniform brightness , corresponding to the small portion of the slit illuminated by the whole of the spark , accompanied above and below by a ragged edge in which the illumination fades away , as more and more of the spark is cut off , till nothing is left but the rays coming from its very extremity .
The distance from the central stripe to which the illumination would be perceived to extend would depend on the intrinsic brightness of the part of the spectrum regarded ; and , bearing in mind the relative brightness of different parts as laid down by FRAUNHOFER for the solar spectrum , we should get for the visible boundary a curve similar to that represented in the figure .
-March 1863 .
] Received September 1 , 1862 .
On comparing these Tables , it is seen that the places of a very large proportion of these lines are identical , or at least differ only within the limits of observation , although they may be very unlike in brilliancy and magnitude .
Such lines must be considered the same ; for the essence of a ray is its wave 's length , and the difference of intensity is but an accident depending on extrinsic causes .
That this comparison may be more easily made , I have given in the following Table the mean places of those whose identity is probable : except in the spectra of iron , copper , and many of carbonic oxide , such are the great majority of what I have observed .
For the conspicuous ones , which I have denoted by Greek letters , and which are recognized by peculiar characters , as also for some others which stand in definite relation to them , the means were taken without strict reference to probable errors ; for the rest those limits were taken which I have already mentioned .
It is possible that in several of the latter class one or two of the outliers may not belong to them ; but the general agreement of each is so close as to preclude doubt .
I include in the Table none beyond FRAUNIIOFER'S H , though several were observed , and also omit the very faint red lines which were visible in some spectra outside the bright part* .
It gives for each the mean place , the extreme deviations from that mean , the number of C.P. spectra in which it is found , in how many of these it is a * , and the same data for the R. spectra .
the ordinary character of which is bn .
When examined with the S2C prisms it is triple , the first being very narrow but bright , with a slight tinge of orange , the second and third pure yellow , the second broader than the third , and both sharply defined .
Nos. 26 to 29 form a very conspicuous group .
The first , s ( which begins the green ) , is marked by the very obscure interval which separates it from the others ; it as well as si are double , and the S2C prisms show many fine lines in their neighbourhood .
With iron , copper , and in many of the CO spectra , the number of these is very great ; it may be that the bright green from this to No. 44 is composed of such lines too close to be resolved ; but I think it is continuous .
No. 36 is nearly in the place of E. No. 38 v is very conspicuous on this bright ground ; it is often intensely brilliant , sometimes more so than 0 , and it occurs more frequently .
No. 37 seems to belong to it ; with the S2C prisms , and sometimes with the glass one , it is double of two equal .
Its brightness is scarcely developed in 0 , C.P. , very much in N , but quite as much in 0 , R. , as in N , R. CO , R. exceeds them both , though it has only 2 *s in C.P. With platinum , CO , C.P. , it is replaced by a bright bluish-green band of singular appearance ; this by careful focusing shows eleven or twelve fine lines , of which No. 37 is the first and No. 42 the last .
In this spectrum the group s , s " is also replaced by some twenty-five very fine lines from No. 23 to No. 36 .
The place of X is identical with the centre of the double line of b. No. 44 D , with its companions 42 and 43 , is very common , though the latter , 43 especially , are hard to see in the blinding glare of the first , which is often most intense .
They are always seen with S2C prisms , which also show 0 double ( as Duboscq does occasionally ) .
The components are generally , but not always , equal .
Though frequent in R. it is seldom bright .
In C.P. , N brings it out best ; it is wanting in 8 of 0 , 8 of II , 5 of CO ; and of its 50 *0 has but 4 , H 1 , and CO 3 ; one belongs to mercury in mercury vapour .
It ends the green .
No. 49 is remarkable , not only from its being in the place of F , but also from its being ( or being in ) the characteristic blue band of H , C.P. In the other gases and vapours it is scarcely ever a * , and often faint ; here it is of great brightness and great breadth , often 6 ' or 7 ' .
It has a cloudy but irresolvable look , and its edges are not sharply defined .
Even with 2 S2C prisms I cannot resolve it ; and it gives me the impression that it consists of light whose wave-length varies continuously .
Its mean by 221H , C.P. , is 33 ? ?
41'18 , corresponding to a wave length 1798 .
It is singular that the N , C.P. , spectrum of tellurium has a band possessing this peculiar type so decidedly that one who saw it without knowing its origin would undoubtingly assert the presence of H ; and my first idea was that this gas must be a component of the metal .
If , however , the eye be directed to the negative boundary of the spectrum , two brilliant points are seen there not found in the ordinary H band .
At the same time they may be due to the grosser metallic vapour which abounds there ; for Te volatilizes under the discharge more than any other metal which I have tried .
Iron in N , if the pressure be diminished a few inches , shows a similar band , though the other characters of the C.P. spectrum and discharge are unchanged .
Zinc in CO , C.P. , has one of the same sort 3 ' broad .
Nos. 51 and 52 are a well-marked pair ; but a much more remarkable group is that from No. 54 i to No. 59 z " .
Of these , 55 and 58 are very narrow , and , though bright , are often overpowered by the splendour of their neighbours ; the first has a still fainter follower seen in the S2C prisms , which also show each of the brilliant ones to be compound , i and X double , z " ( which begins the violet ) to consist of six fine bright lines .
In thirty-two cases x and i were observed as one , or as close double .
When taken separately , their distance was found 0 ' 96 ; and hence the place of X should be increased , and that of ' diminished by 0'14 .
H is unfavourable to this7group , which is most splendid in O and CO , except in Cu , CO , C.P. , when it is but a shadow of itself .
The ground is covered with very fine lines from No. 61 to No. 68 in the CO , C.P. , spectra of Cu , Pd , Fe , Bi , Sb , and Mg .
No. 72 v , besides being important in the spectra of other gases , is apparently developed in H , C.P. , to the broad violet band , the third of the three brilliant ones of that gas ; 22 of H , C.P. , give its place 34 ? ?
19 ' 82 .
It is not nearly so luminous as No. 49 is in H , but is quite as broad and even more undefined .
Unlike other violet bands , this is the first to disappear on rarefying the gas .
I have occasionally seen fine lines in it , which , however , do not seem to belong to it .
No. 73 is in the place of G , and No. 85 very nearly in that of H. If the individual spectra be compared with this Table , the result will , I think , throw some light on the questions which I mentioned at the beginning of this paper .
1 .
As to the essential connexion of each line of a spectrum with the chemical nature of some one metal , or some one gaseous medium , I think these observations are against it .
For instance , if any line , say No. 22 6 , be found with electrodes of aluminium in all the gases which we have examined , our first inference might be that it is an aluminium line .
But when we find the same thing is true of nickel , palladium , antimony , and magnesium , and that all the other metals have it in some of their spectra , we must conclude that it does not belong exclusively to any metal ; nor can it be considered a gas line .
In N it is indeed found with all the twenty-two electrodes ; but it also occurs nine times in 0O , fourteen in H , as well as in the vapours of mercury , phosphorus , and bisulphuret of carbon .
There are many similar cases ; and though in general the lines of this Table are not seen with all the metals , yet it will be found that , with respect to gases ( omitting No. 1 , which on account of its bright neighbours is seldom seen ) , out of the whole eightyfive only Nos. 5 , 6 , 9 , 73 , 79 , and 85 are wanting* in H , C.P. ; only No. 6 in CO , and that of these all but No. 5 are found in the R. spectra , and none are wanting in A , N , and 0 .
As to metals , the lowest number , 12 , occurs in No. 85 ; but the average is fat above this .
These facts may be explained in two ways .
We may suppose that the action of the electric discharge on the molecules which transmit it has in itself intermittences which tend to produce maxima of light recurring at intervals , and which are effective in doing so when the forces inherent in these molecules are in accordance with them .
If that accord be perfect , the development of light will be intense ; but if imperfect or wanting , the corresponding light will be feeble or will vanish .
Or it may be supposed that our metals and gases are compounds of unknown elements which are separated by the dis , charge , and exhibit their appropriate lines .
This hypothesis is tempting , for if true it would at once lead to analysis of many of our supposed simple substances ; and the facts which have been stated respecting the band No. 49 give some countenance to it .
I however think the first view of the matter more probable , for reasons which I shall soon state .
According to it , the existence of a luminous line merely indicates the presence of matter in the circuit ; but its intensity depends on the nature of that matter , which may either make it extremely bright or obscure it , even to invisibility .
It is generally supposed that the presence of a metal gives brilliant lines , and that those due to a gas , or vapour are less bright ; but as the two are always simultaneously present it is not easy to separate their influences .
A promising plan of effecting this has been proposed by PLUCKER .
Two balls connected by a capillary tube were filled with any gas and then exhausted ; the balls were coated with tinfoil , and when they were connected with the terminals of an induction machine reciprocating discharges took place , which , though very faint in the globes , were bright enough when condensed in the capillary part to .
give a spectrum .
Here nothing but glass is in contact with the rarefied gas , and therefore he expected to obtain only gas lines .
I repeated this experiment , adding a glass stopcock to one of the balls , that I might use the same glass with different gases.:The spectra observed were the IR .
of N , 0 , and H ; the first had 13 , the second 23 , and the last 13 , besides several in each too faint to be taken .
Of these wereCommon to three .
In N and 0 .
In O and H. In 0 .
In N. In H. Nos. 18 Nos. 27 Nos , 49 Nos. 16 Nos. 14 Nos. 5 39 68 75 24 76 13 46 33 84 82 57 36 61 41 65 42 78 44 48 52 54 72 980 None of these , however , are peculiar , except that Nos. 11 , 17 , and 54* appear as double .
That the presence of a new body in the circuit developes oftener than originates lines , is further shown by moistening the electrodes with a solution of some salt ; for instance , chloride of bariumf .
In trying it , platinum electrodes ( carefully washed with nitric acid and distilled water ) were lapped with clean cotton thread to retain the fluid , which , however , soon evaporates by the heat of the discharge .
The change of the spectrum is almost startling from the instantaneous and intense development of many lines , and the increased lustre of some which before had been noted as *s .
The most conspicuous areo l s. Very bright ... . .
32 38'1 a ' .
* very intense. .
40'0 ' .
* dazzling ' ... . .
491 y. Very bright.* ... 54*7 No. 16 .
Orange intense ... 56'3 No. 17 .
Very bright ... .
57.7 No. 34 .
* very intense. .
33 18'4 Nothing striking till 0 , then No. 47 .
* brighter than 0. .
37-4 No. 48 .
Almost *. .
39-4 X. * of exceeding intensity..34 2*3 X ' .
Very bright ... ... 4'6 No. 80 .
Extremely bright ... 40'6 All these are found in the platinum air spectrum , though in most cases with a totally different aspect ; so that the barium makes no change in the place of the lines , but only in their brightness .
2 .
That the peculiar character of a line is modified by other circumstances than the chemical nature of the bodies engaged in producing it , appears from the difference between the spectrum at common-pressure , the transition spectrum , and that in rarefied gases , to which I have already referred .
The chemical conditions are unchanged ; but at first sight nothing is more dissimilar than the three ; the first with its numerous bright *s ; the second a shadow rather than a spectrum , with a bare suspicion of a few lines ; and the R. one much brighter , but with a seeming discrepancy of lines that marks it as peculiar .
The R. has , on an average , 0'53 of the C.P. lines ; it extends as far in the violet , but is cut off in the red part .
There , especially in N , are often found a set of equal bands reaching from No. 11 to No. 19 inclusive ; there is also another set , that evidenboth that the spectrum of air is not formed bythe mere superposition of those of N and 0 and that certain electrodes can excite lines in a mixture of two gases which axe not visible in either of those gases taken separately , or cannot excite them in the mixture , though they can in either or both of the components .
This unexpected fact is illustrated by an attempt which I made to obtain the lines of platinum , by comparing its spectrum in mercury vapour with that of mercury .
I expected , as I had done in the casesalready mentioned of gases with Hg electrodes , that 1 should get the spectrum of IPt+that of Hg ; but the result was otherwise .
Hg Hg had forty-eight lines , Pt Hg only twenty-five , so that the presence of Pt instead of Hg as electrodes pu out twenty-three .
It however brought out four new ones , of which No. 15 y is found with every solid metal , Nos.'33 , 59 z " , and 76 1 with most of them .
There is here no superposition , but , instead , either an antagonism of platinum and mercury , or the curious fact that:tha fluid electrodes act differently from the solid .
It-would be interesting to compare the spectra of tin in these two states .
The case is the same with a chemical compound as with a mixture .
On the principIe of superposition , the spectrum of CO should be merely the sum of those of C and 0 .
To avoid all uncertainty about metallic lines , I used electrodes of graphite in the two gases .
In each case nothing but C and 0 were present , and I expected to get the same spectrum in both .
They were , however , very unlike .
Of the lines which they had in common , No. 3a is nc .
in. .
0 , but is a in ... .
CO No. 10 f is a in .
0 , but only c. in ... .
CO No. 15y is a in .
0 , but only b in ... CO No. 22 ' is intense e in 0 , but only b. in ... .
CO No. 26 s -is wb .
in .
0 , but one of many very fine in CO No. 39 is faint in .
0 , but a ' as bright as X in .
CO The , group is not nearly so bright in O as in CO. Besides these differences , 0 has ten wanting in CO , among which are No. 38 8 , very bright , and No. 44 0 , in all its brilliancy ; while CO has eight wanting in 0 , of which No. 75 is a * ; the difference being greater than I have sometimes found with much less chemical agreement .
The only conclusion which such facts permit is , that the spectrum is a simple resultant of all the actions present , some of which may combine to produce an exalted effect , while others may be antagonistic in any degree .
4 .
As to the manner in which electricity produces the rays of these lines , whether by merely heating the medium , or by some luminiferous action analogous to its heating , nothing is really established .
Heat is known to produce some brilliant lines when metallic vapours are introduced into flames ; and it is possible that a temperature far above that of our hottest flame might bring out all those which I have enumerated , and the multitude of others which I did not attempt to measure .
But is the temperature of the elec-

