Fluorescence and Resonance of Sodium .511 When ylr differs from zero , the calculations are naturally more complicated .
The most interesting and instructive cases occur when yfr is small .
I have chosen ^ = 1/ 10 .
The corresponding values of f , x , and y are given in Table III , calculated from equations ( 10 ) , ( 11 ) , and a plot is shown in fig. 3 .
As in the former problem , where the liquid is supposed to adhere to the walls notwithstanding the sharp edges , the pressure in the recesses of the tube is simply that due to the velocity at a distance .
At other places the pressure can be deduced from the stream-function in the usual way .
Observations on the Fluorescence and Resonance of Sodium Vapour.\#151 ; II .
By the Hon. R. J. Strutt , Se .
D. , F.R.S. , Professor of Physics , Imperial College , South Kensington .
( Received June 12 , 1915 .
) [ Plate 8 .
] S 1 .
Introduction .
Prof. R. W. Wood has made many interesting observations on the fluorescence of sodium vapour .
They are conveniently summarised in his ' Physical Optics ' ( MacMillan , 1911 ) .
The fluorescent spectra he has obtained are apparently connected with the banded absorption spectrum of dense sodium vapour .
If white light is employed , this banded absorption spectrum is re-emitted completely as a fluorescent emission spectrum .
If monochromatic light is used , a portion only of the complete band spectrum is emitted , this portion consisting of a moderate number of lines approximately equally spaced along a normal spectrum , and including a line coincident with the exciting line .
These observations on the band spectrum are of great interest and importance , but observations on the line spectrum of sodium , in absorption and in fluorescence , are more within the range of theoretical discussion at the present time .
The line spectrum is observed in absorption and in fluorescence at a density of sodium vapour small compared with that needed for the band spectrum .
If a beam of white light traverses such vapour , the D line and the other lines of the same series in the ultra-violet are seen in absorption .
At the Hon. R. J. Strutt .
Observations on the same time there is a lateral re-emission of the D line , discovered by Wood , and called by him resonance radiation .
The first paper of the present series* * * S dealt with this resonance radiation .
It is excited most conveniently not by white light but by sodium light , which contains the only effective constituent of white light .
The D line is the first member of a series of lines known as the principal series of sodium .
The second line is at wave-length 3303 in the ultra-violet .
The lines of this series up to a very high number can be seen in absorption , and since , in the case of the D line , the whole of the absorbed light quantitatively reappears in resonance radiation , f it would seem at first sight very probable that resonance radiation would also be observed at wave-length 3303 , and at all the other lines of the Principal series .
The question arises also , would stimulation by a higher member of the series give rise to emission of a lower member ?
For instance would stimulation by \ 3303 give rise to D light ?
This question is not new .
* It has been proposed by Prof. R. W. Wood , and he has looked for such an effect , with negative results.^ There was no doubt , however , that if anything of the kind could be observed , the fact would be important for the theory of spectrum series .
For this reason I have made a fresh effort , which has had a successful issue .
S 2 .
Excitation of D Light by A , 3303 .
In planning the experiments , a source was looked for which would give a sharp and narrow line of great intensity at wave-length 3303 .
Preliminary experiments with a salted Meker flame showed that this source was extremely poor in the radiation desired .
An oxy-hydrogen sodium flame might have done better , but I had recourse to a sodium vacuum arc , in quartz , which gave the line 3303 in equal intensity to the D line , measuring the intensity by the action on ordinary ( not orthochromatic ) plates .
S The silica envelope of the lamp ( fig. 1 ) consists of two bulbs , , c , each of 50 c.c. capacity , joined by a tube b , about 13 mm. inside diameter .
The cathode is a pool of molten sodium d , and contact is made to it by an iron wire e , winch passes out through a narrow quartz tube into which it is * 'Roy .
Soc. Proc. , 'A , vol. 91 , p. 388 ( 1915 ) .
t Dunoyer and Wood , ' Phil. Mag. ' [ VI ] , vol. 27 , p. 1025 ( 1914 ) .
J * Phil. Mag. ' [ Vlj , vol. 18 , p. 533 ( 1909 ) .
In this paper it is shown that excitation of dense sodium vapour by the group of zinc lines in the neighbourhood of A 3303 gives rise to an ultra-violet band spectrum .
But no definite evidence is given of any re-emission of A 3303 , nor have I myself been able to observe this , as will appear .
S This measure of- -relative intensity has of course no scientific yalue .
It is merely used for practical convenience in comparing different sources .
An energy measure would be enormously more favourable tb the D line .
Fluorescence and Resonance of Sodium Vapour .
/ - C Fio .
1 .
cemented with sealing wax .
The anode is an iron wire f about 3 mm. in diameter .
It passes up the glass fitting gg , into which it is cemented ; gg is cemented into the silica tube at h. A stopcock h provides for exhaustion .
The constriction at the bottom of gg , where the wire f passes out , prevents too much distillation of sodium into the upper part of this tube .
This form of lamp , arrived at after repeated trials , is fairly satisfactory in use , lasting long enough for a considerable number of experiments .
The large bulbs are a most important feature .
The lower one allows of the effervescence of the molten sodium , which always occurs at first .
The upper one serves a different purpose .
The anode f gets red hot , and if space is not allowed round it , the surrounding silica gets so hot as to be rapidly acted on by the sodium , eventually cracking .
The lamp must be kept well exhausted , preferably by continuous action of a Gaede mercury pump .
The sodium gives off much hydrogen , and if this is not continually removed the lamp gets very hot , and the silica is rapidly blackened by reduction .
The current passed is about 5 amperes .
The lamp is started by warming it and connecting an induction coil to the sodium cathode , and to a wire twisted round the tube b. After prolonged use the silica gets brown .
The lamp can then be dismounted , cleaned out with alcohol and water , and then with dilute hydrofluoric acid , which readily removes the reduced silicon .
It can then be recharged with sodium , and is almost as good as new .
It may be mentioned that the lamp gave eight lines of the principal series of sodium with half an hour 's exposure on a small quartz spectrograph .
The D line was exceedingly brilliant , and the strongest salted flame looked a very pale object beside the lamp .
The lines of the two subordinate series were also shown very beautifully , the series relation being conspicuous to the eye .
Such lamps show the resonance radiation of D light 2 s VOL. xci.\#151 ; A. Hon. R. J. Strutt .
Observations on the incomparably better than any salted flame will do .
The advantage lies not only in the brilliancy , but also in the narrowness of the D line .
As Dunoyer has shown , only the central part of the broad line given by a flame is effective .
In studying the fluorescence produced by the line 3303 , it is necessary to suppress the light due to the visual lines , particularly the D line .
This was accomplished by a cell of cobalt blue uviol glass , filled with a dilute solution of nitrosodimethylaniline .
This combination is very transparent to \ 3303 , and suppresses everything in the visual region except a small part of the light of the red sodium line 6161 .
It is necessary to tolerate the latter , for if the nitroso solution is made strong enough to suppress it , serious loss of 3303 light is incurred .
In practice not enough red light came through to cause inconvenience .
The light coming through the filter was examined with a quartz spectrograph .
The only line on the plate was 3303 .
The visual spectrum showed only the red line 6161 , and this could only be seen by directing the spectroscope straight at the lamp through the absorption cell .
The arrangements are shown in plan in fig. 2 .
The lamp is at A. It is focussed by a quartz condenser B on the wall of the quartz flask E , the ultraviolet filter D intervening .
B is set in the wall C of a dark cupboard , the axis of the lenses being inclined at about 60 ' to the wall , whioh makes the apparatus more accessible to an observer .
The observations are made inside the cupboard .
E is a tound-bottomed quartz flask of 100 c.c. .capacity containing a little O A V/ Z/ // // // / c Fig. 2 .
Fluorescence and Resonance of Sodium Vapour .
515 sodium .
It is connected by rubber tubing to a Gaede mercury pump , which is kept running , and which serves in addition to keep the sodium lamp A exhausted .
Before starting observations , the eyes are well rested in the dark .
The bulb E is then well heated in a Bunsen flame , to nearly a red heat , so as to generate sodium vapour .
When the lamp is started , the front surface of the bulb is seen to emit visual light .
In the earlier experiments this was too faint for spectroscopic analysis , but it was examined by absorption methods .
First , a cell 1*2 cm .
thick filled with saturated potassium bichromate solution held in front of the eyes was found to transmit it undiminished .
This shows that the light must be in the yellow or red region of the spectrum .
Secondly , the light was entirely cut off by substituting for the bichromate a strong solution of prsesodymium nitrate .
The latter solution transmits the red , but has an absorption band extending from \ 5820 to X 6020 .
The D line at X 5890 lies in this narrow region , and considering that the emission takes place from sodium vapour , there was practically no doubt that it did consist of D light .
In later experiments , under improved experimental conditions , the light was bright enough for its orange colour to be very conspicuous .
Examined with a direct vision spectroscope it was found to be monochromatic , consisting entirely of D light .
So far I have supposed the bulb to be quite hot , say at 350 ' C. In this case the 3303 light is not able to penetrate far into the sodium vapour , and consequently the D light excited is confined to the surface of the mass of vapour .
As the bulb cools and the vapour diminishes in density , the excited D light gradually penetrates into the interior of the vessel , and fills it entirely .
The light then slowly fades away as the vessel becomes cold .
These effects are exactly the same as are observed when D light is used as the exciting light .
Possibly suspicion may arise that a trace of D light had been able to penetrate the ultra-violet filter and produced the observed effect .
A simple test experiment entirely negatived this possibility .
When a piece of ordinary plate glass , 1*2 cm .
thick , was interposed in the path of the exciting beam so as to cut off ultra-violet light , the excited D light disappeared altogether .
S 3 .
No Excitation of the D Lines by Other Kinds of TJltra-violet Light .
Another* doubt may possibly remain .
Granting that the excitation of D light is due to ultra-violet light , is it essential that this particular wave-length 3303 should be used ?
Would not any intense ultra-violet radiation serve ?
Hon. R. J. Strutt .
Observations on the To test this a mercury arc in quartz was substituted for the sodium arc .
It gave very strong radiation through the ultra-violet filter , lighting up a piece of uranium glass much more brightly than the sodium lamp had done .
It also caused the quartz wall of the bulb containing sodium to emit a little blue fluorescence , but when this was cut out by holding a bichromate solution before the eye , nothing remained : D light was not emitted .
The strong ultra-violet mercury lines present were probably the groups about X3650 and X3130 .
S 4 .
Search for Evidence of Resonance Radiation at X 3303 .
It appears , therefore , that light energy of wave-length 3303 absorbed by sodium vapour is , in part at any rate , re-emitted as D light .
Clearly then the relation found by Wood and Dunoyer for stimulation by D light , * that the whole of the absorbed energy reappears as energy of the same wavelength , cannot apply to the line 3303 .
It must be a peculiarity of the D line , distinguishing it from other members of the series .
The question arises , however , whether any part of the energy of X 3303 is re-emitted at the same wave-length ; whether , in other words , there is any resonance radiation of X 3303 .
An experiment was arranged as in fig. 3 .
In this case a cylindrical silica tube , about 2 cm .
in diameter , kept in connection with the pump , contains the sodium vapour .
Light of wave-length 3303 is incident perpendicularly upon it , as shown by the arrow .
Some ultra-violet light is found to proceed obliquely from the point of incidence .
It is focussed by a quartz lens B upon a fluorescent screen of uranium glass .
The image is at D , and is shown by the fluorescence .
This fluorescence must be viewed obliquely .
If viewed in the prolongation of AD , yellow D light from A , stimulated in the way already described , enters the eye , and masks the fluorescence on the screen at D. It is necessary to decide whether the ultra-violet light emitted from A is due to resonance radiation , or whether it is merely light scattered by the striae in the walls of the silica tube.*f A comparison fluorescent patch was obtained at E by forming there the image of a small smokeless gas flame F. E was adjusted to be of the same brightness as D when the tube A was cold , and contained no sodium vapour , and when the solid sodium in the tube was out of the way of the incident light .
On heating A so as to raise sodium vapour no increased brightness of the * ' Phil. Mag. , ' vol. 27 , p. 1025 ( 1914 ) .
+ The arrangement adopted excludes specular reflection from the silica tube , as will be evident from fig. 3 .
Fluorescence and Resonance of Sodium Vapour .
fluorescent patch at D could be observed .
E was still a match for it .
This proves that the resonance radiation of \ 3303 , if any , is but a small fraction of the light of this wave-length scattered by the walls of the vessel .
A similar experiment was made with the same vessel but with D light a6 the exciting light .
The ultra-violet filter was removed , a sheet of ordinary ground glass was substituted for the uranium glass at C , and a salted spirit flame for the smokeless gas flame at F. Starting with the sodium tube A cold , the yellow image at E was adjusted to equality with that at D. On heating the tube A , so as to raise sodium vapour , D became D E C Fig. 3 .
many times brighter than E , owing to the resonance radiation , which entirely outshone the scattered light .
There is no reason to suppose that the fraction of the incident light scattered by the striae in the quartz tube would be very different , whether this light consisted of D light or of X3303 .
Assuming that the same fraction is scattered by the walls of the tube in each case , it follows from the experiments described that the resonance radiation is a very much smaller fraction of the exciting radiation for X 3303 than for D light.* * -X tube with a window of clear optically worked quartz fused on would make it possible to push further the search for resonance radiation of X 3303 .
Hon. R. J. Strutt .
Observations on the As already mentioned , the experiments of Dunoyer and Wood indicate that the whole of the absorbed D light is re-emitted as D light .
My own experiments prove that this is not true for light of \ 3303 .
The interesting possibility remains for future investigation that not only a part but the whole of the absorbed energy of \ 3303 is re-emitted as D light .
Such a result would , howevbr , be difficult to reconcile with the quantum theory .
S 5 .
Effect of One Component only of the Ultra-violet Doublet \ 3303 .
The first ultra-violet sodium line , though hitherto spoken of as single , is in a fact a doublet like the J ) line .
The components have wave-lengths , on the International system , of 3302 96 and 3302*35.* There are reasons for regarding these lines as belonging to different series : in fact , for regarding the principal series of sodium as really including two series .
The series including Di and 3302*96 , the less refrangible members of the respective doublets , are each split into four components in a magnetic field , while 3302*35 and D2 , the more refrangible members , are each split into six components .
This point of view is strengthened by the observation of Wood and Dunoyerf that if sodium vapour is stimulated by the D2 line , Di does not appear in the fluorescent radiation , which consists of D2 light only .
It thus appeared of considerable interest to determine whether stimulation by the line 3302*96 would give rise to Di only , or whether it would give rise to D2 ( which may be regarded as belonging to a different series ) as well .
An accidental circumstance , to which attention has been drawn by Wood , made it possible to attack the question .
This is the extreme nearness of certain zinc lines to wave-length 3303 .
The zinc arc spectrum shows a doublet which lies inside the sodium doublet .
The wave-lengths given by Kayser stand thus : Intervals .
Sodium 3303*101 0*05 Zinc . .
, 3303*05 J Zinc 3302*701 0*21 Sodium 3302*49J Kayser 's table has been compiled by critical collation of the data given by different observers .
These observers have not of course felt any special interest in the exact relative position of the zinc and sodium lines mentioned , -* which has no importance except for the technique of the present investigation .
It was thought therefore that direct comparison would give greater certainty * I quote these wave-lengths from the Table given at the end of Kayser 's ' Spectroscopy , ' vol. 6 .
'm t ''Phil .
Mag. ' [ VI ] , vol. 27 , p. 1018 ( 1914 ) .
Strutt .
Roy .
Soc. Proc. , , vol. 91 , PI .
8 .
Fig. 4 .
Fluorescence and Resonance of Sodium Vapour .
as to the exact intervals .
The spectra were compared in the third order of a 10-foot Rowland grating , using a quartz vacuum arc both for zinc and sodium .
With short exposures the lines , especially in the case of sodium , were very fine and sharp .
The intervals above mentioned were measured as 0 043 Angstrom and 0*198 Angstrom , in good agreement with the numbers adopted by Kayser .
In all the photographs there was some overlap of the images of 3303*10 ( sodium ) and 3303*05 ( zinc ) .
On the other hand there was definite separation of 3302*70 ( sodium ) and 3302*49 ( zinc ) by a narrow interval of clear glass .
When there is separation , it is definite proof that the lines themselves do not overlap , but on the other hand an overlap of the photographic images is no proof of overlap of the lines themselves , for the images may be , and with any but the shortest exposures certainly are , broadened by irradiation .
For this reason 1 have not been able to definitely establish from the spectrograms whether , and to what extent , 3303*05 ( zinc ) overlaps 3303*10 ( sodium).* It is scarcely doubtful that , as is known to be the case with mercury , the zinc lines broaden with increase of current .
If therefore the zinc line 3303*05 does not overlap its sodium neighbour for small currents , an increase of the current will probably make it do so ; on the other hand , up to 30 amperes at all events , the spectrograms still showed that there was no overlap of 3302*70 ( zinc ) and 3302*49 ( sodium ) .
Fig. 4 ( Plate 8 ) shows an enlargement of a photograph taken with 30 amperes passed through the zinc lamp .
The closer doublet in the middle is of course the zinc one , the sodium doublet appearing above and below .
The net result of this study of the spectra is to show that there is a zinc line in practical coincidence with the less refrangible member of the sodium doublet , but no zinc line in coincidence with the more refrangible member .
If we use a zinc arc to excite sodium vapour , we apply excitation at one line only of the doublet 3303 .
The form of lamp which has been found useful for zinc vapour is shown in fig. 5 .
It is a slight modification of that used for sodium .
The envelope of the lamp consists of two lengths of rough silica tube each about 13 mm. inside diameter and 30 cm .
long , united by a short piece of clear silica tubing about 8 mm. inside diameter .
The cathode is a layer of molten zinc d resting on an iron plug e , which of course must not fit the silica tube too tightly .
A mica washer at f prevents the molten zinc running down below e. Connection is made to e by an iron wire g passing out through a glass fitting A , * Probably this could be done if the spectrograph were arranged with a suitable lens to form an enlarged image on the plate of the original grating image of the group 3303 .
Hon. R. J. Strutt .
Observations on the -a cemented with sealing wax .
The anode h is also a loose iron plug , with a pointed end , as shown .
The connection is brought out in the same way .
A side tube l in the glass fitting provides for exhaustion .
V 's .
The arrangements for the experiment are seen in plan in fig. 6 .
The zinc lamp is shown in section at A. The rays from it pass through the quartz condenser B set obliquely in the wall as before , then through the ultra-violet filter D , coti- ' verging at E on to the wall of a quartz bulb containing soditun vapour .
The rays impinge on the quartz bulb almost at grazing incidence , and set up a fluorescence of D light , as when the , , sodium lamp is used as a source of 3303 light .
The intensity of this light depends very much on the current through the zinc lamp . .
This is no doubt due to the fact that at small currents the zinc line is too narrow for any overlap with the sodium line .
With larger currents such overlapping begins , and the vapour is stimulated .
With currents of 4 amperes no D light is seen .
Increasing the current to 6 amperes it can be seen , and identified by the absorption methods before described .
At 15 amperes it is bright enough for spectroscopic examination .
The current may be further increased up to 100 amperes , with a great increase in the fluorescent light , which is now , perhaps , as bright as a slightly salted flame .
The zinc lamp , however , will only stand these currents for a few seconds .
It has been found convenient to start up the lamp , and run it at , say , 6 amperes , by means of a regulating resistance .
When all is ready for observation this resistance is shunted by a smaller one , so as to pass a heavy current for a short time .
For the heaviest currents a fuse wire without other resistance is used in the shunt circuit , and allows the heavy current to pass for a few seconds only ; enough , however , for observation .
The lamp is not extinguished when the fuse melts in the shunt circuit , thus the observation can easily be repeated after replacing the fuse .
In order to decide whether one or both D lines were present , a special spectroscope was employed , consisting of slit , direct vision prism , and telescope .
The slit , 1 mm. broad , is a fixed one , made in a brass sheet F. The slit itself is horizontal , and is in the plane of the diagram ( fig. 6 ) .
A vertical strip of brass G divides the length of the slit into two parts .
One of these parts ( that towards the top of the page on the diagram ) is backed by Jihe fluorescent light of the sodium vapour in E. The other , part Fluorescence and Resonance of Sodium Vapour .
521 is backed by a minute soda flame H , which affords a comparison spectrum .
This flame is simply a coal-gas flame burning from a drawn-out soda glass " , pA Fiq .
6 .
jet , and turned down " to the blue .
" Its spectrum shows the D line superposed upon the Swan spectrum .
The prism used ( indicated at J ) is a powerful combination designed by Lord Rayleigh , * to whom I am indebted for the loan of it .
Ten right-angled flint glass prisms , with their refracting edges horizontal , are immersed in carbon disulphide , containing enough benzol to exactly compensate the deviation for D light .
The combination is 20 inches long .
It was mounted with its nearest end 48 inches from the slit , though represented much nearer in the figure , for convenience .
Under these conditions the D lines in the comparison spectrum were seen well separated with the naked eye , but in the experiment a telescope K was used .
It consisted of single convex lenses , object glass 5 inches and eyepiece inches focal length , giving rather more than three-fold magnification .
With this amount of magnification the images lose nothing in brightness , while they gain in angular magnitude\#151 ; always an advantage in dealing with very faint objects .
The telescope tube was covered with phosphorescent paint near the eyepiece , which proved a convenience for quickly finding one 's way to it in the dark .
The spectroscope thus arranged proved most satisfactory for resolving the D lines from a very faint source , and I do not think that it could be improved on for this purpose .
The great advantage of the large dispersion is that a wide slit is admissible .
In making the observations the zinc lamp was started with a small current , the quartz bulb was heated , and the observer at the spectroscope * See his ' Collected Scientific Works , ' vol. 1 , p. 459 , and vol. 4 , p. 394 .
VOL. XCI.\#151 ; A. 2 T Hon. R. J. Strutt .
Observations on the got his eye rightly directed and focussed so as to see the D lines well in the comparison spectrum .
He then knew exactly where to look for these lines in the usually fainter fluorescence spectrum , during the short time it was on .
This use of a comparison spectrum is practically essential .
It is strongly to be recommended in all spectroscopy of faint sources .
At the first test , 100 amperes were passed for a moment , and both D lines were seen in equal intensity , as bright as the comparison spectrum , if not more so .
In subsequent tests the current was reduced by successive stages down to 15 amperes .
At this point the result was still the same , quite definite and distinct .
With smaller currents than 15 amperes the fluorescent light was too faint for spectroscopy .
In all not less than a dozen observations were made , not all on the same day , and each one of them separately was conclusive as to the general result .
I was not able to decide definitely which , if either , of the D lines was the brighter .
In a flame Da is somewhat brighter than Di , but in the spark no definite difference is recorded .
As already shown , stimulation is confined to the component 3303*10 at 30 amperes , and possibly for much larger currents still .
Both the D lines appear in equal intensity at 15 amperes .
There is , therefore , an ample margin of safety in saying that stimulation applied at one member only of the first ultra-violet doublet gives rise to emission of both the D lines .
S S 6 .
Search for Fluorescent Emission of the Subordinate Series Lines .
As stimulation at 3303 had proved capable of bringing out the D lines it seemed possible that it might also bring out the subordinate series lines .
To look for these a zinc lamp was first used , loaded momentarily with 100 amperes .
The arrangements were as in fig. 6 , except that instead of using the prism J and the telescope K , the slit was viewed without a telescope , through a much less dispersive direct vision bisulphide prism , the eye being only about 8 inches from the slit .
The spectrum obtained in this way was short and bright .
The swan spectrum of the little gas jet H served as an approximate guide for the eye , and the lines 5685 and 6157 of the first and second subordinate series specially looked for .
No trace of them , or of any other line except the D line , could be seen in the fluorescent spectrum .
A second attempt was made , using the sodium lamp as a source of A 3303 .
A quartz tube , 2 cm .
diameter , was used to contain the vapour , and examined with an ordinary pocket spectroscope .
In this case the brightness was " ^perhaps somewhat less , but the observations could be made at leisure because the^odium lamp does not need to be heavily overloaded in order to Fluorescence and Resonance of Sodium Vapour .
523 bring out the fluorescence .
No trace of the subordinate series lines could be seen .
The green one , 5685 , at all events , could have been seen if it had had the same intensity relative to the D line as it had in the original sodium lamp .
My colleague , Prof. Fowler , very kindly made these observations with me .
S 7 .
Summary* ( 1 ) Sodium vapour , illuminated by the second line of the principal series at wave-length 3303 in the ultra-violet emits the D line , which is the first member of that series , in fluorescence .
( 2 ) Stimulation by ultra-violet light in general does not cause emission of the D line .
( 3 ) If one member only of the ultra-violet doublet 3303 is stimulated , not one only , but both of the D lines are emitted , in about equal intensity .
This is an unexpected result , in view of the work of Wood and D unoyer , who found that stimulation by Da light was unable to excite Di light .
( 4 ) Stimulation at X 3303 was not found to give rise to any observable resonance radiation of the same wave-length , nor to any observable emission of the subordinate series lines .
I have much pleasure in thanking my assistant , Mr. P. Thompson , for valuable help in carrying out the experiments .
* The chief results of this investigation were indicated in preliminary .communications to * Nature , ' May 13 and June 3 , 1915 .
VOL. XOI.\#151 ; A.

