]\gt ; 16 Prof. A. Gray and Dr. W. Stewart .
Radiation of [ Apr. 27 , " " On the Radiation of Helium and Mercury in a netic Field By Professor ANDREW GRAY , F.RS .
, and D.Sc .
, with A. HOUSTOUN , M.A. , and 1 ) .
B. M.A. , Research Students in the University of ceived \mdash ; Read May 14 , 1903 .
The experiments described in the ring paper had for their primary object to test , for as many lines as possible of different substances , the proportionality of the change of -length , for eaoh of the components into which a single spectral line is resolved ) the application of magnetic field , to the field intensity , and deduce the corresponding values of the ratio of charge to mass of the electron .
It has been found difficult to obtain a ready lpply of good discharge-tubes ' .
and up to the present our observations ) ecll mainly of the spectra of helium and mercury , in the latter case chiefly the green line of wave-length approximately 5461 tenth-metres .
With respect to this line we incidentally observed fully a year ago , as we afterwards had also been done a little earIier by Zeeman , *that in its immediate neighbourhood there existed a number of comparatively faint lines , so that what appeared in an ordinary spectroscope as a single bright line was in fact a group of six or seven , consisting of one bright l a triplet of faint satellites on the side next the violet end of the spectrum , and triplet or at least a doublet of faint lines on the other side .
These faint lines were very difficult observe and to determine , and the measurements of wave-lengths which we obtained do not agree with those snpplied to by MM .
Fabry and Perot .
The lines , however , we found were only visible in our tubes , when the discharge had beeIl kept going for a time , and when , therefore , the vacuunl had been considerably changed , and from glimpses that were obtained at times of what seemed to be further lines , its appears probable that the full complexity of the greell line , when unaffected by a magnetic field , has not yet been disclosed , and oan only be investigated under peculiar conditions of the vacuum in the tube , and possibly of the action of the discharging coil .
An important question is raised by these facts as to what it is that is resolved by a magnetic field ; it may not bc correct to regard the complex of lines , which are inly seen when a powerful field is applied to this light , as a resolution of the simple line .
It lvill be convenient to defer fnrther particulars of these observ mtil after short description of the pparatus e * ' Astl osicol J.1d , ' vol. lo ( 1902 ) , ) .
1903 .
] and a Ma .
cfro-/ npfr/ no f.\mdash ; The apparatus consisted of a set composed of a large electromagnet and an echelon of twenty-six plates auxiliary .
The magnet specially made for the purpose of these experiments and some others .
The iron part ( yoke , limbs and poles ) , is rectangular in form , and is about tlIree feet long by about a foot and a quarter deep .
The permeability of the cast-steel used for the yoke and limbs coincides ] with that given } ) Hopkinson for wrought iron : for C.G , S. the no of is 15:000 C.G.S. large magnetising coil the yoke and extends over its hole length .
The poles are provided with smaller coils about a foot in length , having respectively 1140 and 1170 tnrns of copper of No. 9 standard wire gang .
The coils can stand easily about 16 amperes , and for short periods 50 01 60 amperes .
The field intensity between the poles , as tested by the ballistic method , is great , but mlfortunately we not hithcrto been able to ) serve the Zeeman effect in a ficld of much intensity than 10,000 C.G.S. , as in all the ) with which we worked the light so faint the highel fields that it is not possible to obtain readings .
.\mdash ; The echelon , hich is by Hilger , ondon , has twenty-six plates ; the of is mm. ; the width of step is 1 mm. The difference in between two lines under observe tion is by is the angular displacement between the two lines that between two sucoessive orders of spectrum , S the thickness of one of the echelon plates , and the refi.active index of the glass for wave-length .
We have the ratio by means of a micrometer eye-piece fitted to the observing telescope .
The factor , which multiplies , was obtained in the following manner .
The refractive index of the glass , given by the maker of the instrument for the four lines , G. For it is Cauchy 's relation between refractive index wave-length assumed , namely , and in this equation were substituted for and their known values for three of the principal wave-lengths , giving three simultancous equations to determine the three , C. verified that these values of gave to less than a in the fourth of decimals the values of furnished by the makers for the four VOL. LXXII .
Pruf .
A. Gray and Dr. W. Stewart .
Radiation of [ Apr. 27 , lines , G. Then from Cauchy 's equation were calculated the values of corresponding to the wave-lengths 4000 , 4500 , 6000 , 5500 and so on up to 8000 tenth-metres .
From these values a curve was plotted which gave for any wave-Iength .
In a similar manner a curve was otted which gave for any value of .
Finally we calculated a stlfficient number of values of the multiplying factor , and plotted a curve giving the value of this factor for any wave-length .
This curve is shown in fig. 1 .
Arrnngem of Apparatus.\mdash ; The source of light was an exhausted tube , containing the vapour the spectrum of which was under examination , provided as usual with tsrminals and joined up in the secondary circuit of an induction coil ; the capillary part of the tube was placed between the poles of the electromagnet .
A convergent lens was placed so as to bring the light from the tube to a focus 011 the collimator slit of the auxiliary spectroscope , which effected a preliminary partial resolution of the light before it fell on the slit of the collimator used directly with the echelon .
This auxiliary spectroscope was of the constant deviation kind , and had the collimator and telescope fixed at right angles to one another .
The prism was a species of internal reflection prism , made up of two prisms and one right-angled prism , cemented together as shown in fig. 2 .
The light entered at the face , underwent internal reflectio ] ] at the hypotenusal face , and emerged at the face .
With this arrangement the part of the spectrum in the middle of the field of view 1903 .
] Hdium and Mercury in a Magnetic Field .
to be seen mder a sion equaI to that of a ray passing through a prism in the position of minimum deviation .
Any required part of the spectrum could be got into the middle of the field by a rotation of the prism-table .
The eye-piece of the observing telescope of this auxiliary spectroscope was removed , and the collimator of the echelon was placed coaxially with that telescope , and at such a distance that the light rays , after passing th ough the object glass of the telescope , were brought to a focus on the collimator slit of the echelon .
After emerging from the lens system of the collimator , the rays passed through the echelon grating , and were viewed by means of a telescope focussed for parallel light .
This telescope was provided with a micrometer eye-piece , whose indications , as already stated , gave the ratio FIG. 2 .
By means of a suitable rangement the prism-table of the auxiliaI'y spectroscope could be rotat , ed by the observer when king readings without moving away from his position .
In a similar way the width of the echelon slit could be adjusted .
esnlts.\mdash ; The first observations were made at right angles to the magnetic field on several of the helium lines , and on the green line of mercury .
The results were used to calculate the values of , and of by the formula , where is the of light , .
per second .
In every case the normal triplet was obtained , and the separation between the extreme components found to be proportional to up to fields of 10,000 C.G. beyond which it was difficult to obtain readings for the reason already ated .
The numbers obtained are stated in the following table , and fig. 3 is shown plotted against for different values of the latter , for the red line of helium .
A similarly close proportionaJity was found in the other cases:\mdash ; 20 ray and Dr. W. Stewart .
Radiation of [ Apr. 27 , ' Heliumcell ) ) ( yellow ) , 6678 ( red ) 5461 ( green ) Each result the above table is the mean of several determinations ; of three , seven , and five respectively in the case of helium , and of seven in the the mercury line . .
3 .
OOO 5 OOO lODOO Msgnetio S. , with che green liue of mercury unde ] observation , the field strength was increased to 13,000 C.G.S. , the centre component of the normal triplet was doubled , while each of the outel components was itself tripled .
The p of the two triplets and of the central doul ) was the same as that of the lines from which they originated , namely , of the lines of the normal triplet .
At all fields to 13,000 the faint companion to the yellow helium line was not tl.ipled , but only doubled .
For all the lines specified above , observations were made also along the of , one of the cores beingleplaced for this purpose by a core drilled from end to end with a hole about a centimetre in eter .
The normal doublet was observsd up to a field of about TO,000 C.G.S. , the separation between the two 1903 .
] and a Fi components found to be proportional to the field .
In this , however , it was more difficult to obtain relial)le results , owing probably to the disturbance of the field produced ) the hollow polc .
The following table gives the numbers ined : nbstance .
Helium .
5016 \ldquo ; Iercury The results here given for helinm are the means of six , and five determinations respectively , those for the green line of are the means of eight determinations .
The observations of Zeeman on the green line of , when the field was zero , are referred to above .
They followed an observation of Perot and Fabry , that the line to be a triplet consisting of the main line with a fainter companion on each side at about the same distance .
Our ) servations gave three companions on the violet side , ( we seemed at times to see three ) on the red side .
The values of for these are for the first time , and for the other two .
But , as has .stated , we have had what seem indications that this line system is of mexpected complexity , which is only disclosed under special conditions of the tube .
Hence , though these values of do not agree in several cases with those observed by Perot and Fabry , ib is not impossible sets of observations are correct .
It ought here to be noticed that Rung have obtained a resolution of the green mercury line into three triplets .
This observation is entirely confn med as to the side triplets by ours ( which were made before Messrs. Rung and 's paper came to , but we have not been able to verify their result for the middle group , which appears to us to be it doublet .
essrs .
Eunge and 's instrument was a large Rowland ting of metres diameter of circle , and the spectrnm was phed , so that their observations were , no doubt , more celtain than onrs .
omptes Rendn , p. 409 .
, 1 ) .
23

