The Fine Structure of the Line λ4686 of Ionized Helium Citation Chu, Djen-Yuen (1939) The Fine Structure of the Line λ4686 of Ionized Helium. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/bk2t-5f43. https://resolver.caltech.edu/CaltechTHESIS:12032024-221633745 Abstract No abstract. Item Type: Thesis (Dissertation (Ph.D.)) Subject Keywords: (Physics) Degree Grantor: California Institute of Technology Division: Physics, Mathematics and Astronomy Major Option: Physics Thesis Availability: Public (worldwide access) Research Advisor(s): Houston, William Vermillion Thesis Committee: Unknown, Unknown Defense Date: 1938 Record Number: CaltechTHESIS:12032024-221633745 Persistent URL: https://resolver.caltech.edu/CaltechTHESIS:12032024-221633745 DOI: 10.7907/bk2t-5f43 Default Usage Policy: No commercial reproduction, distribution, display or performance rights in this work are provided. ID Code: 16896 Collection: CaltechTHESIS Deposited By: Benjamin Perez Deposited On: 10 Dec 2024 20:54 Last Modified: 10 Dec 2024 20:57 Thesis Files PDF - Final Version See Usage Policy. 13MB Repository Staff Only: item control page

The Fine Structure of the Line A. 4686 of Iom.zed Helium Thesis b;y Djen-Yuen Chu In Partial Fulfillment of the Requirement& for the Degree 0£ Doctor of Philosophy California Institute of TecbnologyPasadeu., Cillfornia 1938 CONTENTS Introduction l Th~.,ory I. Positi.-ons II. Intensities 6 Previous Exparir~s 7 ~ e n t e . l Method· L AppartlW.S IL ·Experimental Difficulties tmd the Methods of Overcoming Them III. 15 17 Method of Meawrement A. Fomula.s of Fa'bry--Perot Interferometer · 20 a. Procedure Results I. P0ceitions end .Intensities _,50 II. WaT&-lengths 52 III. Compu.tation of Rydberg c.onstant R BS IV. Compu.t&tioo of e/m 54 Conclusions 17 J.cknow1edgements S9 The fine ~ructure of the line ).,_4686 of ionized helium emitted from a discharge tube with hollow cathode coo1ed by liquid air we.a measured with a Fabry-Perot interferometer. The formula for the Fabry... Perot interferometer pattern expressed in Fourier series was applied in the analysis. Of the eight CODrt')Onents predicted by theory four were definitely observed, while the remaining weak ones were inferred to be present. Relative intensities and positions of the eight components were found to be in general agreement with theory. The separation between the two stronge~ components was .0995 ¼. .0018 I.A. or .4529 • .0082 cm~1 cl.ii'f"ering from the theoretical value by less than one percent. The Rydberg constant of helium was caleulatecl with two di.fforent ve,luelil of indices of £rs.ct.ion,. e/m was also eeleulateci vri th both the old and tbe latest values of.' e;to.mic llUlSses. INTRODUCTION. The study' of the tine struo ture ot hy<'lrogen. lines and that ot ionized helium llnea haa been oont1dere4 eo lmportant eyer einoe the ad• 'ftnt ot the Bohr-Sommerfeld t heo:ry on atomic a.tructure. that the general t heor, ot at<>mio phe~na hu been 'built on it aa a oorneretone. Al• though many more 1 n·nu1t1gattona haw bMn devoted to hydrogen than to ionized heUUl)l the apeotrum of the latter 1• theoretically more favor• al>l• than that ot the former 1n aewral :rea:peote : tuat, ffe+ 1a hea- Yier thu B and 1 ta llnea are t herefore leea broadened by the Dol'J)ler etteet1 aeoon4, the He• aucleue haa twice t he aharge ot the B nucleus and th4t apeot1"Wll thereron la lees 1nfluence4 b7 th$ Stark etteat, t hird, helium 61hawa no nuolear spin and no oomp11eat1on will arise from that oauae~ an<l fourth, much greater than 1n the aeparation ot the tint et:ructure components ta n due to t he larger ,nuclear charge. theory ot tine ,structure wu firat confirmed by Peacben1 Sommerfeld•• in hie imaaurement of the tine atructure ot helium line•• Ot all the meaeured linee of t he , . Fowler aeriea an4 Pickering aeri~• the line 4686 wu the brigh test and lte struotun. eepecially that excited by oontinuoue 0U1"ren11 1 aooor4- Later, Leo2 worked. on t he tine ed with the theor., moat aat111:taotorU.y. atructure of the line 4686 a.rid obtained al.most the same r•ew.t. In the present work it 1• attempted to make e. fresh measurement on the aame line with a Fabry•Perot interferometer., 'l'h ~ instrument oe.n gift a greater 1. Pasohen F., Ann. d. Fhysik 2. Leo. W., Ami. d. Phyetk fil. ~ 76'1 901 (1916) (1926 ) pwoision in the determination of W[!ft•length and ot the aeparationa 'be• tween. the components than the conoa1re gratings auoh u that used by Pu• chen and Leo. It can alao enable one to atuc!l1' the relat1Te inteuitiee quantttatiwly by meane ot a. reoentl:, a.nelope4 method ot analyata3 eyen if there ie o"Nrlapping of different order• ln a pattern ot mob complex atructun aa the line 4006 • 'l"he aouroe ot illumination 1• a hollow cathode glow tub•• eaaentlally like thoae of Paachen and Leo. but so oeutructed that the cathode ts 1llitlll3ree4 1n liquid air in order to dlmtn• lab the Doppler ettect, ( 193'1) THEORY. Poaitiona ot the component•• I. 4 Scmmtrteld / applying the apeoial t heory ot relativity to the eleotron maa•• obtained the term value& tor a hydrogen-like atam aa tollowa : 'r • + 2 Rcx:. . ~ (: - ¾) + hi gh er orders of'cx. 2 (1) where R is the i~derberg oonstant Yary1ng w1 th t he mass of the nucleus , z the atomic number of the nucleus. n the total quantum number, azimuthal quantum aumbel", and oonetant. (X..... ( • 27Ce 2/ h o ) The firat term oorrespOndll to .Bobr 1 non-relat11'iatio meohanto•• k the the fi ne etruature I formula derived from In the eaae ot small z , the relattwty eorreetlon tor the term value beooinee (8) when the terms after the seeon4 or equation (1) are negleote4. t11th 'the intmucti:011 c,f el$C~o1'r :sP1.n'",another quantum j number ie used to apeo1ty the state ot an -atOPl in addition ton and k ( or { ) , where and !'rom a quantum mechanical treatment, Ueisenberg and Jordan 6 ha•• cal• eulated both the relativity oorrect1on and the oorreot1on due to spin... orbit 1nte:raot1on. The combined result ia 1. SO!rlllerteld, A., Ann. d. Pbyalk 51 1, ltructure and Spectral Linea P• 45'1 (1923). (1916) or Atomic T - n • J +1 3 i ) (3) 2 Thia 1a emotly t he aame e.e ~ommerfeld'a formula ts to be used instead ot k. Later. Gordon7 (2) • except t l1at j obtained t he same. reault from Dirao•a t heory ot t he ~ydrogen ltom whioh 1a now regarded aa the accepted method ot derivation. Stnae t here are in general two veluee of k ( or i ) oorreapond1 ng to eaah g1Yen j and l • ta really a eolncide nee ot two levels for j - l ._ • ~ k ponds to one Yalue of' k only. found by Paaohen in the line • • l, ~ j + l i As than w11 en each level cone•• a matter of fa~t, the ,component b;y t he· selection rule• AL• .!. 1 and (3}; • III 4 4G86 ie not permi tted by auob a rule. .! 1, uaed in conneotion with fozmu.la 'tion with the tormule. / • Thus more possible transl tions are pemit- 2 ted by the selection rule ~ J< eaoh energy level tor a j , ~j • (2) O, hut 1• l)ermitted + 1 ueed in conneo- This aeema to be a triu:ilph or the quantum mechanical theory over the older quantum t heory. r!'he struoture of ~he line between the levels for n • 3 and t hoee tor· n • • aeleotion :rules .t>L • :! 1. .i.j • ( ot same oo1ncident levels 4686 consists ot the trana1 ti one O/ .±. 1. by taki ng --- Br.te .. 109722.4 es shown in Fie • 1. j bat different ly in order to be d1et inotly visualized. 8 and permitted by the L ) are drawn separate• 'l'he Yi<:> l uen Ll T fU'8 ca lculated ex_ 2 • . 5320 • 10- 5 (1913) Gordon, w., Zeite. f. Phya. 48 11 a. Houston, w.v.~ Phf.• • Rev, ! 30 60 (1927} • (1935) 7. 0. .Birge, l-t. T., r.cY.f•.c1ev. 40 ·01s The ( 19Bf3 ) g It is easily seen t ha t i'rom t he 1:3 possible t r·o.nsitions one oen get components \'th ioh are d<:~ noted by a, b, c, d , e, f, g , and h. 8 Below t he energy diaf.:'. rem, these c0<--nponents a:re drawn eccordl np; to t 'heir t heoreti- cal poai tions CBlculated from ..c. T, and their relative 1ntena:i ties are r oughl y renresen.ted by t t e 1·e1gt:ts ( eee the f ollowi.ng section ) • states and t he trer..sit1ons between t hem, result i ng i n 5 '1:ne diffe:rent mul• liplets II. Relativ0 Intensities. Theore-t 1ca l raletive i ntansi-Ues of t he f'ine structure of were first calculated by Somr. !l8rfelll 0nd ··nnold 10 4686 from .Jhroedinger•s :f<>rni of wave mee!,anics, nnd t h e same result was obtai ned by Sahe. and Bei.nerji 11 from Uirac's theory i n t he form developed by Darwin and Weyl. Accordi ng to them, tl e rela t ive i ntensities of tbe coo1ponents or a multi- plet fo llow the s1.un r .u le, namely (PD} (DT' } ( :PS ) 1 : 9 : 5 5 : g ; l 2 : l ( DJf } 1 : 20 : 14 a ,d t ho relative i ntensities of' t he 5 mul tipl e ts are ( DI~~ ID ) 15.f' 35 J~ where 0 . 851, ~~~p ) 30-- a- .. 1.113, --C1111 0 . 025 , v..., :Y' J~1 15"Q. 3 u- In Fi 1;;: . l, t he 0 .221. i ntensities of eao h tra nsition ere i ndic ,ted i n t be middle ,i f tho above d iagram., and the numerical va L,ea of t he i nte nsities of the components are t-t1ven at t he bottom. 10 . Sor:imerfeld, A., and Unsol d. , A., 2.ffite. f. 1'h.:•e• ea as 2i)9 ( 1926) ~ ~:37 {1926) 704 {1931) PREVIOUS EXPEnIMmTS. The llne 4686 wu first cUaoovered in the apeotrum of !f -Pup- pte and •u thought to be a hydrogen line. Fowler in 1:a12, o\>taine4 a series ot linea, among which 4686 wu the strongest, by paaaing strong condensed d1schergea tl•rougb helium tubH oontedning hydrogen u an 1.m• purity, but they were not ooneidered as helium lines until Bohr5 found• ed hie t heory of' tlte hydrogen atom. Enns12 .tirst obtained the line from the apeotrum ot ~ helium tube free 1':rom it hydrogen, thua g1v• 4686 1ng strong support to Bohr's theory. However. J!'o wlerU pointed out that the line& obeened by hS.m were not aocsurately represented by Hohr'e formula. Then Bohr14 showed that the deviations could be aecounted for by takir-~ into accoWlt the mass of the nucleue which had been assumed to be inf'1:ni tely large 1n oomparieim wi 1:h the mass of an el"ctron in hie orl- g1nal paper; and Fowler 14 admitted t t e olose agreement betwe•n the theoretical and the obsened values. 15 Attenerds• l!'owler undertook e. further inTOat1gat1on and concluded from. the analogy wt t h apar·:-: spectra of other elements aueh u Mg. etc. toot t he aeries beeinnins with t he line 4686 and the e.ssooiated. Pickering series are on'tanced lines ot he• line Additional evidence that t he line • 4686 te a11 ionized heU.um/\in 11wn. oontormity with Bohr's theory was also gtven by Stark, Ra11., and Eva:rus 16 • 12 I:."vans, it.J'. t Nature lS Fowler, A,.• 14 !,foture 92 15 li'owler A•• 16 Evans, E.J • • Nat.ure PP• 92 (1014) 5 u 9fi 231•232 (1914) J~oy. t1oc. Phil. ]"rans. Hoy. ~oe, ? r--va •. ;~en. A Phil. Mag. ~ 284 A 214 90 426 ( 1915 ) 225 (1914} ; ( 1914). and t he referencer there . Before Somrr.ierfeld 's t ! eory of fi ne structure was publis!·,ed, Evan.a and Croxson17 , l' op1ng t he.t a knowledge of the etruot11re () f would serve as a guide in testi ng d1:ffere;1t h~tpotheses e chelon spoctx-oseope. current. the l ine 468 1) tor explaining 'Tt,ey found the line emitted by n condenser die• All the pho tof!r ophe ah:'iwed t l,at t his 11ne wae a close doublet wi t r· component s <1:f. nearly equal intenai ty • t he beet photogr aphs was 0.094 l. 'l'be separation measured trom After Sommerfeld'e paper18 on fine structure waa published, in whiot he quoted certain unpublished measure• me nts by Paachen on the fine structure ot t he 4885 aeries and a~e others, .b ans and Croxson re-examined all t heir photograpt•s and found e third com• :ponent. A few mont'ho later, ~esc1' en1 • publ1st:ed Ms wor k and _gave atrong sunport to Som.mert'eld 's theory. Instead or a n ecYelon which has t he die• advan.t-ge of overlappi nc;: of different otd.ere, he used a large eonoave grating having a diapere i on 2.z A/ mm. The we.ve lenet }'. s were meaeure<l with reference to otl1 er helium li nee calibr ated wlth iron standard of Huisaon end Fabry• and th e obsened v alues were correct with in a few thou- ea·,.dtbe of an Angstrom. was t he 11 ne 4686 The bett resolved line of s.11 those observed and ttia results were as follows • -17. .ii.v ans, R.J., and c . Croxaon Nature ~ 56 :,·:arch (1916) _,_ Spark Piotl.ll'e. Cont1nuoua current PiotUJ.'9. - ···-- --·-·--·--- - ---· -__ .,_ _ Probable Error Tlleor. Int, A.a,, 4686.050 0.01 Id 0.5 4&eo.oae Int. 'lbeor. A. air it, Id 0 II4 o.s s.92e 0.004 IatIId 2 5.905 Ia,II 0 'I 5.803 o.oo~ I .b Ile •• 7 t) • ~ Ia 'I s.eoa 0.003 II(a}b 7.5 5.VOJ Ito 6 5.774 0.006 Illa · ·1 5.549 IIa•b 0 e.•no 0.005 III0 3 5.388 !Ib 5 5.717 Ila G 5.687 Ji; 0.01 - faint but still visible. o.ooe 'l·•roba ble error ot ~ s:.r~lar than JI, III4 0 5.M9 0.005 0 .002 A and t hat ot IIIo o.5 s.397 0,005 t han o.OOl Ae IIIa(b} 4 e.307 0.002 II!< :t, AA. smaller hardly obsenable. Be aleo calculated th e t heoretiesl wave lengt hs of t '. ;e 12 components ( no Scheme 4G86. ,. 0 • m. m• n, n• s. l 2, 2 b 0 d 4685.810 5.63'1 5.890 e.ooo I l, 3 ·-----·- ---· 3, 0 I 2, 1 II 5e6M fh710 5.764 6.924 1, 2 III 5.3()4 5.331 5.384 5.544 where n, n• denote t he azimut ' al e.nd radial qua '.1tmt1 numbers respectively, ( so.me with m, m• ) • 'l be underlined components were obserTed in the spark picturei while in triEi continuous current picture, fewer omponenta were ob- aorved. Comparing these two e;roupa of ccmponents, one ma;y notice that the oompoMni o.30'1 did not appear in the centinuous current picture the component we.a the otrongeet 1n this picture &.ooo 5t'I03 e.oea in the first picture and (3) (l) (S) the coaponenta in the second were very weak• and (4} the relative intensities or the component• were different in the two p1o- turee, 1'rom the ,c,\ ot tlieae component•, :?uohen calculated the separation of the hydrogen «10 ,blet : 1 A 'JJH • 0,3545 !_ 0t0045 wbioh ie exactly the value required by !lo;.wnerteld • s t heOJ"Y, velue deviat es from thos~ obtained from th e direot 1neasuremente of the hy- In view of tluo deviation, Kun:.r.e 19 drog,-en doublet by other investigators. measured t he fine etr.ueture of Howe't'er, t111a with four different arre.ngen»nts ot 468ti apparatua and tour different l igh t sources and fouru\ it to be a doublet or a triplet. His results of five 1ueaaurements out of the sixteen possible ar• rarigements of so .rcee and apparatus may be aunJri1a.rized as follows . l . . Hollow oat:Lode t ul:ie with condensed diecharge IEollow cathode t ube th oontinuoua current. wt 1 ' _ ...,..,...,._ _ __ ··-·-·-·--1---·-·1 ·_- _Ilu ; ~ - IIlab~ · _ _ Ia=- - 1/!~-HU.ger eohalon ;i .0995 i .49Ql : Goerz echelon Hi l f1er LUJ?1J1er - plate l I .117 ---~~--.:...!.!.~!_, ___ -- .1025 - I i 0 ( in A ) . - - - - --- -- - ----·-·--··· . I··-·· -- ·-•- ·- ·-·----·--·"'· -----·-- • ... ·•·· --- -· --·- -- He also concluded t r·at t he component III 8 1, was not properly a part of t his group of lines, Leo2 • in view of both t r-e above mentioned deviation and t he occur• rence of a helium band at 4686 1 i nvest iga t ed t he fi ne structure or tho first 1~ t unze, P.. Ann. d, J)h:-i'e• 79 610 ( 1926 ) two membere of tbe li'owler eerie• •1• grating. ( 4686 and 3203 ) by mea.m ot a higt1 4:taper- Of t he line 4686 the structure was found as 1·01lowa : Qontinuo\ia Current Picture Spark Picture Int. \. air Int. ;\air 0 4685 ,9t;1 0 4085.922, ? .812 e , l313 .,10 6 .70Q 1 .659 1 .55'1 3 .,ue 3 .3-95 5 .309 ,, Tho compommt 44, • :-s1 sharp 4'18 5.309 b 17 r aaohen, poorl:1 localized ( IIIa ) • ffhich was supposed as due to the transl tion was also round only in the spark apeotrum,. 1t to a helium. band, 1n ~em.ent witb other inTestlgat:1.oru.. A JJH from t he mean separat ion.a Leo ascrf~ea He caloulnted 1n the oontinuoua current picture and in the epark picture• and got a smaller value : A~ • .3459 + .005 en1•l He also point"d out that 't be continuous curront piature and the epuk picture differ from eaoh other only in line braadtb and 1nte11Bity graduation and not in the number ot aomponenta. i~tter the papers by Sonm1e1•teld e.nd Unsold lO wero published, Pas• ohen20 re-examined l1is :previous \'lorlt by taking photometer curve ot t he con• ttnuous ourrent picture, and recalculated A~ ( the t heoretically prohibit• ed compouent Illa which had been taken as a most iP.Iportant factor in the earlier oaleulation was re Jeoted in theao later a&l.0u: at1ons. 'fhe laat r.aul t waa obteined 'by taking Leo's :meuurentnt in the oon.t1nuoue Ollft'ent ease only• Aa to releti Te 1ntene1 tiea, he pointed out the general agreement of hla auurement wUh the t heoretioel calnl.atton of Sonmerteli and Unaold• Finally, he concluded that b.ta measure1'!18nt of the a.beolute waft. length tn the eontinuow, current piotun 1s in acoor4. wt th t heol'Y' and that the aeparattona lead to the t neo:retloal ruue of t.'J{; • o.3M5Cn1-, Eouston8 tint UHd a Fabry-!'erot interferometer 'for . meuui-emant of t~ line ~86. Aa he wu tntereeted in the determination of DiJj ; ::.onJ.y- thei':. two otrongoat oomponente Iat Il:b were ~S.SU1'19d ( 468th803 and 4685.703 ) He conclude4 that those value• are not in eaeontial dieagreemem wt t h the ge-n- .ere.l scheme of Puehen's meUUJ'enent. lli.8 value of 8n,, 21 •alue is considered to be the most aocUl'e.te one. -- calculated from 8\\f.1~ I. • In Fi , ., 1s eh tu ter I, t tue w ich o 1ato ot , • er t tn- ( on t rtgbt side ot the · diagram ) e¥nd t he c1rculei t1on system. ( see Fig. 3 ) , The discharge tube stands vertically and t he l1gl1t tram its cathode is retleoted by a mirror M to the colllmo t cr slit. M' ls another mirrow, to be put 1n place o~ whttn 1ntena1ty marks 81'9 taken. Let ua describe the different p:arte tn the fol- lowing eeot1ona • ...\. Discharge 'rube. 'l'hia is a hollow cathode discharge tube cooled by liquid air• Th• ci:athode made o:f copper t ,:.:rma tho bottom ot t he tuhe ( J'ig, 3 ) , 1 ts tapering top being aealecl into 'the glaea tube. its 'tiop, It ie inmeraed in liquid air almost to 'l'he anode ie an sl.mninum rlng just above the cathode and ie 01' auoh a size t hat ·tt will not intercept the light trom t he cathode to the window ot ih.G tube, It 1• also oooled oofl8iderably by the liquid air vapor, '!'he di•• tanoe between the anode and cathode iamade as small u poNible because it waa found t hat in an old tube so conetrueted that the distance beWnn t ne eleotrodes waa Tery 8l'eat, the epaoe 'between them -.u strongly luminous, 'lhua probably affecting the relative intensity or the line 4~00 with reapeot to tbe othel' hel1wn lines. t\s a ·matter ot tact, this new tube gt ves the line 4~0 much brighter t b.an the old one both in abaolute intensity and ln relative inter.sity with reepect to the other lines. A leru, L ia -sealed on the top, qondensing tbe U.ght frcm the ee.thOde upon the oollime.tor eU t and serving u a window o.t t he eane time, The length of the tube end t he a1ze of 1ts top are so designed that when the eathode and altt are in conjugate poettiona, the solid angle subtended by the top of t be tube at the sltt is nearly equal to that subtended by the eollimetor leru1. The dia.11eter of t he cathode is suoh tbat t he image on tm slit is sufficiently large to secure fairly unii'oni 11• -- tu e 10 r !l 111 t tton. wi.,~ volt 1 i r1 e ot r- • • Ctroul tio y t 1 ot .o. • • to e up 1 ed • 1 c1roul tion ti 1 1f o -.1 11, 1 l t a t> • 0 bul 2 8B b t l ut 1 • t • ell la f1rat It 1 t enpu by the cbii.rooal absorber A2 cluring circulation wh ich 1s produced by the d1:t'f11• The spectl"\ll"l'I is quite :p ure u the eireult:1tlon ia going on or it a ion .PllllP• 1t hae been going on for quite a while ooton discharge, It is tound t hat u t he pressure is :reduced t he relative intensity of the line 4~86 beoomee greater as oompe.red to t he other lines. but t here 1e a lower limit of pres• sure ( a.bout . 2 - . • 3 n:r,, .. ) belO\P." which t he (Uscharse oeasee to atan. Du.ring discharge t he heliUN is abi,orbed so t ha t finally 'the discharge stops when the pressure drops t o about .1 um • . or below. 1• usually eum.,lied trm R1 Renee beto~ each elq)osure, helium to euoh a preesure t hat the diaoharse can be started. l1he interferometer plates are 11 cm. in dianeter a nd en half au- . 1 vered on the inside surfaces by evaporation. The1 are 114,ld ta a hollow cylinder with a separator which oonsi_ata of $ invar poet• held in a braaa ring. 'l1he parallel181ll between t he silTered surfaoea ie secured by adjuattng the preeaure esainst each other by means or three springs oompreased wit'h aoren. D, Spectrograph. 'l 'M a 1a a priem epectrogre.pb with a halt prism and a 2-meter camera. 'l'he r s tio ot the focal lengths of the collimator lens and ot the camera lens ts about 5.5 so that t h~ ali t 1:rnage en the plate is wide enough to leaeen the efteot of photographic grain. 'Ille opening of t he C8Ill$l'a la a.a om. high •o that the maximum 1no11natton of the beam la only 4.1 / 200; thus juat1fy1ng subetitutton of eos 8 by- l - :a 2/2 in our calculation. 'the 1i. Pbotometrto /\rrangement. I ntensity marks were put on ea.oh plate by meallS of the arra~ement abown on the right ot Fig. 2. ss was a aoreen wit~twelve alote ot length• from o.6 to 4-8 mm. and was illuminated by a mercur,. · d1soi- arge tube condensing lena Li• A oylindrioal lena L0 Hg and a situtated With uia horizontal at the conjugate position. of Hg, formed an image ot SS at the collimator slit s2 The light ••• retleoted by a mirror M' to the prlem and 12 maru, whoee in.• tenaitiee are proportional to the length of the corresponding alot,were taken on the plate. 1M marks trom the line 4916 whioh t• the closest Hg line to the line 4686 were uaed in taking the miorophotometer detleotione whtoh were thenpiotte4 againat their :relative 1ntena1t1es to give a aurJe tor :re• duotng t he m1crophotometr1o detleotiona s ot t he line II, 4686 to 1ntena1 ti••• Experimental Dlfticultt•• and the methode of overoom1ng them. It wu tounc.\ that the apeotrum ot thla type of diaohuge tube oon.• tatned a numbel' ot line• due to ionized oopp•r• Ou ot the copper ltnee bad a waTe•length 4682 ao close to the line 4688 that the pr1am could not •$:Pe.rate them. To avoid this contusion ot t '1 e interferometer pattern~ it . na found satisfactory to aluminin the 1na14e ot the cathode by evaporation. . Aft eNJar4• not oopper but aluminum l1nee appearecl, which however did not inter- fere 1'1 th our e:xperiment but had 8ome use t( see -·seet-ion d p.25) 9 B. Shift of Fringes. '1'be ohange ot interferometer separation resulting in a shift of h seemed to b4t due to meehan1cal cause• rather than termal. For it oan " be shown from the ooetticient of expansion ot invar t hat o. change ot 1° c. in fringe■ • will c r Ot & C t OU t t 1 te t ro 0 n r box ly exc edi on• • t in " co tlo _ w o 1- w 10 f lt t 11 or t e t 0 nt onl an zed. ehit 0 and t t i • by a1 e on and ere e w du tot r ct tat t t 1 co w a A. te of • 1 • 4 o a u·t. a t e e•i de on a seco tt r the th 8 ea 1 tollowi ot do int rtero nos 1ft or at leas Ter small• i t GXl)oeuree ot about l or 2 mtnut • w re tor th • r ot icturea . e C • pl te w • WO pi littl side is uttici to o or t 0 l 1 • 4 ) 11 ta of this o were late . one j et 08 oe a• tter 11 . narr w rt ttrst xp - rTa i on 1 two c •• t o. C.troular fringes aeeooiated with other etrong lines. At a tii-at eight we should expeot the image formed by the 1ntertero• ·•ter ot a narrow et:tip source euch aa a slit to cona1st ot only a ~ eee• tion ot the eonoentrio. circular tringea 11~• t hose shown tn Fig. ••• a matter of fact, but •• the whole oirclea tr1 t h a ftry bright narrow aeot'i on apl)e~d on the plate whe11 the exposure waa eu:f'tieientl..v lOft8• rh•• eirelea due to 1 the adjacent strong lines aueh as 5015 and 4471 oroaaed the line 4r886 and the:nfo• tntroduoed disturbance, T.beae oirelee were d~e to eoatter1ng d light, aa could be do:monst:rato4 by the taot that they beoame ve17 bright on breathing over the intertorometer plate and gradually taded down u the moia• tUN evaporated away. In spite of 'fer, oa.Ntul cleaning ot the interferometer plates "n.d eY&n replacing by a new pair or pla.tea eliminated. • th~ oiro~•• were never In the attempt to nduee the i ntenaity of theee fringes t r:e Wrat• ten light tilter M:ro, No, l hd been use4. Thia transmuted oQ.v 4% of t he light ahor-ter .tban 4500 A and t hus should eliminate the cireul~r tringea mm the line 4471• At the aeme time the Butman - 40 photographic plete wu eeleoted; because 1 ts aena1 t1 vi ty -tor t he long wave-length i i amall and thus the cirouler tringea trom the line 0015 appeare4 lese atro!1$11• Bow• . ever; the tilter also absorbed about 351& ot the line 468d ao that the ezpoaure ,raa neoeaaer11Y prolonged and the otroular fringes from the line 5015 proport1ona11, strengthened. wen The dieturbanoe .ot these fringes ia the great• eat aouree ot el'l'Or in th1e experiment; but 1t le bel1en4 not very aeriou ae no apparent dit'terenoe can be obaened rrom· the eom;pariaon ot the miorophotometrio tracinge ot two pictures one ta1-n with t he filter uaed and ■the - * The author w!shee to than.'!( l.i,1Jw• Julius T'earson tor t he polishing ot t he new pb:te.i. other taken without 1t • D, fiale.tion. Linea usually apread on both aides whe.n exposed sufficiently long. Sinoe the line 4713 the latter. is quite close to 4686 • 1ta spreading will i nterfere with To aTOid this, the plate ta pal nted with ".Eastman Opaque" on the glass surface. III. A• Method ot Measurement. Formulae ot the Fabry-Perot Interferometer. (a) The optical path dif:f'ere:'!.oe between two 1meceaaive rays which he.Te been reflected t wiee 1s 2 d coe ¢ where d is t he separation between t he platea and '/> 1a t he angle of' 1no1denoe. where n 1a an i ntoger. }tor a maximum inten s dfty, we haft 1 Let f be t he tractione.1 oder at 1nterterenoe at tho oente:r of' t he fringe s ystem and 1 its integral order. then i ( 1 + 't ) /\. • ad (5) For the mth and the ( m+l )th trJ-.-,, ea, we have from (4) • ( 1 • m + 1) A • 2 d ooa ~ ( 6) and ('I) Taking their ditterenee ,'ie 8G t or : /\. = 2 d ( 'X_ ~ d ( aoa ~ m - oos ~ -tl 'P:t, - f-m) (8) .F'rom (5) and (7) a ··d substituting t he value of \ _, from ( e) we get r (9) Let -Xin be tbe distance of t he mth fringe from. the cent•r or the ;:,attel'n, " haTe,ainoe Xia proportional to 4, •m t• (10) whic.b enables w, to oal c11lete f from. the meaeu.red re.d.U of t he fringe■, or aey multiple of them. (b) A is then oalcula ted trom ( 5) it d 1• known. !.et 1 be the wave nu.n.ber of a line, then tran (5) we have : f1tom which 1t ia easily der1Ted t'b.s t for two line• ot nearly equal wave-length, t heir separati on is given by .t;:.J) (o) 2 d From equation (8) we ••• that 1a a constant for a ~iven /\. (11) A f' <t>:+, - <p: and therefore X~+i - X~ '·his latte,; quantity is called t be period ot the 1n.terfeNmeter pattern. because it t he intenai ty • be plotted. with x2 aa abao1sea, tt will be a periodic tunction of the absotaaa with p.r1o4 equal to x;t,-X~ ( Fig. 5 ). Moueten2'~ 1ih1a p~perty to express the pattern u Fourier aerie•• of pe:riodieity auggeats Prof• a I~t called angular or«er ot :t inter:ferenee, be taken as the var.i abl•• then the Fig. 5 patten oan be eXpreue4 a• , (12) e. is the position ot muimum intensity and r 1a t he refleot1ng power where or the .silvered eurta.ce ot t he plates. 'tftls is true only tor strictly :monoohro- broade ning, the pattern for s eJ ~~.,.le- line may be exp:ressed by 23 ; (13) where: i,I _ .Do - _.,,,3/z ✓L ' B,., ·- 2rc ~r~e-'>'l.~o,. ... .,../3CCIS ne" , l - 11 A~= 2n 11i . r e-~ct--Xfl SIY\·n&o , and ci... = l!1.. ( L)2 which is due to Doppler efteet, '-RT +n-d , f: fa .= natural r alf 1 ••;t enet ty >Jraadth, which 1s .58 10 ""4 A ( or 41rd x 5 r,.1dians whe1>e 0 A ia tn A) in t he elauie&l. thoery srd about t be eeme order of. mcgnHude 1n the quantu.."11 t heory, M ,. moleoular weigr;t of sm.i tti ng gaa :; li • T • eu co•: atar1t J absolute te--:nperature • Since e ·.~ay be eh mged ~, e.ny integre.l multiple of cal. tunetiona we ean eount 1mum i ntensity. e 2 ?'t' 1n the t rigonometri- from tl-e abse1aaa which co.rreaponda to any- max• 'I'be the formula (13) is simplified to i (14) If t here e:re eeYeral components of nearly t he same wave-length, the pattern must be a s·u_perposi tton of ;~atterns given by (14) with 10 pro• portion:ll to t lH.9'.r btens i ties and w1 th ~wmu..<n .:ositiona at interval.a pro• 1 - 2:-.s. Jee formulae (l) - {4) of Nfe:renee z. portional to their se,paz•at1ons, since (15) (B) Procedure. a. B'. xpoeure. Before mllk ing an exposure, t he heli um was circulated t hroi.1€h the ohe i'~«il until ell traces of &ir G. rd other i mpurities had disappeared from t he epectrum. ae obeerved with a. direc t v1a1on spectroaeope. The current ust'Jd was around 140 m.a., as t oo sn11ll a. current would necess1 tate .long exvoaure and too large e. ourrent would i ncrease the lin.e breadth, time ot •~oaure wae 4!J mi nutes for t he plate plate 116, 50 minutes for t he 134 and 6? minutes for the other two plates, tion II (li) The As mentioned tn Sec- P• 17, two short exposures were usually made on e second plate to detect a shift of the f r inges. 'l'bis plate . also senee tor ~he measurement o:t other helium lines in o:-de:r to detennine t h~ 1nterterometer separation, for the 1ntenalty of other lines 011 the maln plat e was eo atrona that the fringes were not clearly defined an4 t l:eretore could not be measured accurately. be Micropbotameter Treeing and Mastor cune. For each plate to be measun4, a m1orophotometr1c tracing waa taken of tbe line 4686 with tho .Koch --Goos p1'otoelectr1o miorophotometer ( A. Kl'\188 Co., Hamburg ) • The magn ification was us ually e. of 1 ts tracing 1e sho,m in Fig. 6 A SJr.aple of the line and ( here t'he treeing ie of magn1f1c,J t1on 1 ) ; where four componentc are disti nctly seen. Both coordinates of the tracing were then messu1"11u! •1th t !l e comps:re.t or,: and the ~ deflections ( ordinates ) were converted int o 1nt ens1ttee by meane of the de:f'lect1on•int ens1 ty Ctl.M'8 .. - • 01" ot w • . •1 l • 1 0 • • 0 . uo • C 0 calculate the i ntensity distribution of a single component, by means of ~ formula (14) in which r and CL were tound from another single line tallowing section ) and fa was g1Ten tbe olasa1cal val ue. ( aee the It iG so etnell that it is inmater1al whether the o1aea1ool or the quo.ntum maci,a.nical value is tween. Of the 1nfin1 te aeriee only the first tew terms wen used, e terms in the ease of largest CL and l U terms in ·the oaae ot S!lialleat ():. , beoauae ill the tollowirig terms omitted oontribut at most 0 41 001 o:t' the max1,11um in• tenai ty whioh is wi tbln experimental uneerta.1~. ~1e valuu ot e tor wbioh t he function was calculated wen taken e.t intenale of 5c or 1.5° . depend• ins upon t he period P• 'l'be assumed relative intensities gave the 10 •a• l.'he pca1t1on of each oanponent was calculated from the assumed saparattone bjr meau of equation (U:i ) • 1'or convenifmoe or faleulat1ca, each poeition was uaue.111 replaced by the nearest value wh ieh ta a multiple of a 0 • 2° or i.5° • 1'he error thua introduced wu &till within exverimental uncertainly. After . getting the tntenaitiea tor ell component•, t he t beoNtical ?&tten was eaeil.y drawn and compered w1 th the :& expe:r1mental roe.ster curve. After aeTeral trials, a pattern was finally obtained which n tted the meeter ourve best. ( See :ti'tg. 7 P given in Table I d. It ts 1n this way that the 1ntens1ttea I and podtiona were ohtained tor each ~plate. D&termination of 0(. end r Stnoe t he structure ia ver;ir complex, 1 t ~•1th:J.d be e;;;tro:n,:J.y tecUous 11" besides the inteneitieo and positiona we hove to change also trying to get a pattern to tit the master ourve;e determine d., and r, from another single line. a end r ta Bence it ie desirable to For t hb purpose. an alminum In each of the :fi gures the upper- t hick curve is 'the e::i.perlmen:t.ru. • mast er curve.? the lower t rdck curve i s the theoret,ical one obtainpd Uff swntlling the t,llin curves rep:ros~nting thc1 ei ght oompom~ntw a;o,c,a., e, f' , gph.. Of' the com:pohen-t curves only a part of t,hc whol e pe:d.od is shown,, e~ce:pt t host~ of' the t wo strongest componouts cl,f. • Fig.. 7 A. For Pl a-t e 118 Fi g. 7 B For Pl ~te 124. Fig. 7 C. For Plat e 126 Fig. 7 D. For Pl at e 1~4. line 4665 of about the seine intc~nsity and wave-length as 4086 we.s selected. e. rietermilui.tion of d. The interferometer separation dis determineti f1>om other helium LJ.471 lines 50<18., 5016, 4922, 4713., ·"4457, and •1588. order of interference. Similarly, A 1 , /\.2. ckmot,e:J those of th0 ot.her lines. Let 'A. be the vmve-length of ••.•••• ., 0 eil'ld f ,:rQ ••••••• Then we hEi.ve: from which we get; (16) i1 + fl := ( io + fo ) ~• I From the appro::d.nmta w.l.uc <..lf d mea~ured ,nth orclinary micrometer i 0 can be found e.ppro::d~tely. Using thi~ wJ.u.e ood. Am ( m ::: 1,2, ••. ) the orders of.' interference of all oth,~r lines can be ettl.culated By equ,i:t'tion (16). If these ceJ.c-uleted i'm &re not equal to the ob$0rvcd values, 0th.or values of 1 0 are tried until they are nearly equal. The le:et tried v&l.luEi o.i' i 0 and the ob1:ierved f 0 give the value o.f d from the (:KJ.W?.tion (5) namely ( 10 + f 0 ) /\.o:::: 2d. f. Ce:J.cula.tion of we.ve-lengths. Wavt-::-lc!ngth is ue.ue,lly eru.~te.cl from cl, by equation (5) so that its aeouracy d~p<mds upon that of. d. In the lest sec·tion, only D. row..,h way of det ex-i'Ilining . d W(tS given for the change o:f phese upon r~flt?ction was neglected ruid tlw stmidm•d wave-lengths were giveJ:t for sta1wi.rd conditions, ( u,0 c. and '700 tnm. of Hg ) while the conditions under which the oxpo~"UrS •28- was te.!ren wen usually different. 'l'o m•ke the neoeesary eori-.otions t ~·.Mtoh efteote let us conaider a standud line ( subsori:pt o ) and a line of Uif• knon we:,e-length ( aubsoript u ) • If ).,, , A\L be wave...lengtb at standard eondi tiona ( 15° c. and 760. mm. Hg ) 'b -,,._o I SA\l be their change due to variation of temperature and ;>resilure ,\ rto, nu. be the obnrved oder of interf'ereMe b' n01 8n"' be correction to be added on aeeount of cbnnge of Ji'ose upon N• neotion, we should have the :f:'und~rnental equation or t.he intarfero1:iete:r; (17) or neglecting the.seo0ud order, • ' • noAo+nJAo+l\~Ylc= YluA\l+ YluS)..ufA.IL'Bnl4,. (18) • Since S.:\. may be expressed in terms of S Yl by t he relation the lut two terms on each side of ( l t: ) may be combined into one term, although their origtne are different. • Hence we may write ; (10) mere L> n denot•• tl1e ool'l"ection tor both causes. Let .DJ b$ a value defined (20) . then by subtraoting fl'pm (19) we have .A.,A'Ylo. (nu.- n~)AIA. + flu.LI n~' or (21) (22) nu and nu ... ~ , So 1.u.ean be calculated trom s1noe no· and Ac are ktWWn. The fractio nal part of Du 1e t he observed value and 1 te integral pert oan be oalculated from t he reletion Yt.J1).• n).0 efnce Av.. ts e.pproxims.tely known. find To nu - ~ , we b.ave to use other 11n.ee of known waTe•lengtha ,A., A1. ...... For trom. ( 21) we have s •••••••••••••••••••••• The 'Value on the left aide ot each equation is a difference between the ob• aened valuee and thoae calculated trom (16) ( we mu.at remember t hat 1 • t there oorrea»onds ton• hc,re) or simply tho dttterenoe between the observed t and ealoulated r. 'i'he value on the right eide is un.1'novm, but evidently it is a funet1on of riave•length. e.N pl.t)tted against Ii' t hese ditf'erenee• Am, t he di f ference llu - ~ tro:m t}ie C'il'VO and Au. can be culou :i fited from ( 2.2) • I'm • nJi. ( m • 1,2 •• ) tor t he unknown ean be tound Thie value gives t he wave-length dire~tly in 1 aiL• nc. matter under what ~ondi tione the exposure was made. ll'or the value 0¢ J\. for eaoh of A.,,\.. ••• wl ere f 0 we usually ealeulste tbe product ( i + f ) A 1a the o . sened value plus the aorreotion from t be curve Juet menti oned and is Tei,,· nearly equal to the calculated 'Yfllue. The mean v~~l ue of t hese produata wu U1Jetl for the ealoulat1on ot wave-lengths, I. Positions and Intenait-ies In this experiment- four plates o.f different inter:feroineter separation~ were measured.. •.Because of ove.r lapping not al.l of the eight components cou.ld be observed. On plate ll8 .a nd 12.1.1, the two strongest components f,d were observed together with a bump which corresponded t& the posi ticm of the third component. h. On · plate 126, a t.hird maximum appeared. which corresponded to the position of the fourth •com:ponent g, while h gave neither meld-mum nor . hump because of i t.s pom ti.on lying between the tw.o strongest comporu.m.t. On plate l S4, o£ snalles.t separation, . . 7D or the four eompouent.s d,f,g,h, were mruufeatly observed (Fig. 6). . The " existence of the four weak components a,b.,c,e could be inferred only by the fact that the theoreticsl. pat.terns. deviated too much from the mast:er . ctll'Ve$ when they were aa$Wil00 absent. Of these four, c and e could not. be given too great intm'lt;ities .alld were, theref'oria, alwa.ys r~St.Utted to have the theo.r etieal values tbrougbout the aruuy-ris. After several trlru.s fo-r ea.eh :phte the relative intenai t i es were . found tG have the values giv:en in Table I. Of' course these values ~ not wdque ones; since the intensi~y of one component wu so::.relited to:its neighbors that increasing the one imst deereo.se the ot-her in order to tit the obsenational mastex- CU1"\f'e,. For eumipl~, the couti:xment a :WtU3 very close to the component h .(next orde.r) on plates 124 , lM, to the component • g on pl.ate 126 , and was bet·ween g~.h on plate US. Th11s its tru.e intensity might be different frora the values given in Table I. For the t.wo s-trongest Table I Rel.ative Intenaiti:es -and Posit1ons Plate d p C .l!lll1 ,Tiill of' Hg ma. a b I 1! . 07 . 06 C d e f I. 118 l 1 . 00 . 09 .ZSB-. 29 140 ~ ;, ._~~- - - - -- - - - -- - - - -- - 5.9814 Ii P l - !I .15 .31-.26 l 1,ts 124 5 .. 9'A,~i l .lO . 10 . P, I O .27 • 7 !:>93 l .. 4907 .ro 1.00 .ll -. 4555 l i. ! : - 1 l.26 2.:9fJ'/ 8 . 55-:-. SO I .10 145 p T~~etictl I T1J~f~tical P 1. 00 .m, -.4393 ,. •0 8 lW 1 -+ - - -- - - - .04 . 05 '• ..:;. ,. . • .os .041 !) cm - 1 In the second column are given .t he interf'erometer. third, the preasure.s both. Ett the start and rit the end of exposure; in the ;fourth, the cur-.t,en·t . I denotes intensity, that of comy:Y.,ment f being taken us l.00. The :t heore-ticaJ. fut e..".lsit.iea ttre also given in t he second row from the bottom. The theoretica.l' positions are given in tf'4;l, last row, tll r eferred t,o f . P dr::no.tes : the position.. When no vali.e i s ~iven, it means thet the position i l l the is the same as pr edictE.'Cl &.f the:ior-y or as2u.med to be ro. components, f,d, however., the int~nsities could be asserted with cer- As to positions, those of the components a,b,c,e could hardly be determined with certainty, and tbe--,f were 6SS\lut€,~ to hav-e the tlleoretiecl. ve.luesin the analysis-. .The· compo11~ts g,h were found to have separe.tioos from f greater t!um those predicted by theory. The two strongeat component~ were f0:und to have a sepaz>ation in agreement vd.th theory, except in plate 126.where d was too , close to r. Tbiei was probably due to experilnen~,l error, for the maxima corresponding to tbi$ component were uncert&.in II. on the Photometric tracing.. Wave-lengths.. • . . . . .Only the wave-lengths of 5 compolients, namely d•.f ,g could be calculated from the maxima on the intensity curv~. Those of the· two· stronge$t co~ponents ~ been _accurately determined·and are given in · fable II. They were fir.st det~irmiood from the m,axi:m&, 0;ue corrections o-£ fractional order of' interrerenee ha'irlng been given 1·or the chffllge of'- phase upon reflection and· for the deviation of temperature and.. pressure ' from s ~ conditions. ' ' .' The ·p1-obable errors vrere estimated from the . uneertainties of the µosi tione of ~ . . Then another eorreetion due to the displacement. of ~ by other components were taken into aecount •• .from the :theoretical patt.erns. . The coluroncS headeci by cor. in Table II give this sort o:t correction. the weighte were-, taken f~n the numbe:r o r ~ from which the wave;-lengths were calculated. . . . .• The wave-length oft.he cuponent g could be ealeulated only £rom t wo plate$. It is Ag== 4685.5330 ! .0050 from •plate 126, and l. g = 4685.52.66 .!: .0050 from plate 134. I . . llav-e-l.engths,. I. A. I A. d A.r Plate 1------'----- ~ - - -- - - . , . . - - -- ·~--~ - - - -----·--·-·--·-··-- - - - -- - cor .. wt. from max. 0 124 • 1001±.0015 + . o o u • 101~~001s i • . a o s 1 , : t ~ o o 1 0 -~001s . 8 0 1 6 ! ~ 0 0 1 0 • 701:9±.0011 . 2 • 796~.0020 .00:18 • 7980±..-00:to · 1 • 7052!-.0016 -.002a •1024±•.001& 2 .. aoi~.0014 .0012 .ao~.0014 4685. 8012 + .•0012 The probable errors we-re gi-eat bees.use on plate 126 the :r.,-osition of the mmmum eorrespon<ling to g ls considerab.t y effected by its neigh- boring component$ a,b, and on plate .154 the maxi.IJJ.lfilis too broad mx1 uncertldn. Computation of Rydberg Constmit RHe. In reducing the wave-lengths in I.A. to thoso in vacu.um, both the .formula o.f Meggers antl Pe:ters24) and that of Bencier25) were used. 24) 25) s • 7049±.0017 -.0029 4685 .. '7017 .± .0014 III. • ~;!aggers & Peters, Nat. Bur. ~ - Bull.. l,4• 697 (191.8). David Bender, Phys. Rev. 54, 179 (1958). · 2 From the wa-ve members the values of' Rae were calculated by the formula Table Ill. l. 2.. (u .- J.)•107 Computat.ion of Rae . = 2(89.9 from Meggers & Peters A_,dr · (u -1.>1,. z·el. eor, ft.de -4&85.?011 l.,S072 -.6!;50 109722.425 + .032 ,4685.80i2 ff •.1974 109722.437 .± .028 Mean 109722.450 + .0:50 ") ...."o 7 -- ~u ·· (u - ... ,-.l.J • Ae.1.r (u iz ·•.;; - .p.,.;._ 'A-~ ..:.i.vm ~a.er I)\ rel.. cor. Rtte 4685.7017 1.51.75 .6550 109'122.182. ± .032 4665.8012 l..Sl75 .1974 109722.].~6 + .028 Meen 109722.189 ± ~030 The uncertai.nt.y of the mean i.E not the m.e&m deviation but the me2..n of too ind.ividual uncertainties. I.V. •. Calculation of e/m. From tbe difference between the BJ.db.erg constant for Re+ and that for U the value of e/m •OlJJ.1 be ealcul~tt.">O. by the fe:rmula26} 26) Birge, R. T •• Phys. Rov. Supplement b P• 46 (1929). (24) where F is the Faraday confitsit wnich i .s 9651.l + o.a i f the atomi.e masses H and He ere eJi.-pressed in A.¥to.n scale. 27) For Rs, P,rofesaor Houston's value, ~ , 1O9o'17.759 ± .01028) is used. For ·RJie, we should use the Yeltu~ obtained by ueing the index of refraction deto1"!mined by ltteggers and ?eters, nmaely 109722.450 :t ..oro, since RH was obtained by using the seme inde1' of refraction.. Then Rffe - RH = .44.u'll ± .054. H = 1.,.-0<)612~}.. Subati tuting these valuefl in ( 24) we get. e/m = (l. 759G ± .0013) x 107 e.m.u. where the r,robabl.6 error is dete1'ruinod from. the probable en-ors of I i" we use the old val.uea o.f atomie ·masses as used °blJ Shane and Spedding27) in their calc.ulation of e/m, namely B ~ l..007775 27} Shane, c. D., md Spedding, F. H., Phys. Rev.. 10.., p. t>S (19S5) 26) See ref'erenoo 8. The uncertainty is given by the mewi of the uncertainties of il'ldi.vidual }qi obtained from the lines 6565 and 4861 instee:d of by the mean deviation. which is .ooa and , a.e pointed out by Pro£. Houston, can hardly be taken as r epresenting, the preoision of the value of R. 29) Aston, F. W., Roy • . Soc. P1'0c. 165A, ~91 (1957) . which was determined i,y' Kenneth nth reference to Ue = 4.00216, then we get COl':WLUSIONS Of' the eight- components prooi~ed by theory,. four h trve been definitely observed, :ma the exi.etence gf the remaining we~er o-nes ean be inferred from the master curves. accord . The position$, in general, with theory, esrieeially for the two strongest eamponents • . The . relative intensities are different for different pJ.~tes and not in precise agreement nth the theoretically calculated values• yet the deviations are not more then cen be attributed to the eoridit.i ona _of' excitation and to photo.metric .e 1TOrs. The too high . :.nf :.;m,it,-1.ef of t he eonrpon.ents g,a,b .suggest t-he hypothesis that 4p state is :aore populous d.ue to the absorption of radiation ls - i1p. On the other tumd, eoDll,on,.. ent h does not seem to be sufi;ieiently strong .for this to be correct.• 'fhe unueutl.lly hir,h intensities of f!Om?OO.ente g.h. on plate 134 are perhs,ps associated with the low presaure used in the same way in whic!:l . corresponding v~triations of' relat:i.ve intensity with pressure are observed in hydt"Ogen. BO) That the inteneity of £ is al.wfl.ys e'TO!l'ter then that of' d i s in agreement with Pasehen 1 s observation rather than with that of Leo ·wno found them equal. The observed •.separation between the two stronge~t. components . vthich i .s · 0.0995 I.A. is in complete agreement both wit,h theory within experimental. en10r and with th&t obse1~ by Prof. Houston, while the vel:u.es obtained by Pasch.en, Kunze, and Leo are respectively 0.106, O.l023, 30) • See 1..efe:rence 27) • value obtained f'rom the raicro:photometer CU1"V'e, .098, is a little lower than the theoretical vru.ue. In this present me&Sl.ll'elflent. it is believed .that the value observed is more accurate th.an all the r,revious ones not only because of the accuracy. which is usutlly attained by the interferometer but aJ.m because of the careful consideration being The Rydberg constant is l&rger tllan the value obtained b'-j Proi'. . . .Houston.. This is bee&use the wave-lengths are smaller than hie - values. This di.t'ference, however, is less than the difference bet.ween the Y.alues determined with the differe:a t indi.cea of refraetion.. The value of e/m obtained. by 1.1eiug the latest v-alues of atomic masses io in close agreement with Dunnington•s. value wbo obtained 1.7597 x 101 for e/m from the deflection method•.Sl) ·· - - In ec:.mclu.-1?.ic,n it is worth noting that the obsened. separation. c:orrespends to a. value of - • 1 - .f or the 1>.r..,..e structure con.stsn: _- t. l.o?.4 ± 1.2 - ,L,1,U 52 31) Dwmington, F., G., Phys. Rev.,-; 475 (1951). The t/U.tho:r wish-os to expr ess hi s d.oe1:,c st gratitude to Pt'Ofessor W. V. Houston for his oogg&f;tion .of the probl0m, c.::.rust&nt_C:J.rectioo. f_;.ncJ. h 0lpf ul ii,' Uidance throughout t,he COUl.' Si;; of the v.-:o rk. The r-n.:rLl:10:r eloo wishi~s -to thW'.lk Proi'o s-sor I .a S. &r~,on for his many vru.u.able suggestions.