CIHM Microfiche Series (Monographs) ICMH Collection de microfiches (monographies) Canadian Institute tor Historical Microreproductions / Institijt canadien de microreproductions historiques Technical and Bibliographic Notes / Notes techniques et bibliographiques The Institute has attempted to obtain the best original copy available for filming. Features of this copy which may be bibliographically unique, which may alter any of the images in the reproduction, or which may significantly change the usual method of filming are checked below. D D □ Coloured covers / Couverture de couleur I I Covers damaged / Couverture endommagee Covers restored and/or laminated / Couverture restauree et/ou peiliculee □ Cover title missing / Le titre de couverture manque I I Coloured maps / Cartes geographiques en couleur □ Coloured ink (i.e. other than blue or black) / Encre de couleur (i.e. autre que bleue ou noire) □ Colou PlancI red plates and/or illustrations / lanches et/ou illustrations en couleur □ Bound with other material / Relie avec d'autres documents Only edition available / Seule edition disponible Tight binding may cause shadows or distortion along interior margin / La reliure serree peut causer de I'ombre ou de la distorsion le long de la marge inlerieure. Blank leaves added during restorations may appear within the text. Whenever possible, these have been omitted from filming / llsepeut que certaines pages blanches ajoutees lors d'une restauration apparaissent dans ic? texte, mais, lorsque cela etait possible, ces pages n'ont pas ete filmees. a Additional comments / '"'' Commentaires supplementaires: L'Institut a microfilme le meilleur exernplaire qu'il lui a 6te possible de se procurer. Les details de cet exern- plaire qui sont peut-etre uniques du point de vue bibli- ographique, qui peuvent modifier une image reproduite, ou qui peuvent exiger une modification dans la metho- de normale da filmage sont indiques ci-dessous. Coloured pages / Pages de couleur I I Pagei damaged / Pages endommagees n Pages restored and/or laminated / Pages restaurees et/ou pelliculees Pages discoloured, stained or foxed / Pages decolorees, tachetees ou piquees Pages detached / Pages detachees /| Showthrough / Transparence D □ Quality of print varies / Qualite inegale de I'impression Includes supplementary material / Comprend du materiel supplementaire Pages wholly or pa-lially obscured by errata slips, tissues, etc., have been refilmed to ensure the best possible image / Les pages totalement ou partiellement obscurcies par un feuillet d'errata, une pelure, etc., ont ete filmees a nouveau de faijon a obtenir la meilleure image possible. Opposing pages with varying colouration or discolourations are filmed twice to ensure the best possible image / Les pages s'opposant ayant des colorations variables ou des decolorations sont filmees deux fois afin d'obtenir la meilleure image possible. Pagination is as follows: p. [173]-199. This item is filmed at the .eduction ratio checked below / Ce document est tilme au taux de red'jction indique ci-dessous. lOx 14x 18x 22x 26x 30x 1 ■v' 12x 16x 20x 24x 28x The copy filmed here has been reproduced thanks to the generosity of: Ecole polytechnique, Universite de Montreal, Bi bl iothcque The images appearing here are the best quality possible considering the condition and legibility of the original copy and in keeping with the filming contract specifications. L'exemplaire film6 fut reproduit grace i la gdn6rosit6 de: tcole po1ytc';hnique, Univcrsit.c de Montreal, BIbl iotheque Les images suiv?ntes ont 6t6 reproduites avec le plus grand soin, compte tenu de la condition et de la nenot6 r*c, rexemplaire film6. et en conformity a.»>c les conditions du contrat de fiimage. Original copies in printed oaper covers are filmed beginning with the fro -rover and ending on the last page with a pru.c i or illustrated imprai- sion, or the back cover wr .1 dppropriate. All other original copies are filmed beginning on the first page with a printed or illustrated impres- sion, and ending on the last page with a printed or illustrated impression. Les exemplaires originaux dont la couverture en papier est imprimie sont film6s on comnencant par le premier plat et en terminant soit par la dernidre page qui comporte ur.a empreinte d'impression ou d'illustration, soit par le second plat, selon le cas. Tous les autres exemplaires originaux sont filmis en commenpant par la premidre page qui comporte une empreinte d'impression ou d'illustration et en terminant par la derni^re page qui comporte une telle empreinte. The last recorded frame on each microfiche shall contain the symt-ol — ♦■ (meaning "CON- TINUED"), or the symbol V (meaning "END"), whichever appliab. Un des symboles suivants apparaitra sur ia derniAre image da chaque microfiche, selon le cas: le symbole — •- signifie "A SUIVRE", le symbole V signifie "FIN". Maps, plates, charts, etc., may be filmrd at diffirant reduction ratios. Those too large to be entirely included in one exposure are filmed beginning in the upper left hand co.ner, left to right and top to bottom, as many frames as required. The following diagrams illustrate the method: Les cartes, planches, tableaux, etc., peuvent etre filmis ^ des taux de reduction diff^rents. Lorsque la document est trop grand pour etre reproduit en un seui cliche, 11 est film6 d partir de Tangle supirieur gauche, de gauche ^ droite, et de haut en bas, an prenant le nombre d'images n^cessaire. Les diagrammes suivants illustrent la mdthode. 1 2 3 1 2 3 4 5 6 I DEPARTMENT OF THE INTERIOR Hon. \V. J. HocBI, MMtltr. CANADA W, W. toKt, C.M.G., Dcpuli MM>ttr. PUBLICATIONS OK TllE Dominion Observatory OTTAWA ^^^^^^ W. F. Kino, CM G. T.L.r)-. Director. Vol. 1, No. 7 Experiments regarding Efficiency of Spectrographs BT J. S. PLASKETT, B.A., D.Sc, F.R.S.C. 0) CM MT UO-} POl-07 t ) T l' A W A COVERNMENT PkINTINO BuUEAU l<,tl 1 toJd HULYILCHNIQUE BIBLICTHEQUE icou pomtma. DKPARTMF'NT OF THK INTIIRIOR CANADA II. s W .1 l;,H HF, Mi»iWrr \\ \\ ( nnY, C.M.O., Drpiiir Minuli^. PUBLICATIONS Dominion Observatory OTTAWA W. F. KlN(i, C.M.C. I,I..n., Diriclnr. Vol. I, No. 7 Experiments regarding Efficiency of Spectrographs v\ BT J. S. PF ASKETT. B.A., D.Sc, F.R.S.C. O T T A \V A linvF.RNMENT I'lUNTINi; BinKAU 1914 . -! I =. I i ■V/.,.jj-. /J L/ i;xri;i;i.\ii:.\rs ON .\ii;riini).- loi; :\( i;i:.\s;n(, tiii; i;i i kiiacv <>i -n.( ii!()(.i;.\iMi> IN i!.\i)i Ai. \ i;i.(u irv i)i:ri:iiMiN.\'rFoNs BY J. S. PI.\>M,II, H, \ , I). SI K.H.S.r IN riiout I I ION. Tt is !i well kimwii f;ict ami Iim- liccn repeatedly stated that the modern star spectni^'rapli i- e\eee(liiii;ly wasteful (if liiihi. If wv leave out of consideration lor the prcM.nt the linhi oceiilti^il hy the -lit jaw--, careful computation ol the losses hy alisorptioti and rellertion in the optical train shows that at 11-^ alioui _'.'> per cent, of the litiht is transmitted liy :>. one- prism atid ahoiit N |)er cent, hy a three-jirism spectrograph. The loss at the slit will depend u|ion its width, the focal lei^th of the t(>lescope ohjoctive and the conditions of -seein^i" and will \ary from .jO to !)() per cen;. If wo add losses in the ohjective of the telescope, we ^^et from '_' to S per cent, of the incident star lijiht in a one-prism and from ti to .'! per<'ent. in a tliree-prism spectrograph eflVctivc in producing the siM-ctruni. This was recognized at tlie meeting of the ronimittee of the Astro- nomical and Astrophysical Society of .Vmerica on ( 'o-ojieration in lladial Velocity Determinations at Mt Wilson in l',)l(). and methods of increasing the efficiency of the spectrograph were di.scu.ssed and various suggestions offered. As the writer's part of the work, and with the consent of the Director, Dr. King, experiments on the use of the plane grating as the di.spersion piece of a stellar spectrograph were undertaken, and the result.s are recounted in the first p.art of this puhlication. .■>H77(i-l' Ill I'l 111 II Al I<>N> c i| Mil iM.MiM.iv 1 ili-i |(\ Miiin A ili-riis.Kni witli |)r. I'nmk Schlrsinjici IiikcIoi ni" the Allculiciiy ()l)?iorviii(iiy. at ilic A^trDiioiiiical iiikI \ -troplivsiral Society iiit'otiim ;it Pittsl)urji in Au^jii^t I'.Ml.' rfsulln! ii; the cxpcriini-iit.s on I lie IniKtli ol x|TCPtruii» ilctnilcd in the ^ccoikI |i;nt I In n.arkrd irnin in itiitn-it\ .'it the violet, of the .-iieciia iiiiMJiirrii li\ ilir ^raUnt: ami li\ uiijiitivr iuimh.s of lisilit tlini its cotniiaiid uith iIkpm' nlitaiijcd l)\ ((,!■ \ ImJii m ,,iir order, (■latniL;- lia\i' I. cm iiilc.l li\ Ituwiand msiirz a \rv\ -inum lir-t (inin -l'''|';inni and i'oili MiehcUim and Ainr> couMdi'icd ii pii--iMc' in rule gratings tLuinji a- much a- 7.'> per rmi. of ilie inicii~ii\ in one -pecriuni Kvci\ il thi- ctlicirhi'v i^ lint ica''hrd. uraliiiL; -piM'ira -.lidiild cdinpaiv t'avniirahiv ^^i'h ihn-c pi-d.jilri.l l,\ llilTr -pri-lli di-piT-imi r-prciaiiv toward thr violet where the den.-e liint jila>.~ ii-ed m nn i pri-in train- i- -ironsilv alixwljin". A plane ;:ratint:, to ;;i\e ihi' nioM inieti-c po--il,le lir>t order -pecirinn "• "111- -ide, \\:i- oidiivd iliiuii-li the .loiiii A. lira-iiear ( ompanv, inliriolHr '■'I". "'I- inlid l.\ Dr .1 A Ander-oii ..i .l(,hn- liopkin- Inixer-ilx and received m .laniiaiA. I'.ilJ. Ih.' uraiiiiL' i- a .'.-inch plane uilh ruled Mirlacc "•4X',) I cm (2-87x:<-7."i inche- . nihni; iriOdd to the inch, d'he total '"I'nhcr of line- i- .'..'jSTo, which v.ill \h' t)ie re-oU mcr power in tlie lirsi "I'der n ihc whole aperture I- !i-ed. Dr, Anderson e-timaied that al'Oiii •'" P'l' '■'•"! "I 'he liiihi ua- concentrated in one hr.-t order, Imi a mis- undcr-iamlinu lix' the writer led him to think thi- was {\i- perrentace •,\w aljdiirnil. 7.|,"r7, IsKv 'Ibid 7. ISiH, 1>S!»S. KrriciENfv or HPf' ..iapiih I7'i of tlic iiicidfiit liiilit. wli.ira- ii ii'fcrrc.l to the linhf rfturin-'l from llic siH-i-iiltiii. metal .surfucr As the icflcptivity of !s|MTulimi metal is about 05 p7 Mini J 2") in.i a|Hiiiirc in d'firaiird l)ai k inakiiiii an armle of alx.iii .!() uith il,.' initial diivction. the s|) in i iucii- Tiic -|iiMirmii is f"llii''l 111 the |ilaiic nf, and a- clo-r a- ))u>-iMr In. (iiic rlid (i|' ill.' >lit. The -anir prinriplc |. .■-.,■.{ in ihr third nictlidd, ihc mating lirin^; k placi-d I'.'' •' liall' pii-ni 111 (Mil-' ula— iia\ iiii; aii'ilc- n\ .W .",. .",s ,",, mid Mil . and siKfivd o'l ill,. >id.. (ippii-ilc the .Vs ■.", allele. 'I'lir lijihi i- ivtrai'tcd into tlif jiriMii (in tlir li\ piitcnii.vc and, \\licii at mitiinnini dc\iatHin. i.- in<-idcnt noiinalh' nil llic ~il\c-rcd .^idc ,and ntuiiicd ainnji it> nrijiinal patli, cxcopt for the >liu:lit inclinaiinn ncccs>afy tn lirimr it nn the plate. 'I'lu' use of the iialf-piism for radial velocity unrk wa-^ prniin.-ed hy I'rnl'osnr I'amiiheil at liie meeinii; alin\c n'f(>rreii tn. ami I imder-ta;,d a -teilar spectroKraiih ha- lieeii cnn-tructid al tiie I. irk OliseiA ainry irnin iii> d,-itins uiiich tii\ es very prniiii.sinn rouli.-. Neveiiheie.^v. a- no chan.iit' was neeiied in the tii-aliiiu --iieelrnnraiiii except the siilistitntinn (It' the hall'-i)ii-m Inr the i^iatinK, it \va.- tii(iiii;hl wnrtli uiiile to make tot^ of ihi- 1 s pe. Th( linear di-per.-inii- of the three forms are a:5()A, 17').\. and 17:).l per millimetre at ll-y. Tlie.-e v.alues are almost identiciil with those ^;i\('n hy spcctrograiih 1 ^'^iilA! and spoctroj-raph III S .17:)Ai, one- and three- prism siiectropraphs of the Dominion Oli.servatnry. thus enaliiinfr accurate comijuri.sons of relative intensities to be easii\ made. 17(1 ITlll.li A niiN> 111- llil- lIliMIMliN li|i>l.l(\ A n ilCl rill' Iliri'li;ilii(;il ilr^i'.^ii (if llic >| mtI n iij.r;i | ill idllnvv^ llir ^i|p] ic illril liii\ f"nii. Ilr-I iiiliiMliicrd iiy ( aiiiiilM II at il:r Lick < ilix'Tv aiiirv, nl' wliich >ii('rc~>t'iil cxainiilr- li(-ii|(- tli(i~c ai Mmiiii I laiiiiliiui ami in ( 'hiic aiT tlic Mi'llmi -pci-i r(ima|ili ai .\llr:^licii\ ainl thr -iimlc-pri.-in ^pcctrdiiraiiiis ;it l)na\\a ami Ann AiIhu-. Thr |i,i\ i~, mailr ui l)i-a>- plate- iirnily sci-cwcil liijiothcr a! llic aiiLiK- aiiil t liiii(iiii;til\ hraiTij inicnialh, (Iwini; lo its relative eiiiiipaetne>~ a- cuinpareil wiili tlu' ar\ !(i int I'liiluce a lliiiil rdiinterlialaiicinii -uiiport as in the laller in-inii;ienl . iml the Imx i^ held lle\ilil\ and \et linnl\ in the ln'aeed T-ii'(in I'laiiM li\ iwu -upjHiii- ~(i placed a> to reduce llexiiie to a ininiiiniiii. 'I'iic cilijeiti\e~, which dehiii' 1 ii aiii iliillx . and I ni icllect iiii: slit were made l>.\ the .1, A. Ilia-heaf ( 'oinpaliN . Imt all other parl> ot' the in-trtiiiient were con-tnicteil in the (iliser\ator\ workshop. d'he lorin ot' the instraineiit. llie po>itiiinol' the plate-holdei- in the reiinlar and l.ittrow forms and its method of atiachiiient to the frame and tele-cope are well howii in I'lat" I. Till speeiroi;iaph \\a- -el up and thoroiiiihiy toted and aujir-ted in the lalioralory liefoi-e ii-iny: it in -lellar work. 'l"he -oiiree of liiiht einployeil was till' eailion are. which i- not only ea>il\' produced, hut which f^ives a lar^e nuiulier of well di-i rituued line- and llutiiii^s, and onl\ rei|uires \ery shorl e\po-ure-. It wa- expected that >ome dilhculty would lie encountered in the l.ittrow form from the linht relleeted fiom the lens -lu'laces foj^fiing the plate. I'liri uiialely the elTeci , if |ire-ent, was (|uite inapiire<'ial)le, and it was noi found iieci--ary. as in the .-olar -peetroi:rai)!i. to use an occult iiig liar or tilt the len-. The ilehiutinn wa- found to he -lightly superior when the a\i- of the oliiecti\e jias.-ed iluou^ih the centre of ti;e plate rather than thriHmh the ceiitre of the -lit, the aliule lietween the twii liein^; less than a dcfiree. The lield i- practicall>' flat o\er the raniie of -pectrum included on the plate, and the diTiiiition excellent with hoth uraliiiK and half-prism. I'he iiiethiid adopted hir olitaiiiiiiuf the eoiiiparat i\e inten-ities of prismatic s]i'ctra w.as to make e\po-ures ,-ucce--i\ ely on the >atne cele-tial ()l)jects, hoth >un and -tar> heini:' u-eil. l'i\e sped ra. side h\ -ide oti the same plate, were made of the ohiect li\ each in-t rutneni , the "xpo-ure times in every ea.-e Ijem^i jjroport ional to thi- iiunilMr> 1, 1."), 2, 'A. I. I'or e,\uiiij)li-, on EFFICIENCY OF SI'ECTIOfinAI'Il -i. 177 Pmcvdii. •") cxDosurcs wcrr made with III S, .'> witli tlic I.iltTDw iiratiiiii-. 5 with the i.itti()>. hiiir-prisiii, and tlun .") with III S, all tiircc iiistnunciits luniiifi the same disixTsion. A similar lest was larricij thi'diiiih with the first form uf the uratiiifi spcctrojiraph and the i>ii(?-])rism iiistnimcnt, l)"0 3S(XI Sun. G. 3 P. MP. Prcwyon. RiKel. 3 I' f I I' (I i| 1 :i 1 .■! 1 :i I 1 1 II 1 II 11 11 1 1 I () 1 II 1 II I) 11 OS 3 3.0 1 t 4 :i II Js JO IS 1 II 1 :i 1 II 7 J 6 1 1 :t 1 :i 1 ■-' I 1 1 1 II 10 1 1 1 1 II 1 1 II I II 0'2 () 4 () -) 07 ■ if I . I' •0 3 Mean (Stan). 3 1'. I.. 1'. 1 1'. 0-7 3 JO 1 S II I 3 I I 7 .'> I) 4 II 3 ■J 3 l> .> .") II 3 II 3 3 3 3 J J 1 JO 17 1 I 1 II S r, 1 J II 17 1 3 I 1 > 11 J 10 IJ 1 J 1 J I I ' II 10 1 1 (I 10 10 1 (I 1 1 '.I OS (I li 3 .'. t J 3 1 '.I 1 ."> I I II O s 3 1 .") O ."> O ."i 3 s J s J 3 J IS 1 li II 1 J 1 OS 4 0-2 li (10 ,"lli ."> (1 4 3 3 -.5 30 J 4 JO IS I .") 1 J 1 OS J 2 i.6! 1 -2 i 1-3 1-2 11 10 10 10 1 () 10 10 10 10 Oi» 0-7 0-4 2 2 4 3-3 2-5 IS 1 ,■> 13 I 1 s 3 1 3-7 ,■> 2 ,i 3 5 3 » '» '.» .) 4 J 4 J J ■J 1 ■S 1 1 1 3 OS 0-4 1 50 ti-2 6-3 5-8 5 41 3-4 30 2-."> j'l 19 1 (i 1 3 1 () OS 5 2 The most strikinir feature in the appearance of the jiratinp: as coniiiared with i)risiiiatic spectra which is well shown in Plate II. as well as evident from the talile, is the remarkahle iiniforniity in intensity lutwceii II.^ and X;5S.")0. This unii'ormity is perhai)s the most useful i)ropcrt,\- of the grating spcctroprajih, Fhe contrast hetwcen (rratinp; and prism in this resjioct is very strikinj;, as prismatic spectra are over ton tim(>s, ^ratin^^ spectra only 178 ptiu.ii \niiN> (iK rm. ikimimhn i)H>f.h\"aitaiits for () 102 ^iass sliow ilial ihrotmli ,"> cm, of lhi< '^las>, aiioiii liie mean lenjitli of path throiiiih the >ini:le and tlie lialf-pri-m. :i:{ per eeiii.i-. transmitted at the K line a> c(im|iared uiih 71 pel' ,.ni. ai li-, and n."i per cent, at lb. The etl'ect ol tlii> ali-ur|iiion i,- -Irikiniilx -houn in liie tal.le, where tiie rehitive intensity ot three-prism a,- compared uitli single- and hall-prism >pertru toward tlie violet is ni\-en, Discus.-inti lir>i the relatixc inten>ity of three-prism aiicira. ^^■'' fiii'i 'li'' lormer lia> the ad\anta:;e from Il.j to aljout ,\i;i()(). while from ML'UO down iiraliiiL; -pecira are decidedly >iiperior, ihree-prism >pectra disappear- inu: li(l(iw lis. If a specirimi in which the reiiion from ,\l;j()l) to the tiltra- \iolet i- re(|iiired, only the uratniii could he ii.-ed, while from Hi to II-, the jirisms wduld lia\e the ad\antat;e in other words for early-t vjk' .stars use the uratiny:. tor solar tyi)e, three-prism dispersion. Comi)arin>5 next, cnitin;: witii sinule- and halt'-prisni specfra we find the advantage lies decidedly with tiie latter alioxc the II and K lines, hnt if tiie K line i> re(juired. a,- i,> the ca>e in many early-tyiie stars, it can lie olitained with the Hratinjj; in the ,-ame time as with the one-prism spertrofiraph withoiil making: the lines Ijctween llj and I1-, inuneasiirahlc by over- C'\])osiire, It now the prismatic sjiecira are intei-compar( d. the superiority of one- o\er tiiiee-prisin i,s niarkedl\ diown, '! he fornuT fiive> nearl\ three times the intensity alio\e ,\l_'(i(l nnd a ii;iich ;ireater ratio helow. a ditTeri'iice uf "■oil than a mai;nitude it: the >lars within reacii nf the telescope. It iiiu.st no! !)e fort;()tti'!i, liowe\cr. that thi,s diff( renc( i.s pariialU ollsei hv the KFFU IKNTY OK SrECTROGFi APHS. 179 tliiTcfoM proatrr rosolvinq; power of flirco prisms, altliotiiili in jiliotoirrdphif spectra liicir lii^iii resoiutioii eaiuidt he etTectiveiy eniploynl. and it is of no value in early-type stars. .\ final interesting comparison is that l.etween the one-pri>m and tiie halt-prism spectra. Tlie Former tiives spectra aljdiit •_'() per cent, i.iore intense than the latter, a result to he e\])ecie(l when the los> at the silvered refh'ctinti .-uriace is taken into aeeounl and uhin the lari^er aperture of the !iall-|irism itiHl mm. as compared with.')! nun.i and the co!isei|Ueiit eviileiit that um re inten,-ecuri'(i hy u-^inji the re;;ular one-prism in.-trumei.t rather tiian the half-pi'isiu. making the camera the .-anie length a> collimator. Tlie furtlier advantaL'c of narrower lines for the same >lit-\vidth would he ohtained hy incrcasinji the ien.iith of collimator, camera remaining; the >ame. The advanta^ies of the i]alf-i)rism instrument are its much erature diliiculties occurrinj; with the extendcil one-prism speetro- jirapli. In conclu-^ion. it nuiy he said that, althouiih the spectra ohtained from the tiratiuK are (hsajipoim ini^ly weak and show that the proportion of tlie incident linht difTracted into the ,-pectrum used is certainly not jireater than that estimated hy .Vnder.-^on and does not nearly ai)proaeh that considered possihle. yet excn under this handicap it can lie used to advantage when the K line is re(|uired and if spectra of uniform dispersion are needed. It would also he u.nsity couhl he ohtained, it would he superior even to sinfj,le-prism dispersion l"or most work. Tlie relativi> tiexure of prism and iiratinj; spectrosirapli should also he considered. Ii is well known that a ^mall clianue in the po.^ition of a prism when at minimum deviation doe:- not disiilaee tlie spectrum, while the arifiular movement of a difTracted pencil will he twice that oi' the prating. Tests of the sin^le-jiri-^m oV rilK noMIMiiN u1!S1:K\ A I't IHY. order of 2 or I? km. per sorond. A similar test f)f flu- Imlf-pri-^m sprrtrn- firai)li -liiiwcd iiiaxiiiium Hcmii'i' nf alMrii 7 km. and <>\ ilir i;ralini: ~|iccli'(i- ;j.rapii of alKJUi _'() km. |irr -rcdud In mii>i cxpn-urc- the di^pkiccmciit caiix'd li\ liir rliaiii:r ut' p(i>itiiiM in liciur-an^ili' will nm mnrli cxcci'd onc- tt^ntli "\ iIh-sc amnimt-. and it i- fxident tliat willi pi'iipcr di-inl >iitioii of tlic comparison t\]io.-iir('.-. liiiTr ^liuiild \>r no aiipi'ccial ilc clTfi't on tiie velocity m('a>iiic.~. llriicc tlir litic^iiiin c oi a ni'atinii >p('cti'oLi.rapli li. K.xi'i.iuMi'.N IS (]\ i.i;n(,iii 111 >i'i:( riuM wirii oi!.ik( ri\r, imusms. While the uratinti six'ctfo^raiili \va> under construction ami wiiilc waitim: for it to lie mnunteil on tlir telescope, if -rcmed to lie worth while to ir\" ~nme (xpei-inn'iit- on tlie Icniitli of -pectnnn method of olitainiiiK radial wlocitie- witli an ulijectixc pri>m d'lie po--iliility of ilcierminiii^ \elocitie~ in thi-- wa\' wa> lir^t pointed o\it Wy I'ii'kerinjj; and later inde- pendently proposed iiy ( )!iiinsk\ and l'ro>t. 1)UI so far as known no actual deter!niiiatioii> of .-tellar ladial velocities ha\c e\cr licen iiiade l)y this method. If we consitar which i- at rest relatively (o the sjiectrosiraph. the ^lar line-- wii! he in their normal |)osition,~. [f the star is aj)proachilii:, all the lita ^ ure -liii'i(.(l inward- the \ iolet and it receding;, towards th' red In iirismalic -pectra. liowf\(|-, the dispersion in the red is only a small fraction (if that in the \iol(t. lieiween XO.'iti:;, Hq. and \;5'J34, K. tlie latin i- ie-~ than mie to -i\ while ilie>liifi in the lines, the displace- ment ilue to \ilocity, 1- almost e\acil\ niie-fuuri h at Hq to that at I\. Consequently the -]iecira of .-tars ap|iroachi:iL;: us are leicj;thened and of tho-e recedinji' shor'elied lietween il„ and K \v lodkcd after at lea~t e(|Uailv as careluliy in (jUjeciixc pri.-m as in >lit >pectriii;ra!)!is. Murcdver tliedeter- niinaiion- nl' the \cl()cit>- displacement- nf -tar lino with -lit >pcctr(i;iraplis depend upim linear mcasnreinent- (if the po-itiuii (if liic >iar line with I'espcct t(i cl(i>cly adjac('nt cumpari-dii line,-, whci'ca- in Kii^fh (if -peetrnm delermiiiat i(in,> (Hily tneasurement- (i\cr a iunc laii'zc nf .-[icctriun ,iii\(' appre- eialilc (litTcrcnce-. and it i- ( \idcnt what etTc' '. an' chaiiy;c (jf temperature, especiall\ . uould have (in tin- length. Xevcrthelc— it was thdiitiht desirable to I r> -dine e\|icriinciits aldtiu this line. |iarticnlarl\ as all the api)araius needed was a\ailalilc. Twd eifiht-inch aperlmc (ilijective pii-m- df alunit !.">' and '2'> refractinj: an^jle.- liad recently lieeii received fnim the .1. A. lira-hear ( 'dinpanv fdr use with the remounted photojiiapliie telescdpe. A l'_.-ii"''' ' ''Hii^e 'rclescdpc with a l'h()to-\'isual ()l)ji'cti\c had liccn in the [idsses-idn iif the Astronomical ISranch fur alioiit 2(1 years and as the S-incli stellar camera had liecn dctaclied from the tulie of the i.'i-inch eiiuatdiial it was cdinparati\cl_\- easy to attach the tube df the < 'ocike 'relcscope, at tiie proper aijiile, with the objective prisms in front, in place of the camera. hi order to det'Tmine the constants of the prisms, the deviations for different wa\('-lenf;ths. l\j at minimum deviation, were measured for eacli prism !)>■ a spectrometer with the folldwinn results*. l)i;VI\Ti(i\ .\N1) Hlsi'KliSInN (il iiIUKTlVl'; I'ltlSMS. 25° Prism. 15' Priam. Wave-Length. j Deviation. Dispersion D to [ Wave-Length. Deviation. Dispersion D to I / o ,">s<.t:i 11 1.". (1 ,"iVt:i 1) 8 47-5 47;)7 l.-i II 5 u .-1 i7;i7 .s ,^.7 ■ n 10 IJIC) 1.- :ii .il •IL'll) '.( (') 1S.5 311.) 1 1\ ir> us 41 j 3U34 K 9 lJ-5 25 56779—21 182 rlllLli \ I li i\- (>l IIM, lliiMIMilN ull>l ll\ A riiUY. The focal Iclllilli nl' llic l'l;iiti>-\ i-il;il ( >lijcn i\ c i< M N inrlic- lirlirr tlli> makr- the 1ciil;i1i nl' ihc >prcll'inn tnr llir 1,") pl-i-lii. I'nilri I) t(i K l.")i:5 and lur (lie _'."i pri-in •_'."> 1 1 iniii.. a Idal Irii'^lli willi Imtli |ui-iii- i>l' HI -4 iiun. 'llir ili-taiicr Ih'Iwitii 1> aii'l K with tlic -ini-in >|mtI mjii'apli is 4r."i mill.. fuiisc(iii(i!ily iliry liavc iirartically tlir -aim' (li>pci~i(iii The coliiur (■ui-\T (if till' iiliircti\(' \\a> ilctTinimii liv Ilartir.aiin's cxtra- fdcal iiKiliiMl li\ piai'iim a iliaphramn nvcr tlir dliin-iiM' with iwo >lits al«Ult lialf-ilicli 1>> ilirli llial' llic cillio. tiir lniil': r'i'j,!' t'l rlir slits pr|-pcliiliclllal- to tlir lilir loiliillLT tlicif crlit lr~. aiu 1 llli~ latter pai'allrl to llic rrfraciiiit: rilin' of thr pri-m-. I'lMitoiirapli- ot the -pirtiutii of a lifiulil star, with ilic plair some HI imii. in-iilr ami n\il>i(lt' focii-. ciiaMcil llu> \ariou> point- oii tlir nirxo to lie drii I'lniiiril faii'h' accufatclv A plat of II. 1) II. II. 11, ILII Colour curvp of Cooki- I'liolo-Visiuil— rppcr curve. Curve of SiMMtruI Forui', inrliniij Pkite— Ixjwit ourve. llciriionlal lini's are oni; iiiiUlnictrc ajKirt. the curve is y:i\rii in thi upper ciirxc of the acconipatix int: timirc 1 lie niiniinuiii focii,- is atioiii ll,-, while 1) ami II7 are each aliout II7 mm. Ix-yoiid it. ll,j is onl\ aliout a iiiilliinetre lie\oud the iniiiiuiiitn liul 11, is 2 mm. and K at>oiit 2') mm. lieyoiid it. .Mtliousih well achromatized KFFK IF.NCV m M'K(|l(M(.ltAI'l|s. 183 \i^u;ill\ Mini iTiiiarkaMv I'rrc I'miii a m (■(Jiidar) ~prcM nirn. it is not so wrll (•(ilTrrtcil |iliiiliiuia|iliii'all\ tlmllLlli nf (■(iiirsc iiiuc'li -llprl-ior In llic ofdi'Kiry vi-iial ()l>ji'rii\c, wliiTc till' ili-iaiirc iiciwccii the iniiiiiinini ami I\ i> alxiiit ;{ time- that of the l'holo-\i-iial. I'oi- (he WKvk to lie aci'oiii|ili.-h('(i In thi'sc r\jirririif'iit -, it would he lUrh'ralilc il' the tliiliiinillli Wi'lT >liil"tc(l to\\afd> the \iol(i XI that thr liirht al J) and ll,,>a\. would lir at thr -aiiic I'orus. 'I'hi- would j^ivc a loiiy,('r rani:i' of >|iirtruiii ii\rr whirh to mca^urr iiilTrn'iicr> of Irn^jth. a yircatcr i|Uatitity of .ii-ri\ at ioiial data and pn -unialil> liicaicr acrui'arw This shift ran hr clTirtrd pract ical!\ whrii a s|irrli-iini i~ to hr photot^raphi'd hv suitalil\- inrlininu thr platr and thr doltrd linr shows a faNdiifahlc |)ositi(in, an inrlination to the prrprndiriilar or oi'iLiinal |)o>itiiin of almut :i ' .",. \\ h(Mi the oilout' rurxc Wiis rrdrli'rniinrd fof the nrW" position of till' platp, it took the hum of the liiwir fui\i' iilmtiral oi' couim' with thr upjicr except for thr chatifj:!' of thr slo]ii'. It will lie si'it. that the niiiiiinuin focn^ is now at ahout XKiOII. that II ^ and II, arc ahout i!'J nini., H^ about ()■('), 1) and lU almui II and K about 1-7 mm. bcvond the initiimuni. ]V: ai'coinmodation, it i- r\ idiiit that lu'actirall. the whole spectrum would be in ;;{Kid focii.- when thi' angular aprrturr is 1:1S. I'ut it is evident that the Icniith of the star sprctnini will chan!j,e much more rapidly for slifilit dei'artutes from the axis than whrn the i)late is normal, and con- sequently this method was not used. It is e\idcnt that tlie leniith of thr sprctrum |iroduced will be some function of its position-anuie with respect to the refractinu; edfie of the |)risms and its anuular di.^tance from the optical axis, and althouiih it is jxissibli' to calculate the clianges in lenjith in\(>!\-ed. the uncertainties are such that little coniidence would bi^ felt in the results obtained. Consequently onl.\' one measurable s]iertrum can be made at a time and one preat aihantauc of the objecti\(' prism is lost. I'urtluTmore, a- tiic leniiTli of tliis spectrum will drpend upon lh(> tcnnperatiu'c and focus, it is neci>ssary to make besido it, pirt'erably b\- shiftinjj: tlie plate sidew ise sliiihtlx' so that both spectra nia\- be alon;; thi^ axis, a spectrum of a star of known velocity 1st iM'ni.ic vrii'N-^ iiK nil, duminiiin iiii--ki!\ \ roiiv. to act as a -tandnnl H is tviii moro r^sontial thnn with thr s-|ir iar 'n aiiotlHT iiii niiliHTs a ciiaiitic of posiliun. it is dosirnhh' iliat lli\urf shiiiil.l lir inaih a- -mall a- pos-ililc even th(Hii:li the cliall'^ir 111 lr!li;lli pnnluccil I >,\ tlii- raUsr will Iw ,-iiiall coinpanil in tin- ili-placina iii of tiir spiTiiiii.i as a wlmlc it \\a- lint iVa-il'lr at tlii- tiiiic \n iricln-c tlir apparaiu- witliiii a constant tcnipiraiurr ca-r. w hirli wm .1 lia\i' rc'iuiivd a piccr i<\ phiiic iiiallcl iila— ami an clalnii-air hratiiii: ai lamii'imiit . ncir was it p(i->iMr wiiii ilic -iiiiplr fciriii nl' iiKuintin;; id mak( the llc\iirr \rry -mall, Hwi "Uli luiulit ,-lar- aii'l -liort I'Npixiiic- llic liiiipiiaiiiiv cliaiiL^c- \\{ the rxpo-iircs liy makim: an r\pn.-iiiv (111 llic I'limpari-iUi -tal' I'nili inl'mv ;'inl al'tcr the cxpo-uiv uii tlir star \\li(i-r \i'l(Mit\ was dc-irrcl, .\ll ilinc ,-pi'ctra wrrc mcasiircc! under the a— uiiipiiiiii that ;in> rhaiitir wmild lie lairl.\ uiiil'orni ovrr the >hiirt tiiiir rci|iiiivd, -(I that the crrnr due to dJtTcrciicc ot' tciiipcratiin' would ut least be much reduced. Spectra with the two pri-m- in front of the Pj-indi Cooke I'lioto- \'isiial were made of many of the tiriiihter stars, not oiil\ to lest the expo-lire time ivi|iiiiid. I'lH to compare them with -pedra of only -lii^htiN irnater cii-per-ion made 1>\ ih<- one-prisiii -pectrouraph for -harpne— and power of recordiim deiail. Spectra of <. i,.\rae ami .v < 'yiini with hotli in-iriiimni- ate reprodiicii! m l'hit( 11. .aiid it will !»■ noticed thai the -lit spectra ha\e much -harper dilinnioii. The ciliiecii\e pri-ni -pecira arc made on Wratten and \\ .linw i iiiht '- I'ai. chromatic plaie. ihe -lit on t!ie -lit of the -]iectroMra ph. the -i:ir im.ML'e trailing: across sevi-ral timoF. 'I'he -eeiim was oiil\- fair and tliat would account for part of the lo?? of fiefinition. KiKiiii'.M \ 111 ^I'Ki ri;i '■!! \rii>. is.") Sn far .1- i(l;ili\i' r\|i(p~iiic tin.'' i- ciiiirrnird :i (■.•irc'hil i'~tiin:iii', taking arcdunt III' >|ir( il 1.1 jihiti-. wiilih hI' -|)cilia. ami irlati\c irilcn-it.v fur -rvrral l)aii> (if -pcclia -lii'U (ill llir a\(rai;i' t!iai al II, ilic (ili,jri'li\ i' piiMii -pcctra iciiuiiv :; •-' limes the e\i)osuiv m ilir -lit -prcira Tlie apcrlun- are I'j and 1.') melie-. triUismittini; atiKiiini- ui' liiihi in flie ratio (i< appruxiiiuitely 1 til II riii- wiiulil iiiilicale iliai the In-- at tlir O.')! mm >lil. iissii-- iiiK tlie remaimim lu-e- in tin- tw.i iipiical >>st.'m- in lie the same, i^ aliuut 68^^, PdpMiliralilv li -- than ii-iiall\ e-iimali'il lull uivaier than tiliiaineil in ni\ previoii- I \p(iimenl-J linwevc-, if iitln'r part- uf tlir -peririim are ciMiparnl V.I' lliiil that the nlij..ni\e priMii >pertra '-eiiuire aliiiiit O-T lime.-; at K, _' II time- at II ami li time- at \Kiili) Thi- iinhraie- ihe -irmit; ali-iiri)ti(>n nf the optical tram nf tlie -lit -prri n.^.-apli at the violet, il \\r ion>iiler wa\.-leiiutli MiilM). where the ali-^orplioii-; of the opiical ->^leins aiv not ver\ ditTerent. ue t^ri a lo-> at ihe -lit of onl\ I.")',' whieh a^rei's \('ry clo-elv uiili the former e\pcrimeniall.\ determined lo>> for a 0().)1 mm. slit.- Three pair.- of a half millimetre >erew liivi n in eohimiis .">. 7. '.» of llie table h(>lo\v. The dilTereiiee in veloeiiy of the two stars i> ol)taiued from the differenco in lenjith of the -perira hy di\ idiiej.- Iiy certain con-^tanls. Tlie-e constants were otitained liy -uhtraction of the di-iil:irenients per kilometre at the ditTerent wave lem;ths. a;id these latter were derived in the well known way from ti'i- coii-taet.- of nariie.aiiii's formula for the spectra u>ed. The ncccssar\- \alnes are nixiii in the tahle. 1 Rpport Chiof .\«tronompr, lOaS. p. SI. Astroiihysioal Journal 27, j). 150. 2 It mist iiiil 111- fiiriiotton that the |iri)|)oriion "f si^ir lii:l.t .ic.(.|illi..l :\t tlip slit is ;i fiin.-tion uf tin- serine rin.i oi tlio quality nn.lfo.;ilhii«th of llii'objfctive, aii.l th.~'i. tinur.s arc only v.ilul lor tl:c l.-)-im-h nt Ottiuva mi.icr avcrafie CODliltlODS 186 IMllI II \ rioNS OK im lioMIMOs iiM.-KHV ATOKV I.ow.T W I. I |i|i.r W I ii.i 111 y&f. 9 .liin 2i W 1. por km W I. l)i»|i'mt p miiwss "> mHiMil I 111'.* 31 .'i 03J0 (17 •-• out:) 41 I) t)Oo;ts<» ". IXHIKi^ irj.vi US 1» 0410 IVMI I) (XKCtSS IK.Mt IMMrtlH) o:tJ7 411 11 "d (KKHi.M II. IKKISI'it mil) 1.-. s "fl INVXVJt ». ■ IK) !(»:(> Old.-, .10 •OJOO In : 4400 noosis 40.->n 1)01000 01 III M 1 Mc.ins MR Ml U « Iriiiii I '.nit linker's orl lit i.f .i (hidiii-'. llir velocity at tlir r\\<\ >A I'.M'J w.iillii \>i- al'nu! • _'<» km. piT -rciiliil and ( 'aMIIilHll L^ivr- liir vrldcily nf a Taiiri- as -f-.V)-l km. prr -it.hmI. .\i>pl>iim the annual cunv. 'tinn^ vr ohtaiii the (liffrri'ncr-- in ilir M'lucity of >> Tauri ami .. Hiioni- a- :!^ '.• km. ,,11 ( ),i,,l,,i- hi, !_' '.I km. "11 Nov('iiil)cr '.». aini IJ 7 km, m .laniian li-', V.)l.\. jiivinii n-i(iuals. (> <'. "I •-'■!. -\:\-2 and -2-!1. Tlip airri'cmi-nt i> in the lir~t and la~t licttiT than 1 rsprrird from ttic qtialiiy (if ihc -pi'ciia and tlic pi."ililc .Miiircr nf rrnu'. i-prrially ilia' due to rliaimr of IrniiHTatiirr. which i- very likely the I'ail-e ot the hit;li residual of Xovcmlirr '.). 'Ihc tv.o priiiciiial ditliniltifs in the accui-alc diieriniha! i.ui "i radial voloritios \>y tliis method are in my opinion the poor ijerinitio:, ..i the ohjoctivc pri.Mii .-pectra and tiic maintenance of con-iaiit temperature. 1 he latter of coiir-e can lie overcome liy a siiitalile temperature ca-e with a sensitive tli('nii(»^tat and a fan for eirculatiii!: the heated air. I he poor (lefiiiitioii will of cmir-e lie improved when the optical part- are ke]it at constant tciniK-ratiire and iiece-.~arily then at con-taiit local leiitith. A further iiiiprovciiiciil in the (Icfinition would undouliti dly he i Ifected it tlie focal Iciijith wa- decieti-ed. the di-per.-ion lieinu maintained I y u.-ini; more li f 'A. X. 4467 ■■\. (). B KKH( I1.NM 'l'Kr 1 1(1 i(,l( M'llS. 187 prisms, or pM-iii~ nt' uri:it(r .-ini:!"' Tin ;i'.nhl:iiicc ni IliAiin- uoulil iif m fiirtlirr ;ii'l tln-r uitli ■j.ii\\f sprct I'dnraiih. ]■^,^,,■l, if ih('\- wcfi' as -harp linw cxcr, tlir iiirtlioil of iiirasurciiiciit lUTCssurilx ivndrr- the ariMnacv attaiiiaMc cuiisidi ralily Ir-;-; tliaii tliat IMissitilc \utli ■-p(ctra iiiii-t hr kmiwii, arf quite similar m t \ pr. In ciiirlu-ioii. while tiic^c lApcriiiiciit- show that radial velocities can lie diterniitieil li\ I hi- iiiithod with ati acciiracv likely eoiisiderahly irreatia- than that pn— ililr li> the u-e nf an ali;eneoiis olijective prisms will necessarily limit the aperture that cati lie I'lnploved and, cotise(jnently . the maiitiitudes of the stars attainalile to tliose easilv wit' i tiie reaeli of present (Hinipineiits. a Max I lie allowed to siiiiire-t a direction for future exiierinionts alony; tiii> Hue which anvoiie |)o>ses.-ini.' a larne. aniiular ajierture, apoehromatie ohiective or a paralmlic mirnir and a slit speetrop-apii may undertake. It the slit is placed ai-ciiratel> m the focus of the ohjective. and then opened widelv. you have practically the e(iiii\ aietit . so far as liplit efTirieney is concerned, of olijective prisms of the full aiierture of the nhjeetivo and nivinj; the same disiiersion as the spectro^raiih. It will of course he nee<'ssar\ to have .some means of ^.tiidiiiii, such as by retiection from the lirst prism surl'ace, or liv an auxiliary ^uidiiif: lele.scope, hut one could use as larjie an ohjective as desired without piettins into dilfi"ultics in ropard to ahsorption or lack of homogeneity in the prisms. isv; I'l HI i( \i iii\^ III iiii: i)i>\iiMii\ iiit-rii\ \ri>ii> III. NK.W Ml'l|(\l, l'\i;i- mil nil i.\K-l'l(l-\l >1'K(TI(III1H\1'H It lias Iniin Imcii the <.|)iiiiuii .)l ilu' writer that thf di-iix- cilicatc timt Ulass. thr i> inj ula» ->f ilir Jena (lla^s Works, wliicli lias \m'vu alinoM pxrlusiv«>ly cinployctl a> tlic |)rism niatorial iti iiuxicrii stellar spcctronraplis. is not tln« most suital)l<' jjlass for the i>iu|i(iM . tliat it i> too (Icii-c tcm hinhly cDloureil. ami ln-iicc too al)S(«rl)inn cspcciallv in tin' violet, to i;i\' tin' hest result- I'lii- niiiiiinii wa- cniiliiiiM il li\ iIh r\pcnnieiits pri'viously il.MTihe.l with tlie objeetive pri:^!iis of linht Ihiit iila>-. Thi' >peetra nlitaincil i.xtniili ij iiinrh t'lirthef iiitu the vinlc^i ainl were imirh iikiK' iinitnnii iii intensity .■ilnni: the inen-inahje retiinn than thu-e ^;iveii liy the pn-tii- nl f) l(f_' trla- I. II, , •1111:111,111 apparenrl> with niaii>- nther siiertroscopists, sconi to have chosen the (> 111.' l.daii-e It was (jenerally uxil for tlu' pnrpn>(> and in'iiee piT-suiiialiK the l>i-t. uiiiiMiii carefully goiiifi into 'I"' miitter ol it- lelati\e merit- a- (nnipaivii \Mth nthii' ulasses^ It i- priil>aiil\ thi' tim-l -iiitaMe aiiioni: tlie ileii-i- tliiii-, ami the hiL;h iji-pri -H iii ami iv-nlx im; pciuer ilcmamlcti 11, the miiilern ihi( r-|iri-m -pii-tmiirapii, in'^eihiT uith i:ieat compaetne-- ami .-vmmetry in furm. -ei-miij mn.-t ea-ily -ati-lieil li> liic n-e nl ili'iiso Hint tihi-- I'.ut that thi' eondii inns n (iiiiivij in une-pri-m in-t nini'tit- were lint tlie -atiie wa- mil at lii-t rrciiLnii/ei| Tliiir principal u-efnlne-- has proveii tn i-nii-!-t ill iil.taiiiliit; 'lie -pcclia <>\ nmy t \ pc -lar- ulieie the lino are tru in mimlicr are 1 ach due u< mir clement i>nl> and nni Mcnd- as fre(|iienil\ ucciir- in -nhii' t \ pc >tar-. and ate \er\ nlteii liinad and ijiffii-i Tiider these conditinii- hii^li di-pi'r-inii alid hinh re-nlviliji power arc lint nci'C"ar\- and imleed arc when the line> are ditVu-e, a i|i-ad\ atitatie. 1 iirtliertnnre, in nrdir to olitain a- mm'h material i'nr mea-ureiiiciit a- maii\ liiicM a- pn--dilc III -pectra which yemrallx iia\e hw line- il i- \er\ desiralile that a InHij,- ranue di wa\e-!entith he phnt(ici;i|,hcd at one r\|in-iire. When ilie-e cnn,-!(ierai inn- are laKdi into aci'mmi . ( 1 juj ula-- dne- nK SI>K( IliinWlAI'llrt. IV.I A soiircli into ili<- lnmitiiiv ..n the Mil.j.'ct rovralid liftlc ia|mtiiii. iiial (lata Ml tlir rrhltivr at.-niplH.r. nl ,lllTrlv,it uIm--.- Ill th,' |,l,ot.,i:r:i|.llir renioii. Til.' only inv.'^tiuatKm iMMiiim on tin- .|ii,'-iinn -..■.■inr,l i,, l.r ihr unrk iiiMlrrtakr.i at l'..t^.lain l.v \'<■\n^^ I,, I- Ihr plintnuiapln.- Iri; l.lll as th(>V nnU llirln.lr ill'' MnJiliafV nnmlHr ..1 inaimal- i- n-i- \nv'z>' r.nuiiLli lidw.'V.T. kiinwiiiir lli.' ahMupimii- ..I' th.M tiJasM's. we cati obtain approxiinatr ulra- ul llir atiM.rpti(,ii< of ..ilicr ula>-c- li\ a ciinipaii-iMi of rli.ir (ji-pcr-inn- m ,litT Mronncr a!i-"ri.i ion in the ultra \iolci than ihr other, ■ ami vice \ir-a. In ih,. t.iMr /, i- till' iidex ot rclraction ior 1 ). A i- the mean (ii>- p,.|-.ion from • lo 1 ami a d -, ilio r;itios of the ,li-prrMons hetwccn A' ami 1). l.ctucrii I) ami I' and l.ion A. The la-t live colimiii- iiiM' ili<' ir,an~mi->i oi 111 cm. of tlic i;la~-. lor li\r uavc-lcnuth- m thr violet. I M'.i i: I T.Uil.l nl (cWr.WT.- -I.IIATLU Ul.rJil,>l..NTAl'l\ K < il,.\SSJX Kinil f>f Tnule (■Um. Nu. Ration of Trangmiiwion through 10 cm. for llispomion Wavo-Lcngth B ^ ~ I 4341 400() 3950 3900 3750 ~ .i83 .583 Ordimirv S>1 Cr. O20:! l.'IT.-. OOSTT WJ 702 ,568 | -667 «95 , IWvIa I.iKhl Kl I) 72-J l.-.7'.t7 lO.sT ■<;..-■ ,(i, ._.,, | It-.rVH I.inht II () IJtlf, I (UltJ l:tSI MX ." ■''■>*'. .„„ UnWrvLiul,. 11 Oiltl. I .".77 1 i:i* ..iH Tl.t ;•-« ^ ■•'"'■' «' » lUrMi»l-l <> 71> 1 iKi.-, IM. .... , .1 HV Hoavv Siliciitr l-l. O 102 i I fvts.. IM'. ..IIO i -714 | -Olo -502 , -463 , 187 . 02S 4.56 :i*>S i]!:r{":::ildiu8ae;iufi!';:nJli^ln^^^ ^>0.vl.r 11 li.,v,..,:,.l, T,.n.Uu-.lby,heKv,.r...H. Ma.Milbin 190 prm II \iiii\~ I'F iin: i>m\iimu\ iti!-.i,i;\ aii iin , The (|ii:iiimir~ iran>iiiiltr(l ihnnifili Id cm. ni" () lnj ul;i-> -Imav Ikiw lui-witalili' it i- I'ur prism material, mil only on acTiHiiil ni' its --trnim -jicrial al'Miriiiidii arniiiiil 1\ . luit ai-^f on arcnunl dI' llic ^ciirral a!i-iH-|iiiiiii ail aldiii: tlic iilicitdui'auliii' ~]icclru'ii. I'.wu if accouiii is tal\i n ni' tlir iticrcasi-d tliickiicss necessary wiili tlie liiiiiter y:lass(>s to got the -ame rcMilxinti pdwcr tliey will still have coii^iderahle ailvaiitaije fniiji \ 1(1110 lowarcN tlie red. while towiM'ds and in the ultra \ iolet the ail\anla;ie will i.e \er\ marked. Tlie ()rdiiiary I.ijilit llini n MIO is a c()n>iileral)le iinproxenaiii ovtr tlie () \02 especiall.v in tlie \ iolet and ultra siolet. luit the <)rdinar\- Silicate Crown () ■_'():{ is still lutter. Ilnwever. the dispersion ol' this j;lass is so low that it \\a> not deemed ;id\i>al)le to use it. ( )n lodkim; o\'er the lists (if Jena (jln^'^s''^- the liar>ta tlints apjieared the most prnmi-im: and aopordinfily three of this tyjie ( ) 7'-'2. (» J'JCifi and n7l are tatiiilated for rnnipari .on aliove. If we compare (Mlititi with i>i!lli we -ee that, acroriliiiii to the rule 7 less and a greater, it should he eon-ideral>ly le~s .•ll)^orllin«; in the \iolet. while a coiii|iari>oii ,,\ (» 7_'_' with (>_'():> and als" the mcasureineiits further to the \iolel. the central wa\e length was chostn as M'JdO. This is about inidwa\' limarlx' hetwei'n \t."i."iO and X3tl-'.' 1 the u.-ual limits of measurement propo-ed. or lieiween .\lN(i_' and X37")() tho extreme limits to which the niea~uratili' -pecirum extends. The constants of the material of which the pris:;; is made are )) = 1-.")7S2, A= 01il7^. (i=(iX7, .i=-7().S. 7='.')7t). whii'h are c\cii more fa\oiiralile for t'ansparoiiry in the ultra \ iolet than the laliular constants, ('omputinir in the \.v'll known wa\' the refracting angle and dinionsions of a iirisei of this material of ol nun. (2 inches) dear aperture and with M20(l at minimum deviation we obtain the angle (i.S'^.")"'. the length of the sides lis mni '1 tit EFFICIENCY UF .SPECTKOGUAPIIS. 191 inrhov) ami wi \hv a-c \X\ imii. -'i--'.") iiiclic^i. Tin- moM faxouruMr ant^K I(. niaUc ilii' l(i>~ I'V ri'ilcctinii a tpiiiiiiiuiii for thi~ material is alinul I'.r^lo', l)Ut ihr iiii'ivaMMJ 1(1" li\ llir Miincwliat nivatcr ai:j;l«' is iii:-iy:nilicant. The pri~iii \\a~ inailc 'u iiiin. _" | iiiclirs: hiiili ami coiiscciucntly i- a larjic hlcck of tila— ^ it i- l)cautil'iill> ciilrHirh-s ami ti-aii^parcnt ami nut withstamliii^; its large .-i/e -lidws iKi irace ul' iiiiperlVrt iHiiiKiLieneit y. ('aniul le-ts hy (liaplirajiiiiiriii dilTei'eiit -eeiir,ii- liave .-Ikjwii ilial every part (jl it deliiies e(|ually well and a> a dispersinji piece it is prai'lieally i)erl'eet. In ttu'vy to i)l)tain 'lie full advantaire nf tlie traiisi)arency of this jrjass. the iMikinnatic lolHinal or (.lij.ctixc of the speetroy:r:ipli. u hu-e eentral (•onihoneni i- .lei-idedl\ yello\\i-h. wa- replacid hy a Hrusliear Triplet and a '{'riplet Camera ( »lijerti\ e of -h^;htl>- loniier foral lentith than tlie Siniile Material ua- aUo ..htaiiied. liolh of these ohjeetives were cemented with watch oil, which I had learne.l in I'liiropc hail been >ucce-sfully used for the pnrpo>e N(i -train can he imluced 'ny it in the lens(>s, it does not evai)()rate, and the lo>s hy n lection from the internal surfaces is much diminished. However it was fi)und, at the temiieratiires which a stellar spectroKfajth reaches in the winter, frecpientlv much helow freeziiifi, that tiie (lil crv-talli/.e- and can not he used. Tiie oil was replaced l)y ijlyceriiif. wliich is ■;uiiposeil to remain unchantied at low temperatures, but even this .seemed to undergo, at any rate in the thin layers present between the components, some .sort (.f molecular cl.an.i;e anii ajipeared mottled and liad also to be removed, .\- will be M'en later the loss hy reflection in the internal .-urfai'es i> i-alher a serious matter and c(in>e(iuentl\ it was d(>cided ,,, ,.,.nM'nt tlu' lenses with balsam. Some bal-ain picjiare-d by San.sjer . not so likely to introduce strain. The cementing was successfully accomplished and up lo the present no ill effe( ts (in the delinition has aiipeared. The .spectrograph box was dismounted and the new prism and objectives installed and carefullv tested with artilicial .sources behire being trii-d on the 102 PI'milAIIONS UK mV DOMINIdN i )H^^.H^■ ATI >I{Y. stars. The (li'fiiiit. -lioweil -ti'ikinu aih aiilaiies m eltirii'ncv, es|)(>('ially in tlie \iiilet, (ii ihi' liiihi tliiit A cii' ii> eliaiiiie in the cliaraeler ol liie lielii j;i\('n la liie tiiplel wa- iiolii-ei! Willi ilie I u o pri^nis, t'or. while with the (lelise iHli! the lii'lii Was cuiicaxc In the len-. wilii the litilit Hint, it was (•on\('.\ aiiil ol' siiialler i-iir\ at ure. i'luliablx' wiiii a pri-iii of interniediate (iisper.-ion the tielii Uoulil lie tlat. The (lilTel'elice in focal lcni;th ot ;he centre At.'dll and ol ,\ h')."ii I o e ~ide and K on the other is about 1 iiinL and. ti> accoininodatiiiL;, ilie focii~ of an\ part of ilie ~pectriiin need not lie more than nil") mm. from ilie phiie Thi- i^ not of coiir-e so fiat a tield a~ tlial liiNcn li\- the Mniile .Material, Imt. a~ the one usimI with the Old'J pri-iii wouiil not come tn i'()cii> with ilie lariicr ]iiism without chaniiinjx the camera end, and a.- it i~ not po>.-i!)le to ipiite free this i \ pe iwlien it has such a lariie annular aperture, aiioiit fS. as thi^ from sjiherical aherratii.ij it was llioUL;ht pl-eferaMe to use the triplet. The di-pefsion ni\en li\ the coiiil linatii in i- ."> 1 ."> .\ per millimetri- at H>. is :i \ per millimetre at ,\l_'()l). the central ray, and :57-.') A per millimetre at K, fhi^ i~ alnio.-i exactly t hi'ee-fiftli- the di.-persion of the spectrofiraph with the t» 1(12 pri-m. d"he linear di^pi-r^ion can of course l)e broutiht to ei|ualii\' \>y iiurea^inii the focal lensith of the camera. Init this is not f<'asilile with the pre-ent in.-i riuueiii . I'he >nialler dispersion will pre^uiiiaM\' iii\c a hitler prohaiile error of the radial \-elocity measurements and the -pecira will. perh:i|i-. not -how ■-ome of tlie faini metallic lines vi^ilile wiili ih( den-er pri-m Hut the e\po-urr would, other thinii- lieiiii^ e(|iial. lie onl\ I hree-lift h- a.- meal and ihi- will lie further dimmisheil hy the decrea.-(d lanie intensity around H-, and only aliout on(-iiinili ai K Ihi- mean- that :i maiinitiide laititer m;i\ he reached and. a- the -tars witliiti reach of the pre-ent eiiuip- liK 111 are pi.aelieall', all worked up. ihi- will olT>et -oiiie di-ad\ alltafics. l')efore u.i\iiiti the resiil's of the comparal i\c te-t> of e.\po-uii's. it may he as \vell to ^;i\e lirieily the \alue- reached li\ compulalion. The looses KFFiriKNI'Y OK SPKf TUUCiKAI'H.S. 1U3 l)v ri'dci'tinn aro obtained frnm th<> forimil:ic tiiv<'ii in any treatise on optics, tliose iiMil hriim ohtained Irom Seheiner's Astroi:(Hnieal Spectroscopy. Die losses li\ alir^orptioii lia\e Ixcn coiiii)iit(M| in the well known way from tiie measured value- oi' (i lOJ -las- and t'roin values inferred for (» 7'J'J filass ln,!n the luea.-ured ali.orplions of o 'Mi and OHIO and Ifoin the run of the di>l)er>ioii.~. These v; ' ;es are jirohaMy approximately correct. Let us consider firM tlie lo>ses l.y relieetion in The (jrininal instrument. 1. l>okuinatic Collimator :■> ceiuenied coiiiponents, 2 free surfaces, incidence nearly normal: I raiismis.-ion at each free surface U.^i'). ai each cemented -urfacc '.t'.t. alter emergence sT?,'). 2. () 102 I'ri.Mii, Transmis.'-iou at lirst surface -ST-l. .\fler tw(,. reflections •779. •A. Siniili' .Material Camera 2 separatee} lenses, 4 free surfaces ; transmission at each surface -'Joiir). after 1 retlection.- S217. Total Intensity after lo.-..se^ hy reHecti(in e(|ual.- product of 1. 2 and :i or oflOT. L()s.ses by reflection in new optical pa-ts. — 1. Tripli't Collimator, A cemented components, 2 free surfaces. Trans- mission at each siu-facc O.kS.J. Transmission at each cemented surface -99, After emerffence -STii"). 2. () 722 Pri.~m. Transmission at each surface -SfiS. After two reflec- tions 7tjd cotniionents, same as collimator ■S735. Total Intensity after lo.-.-es by relieetion et|uals products of I, 2 and 3 or ."),S()1. Los>es by ai>sorplion. The mean free path throufrh the prisms was taken a< half the leni;th of the base, beiiiK ■>' 'nm. in the () 102 and GG nun. in the O 722 i)rism. The length of path throutrh the I'riplet 194 Pfni.TC\TI(lV~ OF Tin; DoMIMmN nHSKUVATiinV. (■(illiiiiuli)r aiiil I'aiiifia Ini-c^ ua.- taken a.- 'Jll miii.. tliioiitili the Siiii;lr Material a- Id iniii. uliili' iiic ali.-()riili()ii> were a>>iiiiuil tlic sainc as tlial ol' () :_'■_' lilass. I'lii' ali>iir|itinii 111' llir isukiMiiatir ('cilliinatnr \\a- ilcliiiiiincil. a^ no (lata wci.' oljlaiiiaMc li'i' tlic \fll(i\\ I lun i-ilicair Hint ccniial cunipniu'iil , liv I'limpafini: ilic ciiliiiir 111' tlic roinijlftc Icn- witli thai ci|iicntl\' the ali-iifl'liiili III thi- iilcecti\e \\a- lakdl a> ci|Ual In tha' nl the i) Id'J pri-ni Thi- i~ piiiliaMx nut exact Imi -tiliicicni ly Lrnml I'm- the pci-ciii piupii-e. In the I'lilliiwnin talile aie i;i\cn I'nr ci^lii \\a\ e-leiiLit h~ in the phiitiiiifapliic fcLiiiin lif^t nf all. the inea>iiiei| iraii~mi<-inn <>( Idd iniii, ul' (» Id'J and <) Jii:! !;la~s. ami then \ahie~ n\ (» 7_'_' e~timaieil I'inin t he run ul' the ili>per>ioii>. I'cliiw the-c ale ;:;i\cn the cuniputcii transmissions nl' the \arioii> optical part.- in the two ,-pecirojiraplis. TM'.i.i: It TABLES OF l.NTi:\SlTIK.S .\FTEU MiSi ilU' Tli)V TUlinriill C.I-.VSSES. MaliTiiJ. H. II, 11. um) 100 mm. O inj l(K)mm. 020.) 100 mm. () 72.; ■)7mm. ()102()Iil I'rism titi mm O 722 Ni-w I'rism Trijili't ('(lUimatnr iinii ('jimcra Sini^lo M:tti'n;i! CaiiHTii Isokuiiijitic CoUimiitor 740 ■ Sli."i ■ s7.'i M2 yu> ■ ((7:j slJ 620 • S2.i S2.-. •71)1 ■h.sl >Hi2 <.>74 7i;i ■ -..•io 72.5 72.5 li'.X) soil • '.«.s .'.H'lO I'.'n; .4SII I'lM I ti70 .ti.")S •77.5 itJo .114.5 0.5S 1170 .(«.5.S ■77.5 '.12.5 ■'.)4.5 ■ G.5S 4i;ii li'.HI (LsO •ti.W •7.>:> S»2.5 •!M.5 200 Ii20 i'i(KI 4(K) 71 1 iiltinii intcii.sitios after each lo.-.- hy rcljection ami ali-orpiinii in the components of each svstcm, ami uri' coiuuiiied in the I'ullouiim talile. r.FFIClKNTY OF srECTHnOH VPHS. 195 i AIM.l. Ill mmpi ti I) 1 mi;ki;i nt intknsities of light. lustrument. MOO I H^ 4200 I H| 4000 I K 11, liiii I ran.-inits a (•(iii>i(lcralil\ laiiici |.riv( iiia'j.r n| ilic iiicidciii liiilii tliaii ilic (lid. i;i\iii'.i an (iiitTfiiMit priicil aliniii .'iD' , Midiijii!- ill ilir lilnr. T.'i' , ~1 ioii;icr ai II.s aii>!clll of rvcll the lllOst ctlicilllt form ofs]icctn)- jivaph. and how important it is to watch even appaicntlv minor details, i or example, cemenlim; the tri]i|et collimator or camera makes the ditTerciice hetwceii tranMiii-~ion of n7:1") and 7.")1') a 'iain of \(',' , . while cementing both ui\es ai iiicrea-ed transmission of A'>' , ■ Similarly the loss by absorpliuii in the i--kumatic collimator is soiiu' ;.{!)', fireater than in the triplet a- niveii m the table above, and this estimated loss is fullN borne (jLit b\- the experimental results, ]':\lierimental tests of the ri'lative eHicieiicies of the now and old optical systems have been carried out in two ways I'irst by comparing a number of pl.ates of the same stars, makie.n allowance for the differences in seeinii. b\- the two systems; second l)y makiiifi direct comparative tests directU" followinj: one another on the sun ami stars in exactly the same way a- u:t> iierfornied with the ^ratiim spectro^^raph. Tlie latter method of course nives result^ probably mure reliable and ceriaiiily more directly coniparable than the I'ormer, 'i'he maimer of compariiifi intensities by the tirst method was to make with I he new optical -ystcm three spectra with exjiosures in the ratio ot ]. 2. 'A on one plat, of one star and repeal this for a number of stars. .\bout ten plates of each of the same -tars with the orifiinai spectrotiraph 1% IT'tLUAlIONS III llli; l)i>MINH>N nll>h,u\ A I < ili-i . weri' tlii'ii coiiipaiTii with thc-f, and it wa- cumparativclN (■a>y to oliinatc rclialih thr rdaiivr rspu^iin' titiirs rtM|Uircil i" ii'wv the >aiiic iiilciisily ot !^|1(■(■tnllll at tlif (lilTcifnt ua\ c-lcntii li>. H\ llif -ccoikI ini'tliiMl. M'\t'ii III' ciiilii r\|Mi-un- (Ui (Uic |ilatc witc iiiailc with carli lonn of tlir iii>t iiiiiicii! im llic -aiiic >tai- I'nf cxaiiiplc, oii -, (Iciniiioniiii. inai;. _'■."<. r\ii(isuir> of ' _., 'j, 1 1 ' _>, .'. 'A. 1 1 _. miiiut''s wrrc iiiadr side li\ -idc 1)11 a plair with ili'' < » 7_'2 lui-iii ami ilic triplet (■(illiinatcii' and caiiiera ohjoctixc- in the -pf(ii(ii:iapii, l-Apiisui-o nf 1. 1 ' j. '-'. •'?. 4, li. H niinilti^^ dii-cclly t'ulluw inti cm the >ainr -tar with tin- (» HI-' pri>m aiidtriiilct ol)j('cti\(- were madr mi aimilirf plate and tlic-c hot t Apii-iiii- \wi-c rc])catcd with till- tri]ilcl ciilliinatiir I'cplaccd |i\ the ixikiiinat ic. All ihi-i'r iilatcs were dcxrldprd tci^irl hrf inv the -ami' tiino. 'l"hi- prnrcdiirc was repeated in the reverse order tor i> I.etniis and similar exposures were made on the da>liM:ht ^k\ . It i- e\ident that niimeroii> aeeurate eompari-ons of the relatne inten~it\ at an\ w a\ e-leiiiit li i-an easil\- lie made from >uch a series uf expusiires. Tlie I'ollowinii talile contain-- a siimmarv of the mean vahtcs ol)tained by tioth niethod> r\iti.i. i\'. MKAN VAi.i I-.S (U i;i;i.ATi\i; K\iM>siiu:.s at KK.iir w avi:-i.i:n(;tiis Oiiiiral Svstpms M.>!m.,1 No Coniparcti Kx- pu8uri' Triplit :i () 10.> Triiilel () 102 ls6 isl ll.i j;i'.i 172 I Isl I 1172 07(1 KKFH IKNCY i iK sl'Kr IHodH AI'II^ 197 In itii ni tlir pr.rr.ii.in tiililc we i^ct the relative exposures exp.Tl.ni'iitally ,l,,„.nnine,l for the new and old o|.tie:il sy>teins. and it will l,e x'.Mi liuw marked a saving in exposure time i^ tTeeted. Th.. new >ystein re.,uire,s less than two-lift lis the . \i)oHire tune of th.' old at Hv oiu'-(iuarter at Us and only one tunth at K It must not be torjiotten however that the disp.M-ion of th.' new ^vstein is only three-fifths that of th.' old, and to make thetn ha- been done in the tollo\vin^ table, wliere i- al,-o uiven the ratio uf the coi.ipufd inten-ities of the •■merficnt li^!;ht from the two .systems obtained bv dividin^i the emeriicnt intensities in Table III. TAIil.i: v HI.I..\TIV1. (■..MIMTII) AND i:\l'l .I{1MKNT.\LLY DETERMINED EFFICIENCIES H ff ExiMTiinfiit;il iil.l Ni " ('oiiipiited ( ilil Ni » 4600 ". 4200 ». 4000 1 ""^ »f ■7Xi 1117 tlJ4 ..111) .')S!i ■3»8 ,'>S1I .287 ! 187 ■245 ■".7 019 There is good agreer.ieni in lieiieral in these fiKiires, and the deviation between XI2()() and MOOO is probably mostly due to insufficient data as to the ab. more gradual in its progression from 11.^ to K than is given in the tables. Sueh a i)rogres>ive run of the absorptions would make excellent agreement between .•omputed and experimen.al values. There is another factor which will influence the magnitude of the experimental values so far as the relative values for different wave-lengths are concerned, and that is the fact that all wave-lengths of the light forming the star image are not in sharj) focus on the slit. Owing to the colour curve of objective and correcting lens there is a ditTerence of over three millimetres in the position of the star focus for H-j light and for that at the ends lys I'lHI.K AlloNS iiK \l\y. DiiMlMiiN nH>V, |(\ \ I nH\ (if the >|icclriim ll.-i aii'l K The -lit \\a> placrd -o that liu'lit al wavc- Irimtli- a!>niil \|llll(l anil XHi.'iO wnr in Inni- ii| ii,alii| Talilc \' -h^\^^ t'vidcnci' (if ihi- ill the hiulicr lAiv-iimcnial \aliic ai XltidO. It may t>c |i(iinicii (Hii iliat '.V in 'l"ai>lc W i;i\c~ the relative e\liii-iir(- when lri|ilcl an I i-(il\'iniat cdHnnalor arc iiilei-clianncd. and lliai lhi>y auree well with iln' c-iinialed \aliie- of the aK-drpt ion of the isdkuinat objective in TaMe II until uc Liei near l\ Hire it seem- that ihe ali-erpi icn nf the liKin-iiieate llini of the i-okiimat dit'ler- irom lliat ol liie () 111.' and pi(idiic(- a more t;radiial (halite. I'.oili -el- oi limu'e- toim -tlilum; e\ idenee of the lin-uitahllit \ ol thi- (ilijecli\e for -| leet r( iirraph- In addition to ilie liain in e!lieienc\ another i:reat aii\aniai;e ol thi new ((ptical part- i< the iiniforinit \ in the inten-ity of the phoiojj,rapheil carls -i\pe -peetrum Willi tli( It 1U_' pi'-m. in oriler to olitain -utiicient c\po.-iire on the K line III make it mea-uralilc. the reition around I1-, \\a- -o much onci- e\po-cil :i- lo Mock up the fainter metallic Hik - 1 iirl liermore. as can lie -cell in Plaie II. the -p( ct ra from ihe li^ht Hint pri-m contain four or h\e more mcasuralile h\(lroi:eii line- than the other-, thus con.-ideralilv increasing; tlie material a\ailal)l( lor iik a,-urcnieiii in -tars with few line-, Tiie results of thi- in\ cstiiiat ion ma>' lie -uminarized as follows: — 1. Tli(> dense -ilicate flint, () 10_', t:la— aliiio-t universally cmjiloyed as (jrisiu material in -tdlai- -pci ; incraph- ha- lieen -lioun to i'c loo hiuh!\ absorbinii; all aloim the photoiiraphic -pcctrum and especiall> toward- and in the ullra-\iolet I'or the In-I ic,-ult- in radial \el(icil\ work. This is especiall> the cu.-e in -intile-pri-iii -pect roiiraph- empl(i>c(l on earl; ;. pe stars. •J, 'llie suli-tituiion of a harvta liuht Hint pri-ni <) 722 for the •» ll)_' has cau-ed a decr( a-c in the exposure time- rc(|uire(l lo produci' ihe .-ame iiiteir-ity oi sjiectruni wiieii lioili are reduceil in the --anH di.-per-ion i of 22'', at II-.. 4,s' , at H,- aiid 7(1', at K. besides ci\in,ii con ideralil\' more measurable material in the ultra-\iolet without over-e\p(i-ure .iround H-,. IIKli IKM ^ Ml -CK' IKDCK \I'M-. 199 :i The Mii.-iitiiii..n Ml :i l^ni-licar 'Inplrt for the iMikiiiniit Cullmiatdr Ol.j.M'tivr iiM-, owiiit; I" tlir -troiii: ;ili:-(,i-pti. i ilir liiticr. cITcctcd a fuitlicr si\ iiin 111" aliniii :!()' , _1 ■yUr .ciiHiiiiim "\ ill,. (•(iMiaci .'urv.- -i the inpl-'t n)lliinat.)r and ,,,ln.la nl,;,, .nur- ,lil,.illl-lH'. III.' In- l.S IvHrct ion uN ,T uliat unill.i n.riir with uiictiiicntt'il Icii-c- li'. (i\cr :',(l'^. .-, •I'll,. ratKi- (if ihr iiiiniHiH- of ilic rinniiriii pnicil- fniiii ihr at llo, nl'.t at K, aii.l th.' actual ralio- nf the cNpnMlir- rninilcil nwillli tn tlir ihrrc-Hflh.- .li>p.'i>i()n „f tlu' 11. u system aiv (I :;7 at I! , D -M ai 11. an. I nil at K. AltlKumii thr -niall.M- .Ii>pcrM,,n uill lik.'ly .niail pn.p.ul mnat.lv lar^iir pfoiiaM.' m-or- t|„. fart thai n.arlv all in.' mat. rial a\ailalilc for the present r,|uipiii.'nt lias I.e. II w.irk.'.l .,\.'r an. I ihai ih.' new .'|>li.'al part^ niaM.' .-lafs at IcaM a iiiat;nilii.|c faintt r t.. l"' ira.'hci. -limihl jii-tifv tin .liininislic.l a.'.uracy. C. l-iiiall\ the in\.Mi^:atii)n lia> -hewn tli.' iniiMirlaiMf in >flhir >.)i(rlr(.sc.,py. where th.' litiht i- alway- nieay;re in i|naiitity, nf ><> -electing; the materials aii.l .l.-itrninti the nptieal parts of a stellar si.eetroiiraph that all the l.isse- l.y retlectinti ami aiisnrptinn nia\ lie inininiize.l : ami the ri'stilts imli.vile what a tifeal Nivin^i in expn-iire time ami (•.)iise(iuent increase in outpul ami ranjic may thus he elTecteil. DiiMIMdN On,-KH\ Al'ill^ , Oi rAW A, reliriiatv \'.n\ I — r= : - 'r.- l i--"s- 7 "a 2 5 i S S B