A Study of Electrically Exploded Wires Citation Smith, Sinclair (1924) A Study of Electrically Exploded Wires. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/TCAH-MA45. https://resolver.caltech.edu/CaltechETD:etd-06292004-104502 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): Anderson, John A. Thesis Committee: Unknown, Unknown Defense Date: 1 January 1924 Record Number: CaltechETD:etd-06292004-104502 Persistent URL: https://resolver.caltech.edu/CaltechETD:etd-06292004-104502 DOI: 10.7907/TCAH-MA45 Default Usage Policy: No commercial reproduction, distribution, display or performance rights in this work are provided. ID Code: 2767 Collection: CaltechTHESIS Deposited By: Imported from ETD-db Deposited On: 29 Jun 2004 Last Modified: 03 Oct 2019 23:20 Thesis Files Preview PDF (Smith_s_1924.pdf) - Final Version See Usage Policy. 1MB Repository Staff Only: item control page

California Institute of Technology Pasadena ,California A STUDY OF ELECTRICALLY EXPLODED WIRES Thesis for Degree Doctor of Philosophy Sinclair Smith 1924 HOT ch CIRECT ROTAPI NG TBROR THOTOTRATHS In tre followins paper, & nunber of references are made to direct rotatian.~airror photographs. These »hoto:-rabhs were taken with a roteting mirror cemera built by Dr. tnderson. This camere wes described by Dr. anderson’ at the Pas. dena meeting of the +mer, Phys. foc. of Hay 5, 1923, as the full description of this instrument is not yet in the literature, the foliowin: description is given, The general ervanvenent is as shown on the dijeeram, ANi image of source to be studied is ¢rojected on the slit of the instrument. fhe Lisht coming thru, slit passes thru the camera lens, end after having been veflected from one of tne faces of the octazonal mirror, comes to 6 focus on the film as shown. In use, the instrument isolates a narrow strip or section across that part of the source image waica fells on. the slit, and. registers on the film the changes that take place in this region of the source. 1.Phys. Rev. Vol. 22. p. 206, 1923, RECORDING FILM — 4 SPODY OF LACT IGei LY “SePLODS) VTRus Dr, Anderson hes shown thet electrically exploded wires furnich spectroscopists with a s urce of metallic vapor et an exceedingly high temperature, anc further that this vapor retains a high intrinsic brilliancy for & relatively long time after the excitation has ceased. as the spectrum shows 4 great variety of lines, the question naturelly arises, Bo 211 the lines appear .simltaneously, or ao different lines appear at different tines? In | order to answer this question, a new type of spectrograph was built Which ‘made possible the stuay of the variations of the -spectrim with time. a | | sRODS AND APPARATUS © | hes The essentially new: feature of this 1s spectrograph is a3 rotating mirror SO: "mounted that the besm etter having essed ‘thra. the slit and cea suffered “aiepercton, is reflected from one. of the faces: ‘Of the. rotating mirror, to the “focal plane of the instrument, Two of. these instruments were 2 built, one using - prism, and the mete The arrangement of the» optical parts: of the prism instrument is: shown diagrammatically in’ Fig. 1. Fig. 2 shows the “position of the photographic film relative to the rotating mintor. ‘This kind 7 | of an. ins trunent should be faster than ‘the ‘one to ‘be described, but ‘though a- number of good camera lenses were tried, no “lens combination was~ available > having "the desired characteristics. “That is needed, is & ‘combination which will | “accurately focus the desired range: of spectrum on a iene perpendicular to the axis of projection. A range of omy, about 100 A was the best that could be | done with. the lenses avilable, 80 this form of the - instrument wes abandoned and ; the concave grating type. was adopted, 2. ‘The arran: -ement of the slit, grating, and. rotating mirror, ‘is shown in Fig. By. while the ‘posi tion of the film ‘relative to the rotating mirror. is am vefore. » ROTATING MIRROR AND DRIVING MOTOR RECORDING FILM — = UIGHT FROM PRISM = OE: -GRATING hapa sowIw BWILVLO® : (See Big. B). <A four-inch, one-weter redius, conceve grating having 16,000 lines per inch wes used. The slit wes pleced two meters from the erating which . fixed the distance from.the grating to the focal plane at 66 Om., Z eeve & scale in the first order of 25 A per mm. The film wes held in & circuler holder having a 25 om. radius, The actuél]1 focal curve is not & circle, but a circle was found to approximate it sufficiently well. The octagonsl mirror wes used to insure a record of every exposure, As the film subtends an angle Alichtly more than 90° as seen from the rotating mirror, the light from the erating will elways fall somewhere on the film, regardless of the position of the mirror. &, iihen the instrument is in adjustment, the spectrum mist of course be focussed, and further, the ralings of the gratings must be parcllel to the Blit and at the same time the projected slit imares, i.e. ,the spectrum lines, mast be parallel te the direction in which they move when the mirror is roteted, The focuss ing ean be dore visually with e high vower lens, but this second combinetion of conditions,can not ve met so easily. The proceddure adopted was es follows, The gratiag was first rotated to a position such that the central image fell on the film. A film wes then placed in the holder snd exposed to e series of fi slit images with the mirror in different positions, ené finally a reference line wae put on the film by making = short exposure with the mirror slowly roteting. An aro was used for illumination. After the filn wes developed, by comparing the slope of che slit images with the reference line, the parallel »osition wes de termined. hes 0 ‘ation so oriented the slit that its projected image on the phot graphic film moved acouretely alone #6 its own length when the mirror wes rotated, The erating was then rotated back to its original position, ana the normel adjusted. By watching the spectrum lines with 4 high power lens end slowly 4 * rotating ti 28 grating about its no ered ice parallel te the “lit, 4, It is convenient to introduce a new ht this point As we are soing to-be interested in Gaterminis aifferences in wave as length, let us cell the smellest time difference thet we can cleerly determine, the 'time resolving power” of cur instrument. Wow since time differances are uping along the time axis of ou Ay Hy bate ey ef na ah fe bate ce ot Ri BS ce & - ity bey * 2 ia Ps} cus. , from some arbitrary zero line to one of the points im question, ond then batween these tyro thet the accurécay with wove one cor. rs along measure linesr Gistancesa is the only Limit to our ebility to determine small . oo shall time intervats, Dut if we Limit the time resolving power in this wey, ve =eFt be crediting the Tnetrument with powers it does not possess, A consideration of optics] resolving power males this clear. It is quite the thet regardiacs of the opbical vesolvi the difference in wevesleneth hem, F amy ctrun Lines Tt NES two linss ave quite clese to close son we shall mean by time resolving power, LF C ch gt eS res) rs rs fad gs no i ed iS @e tf er ke ® asurable time intervel between say the eppesrance of two but rather the smallest FRY) in txe behavior of some ~ spac ts “am Line ¢ 6 ill not be ables to éstect the chance, but if greater, wa will be well aware of the temporary iisappearae nee of the time. It eppears then, thet the time resolv: pover.of our photographie film to move its own width, Thus the time resolving power depends both on tre speed of the rotating mirror end,the width of the spectrum in the focel In order to limit the width of the spectrum, a horizontal slit was used as well as the elit proper. As tne Goncave grati oe light, the system is astigmatic. The conjuzte focus of a line in the focél plane, spectrum lines, is farbher from the erating than the consuc ake focus of the spectrum Lines themselveS. The spectrograph slit is of course at the conjugete focus of the syectsum lines, but the horizontal slit must be placed at the conjugate focus of the line in the focal plene perpendicular to the spectrum Lines. he width of this second slit must be GF cours’, governed: by the intensity of the light source. In the present work, the width used was such that the spectrum ocal plane. Fly Le was & half my Wide in t In most of the work to be described, the speed of the mirror was about 150 rev. per sec, At this speed the spectrum travels alofig the film with a velocity of about 450 meters per sec. so that, acsuming that uncertainties along the time axis of the film, due to the finite width of the spectrum and a possible lack of criti icel focus, do not exceed a helf mm., the time resolving power turng out to be of the order of one one-millionth of a second. In other words, if changes take place in the spectrum of the scurce, the order in which events teke place can be dete smmined, provi ided thet the difference in time is not lees then about one one~millionth of a second, | 5. he spectia studied were produced in the manner described by Dr, Anderson. PA Two different condensers were used, one of them having already been described 2 oe a tuderson, It was made wp of £00 sing py Dr. strength window ¢ £0 x 24 inches, having tin foil coatings 17 x #21 inches, The capacity wes shout . one micr ofarad »and in use, this condenser was chenged to 20,000 volts. The wire inches and 7/16 of an inch thick. A closely woven,geuze 32 x 3% laches was = a 2 : 27 Zt, > a 2 * inibeded in the center of the plates during the manufacture, aud e single wire prought out on one side to serve as a lead. ‘These pletes were then stacked in holders with the leads of alternsté plates on opposite sides, and the leads were then joined te comnon terminals. The resulting condenser could be chat eed to a much higher voltage than the ordinary form, and in the present work 4 voltage of 55,000 was used. During the asswmblying, the condenser plates were tested in pairs at 85,000 volts, but it wes not thought save to use the whole condsunser at this voltege on accouht of surge difficulties. 150 plates were used which gave a capacity of 0.6 microferad. As the intensity of the light varies with the oscilletions present in the explosions, for practica purposesy these oscillations furnished a very convenient time scale, After having onee determined the period of each of the volt circuit, as meesured by Dr. Anderson, wes 87,000 and that of the 55,000 volt 2 circuit, owing to much less inductence, turned out to be 185,000. Dr. Anders son cand that the meximwa current in the 26,000 volt circuit was some 10,760 amperes, On the same besis, the maximum current in 55,000 volt circuit turns out to be 65,000 amperes. 6, Some of the explosions studied were produced in air, while others were confined ies a slotted. block of wood. These blocks were ii dsep,wes cut the co electrodes were 2 inches apart. aoe * 1, The spectrum of a nuusber of slewents was examined in the region fron % 3000 tow S000, and very sreat differences were found in the behavior of different ssectrum lines. The spectra of Sm, Al, Cn, Od, and In general, the ere linss were found to be 3 Onn pletely reversed, while the spark lines frequently appeared in the aby ewission. The sperk Lines generally strengthened with ths peaks of the oscillations, anc. some of the spark lines appeared only on the peaks, Orel the spark lines diséypeared from the spectrimm when the circuit stopped os However, as may be seen from the reproduction, the spectrum lines showed almost . possible variations in their behavior. 2 ‘The lines were also found to behave differently with the two condensers. for example, when the explosions were produced with the £0,000 volt condenser, : — . . , 438 ew: . the two sperk Lines of Pb, 4245 and », 228) appeared as sharp lines and shoved no tendency to reverse. With the 55,000 volt condenger, these two lines pacame quite broad end diffuse end showed definite reversal, As a tule, the 1 le pectra opte te jne&. with the second condenser were almost entirely absorption Spectra, %, The spectre also beheved differently when the explosions were confined __ Perm the wa Ke wdacl Qed cane d ine slot when they were pr duced in the open air, ‘The difference was very marked when the £0,0CO0 volt condenser was used, but not so noticeable in the case of the 55,000 volt condenser. With the lower voltave conde ing the explosions increases the strength of the continuous spectrum end als % the tendency for lines to reverse, This tenuency is also present, but not so marked, when the hisher voltage G ondenser is used, for in this case the contin» uous spectrum is elready quite strong in the unconfined explosions and con- firming them ecds but little to its strength. 366r, Gontins . presence | of this continucly spe case of mest of the clements studied, it is quite strong thru out the thitd helg oscillation, end in some cases it persists as late as the fifth. As may be seen from the reprodus tion, divect roteti show thet by the time the thiva half ose t entirely continuous. dust how mach of this excitetion is electrical, and how much is due to tempersture, is of course not known. The fact that the continuous shows such © ite ones vita the electrical oscillations, does net necessarily preclude & cure temmerature 2 excitation since the effect of th the peaks of the cscillations be only to increase tne temperature of the vapor. The fact that the enhanced ] > en the continuous is strongest, suggests én intere sting appear most prominently w hypothesis regarding the mechanism by which continuous radistion is 1 produced, fost is, Bees the cuanta contribrting to centinucus rad re cll ite rau et ie oO 5 civ) @ cS Py So Ou at QO © ee when ionized atoms recombins ® When recombination occurs, we wou totel of the chanse in potential cnerey of the system ,plus the kinetic ex of the slectron relative to the atom, to be rediated in a single quentum. same value. the Einetic ener alnost further, . Mm iv recombine the electron m a I a Th PSO OL AIRE Uae & LSGUYOn rm one of 2 80 _ ite ne oan sovteiputed te the 1 of vers true or false this nut with the fox dias % On 7. o Sead. S572, 95 5689,72 6a 3740.80 B885,15 ue + (29 For convenien No trac to Cale ium ed cS A by tot 38 38 10 to nN & LEE 64,11 toon ae A245 «ts e of 4019.20 or og the lines 3 4168.19 3765.59 67.36 86.81 95.12 8816, G0 20.61 26.07 28.00 24, 40 yption Lines which remai t n throus 3840.61 60.08 72.70 78.19 4045.99 63 77 71,96 4118.70 36426 follows: ent th 4.143, 59 81.94 99,26 4E08.E0 £260.68 TLS 4825.97 SEL 44.04.95 Gir existences. bide Glass It ig - Glasez IIT “Pure absorotion lines which © gertain = the explosions. C2 ei a ta ta ee sal, but quite pronounced 3 with the oscillations Le . ‘Olass VIII Diffuse lives showing no trace of veversa aping only ox the peaks of the oscillations. Class. IZ Sharp lines showing no trace of reversal and appearing only on the peaks of the oscillations. this classification is purély arbitrary, and in weny cases not sufficient, but = deteiled discussion of the behavior of each line is out of the ine weelr, strong. <A more bleborate system seems unmneeessary. In 4 fev cases, Glassification is given after the wave-leneth. In the remarks column, is list also noted in remeres colum, i TIN - a _ Waren 60,000 volt condenser (Explosions con- Length fined in block) Int. Class Remarks 4525 . . sf “Appears at beginning IMPURITIES 3984 Ca m I Appears in 2nd 3969 Ca m I " rn) 4227 Ca om I " nom 4737 C w Band Four heads show faintly in 4th, 5th and 6th MAGNESIUM Wave~ 50,000 volt condenser (Explosions in air) length ; an Int. Class __- Remarks * 3829 s Ili. Appears at beginning 3832 F III " " " 3838 8 III " " " 4481 3 Vi Appears at beginning; very wide at first (possibly 50 A} “ou rY us ITA “ a u us I a 80 «Leeh Ty a u ‘“ T T A au Hu rey yy T Led 74) 69 62 4ST Ul S4eT suTseg ITA 98ST UL o4eT SUT Jeg I m 20 Pe6E uo au a I A BD 64TE pug UT ae SsUutseg I ry 20 «66STEe Sali ducal uo “a “ II & ST“IT 980g “oo a a TI 8 ST-IL 008? “tla a u TI 8 st-dg 8L9R 4ST Ut Buouys 4zeQ IITA wi 4ST Ut o4eT suvdddy A 3 OLPP . | Supuutveq ye saveddy AI 8 sur uurseq 4e@ sageddy LIT s pz-@L Li9¢ Stu0 puz pus 4ST 4e Saeeddy XI wa quesead 4ON 9g¢G28 “ “uo AT 3 B “ rn Tit s pe~dl 9992 FUT UT Seq Je savedd » AI s SULUUT#aq 1¢ saseddy TIL 8 Pz—IL FORE qnonsyy dieys fren Zupuutseq ye saeeddy Jf ut Sdt“ST [9z¢ woo “ u II tw sz-dt egge “ou rm " II w sg~dt Sete 4ST Ul Sel suTzeg It w sz~dT Teo sHLeWOE SSETO . Sa SyieMmey #8e10) “GUT ; (HOOT UT pouTs U4.3ueT {ate Url suctso tdixg) iesuapuod 4704 000g ~u0D suoTSoTdxy) tesuepaoo 4704 000‘Gg | -—OAGA “4S[ 049 UT ATJUTeZ sMoys PUBq “TY OF’ Zred SUL “SUOTRETIEOSO O04 UTM MoT eTsea peounouord o10u SMOUS SENTITP 1ey4ea SI 41 ‘UG ou. WEA Butuurseq edozeq ucy} faye, saeedde gogy *sueis0iqg0eds Mey @ UO se0er, saom se ATUO Sutmous Tort 4ue ysoute aeeddestp Ezy PU ‘SIG ‘OSPP SeUTT Haeds oyy, *dieys a4inb suremez *zeaemoy woTyzod Ter14Ued ely *Buruurseq ey} ut s3uta sptma Sa9n doteasp [96S PUE PPEE BOUTT Bu0s4s 044 ayy *ieulen Tensn ey. UT dm auod seUuT{T eg pue aeedde you Op SeUTT UeF0740U euZ YOoTE UL peutyuoo suorsoTixg Soe, i“ Ty au i’ a “u ua “a u XI a N T g oF a“ ai aay au “a ul ai a nu XT a N LUDD “&TUO 4sT ur saveddy YI A UIP DUS ‘pig ‘pug *4asf zo syeed uo saveddy xT wa N 9662 SaLGTHNdan 48G¢ @XIT soseyeg IA s pig UL 4st saveddy TA s $997 uo “oo u u u XT Ul OSSD uo uo u u “ XT a STSt wg 04 pug go syee’ uo saeeddy TITA A 0877 : ‘perzeddestp 4992 eWT4 eu} {nods SuTuSPTmM peyreu At 8 u u u I s 396E @ MOUS pus quonry4 aeaedde seurt, ssoug AL s BuTUUTSeq 4e@ sageddy Jf 8 PRES “pol 10 -W9G OG} TIQUM S4sTsaed pue suoIgs YY aytab sizecde 4t weum pag 34 [tzu . sapeddestp taqy. pue ast ouy ur &TQuUres “on oa aoo4w uoou4 u XI a ET9¢e dea savedde 47 seseo rey,0 ul ‘saieedde U4y pus pig ‘pug Jo syeed uo saveddy xT uw 3098 — “SED 9T Ueum Upp 6%} Juoge [TZUN sen °u99 40 “UL{UOO PUT SULT SSNJFTP peodq @ SE AST U4G 904 UL SULTT YI sse@lo 8 04 Buz - 64} UT Sivedde sUTT SITY, saseo amos UT TA s -—sorzeu ATTenpers 4saITI UT eptm L104 A s 2892 ‘S32 suey #seTQ “fur SyLouey GSeto “UT , . . Wysuet (ate Uy suotsoTdxg) zesuepuoo 4704 000‘0¢ (ate ur suotsotdgg) Lesusepuod 4TOA 0D0*SG¢S ~OAGH 4ST Ul O18 sutaeg T w 900g “uo a T wa Sotr pig UL sUuLseg T wi LEVD noon u Ita s Moo a us Pr) A s L8EP 4ST UT esnyiTtd ITA g 4ST Ul pesteaer ATOIT4Ua YSOWTY A s QPZP 94°60 STL seaeyod ITA pug Ul surseg I a B9TP "pag eu} 10932 AT SSeTO 03 sazueyod A s qnonzyy sieeddy tI s 8G07 *sounsodxs ews uo TB81sAcd Jo soelGZ “4ST Jo pue 4e sutseg IIA pug UL surgeg I 4 OZOPr a a au il a XT MB r) “a abe? ‘G29 03 Pig Jo syee’ uo saveldy XT a ia “ ESee "U48 02 pag Ut ; Upese 4ST Jo sutuurpgeq Ares 42 Sueeddy XI tw 4 4“ $98e "QSdT 3 UL e4eT Saveddy TIA pi S0ed4 Of 9822 A s u u a pug UT sutseg III s OLS A 8 a “ u pug UT suTseg TII s- $8 9e : : “pig uy A s PAtyZ UT sutsog “II s oA9S -. TVYJR SOULT AT SEBTLO oui0odeq SSUTT esoud A 8 u rf u “ou u III § OPIS re ; mo, A s Pag Ssulanp sueptTa ‘pug UT surTseq TIT s PL9S Fey 801) UT “Har emey SSeTO ay ne (HOOT UT peuTtsuoo UASUT ‘(4te@ UT SuoTsoTdxg} resuepuoo 4TOA 000*0z suotsoTdxy) sestuepuoo 4104 9004Gg “OLE | aat Pr ub ui I A BD 6127 u rr “ I A BD 80ET a u “ T AB BD Ser Fe) is i“ I & 89 6627 Fr) "“ T A BO 6820 Pig UL sutJog I a BO S8er Pig UT suTseg IIA ry pug uy suTseg I uw 0 LBP 4ST UI S1el SsUuLceg TIA a 4ST UT a4eT SsUuTseEg T 34 20 6962 HoH ui ITA AA ao u I a TY 296s pig Ul sutseg ITA A Bac Ut supseg I pt TY Pr6e 4ST UL eyeT SsuTdeg ITA a 48ST UT e4eT sutseg T w SO $262 uu “ u I & Wy pi.2e “pig Url surseg I a W9 8hze SET LT eOdt iasneg Sst “Fur Syaewed Sselo Jul ma : : | (MOOTG UT poursuoo U4sust {af@ Uy suotséeTdeg) rastepuoo 4104 000‘40g SuoTsoTdxm) resuepuoo 470A 000*GG ~SACH aver u “ “ T UW 2098 Surimtseq ye savoddy I w 669 “uo a u T a Olas no “ rr T A PEE “ao ‘a i T & 9272 4ST UL 348] sulrseg I a OSPS pug Ul surseq T pay ogee dagys £zea=Zupuurseq 4e szeeddy Tt a SEee a u u If A 4TSS esnestp ATIUAT [sasuTAUTZeq ye sagoddy I uw g0e¢ esnistp ATIYUSt [sBuruatseq ye sreeddy I m T6ze ui " Fe III 8 PLZ qnonuzy3 epta ATeA 8 TIT r) 82S rr eT} I a 80¢ce Supuutseq ye suveddy T Mh PETS “oo u us II ww SPTe uo “ n It w 92T¢ “oo a u It bs 60TE won “ a II w OOTS 48[ Ul O4BT sutsog II w ¥60E ul u uu I Ly 790 rT u “a T Ww 990¢ a u is T id TLOS ~ rt “ re) T A . 8662 4“ ry} “ I a T962 us is u II A geez SULUUTseq 42 sreeddy Tt A Peg “sreET Sselo °° 4ul HaswOY S880 “Fat oe {ape ay {HOOT ut peuTs _. WoysotdeR} azesuepuos 4704 000*%o2 =u0o UOTsoTdsy) aesaepmoo 4T04 OCO*GgG _ O4suet~eaey “ u “a AT M GOLF “a Fr) u AT 4 SL9F u au a uu " a IIA a “u 4“ a AY Ww TSOP u 4“ a a“ “u “ ITA Ww a “ 4 AT ww L8SP 4“ u u uu uu a ITA wu “a " us AT a OFS u “ “a w 4 “ TIA Ww n u u AT wu OTS? “a ry rr & “ “ IIA A u “a rr AT @ OThY u “ ” Fe) u a TIA = u a a AT mW BL27 au “a “ “ “ “ ITA 5 SULUUTd6q Ye siveddy AT uw Gh3 u i“ a “ rm u ITA w 6937 rr ul “ H u “ IIA wa 6937 a rf u u u au IIA w BLTP ar) u u u u Its w Solr u u u u st i ITA Py POT? u Ts a “4 8 rr IIA A Teor , w a Pr a: u u ITA a 9L07 "£Tuo 439 8 WAG eon a “ Py u TA s eu, ul Sutaeedde eoeay o1eu 7 2909 an *ATuo u19 *% Wag S,TkosO WITH UoTetsea OU Ynonry} saeeddy 14 s em ur Suracedde goer. arem ¥ EZor sR ewOY ase10 “3a, “Steno n SSPID “UT ae (HOOTQ Ur peurzuoo UASuet faye Uy stotsotdxH) azesuepuos 4704 00% og ‘ SUOTSOTAxy) aestiepuoo 4104 000‘ . “OLBM A Wy ABSO-PTION PH«NCKENON 1, In the case of a number cf spectrograms of Al., Pb., Sd., Ca., od Mg.» explosions, a new absopption phenomenon presented itself. For a very short interval of time, usually for not more then about three ‘millionths of a second, that part of the explosion which was projected ; Onnd procuced on the slit of the spectroxraphy became highly absorbing no photorraphic effect. As may be seen from the reproductions, the lirht appears to be completely cut off for a very short time. In the regions very close to strong arc lines, this cutting off occurs a little later than in more remote regions )s0 that the general appear:nce is that of an infinitely wide absorption line rapidly becoming narrow, This delayed absorption is best illustrated by the aluminum spectrogram N® 1 » in which the Licht from the explosion first appears to be completely cut off except in the region within a few angstroms from the two strong arc lines 3944 and 3961. Then 4% time progresses, the vapor becomes transparent to wave-len:ths remote from these two lines, but opaque to wave-lengths within a few angstroms from them. Immediately after the absorption sets in, the appearance is of an infinitely wide absorption line having an emission center some 50 angstroms wide, with the emission core itself showing two sharp absorption lines upon its central revion. In the lead spectrogram N22. the same effect is seen except that in this case the lack of a continuous backround makes the ef*ect less prominent. In this case, the arc lines 3574,3640, 3684, 3740 and the very strong arc line 4058, play the prominent part. ‘The effect shows quite a vuriation, both as regards duration, as can be seen from ' the series of aluminum spectrograms, and ss to time of occurence, In the case of one of the lead explosions, the effect occured about twenty-five a millionths of, second afte: the befinning of the explosion. Ordinarily this Bdedrmieritin, A 3500 - Ato00 Er plracon i Bay beck ABSbo — Xr 4500 Li. effect oceurs st the minime of the oscillations, but not always. In one of the lead spec tro: rams, one of these "absorption bends" was found on the peak of the third half oscillation, ‘his particular one was of very short duration, somewhat less than two millionths of & second, 2. It is very probable that when this absorntion occurs, it only occurs over a very small area of the explosion, “irect rotating mirror photogranhs of the explosions show that in the case of the elements which exhibit thi a’Fact, the explosions have a "mottled" appesrance. Small a eas appear whien spparently cive off no light, or at least not enough to affect the photo. raphic film. The area of there "spots" is only a few syuare mm, It ssems very probable that these"spots" are merely revions in which the drecrrbesh On nh absorption at the preceding paragraph is occuring. it would be very strange indeed, if the whole vass of vapor stopoed emitting li-ht for a short time, ond if this ever cecursdy the Girect rotating mirror photo: repns “fbula certainly show the effect. No such effect was found, so it seems reasonable to conclude that the absorption occurs only in small areas «t any one time, 3. No explanation is offered for this phenomenon. There is a small possibility that it is a kind of photo~electric absorption, but while this misht explain thre eeneral ebsorption, it is difficult to see on this basis why the absorption should be "delayed" ir, the vicinity of stron: are lines. The fact that it appears to occur only in small regions at any one time, mi-ht be explained in this manner? One might suppose that for some reason a large percentay ne f the atoms. in some small resion «ot int> the same state of excitation. °o ¢ This stste micht be cone from which the stom could be ionized by the absorption of a quantum of visible light. In this case we mi ht expect the vapor in this small region to be opaque to all wave-leu:ths below that one having just the 12. Lng ang tl ioniaine eneriyy Em other words , all wave-len- ths below the tee weve, limit, Tn the case of alusinum, this long wave-lensth Limit was looked for. By using a wide slit end red sensitized films, the spectrum was chotorraphed out to A 7500, but no long wave-length limit was found. This of course has very Little pearing on the explanation, for even thouzh the absorption were of a photo-electric nature, one mihgt have a mixture of atoms in various states of excitation doing the absorbing. In this case, as atoms in different long wave-length limits, we would not expect a very definite rei limit to the absorption. However, even if we grant that a Photerciectric kind of absorption can play a part in the above, we have not explained "delayed" action in the neighborhood of strong arc lines. Hence, to the writer at least, this photo- electric type of absorption is very unsatisfactory as an explanation of the above phenomenon. In conclusion, the author wishes to express his deepest appreciation to Dr. J. A. Anderson, under whose direction this work was done, iH as \ Astrophysical Journal 51, 37, 1920. 2 4 ve ¥4, 16, WW2a4 During the study of the explosions, it developed that some metals were more difficult to explode than others. ‘The effect presented itself in the following manner. ‘Shen wires of copper, silver, or gold were exploded, unless the leads to the electrodes which supported the wires were well separated, a spark would wass across the space between the leads and a part off the discharge would thus be shunted ground the explosion, However, it was found that by keeping the leads well separated, all the discharge could be made to go thru the explosion. In order to investigate thds phenomenon, a sphere gap was placed in parallel with the wire to be é¢xploded so that a me=sure could be made of the maximum voltage built up across the vapors of the explosion. ‘The electrodes which supported the wires were arranved so that their separation could be varied, and measurements were made usin; electrode separation of from one to ten cm. The measurements consisted in determining the minimum separation of he sphere gap which would force all the discharge thru the vapors of the explosion, The 55,000 volt condenser was used, Curves wers plotted for @ nuwber of elements; plotting electrode separation, ise. length of wire exploded, against the lernth of the sphere gap. A few of the curves are shown, They all ap: eared to be straizht lines thru the origin, As &@ basis for compsrison, the sphere yap seperation in mm, correspond- ing to a five cm. length of wire, is listed for all the elements stuiied. On. 37 om. Wi. 31 mm, Al. 18 mm, Li. 14 mn Ag. £9 mm. We 2e mm, Pp. 17 om. Fe. 13 mm AM. 2n. ° 2a mn, Sn. 18 mm, As the supply of gold wires gave out before sufficient data was obtained, a definite value for gold cannot be given. However, the evidence on hand indicates that the value is very close to copper end silver. The fact that these three "high" elements belong to the same group in the periodic system, suzgests that elements in any one <roup should bahar. alike in the explosions. However, s g&ance at the elements just listed shows that the avreement rather haphezard, Two one and one hald inch spheres were used for the gap, and the voltaves listed on the curves were computed in the usual way, The fact , that some of the voltages come out much higher than the voltage used on the condenser, is not surprising when one vemembers that we are dealing with rapidly changing curvents. In order to determine just when the voltage across the explosion reached ie maximum, things were so arranved that lizht from both the gap and the explosion fell on the slit of the rotating mirror spectrograph. It was found that as far as the spectrograph could determine, the becinning of the spark and the beginning of the explosion were simultaneous, Probably the wire first vaporized and then developed a high resistance, but did not become sufficiently luminous to photograph, until this high resistance stsge was reached, If the sphere gap was set near the critical position for the wire in question, only part of the current of the first half oscillation passed thy curred bhru the gap. The balance of the current of the first half oscillation and all the subsequent oscillations passed thru the vapors of the explosion. If the gap was made sufficiently short, practically ail the discharge could be made to pass across the 8p. The mass of the wire exploded seems to have very little effect on the determinations, as wires of various sizes of the same metal, ;ave quite similar sphere gap settings. Screening the sap from the light from the explosions, ‘made no noticeable difference, It is thought that these determinations are proportional to the resistance of the various metallic vapors.