Historical Document · 1940
Friedman's report on how the Signal Intelligence Service broke the Japanese PURPLE machine in 1940, written weeks after the break. NSA REF ID A58137.
• :· " PRELIMINARY HISTORICAL GN . S 0 L U T I 0 N ~sEeRET REF ID:A58137 • 0 F REPORT THE T H E 11 B11 / William F. Friedman, Principal Cryptanalyst. October 14, 1940. I Declassified and approved for release by NSA on 08-14-2013 pursuantto E .0. 1352a ..... -SEGREf- -SECREf" li' THE II B11 .ill..CHI~"E Part I. - Technical 1. In the latter part of 1938, messages appeared in a special secret Japanese cipher giving the authorization for travel for' a "Communications Expert" named Oka.'1loto, in order that he might put into service certain cr;-ptographic paraphenalia termed by the Japanese diplomatic offices as the Type "B" cipher machine. This machine was to replace the then currently used Type "A" machine for highly secret com.'llU."'l.ications a.~ong the important Japanese embassiesl throughout the world and the Foreign Office in Tokyo. On February 19, 1939, a message bearing the date of origin as February 18, 1939, in superenciphered code (K-1 tra.~sposed and enciphered by special A-machine procedure) was intercepted and was found to give the effective date of the initiation of the B-machine as February 20, 1939.· The A-machine was still to be used by all holders for certain classes of communication. 2. .Among the first messages received after the effective date of the B-machine were three messages~originating in Warsaw,which had a new type of indicator instead of the normal "A" type indicator. Examina.tion of these messages showed that they v1ere definitely not "A" type_ messages, but due to the fact that six of the twenty-six letters appearing in the text of the messages were abnormally high (as they would have been had the A-machine been used for their encipherment) it was assumed that the messages were prepared by the B-machine and that it was a modification of the basic A-machine. Further intercepts tended to corroborate this theory. 'l'he A-machine was continued in regi.ilar use at Hsinking and Shanghai and very occasionally (apparently when the B-machine was out of commission) the A-machine continued to be eraployed at the places l~1ich had been provided with B-machines. 3. After a brief study it was confirmed that the division of the letters into two categories (one group of six letters and another· group of twenty letters) which was the basis of the cryptographic treatment in the A-machine was retained in the B-machine but with a very important change. Whereas in the A-machine'the 6 letters comprising the 11 6 1 s 11 as well as the 20 comprising the "20's" were enciphered by means of what had been deduced as being a rotating commutator, whose stepping was controlled by a break wheel of 47 positions with certain skips in the cycle (the commutator could advance 1, 2 or 3 steps at a ti.me), in the B-machine the "6's" were enciphered by means of a series of 25 hetergeneous and differently mixed alphabets which were constant in their nature and cyclic repetition. These 25 alphabets were merely a carefully selected set of 25 of the possible 720 permutations or transpositions of 6 elements taken 6 at a ti.me, and a 1 lff SBiRR 'Washington, Berlin, London, Paris, Moscow, Rome, Geneva, Brussels, Ankara, Shanghai, and Peking. SECRET - SECRET deciphering chart or "development11 was C·~nstructed to correspond v.i. th these 25 permutations. This chart was revised and corrected from day to day until it· pecame certain that all its elements were absolutely correct. 4. This having been accomplished- (by April 10, 1939), it became possible; as a result of cryptanalytic· technique elabcrated for the purpose, to decipher the "6's" in practically every message of any considerable length in the B-ma.chine. It lvas found that so far as the 11 6 1 s" between two messages with unlike indicators were concerned, the onl~ difference bet-A·een cne indicator and another was the starting point in the cycle of 25 alphabets. There were 120 different indicators but only 25 different starting points, so that four (in certain cases, five) different indicators r.epresented the same starting point. 5. !lhen the 11 6 1 s" in a given message were deciph~red, the plain text values of cipher Letters scattered here and there throughout the text becanie available, so that the skeletons of words and phrases offered themselves for completion by the ingenuity and the i.maeination of the cryptanalyst. For example, suppose that on a given day the 6 letters forming the 11 6 1 s 11 were E Q A D R H and the follovving text was at hand: Cipher: Plain: 1 BR Ax s F/Q c Ev Q 0 10 x H b c FD L NH Q av Q pp L c ER p ~HE~A~A_E_E ___ ER __ E __ REQ_E ____ ijA_. It is not difficult to imagine that the missing letters are those shown below: Cipher: Plain: B R A X E F Q C E V C. 0 0 X H E C F D L N H Q R V t;; P P L C E R P • • • IHE~AfANE~EQQYER!hlE!IR£~gE§I&IHAT ••• In this process of filling in the plain text values of the 11 20 1 s 11 the cryptanalyst could be guided only by two things: (1) the positions and identities of the deciphered' 11 6 1 s" and (2) the context. For it speedily became apparent that any cryptographic relationship between the plain 'text and the constantly-shifting cipher text values in the case of the letters constituting the group of "20's" had been most carefully eliminated, disguised, or suppressed. This fact corroborated the conclusion drawn from all statiatical and analytical tests made on the cipher texts of the various messages studied. 6. The process of filling in the plain te:x,t values of the "20's" was therefore,as a rule, a very difficult matter, depending usually upon the particular assortment of letters constituting the 11 6 1 s 11 • If the text was in Japanese there was, in addition to the difficulty inherent in that language 'itself, the added perturbation occasioned by the fact that the Japanese Foreign Office had, on May 1, 19.39, instituted a species of "Phillips Code" in connection Vl'i th their use of the B-machine, with a long series of arbitrary letters and abbreviations sta.ndine for numbers, punctuation signs, and frequently used combinations of le~ters, syllables, words, and sometimes co~ plete phrases. For instance, the combination C F C represents period; C C F represents paragraph; the single letter L (not normally used in Japanese) ." . SECRE:l . represents the dir;hthone; ai; X represents ei; P represents ni; V represe~~s long U; Q T ~ represents Arita (shi) itashi tashij B KW represents Beikoku (• United States); T K \'J represents Teikokuseif'u (• Japanese Government); SN W represents Sukunakarazu, etc., etc. The difficulties introduced by this abbreviated or rather code writing alone vrere quite staggering as well as aggravating, for often the 11 text11 even when finally reconstructed appeared more like code.or a random assortment of letters than plain text.2 For the reconstruction of such text, the services of the Japanese experts were absolutely essential, and the w'Ork went very slovrly not only because of its difficulty, but. also because the services of these translators were available only a small part of the time Yihen the traffic for the daily "Bulletin" permitted, which was quite seldom. However, occasi.onally it was. found, after the 11 6 1 s" in a given messa~e had been deciphered, that these letters and their distrj_bution throughout the message gave good indications of the presence, in 1.qiole or in part in the message, of norm.al English text. In such cases,. the "guessing" process was lik~ly to be ccnsiderably easier because of the absence of abbreviations· (except for punctuation signs, in ~"hich case these were a help),because of the 2 A typical example of the sort of 11 text11 usually found at the beginning of messages is the following: 1 Cipher: F G P X P I X U D B D G E C Z L B L N U ZQO~H YNMRQ ARJOP Plain: X F C G J W F 0 V n· D N 0 B B F Y X F 0 C F Y L C C F ~ S G T S J VR DEILO KHIFI AXPPP CGURV ZFGRA FELBK WTLSI LIGDK The correct grouping of letters and the interpretation of the foregoing "plain text" is as follows: ·l·, Ais = 3 x F c = Dai - go G J W = 15 F 0 V = Open parenthesis DD = 2 N0 = "cf" (Jap.plain text particle possessive) BB =1 F Y X = Close parenthesis) F O C = Gokuhi (= Secret) F Y L Kancho fugo atukai C C F = Shin sho ( = Paragraph) = GT S : Getsu (month) J V R K = JUroku (= 16th) H I = Japanese word for "da:'.r" F I C = Begin kana spelling G U R V = Gurii (= Grew) F E L = Close kana spelling B K 1.'! = Beikoku ( = U. S • ) T L S I = Taishi (= .Ambassador) In running language, the message begins as follows: "Number 15 (part 1 of 2 parts) ~e·cret, to be kept ·;d thin the Department paragraph On March 16th the American Ambassador, Grew", etc. - 3 SECRCT SiCRr. l cr~'ptanalyst' s :zrea ter fe..m.iliari ty ..ii. th the lane;uae;e, and because of the availability of the services of a larger number of ~orkers. It happened that in several cases, after a few ~~rds had thus been obtained by pure 11 guessing", a clue v:as afforded as to the 3ener[~.1 nature of the messa:;e and this led to a frantic search for a complete document which might be available either in our ovm files or in the files of other government agencies. One case was found in which the B-machine message contained a paraphrased version of a messaee which had been transmitted in K-1 code. Ad.v&ntage was, of course, i.Ir.mediat.ely taken of tbis circumstance but the entire text of the B-machine messaee could never be reccnstructed from the paraphrased K-1 versjon, possibly because of the excellent paraphrasing, ~ossibly because of the presence.of abbreviations, possibly because of both. Certain ~nglj_sh text messa2es, hov1ever, were reconstructed, some of them to the extent of 90-95%, because the doct:ments being quoted in the messaees -nere fortunately located and obtained, most often through the cooperation and good offices· of G-2. • ?. In all, the ~lain texts for parts of some 15 fairly lengthy messages were obtained by the methods indicated above, and these were subjected to most intensive and e>:haustive cryptanalytic studies. To the consternation of the cryptanalysts, it was found that not only was there a coinplete and absolute absence of any causal repetitions· within any sinc;le me~.sage, no matter how long, or between two messages with different indicators on the same day, but also that when repetitions of three, or occBsionally four,ci~her letters ~found, these never represented the sa..~e plain text. In fact, a statidtical calculation 2avc the astonishing result _that the number of repetitions actually present in these cr;,rptogra.."ils was :..;;is ti:1an the number to be expected had the letters comprising them been -·~··:n at rarldo:::n out of a hatt_ Apparently, the mctchine had with malicious ~.:1: ent -- but brilliantly -- been constructed to suppress alJ.: plain text repetition. Nevertheless, the cryptanalysts h.sd a feeling: that this very circumstance nould, in ~e final analysis, prove to be the 11 uncloingn of the system and mec!ianism. .And so it turned outl 8. In all the foregoing studies, several factors stood out. First, the basic law underlying tbe B--machine was of such character that the ciphering mechanisms seemed to start from certain initial settings and to progress absolutely methodically without cyclic repetition of any sort, strai:;ht throu2;h to the end of the messages, the longest of which for which plain text had been recovered comprised over 1,500 letters. Secondly, two identical plain-text letters in sequence could never be represented by two identical cipher-text letters; nor could two identical plain text letters 26 letters apart be identicaJly. enciphered. This phenomenmwM.ch was termed "suppression of duplicate encipherments at the 1st and 26th intervals" formed the subject of long and arduous stuJy, fruitless experimentation and much discussion. Thirdly, two messages vrith identical indicators 0n the saine day appeared to be identically enciphered, and on direct superimposition shovred themselves to be monoalphabetic within columns, but v'ii. th the monoalphabets constantly, ir·regularly and unpredictably shifting from column to column. Fourthly, two messages with identical indicators . l~ . ' I i' "~I je'")'j' I I ! _, ~J LL~ L~ qi\~\,ll-':::1! .' Ii _..,LJi on different days (different plugboard arrangements into the :nachine) were absolutely different. Fifthly, two messaees ..,.ri th different indicators on the same day (same plugboard arrangement) were absolutely different and shoYied no cr.~';itogra1)hic similarities whatsoever. Sixthly, in each line of 26 letters, two identical letters could be identically enciphered except at the 1st interval, that is, identical encipherments cculd, and often did, occur ·within a line of 26 letters at all intervals, except at the 1st interval, althour:;h this pher.omenon was rare at the 2d, Jd, 4th and 5th intervals. 9. At the same time as the foregoing phenomena wer'e being studied, intensive research Kas.ccntinued in ·an endeavor tc establish primary or basic cipher sequences of ·the nature of those usually found in cryptographs with rotating cO.iJnutators, rotors, and the like, such as in the Hebern and Enigma cryptographs, our hl-134, etc.· For it .-.ras inccnceivable that the macM_ne employed a multiplicity of non-repeating keys of lengths corresponding to the lengths of the messat;es and, moreover, theoretical considerations eliminated the possibility that run..'1ing keys were being used. Somewhere, somehow, the existence of cyclically repeating keys or sequences must he uncovered before solution could be effected. But all efforts to disclose the presence of cyclically repeating sequences were fruitless. In one and only one case was there found even the slightest hint of such sec;.uences as were being sought. In a certain English text message the letter E was found to be represented by Q, 26 letters ay.,ay another E •vas found to be represented by Y, and aga:i.n 26 letters away another E was found to be represented by V, :nakine; the sequence QYV; in the very same message the same trieraph Q Y V was found to represent three I:'s simi..larly spaced. Attempts to add to this 'Q Y V sequence were absolutely unavailin~:. In this long, exhaustive and tedious search for repeated sequences or partially repeated sequences much labor and energy y;a.s expended but it was realized that the difficnlt~,r was probably due to the paucity of the text, despite the number and length·cif the individual messages c:.vailable for stu<ly nnd for which the plain text had been reconstructed. It b~~@.._§..marent .J.h~~-!'.{hat ·.~rould _Q.§..ll~C-~~~Cl~ ..!:@§ I-o obtain, ..Ex-some mann~~r.:._Q!:._O ther:,_~"!!":d:~l •ll.~~~~~J~Il- th.~.....§!1~.iil.dJcatQ..~ ~l'..!~ _on the same dav, or else ~9_cc~eri.__~~-~~:i;'_?.l mess.'.l.rre_s :·;i tf! t~~ ..:W.!!E!. indicator bu~_on differe..11!: ..day~-~Q.-~h.~-~C!i.!!~.J~.~§.e i. 2.~~or~___q_'!_~ th~istence__qJ: such cyg_J.:_~~O.J!.~11£~~~11.J~EL..£~ ec _te?.,!.. 10. In all the thousa.i."ld or 11ore messar;es on hand there were but a mere baker's half dozen or so cases ••here there were two ,nessa3es on the same day ;md in the sazne indica. tcr. ::lore than two had never been found a.nd this was to be expect~d in a system with 120 different indicators available for selection each day. In one case o.f tr.is ra.re phenomenon the plain text for one of th.a two m'3ssages v;as ·available but very little could be done even then as regards the solution of the other member of this pair of messages. For such a method of attack at least 20-25 messages all in the sane indicator a'.-.d on the sai11e day would be necessary and this was of course recoenized as a perfectly hopeless expectation. There re:na"!.ned the possibility of converting several messages with the sa:.""ae indicator but on d.i:'f-arent days to the sa:.ne base and 1"Thile this met:1od of attack looked extremely difficult it did not appear hooeless.._ ... -.... .-• ~ ~ - 5- 11. A method for this conversion to the same 1'a,se Vias (lev~loped and was ter31 ed "the identification of homologs." That ii:i, an attempt was to be made to establish that a given letV:!r on a certain day and another letter on a differe?'.J.t day '':ere treated in an ·absolutely identical or, .nore accura t.ely speaking, ho11ologou~ manner by the machine when .s<:: t to the sai,1e indicator. Tliis convers"Lon process is too involved to explain in this re.:.. port but suffice it to say that difficult thout;h it is, it was successfi.11 in two cases. One of these ~rielded a set of 6 .~essages, ,all in indicator 59173, which could all ~e reduced to the sc;.rne base. .'J.'hese f crmed the cru-. cial set of messages from the study of vrhi.ch success in solution of the 11Uichine \l'Tas fina.lly achieved. .... 12. Distribution tables of the letters constitutine the text of · these six messages were made. It should be stated that in f::mr of ti1ese six crucial messages only fraements of plain text had been reconstructed, here and there; the coinplete o~ nearly compl~te plain texts of only two of these six messages had been reconstructed. Hm·mver, enoueh data •·:ere accurnulated from these t·::o co:npletely and the other partially reconstructed messages to yield distribut:i.on tables vrhich, on careful examination, disclosed the presence of repeated sequences, ·here and there. This, on 3eptem::ier 2:), 1940, at about 2:00 P.TU., was the v3ry first indicHtion that a successful attack :night be possible. There was :nuch excitement at this first ;:rlimmer of licht upon a subject that had for so r:mny months been shrouded in cor:J.?lete darkness and regarued occasionally v:i th sorne d.iscourac;e:nent. The nature of the distribution tables rel:erred to is also too involved to ezplain in this report, but suffice it to indicate that they showed certain symmetries betw~en the successive cipher equivalents of a given plaLn-text letter and the successive appearances of that pla;n-text letter in the cr:-,rptogra::,·inic text. 13. As soon as the existence of cyclic or symmetric sequences became clear, attempts were made to uncover complete ba.sic sequences of the t;ype theoretically predicted. But many conflicts and inco!"lsistencies soon developed, due to the fact that the cr~c.togra.i:.hic laws underlying the shifting from sequence to sequence \'rere still unknovm. Concurrently V'ri th the work connected -..Ti th straightaning ·out and removing inconsistencies in these reconstructed basic sequences ran the v10rk of uncover·j_n~ the cryptographic laws referred to, and very soon the genera~ nature of the latter beca~e quite clear. All efforts were concentrated upon the development of the specific laws and specific basic sequences a.;!plicable to the indicator under study, viz, 59173, with a view to uncovering all the cryptographic phenomena in this case and. then 'searching for analogous phenor.iena in the case of other indicators. Certain qualified personnel .from other sections v:ere brought in to assist, a.."1d a consid.erable amount of night work vras · found desirable in order to push this study to completion at the earliest possible :noment. 14. :ay September 27, just one week later, the ViOrk had progressed to a point where it bec&i1e possi<:ile to hand in t•~o translations repreGenting the very first "solutioll11 to the 3-.m~ch:i.ne. Two .:nessages of recent dates, both in the 59173 indicator, were avaj_lable and vrere solved by ,, I' 0 - SEG'.iET SECREf- applying the principles of solution by homologs, guided by the aid of the reconstructed basic sequences. It was all the more gratifying that this could be done on the very day that annou..'lcer,1ent viras ;n.ade of the signing of the Tripo.rti te Agreement a-:i.ong Germany, Italy, and Japan. 15. Much i'irork re:nained to be done, ho-.•,ever, since only the datR. applicable to but one out o!' the w!10le set of 120 indicators were at hand. Tc solve the remaining 119 indiGators appeared still to present quite a large problem. These solutions ccnsist of findine; the initial settings of three 20-level rotary, electr~cal cryptcgra~:l:ic elements of 25 points each, and finding the crd.:3r in v;hich these three elements are brought i:ito play within each indicator sys tam. ~·;1th but little slackening of the pace set by th.a personnel the.11selves, work has progressed with vigor and at this moment soluti0ns for over one-third of the 120 indicators are available. 16. As to the 1n~chanics of the B-machine, naturally the basic principles of its construction and operation were deduced from the cryptographic phenomena observable in the messages, and immediately plans were initiated for the construction o.f an equivalent machine for our purposes. Orders for the materiel for 2 fully automatic machines were placed and ex!Jedited. \'ibile awaiting the arrival of this materiel a hand-operated machine vims desi•gned by personnel of the J-B section, constructed by them,· and is at this writing being used to assist in the decipherment of messages. Basically, the B-iaacl:"ii.ne consists of 13 rotary, 6-level, 25-point, switches of the type employed in automatic telephon;;r. One of these 13 switches controls the encipherment of the 11 6 1 s 11 and it goes through the same 25-point cycle over and over again as many tL~es as is necessary to encipher the messages. It has 150' crossconnectl.ons, \vhich, 8S stated above, had been established long ago. As to the 11 20' s 11 , these are enciphered by means of 3 banks of 4 s~vitches each, each bank having 500 cross-connections, makine a total of 1,500 sub-circuits available for the encipherment of any given letter. The arran~ements for advanc-· ing these svli tches is such that for a given indicator, say 20346, bank number 3 steps ccntinuqusly, bank number 2 steps once for every 25 steps of bank number 3, and bank number 1 steps once for ev;;ry 625 steps of bank number 3. This type of motion has been desien~ ted as a 11 32111 motion. For a..'lother indicator the order of this stepping may be different and in all there appear to be 6 different types of stepping: 1-2-3, 1-3-2, 2-1-J, 2-3-1, 3-1-2, and 3-2-1. Althoueh there .are 120 different indicators there ar~ only 6 different types of motion or step,t)ing of the three ciphering switches, so that it appears that each type of motion is representf!d by 20 different indicators. Vfr1at differentiates one indicator from another within such a set of 20 indicators are the relative starting points within the three banks of s·witches. These starting points appear to have been very carefully selected so as to preclude or reduce the possibility of 11 overlaps," that is, the production of two messages wr!ich in whole or in p.art have been enciphered by identical keying ele.~ents in identical sequence. Once these factors concerning all indi~ators have been established, and our machine is in operatLon, the reading of B-:nachine messages resolves itself into the establishment of the dailv plugboard arrangement, that is, the order and identity of the wireo leading from the keyboard into the cryptograph and thence out of the cryptograph into the printing unit. Cryptanalytic procedures for thi:::: purpose have al- " ready been established and tested so that this should not be a serious - - ~=;) 11 "i)i 11::-'\ r ·P ', ,) ! ~ r=t~ • , IJ ~ l.! • , I.I · 1 .Q. .. ! ' SECRET problem. Given a long message or several short ones of the same date, solution should be possible in an expeditious manner. 1'7. The solution of the B-machine has, as a concomitant, thrown considerable lj_ght upon the mechanism of the A-mach:l.ne ir.'i th the result that our present A-machj_ne will be modified in the light of these disco:veries and V'd.11 be made more efficient. The whole situation ·with respect to the cipher machines em.:.loyed by the Japanese Foreign Office now appears to be interrrated into a consistent scheme of development from its earliest and ~implest beginnings, about 1930, to its present quite complex fcrm. Problems concerning the l\-machine and questions for which logical and 11 reasona.bleu ans\1rers could not be four1d are now explainable on the sim}lle grounds or" the typ~ of cryptoerE.phic mechanism employed in the earlier machines, which seems to be the rotary switch used in automatic telephony, rather than the revolving commutator employed in our present A-macM.ne. Part. II. Credits 18. The successful solution of the B-machine is the culmination of 18 months of ~-ntensive study by a group of cryptanalysts and assistants working as a harmonious, well-coordinated and cooper.:.tive team. Only by such cooperat:i.on and close collaboration of all concerned could the solution possibly have been reached, and the name of no one person can be selected as deserving of the major portion of credit for this achievement. The parts played by the individual members of the t.ea."Tl. may, however, be ind.icated. 19. The specific direction and coordination ·of all studies on this project was the joint work of Cr;y-ptanalyst Fra..."lk B.Rowlett and Assistant Cr!:ptar,alyst D.obert O. }""'err.er. Their indefatigable J:aOors and brilliant analytical work testify and are a credit to thei.r cryptanalytic skill, training ano ex?erience. To their joint uirection and effcrts are due the extremely fruitful analysis of the cryptographic mechanics underlying the operation of the B-machine as r.t whole, the theory of its opere.tion, and tne developm:mt anC. solution of the 11 6 1 s 11 at any early date in these studies •. They were also extremely active in pushing the solution to a successful ccnclusion by orga.nizinc; a.11d d.irectinc the reconstruction of the developments or wirings of the switches for tne 11 20 1 s 11 • Junior Cryptanalysts Gcr.evieve ;;;,. Grotjan, Albert·.:. Small, and. ::3amuel S. Snyder did :uost excellent work in recovering the 11 6 1 s" during many months of apparently hopeless effort. In this work they were occasionally assistGd by Cryptogra:t:ibic Specialists Cyrus C. Stureis, Jr., Kenneth D. ~...:ill er, and Glenn S. Laudig. Of the la.tter two mentioned, it should also be stated that their very large output of 1'lork in the decoding of J code material, assisted part ti.11e by other inembers of the J-Section, lifted much of the heavy burden of •tr_is absolutely necessary current translation work from the - 8 - -SECREf ~· 11 . . . -- . c.-' 7" ' :~~ .......... u~ '\ • shoulders of the other members cf the J-section, thus giving the latter more ti.il.e for research on the B-machine than would have otherv:ise been possible. For the original suggestion of the electrical telephone S!"d.tching mechanism for duplicating the encifJhernent of the 11 6 1 s!', the d~sign and construction, in collaboration iTi th ~;:r. ftowlett, assisted by other me:nbers of the sect:;.cn, of the !lachine for deciph•3ring the 11 6 1 s:' and for excellent wor~ in the deciphzr:nent of the 11 6 1 sn in current traffic over many mcnt~s, CrJrptographic Specialist (no"'.'i' 1st Lieutenant) Leo aosen is to be mentioned. He also supervised and assisted in the construction of the hand-operated E-machine designed b;{ the cryptanalytic staff of the J-section; in this latter _:Jrojsct Junior C:r:'Ptanalyst H. F. ::3earce, Cryptogra.!,)hic Specialist E.dwarc.l J. !!aw::dns, Sets. -..-.!..lder end J.oy also assisted. In ti·,is connection, it should also be :aent:!.onec'!. that the shop f.::.cilities of the I:.adio Laboratory, :t~avy Yard, ,,.~ere kindly pl.s.ced at our C:isposal and certain .:nateriel furnished thereat, through the courtesy of Commander L.F. Safford of the Co:mnunications Security Group, Office of Kaval Communications. This ,::reatly facilitated the construction of the hand-operated B-machine. The excellent work of ·Tabulating :Uiachine Supervisor Ulrich J. Krapfl end Cryptanalytic f\ide i.:ary J. Dunning in p:::rfcnning countless tabulating ma.chine operaticns deserves mention, as well as the painst.a.king ~ork of the various card punch operators under their supervision. hliss uronenburg perfcnned and assisted in many clerical jobs of routine and special nature in connection with these st'\;..dies. The work of Resea.rch Analysts John .a. Hurt and Paul S. Cate must be mentioned in connection ·with our efforts to reconstruct the texts of messaees in Jarc·ei.nese. This di.f"ficul t \'I Ork had to be done in ;;hat little time could be spared from their regular and arduous duties as translators. The work of Cr~rptographic Clerk F'rances ~. Jerome in maintaining the f j les of the intercept traffic necessary to these studies proved. to be extremely helpful; her everytlay work in operating the 11 6 1 s 11 deciphering machine was painstaking and accurate. To Cryptographic Clerk Mary Louise Prather credit should be given for the careful keeping of the records and index of all messaE;es; it was also she who found the para.r1hrased K-1 message mentioned in Par. 6, and wrd.ch plc:.yed an important part in the final bre<ik into the system. After the initial solution of the 11 20 1 s" had been' made there was great pressure to hurrv the v;ork alone as fast as possible.~ / In addition to considerable overtime ;rork by members of the J-section; certa.in personnel from other sections of the ~.I.S. "1.rere brought in to assist. Some of them performed this vrork only c:s voluntary overtime. In this special assistance the na."!les of the following persons are to be mentioned: J._ssociate CryptaTlalyst Abraham Sinkov, Assistant Cryptanalyst Lawrence Clark, Junior Cryptanalyst Delia Ann Taylor, Cryptographic Clerk 11ilma z. Berryman, ancl·Cryptoeraphic Specialist Edward E. Christopher, .Jr. The vigilance and excellent work done by our various monitor sta.tions in intercepting and copying the necessary traffic also deserves special mention. The assistance rendered by G-2 in obtaining certain data bas already been mentioned. The undersigned was directed to participate in the "B" machine · studies in August 1939,· and from that time on these studies were under :bis ;;eneral supervision, at the sa'!le time that he carried on some of the duties fro-::i. which he had riot been relieved. In addition to this general supervision --;.a ·· 1·0·· {U ·f 1r11 1~1l IF '.;;i. .).J ~ ' ., .,. ~ ,) . rl .:! •r (_\. • " REF -SECRCT he also directed and himself conducted special studies T.i th a view to un- covering the cryJ,:togra.:.Jhic principles underlying the B-mscr.ine. He rd.shes to take t:·:is opportunity to indicate 11is-·zratification of the demonstration, by his assistants, of their i:;rasp of the cr::pta.11alytic techniC1ues taught them as a res1!lt of their participation in the trainine afforded by the Signal Intelligence School, as well as bis pleasure a.t the manner in v;hich the individual members of the team earnestly and wholeheartedly collaborated in this joint effort. He also wishes to express his appreciation of the important assistance rendered by the Officer in Charee of the Signal Intelligence Service, and his commissioned assistants, in expediting the procurement of personnel, supplies, and special information when needed. , October lL,., 1940 F. Friedr.1a.n, Principal Cryptanalyst. ~~:illia.m I / - 10 - •• REF ID:A58137 • ~, ... ._,