Reference / Paper · 1946
Summary version of the above; patent applications 682,096 and 70,412; SIGABA at Hyde Park and Yalta (6 pp.) [writer A]
• RE~~f)528 CQtJ'iRIBUTin~;s • IN THE FIELDS OF crn.n,;mucATIONS SECUhITY MID COl.iMU['.1;9~!.!QES IHTELLIGENOE 1. .&s Principal Cryptanalyst (1939-1940), Head Cryptanalyst (1941), then Director of Communications Research (1942 to date) I have had technical end staff supervision over a large staff (in 1911-5 a~ounting to almost 10,000 people) of cryptographic and cryptanalytic p1:1rsonnel working on many complicated problems in communicat.ions security end communications intelligence before end during '.1orld 1Var II. Hy specific contributions in these two fields are briefly sum··1!;! rized below. 2. ! iy contributions in the Communications Security field during the years 1939-1945 include prPctically all the systems and devices employed during Uorld ~ar II for crYJ'tographic purposes by the Jrmy anil the ~ajority of the systems and devices employed for the same purpose by the Navy and the tepartment of State. A detailed state1 ient is attached covering the fallowing: a. Converter M-134 end ri!-134 A, covered by patent ap~)lica t.ion (Serial No. f>82,096) filed by the Chief Signal Ofricer in my TIEl:r:ie as inventor on 25 July 19.3.3. This machine was the predecessor of the Converter r'I-1.34 C (Sigaba) and represented the first invention of P.Jectrical control, as distlnguished from mechanical control of a set of cipher rotors in cascade, thus getting away froI!l the regular or Metri~ stepping of the rotors. During the iM.portant years 19391941 this machine was used fo:r enciphering the bulk of the highly secret and confidential adcinistrative traffic of the ~ar Department in com.nnnice.tions with the Eeadquarters of Overseas Departr.ients, Corps Areas, Defense Co"1mands, and hee.dque.rters of GHQ Air Force and 2d Air Force. In addition, it was extensively used by the Signal Intelligence ~ervice in forwarding traffic fron our intercept stetions in Honolulu Fnd P,ianila. It uas also used during 1940 end 1941 for communications hetween the War Departl!lent and the u. s. Military Attache in London. In 1941 the '"ar I'epartr.tent pro\rided a nunber of these machines for the Oepartment of State, for use in secret and confidential comr~unications between the Secretary of State and the American Ambassador in London Pnn these were used from 1941 to 1944 for that purpose. It was also used in a special circuit for a number of months in 1942 for direct comnunication between the President and the Prime Minister in London. P.fter t.hese nachines were taken out of ,..ar Department service a number of them (29 or 30) were provided the Office of the Coordinator of InfoJ•mation (leter OSS) for secret COl!lll!Ullications betvteen Washington, J,cndon, and other capitals \'lhere the OSS maintained headquarters. Some of these machines are probably still in service. Declassified and approved for release by NSA on 05-31-2013 pursuantto E .0. 1352a ~\EdlET • b. Converter r.!-134 c, covered by patent application (Serial lfo. ?0,412) filed on 23 1.!arch 1936 by the Chief Signal Officer in the name of Friedman and Rowlett as joint inventors, arose as a result of studies having the aim of i.'llprovin:; Converter M-134 A. About 15 June 1935, Rowlett conceived the idea of using a set of rotors in the M-134 A. Rowlett and I then jointly developed the idea by setting down on paper various methods by ,.,.hich it could be applied in practice to the U-134 A. All Qf these methods were disclosed to the Navy, then engaged in attempts to improve their own unsatisfactory ?~rk I Em,1. 't'he l{avy took one of these Methods and incorporated it in tho design of their Hark II ECU, work on which was begun in January 1938 by Navy contract with the Teletype Corporation. This was done, however, without advising us or anybody else in the Signal Corps until Uarch 1939, when the Teletype engineers brought to ';'l'ashington the first completed set of dravlings of the Mark II Ecr.r, at \lhich time Rowlett and I were invited to the conference with the engineers. A first model was built end delivered on 3 February 1940. Further development was on a completely joint Army-Navy basis and on 19 June 1940 the Signal Corps added its order of an initial 85 machines to the Navy order. r n 17 f./Iarch 1941 the first 10 machines were delivered to the Signal Corps end were given a prompt service test, proving the machines highly setisfactory. In successive contracts the !rmy procured a total of 3392 machines and almost 2000 were in service by ~1arch 1944. The Navy also procured a larger quantity. In the Army the machines were distributed to all commands down to and including HQ of Divisions. They were also used in all important fixed headauarters in the Communications Zone, in all theaters and in the u.s. Tnienever and wherever the late President went during the VTar, the Sigaba went too, on the Presidential Train, at Hyde Park, Yalta, etc. For further information regarding its value in Joint Army-Navy communications, see the detailed notes attached. ~7e know that neither the Germans nor the Japanese were able to solve our Sigaba traffic, though we were able to solve their high echelcn traffic, obtaining intelligence of great diplomatic, strategic, and tactical value. In view of the foregoing, the Sigaba contributed materially to our success in the war. 1 c. Converter M-228 (Sigcurn, Sighua.d), covered by patent application (Serial No. /.43,320) filed on 16 Uay 1942 by The Chief Signal Officer in the nane of Friedna.n and Rowlett as joint inventors, was a cryptographic machine to protect teJetype communicetions, by providing for automatic off-line or on-line {keyboard) encipherment, trPnsmission, reception, decipherment, and printing of ~essages (in a single operation) at the rate of over 360 characters per minute, with high security. On 12 ~arch 1942 the first two models, constructed at Fort Uonmouth, were given a satisfactory service test. On 18 June 1942 the Navy witnessed a demonstration of the machine end decided to procure 200. By 5 June 1944 a total of 3200 machines hed been manufactured a:qd 24gg in service, including 200 by Navy. In ruay 1943 the machines were used in the United Ki11oadom to link together all u. s. Arr.i.y headquarters I 2 • in the Defense Teletype,vriter Network and these machines were used to encipher a tremendous volume of messages, including raw material for cryptanalysis fror.i. all intercept stations. Uost of the traffic that was sent by radio teletype was confidential, but on land lines secret teletype-messages could be sent by this machine. A modification (Sighuad) permitted use of the mchine for transmitting weather data (secret) by the Air Force in two theaters; the same modification permitted use of the machine for secret message~ between ceri,ain headquarters in Washington. In April 1941+ the ';"far Department approved a policy under which the J'llachine could be turned over to the British for use in Combined CoTllITIUnications. For further information on these machines and additional items relating to contributions in the Communicetions Security field, see detailed account attached hereto. d. Cipher Device ll-138, covered by patent application (Serial No. 300,212) filed on 19 October 1939. Thousands of these devices were manufactured. For several years this device formed the basis of the Strip Cipher System, which carried a large part of the secret and confidential coillJllunications of the Army, the Navy, and the State Department. In the AI'ITly it still serves es the back-up system for Converter M-134 C (Sigaba) and as the priPiary system for Posts, Camps end Stations as well as for circular messages to military attaches. In the Navy and in the State Department it is still used to a considerable degree ror secret and confidential traffic. e. Throughout the years mentioned, in my capacity as Head Cryptanalyst and later as Direct.or of Communications Research, many problems in security were brought to my attention and I believe that my long experience in the field formed a solid foundation for mature, sound judgment in arriving at practical and satisfactory answers thereto. Some of the items that may be menticned here are the following: (1) In 1941, as a result of my special study of the manner in which Army ~nd War Department cryptographic communications were then organized, I evolved and developed the idea of the 11 Cryptonet• system, which has worked in a highly satisfactbry manner in practice. (2) The studies and development of Converter ?11-209, over 100,000 of which were produced end distributed in the Army and Navy. (.3) The nstop-gap" or temporary-expedient system of doubleloop key-tape encipherment of teletype transmissions. (4) The "one-tir.i.e tape" or Sigtot syste?D.. (5) The development of voice security equipment, including the "5igsaly11 • . 1lsrm~A6os2a• ®Duib a The development of the ns;yn.chronous Polsrity Reversal System11 of Cifax, wM.ch is based upon an important modification (by Lt. Colonel Rosen) of the principles disclosed in my (secret) patent application (Serial no. 473,193) filed on 3 June 1943. f. I also was a member of the Ad Hoc Oonmittee, consisting of two Navy and tvK> Army members, appointed in 1944 by the Joint OoillI!l.unications Board to look into the rr1atter of communications security in e.11 non-Military departments and agencies; the work of this Comnittee resulted in the establishment by Fresident 'I'rwri.an of the Cryptographic Security Board, consisting of the Secretaries of the State, ~ar and Navy Depart~en~s. 3. My principal contribution in the ca:1JT1Unications intelligence field, directly applicable to our operations in ~orld ~ar II, was in connection ~ith the solution of the Japanese cipher machine (purple system) employed by the Japanese Foreign Office in its highly secret cOI1D11unications with its Embassies and Legations. .C.s Principal Cryptanalyst in the years 1939-1941 I was in charge of the cryptanalytic staff that studied this problem frcm February 1939, when the first traffic in that machine appeared, until September 1940, when we were eble to hand in the first translations. Ry careful analytical reasoning, long and arduous study of the external cryptographic phenonena exhibited by the messages, by correct reasoning, and a wide knowledge of cryptographic mechanisms we were eble to fathom the mystery lL~derlying the functioning of the Japanese ~achine and to construct, without ever having seen the original j_tself, machines which would duplicate the functions of the Japanese machine. So far as I am aware, this is the first ti'f!le in cryptansJ ytic history that a ~achine of such cryptographic complexity was completely reconstructed by pure analysis. As to the ioportance of that solution I need only rofer to the disclosures of the current Joint Congressional Investigation of the Pearl Harbor Attack by the Japanese and to certain state~ents contained in the Chief of Sta.ff 1 s letter to 1.fr. ret1ey. \1hile the solution represents the achievement of a cooperative effort by a nUM.ber of people, it ttaS made possible by good coordination, and proper technical direction of a fair number of skilled cryptanalytic personnel who were selected and trained by Me and who worked under my direction for over 18 months as a harmonious team. I do not believe that this machine was solved by any other cryptanalytic organization. ~e know that the very competent British organization failed in its efforts to solve this problem, for we gave them the solution Pnd a machine in January 1941. Nor did the German cryptanalytic staffs who attempted it gain any success. During the succeeding years, l 9/+l-45, t.he Agency accomplished oany feats in cryptanalysis, too n.u.inerous to mention. The diplomatic cor.iniunications of r.i.any countries were read, some almost in toto; the 4 "''( 'f~ffil'ET . • RFJ.dP;rA60528 n"'llC..l\l}\ffi I; U • secret communications of the Japanese Army and Lir Force wore read to a consii'l.ers.ble degree, contributing greatly to our victory in the Pacific. In my capacity as technical adviser to the Chief of the Agency, and having Staff Supervision over all the technical operations or the Agency, I was always consulted by him e.nd ""e.cted as advisor to all Chiafs of Divisions and Branches in these operations. The extent to which the A~ency engaged in the research, developlllent, Rnd use or high-speed analytic equipments to r~cilitate the application of cryptanalytic techniques and processing is worthy or mention, And my technical advice and collaboration was used in all these cases. 4. From my earliest days of duty in the Office of the Chief Signal Officer I have taken a deep interest in the preparation of texts for use jn training Military personnel in cryptography and cryptanalysis, and the War Department has published a series of such texts which were written and prepared entirely by me. I regard the writing or this literature, which was extensively used at the various An~y Signal or Communications schools, Rnu in the ~rmy Extensicn Courses, ~s one of my very important contributions to the war effort. I believe that this material represents an important contribution to the science of cryptology, because for the first tine its basic principles and techniques, hitherto scattered in a most chaotic, disorganized manner in foreign literature, were set forth in a scientific, logical, orderly and clear manner; and consistent, adeauate and scientific terr.J.inology used in this work. Upon them were also based a long series of graded exercises, with approved solutions, also prepared by me, which were used in conjunction 11ith the texts by thousands of enrollees in the Army Extension School, in the various schools throughout the ArMy during the war, in the special schools in cryptography and cryptanalysis at Fort rlonmouth (later at iJint Hill Farms Station), and at Arlington EaJl Stetion itself, to train thousands of new eMployees. All or nost 0£ these texts were also used by the u. S. ~avy, the u. s. Coast Guard, the Federal Bureau of Investigation, and the Departnent of State; copies were also officially furnished the Canadian and British Government. It ~as at my suggestion that the ·~ar Department, on 11 October 1930, formally established the Signal Intelligence School in 'Tashington, for training Regular Army officers in signal intelligence operations. I served as the ~irector of that School, in addition to my other duties, organized the 2-year course given, and directly supervised the instruction. The fact th.at of the nine Army grElduates ( tl~ere were two officers from the u. s. Coast Guard and they also worked in the cryptologic field later) seven came to occupy top-level positions in communications intelligence and com.munications security work during the war. In addition to the foregoing, numerous technical papers were written by me in my spare time; these were usually published by the ~s.r Depertment as secret or confi~ential docu.~ents, or they appeared ; ........ • as articles in the Signal Corps Bulletin (restricted). ~wo of the most important or these works ere entitled n.Analysis of a ?,!echanico-electrical Cryptograph 11 , in which I set i'orth the basic principles and techniques in the solution of cryptograms produced by electrical rotors in cascade, and nThe Index of Coincidence 11 , a revision of an earlier paper under the same title, in which there appears for the first time in cryptologic literature applicetions of statistical theory and techniques, later to become of great importance. 6