90,3NO. UNITED STATES CRYPTOLOGIC HISTORY CCH-E32-9S-03 TCS-54649-95 American Cryptology during the Cold Waf; 1945-1989 Book I: The Struggle for Centralization 1945-1960 Classified by NSAlCSSM 123-2 Declassify on: Originating Agency's Determination Required HANDLE VIA TALENT KEYHOLE COMINT CONll:l.Ob~m1VlSs1Jo5iIN-NTnLY THIS DOCUMENT CONT. D MATERIAL NOT TO FOREIGN NATIONALS This monograph is a product of the National Security Agency history program. Its contents and conclusions are those of the author, based on original research, and do not necessarily represent the official views of the National Security Agency. Please address divergent opinion or additional detail to the Center for Cryptologic History (E322). This document is not to be used as a source for derivative classification decisions. 'fOP !!(:FU!T tlM81b\ UNITED STATES CRYPTOLOGIC HISTORY Series VI The NSA Period 1952 - Present Volume 5 A merican Cryptology during the Cold War, 1945-1989 Book 1: The Struggle for Centralization, 1945-1960 " Thomas R. Johnson CENTER FOR CRYPTOLOGIC HISTORY NATIONAL SECURITY AGENCY 1995 IIANBLEJ VI'" 'f*LEJtff ItEJYII6LEJ e6!9f!NT e6!(TftaL S'iSTEMS JaIHTL i NO! l'tELfl:AStl:ftLE 'fa P6MJIEJ!\ ?fA:'ff8?HrcS TOil SECAiT ' 'M8FtA 19P SECRET YMBRA Table ofContents Page Foreword xi Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. XUl Acknowledgements xv BOOK I: THE STRUGGLE FOR CENTRALIZATION. 1945-1960 Chapter 1: Cryptologic Triumph and Reorganization. 1941-1949 World War II and the Intelligence Revolution 1 The Way COMINT Was Organized at the End ofthe War 7 TheCJO 11 The Cryptologic Allies 13 Chapter 2: AFSA and the Creation ofNSA The Stone Board 23 AFSA 26 The Brownell Committee 33 Korea 36 The Country 36 The Asia Dilemma 38 The Invasion 40 The Murray Mission 41 Counterattack 43 China 43 AFSA and ASA Operations 46 White Horse Mountain 48 AFSS Introduces Tactical Warning " 48 The Navy " 51 The AFSA Factor 52 Relations with ROK COMSEC and COMINT 52 Korea - an Assessment 54 Chapter 3: Cryptology under New Management Canine and the New Organization 62 The Early Work Force 63 Fielding the Field Offices 67 Civilians in the Trenches - the Civop Program 69 COMINT Reporting in Transition 69 IIM.BhB YIA ~·.hBUT IEB'IIIShB eSMUiT eSUTRSh S;YS'fBMS ofSUiThY N'6T ft~L~kSkftL'! TO It'6!tEICN I'O%:TIONkLS iii T9P§i(;RH~ .§ Ib) (1) .~; Ib) (3) -50 USC 403 Ib) (3) -18 USC 798 Ib) (3) -P.L. 86-36 TO' $I!CRET ~ 71 73 75 75 76 78 80 83 85 86 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 107 108 111 112 118 121 121 125 126 126 127 127 131 132 132 138 139 140 140 143 144 147 148 .....................' .--=....-=-...............,.....-----_..... The Far North Ji . 1 . What It Was Like . ________~~~--------~f· . SIGINT Goes Airborne . 'BLUESKY . Peripheral Reconnaissance . The Origins ofAdvisory Warning . The RC-130 Shootdown . Advisory Warning Is Implemented . The RB-47 Shootdown . . . . . . . o' . ......,.:o;r-""'I'I'Ilr-:-""l""'l":--':""""""......-o:"T~O;--"'I"""-'I"l""""" The T ird Partie.s inthe Ear y Years .•............................. CIA in the NSATrenches . , ......•..•.............................. J 2···/···············,································ NSNs Othe'r'c~~~tit~~~I·.' .' .' .' .'1.'.'[.'.'.'.'.':.'.':.'.'.....'.'.'.'.'.'.'.'::.':.'.'.':.' ELINT and NSA/ ................•..•............................... Building the Overt Collection System . 1 I· /........•...•................................ ~~LInUj'~~'~UD-~~~.•..•................................ 1""""';.;;...,;; ------1;;;..;;;;.="'-=;.....,.;•..;....;.. .;...~ ,' : . NSA Training - the Early Years . Setting Up Security , /./../ . NSA and the U.S. Intelligence System . '." i.i. .••.•.•........................ Consumer Groups Come to NSA / /./...•.•........................... The Struggle for Technical Control .i'i" .,..•.•......................... The Decentralization Plan i.i . Relations with the SCAs i. / /., .•.•.......................... The SCAs Create Second Echelon.s / /.. , .•.•.......................... Watching the Watchers ../ / /./...•............................ NSA and CIA - the Early Years i. i' •.. " . CIA Enters the COMINTBusiness . ~-----..-:-:-~L!:r:::::::::::::::::::::::::::: Ib) (1) Ib) (3) OGA IIANBLB Vb\: 'f2\Ll!Jf'f'f KI!J'df6L)!; e6MINT e6fofT!t6L ~ i ~Tl!:M~ JOiN I L i ~TQ'J' R8L8AS1tBLI!J 'fa faMfffl'f fofi!i:Tf6N2\:L~ TOP5EeRET~ iv Ib) (1) Ib) (3) -50 USC 403 (3) -P.L. 86-36 Chapter 4: The Soviet Problem The Early Days •. . . . . . . . . . . .. . . The Advent ofBOURBON .•............... . . VENONA ....•......... , ...•..................................... "Black Friday" ........•....•............. . . ASA and AFSA Turn to Radioprinter . . The Soviet Strategic Threat .•....•............... . . How It Began .......•.....•................. . . The American Response . . The Soviet Atomic Bomb Project . . The Chinese Threat •.......................... . . The Early Days ofOverhead . . The Attack on Soviet Ci her S stems . . Trackin Submarines . 157 158 160 168 169 170 171 174 176 178 179 184 187 188 189 Chapter 5: Building the Internal Mechanism Cryptology is Automated - The Story ofEarly Computerization 195 Antecedents ..........•.......................................... 195 Postwar Developments .•.......................................... 197 NSA Communications in the Pre-Criticomm Era 205 The COMINT Comnet 207 Securing American Communications 211 The Era ofthe Wired Rotor 211 The Early Years ofSecure Speech 214 Organizing for COMSEC in the Postwar World 215 AFSAM-7 ..........•............................................ 217 The Push for On-line Encipherment 218 From SIGSALy to Modern Voice Encryption 220 TEMPEST ...............•.......................................................... 221 Chapter 6: Cryptology at Mid-decade The Early Assessments ...•............................................. The Robertson Committee . The Hoover Commission . The Killian Board .•.............................................. The Jackson Report •.............................................. 1956 Suez ............•............................................... Hungary ........•............................................... I I··············································· Lebanon,1958 . 1956 in History . IIANBU VIA 'H.cFlP''f IEElYllabEl CaMIN'f CaN'fRaL S-"fSTEMSJ6II('fL'l NUl RELEASABLE 10 F"ltEIGl'4' Nit'fI6f4'1rLS 227 227 228 229 231 232 232 235 236 237 239 v 'fap SECRIiTjJ.MB'1t4 The Reorganization 239 The Move 241 Chapter 7: The Eisenhower Reforms The Post-crisis Centralization ofthe SIGINT system. . . . . . . . . . . . . . . . . . . . . . . . 253 Criticomm 253 The Baker Panel 256 The Reuben Robertson Report 259 The marriage ofELINT and NSA 260 The Kirkpatrick Committee 263 NSA Centralizes the Field System 264 AFSS and the Development ofSecond-Echelon Reporting 265 The Struggle for Control in the Pacific 268 Samford Joins the Agency 269 The Tordella Era Begins 271 Public Law 86-36 272 NSA and the American Public - The Issue ofLegality 272 Public Revelations and Cryptologic Secrecy 274 Classifying CryptologicInformation 275 Breaches in the Dike - The Security Cases 277 L' Affaire Weisband 277 The Petersen Case 279 Martin and Mitche11 280 Non - Responsive IlA:NBLI!l VIA: 'fA:LI!lN''f IlI!lYIlSLI!l eSMIN'f eSN''fftSL SYS'fI!lMSdSn,'fLY N'S'f MlbJ!JA:SA:BbJ!J 'fS FSRflf8N' NA:'ffSN':ALS TapSECRE'f~ VI T9PSE(RET~ Non - Responsive 1I".~TQbJil Vllt 'Y'ltbJilN'f I\:BYIISLB eSMII{'f e6N''fft6h ~ 1MJill'9I~ J61NTL i vii 'fap SECRET~ fepSEeRET~ Non - Responsive IIAHBb8 VIA 'Wtb8N'f' H:8YII6L8 eSMUi'f e6N'fft6b S'IS'fJ!)MS;J61N'fbY 14'6'1' MlLEASABtE '1'8 P8Mll6N Nll:'I'ISNll:LS IOP3EeRET~ viii Non - Responsive 19P5ECR!~ N9'Y' RFlbFJ1rSABM!J 'Y'9 F9RFHSN PfA'Y'I9PH'rbS ix iOP!l!eRET~ T6'5ECRET~ Non - Responsive iii A ;WQU; VI 'z TA15I!HfT KSYII8hS 88MIN'Y' 88N'Y'fteh SYS'fBMS d6mTL i Pf8T RShBlzSlzBhS '1'8 P8RSISPf Pflz'l'I8PflzhS x Foreword (U) The Center for Cryptologic History (CCH) and its predecessors have published thirty-seven volumes - monographs, crisis studies, source documents, bibliographies - concerning the history of signals intelligence and information systems security, the yin and yang of modern cryptology. These pUblications have treated specific events, organizational issues, and technical developments in peace and war; most have been pioneering efforts, based on original documentation, and, in many cases, are the first history of their particular topic in any venue. (U) There has been a strong need, however, for a single work to undertake the full sweep of cryptologic history, providing a context into which the more specialized studies may be placed. Such a cryptologic Cook's tour should incorporate the military-political events of our time and the history of interaction between cryptologic organizations and other components of the intelligence community - access to SIGINT and INFOSEC is limited to "insiders," but it is clear that cryptologic operations do not occur in a vacuum. (U) Thomas R. Johnson's American Cryptology during the Cold War, 1945- 1989 meets these requirements admirably. Drawing on over a decade of study and reflection on cryptologic history, Dr. Johnson deals with three facets of cryptologic history: first he explains how cryptology responded to the landmark events and challenges of the post-World War II era. He next provides profound analysis of how events and personalities affected the development of cryptology institutionally and professionally. Finally, and even better, Dr. Johnson spins a fascinating tale of the success or failure of cryptologic operations in the various crises that have challenged the SIGINT system. (U) With Books One and Two of this projected four-book work now available, American Cryptology during the Cold War is "must reading" for the cryptologic professional. The narrative and analysis in these first two books are essential background for understanding how the cryptologic community progressed to its present configuration. This is the definitive work on American cryptology after World War II. (U) For readers who may wish to explore American cryptology prior to the modern period, I recommend as a companion piece to the present book, Dr. Ralph E. Weber's Masked Dispatches: Cryptograms and Cryptology in IhUfBJ"B VIA 'fltJ"BN'f IEBYIIetJB eeMtN'f eer('fIteL 3 i 13YP;M13 JOINYL i NOT REI FA ca ARbi: 1'9 F9Fl:EI6f( nATIONALS xi TOPSEeRE~ American History, 1775-1900 (CCH, 1993). Two more useful books with background on pre-World War and World War II cryptology are Frederick D. Parker's Pearl Harbor Revisited: United States Navy Communications Intelligence, 1924-1941 (CCH, 1994) and Thomas L. Burns's The Origins of the National Security Agency, 1940-1952 (CCH, 1990). David A. Hatch Director, Center for Cryptologic History IlA~TQbE VIA 1:1AbE~r1:1KEYIlQb8 6QMIN'f 68H'fR8~SYS'fI3MSd8IN'f~Y H8'f RI3~I3*8*BLI3'f8 P8R13f6U UKff8N*LS T6P5ECRET~ xii TOP SECRET liM8RA Preface What It Is and What It Is Not This book is intended to be a general overview of U.S. government cryptology since the end of World War II. It is projected to be a four-book study carrying the story to the end of the Cold War, symbolized by the fall of the Berlin Wall. I have attempted to include the entire effort, which includes the Service Cryptologic Agencies (as they were once called), as well as certain CIA programs. These organizations comprised almost t-pe totality of the cryptologic efforts of the federal government, although other organizations (FBI is a good example) have occasionally dabbled in the discipline. Because it is comprehensive rather than strictly organizational, it contains information about the field sites, intermediate headquarters and the SCA headquarters themselves. It does not cover in detail the organizational aspects of the creation of the National Security Agency. That is covered in good detail in Thomas L. Burns's book, The Origins of the National Security Agency: 1940- 1952, published in 1990. Thus the coverage of events between 1945 and 1952 is sketchy and simply tries to fill in blanks in the record that the Burns book did not cover. This is not a history of private or nongovernmental cryptology. Although it covers relationships with our Second and Third Party partners, it does not focus on that aspect either, except as it contributed to the development of our own effort. Our long-standing debt to the British cryptologic effort at GCHQ should not go unnoticed, however. It deserves a separate book. If you are looking for a history of your specific organization, you will not find it. This is a history of events, not organizations. The importance of the cryptologic contribution to American security is so broad as to obscure individual organizations and, often, the specific people involved. In certain cases, however, I have identified major individual contributors to cryptologic history or those who were, by chance, thrown into momentous events. Two overarching themes characterized American cryptology from the end of World War II to the end of the first Nixon administration: centralization and expansion. The SIGINT system underwent a period of almost unbroken expansion from 1945 to the American retreat from Southeast Asia. These themes dominate the first two books in the set. The end of the Vietnam War and the era of the Watergate scandals that followed marked a watershed, and new themes of retrenchment and decentralization marked the period that followed. These will be the themes that open Book III. THOMAS R. JOHNSON HANDLE VIA TALENT KEYHOLE COMINT INTbY E TO FOREIGN NATIONALS xiii TOP SECRET liMBRA - - _ . - '(b) (3) -P.L. 86-36 TOP SECAr, tJMBR";:- (b) (1) _ (b) (3) OGA Acknow-Iedgements My debt to others 1)egins with. the staff Of the NSA Archives, who dropped whatever they were doing whenever I needed\material.l ~n the Archives helped with photographs, and the staff in L32 produced.hundreds of black and white prints to go into the publication....} / ;;;; FE31 (Gedl\'.;f.·.·.··.hiCS) did most ofthe map work. My debt also includes the CIA staffhistorians, especiall \ lan~ I who guided my work and opened doors to CIA material. My thanks also go to the editorial staffofBarry car1een~ I andl IWho, for days on end, did nothing but edit tis history. It was the longest work that the Center for Cryptologie History has attempted, and I am sure it taxed their patience, although they never said so., Also owing to the unusual length and complexity of the book, the NSA photo laboratory (E23) and NSA's printing services (Y19), which did the photo reproduction. and printing ofthis book, should b.e recognized for their major efforts to get out the publication.1 ~eserves praise for the cover graphics. In the Service Cryptologic Agencies, James Gilbert and Jack Finnegan, the INSCOM historians, were very responsive to my need for Army cryptologic materials. A special debt is also owed the/historical staff at the Air Intelligence Agency. Everyone on the staff, from James Pierson: (now retired) to Jo Ann Himes to Joyce Horns to Juan Jimenez, responded almost instantly to my many requests for information. Their help resulted in a rather more thorough treatment of Air Force cryptology than would have been possible otherwise. The history itselfhas had a large number of "readers" who plowed through the various drafts and revisions offering helpful comments and additional information. Everyone in the Center for Cryptologic History (CCH) had a hand in its improvement, as well as a list of other readers who critiqued various portions. Among them, David Gaddy ~ndl I ~eserve special note for their help with the chapter on Vietnam. ...l The history also had a group of "general readers," senior Agency officials who agreed to read the entire work in draft state. Milton Zaslow, Cecil Phillips, Donald Parsons, Eugene Becker, and David Boak spent long hours poring over various drafts, offering comments and encouragement and correcting information. Finally, I wish to thank all those who, over the years, volunteered their time to sit for oral history interviews. NSA owes them all a debt of gratitude for their contributions to retrieving otherwise vanished information. THOMAS R. JOHNSON HANDLE VIA TALENT KEYHOLE COMINT CON NOT RELE EIGN NATIONALS xv JOINTLY TOP SF'RH tJMBkA - - TOP SEeR'T l:'Jf4IBRA - - Footnotes The text is footnoted throughout with short, abbreviated citations. More complete information can be obtained in the Bibliography. However, a few comments on certain footnote abbreviations are in order. The largest number ofcitations is from the Cryptologic History Collection, which is the working file of the Center for Cryptologic History. This collection is organized into sixteen series, and citations to that collection begin with the series number and a series ofnumbers, e.g., CCH Series V.A.29. Citations from the NSA Archives vary depending on whether the document was part of an archived collection or was still in the Retired Records collection when researched. The former begins with the accession number, followed by a location, e.g., ACC 16824,CBTB 26. The latter begins with a box number, followed by a shelflocation, e.g., 28791-2,80-079. A general bibliography and an index are included at the end ofBook II. HANDLE VIA TALENT KEYHOLE CO OL SYSTEMS JOINTLY NOT E TO FOREIGN NATIONALS xvii TOP SeC:RET UMBRA - --------------~ - - - - - - TOp Si'AET tJMBRA Chapter 1 Cryptologic Triumph and Reorganization, 1941-1949 The combined U.S.-U.K. COMINT operation of World War II was perhaps the most successful large-scale intelligence operation in history. '-- ---JPIA,1971 WORLD WAR II AND THE INTELLIGENCE REVOLUTION The Second World War began a true "revolution" in intelligence. The impact of intelligence on the strategy and tactics of the Allies (and to a somewhat lesser extent on the Germans and Japanese) was truly revolutionary, and it is just now coming to be recognized for what it was. Through the publication of books like Frederick Winterbotham's Th€ Ultra Secret and John Masterman's The Double Cross System and by the massive declassification of war records begun by the British and Americans in 1977, the true extent ofthis influence is now emerging. No other intelligence source had the revolutionary impact of SIGINT. World War II was, in the words ofhistorian Walter Laqueur, "a SIGINT war." The influence ofSIGINT was so pervasive that it is now hard to imagine how we might have fought the war without it. Even prior to the direct engagement of American and British forces against the Germans and Japanese, two of their most complex ciphers were broken. The British effort at Bletchley Park first produced plaintext reports from the German ENIGMA system in September 1940, the same month that a small Army team under William F. Friedman broke the Japanese diplomatic cipher machine called PURPLE. By February of 1942 the Navy had broken the Japanese Fleet Operational Code, called JN25. In 1943 the Army broke the Water Transport Code, while in 1944 a lucky battlefield retrieval of cipher material allowed the Army to read the Japanese Army codes. When combined with successes in direction finding, traffic analysis, and the exploitation of plaintext communications, SIGINT yielded a torrent ofuseful information. British achievements have come in for the most scrutiny (and praise), We know that Churchill "revelled" in his ability to read Hitler's mail and spent hours pondering on Nazi strategy as revealed in the decrypted messages. The -British set up a very efficient and secure system for disseminating SIGINT, the precursor ofour SSO (Special Security Officer) system. Always wary of the "blabbermouth" Americans, they insisted that we adopt their system before they would share everything in the SIGINT larder with us. As the Combined Chiefs prepared for Overlord, they knew precisely how the Germans were reacting to the invasion plans and where they were positioning their units for the expected blow. HANDLE VIATALENT KEYHOLE COMINT MSJOINTLY N TO FOREIGN NATIONALS Ib) (1) Ib) (3) OGA 1 - - - ~ - - - - - - - - - - - - - Moreover, once the invasion was launched, they knew what the Germans were doing and were able to adjust accordingly. As Allied troops moved across France, they moved in sync with the gold mine of intelligence which detailed most of the important German military movements. Their intelligence officers must have looked like geniuses - they were able to predict German moves before they happened and could advise commanders how to react. If every dog has its day, this was the day of the G-2, the military intelligence officer. The product of breaking high-grade ciphers was called ULTRA, and it was so good that when it was not available, as it was not at the Battle of the Bulge, the G2 corps scarcely knew what to do. A few predicted the German offensive, but most did not. They were wedded to the SSO and the bonanza ofinformation that he could provide. The Pacific was the American theater, and the U.S. was as successful there as the British were in Europe. Navy cryptanalysts broke JN25 in time for Admiral Nimitz to use it in the Battle of Coral Sea in May of 1942. The success of strategic SIGINT was so important that Nimitz had become a permanent convert. When the cryptologists at Pearl Harbor came to Nimitz with information outlining a much bigger battle shaping up in the central Pacific, the admiral was quick to believe and quick to act. To his dying day he credited SIGINT with the key to the victory at Midway. This turned the war in the Pacific completely around and launched Nimitz on his Central Pacific campaign which took him to Okinawa. He considered SIGINT as an absolutely critical component, and he learned to use information from both the high-grade cipher traffic and the plaintext messages and operator chatter. Some ofhis subordinates were as successful as Nimitz in the use of this intelligence, some were not. But it is hard to argue with results. SIGINT and MacArthur had a turbulent marriage. The commander in the Southwest Pacific had outstanding success in using SIGINT on some occasions, the most conspicuous success coming in his 1944 New Guinea campaign. There were also some failures resulting from several causes. His staff never came to trust SIGINT as did that of Nimitz. When they did use it, it was sometimes hard to get it melded into the battle plan, as MacArthur was a classical intuitive decision maker. Jurisdictional disputes between MacArthur and the War Department in Washington caused him to come to distrust this strange SSO lash-up which he could not control because it did not work for him. In the battle for the sea lanes, SIGINT again played a decisive role. The Japanese merchant marine was devastated largely because its movements were being given away in the Water Transport Code. Sinking the defenseless and slow-moving merchant vessels was relatively easy when their movements were known beforehand. In the Atlantic, the U.S. and the British used decrypted ENIGMA messages to track German U-boats and to drive their wolf packs from the sea lanes. This was not quite as easy as going after merchantmen, and the marriage between SIGINT information and operational procedures to effect a kill represented a very high level ofmilitary and technological expertise. It may have been the most difficult and delicate use ofSIGINT during the war. II Hl"Qb:S VIA 'I'J.]":S~l'l' K:SYII8hE eSMUff eSlffft6t B'l S'l'JiJMB46U'f'fL I TQPSECRET~ 2 Tap SECRET~RA One other wartime accomplishment would become significant in later years. In 1944 the British and Americans established a Target Intelligence Committee (TICOM) to interrogate captured German COMINT personnel. The major objective was COMSEC - to determine how well the German cryptologists had exploited Allied communications. The flip side of that effort was COMINT - to see how well the Germans were doing against other, and particularly Soviet, communications. TICOM was at Bletchley Park, headquarters for the British cryptologic service, Government Code and Cipher School (GC&CS). Six teams of American and British COMINTers were dispatched to the battlefields of the Continent. They sent their "take" to the Document Center at GC&CS. The original documents remained there while the microfilm copies were sent on to Washington. TICOM teams also captured equipment. One-of-a-kind equipment remained at GC&CS, while duplicates were sent to the United States. The new system was so successful that teams were established in the Pacific, with the British taking the lead in Southeast Asia, the United States in the Central Pacific and Japan, and joint American and Australian teams in Rabaul and Borneo. Although TICOM was formally dissolved in November of 1945, American and British experts continued to exploit the material for years afterward, and TICOM was later re-created in the United States as TAREX (Target Exploitation), minus British participation. If the strength of American SIGINT was in providing militarily useful information, its weakness was in its organization. The Army and Navy were at constant loggerheads over the control of cryptology, and at times the factional disputes were little short of catastrophic. British historian Ronald Lewin, a great admirer of American technical ingenuity which yielded the SIGINT bonanza, was frankly contemptuous of our inability to get along: The old antagonism and suspicion between Army and Navy persisted in a manner that may at times seem infantile, until it be remembered that tribal loyalty, narrowness of vision. and sheer egocentricity can make even the most senior and hardened officers occasionally enter a second childhood.l Army and Navy cryptologic organizations had a long and inglorious history of failing to coordinate their efforts, dating back to the 1920s. In 1940, when the Army's success in breaking Japanese diplomatic cipher systems became known to the Navy, there ensued lengthy and difficult negotiations to determine how the effort was to be divided. They finally arrived at a Solomonic solution by which the Army processed Japanese diplomatic traffic originating (i.e., cipher date) on even days of the month while the Navy would process traffic from odd days. This resulted in a fair division politically, but from the standpoint ofcryptanalytic continuity it was a horror. To make matters even worse, there was in those days no thought, no concept, of centralized and coordinated intelligence analysis. What little analysis and interpretation was done (and there was very little indeed) was accomplished by each service on the traffic which it had decrypted, leaving for each a checkerboard pattern of information in which every other day was left out. This IIltNBbS VIA ¥zItbSfff I(fi]YII6Lfi] e6MHof¥ e6H'fft6L Si S¥fi]MSJ6IIof'fL i NOT RIP Ii: A ~ A IH..Ji: TQ ¥QRJi:lQl'T l'TATIQl'TlzbS 3 almost inconceivable situation persisted until 1942, when diplomatic traffic was, by mutual agreement, left to the Army, while the Navy concentrated on Japanese naval materia1.2 The disaster at Pearl Harbor resulted in a thoroughgoing Army internal investigation. Secretary of War Henry Stimson picked Yale lawyer Alfred McCormack to lead the way. McCormack discovered a scandalously incompetent Army G2 and a nonexistent SIGINT analysis and dissemination system. He set up a separate system called Special Branch, Military Intelligence Division, and was picked as the first deputy. (Colonel Carter W. Clarke became the first commander.) At the same time, the Army and Navy arrived at a joint modus operandi regarding the division ofoverall SIGINT responsibilities. Each service was to work what we now call "counterpart" targets. Since there was little in the way of Japanese Army traffic to work, the Army took on the task of diplomatic intercept. The third partner was the FBI, which shared Alfred McCormack with the Navy the task of working Western Hemisphere agent and clandestine traffic. These three were tobe the only participants in SIGINT for the duration of the war. Roosevelt's directive of July 1942 specifically excluded the FCC (Federal Communications Commission), Office of Censorship, and the OSS (Office ofStrategic Services) from SIGINT production .3 At the same time a standing committee of Army, Navy, and FBI COMINT officials was established. It met only a few times and had little lasting impact on organizational matters. Meetings were frequently marred by vituperative arguments, especially between Navy and FBI, which were supposed to be sharing Western Hemisphere clandestine traffic. It was not cryptology's finest hour. Meanwhile, the COMINT activities of the FCC and Censorship Bureau continued virtually unabated.4 Only the OSS seems to have been temporarily frozen out ofthe COMINT community. Resurrected after the war as the CIA, it IIA~iBbE Yb\ 'f1'[bEn~1E:8YII8bE e8MHff eeU'fftSL S"lB'f'BMBJEltSABL8 'fa :f'6ftEItH( NAIIONALS Tap SECRET u~ 8 - - - - - - - _ . - - - _ . _ . - 1'011 S!CRET ~ Government offices on the Mall Both SIS and OP-20·G began World War n in these temporary buildings on the Mall in Washington. Arlington Hall Station in the 1940s IIMlQUJ YIlt 'I'AbSn'l' It;SYllObS CaMIN'I' CenTRaL BYSTFJMS j6II4TL i N6T RFJLFJkBkBLFJ 'f6 f6ftFJI6?4 ?4k'fI6?41\LS 9 ,.J:OP SEeRE I ~A ; Ib) (1) Ib) (3) -50 USC 403 Ib) (3) -P.L. 86-36 T9PSECRET~ To Army cryptology, as to the Navy, peace was devastating/Most of the work force at Arlington Hall left civilian government service, and within/days /the halls were almost empty. Intercept sites overseas were suddenly confronted/with no Japanese or German intercept mission. One former soldier described the experience/as being left stranded on Okinawa with no Japanese mission to copy and no instructionson a/follow-on assignment. His unit eventually moved to Seoul, relocated to a/former /Japanese communications station, and there got a new mission- Soviet and.Chinese Communist communications. European units tackled French and Greek missions, and/by mid-1946 nearly half the Army's end product was based on the intercept ofFrench communications.12 The late 1940s were a period of damaging retrep.chment. The Army and Navy cryptologic organizations that began the Soviet mission had little experience, less money, and no expertise. Yet ASA was able to survive better/than OP-20-G. The Army had relied historically on civilians, and many of the best, including William Friedman, Frank Rowlett, Abraham Sinkov, and Solomon Kullback, stayed/on. Missing the excitement of wartime cryptology, others drifted back to Arlington Hall after brief, humdrum civilian careers. The Navy, which had r.elied on uniformed cryptologists, lost a far higher number to civilian life and found the transition to peacetime a difficult one. In 1947 ASA and OP-20-G were joined by yet a third cryptologic service, that of the newly created Air Force./The Army Air Corps had actually established its SIGINT service in the Pacific in 1944. /The Air Force acquired an early reputation for parochialism and interservice rivalry./The feuding led Carter Clarke, then head of Special Branch of Military Intelligence Service, to write in June 1944 that "the Air Force insists that these I loperate only for the Air Force and insists further that no personnel can be attached or detached therefrom; neither should the theaters give them any operational directives in the sense that we think ofit." The first Air Force unit in the Pacific was the _____________.....Ilwhich beganoperations in 1944 in New Guinea.13 When the independent Air Force was created in 1947, there was no direct reference to cryptologic activities, and for a time ASA continued to provide these to the nascent Air Force. Yet the Air Force was determined to establish its own capability. Certain Air Force generals were aware of the contributions of COMINT during the war. One in particular, Hoyt S. Ve.'1denberg, who was later to become Air Force chiefof staff, was convinced that the Air Force had to have its own cryptologic service. He saw how the British controlled cryptology in Europe and felt that it was essential to get this under American, and particularly Air Force, control.14 In early 1948 the Air Force fashi(med a transition agreement with ASA. The latter established an Air Force Security Group within its headquarters at Arlington Hall to oversee the transfer. Three c::::::::Jand eight COMSEC units were turned over to the Air Force. The Air Force role was defined as mobile and tactical, and ASA continued to operate all fixed sites. A set number ofASA officers (thirty-two) became blue-suiters, and this group became the "founding fathers" ofAir Force cryptology. Air Force cryptologists I1AUBLEl VIA 'fALElU'f IU..1YII6LEl e6fdHtf'f e6Itf'fft6L S-YS'fElMSd'6Htf'fLY 'lOP SECRET u~ 10 Te"S!CREI~ were to continue to train at ASA schools and were to contribute instructors and financial support as soon as the Air Force had a budget of its own. Significantly, the Air Force assumed all responsibility for "the investigation for intelligence purposes of all types of electronic emissions relating to radar, radio control of guided missiles and pilotless aircraft, proximity fuses, electronic navigation systems, infrared equipment and related subjects." In other words, the Air Force was to take the ELINT and electronic warfare missions, which were at the time too new to even have a name. Needing equipment but not yet having a budget, the Air Force arranged for the transfer of equipment from the Army, which turned out to be cast-offreceivers and antennas that ASA no longer wanted.15 On 20 October 1948, the new Air Force cryptologic organization was officially established as the U.S. Air Force Security Service (USAFSS), still located at Arlington Hall. It was a major air command, responsible to neither intelligence nor communications. Thus from its earliest existence the Air Force accorded a loftier organizational position to its cryptologic service than did the other, more senior, services. And the Air Force did something else that was unprecedented. In May of 1949 it moved completely out of Washington. Security Service set up shop at Brooks Air Force Base outside of San Antonio, Texas. The move was calculated to remove USAFSS from geographical proximity to the central control authority for COMINT - at the time the Coordinator for Joint Operations, shortly to become the Armed Forces Security Agency. Thus USAFSS hoped to be insulated from any sort of outside control, which it regarded as bald interference in its affairs.16 THECJO The lack of central control for COMINT was the most pressing problem of the postwar years. Cooler heads recognized that the uncoordinated and fractionalized efforts that had existed since the 1920s simply had to be better controlled. They had already agreed on a committee system, at that time called STANCIB and STANCICC. The committees could and did arrive at policy decisions which, in the case of unanimity of the board, were binding on the services. What was still lacking, though, was an executive organization to carry out the routine business ofcentral coordination. In early 1946 the Navy proposed such an executive body. They called it the Coordinator for Joint Operations, and it was to work out routine intercept coverage and processing responsibilities between the services. The Navy got Army concurrence, and on 15 February STANCIB approved the proposal. The Coordinator for Joint Operations, or CJO, was born.17 The CJO was to implement general policies on allocation ofjoint tasks as approved by STANCIB. It was to be assisted by three groups: the Joint Intercept Control Group (JICG), the Joint Processing Allocation Group (JPAG) and the Joint Liaison Group (JLG). nA~TQbK VIA 'I'A15KN'I' ICKYII9bB 09111Hf'f eeri'fft8b B-YS'f8MB ifem"fbY 11 .+OP SECRET urRA .IOP§EERE~ The CJO agreement owed its existence to the two most influential sponsors, Joseph Wenger (who commanded OP-20-G) and Preston Corderman (chief, ASA) for the Army, and it was in those days referred to as the "Corderman-Wenger Agreement." But when the first CJO was appointed, it turned out to be Colonel Harold G. Hayes, a long-time Army COMINTer and the new chiefofASA. The first task of the CJO was to allocate intercept tasks. This was not as easy as it appeared. Agreement was reached that counterpart targets were to be copied by the respective U.S. service cryptologic organization. All other targets, even those being intercepted entirely by a single service, were to be considered '1oint." The CJO then reallocated the intercept responsibilities. This had the largest potential impact on the resources of the Navy, which during World War II, as previously discussed, completely gave up '1oint" targets (with a few exceptions) to the Army. Intercept allocations really got down to priorities. With limited resources (and in 1946 resources were constrained), the key to obtaining copy was in the priority system. In September of that year USCICC decided to hold monthly meetings to consider priority problems. By this process a standing priority list, in rather general terms, was established. The CJO then made intercept assignments to positions in the field. When the CJO assigned ajoint case to a position it controlled (Le., one which had been turned over by one of the Service Cryptologic Agencies, as they were then called) there was no problem. But occasionally the CJO assigned a joint target to a service-protected position. This invariably met with resistance, and the CJO had no enforcement authority. The Service Cryptologic Agencies (SCAs), for their part, insured that counterpart positions were manned with the best operators, that they were never left uncovered, and that technical data were always up to date. In short, if a target had to be slighted, it was likely to be the joint target. The servicemen never forgot whom they worked for. CJO also allocated processing tasks through the JPAG. Since people and equipment for processing were in very short supply, processing on each major target was to be done in only one place - either Arlington Hall or Nebraska Avenue - no matter which service collected the traffic. In those days communications systems were mutually exclusive rather than common and interlocking, and once traffic was intercepted by one service, it had to pass vertically through those communications channels all the way to Washington. This meant that there had to be communications between Nebraska Avenue and Arlington Hall so that the traffic could be exchanged, and under CJO a teleprinter link was set up. The services had a great deal of difficulty talking to each other (electrically, not to mention in person), and it was a real effort to establish common cryptographic gear for interoperability. In the late 1940s this process was just getting started. Communications security policy was, if possible, even more difficult to meld into a cohesive system than was COMINT. Through the war each service handled its own COMSEC matters with little reference to joint policy. In the Army, ASA was responsible for both . COMINT and COMSEC, a development substantially influenced by such technicians as Frank I1AP1BMlYllr 'flthIUi'l' IEFlYH8hB C8f¥HNT caNTRaL 8YMrJM8 JatH'ft"i' 12 :rep 'I!!!CRE I U~RA Rowlett and William Friedman. In the Navy, COMSEC had begun within Captain Laurance Safford's embrace, but it had eventually become part of a separate organization under Naval Communications, called OP-20-K. After the war, COMSEC policy was allocated by an unregistered executive order to a Cryptographic Security Board consisting of the secretaries of state, war, and navy. This very high-level board quickly became moribund, and the real actor in COMSEC policy was , ) the Joint Communications-Electronics Committee (JCEC) and its subordinate, the Joint Security and Cryptographic Panel. When COMINT was unified in 1949 under the Armed Forces Security Agency (AFSA), COMSEC was still decentralized. The CJO was a compromise between those who wanted tight central control and those who wanted to continue a loose arrangement. It was voluntary, as had been ~ll of its predecessors. It never resolved the conflict over joint targets, much to the dismay of the State Department, which was the principal customer for most of those targets. But the establishment of an executive organization was the first step in creating an organization to control COMINT. It didn't work, but it pointed the way toward the future. THE CRYPTOLOGIC ALLIES America's SIGINT relationship with Great Britain also dates to World War II. In July 1940, the British ambassador to Washington, Lord Lothian, proposed that the two nations exchange information on, among other things, technological secrets related to "submarine detection and radio traffic." This appears to have pertained generally to SIGINT, but the wording of the now famous Lothian Letter did not really say precisely what he (or Churchill) meant. It also appears that day-to-day intelligence cooperation predated the Lothian Letter, for in April of the same year President Roosevelt met Churchill's special envoy William Stephenson to discuss a plan for secret cooperation between the FBI and British secret intelligence. According to a fascinating account in the somewhat unreliable book by William Stevenson (unrelated to the wartime William Stephenson), it was at that meeting that Stephenson informed Roosevelt of British progress in breaking the German ENIGMA system. (This might have happened but was quite out of character for the security-conscious British.) This meeting did, in fact, lead to the establishment of the British Security Coordination (BSC) in Washington, with Stephenson in charge. During its early days this organization dealt primarily in HUMINT and counterintelligence.is The Lothian Letter was followed in August by a visit by Sir Henry Tizard, scientific advisor to the Royal Air Force (RAF). This inaugurated a series oftechnical discussions on a wide variety of subjects. Tizard, not a SIGINTer, was mainly interested in discussing radar and other such technical developments. At the same time, the United States sent to Britain a delegation consisting of Brigadier General George V. Strong (Chief of War Plans), Brigadier General Delos Emmons (United States Army Air Forces II1tNBbB VI1t 'f1tbBN'f IEBYII8bB 08ftnN'f 08N'fR8b S'IS'fBMS d8m'fb'I N8'f RBbBASABbB 'f8 fl8RBI8N NA'fI8NAbS 13 USAAF), and Rear Admiral Robert Ghormley (Assistant Chief of Naval Operations). Though the discussions were to be general, it appears that Strong had, or thought he had, considerable latitude to discuss cryptologic intelligence. On 5 September he cabled Washington to propose a total exchange of information on SIGINT product and technical matters (i.e., cryptanalysis). Back in Washington there was a good bit of concern. The Navy said "No," while the Army vacillated. Their top cryptanalyst, William F. Friedman, was consulted. Friedman favored the exchange. So initial hesitance was eventually converted to approval, and on the day after Christmas 1940, the Army decided once and for all to initiate a complete cryptologic exchange with the British. In February 1941, Captain Abraham Sinkov and Lieutenant Leo Rosen of the Army's SIGINT organization, along with Lieutenant Robert Weeks and Ensign Prescott Currier of the Navy, sailed to London. They brought with them a PURPLE Analog, a machine the Army was using to break the keys for the Japanese diplomatic cipher system. They had instructions to initiate a complete exchange ofcryptanalytic and SIGINT information.19 The British appear to have been flabbergasted. Never had they anticipated that the United States would simply walk in and plunk down their most secret cryptanalytic machine. This was, indeed, an intelligence exchange worth the money. But they were cautious. They did not tell the Army and Navy emissaries everything they were doing, and they did not show them the ENIGMA operation at first. Agreed upon in principal in 1940, the complete exchange of cryptologic information and techniques progressed slowly through the war. Once again the Navy, reluctant in the beginning, produced the more beneficial exchange. This was due largely to historical circumstances. The Army was still mobilizing and clearly would not see action in Europe until at least late 1942, if not later. But the Navy was already engaging German U-boats in the North Atlantic. They and the British had worked out a convoy system, and daily cooperation in intelligence was essential to avoiding wolf packs. Thus it was that Commander Roger Winn, who headed the Operational Intelligence Center in the Admiralty, convinced the U.S. Navy that it must have something similar. Prompted by Winn, the U.S. Navy established the mysterious organization called F-21 (Atlantic Section, Combat Intelligence Division, U.S. Fleet) and its still more mysterious submarine tracking room. The latter used all sources of intelligence, including U-boat positions obtained by ENIGMA decrypts, passed to them by the British. The arrangement worked well at first, but in February 1942 the Germans introduced the four-rotor ENIGMA, and the British at Bletchley were unable to read it. The Americans were already suspicious because the British kept the cryptanalytic techniques so closely held. So in 1942 the Navy embarked on a project to break the ENIGMA themselves, in defiance of British protests. Colonel John Tiltman, a temporary GC&CS resident in Washington, finally convinced the British that the Navy would proceed with or without British help. In June 1942, after Tiltman's intervention, the Navy sent two expert WI )TQU; VIA :~"'15S~T1' KSYU915S EJ9MHf1' EJ9N'FR915 SYS'FrJMS iJ6fN't'LY )TOT RiJ..i • ~h\8U] 1'9 F8RFH8U Nk'fI6N'ALS ....IOP§EeRE~ 14 :rep Sl!eREru~ cryptanalysts, Lieutenant R. B. Ely and Lieutenant Junior Grade Joseph Eachus, to Bletchley to learn all they could about ENIGMA processing. In September the Navy began a project to build a four-rotor ENIGMA processor (called a "bombe" by the British). When, in the summer of 1943, the Navy moved to its new headquarters on Nebraska Avenue, a major portion of the space was reserved for the bombes, which were being employed to break the keys on German submarine ENIGMA traffic. In the end, the two nations drove the U-boats from the North Atlantic, based in part on information provided by the bombe project. Meanwhile, the Army was having its own problems on the SIGINT front. Increasingly suspicious of British reluctance to share cryptanalytic techniques, they retaliated by refusing to share informationon voice ciphony equipment with Alan Turing. Since Turing was one of the top Bletchley scientists (and has been given credit for developing the first British bombe), this was a very serious breakdown in cooperation. It became the subject of a long series ofexchanges between General George Marshall and Sir John Dill (chairman ofthe British Joint Chiefs of Staff), and at one point it seemed possible that the two sides might break COMINT relations. The dispute was resolved in 1943 when the British agreed to allow a total technical exchange. The agreement was hammered out during a series of sessions between Military Intelligence Service and Commander Sir Edward Travis, who headed GC&CS, during Travis's trip to Washington in May. The paper specified that the United States would be responsible for the COMINT problem in the Far East, while the British would worry about Europe. To implement this, it was agreed that the Americans would send a team of cryptologists to Bletchley to work side by side with the British in all aspects ofCOMINT, including cryptanalysis of the ENIGMA. That way the Americans would gain technical expertise on the system without mounting a competing cryptanalytic effort on the American side ofthe Atlantic. To begin the new relationship, the Army sent a three-man team consisting of Colonel Alfred McCormack, William Friedman, and Lieutenant Colonel Telford Taylor to Bletchley. By mutual agreement, Taylor was left behind in London to serve as a liaison officer and to act as a funnel for British COMINT being sent to the War Department in Washington. Taylor's job was not easy, as there was a good deal of second-guessing the British forthrightness in the exchange. But as the war progressed it became smoother and eventually became a very open exchange ofhighly sensitive information. With the Axis almost defeated, the thoughts of cryptologists in 1945 turned with increasing frequency to the Soviet Union. Both nations had maintained rudimentary efforts against the "Communist menace" since the 1920s, and they both kept small efforts even during the war. In June of 1945 ANCIB proposed to the British that they extend their wartime cooperation to the intercept and exploitation of their erstwhile but distrusted ally. They called the project BOURBON, and it was kept compartmented for the obvious reason that the Soviets were still officially on our side. The arrangement was largely informal and involved the exchange ofliaison units on both sides ofthe Atlantic. nA~fBY!l VIA 'fll:hB~f'fKBYII8hB 08MUf'f 08U'fR8b 8¥S'fi!JMS d6IH'i'bY 1'1O?' PIi:J,,1i: A i A RYJ TQ FQREll8n n1r'i'16Ib\L~ 15 Ib) (1) Ib) (3) -50 USC 403 Ib) (3) -18 USC 798 Ib) (3) -P.L. 86-36 Tap.seeRI!T u~ But in September, with the war officially over, the U.S. had a legal problem. Could it now continue to collaborate with its British allies?\.Clearly, the American cryptologists, good as they had become, still regarded GC&CS with acertain awe. In many cryptanalytic areas the British were still ahead of us, and their organization of the COMINT system was superb. And of course there was the problem of the Soviet Union. Already the wartime alliance had disintegrated. In September of 1945 both the Army and Navy suggested to President Truman that collaboration with the British continue for the present "in view of the disturbed conditions ofthe world and the necessity ofke.eping informed of the technical developments and possible hostile intentions of foreign nations...." In reply, Truman signed a brief, single-sentence note sent to him by the Joint Chiefs ofStaff: The Secretary ofWar and the Secretary ofthe Navy are hereby author.iz.ed to direct the Chief of Staff, U.S. Army, Commander-in-Chief, U.S. Fleet, and Chief of NavalOperations to continue collaboration in the field of communication intelligence between the United States Army and Navy and the British, and to extend, modify, or discontinue this collaboration, as determined to be in the best interests ofthe United States.20 Now that the American side was officially unleashed to collaborate with the British, it seemed necessary to write a bilateral agreement for the postwar year$.\ After months of meetings and conferences, the two sides sat down in March 1946 to sign the British-U.S., or BRUSA, Agreement. The paper which charted the future course of both countries was only four pages long. (The policy conference at which it was signed wa.sfollowed by a technical conference which wrote all the fine print appearing later as annexes and appendices.) I With the signing of the BRUSA Agreement, the BOURBON liaison offices on both sides of the Atlantic became representatives of STANCIB and LSIB, T e BOURBON 0 lcer, '":C;::;-"om-m-an-d'e-r--;::Gi""r-a-n":"'t'M·a-n-s-o-n-,-w-a-s-I...·n-v-e-s":"'te-'--w"'lt'l"r"'l"'It~e--ra-t'l"r"e-r-c-u-m""T"-er...Jsome title of U.S. Liaison Officer, London SIGINT Centre (LSIC, as GC&CS was then known) - or USLO LSIC. He reported to STANCIB through the deputy coordinator for Liaison, part of the new CJO structure. In early 1946 the British moved LSIC from its wartime location at Bletchley to Eastcote, outside London, and began using a new title, Government Communications Headquarters, or GCHQ. Space for Manson was provided at Eastcote. The BOURBON I11d~BLI!i VI1\: 'f1\Lrln'f ItrlYIISb8 eSMIN'f eSn'fftSb S-YS'f8MS dSIN'fbY ~TS'i RKbK &.SARbK 'is li'SRd!JIQN ~Tlt'fI8Nlth8 TOP SECRET ~A 16 liaison office had maintained an office in London, and Manson had to cover two locations, in Eastcote and London. (This situation continues to this day, with NSA holding offices in both London and Cheltenham.) USLO never controlled the TICOM group, which also found quarters at Eastcote.21 The British, meanwhile, had a more difficult problem. While the U.S. dealt with only one COMINT organization, GCHQ, the British had two - the Army at Arlington Hall Station and the Navy at Nebraska Avenue. Not wishing to choose, the British diplomatically located their liaison officer in the State Department building in downtown Washington. (They did, however, maintain a technical staffat Arlington Hall.) Their first liaison officer was Colonel Patrick Marr-Johnson, who had signed the BRUSA Agreement for the British side. When he retired in 1949, he was succeeded by Tiltman, who was already well known to the Americans and had served for a time as Travis's deputy at GC&CS. This began a practice, continued to this day, of assigning very senior cryptologic officials to the respective liaison offices, and the USLO eventually became SUSLO - Senior U.S. Liaison Officer.22 And where were the British Dominions in all this? They were mentioned in the BRUSA Agreement, and it was agreed that they would not be termed Third Parties, but they were not direct and immediate partners in 1946. Arrangements that Great Britain might make with them would be communicated to STANCIB. STANCIB, in turn, would make no arrangement with a Dominion without coordination with LSIB. Thus the nowfamous UKUSA Agreement was not that at all~ at least to begin with. It was a BRUSA Agreement. How it became the UKUSA Agreement was a development that spanned another eight years. Of the three dominions with which the Americans eventually associated, the relationship with Canada began first. Canadian-American SIGINT cooperation appears to have begun in 1940, in the form of service-to-service collaboration between the respective armies and navies. These decentralized arrangements were eventually overtaken by a centralized relationship centering on the Examination Unit of the National Research Council, established in 1941 as one of those clever cover terms denoting a Canadian SIGINT organization. Its purpose was to decode traffic to and from the Vichy delegation in Ottawa.I This unit's control was gradually broadened until it was the dominant force in Canadian cryptology. (It was the linear predecessor of the postwar organization Communications Branch, National Research Council [CBNRC] and its successor, Communications Security Establishment [CSE].) By 1943 it had its own submarine tracking room and was receiving plots from the British based on ENIGMA decrypts. When the British began cooperating with the U.S. in 1941, they requested that the U.S. bring the Examination Unit into the scope ofthe cooperation. But the Americans were leery. They knew that the Examination Unit had been established by Herbert O. Yardley, the renegade American cryptologist who had published cryptologic secrets in 1931 in The American Black Chamber. The Signal Intelligence Service, which had been victimized by Yardley's revelations, informed the IIANBLB VIA "ALElN" IlElYII8LEl e8MIN" e8fffft8L SYS'fF.lMStf8In'f'bY 17 Ib) (1) Ib) (1) Ib) (3) -50 USC 403 Ib) (3) -P.L. 86-36 Ib) (3) -18 USC 798 TO' ~E!CRET ~RA British that they were willing to cooperate only if Yardley were let go. The British, holding no brieffor Yardley, had the Canadians get rid of him, and collaboration with the Americans flowered. By April of 1942 details of the Canadian-American cooperation were hammered out. Collaboration was particularly close in direction finding (DF) ofGerman naval vessels. IBut the United states was suspicious; Canada had just been through a major spy....._-_..... scandal, the Gouzenko affair (chapter 4), and USCIB wanted to go slow. Making matters worse was the head of the Canadian olicy committee on COMINT, a rather prickly characte efused for several years to adopt some of the security proce ures w IC e a es and Great Britain had agreed upon at the J3ltlJSA<::()llfereIlce.l\for~over,while the United States wanted a formal document on ~::':un:,~:ri~lt.grJt~r:iti:::~::e:~V:~dl:~l~';~::~:~:~~:Major General C. P. Cabell. Thu on the battle of t e egal documentation while the United States got its way on securl y procedures.23 Furthest from the mainstream were the Australians. British-Australian COMINT collaboration appears to have begun in the late 1930s when a small Australian cryptographic organization under the Director of Naval Intelligence began working with the British Far Eastern Combined Bureau (FECB) in Singapore. In early 1940 an Australian naval commander named T.E. Nave set up the nucleus of an Australian SIGINT group in Melbourne, which was the origin of the modern Australian SIGINT organization. Its most important organization was the Central Bureau, set up in April 1942 as a combined Australian-American COMINT group. When the Americans departed in 1945, the Australian remnant ofCentral Bureau became Defence Signals Bureau (DSB). The British were determined that DSB should enjoy the same status on BOURBON as the Canadian, and, immediately after the war, began including the Australians in their technical exchanges. But in 1947 this procedure became embroiled in a lengthy dispute over Australian security practices. The procedures in dispute were arcane, and the origins were almost as difficult to fathom, but both apparently originated with a spy scandal. In 1947 SIS succeeded in decrypting some KGB messages which had been sent more than a year earlier and which contained certain classified British military estimates. The messages came from the Soviet embassy in Canberra, and it was immediately assumed that an Australian was passing classified information. The British, alerted by the Americans, sent Sir Percy Sillitoe, chief of British Secret Service, to Australia to discuss this with the prime minister. Sir Percy was under instructions to conceal the origins ofthe information, and when the prime minister, a Laborite named Chilley, demanded proof, Sillitoe mumbled something rather lame about a possible mole. After considerable UAUBLtFfftf TZ\LtlUT fitly-uebEl eellflUT eeUTReL B-YSTElllIS r1eIUTLY ueT Rtlbtl1.SABbEl Te FeRBf6If U:ATI6IUtbS 18 Ib) (1) Ib) (3) -50 USC 403 Ib) (3) -18 USC 798 Ib) (3) -P.L. 86-36 discussion, Chilley agreed to establish a new Australian security organization, called the Australian Security Intelligence Organization. With the Australian security house supposedly in order, the British prime minister, Clement Attlee, intervened with President Truman to get a new hearing of the Australian matter. Attlee complained in a letter to Trumari that: The intermingling of American and British knowledge in all these fields is so great that to be certain ofdenying American classified information to the Australians, we should have to deny them the greater part ofour own reports. We should thus be placed in a disagreeable dilemma of having to choose between cutting off relations with the United States in defence questions or cutting offrelations with Australia.24 With matters at the crisis level, Attlee proposed to Truman that Sir Francis Shedden, the powerful and respected Australian defense minister, visit the United States to plead the case. Truman accepted, and Shedden visited Washington in April. But he was unable to sway USCIB, and the British were back to their dilemma - whether to choose the United States or the Commonwealth as allies. In 1949 the outcome was anything but certain. Then one ofthose unexpected quirks offate intervened which was to save the day: the Labor government under Chilley went down to defeat at the polls,\and Robert Menzies formed a new Liberal-Country Party coalition in December. The conservative Menzies was able to successfully disassociate his government from the leftist.elements of the Labor government. This was critical since the actual source of the leaks was known (through the VENONA project; see chapter 4) to be two leftists within the Australian diplomatic corps. With a Conservative government in power, USCIB authorized a limited resumption of cryptologic exchange with Australia. Full resumption ofties did not occur until 1953. The incident tarnished American-Australian intelligence cooperation for years and caused a serious rift with Britain which was made worse just a few years later with the Klaus Fuchs case and the Burgess and McClean defections. It also had a deleterious affec\ on early U.S. SIGINT efforts against the People's Republic ofChina (PRC).25 By 1953 relations had warmed to the point where Australia was reincorporated as a full COMINT partner. The foundations of the Australian participation in the. UKUSA Agreement (the name BRUSA was changed at British request a year later) came at the Melbourne Tripartite Conference of September 1953. New Zealand came in as a rIfth partner, New Zealand had contributed mainly DF to the Allied cryptologic effort in World War II and had sent people to Australia to serve with the Commonwealth effort in Brisbane., \ \ Ih\UBbS VIii 'I\\bSU'f I(SYIISbS OSMIU'f OSU'fRSbS-YS'fSMBdSm'fbY 19 ~MBRA JOp &EEftE ro~ Ib) (3) -P.L. 86-36 Notes Ib) (1) Ib) (3) -18 USC 798 Ib) (3) -50 USC 403 Ib) (3) -P.L. 86-36 1. Ronald Lewin, The American Magic: Codes, Ciphers, and the Defeat ofJapan (New York: Farrar, Straus and Giroux, 1982),24. 2. Robert L. Benson, "A History of U.S. Communications Intelligence during World War II: Policy and Administration," manuscript pendingCCH publication. Hereafter Beneum "History." 3. Ibid. 4. Ibid. 5. Ibid. 6. George F. Howe, "The Narrative History of AFSAJNSA," part I, unpublished manuscript available in CCH. Hereafter Howe "Narrative." 7. Ibid. 8. USNSG, "U.S. Naval Communication Supplementary Activities in the Korean Conflict, June 1950 - August 1953," in CCH Series V.M.3.1.; Benson"History"; SRH 149, Records of the National Security Agency, Record Group 457, National Archives, Washington, D.C.; oral history interview with RADM Earl E. Stone, 9 Feb 1983, Carmel, California by Robert D. Farley, NSA OH 3-83. 9. NSA retired records, CACL 60, TVC 1317; [Edward S. Wiley] On Watch: Profiles/rom the National Security Agency's Past 40 Years (Ft. Meade: NSAJCSS, 1986), 13. 10. CCH Series VI.1.1.1.; X.H.7.5. 11. Oral history interview with Dr. Abraham Sinkov, May 1979, by Arthur J. Zoebeleinl ...lIDale Marston, and Samuel Snyder, NSA OH 2-79; Howe, "Narrative." 12. Oral history interview withCoI. (USAF Ret.) John P. Shean, 18 April 1984, by Robert D. Farley, NSA OH 16- 84; memo to Chief, AFSA-90,14 Dec 1948, in CCH Series V.C.2.12. 13. NSAlCSS Archives, ACC 26350, CBSK 32. 14. Philip S. Meilinger, Hoyt S. Vandenberg: The Life ofa General (Bloomington, Ind.: University ofIndiana Press,1989). 15. NSAJCSS Archives, ACC 26350, CBSK 32; oral history interview with Gordon W. Sommers, Hqs ESC, January 1990, by Millard R. Ellerson and James E. Pierson; Richard R. Ferry, "A Special Historical Study ofthe Organizational Development of United States Air Force Security Service from 1948-1963," Hq USAFSS, 1963. 16. "An Oral History Interview: The Electronic Security Command - Its Roots; Featuring the Founder of USAFSSlESC, Lt. Gen. Richard P. Klocko (USAF, Ret)," Hqs ESC, 20 October 1989. 17. The history ofthe CJO is covered in detail in Howe, "Narrative," and in Thomas L. Burns, The Origins ofthe . National Security Agency, 1940-1952, United States Cryptologic History, Series V, Vol. I (Ft. Meade: eCH, 1990). Hereafter Burns, Origins. IIAm3""S VIA 'fkbHlN'f IEHlYlIabHl CaMIN'!' CaNTRal::; SYSTf!lMS tf6fI"TI::;Y NOTRBf,EASABT F 1'0 i'ORl!lI8N NA:TI6IofA:L5 TOPSEEftE~ 20 Ib) (3) -P.L. 86-36 ..tOP §EERI!T~ 18. NSAlCSS Archives, ACC 16824, CBTB 26; Jeffrey T. Richelson and Desmond Ball, The Ties that Bind (Boston: Allen and Unwin, 1985),136-7. Hereafter Richel$On, Ties. 19. "A Chronology of the Cooperation Between the SSA and the London Offices ofGCCS," 2 June 1945, CCH Series VI.V.7.2. In addition, the entire Army-British and Navy-British relationships during the war are covered in detail in Benson, "History." 20. CCH Series V.A.29. The el;lrly collaboration with the British on the Soviet problem is covered in George F. Howe andl I"Historical Study of COMINT Production Under the Joint Operating Plan, 1946- 1949," in CCH Series V.E.!.!., and in Michael Peterson, "Early BOURBON - 1945. The First Year of Allied Collaborative COMINT Effort against the Soviet (Jnion," Cryptologic QualUrly (Spring 1994). 21. See Howe, "Narrative"; Burns, Origins; Benson, "History"; Peterson, "Early BOURBON"; and CCH Series VI.J.1.2. 22. "Origins ofthe SUSLOs," in CCH Series X.H.8. 23. "Historical Summary ofU.S.-Canadian/COMINT Relations," 12 April 1949, in CCH Series V.J.3.; NSAlCSS Archives, ACC 16824, CBTB 26; Richelson, Ties; oral history interview with Frank B. Rowlett, various dates, by Henry F. Schorreck andl ~SAOH 14-81. See also Howe, Narrative, and Benson, "History." 24. Copies of papers from the Harry S. Truman Presidential Library in Independence, Missouri, in CCH Series XVI. 25. The early problems between the U.S. and Australia in COMINT cooperation is covered in Richelson, Ties; Howe, "Narrative"; Benson, "History"; copies ofpapers from the Dwight David Eisenhower Presidential Library in Abilene, Kansas, contained in CCH Series XVI. Specific information about the Australian spy scandal and its impact on COMINT collaboration is covered in Christopher Andrew, "The Growth ofthe Australian Intelligence Community and the Anglo-American Connection," In.telligence and National Security (April 1989), V. 4, # 2: 213- 256; Robert Manne, The Petrov Affair: Politics and Espionage (Sydney: Pergamon, 1987); and Desmond Ball and David Horner, "To Catch a Spy: Signals Intelligence and Counter-espionage in Australia, 1944-1949" (Canberra: Strategic and Defence Studies Centre, Australian National UniversitY,1993), pending publication. 26. Vincent Las Casas, NSA's Involvement in U.S. Foreign SIGINT Relationships through 1993, United States Cryptologic History, Series VI, Vol. 4(Ft. Meade: CCH,1995). IIANBbB VIIi 'Y'AbBN'F KBYII8bS e8MUH 08n'Y'RebSYS'Y'SMSd8m'fbY NOT REI EAgARI F TO EOREIQ)T ~T-tlTIQPT1\LB 21 lQPSe(REI~ Chapter 2 AFSA and the Creation of NSA The formation of AFSA resulted from both technical and budgetary causes. The technical concerns were first surfaced within the Army Security Agency (ASA) over the conclusions of a study on World War II German SIGINT done by the Target Intelligence Committee (TICOM - see chapter 1). TICOM had studied the German failure to crack high-grade Allied codes and ciphers and concluded that it resulted from a badly fragmented effort. The Germans mounted at least five different cryptanalytic efforts. Each competed for resources and attention, and each jealously guarded its resources and techniques from outside encroachment.1 The result was failure. As Frank Rowlett, perhaps the leading ASA cryptanalyst in 1948, said, "they all skimmed the cream off and they did the easy ones and nobody, none of them, were [sic] ever able to concentrate on the more important and more secure systems and bring them under control." THE STONE BOARD The disastrous results of German cryptologic competition spurred Rowlett and his associates to press for unification of the American effort. In 1948, under the direction of Brigadier General Carter Clarke, Rowlett chaired a committee to write a paper proposing cryptologic unification. The committee included some ofthe leading names in subsequent American cryptology, including Herbert Conley, Benson BufTham and Gordon sOmmers. Rowlett's concerns were mainly technical. With so many good cryptanalysts leaving the services, there was a greater need than ever to concentrate resources. Fragmentation would guarantee the same fate that had met the Germans. This technical argument had been supported in 1946 by the results of the Congressional Pearl Harbor Committee, which, as part ofits final report, recommended cryptologic unification.2 Army secretary Kenneth Royall was persuaded to support unification, but at his level the concerns were mainly financial. Royall was concerned that the formation of the new U.S. Air Force Security Service (USAFSS or simply AFSS) would mean a smaller slice of the monetary pie for ASA. His report convinced Secretary of Defense James Forrestal, who in August of 1948 established a DoD-level committee to look into the matter of cryptologic unification. Although the committee contained members of the intelligence establishments of all three services, it became known as the Stone Board, after its chairman, Rear Admiral Earl E. Stone, the director ofNaval Communications. IIldiBcB VIA 'I'AcBN! IE:BYIISbB eSMIN! eSfi!fteb SYS'fElMS6SIN'fbY NOT REI FA SASlsi: TQ FQRl!iI€lN NIt'fISnALS 23 ~PSEeREr~ Rear Admiral Earl E. Stone, Director, Naval Communications IIMfBbS YIA 'ftrbElN'f I(SYII8bFJ e8?l1Uff eeN'fReb Syt;'fFJMS dem'fbY ~+g'l' R.iJ"K • SASbS 'FQ PQRSIaN UA:'fI8N'AbS 24 Louis A. Johnson, secretary ofdefense in 1949 The Stone Board was anything but harmonious. The Navy was dead set against unification, and Stone was the "chief arguer" (in his own words) against the concept. He got the Air Force behind him, and the result was a majority report arguing against the very concept it had been set up to consider. That report agreed to certain reforms in the current CJO (Chief of Joint Operations; see chapter 1) set-up, but refused to endorse any sort of thoroughgoing restructuring. The Army report favored cryptologic unification under a single agency, but it was only a minority report. The two documents were sent to Forrestal. Since the majority report favored a sit-tight approach, nothing happened, and the results of the Stone Board languished in a desk drawer until after the death of Forrestal in March of1949.3 It is important to understand what was going on at that time. The interservice rivalry which had characterized American conduct of WorId War II had led to calls for service unification. The first step toward a reform ofthe U.S. military structure was the National Security Act of 1947, which established the Secretary of Defense, the National Security Council, and the CIA. Although all three institutions have become very powerful, in the early years they were not, and gaining control of their respective domains was a process marked by fierce rivalry and bitter infighting.4 The new secretary of defense, Louis P. Johnson, arrived at the Pentagon during the worst of these interservice clashes. Cryptologic unification was one of the most hotly contested issues. The protagonists did not leave him alone very long. Carter Clarke pushed Johnson hard on the issue. According to Clarke's own description, he approached one of Johnson's top aides, General Alfred Gruenther, to resurrect the Stone Board documents. Clarke argued that lack of unification was partly responsible for the failure at Pearl Harbor. Johnson, apparently impressed by this, called in General Joseph T. McNarney, a known supporter of unification. McNarney wrote a report which recommended creation of a central organization, called the Armed Forces Security I1ld'fBl=Jll ';1>'. 'i'Abl!lJ'l"'i' KI!l¥IIQbE GQMIN'i' GQJ'l"'i'RQb SYS'i'EfIIS oJ8IN'i'bY WT RiU A ~ A In.i TQ VQREIQW J'fA'i'IQJ'W.bS 25 Ib) (1) Ib) (3) -P.L. 86-36 Ib) (3) -50 USC 403 Ib) (3) -18 USC 798 Agency, but which retained the separate cryptologic organizations of the three services. The report was then discussed at a JCS meeting on 18 May 1949. At this meeting the Air Force chief of staff, General Hoyt S. Vandenberg, changed the Air Force vote to prounification. The minority had suddenly become the majority, and it was clear that unification was to be forced through. The Navy quickly reversed its vote, too, and the decision to create AFSA was unanimous. Why did Vandenberg change the Air Force vote? He may have seen the creation of AFSA as an essential ingredient in better intelligence, but he may also have felt that he could keep the fledgling USAF Security Service effectively independent. Vandenberg's central concern in those days was to establish a strategic strike force (Strategic Air Command, or SAC) which would be supported by an all-Air Force intelligence center. He regarded SIGINT as the key ingredient in such a creation and wanted to place a SIGINT analysis center within USAFSS which would be beyond the control ofAFSA. It is possible that he changed the Air Force vote after assurances that USAFSS would be permitted to establish such a center. (This center, called the Air Force Special Communications Center;. was actually created, and it resided at Kelly Air Force Hoyt S. Vandenberg Base, home of USAFSS, for many Provided the "swing vote" years.) The later creation of thel I that created AFSA I la device to keep intercept facilities independent of AFSA, might also have been part of such a plan. Vandenberg's thinking was probably also influenced by log-rolling in other areas, and may have represented an attempt to obtain Army support for other Air Force programs by yielding on the cryptologic issue.5 AFSA And so the Armed Forces Security Agency was created on 20 May 1949. It was promulgated by JCS directive 2010. AFSA was thoroughly military, and, because it answered to the JCS, its central concernswere all military. Organizations outside the JCS got short shrift in the collection ofintelligence. State Department and CIA were intensely ILurabS VI"'l'AbS~rl'l'fSYIlQU SQMUTl' SQ~rl'aQb 8Y8l'SMS dQUTl'bY MQT KiU 1\~"'Rbi TQ FQaSIS~r ~rAl'IQPiAb6 26 unhappy with this development, but they lacked the power to wrench AFSA out of the military chain ofcommand. AFSA began life in borrowed quarters. Its people, just over 5,000 in the beginning, occupied spaces in Arlington Hall and the Naval Security Station on Nebraska Avenue, sharing space with the Army Security Agency and Naval Security Group from which the space was obtained. Admiral Stone decided that the Naval Security Station would be used by AFSA for COMSEC, while the COMINT mission would be done at Arlington Hall. This decision began a historic physical separation between SIGINT and COMSEC which has never been completely bridged, despite the later move to Fort Meade. It was logical, though. Naval Security Group (NSG; formerly OP-20-G) was strong in the COMSEC discipline. Moreover, the Naval Security Station (NSS) at Nebraska Avenue had only about onefourth the space available that Arlington Hall did, and this disparity in size meant that NSS was about the right size for COMSEC, while the larger spaces at Arlington Hall would be ideal for COMINT. There was a certain amount ofshuffiingback and forth as COMINTers from NSS moved their desks to Arlington Hall and COMSEC people from Arlington Hall transferred to NSS. But when it was finished, all the COMSEC people were housed in almost 214,000 square feet ofoffice space at NSS, while the COMINT operations were lodged in 360,000 square feet at Arlington Hall. Including administrative, storage and machine space, there were only 79 square feet per worker at the Hall, but about 98 square feet at NSS. Workers often sat at tables rather than desks, in large warehouse-like rooms, cheekby-jowl, as they worked complex code or callsign systems. Floors were tiled and the noise level was high. There was practically no air conditioning, and in the summertime it was common to close down for the day when the ratio oftemperature to humidity got too high. AFSA owned two other facilities. The cryptologic school, a rudimentary training ground used originally to keep newly hired workers busy before their clearances came through (see p. 71), reposed in a structure on U Street Northwest in the District of Columbia. The Agency also maintained a courier facility at National Airport, then called Congressional Airport.6 The impact ofAFSA on the services was immediate and severe. Besides turning over more than 600,000 square feet ofspace to the new organization, the Army and Navy had to donate about 80 percent oftheir existing Washington-area billets - 79 percent for ASA and 86 percent for NSG. Although ASA kept many of its uniformed service people, its corps of over 2,500 civilian experts was turned over to AFSA virtually intact. This made the Service Cryptologic Agencies little more than collection organizations, with practically no central processing - all arms and legs, but no body. This revolution was accomplished virtually overnight with only minimal dissension and was AFSA's most noteworthy success. IIlrNBl"B VIA 1f...l"SNIf KBYII8bE S8MIN'f S8?f'fR8l" SYB'fEMBd8fN'fbY 27 Analytic section, Arlington Hall Station The sole exception to this trend was USAFSS. The Air Force cryptologic agency practically seceded, opening its first headquarters at Brooks AFB, Texas, 1,600 miles away from the menace of centralization. Even more startling, it was required to donate only thirty officers, twenty civilians, and eighty enlisted billets to AFSA. So when USAFSS opened its processing center, it had plenty of billets to do it with. If this was what Vandenberg had in mind, it was working.7 AFSA organization reflected service competition. The director was to be chosen from among the three services on a rotating basis, and its first director was its most ardent opponent, Earl Stone. Assisting him were three deputy directors, one for each service. Below them were four major divisions, which have survived to this day - Operations, HltNBLS VIIt 'fltLSN'f KS'IlIeLe CeMIH'f ceNTReL SYSTeMS66IN'fLY NOT ail i blil b2bi: 'l;'Q i'Qil,gIQ}I N....iUHL\L8 28 J!lb) (1) Ib) (3) -50 USC 403 Ib) (3) -P.L. 86-36 Research and Development,COMSEC, and At:lmirtistration. The office designator system was numerical, so that Operations was AFSA 02., R&D was 03, COMSEC was 04, and Administration was 05./Each of the military deputy directors also had a sphere of influence. The NavYideputy director, Captain. Joseph Wenger, controlled COMINT, while the Army deputy, Colonel Sam:ueIP. Collins,\supervised COMSEC, and the Air Force deputy, Colonel Roy Lynn, handled administrative matters.s The field collection effort consisted of the iriterceptsites which had survived the budget cuts after/Wo.rld War II. Army SecuritY\Agencyhad seven sites: Vint Hill, Virginia; Petaluma,/California;1 lHelemano, Hawaii;1 I Fairbanks Alaska' and Clark AFB in the Philippines. The Navy had twelve: 1r- ......._,.....__---JDupont. ~:~hA~....r:~!na~\· \. \ L- .rkaggs Island, California; Cheltenham, Maryland;1 The Air Force had ten mobile units, whose status and location were somewhat vague. Finally, ASA had six SHAMROCK units, whose task was to screen commercial cable messages turned over to ASA by the cable companies under an arrangement which had existed since World War 11.9 Field intercept was the rock that sank AFSA. In theory all the intercept positions were to be under AFSA control. In fact, some were not. Of the 763 intercept positions existing at the time AFSA was dissolved, 671, including all the Army positions, were under some form of AFSA control. Just over 100 were reserved by the Navy for fleet support and were thus completely beyond AFSA tasking authority. But even the positions under AFSA control could be tasked only by treading a complex paper mill by which tasking was routed through the SCAs, rather than being levied directly. This was true especially in the Navy and Air Force - the Army was more accommodating and permitt~d some form ofdirect tasking. Completely beyond AFSA purview, however, were the mobile intercept stations. In theory, these were small mobile efforts for direct tactical support. But AFSS flouted AFSA control by simply designating all their stations as "mobile." Thus even the most permanent and sedentary station was desi ated as a "radio au mo' " , beyond AFSA control. The Army and Navy quickly caught on, and by....._...,...,....,...~-~ 1952 ASA had seven mobile units, while the Navy had three. AFSA's lack of tasking authority over Air Force positions was intolerable, and late in 1950 Major General C. P. Cabell, Air Force director of intelligence, and Rear Admiral Stone signed an agreement granting AFSA the authority to task automatic Morse and radioprinter positions, while USAFSS retained control over voice. The Morse positions (1) Ib) (3) -50 USC 403 Ib) (3) -P.L. 86-36 IIMfBbFJ IfII. 'F/.bBN'f KBYIIS15B 8SMHi'f eSN'fRS15 BYB'fBM8dSHi'fc'[ NQ'i' RBf5BASltBbFJ 'f0 FORSIaN' N'A:fIOftAf:S 29 Ib) (1) Ib) (3) -P.L. 86-36 Ib) (3) -50 USC 403 Ib) (3) -18 USC 798 lOP 5I!CREfij~RA were split 50/50. Still later, in 1951, this arrangement was changed when the new director ofAFSA, Lieutenant General Canine, and Colonel Lynn of USAFS$ signed an agreement dividing the Air Force positions down the middle, regardless ofmode ofintercept. Meanwhile, USAFSS established its headquarters in San AntoniQ - first at Brooks AFB and later at nearby Kelly AFB, on a low rise west of the runway which is now known as Security Hill. Within its headquarters it proceeded to establish a. Stateside COMINT processing center, Air Force Special Communications Center (AFSCC). This was done despite direct orders by Canine that it not be established. AFSA also directed that USAFSS not establish third-echelon processing on theI Itarget, but USAFSS did it anyway. Air Force defiance fragmented the processing effort and had much to do with the demise of AFSA. Despite this, AFSCC continued to process on thel Itarget until the late 1960s, when it was finally turned into an electronic warfare center.10 Service rivalry led to duplication. During the early days of the Korean War, for instance, both ASA and USAFSS covered the Soviet and Chinese air problems in the Korean area, and ASA did not discontinue its coverage until March of 1952, after many months ofAFSA mediation. Likewise in the DF area, AFSA was unable to force a common DF net control for the Korean problem for more than a year. Ultimately the Navy kept its DF system separate. All three SCAs established second-echelon processing centers in the Pacific with or without AFSA blessing. Without firm control of SIGINT, there was simply no way to organize effectively. This lack ofcontrol attracted unfavorable reviews from the generals trying to fight the Korean War and played a part in the COMINT reorganization of 1952.11 The final blow to AFSA was the development of a policy mechanism outside ofAFSA itself. It was called the Armed Forces Security Advisory Committee (AFSAC), and it was created by the same JCS directive that established AFSA. The original plan was for an advisory committee composed of nine members - three from each service - chaired by the director of AFSA. But the JCS gradually changed AFSAC's charter from advisory to directive. Had AFSAC possessed a proper decision-making mechanism, the conversion of its role to that of direction might have worked after a fashion. But the rules required unanimity on all substantive matters.12 AFSAC was immediately immobilized by interservice disputes and was ineffective from the start. AFSA had become a body with no head. Ih\NBbFl 'iIA i'A::bFlf~TRFlYII6bFl e6MH~Te6H"flt6L S'tSTEMS J6H~TL i NOT RE' E A 51' RUi ~Q F'QRFlI8H HJli:T16I{1\LS 30 ~ 00 >~ 00 00 ::~ ~ .0 C II> '"...~ '"Y' ~ ~ ~ t.,) >.... ~ ,p:l II> '"•.... ~ 'l:S ~ ~ ~ ~ S' ~ c -<... 1 c= en ~ ~ ~ Ib) (1) Ib) (3) -P.L. 86-36 Ib) (3) -50 USC 403 Ib) (3) -18 USC 798 One small success during these early years was the development of customer liaison organizations. By 1949 both the Army G2 and the Offrce of Naval Intelligence had established informal liaison offices with their cryptologic counterparts at Arlington Hall and NSS. When AFSA was established, these arrangements continued undisturbed. Both the Army and Navy groups developed a very close relationship with AFSA, and their people often worked in an intelligence production role. By the end of the Korean War, the Army organization, which called itself SRB (Special Research Branch), had some nfty people. Air Force Intelligence had a similar group, which was gradually subsumed by AFSS into a large organization of over sixty people performing both a customer (for Air Force Intelligence) and producer (for AFSS) role. Thus the Air Force group performed both as a producer and consumer, while the Army and N"avy acted only as producers. Both CIA and State maintained small offices within AFSA, under a USCIB edict of 1948. Although AFSA regulations permitted them to see semiprocessed intelligence, they never participated in the production process, maintaining their offices for liaison purposes only. FBI's refusal to establish any office at all reflected J. Edgar Hoover's adamant opposition to COMINT centralization.13 While COMINT was fractious, COMSEC/was relatively serene. During World War II there had been a single authority for joint service communications matters, the U.S. Joint Communications Board, established in July of1942. Its principal members were the chiefs of communications for the Army, Navy, and Air Force. In 1948 it gave way to a new organization, the Joint CommunicaWms-Electronics Committee (JCEC), which reigned supreme in this area for many years thereafter. The JCEC was concerned with communications planning, standards, and interoperability, but its charter by implication gave it a determining voice in COMSEC policy as well. When AFSA was created, JCEC effectively transferred central COMSEC functions to it. The charter did not extend to non.JCS organizations, but the State Department and other civilian agencies with communIcations security concerns had for years relied on the Army and Navy for COMSEC support., and this reliance was transferred to AFSA. AFSA began producing codes and ciphers for all the armed services and many of the non-DoD agencies. In addition, it undertook centralized COMSEC R&D functions, planning and programming, setting of security standards, and technical supervision of the communications security activities of the armed services. The SCAs retained many residual functions, such as distribution ofAFSA-produced codes, security monitoring oftransmissions, and the like.14 While AFSA successfully controlled the highly technical function of COMSEC, it was never able to control COMINT. This lack of control made powerful enemies. The State Department was upset because, under AFSA, the number of positions allocated to I lactually declined in the three years ofAFSA existence, from 64 to 51, and trom almost 1'1 percent ofthe total to only 6.5 percent. nA:UBtiB YIA ~tiBU'fIEBYUStiB eSMUii'f eSN'fRsr:: SYS'fBMStJSIU'fti7 NOT RELEASABI E TO ~~I8!, UkYI('JNAtS 32 - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - (b) (1) (b0l3)-SO USC 403 (b) (3)-18 USC 798 (b) (3) -'P.L. 86-36 TOP SECRET ~M8ftA THE BROWNELL COMMITTEE The entire intelligence community was concerned over performance of the COMINT system in Korea. AFSA had not predicted the outbreak of war. A watch committee established under the wing of CIA in early 1950 listed Korea tIfth on the list of world trouble spots, but this was not translated into action and when the war be an AFSA still had no positions allocated to Korean military. AFSA had no more dangerous opponent than Walter Bedell Smith, director of Central Intelligence. In 1950 the wartime feud between the COMINT empire and Smith's HUMINT organization boiled over. On 10 December of that year Smith wrote a memorandum recommending that a committee be established to "survey" COMINT. Smith was "gravely concerned as to the security and effectiveness with which Communications Intelligence activities ... are being conducted." He pointed to "the system of divided authorities and multiple responsibilities" which was endangering national security. The National Security Council in turn forwarded the recommendation to President Truman, who directed that a committee be formed. Walter Bedell Smith The JCS could not take heart from the Director ofCentral InteUigence composition of the comittee. Its chairman was George A. Brownell, a New York lawyer and layman in intelligence matters. The members were Charles Bohlen, a prominent State Department official~William H. Jackson, special assistant to the DCI; and Brigadier General John Magruder, special assistant to the secretary of defense. Thus the Joint Chiefs, who owned the COMINT organizations, had no one on the committee. It was composed of"enemies," representatives from State and CIA - the two most vocal opponents ofthe existing system. HANDLE VIA TALENT KEYHOLE COMI SYSTEMS JOINTLY NOT TO FOREIGN NATIONALS 33 TOP SECRET ~MBRA- - George A. Brownell The Brownell Committee held fourteen days offormal sessions, which were backed by many days of research and data-gathering. Its report was a scathing indictment of old ways ofdoing business. Its bottom line stated bluntly that The added difficulty ofthe problem under attack places ....._----------_...a greater premium than ever on the quantity and quality of the physical and intellectual resources available, and on the efficiency and clarity of the organization charged with the task. While much has recently been done to provide adequate physical resources for the job, the Committee is convinced that the present organization of our COMINT activities seriously impedes the efficiency ofthe operation, and prevents us from attracting and retaining as much top quality scientific management manpower as this country ought to be investing in so important a field. It is highly significant to the Committee that the return of many of the best wartime COMINT brains to more attractive llAf(Bh~VIA 'fA~N:'I' l<:fJ t 116hfJ e6MH('I' e6I('I'It6t: S7[S't'fJ!tISd'61Irf't't: f NOT RELEASABLE TO FORi:IElN 'PM'fIONALS 34 The committee concluded that the creation of AFSA, coinciding as it had with the creation ofUSAFSS, had resulted in four COMINT agencies where there had formerly been two. It criticized AFSAC for obstructionism and requested that it be abolished. It attacked USAFSS as a virtually autonomous organization not operating under joint control at all. The positive recommendations of the Brownell Committee are worth studying, because they encompass the present-day structure of SIGINT in the United States. AFSA should be greatly strengthened, especially in its ability to control tasking at SCA collection sites. AFSA or its successor should be removed from JCS control and should be placed under USCIB, whose membership should be revised, and whose procedures should be governed by a vote of four, rather than unanimity, as had been the case with AFSAC. AFSA should centralize and consolidate processing operations wherever possible to increase the resources brought to bear on intractable cryptanalytic problems. The director should be upgraded to three-star rank, and should be appointed by the president to a fouryear term. He should have a civilian deputy. Civilian career development should be encouraged to a much greater extent than formerly. The next several months were spent putting the Brownell report into directive language. The result was the Truman Memorandum, issued on 24 October 1952. This memo directed a complete restructuring of COMINT along the lines that Brownell recommended. It resolved an on-going dispute about how to change AFSA by abolishing it and creating in its place a new organization called NSA. Its director would work for the secretary of defense, who would become the "executive agent" for COMINT for the entire government. On the same date the National Security Council issued a revised NSCID 9, almost a verbatim quote of the Truman Memorandum. Both documents were classified Top Secret, thus hiding the official creation of NSA from the American public for many years. All that remained was for the secretary of defense to issue a memorandum establishing the new agency. He did so on 4 November the day that Dwight Eisenhower defeated Adlai Stevenson for the presidency. The creation of NSA was one of the last historical legacies of twenty years ofDemocratic governance. The Truman Memorandum, on the advice of Lieutenant General Canine, had excluded COMSEC. Despite his belief that NSA should have both a COMINT and a COMSEC role, Canine recommended against mixing both in the same document. Lovett's memorandum on 4 November did mention that NSA would inherit the COMSEC functions formerly performed by AFSA. A memo in December spelled out those functions in more detail, and this marked NSA's first formal COMSEC charter.17 il.A :-TQbi: VIA TAUJtfT KKYII8bK 68MIU'f 68N'fR8b SYS'fBMS d8lfffbY NOT REI F A ~ A ibi: TQ P8RrJI8N NA:TIOH'1"L5 35 fe, 5EEREf UMBRA KOREA It has become apparent ... that during the between-wars interim we have lost, through neglect, disinterest and possibly jealousy, much ofthe effectiveness in intelligence work that we acquired so painfully in World War II. Today, our intelligence operations in Korea have not yet approached the standards that we reached in the final year ofthe last war. General A. James Van Fleet, Commanding General 8th Army, June 1952 The Country American intelligence interest and attention, so painfully refocused on the Soviet threat after World War II, were not to be rewarded. The next war occurred not in Europe, where allies and commitments were, but in Korea, a remote Asian peninsula whose name many Americans had never heard in 1950. Korea had, throughout its recorded history, been a battleground between China, Japan, and Russia. Frequently invaded and occupied, its primary purpose seemed to be as a strategic buffer among three conflicting imperial ambitions. The most recent change of ownership had come after the Russo-Japanese War of 1904-05. Russia, the loser, was forced to cede its influence. Korea became forcibly Japanese. The Allied powers recognized during World War II that Korea was one of those geopolitical oddities whose status had to be resolved. It obviously could not remain Japanese, and so at the Cairo Conference of 1943 Roosevelt endorsed a policy that would ensure a "free and independent Korea." At Yalta in April of 1945, the Big Three (the United States, the USSR, and Britain) agreed to an Allied trusteeship, to be administered by the three plus China. Nothing further happened until the USSR declared war on Japan on 8 August 1945, simultaneously invading Manchuria and Korea. The sudden movement of Soviet troops onto the peninsula appeared to portend Soviet occupation, and MacArthur was directed to rush troops to the southern end of Korea. The United States proposed a division of inilitary occupation on the 38th Parallel, splitting the peninsula roughly in half. Moscow unexpectedly agreed, and still more unexpectedly, complied. American forces dwindled down to about 30,000 by 1948. In March of that year President Harry Truman, following the country's mood of dedicated military budgetcutting, decided that America would simply have to abandon Korea to the United Nations, I IlrUBbEl YIA 'f'/rbElH'f IEElYlf6LfJ e6MIN'f e6I{'fft6L S'fS'ffJI'ifSJ6IH''fL1 i OP 5!EREf UMBRA 36 TOP SEeRE j UMBRA MANCHURlA .. .' Korea, 1950 SOUTH KOREA W" )TQblii "I'\: l'Abl!:l'f'I:' KI!:YHQbl!J OQMIN'f OQU'fR8L ST[S'fI!1MSd8HffL { NO1 REU:A~5L~ "6 P6ltEI6N 1b8!:'fI8NALS 37 Tap SEe"'!T tJMBRA rOPSECREi tJM8itA to sink or swim on its own. He decided to end the American trusteeship and sponsor free elections. So in the spring of 1948 American forces marched out of Korea. The South boycotted the elections, which led to a new National Assembly and a government headed by Syngman Rhee, a seventy-three-year-old militant anti-Communist who had spent forty years in exile in the United States waiting for the liberation of his homeland. The North formed its own government, the Democratic People's Republic ofKorea (DPRK), headed by a young thirty-six-year-old Communist named Kim II-sung. The peninsula was divided at the waist. Syngman Rhee The Asia Dilemma Kim II-sung In 1949 catastrophe struck in the Far East. The corrupt and despotic Chiang Kai-shek and his Nationalists were ousted by the Communist forces of Mao Tse-tung. As the Communists marched into Beijing, Chiang fled to the island of Formosa (Taiwan), some 100 miles off the coast, followed by as much ofhis army as could flee with him. By the end of the year, Mao was making confident proclamations about his intent to invade Formosa and drive Chiang and his army into the sea. In Washington, the administration was convulsed over whether the United States should support Chiang and the Nationalists. In the end the anti-Chiang faction won, and Truman, on 5 January 1950, issued a public statement tha~ the United States had adopted a "hands off Formosa" policy. Ambiguity about which side of the line Korea stood on was HMlQbl!J VIA Q\'tl5Bl'l'f IiBYIlSl5S GSMHcf'l' GSN'l'R6b S7{S'f~MSjOUif'fLi NOt RELEASABLE 10 FOREIG~ M1\TII5N'A:LS 'Fe, 'ECRE I UMBRA 38 Ib) (1) Ib) (3) -50 USC 403 Ib) (3) -P.L. 86-36 Ib) (3) -18 USC 798 Ib) (1) Ib) (3) -50 USC 403 Ib) (3) -P.L. 86-36 Tap SE(RIiT UfldBRA resolved a week later when Secretary of State Dean Acheson, at a press conference, described an American sphere ofinterestin the Pacific that implicitly excluded Korea. By June 1950 the United States had boxed itself into a very weak position in Korea. From a full army corps, it was reduced to a 500-man Korean Military Aid Group (KMAG). The U.S. had left behind plans and equipment for a 50,OOO-man ROK (Republic of Korea) "constabulary" (rather than a real army) but devoid of heavy equipment, as the U.S. was afraid that the militant Rhee would use it to invade the North. Rhee drew up plans for a real army of 100,000, and he succeeded in extracting additional American commitments of weapons (but still no heavy, mobile offensive weapons). On the other side of the 38th Parallel stood a DPRK army and air force of\about 135,000 men, equipped by the Soviets with much ofthe heavy equipment that the Americans had denied to Rhee. American military forces, overall, in 1950 were in a weakened state. Defense budgets had continued to decline from their World War II peak, and the defense budget for 1950 was only $12.3 billion, with an authorized Army strength of 630,000 (but an actual strength ofonly 591,000). Of these, only 108,500 were in the Far East, almost all of them in Japan. In line with administration policy, the Pentagon had no plans to defend Korea and no one there to do it. The American contingency plan for the peninsula was basically to evacuate all dependents to Japan.18 Parallel to the national lack of interest in Korea was AFSA's neglect of the problem. There were no documented high-priority national intelligence requirements on Korea, and the only requirement that related~t all was couched in terms of keeping track of Soviet interest in the peninsula. At the time AFSA had "no person or group of persons working on a North Korean problem." During the previous year, SeA intercept sites had stumbled onto som.....~ JNorth Korean messages which were originally collected as suspectedL ,When in May 1949 these messages were identified as North KoreaIl,two intercept positions atl Iand a tactical unit not under AFSA control, were tasked with follow-up copy. AFSA had no Korean linguists, no Korean dictionaries, no traffic analytic aids, and no Korean typewriters.19 No one really expected an invasion in Korea. There was fragmentary HUMINT reporting, generally disbelieved by all, that there could be an invasion by North Korea in 1950. In March an Army organization called the Intelligence Indications Steering Committee cited the possibility of military activity in Korea sometime in 1950. But this was set against a general disbelief in the intelligence community that Korea presented a real problem. Mter the war broke out, there was the usual scramble by intelligence agencies to find the indicators that had been missed. AFSA, for instance, discovered traffic indicating that there had been large shipments of medical supplies going from the USSR to Korea beginning in February. A Soviet naval DF net in the Vladivostok area had undergone a IIAl'fQbS ¥IA 'FAbSl'f'F I'ieSYllebS OeMIN'f eeN'fR8L 8""iB'f~MBJ6U(TL i NOT RELEASABLE '1'0 POREIGN NAtIONALS 39 T9P SE(RET liMBRA TOP SECRET liMBItA dramatic switch to South Korean DF tasks beginning in February.2o This did not quiet the critics. The Invasion About 0330 on Sunday morning, 25 June 1950, Captain Joseph Darrigo, a KMAG military advisor to the ROK posted near Kaesong, was jarred awake by the roar of artillery. Darrigo, the only American on the 38th Parallel, was in the middle of an invasion of North Korean ground forces into South Korea. He managed to make it to the ROK 1st Division headquarters at Munsan just ahead of the advancing North Korean forces, and he spread the alarm. There appears to have been no tactical intelligence warning. A reporter in Seoul got word of an invasion and rushed to the American embassy for confirmation. At the same time that he got off a wire to New York, the American ambassador was cabling Washington. His cable had to be encrypted and decrypted, and it got there late. The Americans learned ofthe invasion from the reporter in Seoul.21 ASA decided to support the fighting with a communications reconnaisance battalion at Army level and three battalions to serve each of the three corps. The 60th Signal Service Company at Fort Lewis, Washington, appeared to be closest to being ready for deployment of any ASA tactical asset, so that organization was selected. But it took time to get ready, and in the meantime ASA Pacific (ASAPAC) in Hawaii rushed a signal collection unit to the Korean peninsula, arriving there on 18 September. The Fort Lewis unit did not arrive until 9 October.22 Meanwhile, the Truman administration had decided to help the fledgling ROK army and got UN backing for the deployment of a multinational defensive force to Korea. Truman directed MacArthur to rush the 8th Army from Japan to Korea, and the first American troops reentered Korea by air on 1 July. But it took time to get enough troops into the country, and the DPRK army charged ahead, pushing ROK defensive units ahead of it pell-mell. By mid-August, ROK defenders had been shoved into a perimeter around the port city of Pusan, the last remaining large city still under the control of the Rhee government. When the first ASA unit arrived in September, the ROK army, bolstered by newly arrived American divisions (the 24th Infantry, 25th Infantry and 1st Cavalry), was desperately hanging onto this slice ofthe Korean landmass, and the American and Korean defenders were in the middle ofa fierce struggle to retain the town ofTaegu.23 ASA's primary concern was to get linguists. Perhaps the only two first-rate Army Korean linguists were Y.P. Kim and Richard Chun, who were both instructors at the Army Language School in Monterey in 1950. Chun had been cleared in World War II, but Kim had never been in the COMINT business. ASA needed linguists at Monterey to train what was expected to be a sudden flood of Korean language students, but they also needed someone in Korea who could translate Korean. ASA hesitated just a brief moment, and IIA~TQbE VIA 'fAbEtT'f ICBYIIeJ=:B OeIllIN'f OSH'fRSJ=: S"IS'frJM86Sm'fbY 14'61:' ftrJbrJJltSJltBhB 'fe FefH"lI8N NNI'IQNAbl!l TOP SECRET liMBI\"A 40 Ib) (1) Ib) (3) -P.L. 86-36 TOP SECRET tJMBRA then Kim and Chun, neither as yet actually cleared for COMINT, were on their way to Korea to assist the newly arrived ASA tactical COMINT unit. Until their clearances came through, they worked in a locked and guarded room every day. Intercepted messages were brought in periodically. They would translate the traffic and then pass it through a slot in the wall to the communications center.24 The Air Force Security Service likewise ~ad one unit in the Korean area in 1950 - the 1st Radio Squadron Mobile (RSM) at Johnson Air Force Base outside Tokyo. This unit had been created in 1942, and it had supported 5th Air Force through MacArthur's Pacific campaign from New Guinea to Japan. In 1950 it was still engaged in support to 5th Air Force, but by then/had changed its mission to I I I lIn late June it scrambled to change over to Korean targets. It had no cryptanalytic capability, and so began with a traffic analytic attack against North Korean air targets. It likewise had no cleared Korean linguists, so it could do little against readable voice communications.25 The Murray Mission The Air Force Security Service actually beat ASA to Korea - their first representative, First Lieutenant Edward Murray, arrived in Taegu on 19 July. But Murray's mission quickly became entangled in one of the most bizzare incidents in the history of American cryptology. When Murray arrived, 5th Air Force already had a COMINT service. The origins ofthat organization are very murky but appear to go back to the days after the end ofWorld War II. At the time a civilian named Nichols, who also had a reserve commission as an Air Force major, headed the local Air Force Office of Special Investigations. Nichols, whose background and training in COMINT are completely unknown, decided that Korea needed a COMINT service. The South Korean government under Syngman Rhee did not appear interested, so Nichols proceeded on his own, seeking out the assistance of some "Koreans with COMINT experience. Among his recruits was one Cho Yong 11, who had come from North Korea, where he had been a radio operator and cryptanalyst with the North Korean Army. Joining Cho was Kim Se Won, a captain in the ROK navy. Kim had served as a COMINTer with the Japanese army in World War II and, owing to having been interned by the U.S. Army in Hawaii, spoke excellent English. Cho, Kim, and those who worked for them did intercept and translation work for Nichols; the source offunding has never been discovered. In 1949 Cho, with Nichols's assistance, obtained a commission in the Korean air force (ROKAF), and his group dual-hatted as a private group working for Nichols and as the ROKAF COMINT service. At about the same time the ROK navy set up Kim and some colleagues from the Nichols group as their COMINT service, so they, too, were dual-hatted. Ih'.NBbEl Yh\ '¥t\bElN'f IiElYHSbFl eSMIN'f OSN'fRSb SYS'fBMS d8IN'fbY NO I RELEAS1l:8LE TO POftElffl, H:A:'f18H:A:bS 41 TOP SECRET tJMBRA T9P SECRET liMBIbIc When the ROK army retreated south in July of 1950, Nichols and his COMINT group retreated with them. As they fled south, fissures developed between Cho and Kim, and in late July or early August the Kim group seceded. Cho stayed with Nichols to supply COMINT to the Air Force, while Kim eventually hooked up with ASA units entering Korea. Nichols was reporting directly to 5th Air Force, which was releasing his reports into USAF intelligence channels at the noncodeword level. Meanwhile, AFSS had sent Murray to Johnson Air Force Base to put together a direct support package. Murray assembled some vans and other equipment from 1st RSM, and on 15 July he flew to Korea to set up a mobile COMINT effort. AFSS was operating under a misty-eyed concept of COMINT as covert operations, and 1st RSM was directed to expunge its identifications from the equipment, and to insure that Murray could not be indentified as a COMINTer. The direct support went under the codename Project WILLY. Murray's first concern on arriving in Korea was linguists. Fifth Air Force offered him eight of them, straight from the Nichols pool. The only problem was that Nichols still controlled them, and the upshot was that Nichols wound up with 1st RSM's equipment for use by his own operators. As for 5th Air Force, they were quite happy with the support they were getting from Nichols and informed Murray that he was no longer needed. First Lieutenant Murray returned to Japan on 1 August, having utterly failed to set up a Security Service unit in Korea and having lost his equipment to boot. The breathless nature ofNichols's coup left USAFSS spinning. A severe jurisdictional battle ensued, encompassing command organizations in the United States, Japan, and Korea. Security Service appeared to carry the day, and Murray was ordered back to Korea on 12 August, armed with a letter of authority from General Banfill (Deputy for Intelligence, Far East Air Force). But the struggle was far from over. Nichols was still unwilling to relinquish control of his COMINT organization, and he had the backing of 5th Air Force. Nichols was a local asset under their complete control, was publishing COMINT without the restrictive codewords that limited dissemination, and already had the expertise that Murray lacked. On 17 August, 5th Air Force ordered Murray to catch the next plane out ofKorea. AFSS was again out ofthe picture. The Nichols effort was limited by its lack of national-level technical support from AFSA and USAFSS, and 5th Air Force eventually realized this. On 20 November, 5th Air Force reversed its earlier position and asked for the deployment ofa radio squadron mobile to Korea to provide support. Cho's group became Detachment 3 of the 1st RSM, and Nichols disappeared from the scene. Meanwhile, back in Tokyo 1st RSM was trying to mobilize an effort against the North Korean air force. When Murray returned to Japan the first time he carried with him some captured North Korean code books turned over to him by Nichols. Lacking Korean translators, the unit came upon a Catholic priest named Father Harold Henry, who had spent a number ofyears in Korea as an Army chaplain. AFSS agreed to give him access to IIAHBLB VIA !fAUH'!' llBYIIeU eeMtH'!' eeJ!'f'flteJL S fS'fJ!:MS JeJn~'fL i lap SECRET liMBRA 42 intercepted materials but did not agree to give him an SI clearance. He began applying the code books to the traffic, and he turned out to be a pretty good cryptanalyst, even though he was doing the work without benefit of formal clearance. Father Henry produced the first decrypts of enciphered North Korean air traffic.26 Counterattack While ASA and AFSS were having trouble getting organized tactically, AFSA pushed' rapidly ahead. Despite an almost total lack of expertise and resources to work the unfamiliar Korean target, codebreakers in Washington succeeded in penetrating North Korean communications by late July. At the time, DPRK troops were being readied for their all-out assault on Taegu, which, if successful, might have caused the collapse of the Pusan perimeter and American defeat. Three divisions of Lieutenant General Walton Walker's 8th Army were on line with the remnants of five ROK divisions; opposing them were fourteen battle-tested DPRK infantry divisions. On 26 July AFSA decrypted a North Korean message which contained much of the battle plan for the assault on the 30th. The information reached Walker on the 29th, and he shifted his forces to meet the attack, thus saving Taegu and the Pusan perimeter.27 It was one of AFSA's most conspicuous successes. On 15 September MacArthur launched the spectacular Inchon invasion, the second largest amphibious landing in history, near Seoul. North Korean troops suddenly had a large American force in the rear of their operations. On 19 September 8th Army began its breakout from the Pusan perimeter, and in a brief month they had pushed DPRK forces back north of Seoul. Syngman Rhee's government formally returned to the capital on 29 September. But the dynamic and committed Rhee wanted to push the fighting into North Korea, and on 30 September, ROK troops crossed the 38th Parallel. Washington viewed this development with anxiety. But MacArthur was confident that Chinese and Soviet forces would not intervene and, like Rhee, lobbied for authority to go all the way to the Yalu River. The CIA issued an assessment that MacArthur was right. The risks of invading North Korea appeared minimal, and in the end the Truman administration backed MacArthur. American forces crossed the 38th Parallel on 9 October, heading north. China The Chinese problem which MacArthur was so blithely underestimating had been building for years. The postwar COMINT effort against Chinese communications began officially in 1945 during the mission ofGeneral George Marshall to try to get Chiang Kaishek and Mao Tse-tung to the bargaining table. Marshall, familiar with what COMINT had IIAHBl>JE 'lIlt 'Y'ltl>JEN'Y' KEYIIel>JE OeMIN'f OeN'fR:eb 8YS'FElMS dem'fbY N81f R:I!lbE/rSABbS 'f8 F8RSIaN NIt'fI8Ultb8 43 TOP SECKET b1MBRA Ib) (1) Ib) (3) -P.L. 86-36 Ib) (3) -50 USC 403 Ib) (3) -18 USC 798 done during World War II, requested CO~IINT information from both Communist and Nationalist communications. ASA mounted a small effortag~instboth the Nationalists and Communists. 0 __~~_~ IASA could still report that the two sides were far apart,\ancl it was obvious from the COMINT traffic that they were determined to settle their differences on the battlefield. The Marshall mission was withdrawn in 1946, and in October 01\1949 Mao triumphed. Following the withdrawal of :he Marshall mission, the COMINT mission against China suffered, as ASA employed all available resources against the So7t tarlret.1\,\ IASA kp.nt only a small section against Chinese civil communications I Collection resources were concentrated a~ I security problems.28 When American and South Korean troops crossed the 38th Parallel, the Chinese had already decided to intervene in North Korea. The decision was taken at a meeting in Beijing from 3 to 7 October 1950. On the first day of the conference, Chinese foreign minister Chou En-Lai called Indian ambassador Panikkar to tellhim of the decision, and Panikkar relayed this news to the West. But Indians were regarded as pathologically leftleaning, and Panikkar's communique was disbelieved. Chou's warning was followed up by Chinese radio broadcasts, but these, too, were disregarded.29 Historian Clay Blair asserts that "when MacArthur returned to Tokyo from Wake Island [in mid-October] he had no inkling ofthe CCF armies gathering in North Korea." so This was wrong. AFSA had clear and convincing evidence of the massing of Chinese troops north of the Yalu and had published it in product reports available to the JCS, the White House, and to MacArthur. As early as July, AFSAbegan noting references in Chinese civil communications to army units moving north.• Rail hubs in central China were jammed with soldiers on their way to Manchuria.• By September AFSA had identified six of the nine field armies that were later involved in the fighting in North Korea and had located them in Manchuria, near the Korean border. Ferries at Anshan (on the Yalu River) were being reserved for military use. Maps of Korea were being ordered in large quantities. On 7 November, in voice communications intercepted and published by the COMINT communityI fstated, "We are already at war here."S! fh\UBLJ!J '4 fA 'fALJ!JI,T I{J!J i'II6LJ!J e6MlIfT e6I,Tft6L B1 STflMB i16IIfTLY I,6T ftflLf!lA:SA:8LEl 'fe peRElISn ?tATI8nlrcS 1'OP SECRET l:JMBRA 44 TOp SECRET UMiRA Douglas MacArthur with President Truman on Wake Island, 1951 That was not news to the ROK army. On 25 October a ROK division had been badly mauled by elements of the Chinese 40th Army, already reported by AFSA to be close to Korea. Five days later MacArthur's chief of staff, Lieutenant General Ned Almond, reported that he had seen Chinese POWs being held by a ROK unit. On the first of November, a Chinese force attacked a U.S. unit for the first time. But Charles Willoughby, MacArthur's G2, preferred to believe that these encounters represented isolated PRC volunteers rather than division-strength regular army units confronting UN troops.32 AFSA reports continued to document the presence ofmajor Chinese forces on the Yalu, but the reporting was subtle. AFSA was regarded as a collection and processing agency, not as a producer of intelligence. There were no dramatic wrap-ups, no peppery conclusions - just the facts, strung through a flood of intelligence reports. The COMINT community had almost the only hard information about the status ofChinese forces.33 Intelligence agencies were beginning to pay attention. The Watch Committee of the JIIC, which began noting Chinese troop movements as early as June, concluded by September (probably on the basis ofAFSA reporting) that these troops were moving north rather than to the coastal provinces near Formosa. By mid-October, influenced perhaps by MacArthur's opinions, the Watch Committee had concluded that, though there was convincing evidence that startling numbers ofChinese forces were in Manchuria, the time for intervention had passed - they assessed that the Chinese would not intervene. II1rNBbl'l VIA 'fl\bI'lN'f 1i'I'lYIIObl'l eOMUi"f eON'fftOL &'IS'frJMSdOIN'fLY 45 Tap seCRET liMBRA rop SECRET l::JMBRA However, encounters with Chinese ground and air forces in late October and early November caused the committee to take another look. Admiral Arleigh Burke, who commanded naval forces in the region, was convinced that Chinese intervention was imminent and brought up the subject twice to Willoughby, who summoned his very large staffto try to dissuade Burke.34 MacArthur continued to press ahead with offensive operations to reach the Yalu and get the boys home by Christmas. But on the snapping cold night of 25 November with trumpets braying, thousands of Chinese soldiers fell on unsuspecting units of the 8th Army. The American offensive turned quickly into a defensive, and a defense into a rout. The American and ROK armies were overwhelmed, and some units were virtually wiped out. Weeks later the front stabilized near Seoul, and the war settled down to grim trench warfare for almost three more years. AFSS and ASA Operations AFSS operations in Korea continued their harrowing path. The decision in November to send regular AFSS units occurred just prior to the Chinese invasion. Two locations were envisioned: one in Sinanju to intercept North Korean targets in the battle zone and a rear detachment in Pyongyang to intercept related Soviet and Chinese communications. But even as the two detachments were in the air on their way to Korea on 28 November, the Chinese had attacked, and Sinanju was not safe. The unit destined for Sinanju was diverted to Pyongyang, much further south, while the detachment commander was flown to Sinanju to assume command of the troops on the ground (the Cho detachment) and to get them to safety farther south. AFSS in Korea operated as Detachment Charlie of 1st RSM until 1951, when the 15th RSM was activated to control all AFSS Korean operations.35 The Cho group made it safely back to Allied lines, and by February of 1951 the front had stabilizedjust south ofSeouI. ASA tactical units dug in for the winter. ASA manual Morse intercept efforts in Korea were having very modest success. Most intercepted material wa~ I "'kI"~viding little of tactical value. But sometl.:-im-e-in"'--F=""eb"'r-u-a-r-y---' reports began to filter to ASA that UN front-line troops were hearing Chinese voice communications. ASAPAC (Advance)s~nt an investigating officer to IX Corps, and he reported that there was a good volume ofspokell Chinese interceptable. ASA already had some Chinese linguists, but wha~ they needed to exploit this type/of nonstereotyped communications was native linguists. A:n~rrangement was made with a former Nationalist Chinese general working for the U.S. in Tokyo to begin hiring former Nationalist officers from Formosa. They were enticed to Korea by the promise of earning GS-6 pay as Department of the Army civilians, and they were to enjoy officer status while in Korea. Competition was keen, and by the summer of 1951, Chinese linguists were flockingtoASA units in Korea. Ib) (1) Ib) (3) -50 USC 403 Ib) (3) -18 USC 798 Ib) (3) -P.L. 86-36 IIl!:f4'Bbl!J 'Ill!: 'fl!:bl!Jr,'f I[I!JYII6bl!J e6MIn'f e6r,'fft6b SYS'fI!JMS if6In'fbY r,6'f ftELI!JA8ABLI!J 'f6 F6ftt1lffl, f4'l!:fi6NcAbB Tep SECRET l::JMBItA 46 - _ . - - - - - - - - - - - - - - - - - - - TOP S!eK!'f t:JMBRA DF operations - an ASA DF unit in the mountains of Korea The linguists were formed into Low-Level Voice Intercept (LLVn teams and were positioned as close to the front lines as possible. The effort was expanded to include Korean LLVI, although that part of the program got off to a slower start because of the difficulty of getting good linguists in a cleared status. Low-level voice quickly became the prime producer of COMINT in Korea, and the demand for LLVI teams overwhelmed ASA's ability to provide enough good linguists. The program expanded from one unit, to seven, to ten, and by the end of the war there were twenty-two LLVI teams, including two teams dedicated to tactical voice intercept.36 In September of 1952 the 25th Infantry Division began picking up Chinese telephone communications from their tacticallandline telephones. This was accidental, of course, and apparently originated from a sound detecting device normally used to indicate the approach of enemy troops. When the unit moved off line, they passed on the technique to the relieving 40th Infantry Division. The 40th improved the equipment but did no analysis. In November, an ASA liaision officer at division headquarters was notified, and ASA proceeded to develop the technique on other sectors, supporting it with LLVI teams IINlq'I'LI!l .. IN TNLI!:Iq'T KI!:YI f6LI!: e6MmT e6N'Tft6L SYSTI!:MS d6mTLY NO I RELEM2\:ftLE Te peftl!:I6N' HA'f16NALS - - - - - - - - 47 TOP SECRET l:JMBKA Ib) (1) Ib) (3) -P.L. 86-36 Ib) (3) -50 USC 403 Ib) (3) -18 USC 798 10' 5!CR!T l:JMBRA consisting of either Korean or Chinese. linguists, depending on which type of unit was on the other side of the line. The Americans had accidentally redisc=overed a techn~que for gathering intelligence which had originally been developed during World War I and which had been a prime producer of tactical information. I I \\\\ These LLVI teams were quite small, consisting only of an ASA officer, a couple of enlisted men for analysis, and two or three native linguists. Their value to front-line commanders, however, far outran their cost, and LLVI was hailed as one of the most important producers oftactical intelligence during the war. White Horse Mountain As the conflict settled down to unremitting trench' warfare, highlights were few, and peace talks gradually replaced warfare in American newspapers. But the front \lines continued to shift imperceptibly as the two sides bludgeoned each other in a series of bloody encounters to take high ground. One ofthose, the battle fot White Horse Mountain, illustrated the use ofCOMINT in a tactical situation. The action was originally tipped off byl. la Chinese Communist military message that was in the hands of the tactical commander before the battle took place. ASA set up a specialI feffort and tactical communications to report information that might bear on the battle. True to the intelligence prediction, the Chinese launched a massive infantry assault on American and ROK troops at White Horse on 6 October and persisted until 15 October. Throughout the battle, LLVI teams kept the American commander informed of the position and activities of Chinese units. In a precursor to Vietnam, the American units were able to call artillery fire on Chinese positions on the basis of the LLVI-provided information.38 The Chinese suffered nearly 10,000 casualties out of some 23,000 committed to the battle.39 AFSS Introduces Tactical Warning Like ASA units, AFSS operations in Korea depended increasingly on intercept of lowlevel voice communications, using this for tactical warning. The concept relied on the Joint Training Directive for Air-Ground Operations published in 1949, which stated that the primary purpose of radio squadrons mobile for tactical support was to collocate with the Tactical Air Control Center (TACC) so that direct tactical warning could be supplied. (This followed World War II COMINT doctrine used effectively by Lieutenant General Kenney at 5th Air Force.) I1Af,fiLtl VI:A 'f:ALtlIff KtlYI16Ltl e6Mn,T e6N''fft6L SI SUMSJ6IMTL I !f6'f ftl!lLEA8ABLI!l 'f6 fi'6ItEfeI4' 14'i%:Tf614'ALS TO' SECRET l:JMBRA 48 TOP 5&CAET YMBftA Because of the lack of linguists, AFSS was slow to set up this service in Korea. However, in the early spring of 1951 AFSS units began intercepting Soviet groundcontrolled intercept (GCI) communications, and this spurred Far East Air Force (FEAF) into requesting AFSS tactical support. Fortunately, AFSS did have some Russian linguists, and eight of them were on their way to Korea in April to form the first linguist team. They originally set up a mobile intercept and processing hut at Pyongtaek in central Korea, and communicated with the TACC by landline. No one in the tactical air operation was cleared for COMINT, so it was disguised using a simple substitution code to identify enemy aircraft and ground checkpoints. Arrangements were made for the TACC controller to pass relevant COMINT, intermixed with radar plots, to fighter pilots. The operation was nicknamed "YOKE," and became highly successful because it significantly expanded the range of control of the TACC and improved the air controllers' ability to warn pilots ofimpending threats. As the front advanced north ofSeoul, so did the air control operations. In June of 1951, the entire air control operation moved forward to a hill four miles northeast of Kimpo Airport near Seoul. But in August hearability deteriorated, and the operation, including the TACC and Security Serice operations, migrated by LST to Pyong-Yong-Do island. Only six miles from enemy lines, "P-Y-Do" (as it was called) was in an ideal location. The site at Kimpo was kept open, and linguists were split between the two sites. Soon AFSS was finding tactical voice communications in Chinese and Korean as well as Russian. Two more voice teams were established for the additional languages. The Korean voice team consisted of the Cho contingent of the Nichols group. The Chinese team set up shop on the campus of Chosen Christian College in Seoul (today, Yansei University). AFSS acquired its Chinese linguists in Korea basically the same way that ASA did - they hired foreign-born linguists. In this case, they did business with one General Hirota, a former chiefof the Japanese army COMINT agency during World War II. Hirota hired twelve Japanese linguists who were fluent in Chinese. With so many languages involved, the tactical support operation was unusually complex. The AFSS facility at Kimpo correlated Chinese early warning voice, Chinese Gel voice, Soviet GCI voice, Chinese air defense Morse and Korean Gel voice. Each input was produced by a separate team, and each team was in a different location for security purposes.40 In September of1951 the P-Y-Do operation was closed down and moved back to Kimpo, and that fall all AFSS operations were consolidated at Chosen Christian. This was the first time that all components of the operation were collocated, which made correlation of activity easier. According to one officer involved in the operation, "the present top-heavy success ofthe F-86s against MIG-15s dates almost from the day of the inception of the new integrated voice-CW-YOKE service." 41 WA :-T,gI"i: 171 A 'l'AJ"J!]~T'l' I(;J!]YIISbJ!] GSMIPT'f 8S'Pi'fRSh SYS'fB!'If:3 dSHf'fI::Y 49 lQP SECR!'f tJlYIBRA lap SECRET l:JMBRA In early 1952 much of the Gel traffic that AFSS had been intercepting began to dry up, and AFSS became convinced that it had gone to VHF. Moreover, about that time the Chinese stopped tracking Communist aircraft, and they tracked only "hostiles." These twin changes spelled potential disaster for AFSS tactical operations. From a practical standpoint, the lack of tracking would force AFSS to rely almost entirely on intercepting Gel communications. But since these communications were disappearing, probably to VHF, that source of information was also drying up. The changes also generated a security problem, since the positions' of Communist aircraft had been disguised as radar plots when being passed to the TACC. If there were no more radar position reports, disguise ofthe origin ofthe information would be much more difficult. Delmar Lang on Cho·Do Island in 1952 I1ltPfBr"S 'lIlt 'FltUjPf'F KEiYIISr"S GSMlN'F GS~l'FRSr"8Y8'FSMS d8lN'FbY lOP SElkE I UMBR1t 50 Tep SECRET l:JMBIbIc These developments roughly coincided with the arrival of the first batch of schooltrained American Chinese linguists, headed by Lieutenant Delmar "Del" Lang, in mid- 1952. At the time the unit was located in Seoul, where VHF intercept was hardly possible, while the TACC had moved to Cho-Do Island, near the North Korean harbor of Wonsan. Information had to be relayed from the AFSS unit to Kimpo and from Kimpo to Cho-Do. Lang moved the operation to Cho-Do Island and collocated it with the TACC. Tests on Cho-Do in August of 1952 confirmed that both the Soviets and Chinese were now using VHF for their GCI control activities. To solve the security problems and to make sure that the TACC controller got the best possible support, Lang positioned an AFSS linguist in the TACC in March of 1953, sitting next to the controller. The linguist had a field phone on his desk, the other end of which was attached to the output of a receiver at the Security Service intercept unit threefourths of a mile away. In an era when no one knew much about TEMPEST (see chapter 5), such a wireline was regarded as secure simply because it was a landline.42 Combined with improved hearability, the new lash-up at Cho-Do Island provided the best support that AFSS mustered during the entire war. In one day, which Lang described as the "great Korean turkey shoot," American F-86s downed fifteen MIGs without a loss, even though none of the MIGs was ever seen on radar. The information came, of course, from the COMINT operation at Cho-Do. A visiting ASA colonel commented that "it was just like shooting ducks in a rain barrel." It was a model for tactical COMINT operations and was resurrected by the same Del Lang years later in Vietnam. (See chapter 12.)43 The Navy Naval cryptology was a bit player in Korea. The DPRK had no blue-water navy, and it was so weak that the Inchon invasion went unopposed from the naval standpoint. The naval COMINT unit in the region wasl I Butl !was not concerned with the small collection ofDPRK coastal patrol craft. The organization concentrated instead almost entirely on the Soviet navy in the Pacific, to determine what moves, if any, the Soviets would make toward the U.S. presence on the Korean peninsula. The unit was/housed in cramped quarters in a former Japanese artillery training school, entirely too small and inadequate for the found an old Japanese ammunitiollstorage building about ten mile~fro Rehabilitation began in 1951,and in November 1~f)21 II'Oove~to here it remained for many years. Mostofth~N~gsupportto the war effort came from its afloat detachments. O~igi~~~iIlgO:l.ltofHawaii,detachments were placed aboard 7th Fleet vessels beginning in A:llgust 1951, and at the end ofthe war, 7th Fleet had three such units.44 Ib) (1) Ib) (3) -50 USC 403 Ib) (3) -P.L. 86-36 Wi\MQbK VIA 'I'eAbKN'f IE:KYIlSbFJ eSMm'f eeU'fRSb S'IS'fI!JMSJ6UofTL i US'f RFJJ::Ji!lA:SkSbl!J TO P'6ftI!:IGN NATIONALS 51 'FOP SECRH tlMBR-A TOP SECRETtJMBItA The AFSA Factor On the home front, AFSA provided significant help to battlefield commanders. AFSA's quick work I lin time to turn the tide at Taegu appeared to portend the same kind of COMINT effectiveness that the U.S. had enjoyed during World War II. But it was not to be. \ In November 1950, with Chinese Communist troops flooding into North Korea, AFSA turned its attention to Chinese communications. I In 1952 the painfully slow progress on traffic analysis of Chinese army nets finally began to bear fruit. There were indications through traffic analysis\that the 46th Army was moving northward. The army eventually arrived in Manchuria and crossed the border into Korea. As it did so, AFSA began exploiting People's Volunteer Army (PVA) nets from a traffic analytic standpoint, and it achieved a level of competence on PVA nets that allowed extremely accurate order of battle determinations, unavailable through any other intelligence source. Through traffic analysis AFSA noted the build-.up of PYA units on the eastern front, and this allowed 8th Army to reinforce its right side prior to a major PVAassaulton 15 July 1953.48 Relations with ROK COMSEC and COMINT COMSEC assistance to ROK forces began almost as early as COMINT collaboration. In September 1950 ASA was asked to furnish low-level cryptographic assistance \for use by the ROK army. After conferring with AFSA, ASA shipped some strip ciphers and Playf~Hr squares. It was soon found, however, that these very time-intensive systems would not be fast enough, and in 1953 ASA provided the first electromechanical cipher equipment, the BACCHUS system. Later in the year ASA also released the DIANA one-time-pad system.49 Cryptologic cooperation with the ROK COMINT organizations continued throughOut the war. USAFSS continued its relationship with the Cho group, while ASA continu.ed to do business with the Kim group. In November 1951 ASAPAC proposed the consolidation Ib) (1) Ib) (3) -50 USC 403 Ib) (3) -18 USC 798 Ib) (3) -P.L. 86-36 IIANBbl!J VIA 'fJ.bl!J~''fKB'III8bE 08MIN'f 08N'fR8b BYS'fBfIISlf8lIffb'I NO I RELEAS"AftLtl 'f6 ~6ftf)I8NfM'fI8~WlbS T9P SECRET UMBItA 52 Ib) (1) Ib) (3) -P.L. 86-36 Ib) (3) -50 USC 403 Ib) (3) -18 USC 798 'fep 5E!eRE!'f t:JMBItA of the two efforts, but AFSS firmly rejected the overture. This wllsprobably based on Air Force fear that ASA would dominate the relationship and get back int() the business of copying North Korean air targets, but this may also have been based on thevj;lry realistic appraisal that the animosity between Kim and Cho.was unbridgeable.50 The situation continued unchanged, and late the next year an official for the newly created NSA/ 1"51L...- ..... By charter (NSCID 5), CIA had control of all foreign intelligence relationships. But the "battlefield marriage" between the American and South Korean COMINT organizations represented a significant exception to the general rule. Korea wasJCS turf, and military commanders were cool to CIA participation in their arena. II"[~IBIoB '1IA flrbBNf IEBYH8LB 88MInf 88N'fft8b S-YS'fBMS 68In'fLY I(()Y ttELE:AMf'LrJ 'f8l"8ftrJI8N NA:'fI8NAIoS 53 1=9P SECRET t:JM8RA Tap SECRET tJMBItA Korea - An Assessment Ib) (1) Ib) (3) -50 USC 403 Ib) (3) -18 USC 798 (3) -P.L. 86-36 """'\\\\\\\ The Korean War occurred during a period ofstruggle in the cryptologic community. It began a year after the formation of AFSA and concluded after the AFSA ship had been finally scuttled in favor of a new vessel, the National Security Agency. The demands of war highlighted the fissures in the structure, and those fissures in turn made prosecution ofthe war more difficult. AFSA wrestled with the SCAs over control ofintercept pOsitions and targets throughout its existence, and many of those battles were related to the war effort. The Brownell Committee was convened in part because of complaints by organizations outside the Department of Defense over degraded cryptologic support resulting from the war. The committee stressed in its final report that the cryptologic community had been shown deficient in its effort during the war. NSA replaced AFSA partly because ofwhat was happening (or not happening) in Korea. But after forty years the picture does not look quite so bleak. Actually, AFSA and the SCAs provided good support to the war effort. Although AFSA (along with everyone else) was looking the other way when the war started, it did a remarkable about-face, and within a month it was producing large volumes of decrypted information from North Korean communications. Its accomplishments during the battle for the Pusan perimElter, I ~~~ information to support tactical commanders, were considerable and important. The reporting program, although hampered by restrictions on AFSA's production of "intelligence" as opposed to "intelligence information," was farsighted and effective. AFSA, almost alone among intelligence agencies, foresaw the Chinese intervention. The development of Chinese and Korean order of battle owed much to AFSA's high-powered traffic analytic effort. After a slow start occasioned by lack of mobility, tactical resources, linguists, and working aids, ASA and USAFSS put together highly credible battlefield COMINT organizations. ASA's LLVI program produced more valuable information for ground commanders than any other source. AFSS put together a system for warning fighter pilots which was partly responsible for the much-ballyhooed kill ratio in that war. , ltAflBLE "fIA 'f:ALEfI'f' I£EYIISLE eSMIn'f' eSn'f'RSL SYS'f'EMSifSIN'f'LY HQT Kl1bl1ASABbFJ 'fe FeRFJISfi fiA'fI6nkLS i OP 5I!(R!T t:JMBAtr 54 (1) Ib) (3) -P.L. 86-36 Ib) (3) -50 USC 403 Ib) (3) -18 USC 798 TOP 'EeRET liMBRA AFSA's quick start was not sustained. Beginning in July of 1951, the North Koreans began a total changeover of their communications procedures In the first month of the war, AFSA read more than one third of all North Korean her messa es received, and by December AFSA was reading more than 90 percent. The new North Korean secun y '----------:-;--:-;---;----:---.,:-:--:-;-~---:--:"":;-;--:--~ measures were evidently inspired by the Soviet Union, whose communications had in 1948 undergone a similar transformation in the face of possible American and British exploitation efforts. (See chapter 4.) It was accompanied by a decline in North Korean radio messages incident to the beginnings of static trench warfare roughly at the 38th Parallel, which gave the enemy a chance to divert radio communications to landline. W• )+QU: VIA 'I:\\bIHf'F Kli.iYIISU OSMlffT OSNTltSI:J SYSTEMS ttSn"YI:J f ns'f RBI:JBASttBbE 'fa f'aftEH'IU I,,1\"I6I{1\I:JS 55 lap SECRET liM81tA Ib) (1) Ib) (3) -50 USC 403 Ib) (3) -18 USC 798 Ib) (3) -P.L. 86-36 TOfl SeERET l:IMBAA Security was a problem in Korea, as it has been during all wars. Occasional press releases exposed COMINT support to battlefield commanders. The release of information about AFSS exploitation of GCI communications became so serious.that in October 1951 Detachment 3 of 1st RSM took the extraordinary step of suspending opera.tions for a few days until they got the attention of key officers in 5th Air Force.56 The employment of tactical GCI voice and tracking information in the air war c.aused AFSS to devise new measures to cover the information, and it set a precedent for use of similar information during the war in Vietnam. When NSA was created in November 1952, immedia.te steps were taken to sort out the effort in Korea. NSA's recommendations amounted/to a classic "lessons learned" about war. Most pressing was a program which would allow the use of indigenous personnel with native language capability. Almost as urgent was the need to sort out the tangled relationships with the various ROK COMINT efforts. It would also be necessary to increase NSA representation in the field and to expand existing field offices with technical experts assisting the SCAs. Finally there was a call to develop new special identification techniques that would allow NSA and.the SCAs to track target transmitters I I INSA sponsored these themes for years, unt-il-t~h-ey-"" became tantamount to COMINT doctrine on warfighting. One beneficial effect of the Korean conflict was to begin a rapid rise in cryptologic resources. In July 1950 USCIB recommended to the National Security Council that COMINT receive a hiring jolt. The NSC approved this on 27 July in a meeting attended by the president himself.58 Korea was America's first stalemated war, and recriminations resounded for years later. But even an acerbic CIA critic of the cryptologic community had to admit that "COMINT remained the principal source ofintelligence for threat until 27 July 1953, when the armistice was signed at Panmunjom."59 Notes 1. Rowlett interview, OH 14-81. 2. Sinkov interview, OH 2-79; oral history interview with Herbert L. Conley, 5 March 1984, by Robert D. Farley, NSAOH 1-84. 3. See both Burns. Origins, and Howe, "Narrative." 4.William L. O'Neill, AmericanHigh: The Years ofConfidence. 1945-1960 (New York: Free Press, 1988.) 5. See Burns. Origins, 65. 6. CCH Series V.F.5.1. IhldiBbl!J VIA 'f1tLl!Jfff f(I!JYII6LI!J e6MIN'f e6fi'fR:6b S'{S'fI!JMS J6IWft:: i H'6'f KI!JLI!JASABLI!J 'f6 f'6R:I!JI8fi ftA'fI8fiALS Tap SECRET l:IMBRA 56 - - - - - - _ . _ - - - - - - - '(b) (3) -P.L. 86-36 Tap SECRET l::JMBR-A 7. Howe, "Narrative." 8. Burns, Origins. 70-71. 9. Howe, "Narrative." 10. Burns. Origins, 75-77; "Report to the Secretary ofState and the Secretary ofDefense by a Special Committee Appointed Pursuant to Letter of28 December 1951 [Brownell Report]," in CCH Series V.F.7.13; NSA Archives. ACC 26350. CBSK 32; "Analysis ofAFSS Effortin the Korean Action," unpublished draft. USAFSS. n.d.• in CCH Series V.M.4.1.; A Reference Guide tp Selected Historical Documents Relating to the National Security Agency/Centra] Secrwty Service, 1931-1985, Source Opcuments in Cryptologic History, V. I (Ft. Meade: NSA. 1986).36,38. 11. Wenger comments on Howe dratt history in CCH Series V.A.13. 12. Burns. Origins. 59-96. 13. Burns, Origins. 89;1 '''Consumer. Liaison Units, 1949-1957," in NSAlCSS Archives ACC 10684. CBRI 52. '--_----" .....J 14. Howe, "Narrative"; "JCECMemo for Information No., 1, Charters," in CCH Series V.G.2.; Brownell Report. 15. Brownell Report. 16. Ibid. 17. Ibid. See also Howe "Narrative"; Burns, 107-108. 18. An excellent account of the diplomatic background to the invasion of Korea can be found in Clay Blair. TIu! Forgottell War: America/ill Korea. 1950-1963 (New York: rimes Books. 1987); and Joseph Goulden. Korea: TIu! UIltold Story oftlu! War (New York: Times Books. 1982). 19. Seel I"The U.S. COMINT Effort during theKorean Conflict - June 1950-August 1953." pub. on 6 Jan. 1954. an unpublished manuscript in CCH collection. series V.M.l.l. See also Howe, "COMINT Production ... "and The 'Brownell Committee Report'. 13 June 1952, in CCH series VI.C.1.3. 20. [AFSA 235] no title [report on significant activity conne~ted with the entry of Chinese Communists into the Korean conflict], 25 March 1952. in CCH Series V.M.7.1.; anOThe U.S. COMINT Effort...." 21. William L. O'Neill, AmericallHigh: TIu! Years ofConfidence. 1945-1960. 22. Howe, "COMINT Production ..."; Dick Scobey (NSA), draft study of ROK SIGINT Effort. no date, in CCH series V.M.6.1.; Assistant Chief of Staff. G2. "COMINT Operations of the Army Security Agency during the Korean Conflict. June 1950-DeceII:lber 1953," in CCH Series V.M.2.1. 23. Blair, Forgotten War. Ch. 2-4. 24. Interview Youn P. Kim, 22 February 1982, by Robert D. Farley. OH 2-82, NSA. 25. Hq USAFSS, "Analysis of AFSS Effort in the Korean Action," unpublished draft manuscript in CCH series V.M.4.1. 26. Summaries of Project WILLY can be found in the foll~wing sources: Hqs USAFSS. "Analysis of AFSS Effort ...... Dick Scobey. "Draft Study ofROK SIGINT Effort"; ["Hop" Harriger], "A Historical Study of the Air Force Security Service and Korea, June 1950-0ctober 1952," on file at Hqs AlA in San Antonio. 27. Manuscript entitled "SIGINT in the Defense of the Pusan Perimeter: Korea. 1950," (SC) in CCH series V.M.1.lO. See also Clay Blair, The Forgotten War, 240. HANBL:I!J VIA 'fALeI4'T K:I!J tH6Le e6ItflI4'T e6I4'Tft6L S 1S't':l!JMS tf6m'fLY N6T ft:I!JL:l!JAS1tBL:I!J 'f6 F6ft:I!J16!4' !4'A'fI6!4'ALS 57 1'6' 5ECRET tlMBRA Ib) (3) -P.L. 86-36 lOP SIiEAET l::JMBRA 28. Guy Vanderpool, "COMINT and the PRC Intervention in the Korean War," paper available in CCH. 29. Roy E. Appleman, Disaster in Korea: The Chinese Confront MacArthur (College Station, Texas: Texas A and MPress, 1989). 30. Blair, Forgotten War, 350. 31. [Drake, Robert, and others] "The COMINT Role in the Korean War," unpublished manuscript in CCH series V.M.l.9. See also Howe, "COMINT Production in the Korean War ..."; otal history interview Milton Zaslow,14 May 1993 by Charles Baker and Guy Vanderpool, NSA OH 17-93; oral history interview Robert Drake, 5 December 1985 by Robert D. Farley and Tom Johnson, NSA OH 18-85; oral history interview Samuel S. K. Hong, 9 December 1986 by Robert D. Farley, NSA OH 40-86. 32. Blair, Forgotten War,375-78. 33. Zaslow interview; Drake interview. 34. Department of the Army G2, "Indications of Chinese Communist Intentions to Intervene in Korea," 7 May 1954, in CCH series V.M.7.4.; oral history interview Admiral Arleigh Burke, 9 December 1981, by Robert D. Farley and Henry F. Schorreck, NSA OH 13-81. 35. "Analysis ofAFSS Effort ..."; George Howe, "COMINT Production in the Korean War...." 36. Assistant ChiefofStaff, G-2, "COMINT Operations ," contains the best summary QfLLVI operations. 37. See Assistant ChiefofStaff, "COMINT Operations ," 56-57. 38. Assistant ChiefofStaff, G-2, "COMINT Operations " See also oral history intervie~L.- 124 April 1982 by Robert Farley, NSA Oral History 9-82,122. 39. For a description ofthe action, see Walter G. Hermes, Truce Tent and Fighting Front, United States Army in the Korean War (Washington, D.C.: Office ofthe ChiefofMilitary History, United States Army, 1966), 303-08. 40. Summaries ofAFSS tactical operations can be found in the following: USAFSS, "Analysis ofAFSS Effort in the Korean Action," unpublished draft in CCH Series V.M.2.1.; NSA, "Review of U.S. Cryptologic Effort, 1952- 54," in CCH series VI EE.1.3.; and [Hop Harrigerl "A Historical Study...." The latter document contains the fullest explanation ofthe Yoke operation. 41. [Hop Harrigerl "A Historical Study... ," 72. 42. The new operation is described in USAFSS, "Analysis of AFSS Effort ..."; "Historical Data Report for the 6920 SG,I January 1953,"; interview with Delmar Lang [undatedl, in CCH Series VI, AFSS section; and Major Chancel T. French, "Deadly Advantage: Signals Intelligence in Combat, V. II," Air University Research Report # AU-ARI-84-1,1984. 43. French, "Deadly Advantage ..."; oral history interview Col (USA, Ret.) Russell H. Horton, 14 March 1982 by Robert D. Farley, NSA Oral History 6-82. 44. U.s. Naval Security Group, "U.S. Naval Communications Supplementary Activities in the Korean Conflict, June 1950-August 1953," in CCH Series V.H.3.L 45. Richard Chun, unpublished manuscript in CCH Series V.M.LIL 46. Drake and others, "The COMINT Role in the Korean War." 47. Assistant ChiefofStaff, G-2, "COMINT Operations...." 48. [Drake and others] "The COMINT RoJe in the Korean War." IIANBbB VIA 'fA:hEl1ff I(ElYII8hEl e8MUff e8r,'fR8h Sl{S'ftJMS d'8In'ff::H NO1 REL~'AS1iBLI!l 'f8 F8RBISN nA'I'IQ~A bS TQP SECRET l:JMBRA 58 Ib) (1) Ib) (3) OGA +9P SEeR!!!T UMBRA 49. "U.S. Cryptographic Assistance and COMINT Collaboration with the ROK," 24 February 1955, in CCH Series V.M.6.5. 50. File ofmemos related to the history ofAFSAJNSA communications center, in CCH Series VI.H.1.2. 51. NSA, "Study ofthe COMINT Situation in Korea," undated memo (probably December 1952) in CCH Series V.M.l.14. 52, "Agreement on COMINT activities between the U.S. and the Republic of Korea, 1956," in CCH Series V.M.6.3. 53. Brownell Committee Report, G-I-G-2; see also Mary E. Holub, Joyce M. Homs and SSgt Kay B. Grice, "A Chronology ofSignificant Events in the History of Electronic Security Command, 1948-1988," 1 March 1990, in CCH Series X.J.6. 54. CCH Series VI.A.1.3. 55. "Study ofthe COMINT Situation in Korea." 56. "Analysis ofAFSS Effort in the Korean Action." 57. "Study ofthe COMINT Situation in Korea." 58. USCIB memorandum, 20 July 1950, and NSC memorandum dated 27 July, in Harry S. Truman Library, Independence, Missouri (contained in CCH Series XVI). 59~ ~The History ofSIGINT in the Central Intelligence Agency, 1947-1970," October 1971, V. 1.,86 in CIA history collection,Ames Building, Rosslyn, Virginia. HANDLE VIA TALENT KEYHOLE CO ROL SYSTEMSJOINTLY NOTR ETOFOREIGNNATIONALS 59 Ib) (1) Ib) (3) -50 USC 403 Ib) (3) -18 USC 798 Ib) (3) -P.L. 86-36 TOP 5"RET liMBI'\A ChapterS Cryptology under New Management There is something about cryptologic work that gets into the hide.... Ralph Canine, 1968 NSA began life under a pall. The Brownell Committee had declared its predecessor to have been a failure. Outside the cryptologic community the1 was a common fee~ng that COMINT was broken and in serious need ofrepair. According t ho was appointed by Allen Dulles to ride herd on the cryptologic effort, The early 1950s were the dark ages for communications intelligence. Intelligence officers who had been accustomed to providing information not only on the capabilities but also on the intentions of the enemy during World War II were reduced to providing the government with estimates based on frail fragments of information rather than factual foreknowledge. The creation of NSA was an attempt to address the problems of cryptology as the Brownell Committee saw them. (As we saw in the section on Korea, that perception was not 100 percent accurate.) That is, it attempted to institute a firm control mechanism that would unify the system and create an organization which was, in and of itself, responsible for getting the job done. No longer would consumers have to go to four different organizations to get answers or to fix blame for the lack of answers. It did not give the organization resources, improve its personnel situation, or give it adequate working space. Ib) (1) Ib) (3) OGA When NSA began life, it simply inherited its resources from its predecessor. It got the AFSA billets and the people in them, the AFSA spaces at Arlington Hall, and the AFSA rooms at the Naval Security Station. And it inherited an idea, that unification worked better than division. The difficulty was in trying to implement the solutions that the Truman Memorandum imposed. AFSA, despite its failings, had been a step in the right direction. NSA now had to take the next step. To the AFSA population, the name change must have seemed more for appearance than for any practical value. There was no immediate change in their condition. They stayed where they were - if they were COMINTers, they remained at Arlington Hall, and if they were COMSEcers, they stayed at Nebraska Avenue. Lieutenant General Canine, who had replaced Admiral Stone as AFSA director, stayed on as director of NSA. When Canine first gathered the NSA work force together on 25 November 1952, he alluded to the conflicts which had preceded the establishment of NSA, but they must have seemed remote to those who listened. It looked like business as usual. HANDLE VIA TALENT KEYHOLE LSYSTEMSJOINTLY ASABLE TO FOREIGN NATIONALS 61 TOP SECRET tJlViBRA - Lieutenant General Ralph J. Canine went to bat for the new organization at a time when its existence was challenged and its longevity was far from certain. Canine and the New Organization But it was not to be business as usual, largely because of the personality of the first director. Lieutenant General Ralph Canine, who dominated early NSA policies and stamped his character on the Agency, had been a line Army officer with no intelligence experience until he became deputy assistant chief of staff for army intelligence in 1949. Prior to that he had been an artillery officer, with wide experience in combat (both world wars, serving under Patton in World War II) as well as logistics. Although he brought no technical education to cryptology, he exerted his influence through a hands-on management style. He was forceful and determined and tenaciously enforced the Brownell recommendations on the reluctant SCAs. His whimsical personality produced legions of "Canine stories," which simply embellished liis reputation as a maverick. Collins proclaimed him a "fortunate choice," and said that "he ... raised the National Security Agency from a second-rate to a first-rate organization."z Canine was no diplomat, I1Al'lBbE ¥h\ Y\'.bEl'l'f KEYH9I:JE 69MlN'f eS?i'fRSb SYSTEMS ofSUffbY NOT REI,EASABI F TQ li'QR:BI8fH,ATIONALS 62 and he might have failed had he come along ten years later. In 1952, however, he was the right man for the job. One of the first things Canine did was to get rid of the triumvirate of service deputies who, under AFSA, had represented their own service interests rather than the interests of the central organization. He replaced them with a single vice-director, and named Joseph Wenger to fill the position. But Wenger was probably not very happy as the vice-director. By all contemporary accounts, Canine served as his own vice-director. He tended to make all key decisions himself. He had no patience with long vertical lines of control, and when he wanted an answer, he went directly to the person involved. He relied on his staff to keep others in the chain of command informed of his comings and goings but did not feel bound, himself, to use the chain. The system smacked of paternalism, and one of Canine's subordinates once said, "Whenever I see him nowadays, I expect him to pat me on the head."3 Canine organized NSA rather like AFSA had been structured, with Production, COMSEC. and R&D being the major divisions. But he broke Administration into its component pieces (security, personnel, training, logistics, and plans and policy) and placed them on his "special staff," a classically army way of doing things. The office designation system was a trigraph, NSA followed by a dinome: for instance, NSA-02 was the Office of Operations. In February 1953 Canine changed Operations to Production, or NSA-06. Production was structured much like a factory, in which the parts of the cryptologic process were organized functionally rather than geographically. The major divisions within Production were Collection (NSA-60), Analysis (NSA-70), Machine Processing (NSA-80), and Exploitation (NSA-90). Although NSA has since changed over to a more geographical approach, the original organization more closely corresponded to how cryptologists viewed their profession at the time - as part of a complex process suitable primarily for highly skilled factory technicians. What made cryptology different from other intelligence disciplines was both the intricate technical challenge and the assembly-line processing system. It also represented NSA's way of conceptualizing the process of intelligence - as underlying data revealed through mathematical attack rather than as cognitive insight arrived at through inspiration.' The Early Work Force The Korean War had ushered in a period of explosive growth in the cryptologic population. This was followed by a long period offairly steady personnel growth, as Table 1 shows. I11d',BbE VIA ~ltbEPPf I(EYIISbFJ eSfllHff eeU'fft8J:: 8'1 8'f~M8 J8II('fL i NOT REI,E A SAHI F TQ i'QREI6N I4"*'fIONALS 63 Ib) (1) Ib) (3) -50 USC 403 Ib) (3) -18 USC 798 (3) -P.L. 86-36 Table 1 Cryptologic Population, 1949-19605 Year AFSA NSA Totals (includes SCAs) Dec 1949 4,139 Dec 1952 8,760 Nov 1956 10,380 Nov 1960 12,120 10,745 33,010 50,550 72,560 The work force in 1952 was double what it had been under AFSA, but it Wl:l.:~··:>~Hl smaller than either ASA or USAFSS and larger only than NSG. 6 The Hoover Commission, which was probably L...;-,thr-e-m-o-s~t-e-x':"'"te-n-s""i-ve---'i-n-v-e-st:-:i-g-a':"'ti:-o-n-of'l:"""':'t':"'"h...Je federal bureaucracy ever, estimated that cryptologic costs amounted to about halfa billion dollars.7 In the early days, the work force was about one-third military and two-thirds civilian. A snapshot ofNSA's work force in 1956 (Table 2) showed most of the population working in Production. Pay tables were not quite as generous in those days, as Table 3 clearly shows. A grade 5 employee (the most numerous group of NSA employees) started out making $3,410, which smacks of impoverishment. But with houses costing below $10,000, and frequently below $5,000, employees may have been just as well offin real terms then. I1AP'tQ:b~ YIA 'I'A:b~ilf'F l\;EJYII8rsFJ 88MINT ceuTHers SYMFJMS den,'fLY 64 Ib) (1) Ib) (3) -P.L. 86-36 Table 2 NSA's Work Force by Organization, 1956 Element Production R&D COMSEC Communications Training Directorate, admin, overseas Totals Table 3 Pay grade allocations and salary (basic level) 1952 and 19938 Grade Salary 1952 1993 Grade Alloc 1952 1993 1 $2,500 $11,903 0.2%) 0%) 2 2,750 13,382. 0.7) 0.07) 3 2,950 14,603 6.5) 0.5) 4 3,175 16,393 13) 0.7) 5 3,410 18,340 26) 1.9) 6 3,795 20,443 7) 1.6) 7 4,205 22,717 18) 4.9) 8 4,620 25,259 1.5) 1) 9 5,060 27,789 12) 6.8) 10 5,500 30,603 0.5) 0.2) 11 5,940 33,623 7) 12.1) 12 7,040 40,298 4) 22.1) 13 8,360 47,920 2) 26) 14 9,600 56,627 0.8) 11.5) 15 10,800 66,609 0.6) 6) 16 12,000 .02) 17 13,000 .02) 2) 18 14,800 .02) I1MiBbB Ifb\ 'f1\bBti'f IEBYIIebB eeflmi'f eeU'i'R8b SYS'i'BflfS of8m'fbY NOT REI E AS' ~u 'fQ P'8R1!lI8U }(1d'IOM1 BYS'fl!lMB d8HffUI WQ'I' iWilbl!iASltBhEl 'f8 F'SRl!lIBN' N'A'i'I8R,ltLS 81 By 1956, however, Canine apparently understood enough of Navy COMINT organization to object to its entire philosophy. He took aim at the subordination of NSG detachment commanders to naval communications: "This is an unsatisfactory arrangement; there is always a conflict ofbasic interests in the direction of the units. The. superior officers in the chain ofcommand ... are primarily concerned with general service communications; they are generally inexperienced in COMINT functions...." He related the submersion ofNavy COMINT to Navy communications with SCA position totals, in that from 1953 to 1956 NSG grew by only 7 percent, while ASA expanded by 380 percent and AFSS by 410 percent. This, he contended, resulted from deficient naval GOMIN't organization.50 In contrast to NSG, AFSS growth was breathtaking. From a tiny cadre of 156 people in 1948, AFSS grew to 23,128 people by the end of 1960. The command had over 1,000 positions, a budget of more than $26 million, and it had surged ahead of both ASA and NSG on all counts in only twelve years.51 NSA's relations with AFSS, however, were the worst of the three. Although COMSEC relations were smooth, COMINT was not. Under the hollow gaze of AFSA, AFSS had virtually seceded from the COMINT community, carrying its entire field site list with it.• It had called the field sitesl Iso as to exempt them from AFSA tasking. (Major General John Morrison [USAF], a for-mer NSA assistant director for production, once said thatl kithvery isolated exceptions, were about a~ ~sthe Eifel Tower.")52 Canine's dicta onoperational and technical control were intended largely to corral the errant AFSS resources, This was effective but did not make AFSS very happy. The biggest row of the decadc.was over the Air Force SpedalCommunications Center (AFSCC). Officially created in July 1953 as the 6901st Special Communications Center, AFSCC was intended as a third echelonprQcessing center to satisfy Air Force desires for an indigenous Air Force COMINT center. The organization picked up such miscellaneous functions as the ssa system and the USAFSS training school but wasintended all along as an analytical center and began functioning as one froJ:ll its very first day of existence. Canine had said "No" to the Air Force plan but lost the battle;Jn January 19.54 he gave up and, under the aegis of the decentralization plan, AFSCCacquired the mission of processing and reporting on the=: :==J'J.'() this Il,ucleus""as added, over the years, virtually the e~eI lasWEll1~s,begillIlitlgin1961, the I rRelatiom)hipscoIltin.tle~t~~et~ri~r~te,a.Il~~yt~e end ;., Ib) (1) Ib) (3) -50 USC 403 Ib) (3) -18 USC 798 Ib) (3) -P.L. 86-36 IIkHtiLJ!J 'ilk 'fkLJ!J!fi' UJ!J'it16LJ!J e6MtH'f e6H'fft6h SY'S'fEMSd6IIffLY NOT REI E A S A BT i: TO :r:QAEI€lN N-A'fl8!fJd:.S 82 ofhis term as DIRNSA, it was rumored that Canine was barely on speaking terms with the AFSS commander, Major General Hunt Bassett.54 The SCAs Create Second Echelons The decentralization plan spawned a second concept, frequently wound up..... .....1 controlling related intercept positions at smaller units. The arrangement amounted to a de facto layering system in which large units controlled operations at smaller units, and in some cases the smaller units were officially subordinated to the larger ones. The intermediate tier came to be known as "second-echelon," while NSA (and in the Air Force, AFSCC) operations were called "third echelon." (1) Ib) (3) -50 USC 403 Ib) (3) -18 USC 798 Ib) (3) -P.L. 86-36 nA~JQbK'lIA 'fltbKPf'f f(BYH8bB eSMfH'f eSn'fftSb ~YS'fBM~JSfH'fLY NOT BE! E ASUU"K 'f8 FSftB[f}N' NAIIONALS 83 l Ib) (1) Ib) (3) -50 USC 403 Ib) (3) -18 USC 798 Ib) (3) -P.L. 86-36 All three services created administrative units to supervise theater intercept sites, and to serve a liaison function with the supported commander(s). However, they all showed a disinclination to combine operational and administrative functions in the same organization, believing those to be separate tasks.58 Watching the Watchers DIRNSA's supervisor was not really the secretary ofdefense, despite what the Truman Memorandum said. In 1953 the secretary of defense assigned that job to General Graves B. Erskine, a Marine Corps four-star who was already assigned to his staff as head of the Office of Special Operations. Erskine monitored the CIA budget, which was hidden in the DoD budget, and after July 1953 he also monitored NSA. His deputy, Air Force colonel Edward Lansdale, later became famous as the author ofcovert actions projects in both the Philippines and Vietnam. The monitoring that Erskine did was rather loose. He always retained professional cryptologists on his staff to work the details of cryptologic money, and under such a 85 leap SECRET tJMBkA system, oversight was not detailed. Occasionally a big-ticket item would come up, like LIGHTNING (see p. 204), and Erskine's office would become involved. But Congress had not yet instituted an effective review of the intelligence agencies (and did not until the mid- 1970s), and CIA did not yet have the authority to ride herd on the finances of the DoD intelligence organizations. So by the standards of later days, no one was really paying much attention to the intricacies ofNSA's money.59 NSA AND CIA - THE EARLY YEARS Will you please have the proper instructions issued discontinuing the cryptanalytical units in the offices ofthe Director ofCensorship, the Federal Communications Commission and the Strategic Services. Ifyou are aware ofany other agencies having services ofthis character. will you please have them discontinued also. Franklin D. Roosevelt, Memorandum for Director ofBudget, 8 July 1942 The origins of CIA were rooted in World War II. Roosevelt, under the pressure of wartime exigency, created an espionage agency in 1942, called the Office of Strategic Services (aSSn, under New York lawyer and World War I battlefield hero (winner of the Medal of Honor in France) William Donovan. Donovan's agency both collected and produced intelligence and mounted covert operations around the world. It was a mission that CIA was to inherit several years later. NSA's difficulties with CIA stemmed from decisions made in the 1940s, almost all of them bad. JCS, which owned most of America's intelligence assets, opposed OSS from the beginning and did everything in its power to deny to ass the resources to do its job. The Joint Chiefs failed to keep OSS out ofthe HUMINT business, but in one area they succeeded almost totally: COMINT was denied. Roosevelt's order (above) resulted in the closure of a small OSS COMINT organization. Even worse, it was used by the JCS to deny to OSS access to ULTRA. Thus ass reporting was crippled from the beginning. It had access to agent reports, photoreconaissance, POW and defector reports - everything, in short, but the most useful and reliable information. If World War II was, as has been claimed, a COMINT war, OSS remained on the intelligence sidelines.60 And it rankled. OSS seniors who later served in the higher ranks of CIA never accepted the JCS policy. The British intelligence services, which dealt closely with OSS, were appalled. Their own intelligence community was unified, and HUMINT was routinely integrated with COMINT in highly specialized offices, in order to reap full value from both. (For instance, Ian Fleming, a British naval officer and later author of some note, was responsible for the integration of Bletchley-produced ULTRA with the Navy's HUMINT and HANDLE YIA TAJ ENT KFYUObK C9MHi'f €8fffltfJLSiSIEMSJOINTLY NOT RELEASABI F TO li'QRBI8N' I(iiiIONALS leap SECRET tlMSR'A 86 Ib) (1) Ib) (3) OGA T~ special operations.) JCS had used security as justification for the denial of ULTRA to OSS, but the British were at least as security conscious as the Americans, and they seemed able to get COMINT of the highest sensitivity to those in the HUMINT business who needed it. The outright denial ofULTRA to OSSjust did not make sense.61 Truman discontinued OSS immediately after the end of the war, partly to rid himself of Donovan, who was not in favor with the president. But within six months Truman once again had himself an intelligence organization, called the Central Intelligence Group. CIG was bedeviled by the same problems that submerged AFSA - lack of its own budget and personnel resources (people were loaned in from other intelligence organizations), absence of a congressional mandate, and lack of firm direction from the top. But the idea was the same as that ofAFSA - to establish central control ofU.S. intelligence operations. When CIA was created in 1947, succeeding CIG, it got its congressional mandate, its budget, and its own personnel. It still lacked firm leadership, but that was remedied in 1950 with the appointment of General Walter Bedell Smith as DCI. Smith had been Eisenhower's chiefof staffin Europe, and he knew how to run a tight ship. Tussling with "Beetle" Smith was like landing in a cactus patch. In the early days the only high-level COMINT available to CIG was a copy of the MAGIC Summary put out by the Army, which was available in the Pentagon. In the very early days, only fifty people in CIG had a COMINT clearance. But in June of 1946 Hoyt Vandenberg became DCI. Vandenberg was fresh from a tour as chairman of USCIB and knew the value ofCOMINT. In December he created an organization within CIG, called the Advisory Council, to deal with what he hoped would be a flood ofCOMINT reports. For a while there were few reports to disseminate. Requests for access to COMINT reports were generally denied. But in early 1947, two CIG organizations began to get involved with COMINT operations. The first was OSO (Office of Special Operations, the clandestine organization), which in March proposed to the Army and the Navy that they begin a Joint Counterintelligence Center (JCIC), using COMINT as the basic source of information. The services received this enthusiastically, and JCIC was established at Nebraska Avenue, with the understanding that it would eventually move to CIG. (It moved to CIA in 1949.) At about the same time, Colonel Robert Schukraft, chief of the Communications Division at CIG, was establishing a relationship with ASA. Schukraft had been a key figure in wartime Army COMINT and knew many of the people involved in the COMINT business. He began a relationship with Frank Rowlett at ASA\ HANDLE VIA TALENT KEYHOLE COMI L SYSTEMS JOINTLY NOT R TO FOREIGN NATIONALS 87 Ib) (1) Ib) (3) OGA lOP SECRET '='MB~ The operational aspects of these budding relationships eventually came under the aegis of OSO, and specifically one William ("Bill") Harvey, a former FBI agent who became legendary for his clandestine operatioIlJi.....J.l.D.d.e:t..lia.l:JUli.1L....w.IlQ..~l.k..~a.r..iD....l.5I.5JJ~ n rs became centralized Ib) (1) Ib) (3) -50 USC 403 Ib) (3) -18 USC 798 Ib) (3) -P.L. 86-36 Meanwhile, CIA requests for COMINT reports were still being routinely turned down. But thel rcontributed to breaking the logjam, and ever larger volumes 01 COMlNT report senes were orwarded to CIA. Once at CIA, the material was subdivided according to subject matter and farmed out to analysts through the auspices of the Advisory Council. CIA was determined to base reporting on all-source information, rather than to strictly segregate COMINT from all other sources. Ofnecessity, then, the number ofCIA COMINT clearances rose rapidly, until by 1970 most intelligence analysts were cleared for the source. (See Table 5.) CIA policy stood in contrast to that of the Pentagon, which generally chose to compartment COMINT and to deal with two separate handling systems - COMINT and all other sources. HANDLE VIA TALENT KEYHOLE COMI ROL SYSTEMS JOINTLY NOT RELE 0 FOREIGN NATIONALS 88 TOP SEcRET tJMBAA - . (1) Ib) (3) ELINT was above HF, so sitestended to be located in great profusionl I lIn this field each SCA had been given the prim~a~r--y-c~o"'n-e-:etl""'lO-n--'-Jo~b-br-Y---"lt~s~.r~e-sp-e-c~t"'lv-eservice, and each moved quickly to establish sites. In many, if not most, instance.s ELINT preceded COMINT, and a ain in most cases ELINT sites already existed where COMIN'l'sites were later added. Added\.to this was a burgeoning airborne "---=----:'-----=-""':"":'"""":'"-:--~:::-:-=:-:::-"":":'~ collection system, fielded by USAFSS. NSA played no role in ELINT, either in collection or processing. When it carne to COMINT, though, NSA employed its guiding hand. Even before NSA was created, AFSA had a master plan for the establishment of SCA intercept sites which I I'l.'his plan was passed on to NSA, which refined it. NSA worked very closely with each SeA to determine collection requirements and determine the best candidate locations. In the early 1950s NSA asserted control over site surveys, without which no collection site could be established. NSA balanced customer requirements against existing overtl Isites, documented hearability, and Second and Third Party contributions. If the project did not make sense, DIRNSA could be counted on to oppose it.101 HA~H;)bK VIATAU)l"T KKYHQbg eQMI~tTeQ~tTRElJ,;;SY8lPSMS dElIN'fbY NOT REI E ASA BI Ii: TO i:ORgIG)l" ~h\TIElWI:LS 112 \ \ ( 1) (3)-50 USC 403 (3) -P.L. 86-36 \ "'" ( 1) (3)-50 USC 403 3) -P.L. 86-36 \ ~ ( 1) (3)-50 USC 403 (3) -P.L. 86-36 \ "'" ( 1) (3)-50 USC 403 (3) -P.L. 86-36 \ '" \ ~ ( 1) (3)-50 USC 403 3) -P.L. 86-36 '(b) (1) (b) (3) -50 USC 403 (b) (3) -18 USC 798 Ib) (3) -P.L. 86-36 Berlin was in an entirely different situation. Its status as a four-power occupied city meant that Soviets could walk relatively freely even in the American Sector..Stalin's attempt to squeeze the Westerners out of Berlin (resulting in the Bedin Airlift) in 1948 placed the city in a uniquely precarious position. In such circumstances the first COMINT intercept organization, a detachment of the ASA sit~ Iarrived in a covert status and stayed only a few weeks in 1951. But ASA covert detachments kept appearing in Berlin, and in the following year the command established a permanent unit there, and the troops moved from tents to covered buildings. In 1953 the Army G-2 concluded that the results had been paltry and recommended the site be closed, a strange finding given the later reputation of Berlin as a SIGINT bonanza. Fortunately, no one listened to the G-2, and ASA continued to occupy a variety of locationsI tAFSS followed ASA into Berlin in 1954, beginning a presence in the city that would last until after the fall of the Berlin Wall.102 Berlin became a SIGINT gold mine, a window into the heart of the Communist Bloc military system. In the mid-1950s the collection sites began to report the t~istence ofYHF communications, and NSA moved in to investigate. An NSA technician _ \ I discovered that Berlin was crisscrossed with above-HF communications that the West had never before intercepted, including Soviet high-capacity multichannel and microwave transmissions. The discovery was to have a profound influence on the development of the SIGINT collection system.103 • (3) -P.L. 86-36 I1AJ,BbEl lIlA 'i\\bElNT ICE'll lObE COllHUT COUTHOb S"[S'I'ElMS dOIN'fbY NOT RBbEh'tBA:BLS 'fO FOftEl6!( !M:TIONALI5 118 - - - - - ------- ----j Ib: (1) lb: (3)-50 usc 403 (b: (3)-18 usc 798 (b) (3) -P.L. 86-36 ------- / -------- ( 1) (3)-50 USC 403 (3)-18 USC 798 (3) -P.L. 86-36 OGA ( 1) ( 1) (3)-50 USC 403 (3) -P.L. 86-36 ( 1) ( 3) -------- ~- ( 1) (3)-50 USC 403 (3)-18 USC 798 (3) -P.L. 86-36 -------- (1) Ib) (3) -------/ ------/ ( 1) (3)-50 USC 403 (3)-18 USC 798 (3) -P.L. 86-36 \ "" ( 1) (3)-50 USC 403 3) -P.L. 86-36 \~ \ "" ( 1) (3)-50 USC 403 3) -P.L. 86-36 \~ ASA PAC in the 1940s ASA's first p()stwar Far East headquarters was in a relativ:ely intact building in downtown Tokyo. Japanese nationals staffed the support services. The Far North All three services established collection sites in Alaska. The Navy site at Adak dated back to World War H, and the Air Force and Army soon followed. The USAFSS site grew out ofthel • eWorld War II Army Air Corps asset. Security Service established its first collectir'tNBr:.E YIA 'fAbSf'f'f KSYIISr:.SOSMHf'f OSf'f'fR8r:. BY6IfSftlSd8In'fbY N8'f RFlbBABltBbB 'f8 P8RSIaN Uh'fISf'ilrbB (b)(1) (b)(3)-50 usc 403 (b)(3)-P.L. 86-36 141 '(bl (11 (bl (31 -50 USC 403 (bl (31 -P.L. 86-36 rb}(11 (b) (31~50 USC 403 (b 1(3 1- P .L,. 8 6-36 Table 6 Summary ofIncidents by Type, 1949-1985 Type Number '--- -----.1 Non-SIGINT aerial reconnaissance Other military Commercial/private air Military ship Commercial ship (Mayaguez) 22 56 18 3 1 The number and pattern of peripheral reconnaissance flights over the years, and the nationalist sensitivities of the Communist nations, produced a lively time. Some of the shootdowns became international incidents which heightened the Cold War tensions~nd seriously affected international diplomacy.130 All three services developed their own aerial reconnaissance programs, each using different types of aircraft. Of the three, USAFSS had the largest program. Security Service began laying plans as early as 1948, but it was not given the go-aheadfrom USAF until August 1950. Originally USAFSS hoped to use the C-47 as an airframe, and it actually tested that aircraft and a C-54. USAFSS decided on the RB-50, a modification of the B-29, as its long-range airframe, but none was available, and in the early 1950s the command used an RB-29 as an interim measure. The single J:tB-29 went operational in the Pacific in 1954, flying out o~ tbut this was never more than an experiment. AFSS finally ended up with a group often RB-50s in 1956, and by the fall of 1957 all ten were distributed - five to Asia and five to Europe. The program was a joint effort between AFSS and the theater commanders, who operated the. front end of the planes. In the early years ofthe program, only the back-end crew was COMINTcleared. All positions were under local control, and tasking was done by USAFSS with little or no NSA input.131 The Navy program developed from the early VQ-l and VQ-2 squadrons originally established in World War II. VQ-l was originally based at Sangley Point Naval Station in the Philippines, flying P4M Mercators, P2V Neptunes, and A3D Skywa.rriors. In 1955 the In Europe the SIGINT reconnaissance mission, VQ-2, evolved out of a World War II naval unit at Port Lyautey, Morocco.I t32 HltJ1BbEl VIA 'f1tbElU'f ItElYHebEJ eeMIN'f eeU'ffteL S';S'fElMS dem'fLY ~T9'f RjjJb8ASA8U '1'9 F9RlilIOU UA'I'IOUAJ"S 142 Ib) (1) Ib) (3) -50 USC 403 Ib) (3) -18 USC 798 3) -P.L. 86-36 Ib) (1) Ib) (3) -50 USC 403 (3) -P.L. 86-36 RB·5!) When converted for reconnaissance use, the World War II B·29 was renamed the RB·50. The SAC Ferret program co.ntinued in the postwar years with only minimal involvement by the cryptologic community. By late summer. of 1951, both AFSS and AFSA had become ....._---------_...... interested in the program, and by September the plans were expanded to include The Origins ofAdvisory Warning The AFSS unit atl tby now renamedl Irealized that the in their hands information that could save an aircraft from bein shot down. ..... ......In early 19521 Iworked out a plan to warn aircraft in imminent danger, by passing III coded warning to the Air Control and Warning (AC&W) sitesI IThey wrote down their plan into a document which they I1Am3I"S VIA 'f1tH~m'f KElYUSLFl OSMHff OSN'fHSL S';S'fFlMSd'tmif'fL i NOT REI E ASA aLi TQ FQR8I8U rhlt'f16IifA:LO'Il 143 ,(b) (1) '(b) (3) -50 USC 403 (b) (3) -P.L. 86-36 called Project BITTERSWEET and sent it to USAFSS for approval. In May 1952 AFSS approved the plan f.r temporary imPlementsli.; ! \ Idetails ofthe new warning procedure were still being worked out when,L...--::-:~_ _-;::-;::-I. on 13 June, an RB-29 SIGINTcollection flight was shot down over the Sea ofJapan. The two AFSS operators who were killed might have been saved had a system been in place; the event added a real sense ofurgency to this, the earliest advisory warning plan in American SIGINT history. At this point BITTERSWEET got bogged down in the tangled thicket of COMINT classifications. The problem revolved around the ossible USCIB approved the USAFSS advisory warning plan for the Far East, but LSIB was reluctant t<;l go along except in a war zone (Le., Korea). It appears that at least one version of the plan was given interim approval by USCIB, and a former USAFSS operator claims that it was actually implemented in the early 1950s for at least one mission. Variousmodifications were introd\lced to make it more palatable, such as the use ofbogus messages disguised as warning messages by AC&W units. In 1956 President Eisenhower, concerned over the number of incidents and loss of reconnaissance aircraft, directed that positive action be taken to remedy the situation. The only change that resulted was the implementation of a Navy warning program in the Far East, which contained certain safeguards, chief among these being the initiation of "blind" (unacknowledged) broadcasts. Through the Summer of 1958, there existed no universal advisory warning program. 133 The RC-130 Shootdown The RB-50 program lasted only a few years. The aircraft were old and difficult to maintain and had room for only five positions. The success ofAFSS collection against the growing VHF problem led to a new program on the heels of RB-50s, in which the new McDonnell-Douglas C-130 would be converted to a collection platform. The C-130 had room fotOPositions, could fly longer and higher, and,being new, had few maintenance problems. AFSS planned for a fleet oq lin each theater, to begin in 1958. The firstDvent to Europe, and in September AFSS, in association with USAFE, began to fly trial reconnaissance missions inl lareas. 134 I1hNBIoEl VIA 'fAIoElN'f U:ElYII6IoEl e6MIN'f e6H''fft61o S'iB'ftilMBd6m"fh'i NOT RKIsK • is' au; 'Fe Fe~ISN fM'ff6fO\hB 144 RC·130 Then disaster struck. On 2 September 1958 an RC-130 on its initial flight out of Adana strayed over the border and was shot down. Two pairs ofMIG-15s (or 17s; there was not enough evidence to determine which) attacked the reconnaissance aircraft in waves in a well-coordinated operation which left no room for doubt that their intent was destruction. The voice tapes were as dramatic as they were damning. (See p. 146.) The Soviets said nothing, so the State Department on 6 September sent a note to the Soviet government requesting information on an unarmed C-130 carrying a crew of seventeen which had disappeared during a flight from Adana to Trabzon, Turkey. Finally, on the 12th the Soviet embassy in Washington replied that the missing transport had crashed in Soviet Armenia, killing six crew members, but that Moscow had provided no information about an additional group of eleven. An exchange ofdiplomatic notes over the next ten days shed no further light on the missing eleven bodies, so on 21 September the State Department admitted that they knew the aircraft had been shot down and appealed for information on the rest of the crew on humanitarian grounds. The Soviets replied that they considered the flight to have been an intentional violation of their borders but made no reference to the involvement offighter aircraft or a shootdown. IIltNBhS Yilt 'fALSN'f I(SYIISLS eSMlH'f eSN'fftSL SYSTEMS J6UifTL i ~OT Ri:U "~,, JilU 'i'Q peRSiaN N*TI6!ifJ(L~ 145 PBS [pilot billet suffix] 582 201 to 218 201 201 201 218 [missing] [missing] [missing] [missing] [missing] CONVERSATION The target is a large one.... Roger Attack! Attack! 218, attack. I am attacking the target Target speed is 3000, I am flying with it. It is turning toward the fence [Le., border]. The target is banking.. " It is going toward the fence. Attack! Yes, yes, I am attacking. The target is burning.... The tail assembly (b% is falling oro the target. Look at him, he will not get away, he is already falling. Yes, he is falling (b% I will finish him off) on the run. The target has lost control, it is going down. A crew of seventeen men, including eleven USAFSS airmen and a front-end crew of six, was lost. In October the Soviets produced the bodies of the six members of the front-end crew, but the bodies of the eleven USAFSS airmen were never turned over; and this strange circumstance produced a spate of conspiracy theories regarding the possible capture and long-term incarceration, not to mention forceable interrogation, of the COMINT crew. The evidence of the voice tapes makes it quite clear that no one could have escaped the fiery crash in a mountainous region ofthe Caucasus, but what happened to the bodies remains a mystery to this day.ls5 In November, after more than two months of Soviet "stonewalling," Deputy Under Secretary of State Robert Murphy summoned Soviet ambassador Mikhail Menshikov to his office, told him he had the voice tapes of the shootdown, and said he would play them immediately. Menshikov declared that he was not a technician and walked out of the office. In January of the following year, Vice President Nixon and Secretary of State Ib\f'fBLfi:l VIA 'fALfi:lfff Kfi:lYIISLfi:l e8MIU'f' e8N'f'ft8L SYS'f'fi:lMS7" April 1957, ACC 10684, CBRI 52; and ~AFSS-NSA Relations, October 1952-September 1954," V. I, USAFSS Official History available at Hq AlA, Kelly AFB, San Antonio, Tx. 35.1 I~EarlY BOURB<)N - 1945: The First Year of Allied Collaborative COMINT Effort against the Soviet Union," Cryptologic Quatterly, Spring 1994, 1-40. 36. CCH Series VI.EE.1.3J tSIGINT Directives,: 28 Sep 1992 NSA(P0443) paper; ~Mechanization in Support ofCOMINT," collectlOn ofpapers dated from 1954-1956, in CCH Series XlI.Z. 37. "Mechanization in support ofCOMINT." 38. "Site SurveylHearability Tests," in CCH SeriesVI.I.l.l. 39. George Howe, ~Narrative History," Part V. 40. CCH Series V.A.28., and VI.M.1.5. 41. Letter, Canine to SCA chiefs, 14 June 1955, in CAHA, ACC 26418, CBOM 16. 42. CCH Series VI.M.1.5. 43. USAFSS History, "Historical Resume: Development and Expansion of USAFSS Capability in the Pacific Area, 1949-1957," available at AlA hqs, Kelly AFB. 44. Memo fm MGen Samford 1)2g;A.?~26424, CBOM 22. ~~Meade:~~~'19::r1itiQ5oLl~he Berlin Tunnel, U.S. Cryptologic History, Special Series, Number 4 I "Newspaper Items Relating to NSA," in CCH Series VLII.1.2.; Kirby interview. 93. Thomas Powers, The Man Who Kept the Secrets, 83. 94. Interview with Richard;Bissell, Jr., by Dr. Thomas Soapes, 19 Nov 1976, CIA history staff. 95. "CIA·NSA Relationships ... ,"CGH Series VI U.1.2.; Tordella interview. 96. Robertson Report, in CCH Series VI.C.Ul; NSAlCSS Archives, ACC 29965, HOI-0706-5 97. See Jones, The Wizard War: British Scientific Intelligence, 1939-1945 (New York: Coward, McCann and Geoghegan,1978). 98. Draft Robertson Committee report in CCH Series VI.X.1.7. ~- 100. "Background to the Robertson Report: Potentialities of COMINT for Strategic Warning," in CCH Series VI.X.1.7; Draft of Robertson Committee Report in CCH Series VLX.17.; "ELINT History and Background Ib) (1) Ib) (3) -50 USC 403 Ib) (3) -18 USC 798 Ib) (3) -P.L. 86-36 HANDLE VIA TALENT KEYHOLE COMINT CONT SJOINTLY NOT RELEAS IGN NATIONALS 153 ~ ( 1) (b) (3) OGA TOP SFCAH tJMBAA Papers," in CCH Series VI.O.1.1.; Hoover Commission Report in CCH Series VI.C.1.8.; "History ofthe Electronic Intelligence Coordinating Group, 1955-1958," in CCH Series VI.O.1.6. 101. NSAlCSS Archives, ACC 9092, CBIB 24; "Acquisition of COMINT Intercept Stations Overseas," in CCH Series V.F.6.1. 116. Howe, "Narrative History," Part V, Ch. XXVI-XXX; Ferry, "Special Historical Study... "; "Synthesis ofASi\ Programs, FY 53," in CCH Series VI.Q.1.14; NSACSS Archives, ACC 9092, CBIB 24; "History, Location and Photos of ELINT sites," in CCH Series VI.O.1.4.; "INSCOM: and its Heritage - a History," in CCH Seri.es VI.Q.1.15; MSgt William R. Graham, "Misawa - Air Base and City, " 1982. (b) HANDLE VIA TALENT KEYHOLE COMINT C STEMSJOINTLY NOT RELE OREIGN NATIONALS I -10 USC 130 -50 USC -18 USC 798 TOP 154 Ib) (1) Ib) (3) -P.L. 86-36 Ib) (3) -50 USC 403 Ib) (3) -18 USC 798 Hb) (3) -P.L. 86-36 119. ~A Chronology of Significant tveJ;ltsin the History of Electronic Security Command, 1948-1988," 1990, available at AlA Hq, KellyAFB, Texas; Interview with Gordon W. Sommers by Millard R. Ellerson and James E. Pierson, January 1990, available at AlA Hqs. 120.1 ~rticlein NCVA Cryptolog, Fall 1989; ~INSCOM and its Heritage - a History," in CCH Series vl.Q.1.l5. 122·1 ;::::::=====~---123.1 • • ~RadiODirection Finding in the U.S. Navy: The First Fifty Years," unpublished manuscript in the CCH Senes VII.85. 124. Don East (Capt, USN(R»), ~A history of U.S. Navy Fleet Air Reconnaissance, Part I: The Pacific and VQ-l," The Hook, Spring 1987, 16.; Oral in~rview with Samuel K. S. Hong, Dec. 9,1986, by R. D. Farley, Honolulu, Hawaii, NSA OH 40-86. 125. East, 15-17. 126.1 tOld BOURBON -1947: The Third Year ofAllied Collaborate COMINT Effort against the Soviet Union," Cryptologic Quartetly, Vol. 13 No.3, Fall 1994. 127. William E. Burroughs, Deep B{qck: Space Espionage and National Security (New York: Random House, 1986), 58-9.; "A History ofthe USAFSS Airborne SIGINT Reconnaissance Program (ASRP), 1950-1977," in CCH SeriesXJ. 128. ~Analysis ofAFSS Effort in the ~orean Action," unpublished manuscript available in CCH Series V MA.1.; "A History of the USAFSS Airborne SIGINT Reconnaissance Program (ASRP), 1950-1977," 20 Sep 1977, USAFSS history available in CCH Series XJ. 129. See BuroughsDeep Black, 58-59, and Thomas Powers, The Man Who Kept the Secrets, 24. 130. The definitive study of reconnaissance incidents was done by NSA historian Donald Wigglesworth in an unpublished study entitled "A Summary of Peacetime Hostile Air and Sea Actions 1949-1985," March 1986, in CCH. In addition,1 hcentIy published an article in the Cryptologic Quarterly ("Maybe You Had to be There: The SIGINT on Thirteen Soviet Shootdowns ofU.S. Reconnaissance Aircraft," Vol. 12 No.2, Summer 1993, 1-44), which provides an excellent summary ofthe shootdown ofreconnaissance aircraft.. 131. "A History of the USAFSS Airborne Sigint Reconnaissance Program ... ," Mary Holub, Jo Ann Himes, Joyce Homs, and SSgt Kay B. Grice, ~A Chronology ofSignificant Events in the History of Electronic Security Command, 1948-1988," in CCH Series X.J.6. 132. East, "VQ-I. .. ," and East, "The History of U.S. Naval Airborne Electronic Reconnaisaance: Part Two, the European Theater and VQ-2," T he Hook, Summer 1987, 32-35. 133. ~Analysis ofAFSS Effort in the Korean Action," USAFSS history available in CCH Series V.M.4.1.; CBTJ 44, NSA/CSS Archives, ACC 19220; USAFSS, "A Special History of the Advisory Warning Program, July 1961December 1964," in CCH Series X.J.3.1. 134. USAFSS, "A History ofUSAFSS Airborne SIGINT Reconnaissance...." IIlrPiBLB TlIA 'fALFJ!:f!' I{ErtII6L~ eO!9III~ I CON lRUL Sf SIEMSJOIN ILx 155 Chapter 4 The Soviet Problem THE EARLY DAYS From Stettin in the Baltic to Trieste in the Adriatic, an iron curtain has descended across the Continent. Behind that line lie all the capitals ofthe ancient states ofcentral and eastern Europe ... all these famous cities lie in what I might call the Soviet sphere, and all are subject, in one form or another, not only to Soviet influence but to a very high and in some cases increasing measure ofcontrol from Moscow. Winston Churchill, 6 March 1946 The end ofWorld War II did not result in a large number of unemployed cryptologists. That it did not was due almost entirely to the advent of the Cold War and an increasing concern with what came to be called the Soviet Bloc. (In the 1950s, believing in a worldwide Communist conspiracy, Americans called it the Sino-Soviet Bloc.) Wartime cooperation with the Soviet Union began to break down in early 1945. Through a series oflate-war conferences among the Allies, it became clear to the West that the Soviet Union did not intend to retreat from Eastern Europe at the end of the war. An increasingly frustrated Roosevelt administration became less and less constrained about public references to the rift with Stalin, but Roosevelt himself remained convinced up to his death in April 1945 that the rift could be healed by diplomacy. His successor, Harry Truman, did not share this optimism. The administration moved to check Soviet expansionism abroad. As a result of strong pressure, Stalin removed Soviet troops from Iran later in the year. Meanwhile, Greece was faced with a USSR-inspired internal Communist threat, while neighbor Turkey faced an external threat by Soviet divisions massed on its borders. Truman again faced down Stalin, announcing the Truman Doctrine, a promise to come to the aid of countries in that area faced with Communist subversion or external threats. Administration policy toward the USSR hardened with the publication, in the magazine Foreign Affairs, of an article by George Kennan, late deputy chiefofmission in Moscow, postulating the Cold War doctrine which became known as "containment." The next year a democratically elected government in Czechoslovakia fell to a Communist coup, and the new government became an effective satellite of the USSR. Meanwhile, Soviet troops remained in Poland and East Germany, while Communist governments took over in Hungary and the Balkans. In June 1948 Stalin tried to cut Berlin offfrom the West, and Truman initiated the Berlin Airlift to resupply the city. The H:A:NBbEl VIA 'f:A:bElfff IlElYH61:JiJ e6MUff e6U'fft6b Si S'fI!JMS tf6U,'fb i ~TQT RKbEASABcE] 'fe peRSISN ?M'fI6U:AbS 157 Truman administration also saw the Korean War as the first move in a Soviet-inspired military offensive. The Advent ofBOURBON American cryptology had dabbled with the Soviet problem over the years, with indifferent success. Yardley, in his book The American Black Chamber, claimed to have broken Soviet diplomatic codes. The truth is that, though Yardley's MI-8 worked Soviet diplomatic traffic, only a single instance of success was ever recorded, and in that case the transposition being attacked was based on the German language.1 Friedman's Signal Intelligence Service obtained MI-8's traffic upon MI-8's demise in 1929 and made a brief, unsuccessful attempt to solve the codes. Then in 1931 a Soviet espionage front posing as a trading company called AMTORG came under the glare of Representative Hamilton Fish of New York, who subpoened some 3,000 AMTORG cables from the cable companies in New York. Fish turned them over to OP-20-G, which, having at the time only two cryptanalysts (Safford and Wenger), failed to solve them. The cables were then transferred to SIS, which also blunted its spear. This was virtually the only attempt at Sovcdiplomatic traffic by the services during the 1930s, and Friedman's people doubted that any Soviet codes could be solved.2 They were, in fact, wrong. 1-----------------------.....,1 attack on Soviet military systems throughout the 1930s. The primary target was COMINTERN (Communist International) traffic,1 I But when, in June of 1941, Hitler's army invaded Russia, the British allowed the Soviet problem to wither. GCCS made a brief attempt to turn the USSR into a COMINT Third Party, and even established an intercept site in Russia near Murmansk in 1941. The dialogue came to a quick halt when the Soviets began inquiring into British success against ENIGMA. In 1942, the Radio Security Service and the London Metropolitan Police discovered an extensive Soviet illicit network in Great Britain, and Stewart Menzies, head of British intelligence, directed that work be renewed against Soviet communications, especially KGB, GRU, and COMINTERN nets likely to carry information of counterintelligence value.3 In the United States, SIS was collecting a small amount of Soviet traffic on a casual basis as early as 1942. On 1 February 1943 the Army opened up a two-person Soviet section. The inspiration for this effort was the Army's successful attack on Japanese diplomatic communications, in which the Japanese discussed their efforts against Soviet systems. The Japanese material gave SIS some handholds into Soviet systems. OP-20-G came in later, opening both intercept and cryptanalytic study in July 1943. Because the USSR was a wartime ally, the effort was rigidly compartmented and known to only a few. In August 1943 the Army and Navy cryptologists began cooperating on the new Soviet IIltHBLtJ VIlt fltLtJNf If:tJ r tlI6LtJ e6fvlIWf e6N'fH6L SYS'fEMSif6IH'fLY NOT REI E A it A IU"I!: TQ F9R~I8H NA'fI6l(2\L~ (b) (1) (b)(3)-50 USC 403 ~~BRA 158 problem and, during 1943 and 1944, cooperatively worked a number of Soviet cryptographic systems.4 Meantime, the Navy, in order to collect Soviet naval traffic, had opened up an intercept effort at Bainbridge Island in Washington State. Tightly controlled, it was headed by Louis Tordella, later the deputy director ofNSA,5 By the end of the war, both cryptologic organizations were mounting extensive efforts against Soviet communications, despite the official designation of the USSR as an ally. OP-20-G, concentrating on Soviet naval communications in the Pacific from Skaggs Island and Bainbridge, employed 192 people, while ASA had almost 100. They had both been surreptitiously training Russian linguists for a year. But the effort was charged with political implications. Roosevelt was trying to maintain the fragile alliance with the USSR and was being challenged on the left by Henry Wallace, a potential political rival who felt the administration was anti-Soviet. In this atmosphere Brigadier General Carter Clarke of the Army G-2 paid a visit to Preston Corderman (chief, SIS) and Frank Rowlett several months before the end of the war. Clarke said that he had received informal instructions - allegedly from the White House - to cease any effort against the Soviet problem. It appeared that someone in the White House had gotten word of the compartmented Soviet problem and had concluded that it did not accord with the current diplomatic situation. (It was discovered years later that the White House staff was in fact infiltrated by a single Communist or "fellow traveler," who may have been in a position to know about the Army program.) Clarke did not desire that the program be closed, and in fact SIS (soon to be renamed ASA) received a steady increase in resources for the program.6 In June 1945, with the war coming to an end, the Navy proposed formal collaboration with the Army on the Soviet problem, which was then referred to as the RATTAN project. The Army wanted a more integrated effort, but they eventually agreed to organize under the more decentralized Navy scheme. At the same time, ANCIB proposed to LSIB that their cooperation against Germany, Italy, and Japan now be extended to include the USSR. The Americans proposed that the cooperation be fully as close as it had been during the war. This included sharing all details, including the status and method of cryptanalytic attack, and the exchange of raw traffic and code/cipher recoveries. The British agreed, and in August the two sides arrived at an unwritten agreement predicated on an understanding arrived at in June between Rear Admiral Hewlett Thebaud, chairman of ANCIB, an1 Ifor LSIB. This historic agreement extended bilateral cooperation into the Cold War and established the basis for what became known as the BRUSA Agreement. The two sides agreed to call the new project BOURBON.7 (b) (3)-P.L. 86-36 Ih\!fBLfJ VIA 'fALl!JN'f t{fJY1I8LfJ 68MU,,'f 68N'fft8L S'[S'ffJMSd8m'fLY NOT REI F A ~ A iU 1'9 F9REH8N' r.qA'fI8NALS 159 (b) (1) Ib) (3) -50 USC 403 Ib) (3) -18 USC 798 Ib~(3)-P.L. 86-36 During the mid-1940s the two sides mounted a relentless attack on the w&rtime generation ofSoviet ciphers. The British provided much of the cryptanalytic expertise, the Americans most of the processing capability. They used whatever material they could get their hands on, including information on the Japanese cryptanalytic attack. TICOM debriefings of German cryptologists also gave the two partners useful information about Soviet systems. VENONA Alone and compartmented, the effort against Soviet diplomatic traffic had continued throughout World War II. In the long run this tightly held problem would have the greatest impact on American history in the postwar period and would become the most widely known. It was called VENONA. In the early years of the war, the Army received incidental Soviet diplomatic traffic, most of it through its arrangements with the cable companies, which carried a large bulk ofcommon-user communications. Since New York was the terminal for the transatlantic cable, Soviet diplomatic traffic was routed through that city. The Army arranged with the cable companies to get copies of most of the cables that the Soviets were sending, both to and from Washington and, more important, to and from AMTORG. Much of this traffic was believed to be KGB-related. In 1943, ASA had mounted a secret effort to attack these communications, but they looked impossible. They were produced from codebooks enciphered by means of one-time additive tables. Assuming no re-use, there was no point in continuing. But ASA was not assuming anything, and Lieutenant Richard T. Hallock of ASA directed that his small section machine punch and process the beginnings and endings of some 5,000 messages to test for depths. In October 1943, ASA found the first indication that the additive pads may have been used more than once, a find which was to change the history of the postwar world.9 Hallock and his small band of cryptanalysts had found what is called "manufacturer's re-use" caused by the first German invasion of the Soviet Union in 1941. The KGB's additive pad generating facility produced two sets of some pads, presumably because ofthe II*!'BLB YIA 'hLB!" I(BYII8LB 86MIN'f 86N''fIt6L SYS'fEMS J6fIq''fLY 160 pressures associated with the rapid German advance toward Moscow. These were disseminated to widely separated KGB organizations, which were unaware that they had duplicate pads. ASA never found depths of more than two, and at that depth, decryption was only theoretically possible but practically a back-breakingjob, assuming one ever got hold ofthe depths themselves. Months went by, but finally ASA cryptanalysts, in November of 1944, were rewarded with their first depth. This was followed by others, and it appeared that they might be able to eventually break some traffic. But the job still looked gargantuan. While one section worked on identifying depths, another worked on the underlying codebooks that were slowly emerging from under the additive key. This effort was led by a reclusive linguist and bookbreaker named Meredith Gardner. A Texan originally, Gardner had obtained a Master of Arts in German from the University of Texas and had been a Ph.D. candidate at the University ofWisconsin before going to work teaching at the University of Akron. He had joined SIS in 1942, and although he began in the German section, he quickly switched to Japanese, where he proved his linguistic gifts by picking up this extremely difficult language in just three months. At the end of the war, he switched again, this time to the Soviet problem and spent his first several months learning Russian. In December 1946, he had only recently emerged from language school when he made a major break into a KGB message, decrypting and translating a digraphic sequence of a 1944 message from New York to Moscow sending English text. Gardner found that the KGB used the code values for "spell" and "end spell" anytime they needed to encrypt a foreign word or other term that did not appear in the codebook. It was these two values that yielded many ofthe early breaks. In December 1946, Gardner broke a portion of a KGB message that listed American scientists working on the atomic bomb. This message turned heads. Why would the KGB be interested in such information? ASA immediately turned the translation over to the Army G-2, and Carter Clarke had General Omar Bradley, the Army chief of staff, briefed on the message. G-2 expressed a continuing interest in any messages that contained like information.10 Through the war ASA had proceeded virtually unaided, but after World War II several outside factors speeded the tortuously slow process of additive key diagnosis and recovery and bookbreaking. The first was the defection of a Soviet GRU cipher clerk, Igor Gouzenko, from the Soviet embassy in Ottawa, in September 1945. The case caused a sensation because Gouzenko indicated the existence of a possible Soviet effort against the American atomic research effort.a Because Gouzenko worked with communications, Frank Rowlett ofASA was invited to interrogate him. During his sessions Rowlett learned much about the way the KGB codebooks were put together and how the additives were used. This information cut time offASA's cryptanalysis effort.12 IIJltfn~LI!JVIA 'fALI!Jlff l()i)YII6LI!J e6hIIN'f e6K'fft6L SYS'fI!JMS 66IN'fLY PO,S,+, RI!1J"Fh*,s"',BJ,,B 'fe peRBIaN Nlt'fIeNAbS 161 TOP SECRET tJMBRA A second outside source of information was a 1944 FBI burglary of AMTORG, during which the agents carried off stacks of unenciphered messages with their cipher text equivalents. In 1948 the FBI turned over this bonanza to Gardner, who began comparing the traffic against transmitted messages. In this way he could identify some of the code group meanings because he had both plain and cipher texts.is A third outside source was called "Stella Polaris," a Byzantine story which began in the early days of World War II. When, in June 1941, Germany invaded the USSR, the Finns went to war against the Soviets, siding with Germany against their mortal enemy. On 22 June a Finnish unit, presumably security police, entered the Soviet consulate in the Finnish town of Petsamo, near the Russo-Finnish border. Here they found the Soviet communications people frantically destroying cryptographic material. Some of it was burned beyond use, but certain ofthe codebooks were recovered more or less intact. These codebooks were property ofthe First Chief Directorate ofthe KGB - they were, in fact, the same codebooks which, in the mid-1940s, Meredith Gardner was working on. The charred codebook fragments were turned over to the Finnish COMINT service, headed by one Colonel Hallamaa. By 1944 the war was not going well for Germany, and Hallamaa became concerned about an impending Soviet invasion of his homeland. He arranged to smuggle the contents of the Finnish COMINT archives, including the Petsamo trove, to Sweden, where photocopies were made. Copies of the Petsamo materials wound up in the hands ofthe Swedish, German, and Japanese COMINT organizations. Along with the documents went Hallamaa and the entire Finnish COMINT service. At some point information got out to the newspapers, and the fact that Finnish intelligence people were working hand in glove with the Swedes became public knowledge. Knowing that the KGB was almost certainly after him, Hallamaa and most of his people fled to France, where, after the war, they worked nominally with the French intelligence people, but were actually controlled, according to some sources, by the British. So it was that the British got their own copies ofthe Petsamo codebooks. At the same time (1945) an ass representative began working with Hallamaa, and the ass, too, received its own copies (although not, perhaps, a complete set). The codebooks eventually made their way to ASA and AFSA. Since by this time a number ofintelligence services had copies, which source did AFSA get? In the days after the war, a TICOM team obtained a copy from the Germans, and it was this set that first made it all the way to Meredith Gardner's office. Shortly thereafter AFSA began obtaining Petsamo materials from the British under the codename Source 267 and may, at some point, have received copies from OSSICIG, but these were no more than duplicates of materials they already had.14 HANDLE VIA TALENT KEYHOLE COMINT YSTEMSJOINTLY NOT RELE FOREIGN NATIONALS 162 (1) Ib) (3) -50 USC 403 Ib) (3) -18 USC 798 Ib) (3) -P.L. 86-36 The Stella Polaris find did not break the KGB codes. They were fragmentary and pertained only to one version, But they did shorten the time'-- ----1 involved in the laborious bookbreaking process by providing Gardner with "models" of Soviet codes on which to base his own recoveries. After reading some of Gardner's earlier translations showing the scope of KGB operations in the United States, Carter Clarke, the Army G-2, called on the FBI for help. His first contact was in July 1947; Wesley Reynolds served as a link it was with Wesley Reynolds, the FBI liaison in the NSA·FBI liaison and later with Army G-2. Reynolds had joined the FBI became NSA's chiefofsecurity. in New York in 1941 after several years of law practice with hisfather and older brother. He had begun liaison work with G-2 in 1942, and ten years later jumped ship to NSA, where he became NSA's first professional chiefofsecurity. Reynolds concluded that VENONA could turn out to be. a full-time job, and he appealed to Mickey Ladd, head of the FBI counterintelligence operations, for a dedicated agent. Ladd assigned one Robert Lamphere, who, like Reynolds, had joined the bureau in 1941. Lamphere had worked virtually his entire career in counterintelligence, mostly in New York. He knew the territory, but he did not yet know ASA and Meredith Gardner. What ensued was one of the most remarkable partnerships in intelligence history. The shy, brilliant Gardner, speaker of half a/dozen languages, brought to the relationship his ability to break codebooks and produce translations of extremely difficult material. Lamphere brought his detailed knowledge of KGB operations and personalities, along with his contacts within the counterintelligence community. Together they worked over the fragmentary texts ofold KGB messages. One of the first products of/this marriage of convenience came in 1948. It was a decrypt of a message sent in 1944, in which the KGB reported on the recruiting efforts of an unnamed spy. Using the FBI counterintelligence file, Lamphere identified two possible candidates: I ~n employee ofthe Navy Ordnance Department, an~ I an engineer working on airborne radar for Western Electric. Both had been under FBI suspicion for possible Communist liaisons. Neither was ever brought to trial, but it was the first fruit ofthe Gardner-Lamphere relationship. IIftUBI:il!J VIA 'fAI:il!JU'f IEl!JYIISLl!J eaMHi'f eSN'fRab SY&FBMS " Ib) (1) Ib) (3) -P.L. 86-36 Ib) (3) -50 USC 403 HANDLE VIA TALENT KEYHOLE COMIN YSTEMSJOINTLY NOT 0 FOREIGN NATIONALS 175 TOP SECRET tJMBRA - Ib) (1) Ib) (3) OGA as "SMAC" (Soviet Missile and Astronautics Center). SMAC established aL..- ...I new operator-to-operator communications system which became known as thel IJoseph Burke, now regarded as the "father of SMAC," is believed to have originated this system.so To orchestrate the system, SMAC established an all-night watch, virtually eliminating the call-in system for this critical project. SMAC was one of the organizations that eventually got NSA out of the eighthour-per-day mode, and it pioneered in Joseph Burke the development of tip-off systems and quick reaction capabilities. In both concept and technology, it long preceded NSOC.31 NSA had numerous competitors in the missile arena. The Air Force had launched a small detachment of ATIC (Air Technical Intelligence Center at Wright-Patterson Air Force Base, Ohio) in San Antonio. Collocated with Air Force Security Service, SMTIG (Soviet Missile Technical Intelligence Group) consisted of a cross-section of the Air Force intelligence disciplines, but it was dominated by SIGINT people. Its analysis directly overlapped much of what NSA was doing. In addition, CIA was well along on its missile analysis effort and included SIGINT as well as other intelligence disciplines in its program.32 The Soviet Atomic Bomb Project While the Soviets were developing delivery systems, Stalin directed that the development of the nuclear weapons themselves be given the highest priority. Working with information provided by the atomic spies in the West, and with captured German nuclear physicists, the Soviets raced to get the bomb. Their first test site was constructed at Semipalatinsk (now referred to as "Semey"), a remote Siberian location, and for some years the Soviets used that site exclusively. The Semipalatinsk monopoly on nuclear tests HANDLE VIA TALENT KEYHOLE COMINT CONT SJOINTLY NOTRELEAS EIGNNATIONALS Ib) (1) Ib) (3) -P.L. 86-36 Ib) (3) -50 USC 403 Ib) (3) -18 USC 798 TO 176 Ib) (1) Ib) (3) -P.L. 86-36 Ib) (3) -50 USC 403 Ib) (3) -18 USC 798 fib) (1) jj Ib) (3) OGA was finally broken in 1955 with the explosion! OJ an underwater device in the sea off Novaya Zemlya, a large Arctic island north.east/ofthe Kola Peninsula. 33 As with the missile development prograIll, so' it was with nuelear weapons The bomber and missile gap controversies in the late 1950s triggered a search for an operational Soviet strategic nuclear delivery organization. With the launch of S utnik in October 1957, this became a white-hot priority, he Soviets did not yet have a nuclear delivery organization, all the information from Senator Symington notwithstanding. In January 1960 the USSR publicly announced the formation of a new Strategic Rocket Forces (SRF) command.1 / In 1960, DCI Allen Dulles directed that the chairman of the Guided Missile and Astronautics Intelligence Committee (GMAIC) organize a study group to completely evaluate the Air Foree contention that there was a missile gap. U§ingl I PhOloi:p~ie ~;~det collected from. the U-...2....,....t..h....e.. conlIIlitteecOil.cluded that only the test site at was capable of launching a missile. This contradicted the latest nationa mte Igence estimate, which postulated that there would be thirty-five operational launchers by mi/ • IIfthe Umted :states could not penetrate the Iron Curtai1 tthey would have to do it from the air. Attempts hadalready been made. In the late 1940s the CIA had tried to float high g .....1006s over the USSR, equipped with cameras and recorders. This so-called rogram failed dismally. The few balloons that floated all the way froni o lelded little useful information.42 ' - - - More determined were deliberate overflights of Soviet soil. SAC had a highly compartmented (and still obscure) overflight program, carrying a variety of sensors. This dangerous approach to intelligence collection was augmented by the RAF, which mounted occasional overflights. But their participation was limited and ended after one famous incident in 1953. At American behest, RAF aircraft overflew Kap~stirlYarjl I I l'I'heycame back with their planes shot full of holes and allegedly toldt~eAr:nericafisthatif they wanted that sort of thing done, they could jolly well do it themselves. I ~A series of ground-breaking studies in the early 1950s urged Eisenhower to plunge into advanced technological alternatives. One of the most attractive proposals was HANDLE VIA TALENT KEYHOLE COMINT C EMSJOINTLY NOT RELE OREIGN NATIONALS . :Ib) (1) !Ib) (3) 'OGA 179 Ib) (1) Ib) (3) OGA TOP SErBi; ~MB~ suggested by the Surprise Attack Panel, a committee set up under Dr. James R. Killian, president of MIT. Dr. Edwin Land, a member of the panel and inventor of the Polaroid camera, suggested that a camera could be devised which could take pictures from very high altitudes, if the Air Force could build an airplane from which to mount such a camera. In November 1954, Allen DUlle~go~ ~o build some thirty new aircraft which had been designed for justsuch a purpose by Kelly Johnson, the top designer at Lockheed. They were calledU~2s.44 There w~safthe time no guarantee that the U-2 was the answer. In fact the tse"hOWP; ,dm;";,'..';O" Co"ti"u.d to "J" rc~s:;e~al}O~:::ii;'::;:;~j::J:::I ""erefloatedacrosslheUSSR from Europe to Asia in early 1956.1 1 ........................... l;;;;.;;;;== ,... ....,. ..,...----,Jand some of theJI1IIlay have Eu successfultha and of the 500, only forty-four were~ev-e-r-r-e-c-o-ve...red after their long ride from west to east.45 The U-2 project was a very risky gambit by an administration desperate to find out what was happening in the Soviet Union. Advanced equipment was placed aboard an aircraft easily picked out on radar, and defensible only because of its operational altitude. If the Soviets ever got a weapon that would shoot that high, the U-2 could be a sitting duck. This was undoubtedly in Eisenhower's mind when in 1955 he broached the Open Skies proposal to Khrushchev. The U-2 had not yet been launched, but when it was, it would be a target.46 From the time of the firstU-2 overflight on 7 Apri11956, to the shootdown ofFrancis Gary Powers on 1 May 1960, the Eisenhower administration launt.hed twenty-four missions. The objective was photography, and the targets related to Soviet strategic systems. The aircraft also carried anc::::::::Jpackage, but this was probably used for internal defense (presuma f nfri ndi t '. reat and to tar et the cameras. III S eCla rooms - oni a few individuals at each site were cleared. (1) Ib) (3) -50 USC 403 Ib) (3) -18 USC 798 Ib) (3) -P.L. 86-36 HANDLE VIA TALENT KEYHOLE COMINT CON MS JOINTLY NOT RELEASAB IGN NATIONALS TOP SECRt!EIJJ.MB:~- 180 ij I !j I :j I j j j j :j I j ,j , .... Cl ,;j 00 ~- i j j j t j I j j "'" j ~ ~ j j j j j -- - I :(1) (3)-50 USC 403 (3) -P.L. 86-36 / (Ib) (1) Ib) (3) OGA (1) Ib) (3) -50 USC 403 Ib) (3) -18 USC 798 Ib) (3) -P.L. 86-36 .!Q.P5f'iAEf-OMBRA Prior to the Powers flight, NSA began to note increased lJertryFenech, the NSA6ilicial in charge of the operation, state t at a mlsslonJuspnott9the infamous May.Day flight was chased by a Soviet interceptor aircraft all tile wayto Mghanistan. It was obvious to Fenech that the Soviets were loadin u .49 Powers took off on 1 Ma1960 Back at NSA, Fenech reported to CIA that the aircraft had probably been lost to unexplained causes. It was the first loss of a U-2. CIA was desperate to know what had really happened to the aircraft, and in early 1962 General C. P. Cabell, deputy director of the CIA, decided to trade Soviet spy Rudolph Abel for Powers. In March 1962, only a month after the return of Powers, CIA called a board of inquiry, and into the middle of it marched Fenech, accompanied by NSA Director Laurence Frost, Deputy Director Louis Tordella, and Assistant Director for Production Oliver Kirbyl I Boththe Soviets and Powers said that the plane had been shot down at high altitude with anSA-2:1 IFenech told·CIA tha.t it appeared Powers had begun a descent well before the SA-2 hit. Had he gone to sleep? Was it inattention or hypoxia? Did he flame out and search for a lower altitude to restart his engines? All Fenech knew was that Fenech did not belie,re what Powers had told CIA. The CIA crowd was not amused, and Fenech I1r"i""rull~nta long and hostile grilling by the board. WhatreaUyhappened?We",illpr0l>a.bl!ne~er~now.Powers died in a helicopter crash in 1977, so no more information is available from him. Butthel I HANDLE VIA TALENT KEYHOLE COMINT CO MS JOINTLY NOT RELE OREIGN NATIONALS <"'" Ib) (1) Ib) (3) OGA 183 leaned so heavily on was suspect. Moreover, the Soviet officer who was in charge of the SAM "---------------'battery that supposedly shot Powers down stated after the end of the Cold War that the air defense operators were so shocked at the shootdown that they didn't believe it, and for twenty minutes or so they continued to reflect the aircraft on its presumed track to cover up their befuddlement.53 Ifthe Soviet defenders did not know for sure what had happened, and if they covered up information so as not to look bad up the line, the chance at ever arriving at the truth looks very dim indeed. The theory that he was downed by an SA-2 at very high altitude (68,000 feet) appears more plausible today than it did in 1960. THE ATTACK ON SOVIET CIPHER SYSTEMS L...- Ib) (1) Ib) (3) -50 USC 403 Ib) (3) -18 USC 798 Ib) (3) -P.L. 86-36 ____-----'1Baker Report, 1958 When it was created, A.FSA inherited a Soviet problem that was in miserable shape. The first was unenciphered radioprinter, which ~nformation. These links had not There were only two bright spots. carried valuablel L-"""":"'"-:--"""T"---------........- - - - - - - - - - - . . . . , yet begun to go to cipher. I Even the darkest days, however, had their rays of hope. Howard Engstrom, a World War II cryptologist now in the civilian computer business, suggested in 1950 that AFSA might make progress by establishing a research institute comprising eminent civilian scientists to attack the problem, very much in the pattern of Los Alamos of the Manhattan I1MFQbK YIATAbKNT KKYH8bl!: e8MIPfT e8NTR81:J 8Y8TI!:M8d8mTI:JY NOT REI F" S" Phi: TQ FQ~i8HIh'r'f16l(ALS 184 _ . - - - _ . _ - - - - - _ . _ - - - (1) Ib) (3) -50 USC 403 Ib) (3) -18 USC 798 Ib) (3) -P.L. 86-36 Project.54 Then, two years later, AFSA's Scientific Communications Advisor Grou (SeAG, a predecessor of NSASAB), chaired by Engstrom, given sufficient resources L...---,,...------......,,...----:......-.,.....----=---=.,.....-......,,..,,..-.......-l and a strong research and development effort. The need for a skilled civilian work force or the employment of an outside research institute was essential. AFSA did not have a strong enough civilian work force, and the Brownell Committee made this point forcefully that same year.55 I1/diBbB VIIt 'fltbBN'f l(EYIISbE eSMIN'f eSfffRSb SYS'fEMS lfSIN'fbY MO:r IU;bli: 't~ABbl'J'fe peRBISN N*'fI6n1\L~ 185 Ib) (1) Ib) (3) -50 USC 403 Ib) (3) -18 USC 798 (3) -P.L. 86-36 As NSA struggled with the problem, two camps formeq concerning the prospects for success. The first group felt/that the effort was hopeless and should not be funded. The 1957Baker Panel leaned toward this viewpoint. The committee. recommended that an effort be kept alivel I lOut it was pessif-m-l':""'s"':'tI':""'c-a"":b-o-u"":-t"":l-o-n-g--r-a-n-g-e-c":"h-a-n-c-e-s"::fo-r-s-u-c-c-e-ss-."l!'6ir--+A second group felt that the United States would never know whether it would be: possible or not because of inadequacies at NSA. /The organization was too skewed toward military manning, was not hirin the ri ht .k.inds of civilians, and did not have an adequate budget. This opinion was well entrenched at CIA/and was led by former NSAer Frank"-------' Rowlett. A variant on this interpretationwas offered by the Baker Panel, which suggested that the internal NSA structure could /hot cope with the complexities of high-grade systems. That job should be given/entirely ito a Los Alamos-style civilian research institute.62 But within NSA itself there/was a strong undercurrent of disagreement with both camps. Representative of this view was the report ofa committee chaired in 1956 by Navy captain Jack Holtwick. Holtwick felt that a concentrated attack would yield enough I lalone to justify the effort, and he recommended a massive computer attack. Such a super-high-speed computer would cost in the neighborhood of$5 miIlio.h per year, a considerable sum in those days. NSA would ~-:;====::::;7:7'==U-:=;~:":""L:'::":":"'7::-;::-::-:=-===-::;:;::'-=-=~=nee1 ~nd would probably have to have some of the work done at a private research organization (the Los Alamos option again)·1 I I r IIANBLFJ 'lb\ ~""LFJN~U:flYII6Ltl e6MIn'f e6U'fft6L ST[S'ftlMS tf6IIffbY NOT REI Ii: Ai,) il:r..g 'l'Q F8aBI8N HA'f16H}[LS 186 (b)(1) (b)(3)-50 USC 403 (b)(3)-18 USC 798 (b)(3)-P.L. 86-36 TRACKING SUBMARINES - THE STORY OF BURST TRANSMISSIONS Late in World War II, German scientists had once again come up with a serious threat to Allied cryptologic efforts. This time, they had devised a way to compact lengthy manual Morse messages into messages lasting only a few seconds. When played at normal speed, a message sounded like a burst of noise in the receiver. The Germans called it KURIER and intended it to be on submarines, agents (spies), and eventually aircraft for low-probabilityof-intercept communications. Early models were deployed before war's end, and GCCS intercepted transmissions on at least one occasion. Fortunately, however, KURIER was still in the experimental stage. When the war ended, a German submarine surrendered in Argentina, the nearest landfall. Aboard the sub was a German scientist with extensive engineering notes and knowledge of the system, and he was willing to talk to the Americans about it. Even luckier, the British captured an actual KURIER system, and both the British and Americans experimented with it, primarily for the purpose of building burst systems for their own submarines.65 Unfortunately, the Soviets also captured German scientists working on KURIER, and the TICOM teams discovered this during their debriefing sessions. At the time, the Navy I1AP'fQbK VIt. TAbKP'f'F K8YIIeLtJ eSMHff e6N'fR6L S'lS'ftJMSofSnnLY NOT RE' E A i A 2I,.K 'Fe FeR8I8Pi PiA'fI6NkLS 187 JDPSE~REI~ L -------I - ( 1) b (3) b (3) (b (3) 50 USC 403 18 USC 798 P.L. 86 36 1 (1) Ib) (3)-50 USC 4 Ib) (3) -18 USC 7 Ib) (3) -P.L. 86- -------/ Ib) (3) -P.L. 86-36 Ib) (1) Ib) (3) -50 USC 403 Ib) (3) -P.L. 86-36 Notes 1.1 ~Soviet Manual Systems Since 1945: A History of Their Cryptography, Usage and Cryptanalytic Exploitation"; unpublished draft available in CCH. 2. Ibid; see also Frank Rowlett, "Recollections ofWork on Russian," unpublished manuscript dated 11 Feb.\l965, available in CCH; see also NARA, SRH·OOl, 296. 3. RobertL. Benson and Cecil Phillips, History ofVerwna, published in March 1995. The name "KGB" will be used throughout this book to refer to the Soviet intelligence organization and its predecessors, the MVDand NKVD. 4.1 ~Before BOURBON: American and British eOMINT Efforts against Russia and the Soviet Union Before 1945," Cryptologic Quarterly, FalllWinter 1993, 1-20; Benson and Phillips, Venona; Frank Rowlett, "The Story ofMagic," Ch. VII, 53; manuscript available in eCHo 5. Rowlett, VII.53.; Oliver R.Kirby, "The Origins of the Soviet Problem: A Personal View," Cryptologic Quarterly. Vol II, No.4, Winter 92;; Louis W. Tordella1 series oforal history interviews beginning 28 June 1990 by Robert Farley, Charles Baker, TomJohnson and others, NSA OH 8-90. 6. Rowlett, "Recollections ... "; see also Oliyer R. Kirby oral interview, 11 June 1993, by Charles Baker, Guy Vanderpoolt land David Hatcb~NSAOH 20-93. Tordella interview. 7. Howe, "Narrative History ofAFSNNSA, Part I"~ ~EarlYBOURBON," 1994. 9. Benson and Phillips, Venona. 10. Benson and Phillips, Venona; Kirby, "The Origins ofthe Soviet Problem." 11. Robert T. Lamphere, and T.Schachtman, The FBI· KGB War, a Special Agent's Story (New York: Randon House, 1986. 12. Benson and Phillips, Venona. 13. Lamphere and Schachtman, 78. 14. For information on the Petsamo Incident and the Stella Polaris project, refer to the following: Lamphere and Schachtman (not the best source); interview with Hallamaa in Madrid in 1951, in NSNCSS Archives, ACC 7975N, CBRJ 22; Benson and Phillips, Venona, Stella Polaris document collection in NSAlCSS Archives, ACC 1177-79,11369-76,12504, 19043N, 19044,CBRJ 23. 15. Benson and Phillips, Venona. 16. Ibid. 17. See David Martin, Wilderness ofMirrors (New York: Ballantine Books, 1980). 18. Brownell Report, 106-08, in CCH Series V.F.7.13. n*I(6L~ V1* 'f*L~I('l'I<:~YIfeLl!: eeMU('I' e6N'fft6L S'IS'ffiH,lS delN'fLY NOT REI$ASABI E TO FOgIi:IQ~r ~W"iIeUA)§S 190 .Jf'lb) (3) -P.L. 86-36 . .... ,." ~~~ 19. No history ofBlack Friday was compiledatthe time, partly beca1lSeofthe fragmented natllre ofcryptology in those days. The best versions ha,ve 6nly recently been cOlI)piled:f l"Beyond BOURBON - 1948. The Fourth Ye~rgfAlliedCollaborative COMI~TEffortAgainst the SovietUniQ Il,"Cryptologic Quarterly, Spring 1995;1lJ:ldBenson and Phillips, Veno~,Forlidditionalinformation,aee ore.linterview withl I I JOMay 1985 byl ~nd Rpbert Farley, NSA OH 0;Jo"85; oralhistory interview with Cecil J. Phillips, 8 July 1993, by Charles Baker and Tom Johnson, NSA OH 23.93; ore.1Mstory interview with Herbert L. Conley, 5 March 1984, by Robert Farley, NSA OH 01-84; and Oliver R.Kitby, "The Origins of the Soviet Problem...." 20. Details of the early Soviet program can be found i~ ffJ'EarlY History of the Soviet Missile Pro",~(l945-1953)"S"'''''~'V(S=m" 1975), 12-~ ~:~Vi.tL=d- B~dBal"'''' MiaaU, Program, 1945-1972: An Historical Overview," unpublished manuscriptavailable in CCH. 21. he Soviet Land-Based Ballistic Missile Program...• " 39. O'Neill, 279. 38. Kirby interview. 6 May 1994.26. Oral interview with Ray Potts an 29. Walter Laqueur, A World ofSecrets: The Uses and Limits ofIntelligence. (New York: Basic Books, Inc., Publishers),I43-44. 30.DThe Soviet Land-Based Ballistic Missile Program...." 31. Tevis interview. 32. Amato interview. 33~ ~HiStoryofthe Soviet Nuclear Weapons Program," intelligence report available in CCH. 40. Background papers for the 1967 Eaton Committee, available in CCH; oral interview with Milton Zaslow, 14 May 1993, by Charles Baker and Guy Vanderpool, NSA OH 17-93. (1) Ib) (3) -50 USC 403 Ib) (3) -P.L. 86-36 II*NBbB VIA '¥ltbBN'f ItE)YIISbB 8SMIn'f 8Sff'fftSL S'IS'fBMSdSIN'fbY ~TQT REbEA&'tBbE Te FSRBI8N NA'fISfUtbS 191 41. Ray S. Cline, The CIA under Reagan, Bush and Casey (Washington, D.C.: Acropolis Books, 1981), 200-201; "Report of the Secretary's Ad Hoc Committee on COMINT/COMSEC," June 1958 (Robertson Committee), CCH Series VI.C.1.11.; "Tibetan Revolt of 1959," informal paper prepared for Eaton Committee in 1967, available in eCHo 42. Burrows, Deep Black, 62-3. 43. Burrows, Deep Black, 67; Michael R. Beschloss, Mayooy: Eisenhower, Khrushchev and the U-2 Affair (New York: Harper and Row, 1986), 77-79; Oral interview with Henry R. Fenech, 30 Sep 1981, by Robert Farley, NSA OH8-81. 44. Stephen A. Ambrose, Eisenhower, Volume 2: The President (New York: Simon and Schuster, 1984),227-28. 45. Burrows, Deep Black; NSNCSS Archives, ACC 24355, CBOH 36.; Ambrose, Eisenhower, 309-10. 46. Ambrose, Eisenhower, 265. 53. Vox Topics, V. 3, # 3,1992. I54. CIA-AFSA collaboration (Wenger file), in ACC 9142, eBm 27. 55. Collins, V. II, 6; Brownell Report. 61. Baker Panel report. 62. Collins, V. II, 16,23. 63. Holtwick. Ib) (1) Ib) (3) -P.L. 86-36 Ib) (3) -50 USC 403 Ib) (3) -18 USC 798 IIMlBhB VIA 'fltb6Pf'f IC6YIIahEl 8aMIN'f 8aH''fft:8h S';S'fElMSJ8n,'fLY TOPSECRET~ ne'J' RI!lMl/rSJrBbB '¥e FeRBI6f, NoA'¥IenAhS 192 (1) Ib) (3) -P.L. 86-36 Ib) (3) -50 USC 403 Ib) (3) -18 USC 798 .' Ib) (3) -P.L. 86-36 ~~RA 65. NSNCSS Archives, ACC 3838, CHOH 11; O'Ro,l,U'k~~~~rview;oral interview Wi~ 117July 1986, by Robert Farley and Tom Johnson, NSAOH 19-86:; NSAlCSS Archives, ACC 3838, CBOH llL. 66. CCHSeries V.R.1.7; V.B.2.7.// 671 ./../ ./ IRadio Direction Finding in the U.S. Navy: the First Fifty Years," paper available in CCH; NSA/,CSS Archives, ACC 3838, CBOH 11. 68~ 1// 69~~===::::;:t:H:iS'":to';"'ry ofHFDF in the Pacific Ocean Prior to the Advent of Bullseye," 1981, in CCH Series VII 85. 1I1tNBbrJ VIA 'fALrJU'f RElYI16LEl e6MIWf e6H"fft6L Sis'fI!:MS J6Uif'fL i ~O'f IU:U \~A8b~ 'Fe F'8RElI8H HA'f16!41rLS 193 Chapter 5 Building the Internal Mechanism CRYPTOLOGY IS AUTOMATEDTHE STORY OF EARLY COMPUTERIZATION The trouble with machines is people. Edward R. Murrow, 1952 Antecedents Modern cryptanalysis, with its emphasis on the manipulation of large amounts of data, was one ofthe earliest government enterprises to acquire the new office automation equipment being produced by a small company called International Business Machines (IBM). In 1931, OP-20-G obtained some of the new IBM machines and quickly employed them in the cryptanalytic process to sort large amounts of data and determine likes and unlik.es. In 1935 Signal Intelligence Service (SIS) acquired the same type of equipment for the same purpose. By World War II "EAM" (electronic accounting machine) equipment had become commonplace in COMINT processing, and it contributed mightily to codebreaking, especially in the Pacific Theater. By the end of the war, OP-20-G and SIS combined were using more than 700 IBM-type machines.1 Of the two, the Navy seemed to be further along. During the 1930s and into the early war years, OP-20-G had attempted a partnership with Vannevar Bush, the renowned MIT (Massachusetts Institute ofTechnology) scientist, to build a faster comparator for analytic (read cryptanalytic) use. This rather bumpy relationship had so far yielded a number of notable technological and administrative failures when, in 1943, OP-20-G became a partner with GCCS in running attacks against the four-rotor German naval ENIGMA. They ultimately decided on a huge, clunky mechanical marvel which has been dubbed the "American bombe." A technological dinosaur when compared to the devices Bush was experimenting with, the bombe at least worked and was used in the last two and a half years of the war to break German naval ENIGMA keys. The Navy development and contract monitoring operation was called the Naval Computing Machine Laboratory (NCML); it was located on the grounds of National Cash Register in Dayton, Ohio, the prime production contractor.2 IlAUBhJ!J VIA 'l'ALJ!JU'f ltJ!JY1I6hJ!J e6MlN'f e6f,'l'R6h SYS'l'J!JMS d6IH'fhY NOT REI EA i A abi: '];9 f9IU;IQ}1 }T·NflEHlAb6 195 The American Navy Bombe ANavy WAVE checks rotor settings during World War D. Although a very fast comparator, the bombe was not a true computer. It did not have a stored digital program which could be modified. But even as the Navy designed and built the bombes, the British were moving ahead into the era of true computers. To attack systems even more complex than ENIGMA, GeeS was developing a computer which IIANBLI!: \'IA 'fALI!:N'f KI!lYII6LI!l €6MHff €6N'fft8b SYS'f'ElMS iJ8IN'fbY 196 employed an electronically generated key that was compared with the German cipher text. Although it did not have a true internally stored program, the settings were operatoradjustable according to how close he or she thought they were to a cryptanalytic solution. They called it Colossus. Some contend that it was the world's first true computer, although Colossus must compete for that honor with ENIAC, which was being developed at the University of Pennsylvania's Moore School of Electronics to generate complex artillery ballistics tables for the Army. Either Colossus, designed for cryptologic use, or ENIAC, for ballistics, probably deserves the title ofthe world's first computer.3 Postwar Developments OP-20-G could see the technological possibilities in the bombe, and it was decided even before the war ended that the effort should continue. But National Cash Register had no intention of continuing the association. They wanted to return to making cash registers. So at the end of the war, NCML was physically evicted, along with the remainder of its undelivered bombes, and the project came to a halt.' OP-20-G needed a prime contractor with which to work. Months before the war ended, Howard Engstrom, a key figure on the bombe project, decided to start a new company specifically to do business with OP-20G. At war's end, he left the Navy and took with him the best and brightest technicians at NCML. They set up a new company called Electronic Research Associates (ERA), under the wing of an already established firm called Northwestern Aeronautical Corporation in St. Paul, Minnesota. The Navy made no specific promises regarding contracts for the fledgling company, but none were needed. Engstrom and associates had a corner on the technological expertise that OP- 20-G required, and contracts flowed almost immediately.s Howard Engstrom 1I1d'Bbl!l 'iIA 'fAbI!Jli'f IlI!JYIIebI!J eeMIN'f eeN'ffteb SYS'fI!Jf9ISdeIN'fbY NOT REI E AS ABI F 'to i'QRI!JI8H n-A:'f16f4'ALS 197 Ib) (3) -P.L. 86-36 The relationship between ERA and the Navy was emblematic ofthe way relationships had developed between the cryptologists and private industry. During the war OP-20-G had developed a close relationship with IBM, Eastman Kodak, and National Cash Register. SIS had a similar kind of relationship with Bell Laboratories and Teletype Corporation. Those businesses kept a stable of cleared people who could do jobs quickly and quietly for the cryptologists. In the COMINT and COMSEC businesses, it did not pay to advertise.6 Both the bombe and ENIAC had been developed through classified wartime military contracts. Thus computing in the United States began in the rarified atmosphere of tight security. Though the cryptanalytic aspects were not publicized, the Army relationship with the Moore School became a matter of public knowledge in 1946 when the inventors of ENIAC, John Mauchly and J. Presper Eckert, gave a series of lectures on electronic computers. As the two men left the Moore School to establish a computer manufacturing company, they dispersed their knowledge nationwide in what became known as the Moore School Lectures. Many felt that this lecture series launched the computer industry in the United States.7 Howard Engstrom had found out about the Moore School/Lectures, and he suggested that the Navy send a cryptologist to observe. Thus, when theilectures began, sitting in the back of the room was Lieutenant Commander James T. Pendergrass, a Navy mathematician employed at Nebraska Avenue. Pendergrass delivered a report to the Navy on the Moore lectures which focused attention on the emerging new computer technology. This resulted in negotiations with ERA whiCh led to the construction of the Atlas machine.8 Like the bombe before it, the first generation of postwar cryptologic computers produced highly specialized machines, called in those days "rapid analytic machines" (RAMs). Each machine was constructed for a different purpose and attacked a different cryptanalytic machine or problem. Programs were /particular rather than general, and inputs and outputs were of specialized design. A list of AFSA machines, both present and projected, in 1952 contained sixty RAMs, as opposed to only eight that had more flexible objectives.9 An example of a RAM wa~ rhich was developed by ERA to attack I to Even in those early days computer companies were willing to take on difficult developmental tasks. For instance, ()perating under a 1947 contract, ERA developed the world's first magnetic drum storage system~spart of a RAM project called GOLDBERG.ll A successor project, called ATLAS (also built by ERA), (ipplied the drum storage technology to a more general purpose cryptanalytic processor. ATL.A.Sw(iS ERA's first major computer development, and it led to the company's first commercialprQduct, the ERA 1101, produced after the company had become merged with Remington-Rand~Univacto form the first major American computer company.12 Ib) (1) Ib) (3) -50 USC 403 Ib) (3) -18 USC 798 Ib) (3) -P.L. 86-36 IIANBhe VIA 'Hl:heN'f lEe {f I6he e6MHf'f e6H'fft6h S?S'feMS 66Hf'fL? NOT REI E A i A Qbg ~9 F9RI"JI6N lifATIONALS 198 Atlas! While NSG forged ahead, ASA was trying to catch up. In ASA, the role played by Engstrom, Tordella, and Pendergrass was at first taken on single-handedly by Samuel Snyder, one ofFriedman's most talented prewar cryptanalysts. Snyder's 1947 paper "Proposed Long-Range Cryptanalytic Machines Program for Literal Systems" played a seminal role in ASA's first postwar venture into the new technology. In it, Snyder proposed that ASA develop its own analytic computer based on extensive research into existing technology. Snyder himself did most of this early research, drawing at first on information provided by Pendergrass and Howard Campaigne of NSG. He made pilgrimages to the ,fountainheads of computer research: Aberdeen Proving Grounds to see ENIAC, Bell Labs to see its Relay Computer, IBM to see UANB~BlilA 'flt~ElN'fI[ElYUehEl eeMIN'f eer,'ffteh SYS'fl!llVIS Jen,'fUi NOT RE' E A ~ A J!lU: 'PQ FQRBI8H H*'fI6f(AL~ 199 the IBM Selective Sequence Calculator, and MIT to see its Differential Analyzer. He attended a lecture series at the National Bureau of Standards (NBS) which concentrated on Univac products (Univac had been formed in 1946 by Mauchly and Eckert), Raytheon computers, and the Ace Computer (one of the earliest British entries into the commercial computer field). Snyder suggested that ASA team up with NBS, which already had some expertise in the field, and he proposed that ASA form a committee to guide the effort.13 ASA decided to go ahead with development of a general-purpose analytic computer called ABNER. Working through NBS, ASA arranged for subcontracts on mercury delay memory anq for magnetic tape drives from Technitrol and Raytheon, respectively. Snyder contended that ABNER I, which was released for use in 1952, was the first machine that placed primary emphasis on nonarithmetic operations. Although it played a role in the development of later computers for cryptologic applications, one expert in in the field called Abner "barely functionaL" This was an appellation that could have applied to many ofthe early experiments in machine-age cryptology.14 The early cryptologic computers were troglodytic. They were physically programmed in binary instructions input via paper tape. They used octal numbers and words twentyfour bits long. There was no "computer language" as such. Memories were tiny by today's standards - the drum memory for ATLAS, for instance, held only 16,000 words. There being no more advanced technology available, vacuum tubes were used for relays, despite the obvious disadvantages this created in terms of heat buildup and tube replacement. Early computers were usually "down" more often than they were "up." When they were "up," though, they provided answers faster than anything imaginable.15 Vacuum tubes were on the way out, to be replaced by transistors, developed at Bell Labs in the 1940s by future Nobel prizewinner William Shockley and others. NSA scientists were among the first to apply the new transistor technology to computers, and in the mid-1950s it developed an in-house computer called SOLO, the world's first computer to be entirely transistorized. SOLO was subsequently marketed commercially by the contractor, Philco, as the Transac S-1000.16 Other innovations were on the way. In the mid-1950s NSA began making the transition from centralized computer operations to remote job access systems. The first remotejob access computer, ROGUE (for Remotely Operated General Use Equipment), used hardware called Alwac HIE developed by a small firm called Logistics Research, Incorporated. ROGUE had three remote terminalsconnected to a small central processor.17 IIAPfBbEi VIA 'fAbSU'f KSYIISbS eSMIN'f eON'FROb SYS'FEiMS dOIN'FbY NOT REI EASAijI F TO i'OREIQl'i ~TA'fI8U-A:bS 200 SOLO RAMs like ROGUE were good for specific jobs, but cryptologists recognized very early that they would require more generalized systems to process very large volumes ofdata. A study in the mid-1950s depicted just how much material must be massaged. Raw traffic arrived in courier shipments every day at the rate of thirty-seven tons per month. An additional thirty million groups of traffic arrived (in Tecsumized form) via teletype. Traffic from some entities (particularly the mechanization-resistant manual Morse intercept) received less than 50 percent detailed processing - the rest was held in case it was needed.is As early as 1946, NSG began the search for a computer that could hold very large volumes of data. Studies of mass data handling methods led to a contract between the Navy and Raytheon in 1951 to develop and produce a machine called NOMAD that would be physically and financially the largest cryptologic machine yet. But the NOMAD contract went badly offschedule from the first, and the contract was killed in June 1954.19 IIAf(Bbl!J ¥IA 't'AUfff I(I!JYII8bI!J 08MU('t' 08N''t'ft8b &YS't'I!JMB68Hf't'bY NOT REI.EASABI E TO VQHKISn nA't'18nA:L~ 201 The best general-purpose computer in the early days was an IBM product, the 701, designed in partnership with NSA. NSA leaned toward magnetic tape rather than disks, and the 701 had the first truly functional tape drives controlled by vacuum columns. The 701 was followed by the IBM 705, which became the mainstay for general-purpose computing. Coming on line in the mid-1950s, the 705 was a nonfixed-word-Iength machine. It had the best sorter around, an assembler (called a "transembler") that mimicked punched card machines. The 705 had a major impact on data processing, and it made it possible to begin processing massive volumes of data rolling in from the rapidly expanding network ofcollection sites around the world.20 Parallel to the general-purpose processors was a line of special-purpose scientific machines. Notable was the IBM 704, which had a 36-bit word, punched card input, and tape drives for storage.21 Cryptology still needed a general-purpose system.· A committee, formed to review the demise ofNOMAD, specified the requirement for a system that could be ofuse to both traffic analysts and cryptanalysts. For the traffic analyst, it would have to have large storage, have a file capability for collateral information, and be ca able of sorting quickly. For the cr tanal st it should be able to tackle To achieve the requisite flexibility, the system would require a general-purpose mainframe with special-purpose peripherals. The project was called FARMER. 22 At the time, IBM was working on a project to extend the performance of its latest product, the 704, by a factor of 100. They called it STRETCH. IBM approached both NSA and the Atomic Energy Commission (AEC), the two government agencies that it felt would have the most use for such a system. AEC agreed to proceed, but NSA ultimately decided that it wanted something specifically optimized for cryptologic applications. However, IBM was on the right track, NSA concluded, and awarded Big Blue contracts for research in high-speed memory (SILO) and to design a general processing system for Agency use (PLANTATION, later called RANCHO).23 The entire project was eventually folded into a gigantic effort to develop a large-scale computer. It was called HARVEST. The most difficult part of the project turned out to be designing the magnetic tape drives. Under a project called TRACTOR, IBM developed new tape drives and a unique automatic cartridge loading system having 100 times the speed of the IBM Type 721 tape drives then in use. Each of the three TRACTOR units managed two tape drives, and it automatically retrieved and hung data tapes in a robotic environment that was the wonder ofthe U.S. government. It made for great theater and was on the mandatory show-and-tell tour for years. Ib) (1) Ib) (3) -P.L. 86-36 Ib) (3) -50 USC 403 Ib) (3) -18 USC 798 NOT REI E ASA BT E TO FOIUiiIQ~HfA'PI8rHthB 202 HARVEST HARVEST tractor units IIAUBLI'J VIA 'fALI'JU'f I(I'JYlI6LI'J e6Mm'f e6If'ffteL SYS'fEMSJ6n,'fLY 203 HARVEST worked after a fashion and remained the Agency's central processor from the time it went on line in 1962 until it was finally retired in 1976, a phenomenal life span for any computer system. But those who had to make it work remember it as balky, difficult to program, and not performing anywhere near the specifications that had been set for the system. It was a transitional machine.24 NSA's most lasting contribution to computer history was undoubtedly Project LIGHTNING. LIGHTNING resulted from NSA's reaction to outside criticism that it could not brea~ Isystems and to proposals that this part of COMINT be transferred to an outside research organization. Pricked b the criticism, Canine initiated an all-out attack As part of the project, Canine proposed that NSA develop aL.- ...... computer that would advance the state of technology by three orders of magnitude. He decreed that the goal was a "1,000 megaHertz machine," and at a USCIB meeting in August of 1956 he requested $25 million seed money. The sum angered the Defense Department and placed NSA's budgetin jeopardy. In order to get it approved, General Samford took his case directly to President Eisenhower and his top scientists, Vannevar Bush and Jerome Wiesner. Eisenhower came down hard in favor, and he authorized the use ofhis name to push the project ahead.25 Three major contractors participated - IBM, RCA, and Sperry Rand Univac - but Ohi,o State University, Kansas University! Philco, and MIT also performed lesser roles. LIGHTNING never resulted in a computer, but the research teams turned up information that drove the next generation of commercial machin.es. Among the most significant findings were in the field of cryogenics. IBM's Dudley Buck developed the cryotron,.and through his research IBM proved the now-obvious axiom that the lower the temperature, the faster the computer. Sperry Rand Univac concentrated on thin magnetic field devices and, through these early experiments in chip technology, found that computer speed would increase when components were subminiaturized in order to place them closer together. RCA concentrated on applications of the tunnel diode, one of the fastest switching devices known.26 As the 1950s wore on, cryptologists broadened computer applications to include far more than just cryptanalysis. NSA first used computers to generate COMSEC material in 1959, when the COMSEC organization began employing the Univac File Computer for that purpose. And for the processing of intercepted traffic for traffic analytic applications, the IBM 700-series computers continued to be the mainstay. Ib) (1) Ib) (3) -50 USC 403 Ib) (3) -18 USC 798 Ib) (3) -P.L. 86-36 IIAN'BLE VI:A: 'f2\LEfff ItE'i1I6LE e6MHof'f e6fof'fft6L Si STEMS J61fof'fLl NOT REI F A ~ A 1IJ"l'J q:S PSRBI6U Iof2\TIONALS 204 Ib) (1) Ib) (3) -50 USC 403 Ib) (3) -P.L. 86-36 ·1 I__~ ---,I For scanning in the field, CDC (Control Data Corporation, the successor to ERA) developed the concept of key word search under a project called FADE-IN. Text scanning in the field was not implemented until the early 1960s. SOAPFLAKES was believed to be the first key word scan system at NSA.27 Most of these processes were off-line. Intercepted traffic in Tecsumized or diarized form spilled off the communications lines in paper tape form and was carted off to another area of the building to be input to processing computers. But it was not the wave of the future. SMAC first began experimenting with the use of a computer to directly receive inputted messages from the field, and so avoid the paper tape step. This effort used Univac 494s and was in a very early stage ofdevelopment as the 1950s came to a close. 28 NSA COMMUNICATIONS IN THE PRE·CRITICOMM ERA Equipment is obsolescent, insufficient in number and inadequate for the purpose. . .. Such essentials to operations as, for example, a place to put live traffic and operators' logs, are neglected in the installation and are provided, if at all, as an afterthought when operations begin.... Homemade bins in the aisles, traffic piled on the floor or clipped to overhead wires like clothes on a line, logsheets resting on machines, et cetera, are the inevitable result. 1955 study ofthe COMINT communications system Rapid communications is the lifeblood of SIGINT. Cryptologists have grown so accustomed to virtually instantaneous access to remote corners ofthe globe that they could not operate any other way. But in the early days, they operated in a decidedly different mode. AFSA, when created, had no indigenous communications at all. Instead, the organization depended entirely on communications paths and facilities provided by the services. COMINT passed from collection sites to Washington on armed service communications. It was encrypted off-line at the field site, then was passed to a local communications center manned by non-SI indoctrinated people, who put it on common user circuits for transmission. Ifthe traffic originated at a Navy site, it was put onto naval communications; if it was an Army site, it went via Army communications; and so forth. The traffic was long, vertical umbilical, service-unique and electrically sealed until it reached Washington, where the information could then be passed to other services or to AFSA. II*N5LI!: .1* T*LI!:H'T KI!: i II~LI!: e~MI!.T e~HTIt~L SY8TI!:MS J~IHTLY NOT REI E A S A In.,). '];'Q FQREI8Pf N'k'fIfm1\LS 205 The message went via HF single sideband, passing through up to six relay centers before finally arriving at either Arlington Hall or Naval Security Station. It might have to be reencrypted up to five times, and the process required from twenty-four to forty-eight hours to send a routine message to the capital. Because of the many relays and inherent degradation of HF channels, up to 30 percent arrived undecipherable and had to be retransmitted. Messages required several hours for decryption, and the handling time for each message, including marking and routing to the intended recipient, took several more hours. The ASA communications center at Arlington Hall, for instance, was taking approximately fOUf days (on top of the one to two days of transmission time) to deliver a routine message. The fastest possible handling time on the most critical information was not less than five to six hours from time ofintercept, according to information furnished to the Robertson Committee in 1953.29 When AFSA came into existence, the communications system on which it relied was reported to be "in a deplorable and deteriorating state." Arthur Enderlin, one of AFSA's top communications people, conducted a study detailing the decrepit conditions and sent it to Admiral Stone. A disbelieving Stone decreed a full-blown study, which just confirmed Enderlin's contentions.30 Nothing was done under Stone. But when Canine arrived, plans were immediately laid by Enderlin's successor, Lieutenant Colonel William B. Campbell, for a separate AFSA communications center to process traffic destined for AFSA organizations. In July 1952, the new communications handling facility opened in B Building at Arlington Hall, using Teletype Corporation Model-19s. This was a good first step, and it reduced the message Arthur Enderlin handling time for routine messages to One ofNSA's communications pioneers, he helped three hours, while cutting the message develop the system throughout the 19508 and 1960s. backlog to almost nothing.31 IIIdof5LI!J • fA 'fALl!Jlof'f f~1!J i II6LI!J e6MI!of'f e6!of'fft6L S-YS'fI!JMS d'6Hof'fJ:S NOT REJ.EASABJ E TO i'QRiJIElN ?iA't'16Iof*L~ 206 - - - _ . - _ . _ - - - - - - - - - - - - - - _ .. - - - - - - - - - Another Canine push was for secure telephone communications. The NSA gray phone system, formally known as NSNCSS Secure Telephone System (NSTS), began in AFSA's waning days with a total of 200 lines, 100 each for Nebraska Avenue and Arlington Hall. AFSA took possession of two-thirds of the telephone instruments, while the collocated SCAs got one-third. A month later (September) a new microwave system became operational between the two locations, ushering in an era of high-reliability, high-fidelity communications. At the time, the system required an operator to connect the two parties, just like commercial telephone circuits ofthe era. The following April NSA issued its first consolidated telephone directory.32 AFSA began broadening its secure communications contacts with its customers. The Zone of Interior Connectivity (ZICON) net, originated in the early 1950s, consisted of landline communications paths between AFSA and its principal customers: the three services, State, and CIA. Later, the National Intelligence Indications Center in the Pentagon was added, as well as SAC and CONAD (Strategic Air Command and Continental Air Defense Command) for the Air Force.33 The COMINT Comnet By 1952 it was already clear that the growing volume of cryptologic communications would not permit newly established NSA to pursue the old way ofdoing business. Already, the daily group count was considerable and would grow in the ensuing years, as the following table shows. Table 1 Total Mean Daily Average Group Count at NSA34 Year Count 1952 648,000 1953 1,247,117 1954 1,322,552 1955 1,320,073 1956 1,227,158 1957 1,424,351 1958 1,729,430 1959 2,059,763 1960 2,615,377 1961 3,896,211 1962 4,306,910 1963 5,089,777 1964 6,134,601 Il:AUBL8 VIA 'f:AL8NT Il8YIleL8 eeMIN'f eeN'fReL 8'[S'f8MS ·?B'aIL 68. Ibid. 69. Ibid. 70. Ibid. 71.1bid. 72. Ibid. Ib) (3) -P.L. 86-36 IIANBb8 '11ft 'Wfb8N'f I(I~lYIISb8GSMH'f'f GS~l'fRSbSYS'f8MSdSH'f'fb! NOT REI EASABI E TO ¥OR818N N'A:"IO!'4'ALS 225 Chapter 6 Cryptology at Mid-decade THE EARLY ASSESSMENTS Ib) (1) Ib) (3) -50 USC 403 Ib) (3) -18 USC 798 Ib) (3) -P.L. 86-36 It has become exceedingly difficult to obtain significantinformation from coV'ettoperations inside Russia. The security zones at the border, the general restrictions in the interior, the thousands of security police, and the innumerable.informers among the populace.are brutally effective in limiting infiltration, exfiltration, and usefulness ofagents. Therefore, we must more and more depend on science and technology to assist and to complement the·. best efforts of classical intelligence. The KillianBoard,1955 The Eisenhower administration's intelligence focus was not on traditiOnal espionage it was on technical intelligence, whence, Eisenhower knew through personal experience during World War II, he couldobtain vast quantities of information. His concern over the apparent breakdown in COMINT during the Korean War caused him to refocus again and again on NSA. Reportsabout NSA's performance began to flow back to hUn almost from the moment the Agency was created. The reports are important toclaybeca1.l.se they indicate the direction that cryptology was to travel in subsequent years. The Robertson Committee The first reports on NSA were a product ofPresident Eisenhower's concern with Soviet I-----.,fapabilities. In the summer of 1953, the National Security Council began examining America's strategic vulnerabilities, and,with it, the intelligence system that must provide the warning. But Canine adamantly opposed granting COMINT. clearances to the members of the panel, and USCIB backed him. Instead, Canine established a largely in-house examination of COMINT, chaired by Dr. H.. P. Robertson of CaliforniaInstitute of Technology, a member of Canine's advisory panel, the NSA Scientific Advisory Board (NSASAB). Four of the seven members were from NSASAB, and the remaining two were from the Office ofthe Secretary of Defense.1 Robertson reported during the dark days after "Black Friday," whenSovi.etl was still an unrevealed mystery. I II The immediate result of this was the intercept, in 1954, I f This opened up a new world I UA:NBbEl VIA 'f¥tLElN'f !ESYII6LEl e8MIH'f e8N'fR8J:: S"l"S'ffJMS48H,'fJ::Y NOT REJ.EASABI E 1'0 ~QRi:I~H'fmr'fI8N:ALS 227 Ib) (1) Ib) (3) -50 USC 403 Ib) (3) -18 USC 798 Ib) (3) -P.L. 86-36 I ITh~ committee also recognized/the indivisability of COMINT\and ELINT and stressed the effort to fuse both sources into a/single report.Z But Robertson made plain that NSAmust be in the game for the long pull. The long pull was Sovie~ land he urged an all-out attack on the new! I systems introduced in the early 1950s. His committee recommended the development and deployment ofnew interceptl ~quipment.s The Hoover Commission The Hoover Commission was a far larger effort. Established by Eisenhower in 1954 and chaired by former president Herbert Hoover, it was at the time the most thorough reexamination of the federal government ever attempted. Hoover subcommittees delved into every cranny of the bureaucracy seeking improvements and economies. One such subcommittee was a task force chaired by General Mark Clark to investigate intelligence activities. The committee looked closely at NSA." The thrust of the Hoover Commission set the mold for all subsequent panels. Responding to the entreaties of Canine, it recommended increased authority for NSA in virtually every area of its operation. NSA should have the authority to prescribe equipment standards; it should prescribe all intercept and processing standards; it should inspect service cryptologic training and direct modifications as necessary. There was almost no area in which it did not feel that NSA should be further empowered.5 What the panel did for NSA it also recommended for the SCAs. They should have more authority within their respective services, and each should be at the level of a major command. At the time only USAFSS was at that level, although ASA was granted major command status before the report was published. This left only NSG at a lower level within its service. It noted that "largely because of its status as a major command, the AFSS has developed a dynamic and promising program for recruiting, developing and holding on to technically qualified military career personnel.,,6 The committee noted the dismal record of the three services in assigning people to cryptologic posts, and it recommended that security strictures be changed to permit military personnel offices to understand the importance ofthe jobs.7 More controversial was the panel's recommendation that NSA acquire additional authority over ELINT. Canine, who saw himself teetering over the black hole of interservice fighting, opposed this. He was having enough trouble unifying COMINT, without trying to swallow ELINT whole. USCIB noted that NSCID 17 had just been issued, and it urged that this new approach be tried before considering further integration of ELINT. (The impact of NSCID 17 will be discussed in chapter 7.)8 Clark and his committee proposed an all-out attack on Soviet high-grade ciphers, equivalent, in their words, to the Manhattan Project. It would require the best minds in 228 Ib) (1) Ib) (3) -P.L. 86-36 Ib) (3) -50 USC 403 Ib) (3) -18 USC 798 the country, equipped with the finest resources money could buy, but it would be worth it ifeven a portion ofthe Sovie~ Isystems were unlocked. Canine hailed the potential resource augmentation with glee, but cautioned against a total commitment before NSA had thoroughly analyzed the prospects for successs. USCIB supported him.9 Cryptologic personnel requirements weighed heavily on the committee. Clark urged an improved grade structure, including the addition of supergrades, higher pay for consultants, improved assignment of service officer personnel, better perquisites for NSA people assigned overseas (to be the equivalent of those received by CIA), and NSA exemption from the Classification Act. To improve the revolving door nature of military intercept operators (few ofwhom stayed in the service past their initial enlistment), Clark urged the assignment ofcivilians to intercept positions overseas.10 Clark andhis committee were concerned about two other potential problems. The first was the state of COMINT requirements, which were expressed in a document called the Master Requirements List. This, they said, was about the size of the Washington phone directory,/and about as specific. And since customers wanted COMINT to tell them everything, without narrowing the target further, NSA simply specified its own requirements. This had been going on so long that there was danger that the cryptologic community would become completely isolated from its customers and insensitive to them.11 What was occurring in requirements, they felt, was also true in security. COMINT security had become so tight that cryptologists were isolated from their customers. In time of war there was real danger that essential information would not get to the battlefield because of clearance restrictions. Thus the system would defeat itself and become a vestigial appendage.12 It was a debate that would rage for years within the intelligence community. The Killian Board Eisenhower's preoccupation with the Soviet nuclear threat spawned a number of committees to look at American vulnerability. By far the most important of those was the Scientific Advisory Committee, commonly known as the Killian Board. In July of 1954 Eisenhower asked Dr. James R. Killian of MIT to head a study of the country's capability to warn ofsurprise attack. Killian named a panel of the elite from academia, the scientific community, and the military. IffdiBbEl YIlt 'i'/rbBU'f RBYI f6LB e6MIN'f e6N'fft6L S"iS'fBTtfS461N'fLY NOT REI E A ~ A AU '1'9 F8RBI8N ItA:'fI6I(*LS 229 Dr. James Killian, shown here with General Samford at NSA The committee quickly came to the conclusion that spying on the Soviet Union in the classical sense (agents and that sort of thing) was not the answer. The Soviet Bloc was too hard to penetrate. Warning, ifit were to come in time, would have to come from technical intelligence like COMINT, ELINT, and photography. This recommendation was to begin a revolution in the way the government thought about, organized, and used intelligence. From that time on, technical intelligence became the "answer" to the problem of strategic warning. It would remain so for the duration ofthe Cold War. As part of the Killian Board, the Land Panel was to achieve a measure of renown. Chaired by the farsighted Edwin Land, inventor of the Polaroid camera, the panel was to concern itself with the development of new reconnaissance programs. The Land Panel came to have a profound influence on the future of overhead photography, the U-2 I1ANB~B VIA 'fA~BN'fI12BYIISbfJ eSMIIf'f eSU'fftSb S1S'f~MS JOn''fL 1 NOT Hil Ii: A ~,A BbK '1'9 l"9RBI8N NA:'fI8H'}\L8 230 Ib) (1) Ib) (3) -P.L. 86-36 Ib) (3) -50 USC 403 Ib) (3) -18 USC 798 program, and intelligence collection satellites. It was this group that first envisioned COMINT and ELINT intercept packages aboard orbiting satellites. ,...-_....._---------.....,Land believed that science made anythin~ DOssible.1 The Jackson Report The most personal and confidential report on NSA was by William H. Jackson. One of the original members of the Brownell.Committee, Jackson was' appointed by Eisenhower to monitor NSA's progress and to make periodic progress reports through Sherman Adams, Eisenhower's chiefof staff. In meetings with Jackson, the president expressed his personal concern that NSA should be effective, and Jackson kept him apprised ofwhat still needed to be done. Jackson insisted that NSA needed a strong research and development organization, and he regarded the appointment of a director of research in 1956 as a significant step forward. A more difficult matter was the naming of a chief civilian deputy. Canine insisted on running his own show and did not want, and refused to appoint, a civilian deputy. Only when Samford came aboard in 1956 and quickly named a civilian, Joseph Ream, as deputy was Jackson satisfied on this point. Yet a third organizational problem was the matter of a point of contact for COMINT within DoD. Brownell had envisioned that COMINT matters would be handled at least as high as the assistant secretary level. This high-level attention had not occurred, and Jackson reported in 1956 that the nominal point of contact, General Graves B. Erskine, head of the Office of Special Operations, normally turned COMINT over to a lower-ranking staffer. In Jackson's view, this level ofconcern was wholly inadequate to the task at hand. The objective ofall this organizational toHng and fro-ing was to put NSA in position to mount a full-scale attackl l"Only after such an attack has been made," Jackson noted, "can we determine safely, in the event offailure, that the effort is hopeless and the annual expenditure offorty odd millions can be saved."14 NSA was clearly still on probation. It was a probationary period that would not end with a bang but would slowly fade away. The corner was not turned during either the Eisenhower or Kennedy administration. NSA did not come off probation until the presidency ofLyndon Baines Johnson. Ih...H·Bbl!J Vh... 'fitbl!Jfff I{:I!JYII6bl!J e6MIN'f e6H'fft6b SYB'fl!JflfS 46IN'fbY lilQ:r ~ijbI!JAS.·[BLI!J'f6 f'8ftI!:IOIJE VIA 'fAI>JE~f'fKEYII8I>JE S8MUT'f S8~f'fR8I>JSYSlfEMB VIA 'fAbli>n'f KIi>YII9bli> 69MHf'f 69n'fR9b SYS'fElMS of9IU'fLY NOT RELEASABI E TO i'QKIi>I6N" U2\'fI"NALS 241 But then Dyer turned the solution on its head. He recommended that the alternate become primary - this would effectively move the cryptologic headquarters out of Washington. Dyer carried the day, and his committee began to look at possible relocation sites in the spring of 1950. The selection criteria were developed over a period of months, but generally focused on the following requirements: a. Be within twenty-five miles ofa city ofat least 200,000 b. Have work space totalling at least 700,000 square feet c. Possess a "reasonably equable climate" d. Be suitable for complete physical isolation by fences and the like e. Be accessible to mainline air, rail, and highways f. Not be less than twenty miles from the Atlantic Ocean g. Possess dependable and secure water and electric power sources h. Be accessible to commercial and military communications31 Thomas Dyer, chairman ofthe "Ad Hoc Site Board" IIltNBLtJ '1IA 'fltLtJN'f fttJYIISbtJ CSMm'f CSN'fR8L S'[S'ftJMSd8IN'fLY 242 The basic ground rule was that the location selected had to be on an existing military base. The move was to be completed by July 1955. One option the board looked at was to select a site that already possessed a building - like a hermit crab, AFSA could simply crawl in after modifying the shell. Locations in Kansas City, Tulsa, and St. Louis were considered. Another option was to construct a new building on a military installation. The board looked at Fort Knox, Kentucky; Fort Benjamin Harrison, near Indianapolis; Fort Meade, halfway between Baltimore and Washington; Brooks Air Force Base, near San Antonio, Texas; and Rocky Mountain Arsenal, near Denver.32 Then in early 1951 the board sent AFSAC two recommendations - if the existing structure criterion were used, Kansas City should be the choice, and ifa new building were wanted, Fort Knox was the way to go. This produced great controversy in AFSAC. Some pressed for an existing structure, maintaining that the lower cost and quick availability would help meet the July 1955 deadline. Others opposed moving into someone else's offices - that had been tried at Arlington Hall and Nebraska Avenue and had not worked. The Air Force pressed for Fort Knox, contending that it was less vulnerable to a Soviet nuclear strike. Stone and Major General Canine (who would soon become director of AFSA) both opted for Fort Meade. But in the end AFSAC voted for Fort Knox. The JCS approved the Fort Knox option in April, but only after another heated argument about the advisability of moving to a relatively isolated location. Many, including Stone and Canine, were concerned about the critical lack of housing in the Fort Knox environs, and some wondered ifthe their civilians would accept the choice.33 While orders were being cut and contract proposals were being written for the Fort Knox construction, AFSAC members argued vehemently over the functions to be moved. Dyer was the author of a plan to split COMINT into two parts - three-fourths of it would move to Kentucky, while some residual functions would stay in Washington, along with most ofCOMSEC and some liaison offices. He was opposed by Admiral Joseph Wenger, who felt that splitting COMINT would be disastrous. Ultimately, Wenger won, and it was decided to leave COMSEC in Washington, while all of COMINT would move to Fort Knox and Arlington Hall would be closed.34 The board knew Fort Knox to be objectionable to some of the civilian employees because of its distance from Washington. The lack of housing was worrisome, as was the rigid segregation practiced in Kentucky in 1951. But AFSA pressed ahead with the selection anyway, until a startling thing happened: Someone decided to ask the civilians what they thought. No one knows now who originated the civilian opinion survey, but by May of 1951 it was being circulated at Arlington Hall and Nebraska Avenue. The results were a showstopper. Most of the civilians planned to resign rather than go to Fort Knox.35 Without them, AFSA would find it difficult to operate. The problem had to be fixed. Ih\NBhri YIA 'fAhriN'f ICri'lnShri eSMIN'f eSN'fR8h e:'lS'fBMS dSIN'fh'l NOT REI E ASA ALIi; '1'0 FOKI!l18~l ~iNfI8HA:LS 243 The matter came to a head in October of1951. Deputy Secretary of Defense William C. Foster told Canine, the new director of AFSA, that he had a problem. AFSA's civilians were not in favor of the move to Fort Knox, and neither were AFSA's two most important non-DoD customers, the State Department and CIA. Canine went directly to see General of the Armies Omar Bradley, the Army chief of staff. Bradley told him to meet with the JCS. At the JCS meeting in early December the Fort Knox move was cancelled, and Canine was directed to appoint another site selection board. Canine's new selection board, still chaired by Dyer, but including some civilians, held hurried meetings in January and February of 1952. The new site had to be between five and twenty-five miles from the center ofWashington. This placed it within the postulated blast zone ofthen-existing Soviet atomic weapons and thus violated a JCS stipulation that the new AFSA site had to be at least twenty-five miles from the Washington Monument. But Soviet atomiC weapons were progressing all the time, and the twenty-five mile limit no longer made sense anyway. The JCS could have either atomic invulnerability or a skilled civilian work force, but apparently it could not have both.36 The board looked at several sites in suburban Virginia, including Fort Belvoir, some land along the George Washington Parkway inhabited by the Bureau of Roads (later to become famous as the site of the new CIA headquarters building), and Fort Hunt. In Maryland, it considered several sites within the Beltsville Agricultural Research Center, White Oak (site of the Naval Ordnance Laboratory), Andrews Air Force Base, and Fort Meade. Of those, Fort Meade was the only one on the original list. It was twenty-two miles from the Monument, the furthest removed of any site considered the second time around. Despite the distance from Washington, transportation difficulties would be solved by a new parkway then under construction between Washington and Baltimore. There was plenty of vacant land on Fort Meade for construction of headquarters and life support buildings. It was the obvious choice, and on 5 February it became official. (Considering that Canine said he had already selected Fort Meade himself, and had informed Lovett of that, the proceedings ofthe board may well have been window dressing.) 37 Fort Meade, named for the Civil War victor at Gettysburg, inhabited a thickly wooded 13,500 acre tract precisely halfway between Baltimore and Washington. Originating as Camp Meade during World War I, it had been a training facility during both World Wars I and II. During World War II some 3.5 million men passed through on their way to Europe and at the peak of the war 70,000 people inhabited the post. After the war it became a headquarters, first for the 2nd Army and later (in 1966) for the 1st U.S. Army. When Canine first looked at it, Fort Meade consisted of hundreds and hundreds of temporary wooden structures being used as barracks, offices and training facilities, with only a few permanent brick buildings. The corner of the post that NSA proposed to use Ih\H'BLI!J VIA 'fALI!JWf KI!JYII6LI!J €6MIN''f €6tf'fR6L S'YS'fEJMS ot6n,'fLY NOT REI FA 51 Aibg TQ FQRBl8U U*'f16!(xLS 244 was uninhabited, but was near a major intersection - the new Baltimore-Washington Parkway and Maryland Route 32.38 The new building would be U-shaped with double cross-members, designated the center and west corridors. Entry would be in the middle ofthe west corridor, the portion of the building facing Route 32. At 1.4 million square feet, it would be the third largest government building in Washington, smaller only than the Pentagon and the new State Department building. But it was designed for the AFSA population in 1951, and it did not take into consideration the growth that took place up to mid-decade, which left the new building critically short ofspace. The only solution was to leave someone behind, and that "someone" became the COMBEC organization, which remained at Nebraska Avenue until another building was completed in 1968.39 In 1954 a contract was awarded to two co-prime contractors, Charles H. Tompkins Company of Washington, D.C., and the J.A. Jones Company of Charlotte, North Carolina. The contract price was $19,944,452. Ground-breaking occurred on 19 July 1954. When the building was completed, the total cost turned out to be $35 million, an overrun of almost 100 percent.40 Barracks under construction, 1954 HJz~TQl5K VIA 'FJil5Kl'T'F KKYII915K 69MU'T'F 69~f'FR9l5 SYS~MSJ9H'f'FbY NOT RET FA ll'lIJ"K 'F9 F9REH3N N)cTION ALS 245 A few miscellaneous facts wowed the local community. It had the longest unobstructed corridor in the country, 980 feet long (center corridor). At its birth it had a German-made pneumatic tube system that could carry papers at twenty-five feet per second and could handle 800 tubes per hour. The cafeteria could seat 1,400, and the auditorium (later dedicated to William Friedman), 500. As its new occupant, NSA would become the largest employer in Anne Arundel County.41 It was a far cry indeed from the quaint but antiquated Arlington Hall, the stately Naval Security Station, and the firetrap A and B buildings at Arlington Hall. NSA handled the move in stages. There was an "interim move," which put parts of NSA's operation into temporary quarters on Fort Meade. This had the advantage of moving the operation gradually so that large parts of it were not shut down for any period of time. The new operations building would not be ready for occupancy until 1957, and so the interim move also had the advantage of placing cryptologists at the new location in advance ofthe July 1955 deadline. Headquarters construction. 1955, south wing IIMiBbEl VIA ~ALElU~Ie:ElYIISLEl eSMm'f eSU~ReLSYS'fElMS dSUffLY MOT IU':bIUt8ABI:JE ~O FORElI6H N:*'fI6rof:A:~ 246 - _ .. - - - - - - It began with an interim move to four brick barracks constructed for NSA use in 1954 just behind the proposed site for the main complex. The first to arrive, in November of 1954, was a contingent of 149 Marine guards to provide security. The other 2,000 plus people taking part in the interim move included virtually the entire population of GENS, plus enough communicators, personnel, and logistics people to keep them going. Heat for the operation was provided by an old steam engine which was brought in on the old Baltimore, Washington and Annapolis tracks, and was installed in a small copse oftrees, which still exists, between the present OPS2A and the barracks area. (In fact, the original barracks themselves, now converted to living quarters, also still exist.) GENS and its support staffbecame an outpost, connected to the main headquarters by inadequate electrical communications. Most classified material was couriered back and forth four times a day - the electrical circuits were reserved for only the most critical and time-sensitive information.42 The NSA operations building in 1957 I1z\NBbS Vllt WzbSN'Y' IfSYUSbE eSMHf'Y' eSN'TftSt B1S'ff:MSJt)n'f'fL I NOT REI EASA aL.i: '1''9 peRSiaN' r{}\'ftt)NALS 247 To NSNs military population, the move to Fort Meade was a matter of routine. The military moved frequently, and the relatively cloistered atmosphere of a rural Army post was closer to the normal state of affairs. Family housing was of the two-story brick variety, constructed under the Wherry Housing Act. More would be needed, and over 2,500 new Wherry units were planned to accommodate the increased military population occasioned by NSNs move.43 For civilians, however, it was an entirely different matter. The move to Fort Meade was initially contemplated nervously by a standoffish civilian population. Most lived in Virginia and Washington and faced a long commute over narrow and traffic-clogged roads through the heart of a major metropolitan area. There was no beltway. to take traffic around Washington - the trip north would have to be via Georgia Avenue, Colesville Road, New Hampshire Avenue and other city streets. The only plus to this situation was the brand new Baltimore-Washington Parkway, whose projected completion date was January 1955. That would take care of the drive north from Anacostia and would mark a very significant reduction in the driving time. For those who did not own cars (a significant number in the early 1950s), there was public transportation. Although the old Baltimore, Washington and Annapolis Railroad, which had a spur that ran across the street from the planned NSA facility, had closed its passenger service in 1935, the Baltimore and Ohio Railroad still operated commuter train service from Washington's Union Station to Laurel. For $1.82 per day, one could travel round trip to Laurel and back in thirty-six minutes aboard one of the two trains operating each morning and afternoon. Once in Laurel, the commuter could take the railroadoperated shuttle bus to Fort Meade for an additional round trip fare of 50¢ ~ it required twenty-three minutes each way. Unfortunately, the train and bus schedules did not match very well, and there was no bus service at all for a commuter catching the late train. For the early train, the total oneway commuting time from Union Station to NSA was one hour and twenty-three minutes, not including the time required to get from one's residence to Union Station. Both Greyhound and Trailways offered bus service from downtown Washington to Laurel in just thirty-seven minutes, and at 99¢ per round trip, it was a bargain. But neither service brought passengers to Laurel in time to catch the shuttle to Fort Meade, so commuters would be left high and dry in Laurel. For urbanites used to a short commute to Arlington Hall, this was not a happy prospect.44 For most, this meant picking up the family and moving to the Maryland suburbs. To help with the move, NSA created the Meademobile, a trailer parked between A and B Buildings at Arlington Hall. The Meademobile carried information about Fort Meade and surroundings, including real estate ads, school and church information, and locations of shopping areas. On Saturdays NSA ran a special bus to Fort Meade so that employees could look over the area. For those who were still unsure, NSA announced that a move to II:AUBf::e '(I:A 'f:Af::er,'f IEfJ'iH6f::fJ e6~fH,'f e6H'fft6f:: SYS'ffJMSd6IN'ff::Y NOT RELEASABI Fro fOREI6NNA'f16r,:ALS 248 Fort Meade would be regarded as a PCS, and the government would pay to move household effects. For many, that was the decider.45 The closest community of any size was Laurel. Housing prices in Laurel ranged from $8,990 for two bedroom homes to $10,990 for three bedroom homes with basements. There was also a supply of apartments which could be had for rents ranging from $79.50 to $112.50 per month. In the other direction was the waterside community of Severna Park, whose houses ranged in price from $6,000 to $16,000. Waterfront lots could also be purchased in the subdivision of Ben Oaks, but the lots alone sometimes ran as high as a finished house in other areas. A little farther afield was Glen Burnie, where housing prices ranged from $5,995 to over $10,000. South was the planned community of Greenbelt, in the Washington suburbs. This was originally built with government subsidies, and a house there could be had for as low as $4,700. Single bedroom apartments rented for $51 and Up.46 Columbia had not been built yet. The Meademobile at Arlington Hall Station, 1954 IIMiBbFJ YI1t 'i'AbFJn'i' U:FJYIISLEl CSMHff CSn'i'RSL S'[S'i'ElMS dSHffLY NOT REI FA S AUU; '];Q FQ~IS~T lh\'i'ISNALS 249 The Saturday bus to Fort Meade, 20 April 1954 Whatever NSA did to entice civilians out to Fort Meade, it worked. Early estimates of civilian attrition by a panicky personnel office had ranged as high as 30 percent, but the actual attrition rate was less than two percentage points higher than would normally have been expected had there been no move at all.48 By anyone's standards (except for the COMSEC population left behind at Nebraska Avenue), the move was a success. MAP'fQbS ¥!A 'I'AbStf'I' KElYUeLrJ e8MHff eeN'fR8b SYS'fEMSJ~I1'4"fL i NOT BEL i: A ~ '\iU 'I'e Feftl!ll8N rM:'fI6r'4'AL~ 250 Ib) (1) Ib) (3) -P.L. 86-36 Ib) (3) -50 USC 403 Ib) (3) -18 USC 798 Notes 1. Howe,draft report on the Robertson Committee, in CCH Series VI.X.l.4. 2. Robertson report, in CCH Series VI.X.1.6. 3. Ibid. Ib) (3) -P.L. 86-36 4. "Report on Intelligence Activities in the Federal Gi>vernment. Preparedfor the Commission on Organizationof the Executive Branch ofthe Gi>vernment by the TaskForce on Intelligence Activities," [The Clark Committee of the Hoover Commission) App.1, Part 1: The National Security Agency, May 1955, in CCH SeriesVI.C.1.8. 5. Ibid. 6. Eisenhower Library papers, available in CCH Series XVI. 7. Hoover Commission. 8. Ibid; Eisenhower Library papers. 9. Hoover Commission; Eisenhower Library papers. 10. Hoover Commission. 11. Ibid. 12. Ibid. 14. EisenhowerLibrary papers. 15. Peter Wright (with Paul Greengrass), Spy Catcher: The Candid Aurobiography of a Sf!n.ior Intelligence IOfficer (lew York, Viking Penguin, 1987), 98. The details of the Suez crisis are well documll..nted i.n.. I"":':"_~_ The Suez Crisis: A BriefCOMINT Hisrory, U.S. Cryptologic History, Special Series, Crisis. Collection, V.2 (Ft. Meade: NSA,1988). 21. Ibid. 22. T.G. Fraser, The USA and the Middle East Since World War II (New York: Simon and Schuster, 1989), 73. 23. Stephen A. Ambrose, Eisenhower, Volume 2: The President (New York: Simon and Schuster, 1984),469-73. IIIrH'5hl!l • fir 'flrhI!lH''f KI!l1 1I6hl!l e6Mm'f e6r('fft6h S'fS'fl!ltffS 91 In addition, the committee made certain recommendations concerning NSA's administrative practices - for instance, making professional psychological and psychiatric services available in assessing applicants and employees who revealed instability. But almost an the committee's recommendations had already been implemented, and in its final report the committee gave NSA credit for this. The most far-reaching of the changes related to the termination ofthe procedure ofgranting routine interim clearances, and the institution of the so-called three-hour rule, which required that employees three hours overdue for work would be reported to the security office. These and a long list of other changes became a permanent part ofNSA's way ofdoing business.92 IIAUBLEl IfI," T1tbEltiT IEElYIISLEl CSMINT CSNTHSL ST{S'fEMS dSUffLY NOT RELEASABI E TO i:QggISN tfATI61(1iL5 283 JOP 5!CkET U~RA \ As the Walter Committee proceeded, the FBI investigation was winding down. An intensive screening ofon-board employees had turned up a small number of people whose sexual conduct, in light of the sexual mores of the time, might be questioned, and of these some twenty-six had been terminated. The proceedings were not all that a civil libertarian might have wanted, but they calmed the waters long enough for NSA to begin functioning again.94 The damage to NSA's public image was so severe that it overshadowed the cryptologic damage that had been done. Because it appeared that the two defectors had not carried away documents and that they had not had a previous relationship with the Soviets, just what the Soviets did know as a result was speculative. Martin and Mitchell had known about he Soviet problem, but they were in a position to give away information on certain Soviet cipher systems, especially a system called emp ()yees ut no one ever a proof. And unlike Weisband, their defection was not coincident with any sort of "Black Friday." This, the most famous (or infamQus)of NSA's security cases, was not the most damaging. Notes 1. Max Davidson, "The Criticomm System," Cryptologic Spectrum, Spring 1975,11-14. 2. "NSA's Telecommunications Problems, 1952-1968," CCH Series X.H.4. 3. "Report of the Secretary's Ad Hoc committee on COMINT/COMSEC" (the Robel'tsClnReport), June 1958, in CCH Sel'ies VI.C.l.11. 4. "NSA's Telecommunications Pl'oblems...... 5. "NSA's Telecommunications Problems ... "; Tordella oral interview; Eisenhower Library papers inCGR Series XVI. (b) (6) 6. NSCID7. 7. Tordella interview. IIAUBLe VIA 'fALe!", l(eYIISLe eSMH'" eSN''fReL SYS'ftlMS dSm'fLY ns'f RElbElA8.t'.BbEl 'fS FSRfH8N ?u.'fISrfALS (b) (1) (b) (3) -P.L. 86-36 (b) (3) -50 USC 403 (b) (3) -18 USC 798 284 Ib) (3) -P.L. 86-36 ~U~RA 8. Davidson. 9. "NSA's Telecommunications problems ... "1 1108. 10. Eisenhower Library papers, "Report. of the Joint Study Group on Foreign Intelligence Activities," 15 December 1960, in CCH Series VI.C..1.32. 11. "The Baker Panel Report and Associated Correspondence," in CCH Series VI.X.1.9. 12. Ibid.; Eisenhower Library papers. 13. Baker Pane] Report, ACe 1666'7, eaR~ 51. 14. Baker Panel; Eisenhower Library. 15. Baker Panel; Eisenhower Library. 16. Baker Panel. 17. NSAlCSSArchives,ACC 16667,CaRF51. 18. Eisenhower Library papers. 19. Eisenhower Library papers; David Kahn, The Codebreakers: The Story of Secret Writing (New York: MacMillan, 1967), 677; HistoryofIDAlCRD by Richard Leibler, in CCH Series VI.A.1.6.2. 20. Baker Panel. 21. Howe, draft history ofthe Robertson report, in CCH Series VI.C.1.12. 22. Robertson report. 23. Ibid. 24. Ibid. 25. Ibid. 26. Ibid. 27. Memo for Mrlkubject: 0.·verSight..Ofthe National Security Agency by the Department ofDefense, 9 Nov 1967, in CC~.1.27. 28. "History ofthe Electronic Intelligence Coordinating Group, 1955-1958," in CCH Series VI.O.1.6.; Collins, V. III, 12.; Tordella interview. 29. CCH Series VI.O.1.3.; VI.B.2.6. 30. CCH Series VI.O.1.3.1 ~tudY, 16-17; interview with Dr. Robert Hermann, NSA OH 45-94, 2 Sept 1994, Charles Baker and Tom Johnson. 31. CCH Series VI.O.1.3.; VI.O.1.2. 32. NSAlCSS Archives, ACC 39471, H03-03U-4DtudY, 16-26. 33. CCH Series VI.O.1.3.; NSAlCSS Archives, ACC 39471, H03-0311-4. 34. Melville J. Boucher, "Talomatry [sic] and How it Grew," Part I, Spectrum, Fall 1971, 13; CCH Series VI.0.l.3.; ACC 39471,H03-0311-4. IIM,ebB Wit 'fl\±JElN'f ItBYII6LI!J e6Mm'f e6H''f!t6L SfS'fflMS J6m'fLY NOT RELEASABI E TO i'QRBI8tf H1rTIONALS 285 54. Howe, "Narrative History." fib) (3) -P.L. 86-36 35. NSNCSS Archives, ACC 39741, H03-031l-4;1 ~h~ Soviet Land-b~s:ed Ballistic Missile Program,1945-1972: An Historical Overview," manuscnpt m CCR. 36. NSNCSSArchives, ACC 39741, H03-031l-4. 37. CCH Series VI.O.1.13. 38. Eisenhower Library papers. 39. "Report ofthe Joint Study Croup on Foreign Intelligence Activities," ['I'be Kirkpatrick Report], 15 Dec. 1960, in CCH Series VI.C.1.32. 40. Ibid. 41. NSNCSS Archives, ACC 26115, CBNE 48. 42. Informal correspondence between Gary Winch and Mel BQucher, 1977. 43. CCH SeriesVl.I.1.9. 44. Bob Rush, "AFSCC Tasking: The Development of the Three-Echelon Reporting Concept, 1949-1952," USAFSS history available at AlA, Kelly AFB, Tex,as; 45. "History ofthe USAF Security Service; FiscalYear 1955," AlA, Kelly AFB, Texas. 46. Ibid. 47. Official USAF biography, Oct 1977. 48. Historica-'rLliaIoliUlowuw;.lW:~e6901st SCC, Semi-Annual, 1956-1964, available at AlA, Kelly AFB; Oral interview wit 5 March 1993, by Tom Johnson and Jim Pierson, NSA OH 15-93. 50. Ibid. 51. Ellerson oral history;6901 SCC Semi-Annual histories. 52.1 tA Look at the Pacific Experimental Facility," Spectrum, Wi~tei'1974.,18~21. 53. Howe, "Narrative History... ," Part V, Ch. XXVI-XXX. Ib) (1) Ib) (3) -P.L. 86-36 Ib) (3) -50 USC 403 Ib) (3) -18 USC 798 55. CCH Series VI.HH.12.10. 56. Collins, V. III, 40-41. 57. Transcript ofvideotapes offive former directors~L..-_---l~tUdY,16; CCH Series VI.NN.l.1. 58. Ibid.; Tordella interview. 59. CCH Series VI.D.1.1.; Stone interview; Kahn, The Codebreakers, 705. 60. Tordella interview. 61. Tordella biography in CCH Series VI.D.3.4; Tordella interview. 62. Summary of Statutes Which Relate Specifically to NSA and the Cryptologic Activities of the Government, available in CCH. 63. Ibid. tl:1I:I(fiLE v111: T1I:LEIff K:E f1I6LE e6MIl4'T e6HT1t6L S i STEMS J6mTL i NOT REI E4 54 au TQ fQKEI8~1 N/t'fI8?Ml§S 286 64. Ibid. 65. Ibid. 66. Ibid. 67. Ibid. 68. Church Committee Hearings, V. 5,7-8, ACC 25958-25959, HO-02-0405. 69. John V. Connorton (LTJG) and Floyd W. Tompkins (LT), "The Need for New Legislation Against Unauthorized Disclosures ofCommunication Intelligence Activities," June 1944, SRH 016. 70. Ibid. 71. CCH Series V.C.2.8. 72. Ibid. 73. Hoover Commission report. 74. Kirkpatrick Committee report. 75. Benson and Phillips, V. 1,155. 76. Ibid., V I, 158. 77. Ibid., V. I. 78. Dr. Theodore W. Bauer, "Historical Study: The Security Program of AFSA and NSA, 1949-1962," unpublished manuscript available in CCH. 79. Bauer; Kahn, The Codebreakers, 690-92. 80. NSACSS Archives, ACC 2146, CBm 37. 81. Quoted fm Thomas Powers, The Man Who Kept the Secrets: Richard Helms and the CIA (New York: Knopf, 1979),276. 82. Wayne Barker, The Anatomy ofTwo Traitors: The Defection ofHernon F. Mitchell and William H. Martin (Laguna Hills, California: Aegean Park Press, 1981). 83. Press statement; copy available in ACC 27147, CBm 37. 84. NSAlCSS Archives, ACC 24399, G11-0502. 85. Ibid. 86. Bauer. 87. Barker, CCH Series X.H.5. 88. Bauer; Eisenhower Library papers. 89. Bauer. 90. ACC 45399, G11-0502. 91. "Summary ofStatutes ..."; NSAlCSS Archives, ACC 24399, G11-0502. 92. Ibid. WA WQbil VI A, TAbilJ'TT KilYIlEl"'~ 60MHf'f 8eU'ffteL BYS'fEMB iJ6IHTL I NOT RE' E A:a A au; TEl FORBI8N N}l"fI61'4'ltL5 287 ~UrRA 93. Ibid. 94. Ibid. I1lt~fQbE !JIlt ~ltbl!m~KBYIIebE eeMIN'f eeN'fR6b SYS'fBMSd6H('fLY NO:r RKU A i A RUl ~e FeKEte" nA'fI6n:A:LS 288