Chapter 11: Security
The leaders of the American atomic energy program, aware of the tremendous military potentiality of atomic weapons and reports of German atomic research, recognized almost from the beginning the need for maintaining a high degree of secrecy. An important factor in their decision in early 1942 to turn over administration of the program to the Army was their conviction that it was the organization best prepared during wartime to enforce a foolproof system of security. Such a system would ensure that the Axis powers remained ignorant of Allied interest in developing atomic weapons; reduce the likelihood that the Axis states, particularly Germany, would accelerate their own efforts to produce atomic weapons and undertake espionage and sabotage activities against the American program; and, most significantly, from the standpoint of military effectiveness, allow the Allies to employ these weapons against the Axis nations with maximum surprise.1
First efforts to establish security in atomic matters had occurred in 1939, when refugee physicists in the United States attempted to institute a voluntary censorship on publication of papers concerning uranium fission. American scientists did not accept this suggestion initially, but the outbreak of World War II brought home to many of them the need for control over publications relating to atomic fission. To formalize a censorship program, the Division of Physical Sciences of the National Research Council in April 1940 established a committee that succeeded in getting most scientists to withhold publication of papers on sensitive subjects, particularly those concerned with uranium fission.2
In June, when the government-sponsored Committee on Uranium
became a subcommittee of the newly constituted National Defense Research Committee (NDRC), it also became subject to the security measures currently in effect for federal agencies. The NDRC, knowing that it was to be concerned chiefly with projects for the Army and Navy, adopted security regulations that conformed to those of the two military services. Under these regulations NDRC subcommittees were required to adhere to a policy of strict compartmentalization of information, to classify all sensitive materials, and to obtain security clearances for all employees.
Transfer of the NDRC uranium program to the Office of Scientific Research and Development (OSRD) in November 1941 did not significantly alter existing security arrangements, because the OSRD patterned its own security system largely along the lines of the NDRC program. As the OSRD became more involved in negotiation and administration of contracts with industrial and research organizations, however, it expanded its security controls to provide a more adequate coverage, adding security measures for personnel administration, classified information, and plant protection.3
The modest OSRD security system sufficed until, in the spring of 1942, the start of the uranium program’s rapid expansion – the letting of numerous contracts with industrial firms; the employment and interaction of ultimately tens of thousands of workers, scientists, and engineers; and the formation of complex organizations to construct and operate the large-scale production plants and their atomic communities – enormously complicated the problems of security just at the time the Army undertook its new role as project administrator. Although these measures were necessary for the more rapid achievement of a successful fission weapon, they also tended to weaken security.4 Consequently, the Army almost immediately undertook a reorganization and expansion of the existing OSRD security system and, eventually, also endeavored to bring the system more directly under control of the Manhattan District. The system that finally evolved was in many respects unique and introduced a number of innovations in technique and organization that subsequently would be adopted as standard features of government security programs.
The District’s Security System
The security system, as it took form in the newly established Manhattan District, resembled that already in existence in most other engineer districts. Under Army regulations in force in 1942, the security program of an engineer district was limited to routine local security requirements. When broader problems arose, the
district engineer or security officer could call upon the resources of the Assistant Chief of Staff, G-2, in the War Department. Since June 1939, under provisions of a presidential proclamation, the War Department’s Military Intelligence Division (MID) had shared responsibility for matters of espionage, counterespionage, and sabotage in the United States with the Federal Bureau of Investigation (FBI) and the Office of Naval Intelligence. In the latest revision (February 1942) of this Delimitations Agreement – so designated because it set forth the area of jurisdiction of each agency – the MID’s assignment was to cover the military establishment, including War Department civilian employees and civilians on military reservations or under military control, plus a large part of the munitions industry.5
Organization and Scope
Colonel Marshall, in organizing the Manhattan District security program soon after becoming district engineer in June 1942, formed the Protective Security Section. Under direction of a member of Marshall’s staff, this section emphasized such aspects as personnel, plant, and military information security. At the same time, to provide the District security staff with counterintelligence assistance, Marshall arranged with the Assistant Chief of Staff, G-2, Maj. Gen. George V. Strong, for security liaison with the MID’s operating element, the Military Intelligence Service (MIS). From his staff, General Strong assigned counterintelligence responsibility for the atomic project to Maj. John Lansdale, Jr., who had been a lawyer in civilian life.
Because effective security operations required maximum secrecy, Major Lansdale personally visited the Western Defense Command G-2 and each service command and requested that they each select an officer to report directly to him, bypassing both the G-2 and the commanding general of each service command.6 To further facilitate carrying out the internal security functions for the atomic project, Lansdale also organized a quasi-clandestine counterintelligence group. This group operated under cover of the Investigation Review Branch, Assistant Chief of MIS for Security, which Lansdale headed. He reported directly to General Groves, and his group in effect was answerable to the Manhattan Project commander in all substantive respects, even though it functioned from the G-2 office in the Pentagon.7
By early 1943, the pace of the District’s growth – both geographically and in terms of personnel – and its increasing security requirements emphasized the need for a more comprehensive counterintelligence program. In February, General Strong transferred Capts. Horace K. Calvert and Robert J. McLeod to the District headquarters, where they formed the District’s new Intelligence Section. To ensure that this section, which Captain Calvert headed, had full access to the intelligence and security facilities of the Army service commands, Strong requested that each command designate a staff officer to act as a point of liaison with the Manhattan District and, to guarantee secrecy, authorized that each correspond directly with Calvert’s section. At the same time, Groves continued his earlier practice of meeting with G-2 officers to make certain that District security problems were brought to the attention of appropriate Army officials.8
The counterintelligence program became the foundation for a countrywide permanent organization of this aspect of the District’s security system. During the course of the year, the District organized its own Counterintelligence Corps (CIC) and, as its staff increased in size, assigned new personnel to those areas where there was the greatest concentration of project activities. Ultimately, the project had a total of eleven branch intelligence offices at key points across the United States, from New York to Pasadena (California). An officer assigned to a branch usually worked out of an area engineer’s office and, in addition to his intelligence duties, served as security officer on the engineer’s staff. While in matters of command these officers came under control of the Manhattan District intelligence and security officer and reported to him, they also maintained a direct liaison channel with the director of intelligence of the service command that had jurisdiction over their area.9
Expansion and Centralization
Rapid growth also necessitated expansion of other aspects of the Manhattan Project’s security system. In 1942, the District’s relatively modest internal security organization had served well enough for a program that consisted primarily of administering research and development activities carried on in university and industrial laboratories; but, by summer of 1943, a vast program of plant construction and operation had begun.
The move of the District headquarters from New York to Oak Ridge in August provided an opportune time for reorganization. (See Chart 2.) The first step was consolidation in July of the Protective Security and Intelligence Sections. Captain Calvert took over responsibility for the combined unit, designated the Intelligence and Security Section. Although this change was relatively minor from an administrative standpoint (the section continued in a distinctly subordinate position in the District’s Service and
Control Division), it represented a significant shift towards centralization in security matters. This change was consistent with General Groves’ conviction that only through a high degree of centralized control could he and his administrative staff maintain a close and constant scrutiny over the security program.10
Shortly after the District had completed its move to Oak Ridge, a reorganization in the Army’s administration of counterintelligence operations in the zone of interior (ZI) posed a threat to Groves’ control and cognizance over the project’s internal security functions. To economize on internal investigative operations and to concentrate G-2 efforts on expanding counterintelligence operations overseas, the War Department directed the transfer, effective 1 January 1944, of the WDGS (War Department General Staff) G-2 counterintelligence activities in the ZI to the Office of the Provost Marshal General. The effect was to decentralize even further the Army’s ZI counterintelligence functions to the service commands, including maintenance of data files on individuals which Manhattan intelligence officials considered essential to their operations. The change also seemed certain to enhance the difficulties the atomic project already was experiencing in coordinating its internal security operations with the service commands.11
From his vantage point as head of the atomic project’s counterintelligence group inside G-2, Colonel Lansdale endeavored to have the group exempted from the reorganization requirements. When his efforts failed, General Groves decided that the only acceptable solution was to move Lansdale’s unit into the Manhattan District. The G-2 sanctioned this change in December, and Lansdale secured authorization to establish a special counterintelligence detachment. Groves arranged for Lansdale’s transfer to the Manhattan District; however, instead of placing him in charge of the new CIC Detachment, he brought Lansdale into his Washington office as his special assistant for security affairs. Lansdale’s assignment was to keep the Manhattan chief abreast of problems and developments affecting internal security and foreign intelligence wherever they might arise in the project.12
The shift of all project counterintelligence activities to the District required major changes in its security organization. (See Chart 3.) The Intelligence and Security Section in February 1944 became a full-fledged division and, in keeping with Groves’
centralization policy, moved from the Service and Control Division into the district engineer’s own office. To replace Captain Calvert, whom Groves had selected for a special intelligence mission in London, Colonel Nichols – the district engineer since August 1943 – brought in an experienced intelligence officer, Lt. Col. William B. Parsons, to head the new division. In this capacity Parsons administered the District’s security program with the assistance of Major McLeod, the deputy, and Capt. Bernard W. Menke, the executive officer, and with support from a large operating staff of military and civilian personnel. Although Parsons officially reported to Nichols, he personally kept General Groves appraised of all developments.
Expanding intelligence and security activities necessitated procurement of additional personnel to carry out supportive security functions, such as plant inspections and technical and undercover investigations, Colonel Parsons drew 25 officers and 137 enlisted men from the War Department’s counterintelligence manpower pool and the District’s personnel specialists recruited a large number of civilians. In May 1944, to provide administrative services for the expanding security force, Nichols activated the 13th Special Engineer Detachment (Provisional) and assigned Parsons the additional duty of unit commander. Concerned about achieving greater efficiency in security operations, Parsons requested and received permission in January 1945 to combine the 13th with the CIC Detachment.13
By this time, Parsons’ Intelligence and Security Division had become a highly centralized unit, organizationally divided into six separate branches: Clinton Engineer Works (CEW), Security, Administration, Safeguarding Military Information (SMI), Branch Offices, and Evaluation and Review. The CEW, Security, and Administration Branches, for which McLeod had direct responsibility, dealt primarily with security matters at the Tennessee site. The CEW Branch administered the local civilian guard force and the military police contingent that protected the Tennessee reservation; coordinated subordinate security offices in the K-25 (gaseous diffusion), Y-12 (electromagnetic), and X-10 (pile) process areas; and, through a board established for the purpose, reviewed security cases. The Security Branch chiefly monitored activities related to security of project manufacturing plants, especially at the Clinton site, and the shipping of classified materials and equipment. The Administration Branch was concerned primarily with personnel security problems, both military and civilian, but also provided facilities for the special handling of the division’s mail and records and administered certain confidential funds.
The SMI, Branch Offices, and Evaluation and Review Branches, for which Captain Menke had direct responsibility, eventually evolved as a
central clearinghouse for intelligence and security matters that related not only to the Tennessee site but also to the various project operations elsewhere. The principal responsibility of the SMI Branch was that of project-wide monitoring of programs in security education, censorship, and the handling of classified materials. The Branch Offices Branch, as its name would indicate, was responsible for coordinating field security operations in the eleven geographical areas where atomic energy activities were in progress and for reporting the area engineers’ security problems to the division’s Evaluation and Review Branch. The latter branch concentrated in one office functions hitherto performed by several of the branch intelligence offices – most notably, those concerned with the conduct of subversive investigations and the preparation of special reports on District security matters for higher echelons.14
Counterintelligence activities constituted one of the most significant aspects of the District’s security program. Through effective counterintelligence measures, the District sought to provide the shroud of secrecy necessary
to forestall all attempts by the enemy not only to gain information about the American atomic energy program but also to sabotage it.
Yet by its very nature, the Manhattan Project remained vulnerable to espionage and sabotage. The District’s recruitment of thousands of individuals with almost every conceivable kind of background and from all parts of the country made likely the employment of some potential spies and saboteurs, no matter how efficient its clearance procedures might be, and its widely scattered installations made implementation and maintenance of uniform security procedures throughout the project very difficult. The reality of these conditions forced project leaders to assume that, sooner or later, Germany and Japan – and even the Soviet Union – would learn of the atomic energy program and, more importantly, use espionage to expand their knowledge of it and sabotage to destroy America’s military advantage.15
To detect and counter potential espionage and sabotage activities, the District’s CIC Detachment relied primarily upon extensive intelligence investigations. The majority of these investigations were of a preventive character, designed to minimize the likelihood that security might be breached. Of this type, for example, were the many security checks into the unauthorized transmission of classified information. In most instances, CIC personnel found that the information leaks thus uncovered were the result of carelessness or ignorance on the part of the employee or individual with knowledge of the project. But because it was always possible such leaks were surface ramifications of much more dangerous espionage activity, all cases of careless handling of classified data received prompt and rigorous corrective action.
A second type of preventive investigation was the supplementary and more thorough check into the background of employees earlier subjected to routine clearance procedures. Most supplementary investigations were made because preliminary data indicated an employee might be a potential security risk or routine procedures had not produced adequate information about the person’s background. Typical cases were those involving scientists or technicians who recently had come from abroad, especially those who had come from areas under control of the Axis powers. Faced with a continuing shortage of scientifically and technically trained personnel, project leaders early had adopted the policy of weighing the degree of risk against the contributions an employee with security clearance problems could make in development of atomic weapons. “All procedures and decisions on security, including the clearance of personnel,” Groves recalled, “had to be based on what was believed to be the overriding consideration – completion of the bomb. Speed of accomplishment was paramount.”16
Perhaps the most notable example of the application of Groves’ dictum on employing talented individuals
who were security risks was the case of J. Robert Oppenheimer. When the Manhattan commander decided to appoint Oppenheimer as head of the Los Alamos Laboratory in February 1943, he did so with full knowledge that the theoretical physicist, who had worked on the project since late 1941, had only an interim security clearance from the OSRD. OSRD Director Vannevar Bush, S-1 Committee Chairman James B. Conant, and the other scientific leaders were generally aware of Oppenheimer’s past record of association with Communist-related organizations and individuals. They knew that during the 1930s he had been attracted to a number of Communist-front organizations and, while never a member of the party itself, made fairly regular contributions to Communist-supported causes. Communist fellow-travelers, including his former fiancée, were among his friends, and his wife and brother and sister-in-law were former Communists. With the signing of the Nazi-Soviet pact in 1939, Oppenheimer had begun to have serious doubts about the Communists; however, he continued to contribute to the Spanish War Relief through party channels until the spring of 1942 and to maintain a casual contact with his former friends.17
Despite his record of past Communist associations, Groves decided Oppenheimer was the best choice to direct the bomb laboratory at Los Alamos, for since 1941 he had been involved in this aspect of research and development under Metallurgical Laboratory Director Arthur Compton and in the summer of 1942 had become head of the project team concentrating on that work. Hardly had Oppenheimer arrived at Los Alamos in the spring of 1943 when the question of his clearance arose in a new form. At the request of the Manhattan commander, Lt. Col. Boris T. Pash, chief of the Counterintelligence Branch of the Western Defense Command, began an investigation of suspected Soviet espionage in the Radiation Laboratory at Berkeley. Several men known or thought to be associated with Oppenheimer came under suspicion and, as a result, so did Oppenheimer himself.18 On 29 June, Pash submitted his conclusion that Oppenheimer “may still be connected with the Communist Party.” He offered three possible courses: to replace Oppenheimer as soon as possible; to train a second-in-command at Los Alamos as a possible replacement; and, Pash’s recommendation, to have Oppenheimer meet with Generals Groves and Strong in Washington so that they could brief him on “the Espionage Act and its ramifications” and also instruct him that the government was fully aware of his Communist “affiliations,” that no “leakage of information” would be tolerated, and that the entire project would be held under “rigid control.” In recommending this procedure,
Pash was of the opinion that Oppenheimer’s “personal inclinations would be to protect his own future and reputation and the high degree of honor which would be his if his present work is successful, and, consequently, ... that he would lend every effort to cooperating with the Government in any plan which would leave him in charge.” In any event, he suggested, Oppenheimer should be told that two bodyguards were being assigned to protect him against violence from Axis agents. These bodyguards should be specially trained counterintelligence agents who would not only serve as bodyguards but also keep a Check on Oppenheimer.19
Colonel Pash’s report did not change Groves’ opinion. After a quick visit to Los Alamos, during which he presumably discussed matters with Oppenheimer, Groves directed on 15 July that he be cleared. On his return to Washington a few days later, he directed “that clearance be issued for the employment of Julius Robert Oppenheimer without delay, irrespective of the information which you have concerning Mr. Oppenheimer. He is absolutely essential to the project.”20 As he wrote the Secretary of War four years later, “it was apparent to me that [Oppenheimer] would not be cleared by any agency whose sole responsibility was military security. Nevertheless, my careful study made me feel that, in spite of [his] record, he was fundamentally a loyal American citizen and that, in view of his potential overall value to the project, he should be employed.”21
Most security cases investigated by the District’s CIC Detachment involved breaches of classified information or allegations against employees handling classified work of disloyalty to the United States or of affiliation with organizations espousing subversive ideologies. While many such cases presented the possibility of espionage, in fact, investigations turned up only about one hundred instances of such activity. When suspected cases appeared on the increase in 1943, the Manhattan commander selected a number of the District’s own CIC personnel to serve as special undercover agents. They occupied strategically located positions in project offices, laboratories, and plants, set up listening posts, checked intensively into personnel and other records of individuals under suspicion, and took other measures designed to solve espionage cases.22
The appointment of special agents was a move towards greater formalization of the procedure for dealing with espionage, which continued to increase as the project grew in size and scope. Another constructive measure was the establishment of a group of permanent surveillance
squads to carry out supplemental and nonroutine personnel investigations. Members of these squads, as well as other District security agents, soon became adept in employing professional counterespionage techniques and in using such surveillance equipment as cameras with special lenses (telephoto and other types) and concealable listening and recording devices. During their investigations of persons suspected of espionage activities, either District employees or individuals who had contact with project personnel, the agents operated in the guise of diverse roles – to mention only a few, hotel clerks, bell captains, tourists, electricians, painters, contractors, and gamblers.
To ensure effective functioning and control of the surveillance squads and other special security agents on a countrywide basis, District security officials developed new channels of coordination and communication. Through Colonel Lansdale’s counterintelligence staff at Groves’ Washington headquarters, field security teams at the various branch intelligence offices had access to information from the FBI and other government security agencies. These field teams also had to file written reports of their findings and activities on a regular basis with the Evaluation and Review Branch of the Intelligence and Security Division. As these reports accumulated in the files at District headquarters, they became an important source of information for operation of the whole counterintelligence program. General Groves, in particular, made use of the data garnered from these reports in concert with information acquired from other government agencies in preparing his periodic Military Policy Committee and Top Policy Group briefings on intelligence developments affecting the atomic program.
The most serious espionage activity came not from the enemy but from America’s wartime ally: Soviet Russia. Having in the United States a large diplomatic and consular staff as well as other officials for overseeing lend-lease and other assistance programs, the Russians had a more than adequate reservoir of personnel for maintaining an extensive espionage apparatus in this country. Soviet agents, masking as diplomatic and consular officials, turned to members of the Communist Party of the United States and to party sympathizers for assistance in penetrating American wartime institutions and projects. The Russians, making the plea that the American government was withholding important information and thus unnecessarily delaying Allied victory, recruited many native Communists and fellow-travelers to assist them in obtaining vital secrets about wartime activities.23
As early as February 1943, counterintelligence agents of the FBI and Western Defense Command became aware that the Russians were obtaining data concerning activities of the Radiation Laboratory at the University of California. Further investigation revealed that, in October 1942, a leading member of the American Communist Party on the West Coast
had advised a fellow party member employed at the Radiation Laboratory to retain his position so he could obtain knowledge of the secret work under way there. This employee and other Communists or Communist sympathizers working at the laboratory were passing on information about the atomic project at Berkeley to Communist Party members, who promptly turned it over to the Soviet vice consul in San Francisco. Evidence came to light in early April that a high official in the Soviet embassy in Washington had recently given money to a West Coast Communist leader, to be used for espionage. Intensive investigation by Western Defense Command counterintelligence agents resulted in prompt identification of those Radiation Laboratory employees who were engaging in espionage activities. The laboratory discharged the suspects and, where feasible, the Army inducted them into service, placing them in nonsensitive assignments in which they could be kept under regular observation.24
The District’s CIC Detachment scarcely had completed breaking the original espionage chain at Berkeley when, in late August, Oppenheimer reported his suspicion that new leaks apparently had developed in the laboratory’s security system. On the occasion of a visit to Berkeley, Oppenheimer met with Colonel Pash and told him he had learned that a member of the University of California staff, a man who had been a close friend, was acting as an intermediary for transmission of data from certain Radiation Laboratory employees to representatives of the Soviet Union. By Oppenheimer’s account, his friend had been recruited by an official of the Federation of Architects, Engineers, Chemists, and Technicians, a CIO (Congress of Industrial Organizations) union currently trying to organize employees of the Radiation Laboratory. In subsequent questioning, Oppenheimer refused to disclose the name of his friend on the grounds that he was certain the friend was no longer passing information to Soviet representatives. Oppenheimer’s uncooperativeness at this juncture resulted in the Manhattan commander taking personal action. Groves promptly met with the Los Alamos Laboratory chief and, because the security of the atomic project was at stake, ordered him to reveal the name of his friend. Faced with Groves’ insistence in the matter, Oppenheimer named Haakon Chevalier, a professor of romance languages at the University of California. A short time later, the university dismissed Chevalier from his teaching post and he left Berkeley. In retrospect, the likelihood that Chevalier passed any classified information about the project to the United States seems remote.25
The Chevalier case was not the final incident of espionage at the Radiation Laboratory. Less than a year later, another serious security leak had developed there. With assistance from Communist Party members living in the Sari Francisco area, a key scientist from the laboratory met with officials from the local Soviet consulate. The scientist passed on information concerning the pile process, certain chemical data, and the recently arrived British scientists. The District’s CIC Detachment was able to end this espionage activity effectively by securing immediate discharge of the offending scientist, after which, as far as is known, representatives of the Soviet Union made no further attempts to get information from the Berkeley project.26
Meanwhile, probably acting on the basis of information gained at the Radiation Laboratory, the Russians had assigned one of their best men to the Chicago area, with the task of establishing an espionage channel at the Metallurgical Laboratory. By early 1944, this Soviet agent, who was a highly trained engineer with working experience in both Russian and American industry, had made contacts with several Metallurgical Laboratory employees. By the time the FBI learned of his activities in April, the Soviet agent had obtained considerable technical information, which he had passed on to the Russian consulate in New York. Once identified, the laboratory summarily dismissed the suspected employees. Subsequently, the District’s CIC Detachment discovered that one of the discharged workers – a reserve officer who had been called to active duty and assigned to the Northwest Territory in Canada – had taken highly classified material with him when he left the Metallurgical Laboratory. Fortunately, District security officials were able to arrange for confiscation of this material (it was located in the officer’s baggage) and for transfer of the officer to a post in the Pacific Theater of Operations where he would have no opportunity to pass on his knowledge to Russia or the Axis powers.27
Judged in terms of the ultimate utility of the information gained, Russian efforts at espionage at the Los Alamos Laboratory in late 1944 and early 1945 – the crucial period of bomb development – were the most successful of the wartime period. But project counterintelligence agents did not learn of this activity until the late summer of 1945, after the war was over. In a sensational postwar trial, Julius and Ethel Rosenberg and Morton Sobell were convicted of stealing classified data from the laboratory
with the assistance of Mrs. Rosenberg’s brother, David Greenglass, an Army sergeant at Los Alamos, and of transmitting it to Russian agents. Los Alamos, too, was the place where the German refugee scientist, Klaus Fuchs, while serving as a member of the British team sent to the United States under the interchange program, gained a substantial part of the technical knowledge of the bomb that he subsequently passed on to the Russians, first in June 1945 and thence periodically until his arrest by British authorities in early 1950.28
Project leaders also had anticipated that, as the Russians, the Axis powers – particularly Germany – would launch an equally vigorous espionage campaign, but they uncovered no evidence of such activity during the war. In early 1944, at a time when available Allied intelligence indicated that the Germans might well have attained an advanced stage in the development of atomic weapons, the Military Policy Committee reported to the Top Policy Group that “no espionage activities by the Axis nations with respect to this project have been discovered, although there have been suspicious indications.”29
Measures Against Sabotage
In a project where the ultimate goal depended upon continuous progress in intricate and closely related production processes, unscheduled delays or interruptions of any kind could be disastrous. Sabotage in any form, whether perpetrated by outsiders or insiders bent upon slowing down or disrupting a particular process, constituted an ever-present hazard. Recognizing the seriousness of this threat, General Groves directed that any suspicion of sabotage be reported to him immediately. In keeping with Groves’ policy of constant vigilance to detect any hint of sabotage, the District’s CIC Detachment thoroughly investigated every instance of mechanical failure, equipment breakdown, fire, accident, or similar occurrence not readily attributable to normal causes, and kept under constant observation all processes and activities that might attract the efforts of saboteurs. In addition, other security personnel regularly inspected the security systems and personnel clearance procedures at the project’s various installations, with the objective of detecting and correcting possible weaknesses that might invite sabotage.30
Illustrative of Groves’ policy was the investigation into the mystifying failure of the first great magnets installed in the electromagnetic plant at the Clinton Engineer Works. Following a brief period of operation, the magnets began to malfunction. After
disassembling one of the magnets piece by piece, Kellex engineers found that in its oil circulation and cooling system rust and dirt particles were bridging the gaps between the silver bands forming the coil component, which they attributed to the manufacturer’s failure to maintain sufficiently rigid standards of cleanliness. The significance of this incident was that it revealed the inherent vulnerability of the electromagnetic installations and the need for constant surveillance in order to thwart possible sabotage.31
The district’s continuous and thorough efforts to protect the project’s installations and operations against sabotage were signally successful. During the war years, there were no definitely established incidents of sabotage traceable to enemy agents. In most cases where breakdowns or other failures occurred under suspicious circumstances, investigations revealed they were probably the result of causes other than enemy sabotage. For example, during construction of the original gaseous diffusion plant at the Tennessee site, inspectors discovered someone had driven nails through the rubber coverings of vital electric cables leading underground from the power plant to the main production plant. The perpetrators of this act were never found, although the evidence indicated strongly it was the work of disgruntled employees.32
A quite different type of interference with plant operation briefly threatened the Hanford Engineer Works in early 1945. Groves reported to the Military Policy Committee in February that Army and Navy intelligence had recorded more than fifty incidents of Japanese balloons at various sites along the Pacific Coast, some of them carrying incendiary and fragmentation bombs. While none of these appears to have been directed specifically against the Hanford installations, on 10 March a balloon of this type struck a high-tension transmission line running between the Grand Coulee and Bonneville generating stations and caused an electrical surge through the interconnecting Hanford line that carried power to the production piles. Automatic safety devices at the three piles were activated, briefly shutting down their operation. Fortunately, the bombs attached to the balloon did not explode and the transmission line was not seriously damaged.33
One of the most unusual duties assigned to the District’s CIC Detachment was that of furnishing bodyguards for key Manhattan scientific leaders. CIC personnel accompanied J. Robert Oppenheimer, Ernest Lawrence, Arthur Compton, and Enrico Fermi almost continuously. They accompanied other scientists at intervals, when they were at work on projects that required their special
protection. Colonel Marshall had originated the idea of bodyguards, suggesting that they serve also as drivers, to conceal their true function and to reduce the likelihood of accidents. Compton’s bodyguard, a former Chicago policeman, traveled with him in the guise of a special assistant. When Compton was in residence at Oak Ridge, his guard served as a member of the local police force. District security officials exercised considerable care in selecting individuals for bodyguards, seeking those who had demonstrated ability to adapt themselves readily to the kind of situations in which scientists were likely to be involved.34
Safeguarding Military Information
Even though District security officials had planned and implemented a multi-faceted security system to protect all aspects of project operations and developments, they fully realized that maintenance of total secrecy in such a vast project was unlikely. What was more feasible, they believed, was to prevent leakage of any useful knowledge of the program’s special scientific concepts, industrial techniques, and military objectives – or, in Army parlance, “safeguarding military information.”35
Under the provisions of Army security regulations, the basic responsibility for the protection of classified information rested upon “all military personnel, civilian employees of the War Department, and ... the management and employees of all commercial firms engaged in classified work or projects for the War Department.”36 In applying this principle to the atomic program, District security officials placed particular emphasis upon limiting the amount of classified information permitted to any single individual or group of individuals. District security regulations established two basic rules which were to “govern the right to possess classified information”; a person must need the information in order to carry out his job and have access only to the amount of information “necessary for him to execute his function.” To make doubly certain an individual employee was restricted to “the minimum necessary for the proper performance of his duties,” District regulations further directed that “employees ... shall be organized into small working groups or teams so far as possible, each working on its own phase of the job and not being permitted to inspect or discuss the work being done by others.”37
This compartmentalization policy became a far more pervasive influence in the project after the Army assumed
full responsibility for its administration. Where the OSRD had applied compartmentalization primarily to research and development organizations, the Army incorporated it into virtually every type of activity undertaken by the project. Typical was the District’s insistence that production plant blueprints be broken down and distributed in such a way as to reveal as little as possible to any one individual about the overall character of the project. Similarly, the District required that equipment orders to commercial firms specify that an item not be manufactured and assembled at the same location. And when the production plants reached the point of start-up operations, plant managers received instructions to split up orders for raw materials among a number of suppliers so that the purpose for which they were being used could not be readily ascertained.
While project leaders agreed that some compartmentalization of information was necessary, considerable difference of opinion prevailed on the extent of limiting scientific and technical interchange, both between sections functioning within a laboratory or plant and between the various interrelated installations of the project. Military administrators, in contrast to their civilian counterparts, favored the enforcement of stricter controls. These generally took the form of written agreements covering those organizations and installations that needed to exchange data. The agreements specified in detail how and what information could be
interchanged. Inevitably occasions arose when developments required interchange of classified information not covered in agreements. In such instances, project leaders applied directly to the district engineer or to General Groves for special permission to exchange the data needed.38
One of the most important interchange arrangements formed occurred in June 1943, when General Groves met with Compton and Oppenheimer for the purpose of establishing “the principles which should govern the interchange of information between the Chicago [Metallurgical Laboratory] and Los Alamos projects. ...” As a basic criterion determining what information should be interchanged, they set up the test that only data that would “benefit work at both Chicago and Los Alamos” should be exchanged. The agreement that resulted spelled out, in considerable detail, exactly what information could and could not be interchanged (the latter included those categories relating to production piles, military weapons, and the time schedules of various developments); designated by name those individuals at each installation who were qualified to carry on interchange; and outlined exact procedures of exchange – by formal reports, secret correspondence, or visits and conferences. On the most sensitive matters, or where there was serious doubt about interchange, the only channel of exchange was through a visit to the Chicago laboratory by either Oppenheimer or a specifically designated group leader. Although negotiators of the agreement must have been aware of the generally restrictive character of its provisions, they nevertheless emphasized that its major objective was “to maintain as rapid and effective interchange of information as possible.”39
Compartmentalization of information probably aroused more adverse criticism – both from participants in the atomic program and from some of those who, in retrospect, have reviewed its history – than any other single aspect of the project’s security system. Among the participants, the most vociferous critics were the scientists, accustomed to working in college and university laboratories where they could freely interchange the results of their work with scientific colleagues in all parts of the world. Project scientists, such as Leo Szilard, held that over-compartmentalization was a primary cause of extended delays in achievement of scientific and technical objectives of the program. Testifying before a committee of Congress after the war, he asserted, for example, that “compartmentalization of information was the cause for ... failure to realize that light uranium [U-235] might be produced in quantities sufficient to make atomic bombs. ... We could have had it eighteen months earlier. ... We did not put two and two together because the two two’s were in a different
Joining Szilard in condemning compartmentalization in the strongest possible terms was Edward U. Condon, the prominent American physicist who had come to the atomic project from the Westinghouse Research Laboratories. In fact, after spending only a month at Los Alamos, Condon came to the conclusion that he would be of more use to the war effort at Westinghouse than at the New Mexico laboratory. The project’s security policy, he asserted, had a morbidly depressing effect on him. “I feel so strongly,” he continued, “that this policy puts you in the position of trying to do an extremely difficult job with three hands tied behind your back that I cannot accept the view that such internal compartmentalization ... is proper.”42
Most other contemporary critics took a somewhat less extreme position. Concerned about insufficient interchange of data among atomic project scientists causing delays in the solutions of problems related to bomb development, Compton suggested to the OSRD S-1 Committee in December 1942 that it might be wise to increase the number of “responsible persons who are free of compartmentalization. ...”43 Similarly, in June 1943, physicist Richard C. Tolman, in his role as Groves’ scientific adviser, expressed concern that the “proposed regulations [to govern interchange between the Chicago and Los Alamos scientists were] perhaps not quite as liberal as may later prove warranted.” In the weeks following the institution of these regulations, both Oppenheimer and Edward Teller, who was working on a part-time basis at Los Alamos, were troubled by what they viewed as inadequate liaison channels between the New Mexico laboratory and the other installations where related work was in progress.44
When British officials and scientists came to the United States in late 1942, they were surprised to learn that General Groves planned further compartmentalization, which many of them viewed as already having been applied to an extent that made efficient operation impossible. Furthermore, the British soon found that the Americans used the policy as a convenient excuse for withholding information. Thus, the policy became intermeshed with the whole question of interchange with the British, a problem that was resolved only after many months of negotiation.45
By early 1944, most project personnel had come to accept the policy as a fact of life. In looking back after the war was over, even some scientists who had found compartmentalization so distasteful grudgingly conceded it had probably been necessary. The eminent American (German-born) physicist James Franck, for example, while speaking at a conference on atomic energy at the University of Chicago in September 1945, concluded that “so far as secrecy is concerned, they [Army officers] were unrelenting and, in all honesty, we have to admit that they had to be.” But, he went on to remind his listeners that the policy had exacted a “stiff price” in the “wasting of talent and scientific manpower and the loss of precious time. ...”46
From the military point of view, compartmentalization was precisely what was required, both for security and for achieving the most efficient functioning of scientists and technologists. As General Groves expressed his conviction in retrospect:–
Compartmentalization of knowledge, to me, was the very heart of security. My rule was simple and not capable of misinterpretation – each man should know everything he needed to know to do his job and nothing else. Adherence to this rule not only provided an adequate measure of security, but it greatly improved overall efficiency by making our people stick to their knitting. And it made quite clear to all concerned that the project existed to produce a specific end product – not to enable individuals to satisfy their curiosity and to increase their scientific knowledge.47
The District’s policy of compartmentalization of information on the atomic project, in Groves’ words, applied “to everyone, including members of the Executive Department, military personnel and members of Congress.” No one was to have access “solely by virtue of his commission or official position.” Adherence to this policy was possible as long as Manhattan’s funding came from sources already earmarked for the War Department. But project leaders anticipated considerable trouble in the future, because securing new funds would entail congressional authorization.48
By early 1944, the compartmentalization policy was becoming less and less feasible with Congress because of the increasing size of the program, its rapidly rising cost, and the need to begin planning for its postwar administration. Under the original directive from the President, the atomic program obtained funds from the money appropriated under the Engineer Service-Army budgetary category. Funds from this source sufficed as long as Manhattan’s budgets remained relatively modest. But when project leaders estimated that the program would need at least $600 million for fiscal year (FY) 1945, they decided they would have to find a way to provide some information to selected members of Congress who had a need to know. They consulted with President Roosevelt, who thereupon
directed that Stimson, Bush, and General Marshall brief the leaders of both parties in the House and the Senate.49
On 18 February, Stimson, Bush, and Marshall went to the office of Speaker of the House Sam Rayburn, where they were joined by Majority Leader John W. McCormack and Minority Leader Joseph W. Martin, Jr. Stimson outlined the history of the atomic project, including its cost to date, and estimated the total amount needed to complete it; Bush described the project’s scientific background and indicated the likely destructive power of an atomic weapon; and Marshall discussed the potential role of atomic bombs in the Allied strategy for winning the war. The legislators pledged their unreserved support, stating that they viewed its high cost as well worth the price. They promised to work out a system for handling the Manhattan appropriations in committee so that there would be no danger of disclosure of their purpose. Bush found that the “entire meeting was most reassuring, as it was quite evident the three congressmen were exceedingly anxious to be of aid to the War Department in carrying a very heavy responsibility.”50
In June, Stimson, Bush, and Maj. Gen. George J. Richards, the War Department budget officer who was substituting for Marshall while he was out of town, repeated the briefing for the leaders of the Senate. Present were Majority Leader Alben W. Barkley and Minority Leader Wallace H. White, as well as Chairman Elmer Thomas and Senior Minority Member Styles Bridges of the military subcommittee of the Senate Appropriations Committee. Stimson recalled that “the four gentlemen who met with us were very much impressed. They . promised that they would help and keep absolute silence about it and prevent discussion in public as to what it was about.”51
During the remaining months of 1944, congressional leaders succeeded in keeping the vast majority of the members of Congress ignorant of the atomic project. Accustomed to wartime restrictions, most members were willing to accept – without protest – the assurance of their leaders that the work was secret and that the needed appropriations were essential to the war effort. But for a few members this policy was unacceptable, and they directed individual inquiries to the War Department about rumored developments at the atomic sites.
A case in point was Congressman Albert J. Engel of Michigan, a member of the House Appropriations Committee, who in February 1945 was unwilling to accept automatically the War Department’s request for FY 1946 funding from money appropriated under the Expediting Production budgetary category. In a visit to Under Secretary Patterson on the twenty-fourth, the Michigan representative stated that he had heard
rumors of extravagance and waste and that he wanted more information before approving the War Department’s FY 1946 funds. Remembering that in late 1943 War Department officials had dissuaded him from making a proposed trip to the Clinton site, this time he firmly insisted that Patterson allow him to inspect the atomic installations. When Stimson heard from Patterson of Engel’s insistence upon visiting project facilities, he sought assistance from the leaders of the House of Representatives. As Speaker Rayburn was away, Stimson turned to Congressman John Taber of New York, another member of the Appropriations Committee. He and Taber sat down with Engel and persuaded him to forgo objections to funds on the floor of the House, but only after promising him an opportunity to visit some “outside installations” of the project.52
This experience convinced the Secretary of War and the Manhattan commander, as well as other project leaders, that more and more members of Congress would be demanding current information about Manhattan’s activities. Consequently, they arranged to have a selected delegation from each House visit Clinton and, if they wished, also Hanford. With the President’s approval for this plan, Groves and Stimson, accompanied by the Secretary’s aide, Col. William H. Kyle, visited Clinton on 10 April to prepare “for future trouble with Congressmen.”53
Upon the unexpected death of Roosevelt on the twelfth, the inspection trip to Clinton was delayed, but only temporarily. In May after President Truman had given his assent, Speaker Rayburn helped select five members from the House Appropriations Committee – Clarence Cannon, the chairman, George H. Mahon, J. Buell Snyder, Engel, and Taber. Under the careful guidance of the Manhattan commander and the district engineer, the five congressmen spent two days inspecting the Clinton Engineer Works. The legislators returned to Washington convinced that public funds had been well spent and prepared to support the project’s budgetary requests for FY 1946. A visit by a comparable Senate delegation to inspect atomic facilities was not feasible until after V-J Day, when a group from the upper house toured the Hanford Engineer Works.54
As security requirements increased, the Army established a variety of units to administer its highly compartmentalized information security program. By necessity, the program from about late 1942 up until the District’s major intelligence and security reorganization in early 1944 was limited in scope. Faced with a rapid influx of new personnel, both civilian and military,
the District’s Protective Security Section concentrated chiefly on developing ways for instructing them in the meaning of classified information and the correct methods for handling it. To facilitate this education process, the small staff hurriedly prepared and distributed a manual that provided a “statement of District policy regarding Protective Security procedures ... ,” including an extensive section on safeguarding classified information.55
An intensification of protective measures during the first half of 1943 resulted in the establishment in August of the Plant Security Section for Safeguarding Military Information. In an effort to assure attainment of the desired security objectives, the SMI staff developed a new intelligence bulletin. This bulletin, issued in November, set forth in detail the requirements and procedures for safeguarding military information, emphasizing that “matters of vital importance to the government must be protected at all times whether at war or at peace ... [and thus] great caution [must] be exercised in the handling and in the dissemination of all information – written or oral – relative to this Project at any time.”56
By early 1944, consolidation of the District’s intelligence and security facilities opened the way for a more comprehensive information security program, and the establishment in May of a separate SMI Section (redesignated SMI Branch in 1945, when organizationally restructured as a subordinate unit of the District’s Intelligence and Security Division). Under the expanded program, security officials launched studies of all aspects of the atomic project – equipment, material, products, processes, operations, administrative matters – to determine the appropriate classification for their mention in correspondence and other documents. They set up code names (some already in use) for major sites, important materials, items of equipment, and even for the more widely known scientists working on the project. Under this scheme, for example, Los Alamos became Site Y, plutonium became 94, the implosion bomb became Fat Man, and scientist Arthur H. Compton became A. H. Comas. Using the staff and resources of the SMI Section, District authorities directed attention to those areas where security leaks were most likely to occur. Thus, the section regularly reviewed project correspondence with other government agencies, such as the Selective Service concerning deferment of key personnel, and advised on the security classification that should govern each of the thousands of contracts that the District negotiated with outside individuals and firms.57
The establishment and maintenance of effective adherence to security requirements among the project’s thou. sands of contractor organizations comprised one of the most challenging and complex aspects of the information
security program. District authorities oversaw contractors’ security activities through several channels. The branch intelligence offices in principal cities throughout the United States provided a convenient point of contact, and periodic checks of contractor facilities and operations by security inspectors from District headquarters constituted a second avenue of control. These inspectors particularly observed methods of handling classified materials and storing documents. District security officials also investigated contractors’ personnel recruitment programs, written correspondence, stock registration statements to the Securities and Exchange Commission, and similar activities in which security leaks were likely to occur. Finally, when a contractor terminated his contract with the atomic program, District security officials made certain that all classified materials were returned to project control or that the contractor provided for their adequate protection.58
Security problems involving firms under contract most frequently arose where these organizations were carrying out large-scale development of project facilities. Such development, as at the Clinton and Hanford sites, inevitably brought overcrowding of local housing, acute labor shortages, greatly increased road traffic, and other adverse changes that placed a severe strain on normal community activities. The resulting public resentment, generally focused on the contractor firms, created an environment in which threats to security were more likely to occur. In the spring of 1943, for example, Du Pont’s effort to arrange for housing and other facilities for the thousands of employees who would work on the Hanford project stirred up resentment in surrounding communities, already aroused by the Army’s land acquisition program. The spread of rumors, adverse criticism in the local newspapers, and unfounded statements by local officials tended to draw widespread public attention to the project, posing a serious threat to security. Lt. Col. Franklin T. Matthias, the Hanford area engineer, and members of his staff spoke at meetings of service clubs in communities adjacent to the project, in an endeavor to counter the rumors and misinformation concerning Du Pont’s role in the project. By these and similar efforts they laid the groundwork for obtaining the support and good will of the local citizenry – an absolute essential to maintaining the security of the project.59
Efforts to maintain good community relations was an important aspect of the District’s information security program, which had as its prime objective the forestalling of security breaks, first by anticipating them and second by teaching project personnel how to be “instinctively alert-minded and security-wise.”60 Although the SMI Section had primary responsibility for carrying out the program, employee education in security matters devolved chiefly to the SMI staffs at the branch intelligence offices. Each staff, for example, conducted orientation
and refresher sessions for Corps of Engineers personnel; provided each contractor with instructional materials for in-house security education briefings for its personnel; and used a variety of media – training films, circulars and handbills, payroll inserts, telephone stickers, and editorials in project newspapers – to remind District employees of the importance of unremitting attention to the demands of security.61
Because of the policy of compartmentalization, the quantity and variety of educational subject matter available for training purposes was limited. Most workers had knowledge of only the project activity under way at the site where they were employed, and most generally did not even know exactly what was being made in the facility where they worked. And even in some instances, project officials had concocted for employees – those working at the electromagnetic plant – a plausible but inaccurate and misleading explanation of the process involved and the product produced, with the warning that this information was given to them only to help them carry out their jobs. Lacking concrete data on which to base an appeal to employees, security officials had to request that they accept the necessity for strict adherence to secrecy largely on faith and out of a sense of patriotism and loyalty to the men on the fighting fronts.
As did most wartime agencies involved in secret work, the Manhattan District resorted to censorship of various kinds as a means of safeguarding classified information. In the first few months after the Army assumed responsibility for the atomic program, the District and branch security staffs began a cursory review of a few leading daily newspapers and periodicals and gradually enlarged this check of publications until it covered some 370 newspapers and 70 magazines. The censors, several of whom were Women’s Army Corps members, were particularly on the lookout for publication of anything that would reveal classified information, attract attention to the project, or furnish an enemy agent or anyone else with knowledge sufficient to determine the nature of the project.62
While review of newspapers, periodicals, and other publications provided some protection against damaging revelations about the project, the fact remained that once such information appeared in print an element of secrecy was lost. Much more effective was a system that prevented publication of sensitive information. Under the Office of Censorship’s “Codes of Wartime Practices for the American Press and American Broadcasters,” newspapers, periodicals, and radio broadcasters voluntarily agreed to refrain from discussing certain specified subjects and mentioning certain terms. In February 1943, Vannevar Bush proposed that the atomic energy program be brought under this voluntary censorship. At first, both General Strong, the Army intelligence chief, and General Groves had serious reservations about making the atomic energy project subject to this censorship arrangement, fearing that the results
“might be more detrimental than otherwise.”63
Finally, military leaders reluctantly agreed to the voluntary press censorship plan, persuaded primarily by the insistence of Nathaniel R. Howard, assistant director of the Office of Censorship and a former editor of the Cleveland News, that this was the only way to maintain press security of the project. On 28 June 1943, Byron Price, director of the Office of Censorship, sent out a special request to all editors and broadcasters that they extend the previously issued precaution not to publish or broadcast anything about “new or secret military weapons … [or] experiments” to include:–
Production or utilization of atom smashing, atomic energy, atomic fission, atomic splitting, or any of their equivalents.
The use for military purposes of radium or radioactive materials, heavy water, high voltage discharge equipment, cyclotrons.
The following elements or any of their compounds: polonium, uranium, ytterbium, hafnium, protoactinium, radium, thorium, deuterium.64
The aim of censorship was to prevent all mention of the atomic program in the American press; however, on the advice of the Office of Censorship, the District permitted a limited amount of information about certain aspects of the project to appear in newspapers published in communities near the Clinton and Hanford sites. Office of Censorship officials pointed out that complete suppression of information about activities at these locations would actually draw more attention than a policy of judicious release of news of local interest, carefully controlled so as not to reveal any vital secrets. They cited as an example the land acquisition at Hanford, which required relocation of many people and resulted in court proceedings. Stories on these events in newspapers of the Washington-Oregon region would not violate essential security as long as they did not reveal the purpose of the acquisition or the interconnection of the Hanford project with other parts of the atomic program. General Groves assented to this policy but took the added precaution, suggested by Office of Censorship officials, of having Manhattan District representatives visit the editors or publishers of local newspapers and operators of local radio stations to request their cooperation in maintaining the security of the project.65
At Los Alamos, security authorities endeavored to keep all mention of the site and its activities out of the press. Total exclusion was more feasible at the New Mexico installation because of its military administration and geographic isolation from surrounding communities. The policy was reinforced in late 1943 through the use of
regular mail censorship and other measures to minimize the likelihood that knowledge of the site would come to the attention of the press.66
It was inevitable that a voluntary censorship system would not be totally effective, and on those occasions when some reference to the project or atomic energy occurred in the press or on the radio, the District security office and the Office of Censorship took immediate steps to limit its circulation and to run down it origins. A rash of censorship violations occurred in late 1943. A columnist in the Washington Post announced that the Senate’s Truman Committee was about to investigate a “half-a-billion dollar” War Department project in the state of Washington that was “reported to be one of the largest single projects that’s to be built from scratch in the Nation’s history.” On the same day the Post article appeared in the Spokane Spokesman-Review, and soon thereafter the wire services picked up the news item. Almost simultaneously, several newspapers in Tennessee ran a story on the state’s Selective Service that contained a passing reference by the head of the service, Brig. Gen. Thomas A. Frazier, to “the Clinton Engineer Works in secret war production of a weapon that possibly might be the one to end the war.” In both instances, prompt action by the Office of Censorship led to withdrawal of the articles before they had received wide circulation. Subsequent action by the War Department resulted in tracing down the sources of the leaks and in implementing improved security measures to prevent such occurrences in the future.67