"The aspect of the Pearl Harbor disaster which is really surprising is that so many people failed to do either the obvious or the sensible things." Washington Star, 1 September 1945
This is the story of the U.S. Navy's communications intelligence (COMINT) effort between 1924 and 1941. It traces the building of a program, under the Director of Naval Communications (OP-20), which extracted both radio and traffic intelligence from foreign military, commercial, and diplomatic communications.1 It shows the development of a small but remarkable organization (OP-20-G) which, by 1937, could clearly see the military, political, and even the international implications of effective cryptography and successful cryptanalysis at a time when radio communications were passing from infancy to childhood and Navy war planning was restricted to tactical situations.2 It also illustrates an organization plagued from its inception by shortages in money, manpower, and equipment, total absence of a secure, dedicated communications system, little real support or tasking from higher command authorities, and major imbalances between collection and processing capabilities. It explains how, in 1941, as a result of these problems, compounded by the stresses and exigencies of the time, the effort misplaced it focus from Japanese Navy traffic to Japanese diplomatic messages. Had Navy cryptanalysts been ordered to concentrate on the Japanese naval messages rather than Japanese diplomatic traffic, the United States would have had a much clearer picture of the Japanese military buildup and, with the warning provided by these messages, might have avoided the disaster of Pearl Harbor.
This story also records what today must be ranked as an intensely important interlude when the Navy radio-traffic intelligence program deliberately avoided the underlying intelligence of intercepted traffic while exploiting foreign cryptographic systems. Today most intelligence experts would call such a practice naive or ill advised. Yet a policy requiring OP-20-G cryptanalysts to search primarily for unique cryptographic features of codes and ciphers which might later be refined and employed by navy cryptographers was not changed until 1942. Coupled with a reluctance to fire civilian trainees, this policy seriously delayed the training of enough Navy cryptanalysts linguists to deal with a work load which increased exponentially both in complexity and volume after 1939. Ultimately, the resulting shortage of cryptanalysts and Japanese linguists, the problem of misplaced priorities and interservice rivalry issues all contributed to misplacing the major focus of the Navy's cryptanalytic and linguistic efforts
on Japanese diplomatic messages. The unfortunate result of these circumstances was to postpone with fatal consequences an all-out effort on Japanese Navy cryptosystems.
This is not to minimize the value of the pre-Pearl Harbor efforts of Navy cryptanalysts and traffic analysts. Even without the messages pertaining to the Japanese Pearl Harbor strike force, the magnitude of the information they produced pertaining to the Japanese 2nd, 3rd and 4th Fleets and the Japanese 11th Air Fleet was overwhelming. These intimate details concerning Japanese intentions, however, were not based on messages but on analytic judgments drawn from analysis of Japanese Navy communications procedures, patterns, and practices. As suspect quantities from a suspect source, they were not accepted by the very commanders in whose service they had been developed. The lack of confidence in such intelligence made traffic intelligence from the Pacific during the last half of 1941 more an elaborate rumor than trustworthy source material. Commanders at the theater level and in Washington, through lack of early training or insight, were not prepared to exploit the intelligence provided by this source, particularly when the messages themselves could not be read.
In addition to outlining the development of the Navy's cryptanalytic attack against Japanese cryptographic systems, this review also examines other interesting episodes overseas and in Washington which included two attempts, one unsuccessful, to coordinate Navy and Army COMINT activities, efforts to improve fleet communications, and the lessons learned and then forgotten about Japanese naval communications from Japanese Fleet maneuvers of the 1930s. Coordination and cooperation between the U.S. Navy COMINT Center on Corregidor and the British Far East Combined Bureau in Singapore are briefly described.
The origins of the U.S. Navy's COMINT effort prior to 1924 are not entirely clear. However, the Navy established a Code and Signal Section possibly with COMINT interests as early as 28 July 1916. This small organization initially worked against German ciphers during World War I. It also tested the security of U.S. Navy ship-to-ship and ship-to-shore communications during training maneuvers. During the maneuvers of 1917, or example, personnel from the section were involved in an overt attempt to intercept and exploit U.S. Navy communications in order to demonstrate their accessibility to foreign intelligence efforts.3
For some unknown reason, the initiatives apparently ended with the World War in 1918. At that time the Navy voluntarily consolidated its wartime efforts with those of the War, Justice, State, and Postal Censorship Departments, forming a single U.S. Cipher Bureau under the War Department. Commanded by Captain Herbert O. Yardley, assisted by Captain John W. Manly, the consolidated bureau consisted of thirteen cryptographers, twelve of whom were women, and an administrator. It was supported by eleven student
officers, eight stenographers, and fifteen clerks. The Navy was represented by Chief Yeoman H.W. Burt.4
From 1919 to 1923 the Navy seemed to rely almost entirely on the Cipher Bureau. In 1923 the Navy apparently felt that the Cipher Bureau had produced neither the desired cryptographic improvements nor the necessary insights into the activities of foreign navies, in particular Japan's growing fleet. Accordingly, using shipboard communicators, the U.S. Navy in 1923 began an ad hoc effort to listen to foreign radio traffic, which its earlier work had shown to be potentially vulnerable to penetration and exploitation.
In January 1924, Commander Ridley McClean, Director of Naval Communication (DNC) established a research desk within the Code and Signal Section with a complement of one officer, Lieutenant Laurance F. Safford, and one civilian, Agnes Meyer, both of whom were cryptanalysts/cryptographers. Safford and Meyer conducted research into foreign cryptography, organized training in collection and cryptanalysis, developed cryptographic systems for naval communications, and arranged with the Commander in Chief, Asiatic Fleet (CINCAF), and certain naval district commanders to obtain copies of radio intercept of foreign messages. The primary goal of the two was to develop cryptographic systems for the U.S. Navy which would avoid the weaknesses observed in the cryptographic techniques employed by foreign governments.
Early in 1924, Commander McClean, and the Chief of Naval Operations (CNO), Admiral Edward W. Eberle, encouraged the Commander in Chief, Asiatic Fleet (CINCAF), Admiral Thomas Washington, to expand radio intelligence facilities in his area. As a result of this encouragement, the first Navy intercept station ashore (Station A) was established the same year at
Shanghai in the U.S. Consulate. it copied both naval and commercial traffic (Japanese and British). Admiral Washington was responsible for all aspects of the operation including administration, logistics, personnel, and targeting. OP-20-G received all intercept logs, including traffic and messages for cryptanalysis. After the codes were broken and the messages reduced to plain text, the contents were then sent to the Officer of Naval Intelligence (ONI) where, if necessary, they were translated into English.5
Role of ONI
This decision to emphasize a communications security (COMSEC) objective in exploiting foreign communications was to prove costly to OP-20-G. It unwittingly conceded to others, notably the Office of Naval Intelligence, the responsibility for developing and disseminating underlying intelligence, control of language billets and, by thus sowing confusion regarding the nature of communications intelligence, sacrificed much of the long-term initiative regarding direction of the overall effort. The scope of ONI's COMINT-related activities and the magnitude of the cost of this concession in both human and fiscal terms can be illustrated in part by the following story. In 1930, the existence of a secret fund at Riggs Bank in Washington -- at one time as large as $100,000 -- administered by the Director of Naval Intelligence (DNI) accidentally came to light during the transition of ONI directors. This fund was used to buy equipment and material in support of DNC's efforts to intercept and exploit foreign radio communications. A strong possibility exists that the fund was also used to underwrite the costs of stealing, photographing, and translating the Japanese Imperial Navy Secret Operations Code (the Red Book) twice between 1921 and 1927. In 1931, over the futile objections of Captain Stanley C. Hooper, DNC, and Commander J.W. McClaran, OP-20-G, the acting Director of Naval intelligence, Captain William Baggley, returned the entire balance in the account ($65,000) to the U.S. Treasury. Why he did this is not clear. It may have been part of a Hoover administration economy move. Within the Navy the effects were startling. From 1931 to 1933, Admiral William V. Pratt, CNO, in retaliation for this action, ordered that ONI not be shown any of the thousands of deciphered messages available each year -- a policy which must have caused extensive reassignments among ONI's Japanese linguists. In 1929 Lieutenant Commander Safford had advised Hooper that, over a five-year period, it had accumulated 10,000 Japanese diplomatic messages and, in two years, 2,000 Japanese naval messages.6
The early success of OP-20-G in obtaining official and commercial foreign radio traffic and exploiting the knowledge gained from decrypts made possible by acquisition of the Japanese Navy's operation code, had an immediate as well as far-reaching effect on the Navy at large. In March 1926, for example, a secret memorandum from Admiral Edward W. Eberle, CNO, to all major commands noted that current war plans did not "adequately reflect benefits gained from radio intelligence." Accordingly, Admiral Eberle directed that
"Unit commanders both afloat and ashore ... develop their own plans for service of radio intelligence in war." Eberle's letter also recommended (1) "intercept and decoding units" ashore and afloat; (2) direction finding (DF) units in each naval district but primarily in the 12th (Headquarters San Francisco), and in the headquarters of the Atlantic and Pacific Fleet; and (3) translators in the 14th Naval District (Hawaiian Islands) and in Washington. Clearly this represented a major change in perspective regarding radio intelligence and translators which had been inspired by the work of Station A (Shanghai), ONI, and the cryptanalysts in Washington, Safford and Meyers.
The years 1926 through 1928, however, saw little deliberate progress by the commands in implementing Eberle's desires, strongly suggesting that the message about COMINT's value was not widely accepted. Rear Admiral George R. Marvell, the Commandant 14th Naval District (COM-14), did mention radio intelligence in his 1928 war plan, but his attempts to establish an official intercept site at Wailupe, Hawaii, proved abortive. In the Asiatic Fleet area, neither Admiral Mark. L. Bristol, CINCAF, nor Rear Admiral Summer E.W. Kittelle, Commandant 16th Naval District (COM-16), made any moved to enlarge on Station A, although Guam and the U.S. legation at Peking began to appear in correspondence as possible candidates for new sites.
Admiral Joseph N. Wenger, USN
|In the training of intercept operators and cryptanalysts, nevertheless, some real progress did occur during this period. In 1926 Ensign Joseph N. Wenger was the first officer to undergo training in a cryptanalysis "short course." Training for officers consisted of on-the-job training and semiformal instruction by Safford and Meyer. Wenger was followed in the same year by lieutenant Joseph J. Rochefort and Captain Leo F.S. Horan, USMC. After completing his class, Rochefort was put in charge of the Research Desk while Safford performed his required sea duty. Thus Rochefort was in charge in 1926-1927 when Meyer succeeded in the initial solution of the Red Book ciphers and in discovering the "transposition forms." Later the various keys and forms used with a specific cipher were also solved by students.7|
In 1928 the Navy also established a school for enlisted Navy and Marine Corps intercept operators at the Navy Department in Washington, D.C. A classroom and eight intercept positions were erected on the roof of "Main Navy" probably as much for the sake of privacy as for the lack of space. Understandably, student graduates became known as the "On the Roof Gang." The first class began on 1 October 1928. Out of twenty students in the first class, seven finished. All seven were sent to Guam to open that station in 1929.8 Two classes, 5 and 15, were made up entirely of U.S. marines. The school operated until February 1941. Its objective was to train carefully selected military radio operators in specialized radio communications techniques, particularly Japanese intercept, traffic analysis, and simple cryptanalysis.9
Encouraged by these developments, Lieutenant Commander Arthur D. Struble, OP-20-G, early in 1929, drafted a letter to Admiral Charles B. McVay, CINCAF, which directed the expansion of radio intelligence service in his command, including activities in China. Significantly, Admiral Charles F. Hughes, CNO, who signed the letter, again mentioned that "major decrypting units are planned for Washington and Honolulu."10 Responding to this pressure from the Office of chief of Naval Operations, McVay opened shore stations at Guam (Station B) and at the U.S. legation at Peking, and created a position on his staff for a radio intelligence officer. A special cipher was supplied to enable this officer to maintain close liaison with OP-20-G in Washington.11 In addition, in late 1929 Rear Admiral William D. MacDougall, Commandant 16th Naval District, who was subordinate to CINCAF, opened an intercept station at a small naval base at Olongapo in the Philippines on Subic Bay facing the South China Sea. The site (Station C) was destined to move three times in ten years in an attempt to find secure operating spaces, living quarters, and antenna sites where Japanese Navy signals could be heard consistently. (Olongapo, 1930-34; Mariveles, 1934-35; Cavite, 1936-39; and Corregidor, 1940-42.)
Early War Plans
Because of a continuing perception after the end of World War I that war with japan would come sooner or later, the first efforts by the Navy to establish a COMINT collection capability in the Pacific were directed at the U.S. Asiatic Fleet. It was far from enough, however, and fell short of OP-20-G's goal of an even greater radio intelligence capability against the Japanese Navy. U.S. Planning for this eventuality is treated later.
In 1930, OP-20-G planners selected the 13th Naval District, which included Oregon, Washington, and Alaska, as well as Idaho, Montana, and Wyoming, as a prospective location for two new intercept sites: one, a large site to cover Japanese point-to-point traffic with Europe and China on low and high frequencies during wartime; the other, a small site in Alaska ("but not in the islands") to cover Japanese ship-to-shore communications in both peace and war.12 Because of budgetary restrictions, Admiral Pratt, CNO, was forced to wait until May 1932 before directing Rear Admiral E.H. Campbell, Commandant 13th Naval District, to establish the first of these sites at Astoria,
Oregon, where the Navy had a DF station providing navigation assistance to commercial vessels.13 Rather than build and equip a new site, OP-20-G planners were by then reduced to postponing delivery of the new equipment and asking Admiral Campbell to accept a plan in which a COMINT mission against Japanese targets was to be conducted using idle communications equipment.14 The initial COMINT mission was to copy Japanese diplomatic traffic on a commercial RCA circuit between Salinas, California, and Tokyo using idle DF receivers which had been tuned to the commercial band.15
In the 14th Naval District, Hawaii, an unofficial site established at Wailupe in 1925 was given official status in 1931 and authorized one billet by Admiral Pratt, CNO.16 Ineffective because of poor signal hearability, the Wailupe site was moved to Heeia in 1934. After the Japanese attack on Pearl Harbor in December 1941, to further improve reception and communications, the site was moved to Wahiawa.
With the exception of closing the sites in Peking in 1935 and Shanghai in 1940, the geographic posture of Navy COMINT in the Pacific retained this modest form until 1941. Small as the collection effort was, however, it is clear that Naval Communications had succeeded by 1941 in establishing a radio intelligence organization targeting primarily the Japanese Navy in the naval districts in the Pacific basin and the Asiatic Fleet. As noted by Wenger in 1937:
It was not an integrated organization, however, but a number of technically dependent but operationally independent units under the technical control of OP-20-G but under the military command of the two four-star fleet commanders (Pacific and Asiatic). Management control, such as it was, was exercised by CNO and administrative control by many subdivisions of the Navy Department and local commander activities (e.g., Bureau of Navigation, Bureau of Engineering, District Communication and Radio Material Officers, Fleet Intelligence Officers and Station Commanders to name but a few). When you consider that "control" was exercised over vast distances using mail sent by train and ship by personnel who frequently had no appreciation for the special problems faced by those being controlled, the wonder is that the system worked at all.17
From the early 1920s, OP-20-G participated in U.S. fleet exercises by furnishing cryptographic systems for contending fleet elements, the Battle (or Black) Fleet and the Scouting (or Blue) Fleet, and by training individual line officers in cryptanalytic skills. During the exercises volunteer cryptanalysts often succeeded in penetrating and exploiting the opposing fleet's cryptographic and communications systems. The most successful volunteers were officers usually chosen for the assignment by Safford. These officers demonstrated a flair for cryptanalysis by solving puzzles Safford placed in the monthly Communications Bulletins beginning in July 1924. Safford recruited Joseph W. Rochefort, Thomas Dyer, Joseph Wenger, E.S.L. Goodwin, Wesley A. (Ham) Wright, and Jack S. Holtwick in this way.
The U.S. fleet problems of 1929 and 1930 dramatically, albeit briefly, called command attention to the work of OP-20-G and provided important recognition for the work of cryptanalysts both in developing codes and ciphers for the fleet and in demonstrating decisively the vulnerability of insecure communications. U.S. Fleet Exercise #9 in 1929 was marked by successful exploitation by Navy cryptanalysts Safford, Rochefort, and Dyer of the Black Fleet against both the cryptography and communications of the Blue Fleet. Safford, Rochefort, and Dyer read all of the traffic of the opposing force (enciphered by a cumbersome cylindrical cipher) and made considerable progress in solving the signal cipher as well.18
In his summary critique of the exercise on 15 May 1929, Admiral Henry A. Wiley, CINC, U.S. Fleet, ruefully acknowledged that the successes realized by the decrypting units represented a "serious lack of knowledge (of radio security)."19 As a result, Safford
was directed to "make a cipher which could be broken but not allow messages to be read before the problem is over."20
In his memoirs Commander John W. McClaren, who was head of OP-20-G at the time, recalled the chaos created the next year when the fleet was required to use unfamiliar wartime cryptographic procedures, codes, and ciphers during Exercise #10.21 Not only was the fleet's lack of readiness amply demonstrated by the resulting confusion, but both decrypting units again did quite well against the fleet's codes, ciphers, and communications procedures as well. The Black Fleet decrypting unit consisted of Lieutenant Paul R. Sterling, Lieutenant Clarence V. Lee, Lieutenant Frederick D. Kime, Lieutenant Frank H. Bond, and Ensign William H. Leahy. They were assisted by six yeomen and three messengers. The Blue Fleet unit consisted of Lieutenant Llewellyn J. Johns, Lieutenant Wesley A. (Ham) Wright, Ensign R. Bennett, Ensign Lee W. Parke, three yeomen, and one marine sergeant. The Black unit recovered the Blue Signal Cipher. The Blue unit recovered the Black Signal Cipher, the Black Contact Code, and the Black Callsign System.
Fleet Exercise #11 was also held in 1930, and again each fleet had a decrypting unit. Johns, Brown, Wright, and Parke once again performed for the Blue Fleet assisted by three yeomen and one radioman, all Navy personnel. In the problem the fleets used codes specifically constructed by Safford which theoretically could not be broken during the exercise. In his critique of 7 April 1930, however, command in chief of the Black Fleet, Admiral Louis M. Nulton, reflected on the successes of each decrypting unit and leveled considerable criticism against designers of the "Recognition Signals," the "Contact and Tactical Report Code," and, to a lesser degree, designers of the callsign system. His critique stated that the signal cipher and the control code were "simple" and "could have been quickly broken down by expert decryptors [sic]." He further reported that the service cipher used to direct movements was unsatisfactory because of the time required to encipher and decipher messages. The recognition signals were too complicated for quick and effective use, though not impossible to memorize or initiate. The callsigns were also too long and too easily associated with the user, according to Nulton.22 Despite Nulton's criticisms, the exercises reinforced a growing conviction within the fleet that COMINT was a vital tool for commanders, and COMSEC an important prerequisite to success in battle.23
Japanese Fleet Capabilities and Intentions
In addition to working U.S. fleet exercises to make fleet communications more secure, cryptologic personnel overseas copied, analyzed, and, with assistance from Washington, exploited radio traffic from four Japanese fleet maneuvers between 1930 and 1935, demonstrating the benefits to strategic planning of communications intelligence derived from foreign military communications. The stations involved comprised Guam, Olongapo, Peking, USS Goldstar (AG-12), Los Banitos (Mariveles), and USS Augusta (Flagship Asiatic Fleet). Both Augusta and Goldstar normally were mobile detachments
taken from shore stations. At least thirty-three operators were assigned operations-related tasks for the Japanese maneuvers in 1935, which ran from July to September.24
Collectively, these stations intercepted the communications of Japanese ships at sea and from participating Japanese shore stations. The Japanese maneuver activity, at its height, typically extended from fleet anchorages in Japan to Saipan in the Marianas and the Palau Islands east of Mindanao. The COMINT reports prepared by personnel at the sites, and later consolidated by personnel in Washington who often had message text which supported the field's conclusions, superbly demonstrated both the tactical and strategic value of COMINT and established, at least in OP-20-G, the conviction that traffic analysis was an equal partner to cryptanalysis. Not only did these reports reflect the Japanese fleet's strategic capability to wage a large-scale successful war against the U.S. Asiatic Fleet, but they also revealed Japan's intentions to invade Manchuria, to defend the western Pacific in case of a U.S. attempt to interfere, and to conduct electronic countermeasures in the event the U.S. attempted to monitor fleet communications. The 1930 Japanese maneuver was seen by U.S. Navy analysts as a rehearsal for an invasion of Manchuria, which actually did occur in the following year. They also revealed plans for the complete mobilization of the Japanese fleet, a comprehensive knowledge on the part of the Japanese of the current U.S. war plan against the Japanese fleet, and the unpleasant fact that the Japanese Navy was superior in strength to the U.S. Asiatic Fleet. Regrettably, the Navy did not see fit to exploit this valuable planning asset by regularly tasking and funding the resources necessary for its continuation.
The 1933 report, for example, revealed details of Japanese plans to defend the western Pacific from a counterattacking U.S. fleet, actual ship movements, Japanese war plans against China, and a myriad of facts and details about air and sea deployment, tactics, communications practices and procedures, order of battle, and individual maneuver objectives.25 CINCAF, Admiral Frank B. Upham, was particularly impressed by the efforts of the COMINT analysts, whose work was based entirely on traffic analysis, since the Japanese Navy's operational code (the Blue Book) had not been recovered by the time of the exercises. Not only did he visit Station C to personally compliment the men, telling them that one day their work would be of tremendous importance to the nation, but he prepared a unique endorsement for the report. His endorsement, forwarded to Admiral William H. Standley, CNO, on 20 June 1934, contained several significant "COMINT discoveries" based on traffic analysis, including one entitled "Indications of Approaching Hostilities." This prophetic paragraph predicted that "any attack by (Japan) would be made without previous declaration of war or other intentional warning." In keeping with its traffic analysis origins, another finding stated that "preparations would be noticeable in increased radio activity." Admiral Upham also recommended a plan for observing movements of Japanese merchant ships. He believed Japan would try to save as many of these vessels as possible by withdrawing them to Japan prior to any outbreak of war. Ironically, the U.S. Navy did detect such a movement in November 1941.26 Unfortunately, by the time of Pearl Harbor, Admiral Upham was dead, and his report and recommendations lay forgotten in CNO's files.
The 1933 Japanese maneuvers were also noteworthy for two other features which in themselves speak volumes on the state of development of communications intelligence in the Japanese and American navies. Analysts aboard the Goldstar were aware that the Japanese practiced deliberate electronic countermeasures to prevent the CI unit aboard the Goldstar from successfully intercepting their naval maneuver traffic.27 In this regard, it is interesting to note that Operation Problem #IV in 1933 at the Naval War College showed that the United States could not successfully defend U.S. interests in the western Pacific and specifically could not recapture the Philippine Islands or hope to maintain a base in these islands. This action clearly demonstrated Japanese awareness of the value of communications intelligence. On the negative side, however, the final 1933 U.S. naval report on these maneuvers was not completed until 1937! In May 1937 Wenger, in a personal letter to Holtwick at Station C, commented on his final efforts to finish a second report before the end of April. This final report was a labor of both love and curiosity. Wenger had played a major part in recognizing the value of the work done in the field and had collated this work into a final report in 1934. In 1937 he wanted to evaluate the traffic analysis results against Blue Book recoveries not available earlier to see if the messages contained any unique information.28 Thus, largely because of a lack of manpower, five years elapsed between the event and the Navy's final COMINT report on a significant Japanese fleet exercise which revealed many unique Japanese Navy capabilities.29
Another important contribution to the U.S. Navy's efforts to determine the capabilities of the Japanese Navy occurred in 1936 when the cryptanalysts and linguists in Washington translated a message giving the results of the Japanese battleship Nagato's postmodernization trials. This message greatly alarmed U.S. officials because it contained the Nagato's new top speed, which was in excess of twenty-six knots, the same as four new Kongo-class battle cruisers and considerably in excess of the twenty-four knot top speed currently planned for the redesigned U.S. battleships North Carolina and Washington. By inference the Nagato's speed would be the prospective speed for other battleships being modernized and the minimum speed for new battleships of the Yamato class. As a direct result of this message, U.S. naval officials raised the required speed of modernized U.S. battleships to twenty-seven knots and of new vessels to twenty-eight knots.30
1. For many years communications intelligence was also known in the Navy as traffic intelligence if derived from traffic analysis and radio intelligence when derived from decrypted messages.
2. Since 1912, when Admiral George Dewey, chairman of the General Board, decreed that military men should limit themselves to "purely military questions," U.S. Navy war planners did not consider either domestic of foreign motives for war. This situation prevailed until at least the 1960s. Ronald Spector, Professors of War: The Naval War College and the Development of the Naval Profession (Newport, R.I.: Naval War College press, 1977). Cryptography is the development of codes and ciphers; cryptanalysis is the exploitation of codes and ciphers.
3. 18 August 1917 memorandum from Radio Officer to the Commanding Officer, Series II.E.6, Center for Cryptologic History (CCH) History Collection.
4. Notes and Comments on the Necessity for and the Organization of a Cipher Bureau, 1918; A summary of the Organization, Activities, and Achievements of the Code and Cipher Section of the Military Intelligence Division (M18), November 1918, by Chief Yeoman H.E. Burt, USN, Series II.E.62 and II.E.63, CCH History Collection (classified).
5. The Birthday of the Naval Security Group SRH-150; Record Group 457, Records of the National Security Agency, National Archives; hereafter RG 457, NA; a series of declassified ONI memorandums involving Chief Radioman (CRM) H.A. Williams on 12, 15, and 16 April 1924 and 17 July 1924 clearly show their interest in the quality of intercept material supplied by Station A. ONI no doubt would inform the Chief of Naval Operations if anything of naval or national interest was learned from the messages. Chief Williams was stationed at Shanghai at the time. Series II.E.36, CCH History Collection; Jeffrey M. Dorwart, Conflict Of Duty, The U.S. Navy's Intelligence Dilemma, 1919-1945 (Annapolis, MD: Naval Institute Press, 1983), 41-42.
6. Jack S. Holtwick, Captain, USN (Ret), unpublished manuscript, "Naval Security Group History to World War II," SRH-355, 67, RG 457, NA; anecdote about Navy Slush Fund and establishment of Research Desk was taken from Captain Laurence F. Safford, unpublished manuscript, "The Undeclared War, History of RI," SRH-305, RG 457, NA. For a biographical background of Hooper, see L.S. Howeth, Captain, USN (Ret), "History of Communications Electronics in the U.S. Navy," (Washington, D.C.: Government Printing Office, 1936), in which he is called the father of naval radio.
7. SRH-355, 45-48. RG 457, NA.
8. Ibid., 52.
9. Dedication of the memorial to the "On the Roof Gang," Series III.H.26; Interview Duane L. Whitlock, Captain, USN (Ret), 11 February 1983, by Robert Farley, OH-05-83, NSA (classified).
10. 13 May 1929 memo from CNO to CINCAF, Appendix A, SRH-180., RG 457, NA.
11. SRH-355, 223, footnote. RG 457, NA.
12. COM-13 correspondence with OP-20 in July and August 1930, Series II.H.17, CCH History Collection (classified).
13. 22 May 1932 letter from CNO to Rear Admiral E.H. Campbell, COM-13, Series III.H.17, CCH History Collection (classified); see also SRH-355, 67, RG 457, NA.
14. 22 March 1932 memorandum to COM-13 from CNO, Series III.H.17, CCH History Collection (classified) and 27 April 1932 memo same to same, Series III.H.17.1, CCH History Collection (classified); see also SRH-355, 67, RG 457 NA.
15. Interview John H. Gelineau, CWO, USN (Ret), 4 October 1983, by Robert Farley, OH-22-83, NSA (classified).
16. 11 June 1931 memorandum CNO to COM-14, Series III.H.17, CCH History Collection (classified). Without explaining what "unofficial intercept activities" meant, CNO ordered the station "reestablished."
17. Navy records for December 1941 show the following dispositions: Collection -- Total number of receivers, 68: Guam-9; Corregidor-25; Heeia-21; Bainbridge Island (formerly Astoria)-13, SRH-197 "U.S. Navy Communication Intelligence Organization, Liaison and Collaboration, 1941-1945," RG 457, NA; see also SRH-355, 243-44, RG 457, NA.
18. SRH-355, 50-51, RG 457, NA.
19. Memorandum from CINC U.S. Fleet to Chief of Naval Operations dated 15 May 1929, Series III.H.1, CCH History Collection (classified).
20. 10 May 1930 letter Safford to Dyer, Series III.H.10.3, CCH History Collection.
21. SRH-355, 391. RG 457, NA.
22. U.S. Fleet Problems X and XI, Series III.H.10.3, CCH History Collection.
23. Part IV, Communications, in "Command Conclusions on U.S. Fleet Problem 9," Series III.H.1, CCH History Collection (classified). This is a very brief paraphrase of the paragraph on radio security.
24. SRH-355, 155. RG 457, NA.
25. SRH-222, 223, 224, 225, RG 457, NA. Various reports on Japanese Grand Fleet maneuvers 1930, 1933, 1934, 1935.
26. SRH-355, 131, RG 457, NA; see also Gordon Prange with Donald M. Goldstein and Katherine V. Dillon, The Verdict of History (New York: McGraw-Hill, 1986), 134. On 4 November 1941, an OPNAV message reported this fact to CINCPAC, CINCAF, and others. See also Safford, SRH-305, 20. RG 457, NA.
27. SRH-355, 118-119, RG 457, NA.
28. Ibid., 222-223, RG 457, NA.
30. Captain Laurence F. Safford, unpublished manuscript, "A Brief History of Communications Intelligence in the United States," SRH-149, RG 457, NA.
22 June 2001