After an abortive attempt at across-the-board modernization in the early Depression years, virtually all OP-20-G's attempts to increase manpower and improve equipment during the 1930s were directed toward the Pacific basin where the Japanese threat was seen as paramount. In 1937, OP-20-G opened the long anticipated major research unit in Hawaii with the task of supporting Washington's efforts.81 Lieutenant Commander Thomas Dyer was detached from fleet duty and assigned to COM-14 as a cryptanalyst. His duties were to establish a decrypting unit and undertake research work on "M1 Orange Naval Cryptographic System."82 Message files for 1935, 1936 and 1937 were supplied by OP-20-G and sent to Dyer via the USS Chaumont. COM-16 sent copies of all traffic, including messages intercepted by Stations A, B, and C to Dyer, and Washington mentioned that IBM equipment would be forthcoming at once. At the outset, when not occupied by other duties assigned by COM-14, Dyer, who was not a Japanese linguist, single-handedly attempted to recover all keys as they appeared. He naturally forwarded all solutions to Washington for translation. In 1939, when the M1 system had been exhausted, Dyer was assigned cryptanalytic responsibility for the Japanese Navy's Flag Officers Code. Lieutenant Joseph Richardson appeared the following July as language officer.
IBM equipment promised in 1937 was not immediately forthcoming. In February 1938, however, OP-20-G notified Dyer that he was to receive IBM tabulating machinery and two clerks. Taking a little of the glow from the moment, Dyer was also told that after becoming familiar with the equipment he was to train the clerks himself. Typically, he was instructed to provide "material assistance to Washington" as soon as possible.83 Again, the manpower problem imposed severe limitations on potential U.S. successes against the Japanese codes.
In September 1939, all restriction on increasing personnel and installations were removed when the U.S. declared a national emergency. A few weeks later most of the earlier "Neutrality Legislation" was overturned. Nevertheless, manpower shortages continued to plague the Navy COMINT program, particularly in the Pacific, well into
1941.84 In June 1940, Admiral Claude C. Bloch, COM-14, requested more manpower (twenty-one billets) to expand his COMINT operation.85 In his favorable endorsement to the Bureau of Navigation, which handled all Navy personnel matters, Admiral Harold R. Stark, CNO, made the following observation: "The main obstacle to expansion [of communications intelligence activities] is not the matter of increased allowances but finding suitable personnel to fill existing allowances. Many of the reserve personnel in DNC's mobilization slate are reluctant to leave their civilian occupations prior to full mobilization, and COM-14 has been unable to find any suitable volunteers."86 Later in 1940 Safford complained, "We are allowed 75 and actually have 55."87
Compounding the continuing manpower shortage by adding inordinate delays to the system of exchanging crypt recoveries between Washington and the filed was a severe and perennial communications problem which affected all Navy COMINT initiatives, particularly those against the Japanese. Therefore, even before they broke down completely when war began, the primitive methods of U.S. Navy communications and the centralization of cryptanalytic functions proved to be major liabilities which prevented realization of the ultimate organization goal, support for the fleet commander. The addition of officer-cryptanalysts to Station C and to the centers in Hawaii was not successful in this regard because code and key system recoveries from Washington, when available, were delayed by the surface transportation system of commercial vessels and railroads. As a result, completely current English texts were probably a rarity in the Philippines before the war and were never seen in Hawaii until March 1942.
The individual radio intelligence officer-analysts assigned to the Asiatic Fleet and the 14th and 16th Naval Districts between 1930 and 1940 were, by today's standards, almost completely isolated from Washington. Communications between Washington and its far-flung resources in the Pacific continued to be primitive until long after 7 December 1941. Messages and intercept logs, reports and professional correspondence, if classified, were painstakingly enciphered by the radio intelligence officer himself using special equipment and instructions.88 If transmitted as messages on manual morse circuits or landlines, they were delivered to the communications center where they were again enciphered. Material such as traffic logs and routine Japanese messages, however, were always sent home by mail. A package was usually forwarded once a week from Hawaii, Guam, and the Philippines. It would be put aboard a commercial ship or a station-keeping vessel like the USS Chaumont, which traveled the Pacific from California to China. After 1935, a small amount of mail could be sent via the Pan American Airways "Clipper" using a small strongbox built into the hull specifically for that purpose. Keys to open the strongbox were held by authorized officers at appropriate points along the route. The code known as RIP 30 was developed specifically for air mail letters.89 The introduction of air mail service in 1935 reduced travel time from weeks to days, but the need for major improvements in communications before COMINT would become a useful instrument during wartime was clear.90
Circumstances Favor Diplomatic Targets
Electrical communications within the continental U.S. were only slightly better than those overseas, and only in rare cases did they serve to speed the flow of information to Washington from abroad. In August 1940, the Navy had five sites with diplomatic targets which were all linked directly (or indirectly through Army circuits) to Washington via radio and landline communications. These sites were Winter Harbor, Maine; Amagansett, New York; Cheltenham, Maryland; Jupiter, Florida; and Bainbridge Island, Washington.91 Radio communications with Hawaii consisted of single channel morse links between Washington, San Francisco, and Pearl Harbor. Landline communications consisted of the relatively higher capacity government and commercial teletype circuits owned or used by each military department. The Army owned 401 mainly east-west circuits with 1,003 machines. The Navy owned forty-four mainly north-south circuits on the coasts, the first of which was not established until 1941. These circuits served naval activities at Washington, Norfolk, Philadelphia, New York, New London, Boston, and Portsmouth, New Hampshire.92 In March 1941, a commercial teletype line between Winter Harbor, Amagansett, and Bainbridge Island was inaugurated.93 Despite the availability of teletype circuits from the West Coast to Washington, D.C., there is no indication that packages containing Japanese intercept from the Pacific which arrived by either air or sea throughout the 1930s were opened until they reached Washington through normal overland channels. In 1939, triple-wrapped packages from Guam containing the daily bundles of messages were received by the "Courier Station" in San Francisco, where they were opened and placed in registered mail. The inner package was marked, "To be opened only by OP-20-G." In the Philippines during the 1930s, intercept material was mailed to Washington in a similar manner.94
By 1941, the mission constraints in Corregidor and Hawaii improved by their limited cryptanalytic capabilities and the pervasive shortages of all types of manpower in Washington contributed to a growing sense of alarm in OP-20-G. As the year progressed, certain daily summaries produced by Hawaii and Corregidor, particularly those which sounded warning signals, were no doubt marked for electrical forwarding. But the capacity of manual morse circuits and the inherently slow moving manual decryption features of Navy communications between the mainland and overseas stations were contributing to a serious information gap between Washington and the fleet-supporting field activities.
In July 1941, the nucleus of what would become OP-20-GC (communications) in early 1942 was formed using personnel from OP-20-GY (cryptanalysis). The objective of this new element was to encrypt, decrypt, route, deliver and file COMINT dispatches between OP-20-G and the outside world. In theory this organization was intended to be much faster and much more secure than the Navy Code Room. The whole affair, unfortunately, was
undertaken on an incredibly small scale. By December 1941, it consisted of two reserve ensigns, neither of whom were trained or experienced in communications. Under these circumstances -- which seem doomed to fail -- it is little wonder that OP-20-G's customers were attracted to readable diplomatic messages collected by mainland sites (see Chart A). It is also little wonder that, at this time, the principal source of crisis-related communications intelligence available in Washington prior to Pearl Harbor was Japanese diplomatic traffic.95
Three events involving the sites at Cavite and Heeia, Hawaii, from 1938 to 1941 will strikingly illustrate how truly primitive were the communications which served the Navy's COMINT function overseas. In September 1938, Lieutenant Jack S. Holtwick, then the CINCAF Radio Intelligence Officer, complained to OP-20-G about the lack of electrical communications between the unit at Cavite and the flagship. He said that "it now takes days to obtain COMINT information needed to prepare a daily status report." In 1940, Hawaii commented on tracking Japanese naval vessels during annual maneuvers stating that "the only helpful direction finding came from the Philippine unit by Clipper mail!"96 Finally, on 5 January 1940 Admiral Stark, CNO, requested the Bureau of Engineering to connect the site at Heeia to an Army cable which then terminated at Kailua, eight miles away. Stark also requested the engineers to arrange for an intercom between the communications intelligence unit at Pearl harbor, the Lualualei direction finding site, and Heeia, also by Army cable, using "other than teletype instruments." These arrangements were meant to replace the public party line telephone service. By 7 December 1941, the work had not been done, and with the loss of telephone service in the attack, there were no communications between Heeia and Pearl Harbor (about thirty road miles) except by vehicle!97
Cooperation With Allies
Until August 1941, efforts to recover JN-25B code values were restricted to the British force at the Far East Combined Bureau (FECB), Singapore, and four U.S. officer-linguists at Corregidor, working in close collaboration with the British. In August 1941, OP-20-G, Washington, began to help with JN-25B code recovery but was hampered by lack of linguists familiar with Japanese naval terminology and usage and by the slow communications available at the time. The only current JN-25 messages read by U.S. analysts on Corregidor during this period were few in number and were invariably ship movement reports: arrivals and departures, together with some fragmentary schedules. In view of the full collaboration and exchange with FECB, Singapore, there is no reason to believe that the British exceeded the U.S. accomplishments.
The Move to Hawaii
On 7 May 1940, the U.S. fleet moved its headquarters from San Pedro, California, to Pearl Harbor. The move was undertaken with great reluctance by Admiral James O. Richardson, Commander in Chief, U.S. Fleet. Richardson and most Navy officials who opposed the move thought a fleet anchored in Pearl Harbor would be unnecessarily exposed to Japanese naval strength. President Roosevelt, however, considered the move as a necessary countermeasure to growing Japanese bellicosity. Throughout 1940 Richardson bitterly voiced his objections to relocating his headquarters to Pearl Harbor because it challenged the soundness of U.S. policy in the Pacific. He claimed that Pacific naval offensive -- the heart of the Navy's War Plan Orange -- was sure to fail because the U.S. did not have the capability to support an offensive west of Hawaii. He also noted a factor not considered by the war planners: the U.S. was now vulnerable to attack in the Atlantic and the Caribbean. In January 1941, Roosevelt ordered him relieved. His replacement was Admiral Husband E. Kimmel who, at the same time, was designated Commander in Chief, U.S. Pacific Fleet (CINCPACFLT).
During this period of great internal upheaval in the Pacific Fleet, two relatively insignificant events occurred which actually marked the beginning of a close COMINT relationship between that fleet and OP-20-G. On 7 December 1940, exactly one year to the day before Japan attacked Pearl Harbor, Lieutenant Commander Edwin T. Layton, a Japanese linguist with past experience in OP-20-G, became the Fleet Intelligence Officer, and a few months later, Commander Joseph J. Rochefort, who was the only man in the Navy who was both a cryptanalyst and a Japanese linguist, became OIC of the 14th Naval District's Radio Intelligence research effort. Layton and Rochefort were old friends from sea duty and from language training in Japan. Rochefort was assigned to the fleet when he was transferred to COM-14. Both knew of the OP-20-G operation, having served under Safford in the 1930s. Layton served only briefly, but Rochefort had received extensive training as a cryptanalyst. They quickly established a close working relationship, and the liaison would soon prove immensely beneficial to the U.S. Pacific Fleet.98
Both Rochefort and Layton were called upon regularly to brief Kimmel on what COMINT revealed about the Japanese Navy. In one instance Layton's analysis of callsign and address usage, which he had undertaken during 1941 at Rochefort's request, was sent to Washington on the order of Admiral Kimmel.99 His conclusion that the Japanese had begun a military buildup in the Mandate Island (Marianas, Carolines, and Marshalls) was a development which had gone unnoticed by COMINT analysts in Washington, Unexpectedly, rather than foster good relationships between Pearl Harbor and Washington, this episode caused considerable ill feeling toward Layton and Rochefort. It may also have marked the beginning of an unhealthy intramural OP-20-G rivalry between the Washington and Hawaiian centers over the issues of COMINT reporting responsibilities and Japanese intentions which persisted well into 1942.100
Support to the Pacific Fleet
From July 1941 onward, the COMINT research unit in Hawaii under Rochefort prepared daily COMINT summaries for Admiral Kimmel (see Appendix C). They were based on analysis of Heeia collection and to some extent on technical and intelligence information from Corregidor. Hawaii's analytic contributions to the summaries were based on traffic analysis of message externals and direction finding results since the Flag Officer's Code could not be read, and they had no capability against the Japanese Fleet Operational Code (JN-25). These summaries were characterized by Layton after the war as containing "no hard intelligence." This is a harsh judgment. Individually, though it is true that they contained no Japanese message texts, the summaries constituted the substance of Layton's daily reports to CINCPAC. Collectively they revealed a wealth of information concerning Japanese naval activities, particularly those under way in the Mandates, on the islands of Hainan and Taiwan, and along the Chinese coast (see Appendix C).
In many respects COM-14's efforts and achievements in 1941 were similar to what had been accomplished at Station C with traffic analysis against the Japanese Imperial Fleet maneuvers in the 1930s. The COM-14 daily summaries clearly showed that Lieutenant Thomas A. Huckins and Lieutenant John A. Williams, who headed the traffic analysis unit, had solved both the strategic and tactical Japanese naval communication structures. They understood the callsign generation system and were able quickly to reestablish order of battle data after routine callsign changes. This insight permitted unit identifications to the squadron level in ground-based-air and destroyer units. It also
allowed identifications to the individual ship level in battleships, cruisers, and carriers. The capability to exploit these features of Japanese Navy communications lasted until about three weeks prior to the attack on Pearl Harbor when callup and addressing procedures changed abruptly.101 Throughout the period they were also able to use their direction finding capability to produce unique information as well as to support evidence from traffic analysis. The traffic analysis unit was able to identify the Japanese Navy mainline shore establishment from Imperial and Combined Fleet Headquarters to principal line and staff subordinates within each of the fleets in both home and deployed locations. Based on the content of their daily summaries, it is conceivable that communications being intercepted by Hawaii (Heeia) in 1941 encompassed the entire Japanese Navy communications system ranging from Japan to South China, to the Mandate Islands, and to the connecting ocean area.102 Intercepted messages were mailed to Washington for exploitation of their texts.
Japanese Intentions Revealed
As early as July 1941, traffic intelligence reports (i.e., reports founded on traffic analysis) prepared for Admiral Claude C. Bloch, COM-14, and Admiral Kimmel, CINCPACFLT, reflected Japanese air and naval concentrations "awaiting the assumed Southern operations." In fact, from July until 6 December, summaries from Hawaii made frequent allusions to the "formation of Task Forces" and forthcoming "hostile actions" and called attention to similarities between current activities and those which preceded earlier Japanese naval and military campaigns in South China and Indochina.103 Bearing in mind that Hawaii could not read the message texts, the accuracy of these reports was truly remarkable.
Support to Asiatic Fleet
Corregidor too was very active in following Japanese fleet naval and air movements throughout 1941, producing reports for much of the year in technical channels, which included the CINCAF radio intelligence officer and the Hawaiian Research Center. It was not until October 1941 that Station C's technical reports began to appear as daily intelligence summaries. Rochefort's daily reports often contained information derived from reports from Corregidor.104 In late November, because of its scope and the station's central location, COM-16's perspective was judged to be superior to Hawaii's and to the fragmentary and often conflicting reports from other sources such as attachés in Shanghai, Chungking, and Tokyo. On 24 November 1941, Admiral Stark, CNO, ordered Admiral Hart, CINCAF, to receive, evaluate, and combine all reports and conclusions, including those from COM-14, reporting directly to CNO with information to Admiral Kimmel, CINCPAC.105 In the two weeks of peace remaining before Pearl Harbor, this order had little or no effect on events.106
While the United States attempted to maintain a level of strategic equality with Japan in the Pacific by offsetting losses of capital ships sent to the Atlantic with a buildup of long-range air power, the Japanese government formulated plans for war in the Pacific. The Japanese war plan for the Western Pacific campaigns began to unfold well before 10 November 1941 when General Count Hisaichi Terauchi, commanding the Southern Army, and Admiral Isoroku Yamamoto, commanding the Combined Fleet, formally concluded a "Central Agreement" which outlined an ambitious scheme of Japanese conquests.107 According to the agreement, the first operational stage was divided into three phases: (1) attacks on the Philippines, Malaya, Borneo, Celebes, Timor, Sumatra, and Rabaul (also Guam, Wake, and Makin); (2) capture of Java and the invasion of southern Burma; and (3) conquest of all Burma. The Japanese then envisioned pacification of the area, the creation of a Greater East Asia Co-Prosperity Sphere, and probably a defensive struggle against the United States to maintain their hold on the region. A second operational stage also covered by the agreement was to "occupy or destroy as speedily as operational conditions permit," eastern New Guinea, New Britain, Fiji, Samoa, the Aleutians, Midway, and strategic points in the Australian area. According to historian John B. Lundstrom, this is as far as Japanese planning went.108
The portion of the Japanese Navy which was to execute an attack on Pearl Harbor and provide cover and escort for the remainder of these operations had been preparing for its various roles for several weeks. It consisted of 10 battleships (BB), 6 cruisers, 112 destroyers (DD), and 65 submarines (SS). In addition, Japan had large numbers of auxiliary vessels, tenders, minesweepers, and escorts. The fleet was organized into nine naval stations in the homeland area, the China Area Fleet and the Combined Fleet. The Combined Fleet, which consisted of five mobile fleets (1st, 2nd, 6th, 1st Air Fleet, 11th Air Fleet) and three localized fleets (3rd, 4th, and 5th), was destined to carry the burden of the southern strategy as well as to conduct the strike on Pearl Harbor.109
In the opening campaigns of the first phase, the Combined Fleet was divided into four task forces: Force 1 was a carrier strike force consisting of all six fleet carriers, two battleships, and three cruisers under Admiral Chuichi Nagumo. It was to conduct a separate attack on Pearl Harbor. Force 2, the South Seas Force (4th Fleet), extensively reinforced with land-based air units from Japan and submarines from the 6th Fleet, under Vice Admiral Shigeyoshi Inoue (CINC4), was to seize Rabaul, Wake, Guam, and Makin using a reinforced infantry regiment of 5,000 men (the South Seas Detachment). Force 3 consisted of fighting units from the 2nd and 3rd Fleet, the 11th Air Fleet, and the China
Area Fleet, under Vice Admiral Nobutake Kondo (CINC2), carriers from CarDivs 3 and 4, light and heavy cruisers and destroyers from the entire Combined Fleet, as well as hundreds of troop transports, supply vessels, escort vessels, and oilers, and the Southern Army under General Count Hisaichi Terauchi. It was to attack the Philippines, Thailand, and Malaya (the Kra Peninsula and Singapore). It was to follow up this with attacks on the Netherlands East Indies and Burma.110 In addition to providing escort and cover the Malay-Thailand invasion, the role of the 2nd Fleet included being Force 4, a "Distant Cover Force" for the forces invading the Philippines. Command of naval forces directly covering invasion of the Philippines was given to CINC 3rd Fleet, Vice Admiral Sankichi Takahashi.111
In addition to the vessels and their escorts, the Strike Force consisted of three submarines, I-19, I-21, and I-23 on Ship Lane Patrol, 2 DDs as Midway Neutralization Unit (presumably the same unit cited in the message of 16 November 1941, shown in Appendix A) and a train of eight tankers and supply ships.112
Details of the formation, training, and assembling of each of these Japanese naval elements (except for the Pearl Harbor Attack Force), as well as the supporting Japanese air elements involved in the Southern operations, were reported by the COMINT centers in Hawaii and Corregidor. Specifically, they observed Japanese air and naval forces gathering in the vicinity of Takao and Keeling on Formosa and Mako in the Pescadores, a group of islands between Formosa and China. They also noted Japanese assault forces gathering on Amami O Shima north of Okinawa and in the Palau Islands in the Mandates. Air support for the Philippine assaults was also seen assembling in the Palaus and on Formosa.
Because JN-25 messages as well as naval messages in other cryptsystems were largely unreadable, throughout the last few months of 1941 the messages were usually exploited for what their externals revealed (e.g., addresses, callsigns, association with others) and sent to Washington, where concentrated work on code and key recoveries was conducted. With some exceptions, the callsign change on 1 November seriously complicated the work of traffic analysis by introducing at the same time new procedures for addressing messages in which individual units were no longer called or addressed openly in the externals. (See Appendix C, note by COM-14 on 6 November.) Accordingly, when this practice was recognized, the record suggests that both COM-14 and COM-16, while still able to follow developments in the southern areas, had failed to establish continuity on the 1st Air Fleet callsign, which was noted and first identified on 3 November by COM-14 (see Appendix C). Moreover, COM-14 apparently neglected to review October traffic in which this fleet was also active and to make the correct associations regarding 1st Air Fleet organization. Regrettably, the record of Appendix A and Appendix C also suggests that between 1 and 17 November only message traffic which could be associated with pre-1 November southern area activity was examined even for its externals. The residue, including traffic pertaining to 1st Air Fleet activity, was apparently sent to Washington, which had no traffic analysis capability at this time and was concerned with only the cryptographic technicalities.113
Between early September and 4 December 1941, U.S. COMINT units at Pearl Harbor, Corregidor, and Guam intercepted and forwarded to Washington many thousands (26,581) of Japanese naval messages in the fleet general-purpose system (JN-25), a fleet minor-purpose system, a merchant vessel-navy liaison system, a merchant vessel-navy five-letter cipher, and a naval attaché cipher. Hawaii had no capability against JN-25, however, and because shortages of manpower in Washington precluded both code and cipher exploitation, none of these systems were read on a current basis even though Corregidor may have been nominally responsible for their exploitation.114
Had these messages been exploitable at the time, their stunning contents would have revealed the missing carriers and the identity of other major elements of the Strike Force.115 Not only did the surviving messages (which were finally decrypted and translated in 1945 and 1946) provide the identity of the 1st Air Fleet's Strike Force, but they revealed the Strike Force's objective through analysis of its exercise activities and its movements prior to 26 November 1941. (See Appendix A.)116
The method of attack and objective of the Japanese Strike Force were revealed in messages intercepted between 21 October117 and 4 November 1941.118 On 21 October, Carrier Divisions 1, 2, and 5 began a series of exercises and training maneuvers which involved specially modified torpedoes.119 These exercises, which probably ended on 6 November 1941, when CarDivs 1 and 2 were "to fire (torpedoes) against anchored capital ships" in Saeki Bay, amply demonstrated that the Strike Force had a naval objective.120 Furthermore, the extraordinary measures taken by the Combined Fleet to insure adequate fuel supplies for the Strike Force demonstrated that the naval objective was at a distant point far removed from shore-based fuel and even beyond the normal Japanese resupply capability. Between 4 October and 1 December 1941, the Chief of Staff Combined Fleet, CINC 1st Air Fleet (commander Strike Force), units of the Strike Force, and many Japanese navy yards exchanged messages which revealed that three of the carriers (Akagi, Soryu, Hiryu) would carry fuel oil as deck cargo and in spare fuel tanks,121 that additional oilers had been requisitioned into the Strike Force and modified for refueling at sea,122 and that carriers and their escorts would conduct extensive practice of refueling while under way.123
By 12 November 1941, the carriers in the Strike Force had completed necessary repairs and had returned to their respective home ports or navy yards. Virtually all preparations for the Pearl Harbor assault had been completed. Two exceptions were the final deployment of the Strike Force to its point of departure, Hitokappu Bay in the Kuriles,124 and completion of modifications to some oilers which were probably those involved in refueling the Strike Force on its return trip.125 On 11 November 1941, however, CINC 1st Air Fleet issued a routine movement message containing a plan for anchoring at an unspecified future date CarDivs 1, 2, and 5 and several escort units and Maru (commercial) vessels in Saeki Bay in the Inland Sea.126 There was no message confirming the fleet's arrival and, while it is entirely possible that not all elements of the Strike Force deployed to the Kuriles, the routine-appearing message, augmented on 1 December 1941 by deceptive radio broadcasts from Tokyo,127 probably represented an
attempt on the part of the Japanese to deceive U.S. monitors. Other Japanese naval messages now available clearly indicated that the Strike Force would be at sea by that date.
On 9 November 1941, the Commander Destroyer Squadron 1, a Strike Force unit (Chart A), while coordinating his activities with the Naval General Staff (NGS) Tokyo, sent a message which revealed that on 15 November 1941 Fleet carrier Hiryu of CarDiv 2 would be conducting a refueling drill of Ariake Bay while towing the Kokuyoo Maru.128 In addition, examination of movement reports between 17 and 20 November 1941 reveals that the Strike Force flagship at that time was the battleship Hiei and that it was located at Hitokappu Bay (approximately 45-00N 147-40E).129 Finally, on 19 November 1941 CINC Combined Fleet announced to all flagships a communication exercise on 22-23 November 1941, which excluded "the forces presently en route to the standby location."130 Collectively, although not definitively, these messages strongly suggest that since 15 November 1941, instead of anchoring in Saeki Bay, major elements of the Strike Force had in fact been at sea probably moving to the high north latitudes of the Kuriles or, in the case of late departures, toward the east on the 30° line.
While the above information was not available at the time, daily traffic intelligence reports based on traffic analysis of communications of the Japanese Second, Third, and Fourth Fleets concerning events in the western and west-central Pacific areas were produced by both Hawaii and the Philippines. These reports were mailed to Washington where, after about two weeks en route, they formed the basis of biweekly OP-20-G summaries prepared for ONI.131 Although the material was at times more than a month old, a factor which became critical in November and December 1941, officials on OP-20-G did have access to the same Japanese naval COMINT available to Kimmel at Pearl Harbor and Hart at Manila.
On occasion, such as on 26 and 27 November (see Appendix C), COM-14 and COM-16 COMINT summaries, because of their content, were sent to Washington as messages. These particular messages, though considerably less alarming than others issued by COM-14 during the October-November 1941 period, appeared at the same time as the famous "Winds" messages translations (see Appendix B) and contributed to the developing sense of crisis in Washington. Hawaii's report for 26 November 1941 was a comprehensive summary of the Japanese naval and air buildup assembling for a southern operation. It conveyed a distinct sense of alarm at events. Corregidor's report for 27 November identified in even greater detail the existence of both a Japanese Southern Force and a Mandates Force, including several Japanese ground force units (Base Forces) in the Mandates.
Hawaii's picture of the Japanese buildup was not as complete as it might have been, based on the details developed in their earlier summaries. COM-16's message confirmed and enlarged on COM-14's speculation regarding Japanese carriers in the Mandates (i.e., CarDiv 3, Ryujo and one Maru vessel). In a curious and unexplained reversal, however, COM-16 stated that COM-14's reports could not be confirmed. It was also in this confusing context that COM-16 reassuringly and incorrectly reported that as of 26 November 1941
"all First and Second Fleet carriers are still in [the] Sasebo-Kure area." The two summaries from Hawaii and Corregidor on 26 and 27 November 1941, respectively, were thus unique not because of their imperfections but because they clearly showed Washington the current military situation in the Pacific as perceived by radio intelligence centers in the Pacific and Asiatic Fleets. It was entirely possible, as Layton later claimed, that the OPNAV warning message of 29 November 1941 was a direct result of the impact of these summaries on the Chief of Naval Operations. In view of the evidence, however, an even more likely possibility was that all the OPNAV warning messages of November were stimulated by COMINT: Japan's hostile intentions from the diplomatic messages and the likely targets from the daily COM-14 and COM-16 summaries which had inexorably found their way into the consciousness of official Washington.132
81. Both Laurence Safford and A/CNO RADM Royal E. Ingersoll indicated (Pearl Harbor Attack Hearings Before the Joint Committee on the Investigation of the Pearl Harbor Attack, Congress of the United States, 79th Congress, hereafter PHA, part 26, 388 and part 9, 4241, respectively) in their testimonies that they believed Hawaii was responsible to CINPCAC; Safford in his testimony to the Hart inquiry, Ingersoll while being cross-examined by Kimmel during the congressional investigation. it was Safford, of course, who in 1937 opened the office, assigned Dyer and others to COM-14, and allocated all cryptanalytic tasks until 1942.
82. Memorandum from CNO to COM-14, 21 August 1937, "Cryptanalysis, Orange M1 System" Series III.H.17, CCH History Collection (classified). The M-1 was a Kana cipher machine used by the Japanese Navy from July 1933 for about four years. It is described in RIP28. Jack S. Holtwick, A Guide to Foreign Cryptographic Systems: U.S. Navy Short Titles, Cover Names, and Nicknames, 14 June 1971. NSA Cryptologic Archival Holding Area (classified).
83. 14 February 1938 memorandum CNO to COM-14, "Hawaiian Decrypting Unit, Equipment and Personnel for," Series III.H.17 CCH History Collection (classified).
84. Prange, Verdict of History, 212. CNO, Admiral Stark, was committed to the Germany-first strategy, and he hoped to ward off conflict with Japan as long as possible.
85. SRH-355, 368-370. RG 457, NA. See also Prange, Verdict of History, 400, where Bloch is shown as not understanding his COMINT unit as a source of information on Japanese fleet activities when he told General Short that such information came from Washington. In addition, Prange also attributes to Captain Ellis M. Zacharias, USN, ONI that Bloch was antagonistic toward intelligence (possibly just toward Zacharias) and that Kimmel should have either replaced Bloch or corrected the problem.
86. SRH-355, 370. RG 457, NA.
87. SRH-355, 402. RG 457, NA.
88. Radio Intelligence Publications (RIP) No. 39 and 51, SRMN 069 and 072, RG 457, NA.
89. SRH-355, 223, 248, 260. RG 457, NA; NSA interview David Snyder, CWO, USN (Ret), 3 October 1983 by Robert Farley, OH-21-83, (classified); "History of OP-20-GYP-1" Series W.I.5.13, CCH History Collection (classified).
90. See also SRH-180. RG 457, NA; U.S. Naval Pre-WWII RI Activities series.
91. Appendix A, Study of Intercept Activities 23 Aug 1940, Revised 27 September 1940, Series VII, Box 4, Vol. I, pre-1942, CCH History Collection (classified).
92. Howeth, 543.
93. "History of OP-20-GYP-1." Series IV.W.I.5.13, CCH History Collection (classified), and SRH-355, 420. RG 457, NA.
94. Snyder interview (classified); Currier interview (classified). SRH-355, 420-21, RG 457, NA.
95. SRH-355, 420-21. RG 457, NA.
96. NSA interview, Thomas H. Dyer, Captain USN (Ret) Jan-Mar 1982 by Robert Farley, OH-01-82 CCH.
97. SRH-355, 316-17, 394, 351, RG 457, NA. A letter from Nimitz to Halsey on 28 Oct 1942 which indicates that even at that date, motorcycles were still used for carrying traffic from Wahiawa to Pearl Harbor, Series IV.Q.5.5, CCH History Collection.
98. Morison, Vol. III, 42-43; NSA interview, Thomas h. Dyer, OH-01-82; Edwin T. Layton, And I Was There, 51.
99. See Appendix C in toto and "Combat Intelligence Unit, 14th Naval District," TI Summaries with comments by CINCPAC War Plans and Fleet Intelligence." SRMN-012, RG 457, NA. See also Prange, Verdict of History, 453-43, which cites Kimmel's testimony at the congressional hearings when he said he had no reason to suspect the six missing carriers had been converted into a "lost fleet" during November. In fact, before the 17 November procedure change it was clear that these carriers had been assigned to a separate organization, the 1st Air Fleet. (See Appendix A.) However, neither COM-16 nor Layton picked this up as an important development though it was reported by COM-14 on 6 November. See Appendix C. See also "Notes on Communications Security," prepared by J.R. Dennis, 17 November 1942, Series IV.W.I.5.8, CCH History Collection (classified).
100. Intelligence Reports by Pacific Fleet Intelligence Officer, PHA, part 17, 2643.
101. See Appendix C and SRMN-012; COM-14 Daily Comint Summaries for 16 July, 31 July, 28 September, 2 October, 16 October, 21 October, 22 October, 23 October, 6 November, 21 November, 29 November, and 2 December. See also footnote 1 in Prange, Verdict of History, 446, which refers to Kimmel's testimony before Congress (Part 6 beginning on p. 249). Clearly reflecting Layton's assessment of traffic analysis and D/F as sources, Kimmel describes information thus derived as "open to serious doubts" unless supported by readable messages. Examples cited show how closely he was following T/A reports from both COM-14 and COM-16, however.
103. P.H.A., Part 14, 1405, item 33.
104. Considering that Station C belonged to the 16th Naval District, it may seem a bit unusual for CNO to use the CINCAF channel without reference to COM-16. The 16th Naval District, however, was a problem district/command. It was subordinate to CINCAF, but after 1939 the Navy had trouble keeping the commandant's position filled. Between 1939 and 1941, three men, Rear Admiral John M. Smeallie, Rear Admiral Harold M. Bemis, and Rear Admiral Francis W. Rockwell, were appointed. It was not uncommon, as the Japanese crisis grew more intense, to see communications such as this concerning Station C passing directly from Washington to Hart, bypassing COM-16 entirely.
105. Morison, Vol. III, 71.
106. John B. Lundstrom, The First South Pacific Campaign; Pacific Fleet Strategy December 1941-June 1942, (Annapolis, MD: Naval Institute Press) 8-11. [See also Mitsuo Fuchida and Masatake Okumiya, Midway, the Battle that Doomed Japan, (Annapolis, MD: U.S. Naval Institute, 1955), 48 and following.]
107. [Missing from text]
108. [Missing from text]
110. H.P. Wilmott, Empire in the Balance (Annapolis, MD: Naval Institute Press, 1982), 80-82.
111. Morison, Vol. III, 161.
112. Morison, Vol. III, 87.
113. Appendix C concerns the movements of Japanese navy, air and military units of the Southern forces. Its contents were extracted from daily reports between July and December 1941 prepared by the COMINT unit in Hawaii which drew on intercept from Heeia as well as reports from Corregidor. The abundant detail of these reports contrasts sharply with the histories written of the period such as Morison's Vol. III, Rising Sun in the Pacific, because historians such as Morison did not have access to this type of record. The purpose of preparing Appendix C is to show the value of COMINT as a source of indications, intentions, order of battle, and warning, even when the underlying messages cannot be exploited.
114. OP-20-GT-P, the traffic analysis section in Washington, was not established until early 1942.
116. Gordon Prange, At Dawn We Slept, (New York: McGraw-Hill, 1981), 326-33; on or about 5 November 1941, Combined Fleet Operations Order #1, revealed Pearl Harbor as the Strike Force objective; SRN-117687, RG 457, NA; this message imposes radio silence on 6th Fleet submarines on 11 November 1941; SRN-116866, RG 457, NA; this message shows that radio silence was imposed on the entire Combined Fleet on 26 November 1941.
117. See Appendix A. See also Prange, At Dawn We Slept, chapter 40, 320-25, and Verdict of History, chapter 25, which discuss Japanese efforts to modify torpedoes and Kimmel's conviction that torpedoes could not run in the shallow waters of Pearl Harbor after being launched from a plane. Had he seen these messages he would no doubt have changed his mind.
119. SRN-117453, 116476, 117301, and 116323, RG 457, NA. See also Prange, At Dawn We Slept, chapter 40, 320-25.
120. SRN-117453, 116684, 117301, 116323, 117665, 115709, 115784, RG 457, NA.
121. SRN-116239/116901, 115709, 116588, 116131, 116583, RG 457, NA.
122. Morison, Vol. III, 89.
124. SRN-115787, RG 457, NA.
125. SRN-115375, 117683, RG 457, NA.
126. SRN-115787, RG 457, NA.
127. SRN-117673, 117674/117666, 116990/116329, 116436, 116643, 116920, RG 457, NA. See also Prange, At Dawn We Slept, 342-52.
128. SRN-116588, RG 457, NA.
131. Connorton, Vol. I.
132. See Appendix C, where all the OPNAV warning messages of November 1941 are shown in the specific context of the COMINT summaries from Hawaii and indirectly from Corregidor. See also PHA, part 10, 4834, in which Layton asserts his conviction that the two messages in fact stimulated the war warning from CNO on 29 November.