Agents, attachés, and intelligence failures: the Imperial Japanese Navy’s efforts to establish espionage networks in the United States before Pearl Harbor

ABSTRACT The 1941 outbreak of war between the United States and Japan was proximately precipitated by the Imperial Japanese Navy at Pearl Harbor. Yet, the IJN was building and cultivating intelligence networks on the U.S. mainland prior. This article presents a comprehensive view of Japanese intelligence efforts against the U.S. based in California and Mexico. A surprising number of relevant sources have only recently come to light. These complement Japanese Foreign Ministry files, oral histories, and Mexican sources to present a more complete story of the IJN’s intelligence preparations, successes, and failures before and immediately after Pearl Harbor.


Introduction
On the morning of December 7th, 1941, the Japanese Navy radio intercept team in Mexico City was monitoring United States Navy radio communications. American communications would normally be in code, but unusually that morning a message came through in plain, unencrypted English. The message was simple: 'Air Raid, Pearl Harbor. This is No Drill'. Commander Tsunezo Wachi of the Imperial Japanese Navy (IJN) intercepted the message and realized he and team were likely the first Japanese subjects outside of the naval task force carrying out the raid itself to know that war in the Pacific had expanded to include the United States. Wachi said he never forgot that moment. 1 Wachi represented just one facet of the IJN's North American intelligence apparatus operating at a key moment in Washington-Tokyo relations. For more than a decade, the IJN had assembled an intelligence and espionage operation focused on the US West Coast and Washington, D.C. These agents and operatives were supplemented by spies operating in Latin America, most notably Mexico. The IJN in turn recruited informants from civilian backgrounds, former military officers including a retired Royal Air Force officer, and current American servicemembers. Notably, and contrary the dark suspicions of some leaders in Washington, the IJN largely eschewed the idea of recruiting from the Nisei community. 2 Yet, surprisingly, IJN espionage activities in the United States have received relatively little attention in recent decades, with academia and popular culture focused on American code breaking and the U.S. intelligence failure at Pearl Harbor. 3 Some postwar writers including Edwin Layton provided generally accurate overviews of the day-to-day activities of Japanese agents in the U.S. based on U.S. counterintelligence records, contemporary newspaper reporting, and interviews with surviving Japanese officers during the occupation. 4 However, this view of Japanese military structure and objectives on the U.S. mainland was limited by lack of information from the Japanese side and, as a result, much of the resulting narrative was fundamentally told from the American perspective. a small number of American and British agents who were seen as more likely to evade detection. In December 1935, for instance, Director of Naval Intelligence Admiral Shigetaro Shimada listed the salary paid to the (few) American and British agents working on American soil in his diary, including the British war-hero-turned-spy Frederick Rutland. 11 The IJN's plan revolved around keeping Rutland and other non-Japanese 'sleeper' agents in place so that after a war was to start, they would be able to monitor U.S. military installations and industry in the United States and pass this information to agents in Mexico who would convey it to Tokyo. The IJN additionally installed a radio team in Mexico City to intercept and attempt to decode American communications. IJN intelligence did have some minor successes after the war started, including having Mexican agents successfully inspect West Coast U.S. Navy bases to observe damage to ships that was sustained in the Pearl Harbor attack, interception of radio traffic to learn about U.S. shipping, and in obtaining defense-related documents.
Ultimately, however, Japanese efforts to gather significant intelligence on the US mainland after the war started were limited and short-lived. The sleeper effort in the U.S. was dismantled by American counterintelligence in the months before the war started, mostly as a result of lack of planning, lack of training, and aggressive and reckless tactics that came to the attention of the FBI and the Office of Naval Intelligence (ONI). In addition, Japan's hope that Mexico would serve as a safe haven for its agents and radio intercept teams after the outbreak of war proved naive when, after Pearl Harbor, the Mexican government sided with the United States and dismantled Japanese operations within its borders. In essence, the IJN made two big bets -on American and British sleeper agents, and on Mexico based radio intercept teams -that both failed to deliver significant results. The result was that by early 1942, the Japanese government and the IJN had little information from either human or electronic sources in basics such as location or disposition of U.S. Navy ships, with Japanese sources crediting this lack of information as a significant cause of Japanese losses. For example, U.S. intelligence was famously able to locate the Japanese fleet at the Battle of Midway while the Japanese Navy had no idea where the U.S. fleet was, contributing to their loss. 12 Collectively, the new sources utilized here reveal both the trajectory and the reasons for the IJN's mixed results in information gathering before Pearl Harbor, and almost complete lack of information after the war started. The IJN's leadership focused on ship construction rather than the relativelyinexpensive recruitment and training of agents who might have been able to provide useful technical and political intelligence from the United States. Focusing instead on recruiting foreigners, the IJN placed its metaphorical eggs in only a few baskets, relying on the larger-than-life Rutland, a man who was open about his 'business' in Japan until nearly Pearl Harbor, and a small group of American agents including several alleged alcoholics who were caught bragging about their exploits.
While the IJN's backup plan of using Mexico as a base of operations after the war started was at least theoretically plausible, especially given the First World War drama of the Zimmerman Note, it also reflected a fundamental misunderstanding of the geopolitical reality that the government of Manuel Ávila Camacho was unlikely to risk being drawn into the war on the Axis side. As will be seen, after the roundup of the Japanese spy net in Mexico, Tokyo's ability to gather intelligence anywhere in North America was severely restricted, with the exception of some help it received from technically-neutral Spanish diplomats. In essence, the IJN became the victim of its own doctrine, unwillingness or inability to recruit a broad swath of agents and, fundamentally, the fact that once roused to the potential threat, American counterintelligence proved potent in shutting down its networks due to earlier missteps.

Laying the groundwork
In 1907, the Japanese National Defense Policy included the United States on its list of future adversaries for the first time. 13 In these early years, war with the U.S. appeared highly unlikely, but Japan had already defeated other adversaries in Asia, including China and, most dramatically, the Russian Empire. 14 For planning purposes, the Japanese Navy thus compared itself with the United States using the spectre of future conflict to justify increasing budget requests for expanded shipbuilding. 15 Tensions between Japan and the U.S. steadily increased after World War I as Japan occupied German-held colonies in the Pacific during the war and subsequently received international sanction to retain them in the Treaty of Versailles. The U.S., for its part, blocked the proposed Racial Equality Clause in the Treaty of Versailles and subsequently spearheaded the Washington Naval Treaty that set comparative low limits for Japanese Naval construction. 16 Adding racist insult to injury, the U.S. then banned Japanese migration in the Immigration Act of 1924.
Within the Japanese military establishment, the IJN's leadership proved the most aggressive in driving the country's preparations for war against the United States. The Japanese Army remained more concerned with traditional adversaries including Russia and China and continued down the road of conflict with both with the invasion of Manchuria in 1931. This division among the military brass had very real consequences for Japan's military preparedness: with the Japanese army leadership convinced that the real fight lay to the East, intelligence gathering in the U.S. was seen as a low priority. The Imperial Japanese Army created a Spy School in Nakano that offered effective intelligence training, but no naval officers ever attended and none of the Japanese Army graduates of the school were ever sent to the United States. 17 Instead its alumni were primarily sent to China and Southeast Asia with at least one being sent to Colombia. These graduates were instructed in at least one subject that Western spy schools do not teach -ninjutsu, commonly referred to as the art of being a ninja. Yet the fact that none were used in the United States left the IJN to pursue its own intelligence efforts without the alignment or support of their more experienced rivals in the army.
In 1934, Lieutenant Commander Nomura of the IJN set up a radio receiver at a farm outside of Los Angeles in an effort to intercept U.S. Navy communications. Nomura had some quick success in intercepting information that demonstrated the capabilities of U.S. Navy gunnery. However, Naval Artillery expert Haruo Mayuzumi, who analyzed Nomura's information, said it was not utilized in Japanese Navy training or doctrine 'because of the arrogance of the operations staff'. 18 This dynamic would play out to a greater or lesser extent throughout the IJN's prewar intelligence gathering efforts. Even in cases where some measure of success was achieved, much of the information being gathered was simply ignored. Nomura one of the first of a series of agents who would take significant risks to produce reports that would find their way into the IJN archives but had questionable impact on actual naval planning.
This reluctance stemmed in part from the IJN's culture itself. Minoru Maeda, for instance, served in roles from section chief to director of Intelligence Department from the 1930s through the war years and recounted that the IJN was cautious and negative about espionage and regarded spying activities as nearly heretical. 19 Mayuzumi also recalled that before being sent to the United States, Osami Nagano, the commander-in-chief of the Yokosuka district, told him, 'Don't work too hard on spying. If you are discovered, it will be your downfall'. 20 Similarly, the Naval Academy at Etajima was heavily influenced by British naval models with a course in 'being a gentleman' included in the official curriculum, but with no courses offered at all on intelligence gathering or espionage (perhaps an irony given that one of the IJN's most valuable intelligence assets, Frederick Rutland, was in fact a former Royal Navy officer). 21 Even at the Naval War College, which educated the future elite, there were only a few hours of classes on intelligence and no systematic manual like the ones compiled by the U.S. Office of Naval Intelligence in 1933 and 1935. 22 The IJN's general staff, for its part, prioritized ship construction over intelligence gathering. According to Naval General Staff officer Kyugoro Shimamoto, in 1934 the General Staff's Intelligence Department drew up a plan to expand intelligence gathering against the U.S. by spending 4,000,000 yen, which he explained was the same cost as building a single destroyer. Despite the relative modesty of this proposal, the plan was killed by upper echelons. Shimamoto explained the failure by arguing that the IJN's upper echelons simply had little interest in intelligence. 23 In December 1941 there were still only four staff, including the chief, in the 5th (U.S.) Section of the Intelligence Department. These four men requested increasing amounts of information from their American agents but their own analytic capability was limited by their small numbers. 24 Despite this cultural reluctance, the IJN did launch a series of intelligence operations against the mainland United States, Hawaii, and Mexico following the invasion of Manchuria in 1931. In June 1932, Admiral Shigetaro Shimada was promoted to lead the Japanese Navy's intelligence operations and found the portfolio to be close to nonexistent. In his memoirs, Shimada noted that plans to send warships out on manoeuvres were regularly approved by the Emperor himself. In intelligence gathering, however, not only were there no approval loops, but the plans themselves did not exist. 25 Shimada himself was a rare naval staff member with an interest in intelligence gathering. He was a captain of a battleship in the 1920s and in late 1941 would approve the Pearl Harbor attack as Minister of the Navy. In 1944 he was appointed head the Naval General Staff. 26 In an added personal twist, Shimada and Commander Tsunezo Wachi had known each other since boyhood. There would later be postwar suspicion that Wachi was not entirely forthcoming about his espionage efforts in Mexico because Shimada was on trial for war crimes at the time.
Shimada's efforts were focused on the West Coast of the United States, particularly Los Angeles, home base to the Pacific fleet until its 1940 move to Pearl Harbor. Southern California was also home to more than 60 per cent of U.S. warplane production with major warplane companies including Douglas and Lockheed based in the area. In his memoirs, Japanese Naval Attaché Yuzuru Sanematsu outlined the IJN's intelligence objectives in 1932: (1) Observe the movements of the U.S. Fleet, determine its plan of action, and ascertain if they have plans for war preparation (2) Understand the level of fleet training of the U.S. Navy (3) Find the feelings of U.S. Navy crews towards Japan (4) Investigate the construction of warships and updates of existing ones in the U.S. West Coast The highest priority was of course the West Coast of the USA, and it was broken down as follows: 27 More detailed strategies were developed over the course of the decade. When the Office of Naval Intelligence broke into Seattle-based IJN naval attaché Sadatomo Okada's apartment in 1941, they found orders for him to prioritize specific types of information: Given lists such as these, the IJN leadership clearly intended Okada and his colleagues to be busy men. Yet, with only four staff members in Tokyo to consume the information gathered its utility would be limited even when it could be obtained.
Yet however ambitious its goals, the IJN's operations were also hindered by seemingly incessant mishaps stemming in part from a lack of resources, poor training, and communication difficulties. In August 1932, just after the Los Angeles Olympics, L.A.-based attaché Lt. Commander Takuya Torii was ensconced in an apartment in San Pedro from which he could casually walk over and observe the U.S. Pacific fleet at anchorage. Amidst the excitement of the Olympics, Torii enjoyed a boozy dinner at a Japanese restaurant in Los Angeles' Little Tokyo neighbourhood with a colleague. As he was heading back on the 20 miles trip to San Pedro, he crashed his car and was killed. 29 This was not only a major and unexpected blow to the IJN's information gathering efforts but proved a tradecraft failure of its own. After Torii's death, a local Japanese doctor, Takashi Furusawa, made a series of calls to the Los Angeles Police Department requesting Torii's briefcase. His repeated and increasingly frantic calls only served to make officers suspicious. When police subsequently opened the briefcase they discovered incriminating documents related to Torii's intelligence activities. More damningly, it also contained letters to and from Furusawa, who as it turned out was acting as Torii's courier. Furusawa, 58, worked in a medical office in Little Tokyo. He had been in the USA since before 1907 and both he and his wife were active supporters of Japanese military and espionage operations. After the Torii incident the FBI opened a file on Furusawa but he was not actually interrogated until October 1941. He was ultimately arrested on the afternoon of 7 December 1941. 30 More immediately, however, Torii's untimely demise left the IJN without an intelligence conduit in the Pacific Fleet's home port.
With so few agents available, Tokyo saw no choice but to assign another agent named Nakazawa who was based in Seattle to expand his intelligence gathering to cover the entire San Diego to Alaska region. This left him with more than 4,000 miles of coastline and multiple major ports to survey with limited help. 31 This seeming lack of basic operational planning and security extended to even the IJN's largest intelligence operations. The IJN agent who would become most famous, Itaru Tachibana, attracted attention for driving around U.S. Naval bases in his own car, in almost every case trying to look less suspicious by having a girlfriend accompanying him. 32 As a further clue for U.S. counterintelligence, the car carried a license plate registered to Tachibana himself. 33 Preparedness and operational security did improve somewhat by the later 1930s. Naval attaché Kanji Ogawa spent a year at the IJN's intelligence desk for the U.S. and United Kingdom before being sent to the U.S. in 1938, making him something of a subject matter expert. Hawaii-based Takeo Yoshikawa, the agent who observed and reported the position of the U.S. fleet at Pearl Harbor prior to the attack, was perhaps the best prepared after working on the U.K. desk for 5 years. Yet even he recounted that most of his preparation for HUMINT operations came from reading authors including T.E. Lawrence ('Lawrence of Arabia'). 34 The lack of training and operational expertise among the IJN's agents was compounded by communication difficulties compounded by both distance and the Pacific Ocean. Agents who obtained information in California were initially instructed to send occasional letters from Los Angeles to Tokyo by having Frederick Rutland or Japanese consulate staffers simply drop a letter with a visiting Japanese ship captain who would convey it onward. Japanese commercial shipping companies including Nippon Yusen Kaisha were active participants in these plots. 35 Yet as political tensions and the amount of information being passed to Japan increased, this approach began to show cracks. In July 1941, for instance, the Japanese consulate in San Francisco was ordered to obtain street maps of all 48 states, plus a stack of phone books. This was hardly an amount of information that could be quietly passed to a commercial ship's officer in Los Angeles.
The result was that the IJN's agents pivoted to the method of simply transporting huge amounts of material personally. The most reliable method of getting information out of the U.S. was to drive it in a car over the border to Tijuana. From there, agents could more easily load the material onto a ship and send it to Japan. Tachibana himself was observed driving this route on a regular basis. 36 Shipments were also occasionally loaded onto ships at the port of Guayamas on the western coast. Guayamas is 260 miles (420 km) south of the U.S. border but had a nearby railroad, allowing easier communication even during the rainy season when roads might not be passable. 37 It also had a dock and other facilities built by fishing company Nippon Suisan Kaisha and facilities including a Japanese-owned bottling plant. 38 This simple approach was successful for a time. At around the same time as the San Francisco phone book purchases, an IJN agent in Seattle obtained information 'almost beyond description', including maps of oil installations, aerial pictures of the Ford plant in Michigan, air defense plans around the Boeing plant in Seattle, along with thousands of other pages. 39 The agent simply drove the documents from Seattle to Los Angeles and left them at Tachibana's hotel room with plans to have them shipped to Tokyo from a Mexican port.
Yet this approach relied on having willing and capable agents who could provide information at the rate and scale demanded by Tokyo. As already seen, the massive under-resourcing of intelligence efforts in the U.S. made this a formidable challenge, as did the fact that Japanese naval attachés often roused suspicion wherever they went personally. The obvious solution was to recruit Americans or Europeans who could gather information with more discretion. Yet unlike Soviet or Nazi spymasters, the IJN found that the reliable recruitment method came not by appealing to ideological sympathies or an affinity for the Fatherland, but by simply offering cash.
This approach focusing on individuals who were motivated by money carried drawbacks. The Japanese agent in Hawaii at the time of Pearl Harbor, Yoshikawa, told his postwar interrogators that these 'other kinds of people' (non-nationals who were motivated by money) would initially send large amounts of information to prove their worth but then simply stop trying. 40 The IJN's most trusted agent who fit this model was British World War I pilot Frederick Rutland. After the war, Rutland resigned from the nascent Royal Air Force and began his career as a Japanese agent and military advisor. Rutland spent several years in Japan in the 1920s as advisor on technology and naval aviation to Mitsubishi and the Japanese Navy from his house in Yokohama. Rutland gained the trust of key figures in the Japanese admiralty at the time, and when the Japanese Navy decided it needed a sleeper agent in the U.S., Rutland emerged as the obvious candidate.
Rutland was recruited to the Japanese cause by Shiro Takasu, the Japanese Naval attaché in London. As part of the Shimada-led plan to increase intelligence gathering on the West Coast, Takasu convinced Rutland to move to Los Angeles, with the goal of creating a network of contacts and spies that could be activated in case of war. 41 On the surface, Rutland was a perfect candidate for what the Japanese Navy wanted. In addition to being a trusted source to the Japanese navy, he was a strong networker who quickly purveyed his British war hero status into relationships with people in California that could both exploit in the present and possibly use further in the future, including executives in the aircraft industry, naval officers, and the media. Rutland proved to have little difficulty obtaining information from officers at parties and other social gatherings. By 1935, Rutland had purchased a house in the Hollywood Hills, along with two cars, thanks to the largesse he was enjoying from Tokyo. He sent his children to exclusive private schools in the area. Accordingly, he was an expensive hire for the IJN's initially-low-budget operation, collecting 400,000 yen for the four years through 1937. 42 This was more than 10 times the salary of the highestranking Japanese admiral. 43 In 1935, Rutland was the only sleeper agent Japan had in the mainland U.S. The new IJN attaché in London, Oka Arata, stressed in a telegram to Takasu that he did not believe it was a good idea to have Japanese espionage plans for wartime reliant on Rutland alone. He added that, Rutland is not, of course, Japanese, and therefore is working to maximize his own fortune, while reducing the effort of the work. Therefore, we should keep in mind that this is not an easy matter, but we can address it by cutting down the money he is requesting and also specify exactly the fundamentals of his work. 44 Just as it had provided an easy method of transporting large amounts of information to Japan, Mexico again entered the picture in the Rutland plot. Initially, the Japanese Navy had decided that in case of war, Rutland would be able to pass information to the Japanese Embassy in Ottawa, Canada where it could be sent to Japan. However, all involved quickly realized that Canada might not be neutral in an upcoming conflict, while Mexico very well might remain so. In addition, Mexico had a rapidly-expanding Japanese community and a porous border. By the late 1930s, therefore, the IJN's plans for gathering intelligence in the wartime U.S. were focusing to the south rather than the north.

Facing off with the FBI
In the two years prior to the war started, elements of the U.S. government became highly concerned about possible Japanese espionage and sabotage, especially from Mexico. In 1940, ONI officials reported to Congress that there was no truth to the rumor that the Japanese Navy had 1,000 warplanes in the fields near Turtle Bay, Mexico. 45 Yet this same Naval Intelligence office put the U.S. Navy on high alert on 17 October 1940 following a report that Tokyo had hidden 12 bombers in Baja California and was planning to crash them onto ships in San Diego. 46 As already seen, Mexico did feature prominently in the IJN's intelligence gathering efforts and plans in the United States. However, American counterintelligence was looking for non-existent planes, saboteurs, and above all, suspicious activities by fishing boats, rather than the more classic human and electronic espionage that the Japanese were operating south of the border. Japanese fishing ships became a common theme in the American counterintelligence imagination in the middle years of the decade. In 1935, increased activity in Guayamas led the American Vice Counsel to write various memos highlighting Japanese activity in the area. 47 Secretary of the Navy Claude Swanson reported he was sure of this spying by Japanese in Mexico in a memo to Secretary of State Cordell Hull in January 1935. 48 The threat posed by fishing boats even received President Franklin Roosevelt's attention. U.S. naval intelligence soon went over the head of the U.S. Embassy in Mexico City and asked FDR to approve a new naval attaché for Mexico. In addition to his usual duties, this officer was given a Vought scout-bomber with which he or others could patrol the Pacific coast personally. 49 As it turned out, there was little the plane's observers could do other than note the Japanese vessels were fishing -or, perhaps even more suspiciously, that they were occasionally seen not fishing. The U.S. military generally referred to any Japanese or Japanese American-owned boat as simply 'Japanese', not distinguishing between the large shipping companies in Japan, Japanese fishermen, and Japanese-American fishermen. In August 1941 Commander Keisker of the U.S. Navy reported to the FBI that there were hundreds of Japanese fishing boats that could take people and material to any spot on the American shoreline. 50 For its part, the IJN was seemingly more bemused than anything else at the U.S. government's rising interest in Japanese fishing. After the war, Admiral Shimada rated the value of reports from fishing ships to be a 1 on a scale of 1-10. In his postwar interrogation, Captain Wachi commented the American obsession with Japanese fishing boats was comical. 51 There is little evidence of any serious threat to the US mainland from fishing boats, aircraft or any other supposed Japanese saboteurs or plots, mostly because they do not have appear to have existed. Yet the discovery of the Japanese plots and the media maelstrom around the arrests of the Japanese agents contributed to a late breaking overreaction by the US government against the perceived Japanese espionage threat that continued even after it was mostly gone. This overreaction was arguably a cause of the internments of American citizens of Japanese descent during the war. 52 In reality, the IJN did continue using naval attachés and non-Japanese agents to obtain information in the United States. While the IJN's leadership did not place much official confidence in espionage, the officers who were sent to the U.S. as agents and their superiors were highly motivedoften to a fault. This in turn led them to commit operational mistakes that set back their efforts and let the Americans know what they were doing. In 1934, for instance, IJN attaché Masashi Kobayashi was invited to a presidential review of the American fleet. He was invited aboard a U.S. Navy ship and surprised sailors by shamelessly carrying a camera, film, and telephoto lenses on board. 53 Other IJN attaches continuously committed similar or even worse errors. In one early success, attaché Toshio Miyazaki recruited an American sailor named Harry Thompson to obtain information by drinking with sailors in San Pedro bars. The plan yielded dividends, but, strangely, Thompson's IJN handlers assigned him the code name 'Tomison', meaning 'rich village' in Japanese but bearing an obvious similarity to his real name. American cryptanalyst Agnes Driscoll later recounted that this obvious clue helped her solve the mystery of Thompson's identity from intercepted Japanese communiques. 54 He was eventually arrested.
Miyazaki's superior, Tamon Yamaguchi, was even more of a hothead and ignored even more basic security precautions. He instructed his American agent, former U.S. Navy officer John Semer Farnsworth, to simply drop off confidential documents with his apartment doorman instead of using a dead drop or any other method of transmission. 55 Farnsworth was a U.S. Naval academy graduate who had been drummed out of the Navy in 1927 for conduct unbecoming an officer. He was able to use his knowledge and connections to acquire both high value tactical and technical information, including the specs of a new gunnery sight and U.S. capabilities in the emerging area of dive bombing, which Shimada deemed to be valuable information. 56 Yet the lack of precautions combined with some drunken antics by Farnsworth led to his arrest and Miyazaki's expulsion from the United States. In a 1935 diary entry, Shimada listed five non-Japanese people who were clearly identified as spies in the United States. Four were known to have been arrested.
The final break-up of the IJN's West Coast spy network came just months before Pearl Harbor. In March 1941 Office of Naval Intelligence agent Ken Ringle led a burglary of the Japanese consulate. Ringle and his superior officer, Ellis Zacharias, composed two of the twelve total Japanese speakers in the U.S. Navy, and both were working fast as they felt the threat from Japan increasing. The documents seized during this raid laid the groundwork for dismantling the Japanese network led by Tachibana. They also crucially contained instructions demonstrating that Japanese espionage efforts were not focused on fishing boats, and that Japanese agents mostly avoided employing agents from the Nisei community.
The IJN intelligence network's situation by the time of Ringle's raid was already fraught. Tachibana was heading espionage operations from Los Angeles and obtaining information from other attachés in the United States. Rutland continued to work on his sleeper operation, with an increasing focus on communications with Japanese agents in Mexico. The IJN's hope was that he would be able to continue his information gathering on U.S. Navy preparations and aircraft production and be able to get this information over the border and from there to Tokyo. In a breach of operational security, however, Rutland, became increasingly involved with Tachibana's day-to-day operations.
Rutland felt confident in his cover stories explaining his close relationships with Japanese organizations. His employment by Japanese firms in the 1920s had hardly been a secret; his frequent trips to Japan in the 1930s had even been covered by the Los Angeles Times and other local newspapers. 57 Rutland pursued information on American aircraft technology; the information he obtained from talking with pilots and aircraft executives at his house parties and other gatherings was likely very effective. Rutland was a citizen of a third country, the United Kingdom, which was not formally allied with the United States at the time, likely making many of his efforts on Japan's behalf legal.
Yet as the political situation intensified and Japanese espionage became even more brazen, Rutland started sensing danger. He devised a plan to approach U.S. Naval Intelligence and become a double agent for the U.S. while continuing his work for the Japanese Navy. He told the ONI that his efforts for the Japanese navy were mostly legal activities such as collecting open-source information. This dangerous game was initially successful, with Zacharias in mid 1939 sending a message to the FBI to leave Rutland alone. 58 As it turned out, Rutland's fears were well-founded. In June 1941 the FBI arrested Tachibana and Toraichi Kono, the long-time butler of actor Charlie Chaplin. 59 Kono was wellconnected in Los Angeles and was apparently disbursing funds to various Japanese espionage assets. 60 Realizing further arrests were only a matter of time, a number of Japanese agents were ordered to flee from Los Angeles to Mexico along with their paymaster from the Yokohama Specie Bank. 61 At this point, the IJN had apparently detected that Rutland was playing both sides of the equation. A wire from the Los Angeles consulate to the Japanese Embassy in Washington suggested that Rutland was close to U.S. Naval Intelligence and advised not using him to bail Kono out of jail. 62 For its part, the FBI reported it found smoking gun evidence of Rutland's work for the IJN in Tachibana's files. Rutland read the writing on the wall and travelled to Washington to lobby for protection from the ONI personally. FBI agents were closely tailing Rutland as he walked into U.S. Naval Intelligence headquarters. These efforts were in vain, however, and ONI officers refused to help him. Rutland then met with two members of the British Embassy staff who convinced him to return to the United Kingdom, conveniently removing him from the U.S. to avoid any hint of scandal that would embarrass the British in American eyes at a critical moment in the war. 63 Rutland now attempted to one final meeting with the IJN, either in a bid for protection or as part of an elaborate plot to double cross Tokyo. The FBI tailed him to Brownsville, Texas where he attempted to board a flight to Mexico City but was turned away for lacking a visa. Feeling the heat, he boarded a plane headed for London a month later. 64 The IJN had now lost almost all of its assets in the United States. At the State Department's insistence, Tachibana and his associates were deported rather than put on trial to avoid further scandal and a possible casus belli between the U.S. and Japan. 65

The Tijuana connection and Agent S
In 1940, the IJN opened a radio intercept station in Mexico City and sent the experienced and Spanish speaking radio intelligence expert Tsunezo Wachi to head the effort. The Japanese Navy had been systematically intercepting communications since the 1920s, 66 and by the 1930s had deciphered American diplomatic codes. 67 However, the U.S. Navy's military codes [Strip Ciphers] had not been deciphered, 68 and its ability to intercept the U.S. Navy's movements and intentions was lacking. Therefore, as the possibility of the outbreak of war between Japan and the U.S. increased, the Japanese Navy sought to strengthen its communications intercept system. 69 Wachi had already found success breaking Chinese military codes, and now he would take a crack at the American codes. 70 The transmission station the U.S. Navy used in the Atlantic was located in Arlington, Virginia, putting signals on the other side of the globe from Japan. Mexico City's location near the Atlantic and high altitude made it a favourable location to intercept this U.S. radio traffic.
Wachi left for Mexico with his small team of radio intercept and code experts in November 1940. His ship arrived in San Francisco on December 9, with FBI agents watching and taking notes. 71 Wachi then boarded another ship to Manzanillo where he disembarked and proceeded to Mexico City. He promptly set up his intercept station on Calle Orizaba, with a staff of a petty officer and four sailors trained in radio intercepts. 72 Wachi initially had trouble intercepting U.S. radio traffic. He had come with a IJN Type 92 radio receiver which was not able to pick up the longer, 15 kc VLF wavelength used by the U.S. Navy in the Atlantic. He soon jumped on a plane to New York and bought a new RCA receiver with which he was able to intercept Washington's naval broadcasts and the radio waves of the Atlantic Fleet, although he struggled with code breaking due to changes in U.S. Navy codes in April 1941. 73 After being installed in Mexico with the new receiver, the intercept team was able to determine Atlantic convoy schedules and track Panama Canal traffic. In addition to the convoy and shipping info, Wachi mentioned obtaining sensitive information including plans for a new U.S. bomber, the B-29, from a HUMINT source. 74 He referred to this agent variously as Q, Agent S, and Sutton. Wachi was not a cooperative witness with his post-war interrogators and provided various stories about Agent S, many of which conflicted with each other. 75 There is some evidence that Agent S may have been more than one individual. Agent S does resemble Rutland in some ways. Wachi remarked multiple times that Agent S was a disgruntled major who was introduced to him by the General Staff. 76 Rutland was a major, was clearly disgruntled against his home country, and was the only agent personally known to the IJN's General Staff. The list of foreigners that were spies for Japan in Shimada's diary has no one else remotely resembling Agent S other than Rutland. Further, Rutland was known to be traveling to and from Mexico City at this time. Since he was going to Mexico, it would make sense for him to hand carry something of extreme value, such as the B-29 plans. On the other hand, Wachi referred to Agent S as still operating after Rutland returned to London, making the identification questionable.
In February 1941, the Japanese government placed its intelligence assets in the Western Hemisphere closer to a war footing, further reducing its investment in propaganda and focusing on espionage. IJN officials now predicted that in the event of war, its network in all Latin American countries would be centred in Mexico, with additional posts in Rio de Janeiro, Buenos Aires and Santiago. 77 On 2 June 1941, Foreign Minister Yosuke Matsuoka sent the country's ambassador in Mexico a telegram indicating that Mexico City would now be one of its centres of operations against the United States. A few days later, Tachibana and Kono were arrested. The Japanese Foreign Ministry and Navy and not been cooperating on intelligence, and the Director General of the U.S. Bureau of Foreign Affairs Taro Terasaki stated that he had been unaware of the IJN's Tachibana operation and accusing the navy of undermining the country's diplomatic efforts. 78 He offered Foreign Ministry help for bailing Kono out of jail, but added the bail money would come out of the Navy's budget. 79 The Japanese Foreign Ministry's hope that it would be able to run intelligence operations from Mexico did have some logic. During World War I, the Mexican government had been actively hostile to U.S. interests. The famous Pancho Villa incursion from Mexico into the U.S. in 1916-17 highlighted the hostility in Mexico towards the U.S. Unfortunately for Tokyo, by 1941 the new Mexican president, Manuel Ávila Camacho, and his cabinet were firm supporters of the United States. 80 Yoshiaki Miura, the Japanese ambassador to Mexico, appeared confused by the changing situation in July 1941. He reported to Tokyo that he could work with the IJN to enhance communication networks in Northern Mexico, including Mexicali and Nogales, but that there was no point since there were no contacts left on the American side of the border with whom they could engage. On 5 July 1941, Miura sent another strongly-worded telegram to the Foreign Ministry indicating that its plan to use Mexico City as a base for wartime espionage would be fruitless because Mexico would support the U.S. in a future war and would no doubt shut down his activities. 81 The surprise in Tokyo to the messages from Miura was evident. On July 10, Foreign Minister Matsuoka ordered embassy secretary Taro Terasaki to travel from Washington to Mexico to produce a first-hand account of the situation. Terasaki and Miura concurred that if a war with the U.S. was to start, Mexican president Camacho would expel all Japanese officials. He endorsed the continued efforts by Wachi at intercepting U.S. radio traffic and breaking codes, but assumed that, should a war start, Mexico would either formally or de facto join the Allies. 82 Terasaki also considered the pros and cons associated with the plan to operate from Mexico. One Japanese espionage priority was to continue to obtain open-source information from the American press. It would be easy to simply subscribe to U.S. newspapers and magazines and have them sent to an agent's house in Mexico and from there send them to Tokyo via a neutral country. However, agents in Mexico might not be able to obtain other open-source information, including technical journals that might be only available in a U.S. university library. Terasaki further determined that Brazil would also likely support the Allies in the future war, taking that country out of the picture. Indeed, he correctly assessed that any Japanese agents in Mexico would need to send information to either Spain or Argentina to convey it to Japan.
Terasaki also embarked on a political mission of his own: to weaken the Mexican government's support for the United States through a combination of propaganda and bribery. 83 Supporting the former goal, an embassy staffer named Sato arranged to have some young Mexican journalists from influential families be flown to Japan on an all-expenses paid trip. The idea was to have them go back to Mexico and write pro Japan articles in the press and thereby influence Mexican policy. 84 This approach appears to have yielded some success, but was unlikely to change overall Mexican policy towards Japan.
Bribery proved somewhat more successful. In one example, Japan was short of mercury, which was used in the manufacturing of detonators, and was not allowed to buy from Mexico due to international sanctions. Wachi bribed a Mexican general to help source a shipment of mercury and arranged for his Mexican agents to smuggle it to Japan. 85 Their plan was to have the bottles of mercury placed in lower halves of barrels and concealed with bronze scrap above. By this stage, however, Mexican security services were increasingly observing ships departing to Japan and noticed when several barrels ostensibly filled with scrap began leaking mercury during loading. Wachi again resorted to bribery, and his notes show that he planned to give $100,000 to Mexican prime minister Camacho himself to let the mercury go on its way. 86 Whether the bribe was ever delivered in unclear, but the mercury did make it to Japan.
Despite successes such as these, the IJN realized Terasaki's efforts would be of limited help and continued to limit cooperation with the Japanese Foreign Ministry. The Foreign Ministry, on the other hand, was impressed with Wachi's intercept team and requested help intercepting U.S. communications related to ongoing peace negotiations. The IJN refused and continued to focus on purely military interception. 87 Wachi did embark on some efforts to recruit new HUMINT assets to his cause. With operations in California limited to what the heavily-surveilled consular staff could achieve, he and the IJN explored using Mexican agents to infiltrate the United States. Wachi was impressed by some of the Mexicans who had worked on his smuggling schemes and offered them cash if they travelled to the U.S., took some pictures, and came back. 88 He later even used of these civilians to take pictures of the American fleet after the war started in an effort to calculate the damage inflicted in the Pearl Harbor attack. 89 The U.S. Navy later concluded the resulting calculations were reasonably accurate. On the other hand, both the U.S. Navy and the IJN independently concluded that the idea of using Mexican civilians posing as tourists was not a highly effective means of gathering intelligence.
The reality was that by mid-1941, Tokyo and the IJN had decreasing options. The IJN's entire intelligence network in California and throughout the U.S. had been rounded up and deported or imprisoned. Documents confiscated during the Tachibana raid by the FBI demonstrated how deeply Japanese private companies had been involved with intelligence and they were now facing pressure of their own. In August, the offices of Japanese companies Yokohama Specie Bank, NYK, Mitsui, Mitsubishi and Sumitomo were raided by the Department of the Treasury. The stated reason was to confirm compliance with currency regulations, but the presence of FBI agents demonstrated the real intent, though many Japanese civilians involved escaped arrest or prosecution under State Department pressure. 90 The L.A. consulate, for its part, was still able to obtain secret information including warplane production numbers and ships sailings. 91 Yet even these limited successes were short-lived.

Conclusions
Mexico severed diplomatic relations with Japan on 8 December 1941, froze Japanese-owned bank accounts, and began making threatening noises about evacuating or even interning Japanese citizens. Wachi, Miura and the others knew their time as free men in the country was limited. Yet they also knew that getting information back to Tokyo was critical. As a result, Wachi and his network of intercept operators increased their activities while preparing for imminent arrest.
On December 16, Wachi attempted to destroy his radio receiver and his teletype machine. A local farmer observed a group of Japanese burying large equipment and reported it to the police. The ensuring incident was reported breathlessly by the Mexican press. 92 In April 1942 Wachi, Miura and others were repatriated to Japan in a diplomatic exchange. 93 Before leaving, they organized a scheme to convey information from Mexico to Tokyo via the Spanish Embassy and via Japan's embassy in Madrid, keeping at least the hope of further information gathering alive. 94 Wachi was conveyed on the same ship as VIPs including Ambassador Nomura and Terasaki. When the ship heading for Japan docked in Singapore on the way, Wachi was sent back to Japan on a plane ahead of the other passengers to brief the general staff about the U.S. Navy in the Atlantic, indicating his perceived importance. The person debriefing Wachi, Prince Takamatsu (a younger brother of Emperor Hirohito), recounted in his diary, however, that Wachi's team had been unable to achieve great success. 95 Ultimately, the IJN's espionage and intelligence efforts in the United States yielded reasonable amounts of information on American technology and tactics prior to the war, despite its lack of operational security and missteps. Yet the subsequent loss of nearly all IJN assets in California in the months before Pearl Harbor left a gaping hole in Tokyo's information. The Japanese Navy lost its Western Hemisphere radio intelligence, further leaving the Japanese in the dark, beyond traffic analysis of communications and publicly available information from newspapers and other open sources. The IJN's eventual lack of any U.S.-based intelligence network was darkly reflected in the postwar reflections of Japanese intelligence officials themselves. Minoru Maeda, who ran the Third Department (Naval Intelligence) before and after the outbreak of war, remarked in his post-war recollections that: The Third Department was in a state of complete helplessness in the early days of the war. It was almost impossible to get an accurate picture of the enemy's situation. If there had been some information about the enemy's forces after the attack on Pearl Harbor, the Operations Department would have likely reconsidered the Midway operation. The paralysis of Third Department at the critical time of the war was the biggest regret of my life. 96 Similarly, Eizou Hori was an intelligence officer on the U.S./UK desk at Imperial General Headquarters during the war. In his memoirs, he recounted that: The biggest hole during the war in Japanese intelligence was on the American mainland. The Japanese military had spent lots of money to establish an information net, depending only in small part on Japanese. That information net was all destroyed. With the hole created, Japan had no idea what ships were sailing from San Francisco, U.S. military production trends, troop mobilization, aircraft production. You can argue this was a large reason for Japan's loss of the war. 97 Ultimately, the IJN's failure to establish a durable intelligence and espionage network in the United States not only left Tokyo blind after Pearl Harbor but became a contributing factor to Japan's ultimate defeat.