by Ian W. Toll
The Japanese army, and especially the Japanese navy, had suffered a calamitous loss of face. American warplanes had been permitted to defile the divine skies above the emperor’s palace, and not a single intruder had been shot down. “Our homeland has been air raided and we missed the enemy without firing a shot at him,” Admiral Ugaki noted in his diary on April 20. “This is exceedingly regrettable.” A chagrined Yamamoto remarked that the raid “provides a regrettable graphic illustration of the saying that a bungling attack is better than the most skillful defense.” The C-in-C had ordered all available naval forces to sea in a wild goose chase that had only burned fuel and exhausted their crews, while underscoring Japan’s vulnerability to future raids of the same sort. Not until the 20th did Yamamoto acknowledge that the American squadron was long gone. An order flashed out to all units by radio: “Cease operations against American fleet.”
Overcoming their initial confusion, the Japanese military leaders soon reconstructed what had happened. Signal Corps photographs, developed and studied the afternoon of the raid, confirmed that the bombers had been B-25s. By midnight on the 18th, the Japanese army in China reported crash landings in Chekiang (now Zhejiang) and Kiangsu (now Jiangsu) provinces. The first American pilots to be captured gave false answers, claiming either that they had flown from the Aleutian Islands or from a mythological supercarrier. “They never told the truth,” wrote Ugaki on April 19. “We must investigate further promptly so that we can take proper measures for the future.” Two days later he noted, with a chilling subtext, “American war prisoners captured at Nanchang have been sent to Nanking, where they told the truth at last.” And on April 22, “More truth has been added to the statements of the POWs . . .” Now the Japanese knew that Doolittle’s B-25s had been launched from the deck of one of the accursed American carriers.
In China, the Japanese army launched an operation called “Sei-Go,” with the goal of capturing the airfields that had been constructed to receive Doolittle’s planes. The Japanese were well aware that B-25 airmen had been sheltered by local civilians or guerrillas, and the offensive seems to have degenerated into a mass reprisal against the population of Chekiang and Kiangsu provinces. Japanese troops laid waste to the region, routinely slaughtering the inhabitants of entire towns suspected of aiding the American aviators. Biological warfare units spread cholera, typhoid, and dysentery pathogens. Chiang Kai-shek cabled FDR: “The Japanese troops slaughtered every man, woman, and child in those areas—let me repeat—these Japanese troops slaughtered every man, woman, and child in those areas.” It is impossible to know precisely how many were murdered, but the toll certainly ran into the tens of thousands.
Admiral Yamamoto took a newly high tone with the panjandrums of Tokyo. He said that Midway Island had probably been the base of the attack (it had not). Midway was the linchpin of the American threat to the homeland, he said, and if it was not taken quickly, forces committed to homeland defense would have to be greatly expanded at the expense of further offensives in the south. He would brook no further delay.
Opponents on the Naval General Staff and in the army had hoped to kill the Midway operation by a series of delaying tactics. Now their resistance crumbled. On April 20, at a meeting of army and navy planners, it was agreed to postpone attacks on New Caledonia, Samoa, and Fiji until after the Midway operation. Admiral Nagano of the Naval General Staff threw his support behind Yamamoto. General Tanaka of the army, receiving assurances that the Midway operation would not be the first step in a planned invasion of Hawaii, consented to provide an infantry regiment. Vital details that had been previously withheld—dates, orders, the fleet units to be dispatched—were now forthcoming. The Midway offensive was to take place the first week of June, immediately after the return of major fleet units from “Operation MO,” the sea-air-land offensive aimed at capturing Port Moresby, the Australian base on southeastern New Guinea.
Chapter Nine
AT THE FOURTEENTH NAVAL DISTRICT HEADQUARTERS, NEAR TEN-TEN Dock in the Pearl Harbor Navy Yard, a single inconspicuous door opened from the palm-lined sidewalk onto the top of a stairwell. The entrant descended to the basement and encountered a heavy steel gate guarded by a marine. If he showed the right credentials, he was admitted to a large, windowless space bathed in white fluorescent light and furnished with rows of identical desks. A team of yeomen fed punch cards into two car-sized machines known as “tabulators,” primitive IBM computers that collated and analyzed alphanumeric data. Day and night, punch cards raced through the machines, filling the subterranean space with a harsh metallic clatter.
This enigmatic bunker was Hawaii’s codebreaking unit, charged with peeling back the layers of encryption that cloaked Japanese radio communications. Although no sign was posted outside the door, it was formally called the Combat Intelligence Unit, or CIU. Among the staff it was nicknamed “the dungeon.” After a reorganization later in the war it was renamed Fleet Radio Unit, Pacific, abbreviated as FRUPAC. Most commonly (then and in the historical literature) it was known as “Station Hypo”—phonetic code for the letter H, designating the Hawaiian intercept station.
Hypo’s chief was Joseph Rochefort, a forty-one-year-old navy commander. He was a tall, aristocratic man, with dark hair and a keen, friendly face. He had learned to speak and read Japanese while stationed in Tokyo as a language officer in the 1920s, and subsequently joined that small, compulsively secretive coterie of officers who worked in the emerging field of communications intelligence. Rochefort did not much like cryptanalysis, finding that the work impaired his personal life, stunted his career prospects, and gave him ulcers. “I considered myself a naval officer,” he later said, and he had done his best to avoid communications intelligence assignments. Some of the early histories of the war depicted Rochefort as a brilliant oddball, who paraded around in a flamboyant red smoking jacket and slippers. In his oral history for the Naval Institute, recorded in 1969, Rochefort owned up to the jacket and slippers, but sharply denied that he or they were eccentric. Yes, he wore a smoking jacket over his khaki uniform shirt, but only because it had pockets for his pipe and tobacco pouch. Yes, he padded around the office in slippers, but only because his feet were sore from pacing the concrete floor. He did not wear them outside the closely guarded sanctum of his basement lair, and “it wasn’t that I was eccentric or anything.”
Perhaps not, but there was no doubt that the dungeon was a zone in which the usual standards of conduct did not apply. It was an organization of freethinkers, in which a spirit of teamwork blended with an ethic of informality and an esprit de corps. It was built around a handful of talented codebreakers and language officers who felt free to play their hunches. Rank and hierarchy were of little importance. “Not much attention was paid to uniforms or to military punctilio of any kind,” wrote Jasper Holmes, a Hypo veteran. “Although there was never any doubt as to who was boss, a man’s status depended on factors other than rank.” Freewheeling intuition was respected, but so was the plodding line of attack they called “siege tactics.” One man’s desk might be orderly and uncluttered, while he pondered one message intercept at a time. His neighbor’s might be littered with candy wrappers, decorated with pinup girls, and buried under piles of old intercepts. Rochefort would put up with any method or style so long as it got results.
Throughout the era before the Second World War, the field of cryptanalysis was unproven, little understood, and shrouded in secrecy. It had even been condemned as unethical, most famously in 1929 by Henry Stimson, who was secretary of state at the time and refused to read decrypted foreign communications because “gentlemen don’t read each other’s mail.” (Lieutenant Commander Thomas Dyer, perhaps the most naturally gifted member of Rochefort’s team, joked that Stimson’s view did not apply to cryptanalysts “because no one could accuse us of being gentlemen.”) Rochefort observed that many of the best and most experienced codebreakers were slow to climb the promotion ladder because their classified assignments left black holes in their service résum
és: “Every time he goes to Washington and goes in this particular division, nobody knows what he does or anything else, so he becomes known as somewhat of a nut. And some of them were.” Men who got the best results were obsessive, preoccupied, and single-minded—“on the verge between brilliance and being crazy.” Dyer kept a sign above his desk: “You don’t have to be crazy to work here, but it helps a hell of a lot!”
After Pearl Harbor, Rochefort and his key lieutenants worked around the clock. Rochefort went home once every third or fourth night. Most nights he slept on a cot in his office, fueled by coffee and sandwiches. Twenty- or twenty-two-hour days were routine. Dyer sometimes worked two or three days straight, keeping his eyes open by swallowing handfuls of benzedrine tablets that he kept in a bucket on his desk and offered freely to his colleagues. (Asked about this later, Dyer explained: “I figured there were people out there getting shot at. If it should happen that it turned out to inflict some injury on my health in the long run, so what?”) Most of the enlisted men and clerks worked in twelve-hour shifts, seven days a week, or what Holmes called a “normal” eighty-four-hour week.
The unit’s headcount would expand fourfold in the first six months of the war. When he took over the unit in May 1941, Rochefort was promised that he could pick his own men, and the base’s chief personnel officer had honored the agreement. As each new draft of recruits arrived in Pearl Harbor, a chief petty officer lined them up, studied their faces, and chose the candidates who looked bright enough for codebreaking work. The entire military band from the battleship California, knocked out of action on December 7, was selected and assigned the task of transferring the code groups from the raw radio intercepts onto the IBM punch cards. They were soon producing millions of these cards every week, and running the big IBM machines without supervision. The musicians took to the work so readily, said Holmes, that “a theory was advanced that there must be a psychological connection between music and cryptanalysis.”
Before the war, Station Hypo had been assigned by Washington to focus on secondary tasks such as the Japanese diplomatic code, the flag officers’ code, and weather codes. “JN-25,” the main Japanese navy operational code, which held the enemy’s most important secrets, had been reserved for analysts at Naval Headquarters in Washington. It had been in service for a year, and fragments of it were readable by December 1941. But a new edition of the code, which the Americans called “JN-25-B,” had come into use on December 4, 1941. This meant that the second layer of encryption, the “cipher additives,” were all entirely new. All the accumulated work on the old cipher was wiped out, and not a word of the new messages could be read. They were, said Rochefort, “in the black.”
Rochefort received permission to go to work on the new code. A week after Pearl Harbor, Hypo launched the mind-numbing, Sisyphean campaign to decipher 50,000 five-digit numeral groups. As they began the work, it seemed overwhelming, perhaps even impossible. Every significant codebreaking success before the Second World War had been child’s play by comparison; and such breakthroughs (wrote Holmes) “had never occurred except by a combination of lucky accidents relentlessly followed up by men of rare genius.” When Holmes looked around at his colleagues, none struck him as a genius: they seemed like a fairly ordinary group of men. The most difficult stage of such an undertaking is the very start, when not a single code group in any intercepted message can be deciphered. Those are the long, painstaking hours, said Rochefort, when “you see a whole lot of letters and a whole lot of numerals, perhaps in the thousands or millions, and you know that there is a system in there, and there’s a little key to the system that’s something real simple.” A Japanese transmission error in mid-December confirmed that the cipher—the five-digit additives—had been changed, but not the underlying code.
By year’s end, when Nimitz took command of the fleet, Rochefort and his team had made halting progress. It was not much, but it was something, and it would grow. Dyer, who had an uncanny ability to see patterns in the interminable mass of letters and digits, noted that “if you observe something long enough, you’ll see something peculiar. If you can’t see something peculiar, if you stare long enough, then that in itself is peculiar. And then you try to explain the peculiarity.” So he stared, and he stared, and he stared, until an idea took shape in his mind. If the idea was a testable hypothesis, he tested it. If it failed, he went back to staring. Dyer continued:
A lot of it is, I’m convinced, done by the subconscious. Sometimes when people ask me how I can solve messages, I say, “Well, you sit there and stare at it until you see what it says, and then you put it down.” You look—it depends on what kind of a thing you’re dealing with—but you look at it until you see something that attracts your attention, your curiosity. Maybe it doesn’t suggest anything at all. You go on to something else. The next day you come back and look at it again.
Rochefort agreed that it is a matter above all of “persistence, just sticking with this thing day after day after day.”
Readable code groups were few and far between at this early stage of the war, but much could be gleaned from the raw radio intercepts. “Traffic analysis” was the name given to the art of drawing useful intelligence out of nothing more than the rise and fall of enemy radio traffic volume in given locations. Rochefort said it was “common sense, actually. It’s not real intelligence; it’s common sense.” For instance, a spike in radio traffic emanating from Truk, the Japanese-held stronghold in the Caroline Islands, combined with a fall in traffic in the Inland Sea of Japan, might suggest a major fleet movement to the south. Traffic analysis grew more powerful when combined with the first fruits of cryptanalysis, typically the “message externals” or headings—geographic designators (to where, from where) or the identity of sender and addressee (from whom, to whom). Radio call signs for ships, units, and commanders appeared frequently, in predictable parts of a message, and could sometimes be corroborated by sighting reports and radio direction-finding (D/F) data, which provided an enemy transmission’s point of origin. Eavesdroppers recognized the personal quirks of an individual operator—his “fist,” they called it—and by tracking that man’s movements could guess the location of a particular ship. For these reasons, the “externals” fell relatively easy prey to cryptanalysis.
Jasper Holmes wrote that in the months preceding the war, he was not entirely convinced of the value of traffic analysis, finding the reports a “mixture of gobbledygook and vague innuendoes, with very little solid information.” But the war itself brought a satisfying surge in traffic from enemy positions throughout the Pacific, supplying plenty of grist for the mill. In early January 1942, the daily intelligence bulletin reported a sharp increase in radio traffic in the Truk area. This pointed south, and when combined with a few fragments of legible decrypts and a heavy dose of educated guesswork, Rochefort correctly predicted that the Japanese would seize Rabaul, on New Britain, in the South Pacific. Whether or not the Allies could do anything to defend Rabaul was an open question, but if the main body of the Japanese fleet was moving south, that left plenty of other vulnerable points on the long and expanding perimeter of the Japanese Empire, where the Americans might stage carrier raids with little fear of walking into a trap.
None of this information would matter unless it was read and relied upon by decision makers at the highest level of the navy. The admirals of the Second World War had begun their careers at about the same time the first primitive radio communications were used at sea. Cryptanalysis was a black box that few understood and many distrusted. If the codebreakers had already burrowed into the heart of the Japanese code, and were supplying fully decrypted enemy radio communications, no naval commander could afford to ignore them. But Hypo was still months away from producing such airtight intelligence. Traffic analysis was indicative, but could prove nothing. It rested, as Jasper Holmes put it, on “a concatenation of deductions.” It would have been within the rights of any admiral to refuse to place his faith in it.
Fortunately
for Hypo, and the navy, and the United States, Chester Nimitz was not such an admiral. He was briefed each morning at eight o’clock by his fleet intelligence officer, Lieutenant Commander Edwin Layton. Layton also had a standing invitation to walk into Nimitz’s office at any hour of any day if he believed he had important information for the C-in-C. (No one else on the staff, except perhaps the chief of staff, had this privilege.) Hypo provided a daily briefing to Layton, who in turn drew on other sources and briefed Nimitz. Layton and Rochefort had known one another when both men were stationed in Tokyo as language officers in the 1920s. They had shared in the long trial of learning Japanese. They counted one another as friends, and this tended to smooth the contours of their professional partnership, which might otherwise had been complicated by the organizational rivalry between the Fourteenth Naval District (of which Hypo was a part) and the Pacific Fleet staff. Nimitz paid close attention to all the intelligence products that crossed his desk. On his first day as CINCPAC, he told Layton, “I want you to be the Admiral Nagumo of my staff. I want your every thought, every instinct as you believe Admiral Nagumo might have them. You are to see the war, their operations, their aims, from the Japanese viewpoint and keep me advised what you are thinking about, what you are doing, and what purpose, what strategy, motivates your operations. If you can do this, you will give me the kind of information needed to win this war.”
