At Dawn We Slept: The Untold Story of Pearl Harbor

From careful study of local newspapers and information from the Honolulu consulate, a pattern had emerged: The Fleet left harbor on either Mondays or Tuesdays and returned on Saturdays or Sundays. By monitoring radio traffic of American ships and shipborne planes, the Japanese concluded that the enemy Fleet customarily practiced in an area about forty-five minutes’ flight from Pearl Harbor.¹

But for a successful attack on Kimmel’s warships, the Japanese needed exact information about the ships when they were in Pearl Harbor. “It became essential,” Tachibana recalled, “to foresee exactly, at least two weeks beforehand, whether or not the U.S. Fleet would be in harbor on the designated day of attack; to figure out the status of patrols around Pearl Harbor; to have the Hawaiian attack air force crews familiar with the topography of the Hawaiian district and U.S. warships; and promptly establish firm and reserve intelligence channels through which timely information could be obtained.”²

Henceforth, we would like to have you make reports concerning vessels along the following lines insofar as possible:

1. The waters (of Pearl Harbor) are to be divided roughly into five sub-areas. (We have no objection to your abbreviating as much as you like.)

Area A. Waters between Ford Island and the Arsenal.

Area B. Waters adjacent to the Island south and west of Ford Island. (This area is on the opposite side of the Island from Area A.)

Area C. East Loch.

Area D. Middle Loch.

Area E. West Loch and the communicating water routes.

2. With regard to warships and aircraft carriers, we would like to have you report on those at anchor (these are not so important), tied up at wharves, buoys and in docks. (Designate types and classes briefly. If possible we would like to have you make mention of the fact when there are two or more vessels along side the same wharf.)³

In effect this message placed over Pearl Harbor an invisible grid whereon Yoshikawa and his assistants could plot the position of each individual ship in its specific anchorage. Heretofore Tokyo had been principally interested in U.S. Fleet movements. Now the Navy wanted precise information on the exact location of vessels in harbor as well. This dispatch became famous as the “bomb plot” message, so we shall refer to it as such for convenience.

For a number of reasons the U.S. Army did not translate this dispatch until October 9. When the document reached Col. Rufus C. Bratton in G-2, it riveted his attention, as well it might. In no other instance did the Japanese set up what amounted to a grid system for reporting the presence and position of ships in harbor.

Bratton thought that “the Japanese were showing unusual interest in the port at Honolulu,”⁴ but his chief, General Miles, saw nothing to get excited about. He viewed the message as part of the normal Japanese traffic concerning American naval movements. Perhaps such information would hint to the Japanese when Kimmel intended to take his fleet to sea. Even Tokyo’s desire to know when two or more vessels lay alongside the same wharf might mean that “at least the inner one could not come out as quickly or as soon . . . as the outer one. . . .” At most, Miles thought, the message might signify “the Japanese intent to execute a submarine attack on these ships.”⁵

In any case, “the evaluation of those messages that primarily concerned the fleet was the primary responsibility of the ONI.”⁶ Bratton and his Far Eastern Section had consistently rated the probability of war with Japan higher than Miles did. “It was perfectly natural for them that they should think that their particular devil was the big devil.”⁷

Bratton had no authority to warn the Hawaiian Department, even had he considered Hawaii in immediate danger. Although convinced that war with Japan was inevitable, he did not think it logical for Japan “to go out of her way deliberately to attack an American installation.”¹⁰

Twin handcuffs fastened Bratton’s hands: a directive that G-2 would not “send any intelligence to overseas garrisons which might have tactical repercussions without the approval of the Operations Division” and a prohibition against sending out any intelligence based upon Magic because the Navy did not trust the Army networks.¹¹ Occasionally Bratton had tried to buck the system and paid for his zeal with a thorough chewing out.¹²

Commander Kramer of ONI sped the “bomb plot” message through Navy channels—director of Naval Intelligence, director of War Plans, Stark, Knox, and the White House. Kramer viewed the dispatch as a Japanese attempt to simplify communications and cut down on expenses.¹³ He considered it worthy of one asterisk, which he bestowed upon “interesting messages,” but not the double asterisk allotted to “especially important or urgent messages.” He also prepared a gist of it, or flag sheet, reading succinctly: “Tokyo directs special reports on ships with sic Pearl Harbor which is divided into five areas for the purpose of showing exact locations.” In addition, the message and its flag sheet went to Hart. Kramer was under the impression that Kimmel also received it because “everything that went to CinCAF, Asiatic Fleet, also went either as an action addressee or information addressee to Admiral Kimmel.”¹⁴ Actually he never received it.

Kramer’s superiors in Naval Intelligence showed some interest, but not much. The message happened to be processed when Captain Theodore S. Wilkinson, due to become the director of ONI on September 15, was visiting his new domain. The Navy considered Wilkinson a brilliant man. But he knew little about Naval Intelligence.¹⁵ His key subordinates did not hold this against him, however. Directors of ONI came and went with little apparent regard for logic. “In my opinion he had a magnificent mind,” McCollum testified, adding ingenuously, “He accepted my recommendations almost in toto.”¹⁶

Commander McCollum, head of the Far Eastern Section of ONI, was not in Washington when the “bomb plot” message made the rounds, having returned to his desk on October 16 from a trip to Europe. He did not recall seeing the dispatch at that time. If he did, “it did not make much impression” on his mind. He believed that the Foreign Ministry was sending to Kita “explicit directions as the type of intelligence that was needed, much more in detail than any of the other key consulates on the west coast, because he did not have the benefit of the services of a Japanese Naval Intelligence officer within his consulate.”¹⁹ Obviously McCollum did not know that Kita’s “Chancellor Morimura” was in reality Tadeo Yoshikawa, a Naval Intelligence Officer.

If he had seen the message, McCollum thought that like Kramer, he would have considered it an effort to cut down “the frequently voluble type of reports . . . which the Jap Navy did not like. . . .”²⁰ He also considered that how and where the ships anchored in Pearl Harbor “might be interpreted to indicate the facility with which the fleet was prepared to move,”²¹ rather like trying to determine how rapidly the fire department can answer an alarm by charting the firemen’s bunking arrangements.

The way to find out how fast a fleet can sortie is to clock it in the act, which Yoshikawa had done more than once. In this case the “devious” Oriental mind was far more direct than the Occidental. The American intelligence officers—every one a clever and dedicated man—were so busy lopping and stretching the “bomb plot” message to fit the Procrustean bed of preconception that they missed its obvious import.

Nor did the dispatch make any impression on Stark. He could not remember having seen it, and he admitted that if he had, he would have diagnosed it as “just another example of their [the Japanese’s] great attention to detail.” To add another ingredient to this witches’ brew of misunderstandings and assumptions, Stark believed that Kimmel had “the equipment or the forces trained to decode and translate these diplomatic and military messages. . . .” He had inquired “on two or three occasions as to whether or not Kimmel could read certain dispatches when they came up” and “was told that he could.”²²

It was Turner, director of War Plans, to whom Stark posed these queries; he in turn put the question to Noyes, director of Naval Communications.²³ Turner later testified, “On every occasion I was assured that the Commander in Chief was getting as much as we were, and to the best of my knowledge and belief, he was getting it sooner than we were.” Turner believed that this included Magic; that was why he “did not inform the Commander in Chief of the contents of these messages.”²⁴

Noyes, however, denied having told Turner that Kimmel was decrypting such dispatches. “I would never have made the statement that all ciphers could be translated in Pearl Harbor,” he testified.²⁵

When Turner passed his findings to Stark, the CNO “did not consider it necessary to go any further.”²⁶ The question inevitably arises: Why did Stark not ask Noyes or Kirk, Wilkinson’s predecessor? They were on the same staff level as Turner and more directly concerned with the problem. This is a good example of the dangerous misunderstandings that can arise when obtaining important information at second hand. But Stark emphasized that his belief that Kimmel had access to Magic did not influence him in what he sent or did not send to the field commanders. He admitted that it was his responsibility to keep them informed “of the main trends and of information which might be of high interest to them.”²⁷ Certainly the “bomb plot” message fell into that category.

Later Turner claimed that he had “no recollection of having seen” the “bomb plot” message at the time and did not know why he had not seen it. He thought that it changed the picture toward Pearl Harbor sufficiently that he would have taken it up with Wilkinson or possibly Ingersoll, but he “would not have initiated any dispatch on that subject. . . .” He considered such action ONI’s responsibility.²⁸

Both Short and Kimmel were exceedingly bitter when they discovered long after war broke out that Washington had had this important message in hand and had not passed it to them. Short testified:

While the War Department G-2 may not have felt bound to let me know about the routine operations of the Japanese in keeping track of our naval ships, they should certainly have let me know that the Japanese were getting reports of the exact location of the ships in Pearl Harbor . . . because such details would be useful only for sabotage, or for air or submarine attack on Hawaii. . . . This message, analyzed critically, is really a bombing plan for Pearl Harbor.²⁹

Kimmel agreed with Short—and in far more forceful language. He could accept “the general pattern” of Japanese interest in American fleet movements as “conventional espionage,” only to be expected. But, he stated:

With the dispatch of September 24, 1941, and those which followed, there was a significant and ominous change in the character of the information which the Japanese Government sought and obtained. . . . It was no longer merely directed to ascertaining the general whereabouts of ships of the fleet. It was directed to the presence of particular ships of the fleet. . . . These Japanese instructions and reports pointed to an attack by Japan upon the ships in Pearl Harbor. The information sought and obtained, with such painstaking detail, had no other conceivable usefulness from a military standpoint. . . . Its effective value was lost completely when the ships left their reported berthings in Pearl Harbor.

. . . No one had a greater right than I to know that Japan had carved up Pearl Harbor into subareas and was seeking and receiving reports as to the precise berthings in that harbor of the ships of the fleet. . . .

Knowledge of these intercepted Japanese dispatches would have radically changed the estimate of the situation made by me and my staff. . . . Knowledge of a probable Japanese attack on Pearl Harbor afforded an opportunity to ambush the Japanese striking force as it ventured to Hawaii. It would have suggested the wisdom of concentrating our resources at that end. . . . ³⁰

Miles protested that the dispatch, if “taken alone, would have been of great military significance but it was not taken alone. . . . It was one of a great number of messages being sent by the Japanese to various parts of the world in their attempt to follow the movements of our naval vessels.” Miles’s erroneous notion, which apparently all his colleagues who saw the message shared, reveals that Washington did not grasp the contrast between the mainstream of Japanese messages and No. 83. Taken alone or in context, the “bomb plot” message stands by itself. Later, when challenged to produce one other such dispatch, the G-2 had to admit that “if you mean similar in dividing the harbor into sections, there are no such messages that I know of.” He also conceded that it was not “a ship-movement report,” but “primarily a message dividing up the waters of Pearl Harbor into convenient areas for reporting the presence of ships, United States warships.”³¹

Defense against an attack on Pearl Harbor had been the basis of plans, maneuvers, blackouts, and reports for years. The awareness of such danger appears throughout the correspondence of top military and civilian officials during late 1940 and early 1941. It formed the basis of the Martin-Bellinger Report of March 31 and the Farthing Report, which went to Washington in late August. The United States had poured a fortune in men and matériel into the Hawaiian Islands, especially Oahu. What was the purpose if not to be ready for war with Japan and to meet a strike by its Navy against Pearl Harbor?

By itself the “bomb plot” intercept would not prove that the Japanese intended to attack Pearl Harbor. But together with other messages which followed from Tokyo and the consulate in Honolulu, it might have provided Kimmel and Short with a clue. Information which appeared unimportant in Washington might have looked very different in Hawaii. Yet evaluation and dissemination of just such information were functions of the Army and Navy staffs in Washington. They failed to evaluate properly, and they did not disseminate to the parties of primary interest—Kimmel and Short. Still, there is no proof that a full appreciation of the “bomb plot” message in Washington would have changed the course of history, for what Kimmel and Short would have done with it must remain a matter for speculation.

On September 29 Kita replied to Message No. 83 (“bomb plot”) with a suggested refinement:

The following codes will be used hereafter to designate the location of vessels:

1. Repair dock in Navy Yard (The repair basin referred to in my message to Washington #48): KS.

2. Navy dock in the Navy Yard (The Ten Ten Pier): KT.

3. Moorings in the vicinity of Ford Island: FV.

4. Alongside in Ford Island: FG (East and west sides will be differentiated by A and B respectively.)³³

The U.S. Navy translated this message on October 10—just one day after the Army put No. 83 into clear language. Thus, Washington had a further opportunity to read more meaning into Japanese espionage activities on Oahu. But there is not even a hint in the official Pearl Harbor testimony that anyone connected the two messages.

By September Yoshikawa had gleaned all he could about the installations at Pearl Harbor, so he no longer had to waste time gathering statistical data, and the ships snuggled against their docks were old friends. Therefore, he had no difficulty in trying to fit his information into the modified “bomb plot” categories, and his past studies plus on-the-spot experience made him sensitive to any change in the military situation.³⁴

He scouted numerous areas besides Pearl Harbor in trips scattered over many weeks. Mikami remembered taking him to windward Oahu five or six times, both alone and with Kotoshirodo. On these trips Yoshikawa directed Mikami to follow along the Kokokahi Road, with its excellent view of Kaneohe Naval Air Station.³⁵ Twice the agent and his faithful driver visited Wahiawa, which lies almost in the center of Oahu, conveniently near Schofield Barracks and Wheeler Field. On one of these jaunts Yoshikawa tried to enter Schofield Barracks, “but the sentry at the gate refused them permission . . . because Mikami’s taxi did not have proper identification plates.”³⁶

The fact that the Japanese were spying on military and naval activities was no news to the Americans. But they “were helpless to stop it.” Wilkinson later testified: “We could not censor the mails. We could not censor the dispatches. We could not prevent the taking of photographs. We could not arrest Japanese suspects. There was nothing we could do to stop it, and all hands knew that espionage was going on all along, and reports were going back to Japan.”³⁸

Several political leaders in Washington longed to do something about this deplorable situation. And it is another irony of the Pearl Harbor story that at the same time the Japanese intensified their activities on the intelligence front, Senator Guy M. Gillette of Iowa and Congressman Martin Dies of Texas planned to investigate Japanese subversion. Both men had been interested in the problem for some time; both had studied it carefully, were alarmed at what they found, and thought action should be taken without delay. What is more, both men were on the right track.

As chairman of the House Un-American Activities Committee, Dies was primarily interested in Communist machinations in the United States. But he had also been directing an investigation of Japanese propaganda and espionage. By August Dies and his committee had accumulated enough evidence to conclude that Japanese subversion represented a real threat to the United States. In order “to arouse the whole American people into a sense of impending crisis,” he “made arrangements for 52 witnesses to proceed to Washington for public hearings early in September 1941.”

Before taking final action, Dies wrote to the attorney general on August 27 to ascertain whether the hearings “would be satisfactory from the standpoint of the administration’s plans as they related to the Japanese.” On September 8 Matthew F. McGuire, acting attorney general, replied that the President, the secretary of state, and the attorney general all felt “quite strongly that hearings such as you contemplate would be inadvisable.”³⁹

And so the story appeared in the American press on Sunday, September 21, that the plans of the Dies Committee “for exhaustive investigation of Japanese subversive activities” had been called off. The previous day, however, Dies told newspapermen that “the potential Japanese spy system in this country is greater than the Germans ever dreamed of having in the Low Countries.” And he added, “It would be a tremendous force to reckon with in the event of war.”⁴¹

In less than two weeks Senator Gillette took action. On October 2, in conjunction with Senator Edwin C. Johnson of Colorado, he introduced a Senate resolution calling for an investigation of Japanese subversion. He specifically cited “the activities of Japanese consular officials in Hawaii and in the Western States.”⁴²

No one except the Japanese knew how close to home Gillette’s remarks struck. On October 3 Kita “referred to Senator Gillette’s allegations as ‘uninformed rumors.’” And he insisted, “I do not know of any subversive activities that Senator Gillette mentions.” That included his own bailiwick. “The Japanese consulate here was not engaged in any such activities. . . . Therefore, I see no reason why this consulate should be investigated.”⁴³

Kita need not have worried. Gillette’s proposed investigation swiftly ran afoul of the State Department. On October 11—even as the “bomb plot” message bounced around Washington—Hull told Gillette that he strongly opposed congressional investigation of Japan’s consular officials lest such action interfere with the sensitive American-Japanese diplomatic conversations. In view of the tension between the two countries and the critical international situation, he could not approve measures that would offend Japan or might even provoke her to action. So he implored Gillette to drop the matter. “Please, Senator,” he begged, “I appeal to you—don’t rock the boat!” Like Dies, Gillette deferred to the secretary of state’s wishes.⁴⁴

Hull confirmed to the press that he and Gillette had “exchanged information,” but being, as Stimson called him, “such a cagey old bird,” he declined to reveal “whether he had indorsed the Senator’s proposal.” Nevertheless, reporters sensed that Hull “did not wish to do anything at the moment which would stir up further ill feelings between Japan and this country.”⁴⁵

By now, fairly certain that no investigation would topple his espionage apple cart, Kita put in a good word for the local Japanese. He informed reporters on October 16 “that an investigation of alleged ‘anti-American activities’ here will prove Hawaii’s Japanese to be loyal residents of the territory.”⁴⁶ So once again, as in the case of Tachibana in June and the Dies Committee in September, the White House and the State Department went far out of their way to avoid offending Japan. In a speech in Congress in January 1942, Dies asserted that if his committee had been “permitted to reveal the facts . . . on Japanese espionage and sabotage” in September, “the tragedy of Pearl Harbor might have been averted.”⁴⁷ Gillette, too, tells us that “if the investigation had been made, the chances are the Japanese would not have had the nerve to strike us on December 7, 1941.”⁴⁸

Both men could have been wrong—or of course, they could have been right. Kusaka wrote that obtaining continuous information about the enemy was one of the four major problems to be solved in executing the attack.⁴⁹ The cutting off of its primary source of intelligence in Hawaii might well have stiffened backbones in the Naval General Staff to the point of refusing Yamamoto permission to go through with it. Indeed, Yamamoto himself might have paused if he had had to rely on chance and not current intelligence in order to find Kimmel’s Fleet in Pearl Harbor.

CHAPTER 31

“A SIGNIFICANT AND OMINOUS CHANGE”