Blank horizontal row

Deception in the
Invasion of Japan


Dr Thomas M. Huber


In 1945, to end the Pacific war, American strategic plans foresaw an invasion of Japan's heavily defended home islands. Operations Olympic and Coronet, America's proposed landings on Kyushu and the Tokyo Plain, were the largest amphibious invasions ever planned. Although precluded by war's end, preparations for both were extensive. To gain the element of surprise, Washington and theater planners developed Operations Pastel and Coronet Deception, deception operations designed to convey to the Japanese a false story of where the actual assault landings would occur.

In Pastel: Deception in the Invasion of Japan, Dr. Thomas M. Huber reveals the contents, origins, and effects of these two comprehensive, coordinated deception plans as they related to the scheduled invasion of Japan. He also provides the Japanese perceptions as the plans unfolded. This special study reminds us of the vital role of deception in operational planning.

December 1988 RICHARD M. SWAIN
Colonel, Field Artillery
Director, Combat Studies Institute

CSl Special Studies cover a variety of military history topics. The views expressed in this CSI publication are those of the author and not necessarily those of the Department of the Army or the Department of Defense.

Deception in the
Invasion of Japan


Dr. Thomas M. Huber

U.S. Army Command and General Staff College
Fort Leavenworth, Kansas 66027-6900

CSI "battle streamer"

Library of Congress Cataloging-in-Publication Data

Huber, Thomas M.
Pastel: deception in the invasion of Japan.
Bibliography: p.
1. World War, 1939-1945-Campaigns-Japan.
2. Deception (Military science). 3. Strategy.
I. Title.
D767.2.1-183 1988 940.54'25 88-30420











In the summer of 1945, Allied forces were closing in on the Japanese home islands. Costly campaigns on Iwo Jima and Okinawa led Allied planners to estimate a million casual ties in invading Japan proper. Planners urgently sought for ways to reduce the terrific numbers and soon fastened on an obvious one: deception.

At the outset of World War II, American strategists had many choices of where to attack the far-flung Japanese in the western Pacific. Top planners ordered forces under General Douglas MacArthur to move gradually northwestward from the Solomons through New Guinea to the Philippines and forces under Admiral Chester W. Nimitz to move northwestward through the Gilberts, Marshalls, and Marianas (see map 1). These two axes of advance were nearing their objectives by the end of 1944. Washington planners at first assumed that the war with Japan might last until 1948 or longer, and they believed the American forces' next advances after the Philippines and Marianas must be made cautiously against Taiwan and the China coast. Since war aims were increasingly fulfilled, however, planners in June 1944 already began considering landings in Japan proper after the Taiwan and China assaults.

By the end of May 1945, planners had shelved the China operations altogether and considered a direct advance on Japan only. To end the war quickly, planners adopted the am bitious, but potentially costly, strategy of invading Japan's heavily defended homeland itself. Olympic and Coronet, America's planned landings on Kyushu and the Tokyo Plain respectively, represented the largest amphibious invasions ever planned. To reduce these crucial operations' expected high casualties and to guarantee success, planners felt it essential that the operations not be anticipated. Therefore, deception plans were developed to shield Olympic and Coronet from detection. How, then, did the American deception planners, at the peak of their experience, prepare to cover the allimportant invasions of Japan? The pages below explore the content, origins, likely effect, and theoretical significance of deception operations prepared for the invasions in 1945.

Map of western Pacific Ocean Area
Map 1. Axes of advance in the western Pacific, 1944

The deception plans for Olympic were developed under the evocative name of Pastel. Pastel was first hastily sketched by Admiral Nimitz' staff in May 1945 and, by the end of July, had become a mature operational order. By that time, a deception plan for a later operation, Coronet, had also been outlined. The final deception arrangements were crafted by MacArthur's and Nimitz' theater staffs and the Joint War Plans Committee (JWPC) in Washington mainly from the shadows of real operations that had been abandoned. Before trying to grasp the remarkable dynamics that gave rise to these plans, it will be useful to examine carefully the plans' contents.

Contents of Pastel

The final version of the deception plan to shield Operation Olympic (the invasion of Kyushu) was drafted by U.S. Army Forces, Pacific (AFPAC), in Manila on 30 July 1945. It was called Pastel Two. For strategic deception, the plan provided for two fictitious assaults. The first was to be against China's Chusan (Chou-shan)-Shanghai area, with a fictitious landing date of 1 October 1945. The second false assault was to be against Japan's Shikoku Island, north of Kyushu, with the bogus landing set for 1 December 1945 (see map 2). For operational* deception, Pastel Two prescribed fictitious large scale airborne assaults into the interior of Kyushu on the nights before and after the real landings.1

* The documents of the time used the term "tactical" to refer to all combat activities below the strategic level, including the level of activities that today are usually described as "operational."

Map with Korea and Japan
Map 2. Planned assaults in the Kyushu and Shanghai areas

To explain force buildups around the Pacific, the plan's "story" first projected an assault in the farthest plausible place from Kyushu, namely, the Chusan-Shanghai area. This spot was chosen also in hopes of preventing Japanese troop movements out of China. However, since it would soon be obvious that the staging of U.S. forces was not oriented toward the China coast, the story specified that the Chusan-Shanghai operation should appear to have been canceled around 7 September. The target area would then be shifted to Shikoku. The story suggested several reasons for this: the deterioration of the Japanese position in China, satisfaction at the results of the strategic bombing of Japan, and the advantageous acceleration of supply and redeployment activity preparatory to a large-scale invasion.2

The Chusan-Shanghai and Shikoku stories were to be sold with leaflet drops, psychological warfare radio broadcasts, air reconnaissance, bombing and strafing, and a submarine-borne beach penetration landing. The operational deception storythe fictitious airborne strikes at the Kyushu interior-was to be conveyed by placing large numbers of gliders on Okinawa airstrips, by establishing a false airborne corps and division headquarters on Okinawa, and by dropping supplies as if for airborne troops the nights before and after the first day of Olympic.3 In addition, Pastel Two provided for comprehensive communications deception and for an orchestrated sequence of press releases to convey the story. It indicated as well more specific projects, such as circulating rumors in Sixth Army about the use of forces in the Chusan-Shanghai area and sending rubber models of the Chusan-Shanghai terrain, then later of the Shikoku terrain, to the various Pacific headquarters.4

Many particular features of Pastel are of interest. The submarine-borne beach penetration unit was to be provided by Admiral Nimitz, commander in chief, U.S. Pacific Fleet (CINCPAC), on Guam on request of the AFPAC G2 and land in the Chusan-Shanghai area sometime before 18 August to demonstrate U.S. interests.5 Leaflet drops in the Shanghai area were to be provided by the Far East Air Force (FEAF), and the content was to be directed at Japanese military or Chinese civilian morale. Drops were to be at the rate of two per week in early August, three per week in late August, then one every two weeks after the target shifted to Shikoku on 7 September, thus sustaining the threat at a reduced level. FEAF was also to show U.S. interest in the Shanghai region by conducting "aerial reconnaissance, photography, bombing, and strafing missions." These missions were to be flown three times per week before 7 September and once every two weeks thereafter. The extension of leaflet drops and air reconnaissance even after 7 September was required because Pastel called for a latent threat to be maintained against the Shanghai region even though Shikoku had been designated the next main assault zone. The Japanese were to believe that there was some chance the Americans might still land near Shanghai, a belief that would fix Japanese forces in the area. Finally, psychological warfare radio broadcasts were to be transmitted to the Shanghai coast by the AFPAC Psychological Warfare Branch. These broadcasts were to be directed at the morale of the Japanese soldiers or Chinese civilians. Chinese fishermen and junk operators were to be warned to avoid the area north of Taiwan, beginning on 15 September 1945.6

Some or all of these measures were probably meant to be implemented for the second main target area, Shikoku, as well as for the first target area, Shanghai. The Shikoku measures were not prescribed in the Pastel Two text, however, probably because, being a little less pressing, they could be put in directives issued later.

While Pastel assumed several kinds of operational cover, it also advanced one positive ploy of operational deception: large-scale, fictitious airborne strikes behind the Olympic land ing beaches. The drama of the pretended airborne force was to begin in mid-August and develop until the day after X day, the day of the real landings. This elaborate hoax was intended to exploit Japanese fears of U.S. airborne operations.

About 20 August, six gliders and their pilots were to be sent to Okinawa airfields by FEAF to train conspicuously. Thereafter, FEAF was to construct 12 dummy gliders on Okinawa each week until a total of 100 was reached, all to be displayed at or near military airfields. Besides that, a fictional airborne corps headquarters and a division headquarters were to be created on Okinawa around 1 September, and the 11th Airborne, which actually existed in the Philippines, was to be designated as the second division farming the fictional corps. As an added touch of authenticity, 1,000 shoulder patches for the fictional division were to be made and shipped by Joint Security Control (JSC), an agency subordinate to the Joint Chiefs o� Staff (JCS) that handled intertheater deception matters.

These measures were to be followed by parachute drops of equipment into the Kyushu interior on the night before and the night after the Kyushu landings. These supplies would appear to be intended for an airborne unit. On the night of X-1, one drop was to be made behind each of the three landing beaches: at the Kagoshima airstrip behind the Kushikino landing beach, at the Shibushi airfield behind the Ariake Bay beach, and at the Nittagahara airfield behind the Miyazaki beach. On the night after the landings, drops were to be made at the Kitahara airfield in the Kokubu area and at the Miyakonojo North airfield in the Miyakonojo area. Three to four planes were to be used for each day's drops, each plane dropping materials as if to supply an airborne unit. The objective was to hold some Japanese forces away from the beaches on X day (see map 3).7

Pastel Two outlined comprehensive plans for communications deception by manipulation of radio traffic. Communications deception was to be practiced at both strategic and operational levels, with AFPAC, CINCPAC, USFCT (U.S. Forces, China Theater), and USASTAF (U.S. Army Strategic Air Forces) all participating. The Signal Security Agency in Washington was to conduct supratheater traffic analysis to keep the theaters informed of the overall traffic pattern.

At the strategic level, radio traffic to the China theater and the other major Pacific theater headquarters was to be increased to indicate the Chusan-Shanghai assault on I October 1945. After 7 September, these levels were to be reduced somewhat "but still maintained at levels sufficiently high to disguise normal variations resulting from preparations for Operation 'OLYMPIC."' This pattern would encourage Japanese listeners to conclude that the radio evidence was anomalous or else that no large assault preparation was taking place following the cancellation of the Shanghai mission. Radio traffic between the War Department and the major theater headquarters was to be "increased over existing levels by the insertion of control messages." Book messages-namely, the long messages that often were communicated just after a high-level decision had been reached-were to be sent on 4 to 7 September, the moment at which the story specified Washington leaders would cancel the Shanghai operation.8

Pastel Two provided for operational radio deception. Sixth Army radio activities during and after the staging period were to be kept at levels established prior to staging. Naval point-to-point channels were to be controlled to disguise the assault forces' departure. Finally, naval broadcast transmissions were to be controlled so as not to reveal that a large amphibious force was at sea.9

Map of Kyushu, Honshu and Shikoru Islands
Map 3. Parachute drops into the Kyushu interior

In addition to radio traffic deception, the plan called for media deception. Pastel Two set forth an agenda, to be implemented by Joint Security Control, that specified a series of releases to the press or other sources used by the Japanese. These releases were carefully orchestrated so as to suggest and confirm the story indirectly. The plan laid out the program in two-week blocks, and all of the fictional elements-namely, the Shanghai assault, the Shikoku assault, and the airborne assaults-were developed for the period 30 July to 1 November. Apropos of airborne assaults, for example, the program in early August was to reveal the creation of "training areas for rehabilitating airborne units" in the northern United States. September releases were to describe increases in the reconstitution of airborne units and "many gliders passing through west coast ports." October press offerings were to emphasize airborne troops passing through Pacific ports and a particular company's glider production figures.

The strategic assault themes were to be sustained in a similarly indirect way. In early August, the story was to be conveyed that "shipping to China is now being made direct across the Pacific on a small scale" and that a "large number of selected Chinese" were being sent to the Pacific "to be used as interpreters, some being especially trained in government and municipal administration." But already in late August, the press releases were to note that "Shikoku is [the] least defended of the Japanese main islands and provides many potential fighter strips." In early September, it was to be announced that the "U.S. High Command [is] closely watching Japanese dispositions ... in Shikoku." Then, in late October, spokesmen were to tell the press that "Gen. MacArthur has indicated to [the] Joint Chiefs of Staff his inability to improve on the original target date of 1 December for his next offensive." The planners, no doubt wisely, were not going to divulge anything specific in the media about the fictitious strategic assaults, relying instead on quasi-preliminary operations such as leaflet drops and beach penetration parties to encourage Japan's precise identification of the targets. Revealing information too openly would only have made the Japanese skeptical.10

All in all, Pastel Two prescribed two levels of deceptionstrategic and operational-and several different dimensions of deception, including (1) concrete "as if" preliminary opera tions such as beach landings, air reconnaissance, and displaying dummy gliders; (2) radio deception; and (3) media deception in which obliquely misleading information was given to the public media or other sources the Japanese were monitoring. Holding all this complex orchestration together was the story or, rather, the three stories of the Shanghai assault, the Shikoku assault, and the airborne strikes on Kyushu.

Responsibilities for Implementing Pastel Two

Given the great scope of Pastel, its authors saw fit to specify broad areas of responsibility for implementation. Drafted by AFPAC headquarters, the plan specifically allo cated responsibilities within AFPAC. The AFPAC assistant chief of staff for operations, the G3, was to direct all AFPAC commands and agencies in deception activities (see figure 1) and also keep all AFPAC commands "informed as to the status of various phases of this plan." Moreover, the G3 was required to "arrange with Joint Security Control" and the major theater headquarters in the Pacific "for assistance in the implementation of this plan."11

The AFPAC assistant chief of staff for intelligence, the G2, was responsible for deception operations using its own internal means (with the one exception that the G2 was also to "arrange with CINCPAC" for the Chusan area beach penetration party in August). The G2 had its own agenda of media deception items, in addition to that entrusted to Joint Security Control. Besides items for straight release to the press, the G2 agenda included activities such as "initiating rumors in 6th Army regarding possible use of forces in Chusan area" and "canvassing for local Chinese known to be familiar with Shanghai areas."12

Organizational chart
Figure 1. Implementation responsibilities for the Pastel Two

AFPAC's chief signal officer was to direct radio traffic deception within AFPAC and coordinate radio deception with the other Pacific theater headquarters. The AFPAC Psychological Warfare Branch was to arrange the Shanghai leaflet drops with FEAT and make psychological warfare radio broadcasts to the Shanghai area.13

The FEAF theater headquarters also had extensive duties under Pastel Two that included providing leaflet drops, bombing and reconnaissance, gliders for Okinawa, and drops of air borne materiel on Kyushu before and after X day. Naval, Sixth Army, and Tenth Army forces had responsibilities mostly for cooperating in communications deception and providing tactical cover for unit movements where possible.14

In short, AFPAC's G3 handled the main business of coordinating the plan, while AFPAC's G2, chief signal officer, and Psychological Warfare Branch; CINCPAC; FEAF; and USFCT were to carry out appropriate parts of the plan using their own means.

Contents of Coronet Deception

Meanwhile, deception plans for Coronet, the invasion of the Tokyo Plain, were moving along several months behind those for Olympic. These plans had no cover name, and when the war ended on 15 August 1445, the Joint War Plans Committee had just drafted a set of them for forwarding to theater planning staffs. The plans' latest version was thus titled "Staff Study: Cover and Deception Plan for `CORONET"' (Coronet Deception), prepared 17-23 July 1945.

The communications and media deception plans were not developed in detail in this early version, nor were any responsibilities for elements within a theater outlined. The story, however, and the substantive operations conveying the story were clearly set out. Both strategic and operational levels of deception were developed as they were in Pastel Two.

Coronet Deception stipulated fictitious assaults on Shikoku, southeast Korea, and Hokkaido and also local feints on Y day (the day of the real Coronet landings) as tactical deception. The story explaining such assaults was that, before invading Honshu, the United States needed to tighten its naval blockade and intensify aerial bombing of Japan. This meant seizing Shikoku for bomber bases and as an advanced fighter base for southern Kyushu; taking a lodgment on the southeast coast of Korea to obstruct all Japanese force movement from the mainland and to complete the naval blockade of Honshu; and finally, invading Hokkaido to cut off food supplies to Honshu, gain additional air and naval bases, and control the Tsugaru Strait.15

The fictitious Shikoku assault was to take place on 1 April 1946, a month after the first actual Coronet landings, thus holding forces on Shikoku even after Coronet began (see map 4). Like the other fictional assault areas, Shikoku was separated from the Coronet target area by water, which would make it harder to move forces from there back to Honshu for battle reinforcement. "Southern Shikoku in the vicinity of Kochi" was the indicated landing area.

Ground forces were to be manipulated to give the impression that a force was about to invade Shikoku. "An actual commander" was to be assigned to command the fictional Shikoku assault. Five infantry divisions being trained in the Philippines for Coronet and one division in Okinawa were to "be led to believe" that they were to operate in Shikoku. Fictional numbers were to be given to the Philippine divisions so that radio broadcasts using these numbers could maintain these divisions' fictional presence in the Philippines even after their departure for Coronet. This impression was to be reinforced by having service troops show a flurry of activity in the Philippine camps even after the divisions had departed.

Map of Western Pacific Area
Map 4. Planned assaults for Coronet and Coronet Deception

Staff officers in the Philippines were to prepare intelligence studies of Shikoku, and the officers o� divisions involved were "to study maps and models of Shikoku as well as Honshu." Planners also suggested a truck spill of Shikoku maps in crowded Manila traffic, accompanied by a conspicuous effort "to quickly pick up and hide names on maps." Troops were to see movies featuring cities on the Inland Sea, and civil government experts and airfield construction units were to be briefed for work on Shikoku. The division in Okinawa was to be given amphibious training using Shikoku beach names. Japanese prisoners of war might be allowed to witness this and escape.

The Far East Air Force was to provide actual contact operations with Shikoku. It was to drop leaflets telling civilians to avoid airfields and beaches, beginning on 1 March 1446, the actual day of Coronet. "Parachutes and pre-damaged material" were to be dropped near Kochi on 1 January and then at intervals in March. "Time delay bombs and rubber boats" would be dropped off assault beaches "to simulate underwater demolition work." CINCPAC naval forces also had a role, but it was not really specific to Shikoku since carrier and submarine operations were to be applied to the whole Japanese coast from Kyushu to Hokkaido.16

The JWPC's plan for Coronet Deception also envisioned extensive radio traffic deception, media misinformation type deception, and special means deception for Shikoku and the other false assaults, all to be coordinated between theaters by Joint Security Control for all the bogus assaults. This activity is merely mentioned in a few words, however, not spelled out in detail as it was in the final version of Pastel, no doubt because Pastel had already been elaborated for implementation by the AFPAC and CINCPAC theater staffs.17

The fictional assault on Korea was to take place on 1 May 1946, a month after the Shikoku operations, to allow for fictional staging time. The phantom assault's main objective was to prevent Japanese forces' being moved to Honshu before Coronet began. It was also meant to be a plausible continuation of the policy of projecting a fictional strike at the Asiatic mainland that had been a part of U.S. deception since the invasion of Okinawa.

As in the Shikoku operation, an actual commander would be named, and eleven of the Coronet-bound divisions in the Philippines plus one in Okinawa were to be told that they were going to Korea. Fictional numbers were again to be assigned to the divisions and still used for a fictional radio presence even after their departure for Honshu. Staff officers were to study models of Korea, and troops were to see movies of Korea.

FEAF was to drop leaflets, parachutes, and pre-damaged materiel in Korea on 1 November 1945 and to simulate insertion o� agents at intervals after 1 February 1946. Photo reconnaissance was to begin 1 December 1945. Naval deception operations were to include simulated submarine-launched commando raids on the southeast coast of Korea and radio deception measures by submarine in the vicinity of Pusan between 1 December 1945 and 1 March 1946. Besides that, operations against Korean ports were prescribed both to show an interest and to serve as decoy operations for the convoying to Vladivostok that was expected to begin.

Deception for the notional Korean assault included "special operations." Interpreters and civil government experts for Korea were to depart the United States on 15 April 1946. Arrangements were to be made for U.S. currency to be overprinted for use in Korea, and the Red Cross was to be alerted to have personnel ready for transfer to Korea after the invasion. To increase the opportunity for the Japanese to discover the assault, the Red Cross offices in Chungking and in Kunming were also to be advised of this. Sketch maps and intelligence notes of Korea, as well as a complete campaign plan for the Korean operation, prepared by CINCPAC, were to be planted as "lost" in Japanese-held territory by the U.S. China theater commander (commanding general, China [COMGENCHINA]). Finally, beginning on I November 1945, guerrilla forces were to infiltrate Korea and radio out weather information. The weather information was to be useful for the real Olympic operation as well as for the fictional Korea assault. Guerrillas were also to show American interest by blowing up rail bridges and carrying out other acts of sabotage in Korea.18

The fictional Hokkaido landings were to take place last on 1 June 1946 and were to hold Japanese forces on Hokkaido and the Kurils even after Coronet began. Although deception relying on actual ground troops was less developed than in the other fictional assaults, some U.S. divisions bound for Coronet were to be "embarked from Seattle with fictional assignments to the Aleutians." Also, simulated divisions were to inhabit real bases. Aleutian bases were to be kept prepared to stage six divisions. Radio traffic would then simulate four divisions remaining in the Aleutians in the winter of 1945-46 and three divisions staging through the Aleutians in the spring. Radio traffic was also to reveal that a Canadian division would stage through the Aleutians for Hokkaido.

Beginning on 1 February 1946, CINCPAC was to conduct photo reconnaissance of Hokkaido and northern Honshu and dump fast-spreading dyes off the Hokkaido coast where they would be seen by the Hokkaido garrison. CINCPAC and the commander, North Pacific (COMNORPAC), operating jointly, were to have submarines land patrols in the Kurils to place agents or capture prisoners. Leaflets were to be dropped in the Kurils and Hokkaido warning fishermen to avoid certain waters. Radio broadcasts to Hokkaido were to warn civilians to avoid landing areas, and radio traffic "in [a] compromised cryptographic system" was to reveal construction of rocket launchers in the Aleutians aimed at Japan. Radio traffic deception and media misinformation deception were again to be practiced, coordinated by Joint Security Control. The problems of the notional Hokkaido strike were compounded somewhat by the need to coordinate several theater commands besides the usual ones of CINCPAC and AFPAC, namely, COMNORPAC, and COMGENALASKA (commanding general, Alaska)19

Operational deception for Coronet was meant to deflect attention from Sagami Bay, the main assault area to the south, by giving the impression that Sagami was a cover operation and by falsely suggesting progressive landings from south to north (see map 5). The real first assault was to take place on the Katakai-Choshi beaches northeast of Tokyo on 1 March 1945, Y day. The real main assault then was to be launched on the Sagami beaches on the near south side of Tokyo on Y+1Q.

The sequencing of the two real and two fictional assaults was calculated to direct attention away from the real plan. Preliminary operations were to be carried out at Sagami on Y 7. No more was to be done at Sagami, however, until seventeen days later when five other coastal operations would already have taken place, thus giving the impression that the Sagami preliminaries had been a feint (see table 1).

The preliminary operations included, predictably, "heavy bombardment by air and surface vessels, the clearing of beach obstacles, and minesweeping." For the fictional assault areas of Kashima and Sendai, they also included "simulation of destruction of underwater obstacles.... dropping at night of time delay explosives, rubber boats, and other items indicating assault activity," and "simulation of beach reconnaissance, mine sweeping, movement of naval forces, involving the use of smoke, pyrotechnics, radio and radar deception and jamming by submarines or aircraft rescue craft."

Map of planned assualt on Tokyo
Map 5. Planned assaults in the Tokyo (Kanto) Plain area


Table 1
Sequencing of Fictional and Real Assaults in the
Tactical Deception Plan for Coronet

Sequence and Dates of Assault Operations
Y day
Y day
A Preparatory activity at Sagami, south of Tokyo, main assault
B Preparatory activity at Katakai-Choshi, east of Tokyo, secondary assault area assault
C Preparatory activity at Kashima, northeast of Tokyo, dummy assault area
D Preparatory activity at Sendai, far northeast of Tokyo, secondary assault ares assault
A1 Actual main assault at Sagami
B1 Actual secondary assault at Katakai-Choshi
C1 Dummy assault at Kashima
D1 Dummy assault at Sendai
Source: "Deception Plan for 'Coronet,'" Annex D, 24-27.

The second operation was to be an assault preparation in the Katakai-Choshi area beginning on Y-4 and would be followed directly by the actual landings at Katakai on Y day. On the same day, preliminaries were to be carried out at Kashima, northeast of Tokyo, followed by the initiation of preparations in the Sendai area, 200 miles northeast of Tokyo, on Y+4. On Y+8, simulated landings were to take place at Kashima using two divisions in transports. Only then were the main assaults to begin at Sagami, with three divisions on Y+10, building to eleven by Y+30. For good measure, also on Y+10, unloaded transports were to make a landing demonstration off Sendai.

The pattern in this welter of operations was that they moved steadily from south to north, with each preliminary operation followed by a real or simulated assault four to eight days later. This impression was to be strengthened with tactical radio traffic deception that was to "simulate by radio and radar deception the northward, rather than the actual southward, movement of the relief bombardment and Sagami follow-up forces on their withdrawal from Kashima."

The key exception, of course, was the actual main assault at Sagami Bay. The interval between the preliminaries and the assault was seventeen days, not four to eight, thereby suggesting that the preliminaries were a diversion. Moreover, Sagami took place late in the sequence, at the southernmost point, even though all the other operations had moved successively toward the north. The object of all this was to make it difficult for the enemy to read from the preliminaries the existence, timing, and location of the Sagami main assault. The demonstration at Sendai on Y + 10 might seem superfluous since the real landings had already begun, but since it was intended to pin potential reinforcements at Sendai, far from the Tokyo beaches, it still struck American planners as worthwhile. Tactical deception for Coronet, unlike that for Olympic, was an elaborate romance of timing and locale.20

In any case, the content of Coronet Deception, like that of Pastel Two, was highly diverse, relying on radio traffic deception, media deception, and "as ii" preliminary operations, and addressed the deception problem at both the strategic and operational levels.

The Origins of Pastel

Having examined the contents of Pastel Two and Coronet Deception, it is appropriate to ask where these plans came from. As Olympic and Coronet were being developed in the late spring of 1945, deception schemes to accompany these operations were being worked out a month or so behind the plans themselves. Until June 1944, U.S. planners had assumed that south China and Taiwan would be the principal invasion objectives of advancing Allied forces, including the British, and that the war might last into 1948 or longer. In June 1944, however, U.S. planners of the Joint War Plans Committee, subordinate to the Joint Planning Staff (JPS), outlined a new possibility, a rapid advance of U.S. forces by sea, culminating in the early invasion of Japan itself. The Bonins, the Ryukyus, and the China coast near Shanghai were to be intermediate objectives, secured between April and June 1945, with a landing then on Kyushu to take place on 1 October 1945. However, all of this was to begin only after the invasion of Taiwan and the south China coast. These plans, in fact, were titled "Operations Against Japan Subsequent to Formosa [Taiwan]." In other words, the June 1944 plans represented a hybrid or transition phase that retained major invasions on the China coast but added major invasions of Japan's home islands.21

A series of American successes caused planners to take the next evolutionary step, which was to eliminate entirely the earlier preoccupation with the China coast and aim American advances solely at Japan. On 29 March 1445, the Joint Chiefs of Staff set a tentative schedule for the overall invasion plan, Downfall, which called for a 1 December landing on Kyushu. Even so, the Navy leaders -Nimitz, head of CINCPAC, and Admiral Ernest J. King, Navy chief of staff in Washingtonwere still thinking in terms of preliminary invasions of the China coast at Shanghai, Shantung, and Korea, their "roundthe-China-Sea" strategy. General George C. Marshall, Army chief of staff, opposed this, however, and queried MacArthur, who in a 20 April message recommended a direct advance on Japan without a China phase. The JCS adopted this course in principle at the end of April and, after preliminary conferences between AFPAC and CINCPAC in mid-May, issued a directive for Olympic on 25 May.22

The Navy plans for invading the China coast were dead as of the May directive. But their ghost lived on in deception plans for Olympic and Coronet. CINCPAC, on 13 May 1945, produced an early draft of Olympic operations in which the main premise of Pastel was already stated: a fictional attack on the China coast near Shanghai with a fictional attack on Shikoku simultaneously or later.23 Using Shanghai as the target area allowed CINCPAC planners to use their staff studies for the already planned, but now set aside, Chusan-Shanghai operation (Longtom). In fact, CINCPAC planners had just finished drafting the Longtom plans on 18 April. CINCPAC's early Olympic plan also alludes to Operation Bluebird as being successful. Bluebird, the fictional attack on Taiwan and the south China coast that covered the Okinawan campaign, had also been based on a recently canceled real operation. The premise of Pastel thus derived from a long tradition of JWPC planners who had expected to actually invade the China coast and also of CINCPAC planners whose commander had advocated real Chinese operations as late as April 1945. Moreover, the China coast idea was influenced by the recent apparently successful precedent of Bluebird.24

Although the Navy's notion eventually won out, several other agencies besides CINCPAC were simultaneously active in forming the plans for Pastel, especially MacArthur's AFPAC in Manila and the JWPC in Washington. At a conference on Guam from 30 May to 4 June at which AFPAC and CINCPAC planners were addressing Olympic, it was decided that the AFPAC staff in Manila would "prepare a letter plan based upon CINCPAC concept dated 2 June." CINCPAC officers would "visit Manila about 11 June" to confer and then take the plan back "to Guam for CINCPAC concurrence and dispatch."25

The Manila meeting of AFPAC and CINCPAC planners took place on schedule and drafted the first plan labeled Pastel on 13 June 1945. This early Pastel resembled the mature Pastel Two, except that the shift from the Shanghai objective to the Shikoku objective was to be made abruptly on 1 October, the day of the false Shanghai landings. Pastel Two would specify 7 September, not the landing day 1 October, as the less abrupt date for the change of fictional objectives. Also, the early Pastel contained no plans for operational deception other than that it would be provided by the operational commanders' "maneuver of forces."26

Meanwhile, entirely different agencies were generating the overall deception plan known as Broadaxe. This plan was created by the Joint War Plans Committee that, in 1943, had grown out of and was subordinate to the Joint Planning Staff. The Joint Planning Staff, in turn, was a staff agency subordinate to the Joint Chiefs of Staff (see figure 2). The JWPC, in addressing deception matters, often cooperated with Joint Security Control. Joint Security Control coordinated intertheater deception matters, among other duties, and answered directly to the JCS, as did the JWPC's overseers, the JPS.27

Organizational map
Figure 2. U.S. War Department agencies dealing with strategic planing and deception, summer 1945

Early in June 1945, the JWPC and Joint Security Control prepared a "General Directive for Deception Measures Against Japan," which was adopted by the Combined Chiefs of Staff on 16 June. This plan, later called Broadaxe, was then sent to the theater commanders to guide their deception planning. AFPAC and CINCPAC, however, had meanwhile created Pastel with no knowledge that the JWPC directive was coming. In other words, the two plans crossed in the mail. Broadaxe proposed to give the Japanese the impression that invasion of the homeland would be postponed indefinitely. Instead, the United States would try to seize more bases on the homeland's periphery to tighten the naval blockade and increase air bombardment. Broadaxe therefore called for a fictional assault on Taiwan in the late summer of 1945, an invasion of Hokkaido in the early fall of 1945, landings in French Indochina in the fall of 1945, an attack on Sumatra from India in the late fall of 1945, and an "advance into the Yellow Sea" in the winter of 1945-46 to secure bases "for air and inland operations."28

Except for the Hokkaido incursion, the various fictitious operations in Broadaxe retained the China coast emphasis that was the main preoccupation of Washington planners of real operations before June 1944 and that was favored by the Navy for real operations as late as April 1945. The round-theChina-Sea notion was also present in Broadaxe: U.S. forces would invade the south China coast starting at Taiwan, then move their attacks gradually northward. This concept would exercise a powerful influence on the later deception plans for Olympic and Coronet. Eventually, fictitious landings for Taiwan, Shanghai, and south Korea would be planned to cover real assaults on Okinawa, Kyushu, and Tokyo respectively. The phantom landings moved round the China Sea even if the real landings never did (see map 6).

Missing completely from Broadaxe was the Shikoku landing that figured prominently in Pastel. Planners faced the dilemma that, after Okinawa, the fictitious landings targeted on the Asian mainland might not be believed and that fictitious assaults located on the home islands would tend to draw forces to the real assault areas. The designers of Pastel and Coronet Deception would eventually deal with this problem by allocating fictitious assaults both to the Asian mainland and to the home islands, especially those home islands separated from assault areas by water.

After Broadaxe was issued, the AFPAC and CINCPAC theater planners and the JPS and JWPC planners in Washington all realized within a few days that Pastel and Broadaxe had crossed. A meeting was immediately arranged for 27 June 1945, at which theater- and Washington-based conferees could reconcile the two plans. Members of the CINCPAC Special Planning Group, representing also their colleagues in AFPAC, traveled to Washington in late June and met with members of Joint Security Control. Together, these representatives revised the 13 June Pastel plan so that its language was compatible with the larger mandate of Broadaxe. Being for Pacific theater use, the Indochina and Sumatra landings were not mentioned. And, in fact, nothing substantive in the earlier Pastel was changed. The main premises of Pastel, and indeed the format of the 13 June text, remained essentially the same in the 27 June text. The drafters of the 27 June meeting went out of their way to affirm that Pastel was derived from Broadaxe and that Pastel's Shanghai landings were to further the "advances into the Yellow Sea" that Broadaxe advocated. But no detail of the plan was altered in this connection, so the subordination of Pastel to Broadaxe was purely rhetorical.29

Map of Western Pacific Area
Map 6. Planned and real landings in the Japan, Korea, and northeast China areas

All this suggests that, although the theater staffs cheerfully paid lip service to the Washington deception planners' wishes, nevertheless, the actual Pacific deception plans were shaped first and last by the Pacific theater headquarters. Moreover, at least for Pastel, the CINCPAC headquarters seemed to provide the basic ideas, and AFPAC formulated the details.

Still, although the planners at the 27 June meeting made no real changes in Pastel, it did allow the Joint Chiefs of Staff to accept the reworded Pastel as its own. The new Pastel was thus issued as JCS Directive 1410 on 9 July 1945. This marked the point at which the Pastel plan became national policy. Its development was now almost complete.30

Enclosure B of JSC Serial 6117, which was the memorandum revising Pastel after the 27 June meeting, directed AFPAC and CINCPAC to prepare an implementation annex "to show in detail specific responsibilities and timing" for the new Pastel. The result was Pastel Two. AFPAC headquarters elaborated Pastel Two according to the JSC Serial 6117 directive and issued it on 30 July 1945. Following its charter, AFPAC made no changes in strategic deception but abundantly developed operational details for it. These details were probably coordinated with CINCPAC at the joint meeting of the two commands on Guam on 21-23 July.

The operational deception in Pastel Two, which prescribed large-scale fictional airborne assaults against Kyushu, was original to Pastel Two, however, and had no an tecedents in the 13 June Pastel, or in Broadaxe, or in the 27 June Pastel. It is very possible that the operational airborne scheme-all staged from Okinawa, CINCPAC's most forward jurisdiction-was also devised by Navy planners and inserted into the mainstream of joint theater planning at the Guam meeting of 21-23 July.

Thus, Pastel Two, with its extensive implementative annexes drafted by AFPAC on 30 July, marked the completion of deception planning for Olympic. This benchmark had been reached without too much controversy. Up until April 1945, planners had debated where the real assaults of Olympic should take place, but once that was settled and Navy planners offered plausible deception proposals, there was not much controversy over the location of the fictional assaults.

Some discussion continued, however. D. E. Fairbanks Jr., a naval officer, after reading Pastel, complained to Admiral King's staff on 22 June that Shantung was a better objective than Shanghai because it required a more urgent response from the Japanese. Fairbanks also urged that Hokkaido, not Shikoku, should be the homeland target because Japanese home forces would be kept more dispersed. Fairbanks wanted the CINCPAC planners to consider this when in Washington on 27 June.31 These perceptions did not sway the CINCPAC planners and Joint Security Control members at the 27 June meeting, however.

Major General Clayton Bissell, the senior Army representative in the Joint Security Control group, told AFPAC and CINCPAC chiefs in a 31 July memorandum that the 1 December date for the fictional Shikoku landing "would point dangerously close to [the] actual target date for OLYMPIC." Bissell was perhaps reiterating Joint Security Control's concerns that had earlier caused planners at the 27 June conference to change temporarily CINCPAC's Shikoku landing date from 1 December to 25 December.32 It is not clear whether theater planners heeded this complaint either, however. In any case, after two months of interaction between AFPAC, CINCPAC, JSC, and JWPC, Pastel Two of 30 July 1945 would prove to be the final evolution of deception for Olympic.

The Origins of Coronet Deception

The evolution of the Coronet plans was briefer than Pastel's. When the war ended on 15 August 1945, planners had still not conducted the many rounds of consultation for Coronet Deception that had accompanied Pastel. This was probably because Coronet was not to occur until I March 1946, so that resolving the details was less urgent. Moreover, the detailed plans for Coronet itself would not be completed by AFPAC until 15 August.33

When the war ended, Coronet Deception was still only a staff study called "Cover and Deception Plan for `CORONET,' JWPC 190116" that had been prepared on 17 July 1945 by "the Joint War Plans Committee with representatives of Joint Security Control."34 It is not clear exactly what the provenance of JWPC 190116 was. Coronet Deception was detailed, as the work of the Pacific theater planners usually was, and followed closely the deception patterns established by Pastel. This may or may not mean that Coronet Deception was based on material provided to the JWPC by one of the theater staffs.

Presumably, the theater planners would have had little to take issue with in Coronet Deception, since in its premises, timing, choice of targets, and special methods, it closely fol lowed Pastel, which all parties had just agreed to on 27 June. It seems likely that, while working out Pastel, the main deception issues for the invasion of Japan had already been resolved to the satisfaction of all participants.

As it happened, however, an Air Force planning staff group (AFAEP) that had not been a prominent party in the Pastel talks immediately objected to Coronet Deception. AFAEP urged that the plan not be forwarded to theater commanders for information, which was the customary next step, and instead should be redrafted according to its own recommendations. Moreover, AFAEP's recommendations were persuasive.

The Air Force group opposed the deception proposal on the point of Korea as a fictitious target. They felt that, after the real Kyushu landings, the Japanese would not believe Korea as a target. Even if the Japanese did believe it, they would not alter their strategy, which was to move as many troops as they could for defense of the homeland. Moreover, if Japan did concentrate forces in Korea, they would come from Manchuria not from forces earmarked for Japan. Moreover, diversion of American attention from the Japanese home islands might buoy the morale of the Japanese. The Korean ploy, being o� no benefit, was a waste of energy.

Instead of Korea, the Air Force planners recommended an interesting alternative: two fictitious prongs of assault, leapfrogging up the northern and southern coasts of Honshu respectively from the Kyushu beachhead, keeping each leap within range of fighter cover. The north coast advance would aim in several hops at the Fukui-Kanazawa area, and the south coast advance would aim at the area between Osaka and Nagoya. This would seem more plausible to the Japanese than a Korea objective, given the real lodgment already made on Kyushu. Moreover, it followed MacArthur's known leapfrog strategy, which would make it easy to believe. The Shikoku and Hokkaido fictitious landings were also more compatible with a Honshu leapfrogging approach than with a Korea approach. Finally, moving along the coast could be explained by MacArthur's desire to rely as little as possible on naval help because of his known "jealousy" of Nimitz.35

The Air Force ideas were quite plausible. They relied on ground movements determined by air war technology, namely, fighter plane range (albeit fighters older than the long-range P-51). This must have been compelling rationale for the Air Force staff group. Indeed, they acknowledged that they had considered such an actual plan, so the Japanese, too, would be likely to see it as authentic.36

By 7 August 1945, Captain H. R. Thurber, Admiral King's chief of staff, had responded to the AFAEP notes in a memorandum for King. Thurber criticized the leapfrog plan because it would "draw forces up the island of Honshu . . . rather than keeping them dispersed away from the actual objective." He suggested that emphasizing "jealousy" between MacArthur and Nimitz to the Japanese could boomerang and be "harmful to Allied morale."

Thurber felt that keeping Korea as an objective actually would prevent forces there from being moved to the homeland. The Japanese might try to hold Korea even in an emergency in order to contain the "American, Chinese, Russian, and Korean revolutionary threats" there. Moreover, a Korean target would focus attention on the southwestern portion of the Japanese islands, away from the Tokyo Plain. Besides that, Thurber argued, the fictitious assaults on the Taiwan and south China coasts to cover the Okinawa invasion had been believed by Japanese intelligence, fictitious ChusanShanghai assaults for Olympic were already being believed, and Japanese intelligence was also showing concern about an assault on Korea.37

Thurber consequently urged that the Joint Planning Staff approve Coronet Deception in terms of the round-the-ChinaSea tradition that had infused Pacific deception since planning for Okinawa and that had constituted real plans for the Navy up until April. What Thurber did not say but might have said is that suddenly changing the basic concept of deception might confuse not only the Japanese but also the American operators at all levels that were to implement it. In any case, it was once again the Navy that served as the active advocate of the Asian mainland component of deception.

Since the war ended on 15 August, barely a week after Thurber's memorandum to King, it is not clear whether the Joint Planning Staff actually approved Coronet Deception and sent it to the theaters. Given, however, that the JCS, JPS, JWPC, JSC, AFPAC, and CINCPAC had all agreed on the similar outlines of Pastel five weeks earlier and that the Air Force AFAEP was the only group saying "nay," it is likely that the theater commanders and the Joint Chiefs o� Staff would have quickly adopted Coronet Deception as the identifiable companion of Pastel.

On the whole, the development of deception plans for Olympic and Coronet evolved smoothly. The discussion and occasional dissent during the planning usually led to refine ments rather than delay. There were major differences between JWPC's Broadaxe of 16 June and AFPAC's original Pastel of 13 June. Still, the JWPC and JSC gracefully acquiesced to the more detailed planning of the theaters at the joint meeting of 27 June, while also establishing the fiction that Pastel derived from Broadaxe. They performed the planner's duty of supporting the best plan in hand whether it was their own or not. The main conceptual outlines of both Pastel and Coronet Deception probably originated in the CINCPAC staff, which had been developing real round-the-China-Sea assault plans until April. The brief statement of Pastel in CINCPAC's draft of Olympic of 13 May did not differ in its essentials from the mature Pastel Two of 30 July 1445, ten weeks later.

The Likely Effect: Japanese Perceptions

Having looked at the contents of Pastel and of Coronet Deception, what were the deception plans' quality and importance? Pastel provided for fictitious assaults on Shanghai and then on Shikoku, with false airborne attacks in Kyushu as operational deception. Coronet Deception provided for fictional assaults on Shikoku, Korea, and Hokkaido and for a complicated pattern of feints to conceal the timing and weight of the Tokyo landings. Both plans used multiple means to promote their stories, including real preliminary operations, false information released through the news media and agents, and radio traffic manipulation.

Given the great cost in energy and resources, one must ask how much effect these deception operations might have had on Japanese conduct. Even when deception operations are actually carried out, it is difficult to measure their effect, since the enemy's action may derive from considerations other than the deception. Nevertheless, some useful conclusions can be drawn by examining the activities of Japanese Imperial General Headquarters planners from January 1945 to the war's end.

The Japanese appraoch throughout 1945 was to anticipate U.S. plans based either on past American operations or on what Japanese staff members believed to be U.S. interests. Neither method required current intelligence data, and indeed by the summer of 1945, the Japanese had lost their air and submarine capabilities for directly observing their antagonist. So, as American planners painstakingly decided in the first half of 1945 what to do in the second half, Japanese planners also were deciding what the Americans would do. Moreover, the Japanese estimated U.S. intentions merely by simulating the Americans' decision making process: they looked at the facts to determine what military course best served American policy.

The Japanese shadow process reached conclusions broadly similar to those of the American planners. In the Imperial General Headquarters' "Report to the Throne" of 19 January 1945, Japanese strategists announced that the Americans would conduct a "two-pronged advance" from the Marianas toward the Iwo Jima island group and from the Philippines toward either Taiwan, Shanghai, or Okinawa. "The two prongs of the advance would converge in the final assault on the Homeland in the fall of 1945 at the earliest," they wrote.38

The next day, 20 January, Imperial General Headquarters issued an "Outline of Army and Navy Operations," which specified that the "strongpoints to be developed ... include Iwo Jima, Formosa [Taiwan], Okinawa, the Shanghai district, and the South Korean coast." The outline mandated that "preparations for the decisive battle will be completed in Japan proper by the early fall of 1945." Imperial General Headquarters issued an "Outline of Preparations for the Ketsu-Go Operation" on 8 April 1945 that listed seven "Ketsu" (Decision) operations to repel invaders, one for each of six districts in Japan and one for Korea. Ketsu Number Three was for the Tokyo area, and Ketsu Number Six for Kyushu. These two were to receive priority, with final preparations to be completed as soon as possible after 1 October. The Japanese outlined plans to funnel reserves from other areas to Kyushu in case it were attacked or to Tokyo if it were attacked before the Kyushu landings. No other contingencies appear in this series, meaning that the Japanese thought these were the most likely possibilities.39

By early May, battle results in Okinawa led Imperial General Headquarters to conclude that the invasion of Kyushu would come in June. Forces were moved from the Kurils and Hokkaido to Tokyo and Kyushu respectively. Ketsu Number One, defense of Hokkaido, was accordingly scaled down to a holding action only. Meanwhile, shore preparations were behind schedule, and more so around Tokyo than in Kyushu. Fortifications were 50 percent complete around Ariake Bay in `Kyushu; 15 percent complete in the Miyazaki, Satsuma Peninsula, and Fukuoka areas of Kyushu; and not yet begun around Tokyo Bay. These events suggested again that Imperial General Headquarters responded to actual battle developments more than to independent intelligence.40

By mid-July, most of the Japanese materiel had gone to Kyushu rather than elsewhere. Since there were shortages of ordnance, fuel, rations, and transportation to move them, other areas were "drained" so that Kyushu could be rich in supply. Kyushu, though it had only 50 percent of the mortars and antitank guns its quotas required, had nevertheless stockpiled 100 percent of the ammunition, 44 percent of the fuel, and 164 percent of the rations of the established quotas. Even the Tokyo defense zone had supply levels that were far below its quotas, leading planners to worry that effective resistance could not be sustained after the Kyushu campaign. The priority given to supplying Kyushu shows beyond a doubt that Imperial General Headquarters assumed from 8 April on that Kyushu would be the Americans' first objective. This means the Japanese arrived at these assumptions even before the Americans themselves decided what they would do in April and May. 41

The Japanese assumptions were still somewhat inchoate, however, and by no means unanimous. The debate crystallized over the "Situation Estimate for the Latter Half of 1445," issued in July. The main issue was whether the Americans would attack immediately or after a long blockade. A majority favored the former view, curiously mimicking the course of the Americans' actual planning process. If there was to be an invasion, the question was whether the Americans would first take advance bases. The army intelligence staff at Imperial General Headquarters, especially those in Chinese intelligence, were convinced the United States would invade central or north China to "support . . . the Chungking regime." Some officers in the army operations group believed the Americans would land in southern Korea or on Saishu Island in the Korea Strait. Any of these actions would sever Japan's continental supply lines, and a north China or Korea invasion would "check Soviet influence" there.42

The majority of the Japanese high command believed, however, that the Americans would take only the northern Ryukyus and possibly Saishu Island before they landed in Japan proper. A minority felt the first attacks would be in the Tokyo area, but the majority thought that, instead, the Americans' initial landings would take place in Kyushu and Shikoku to acquire the air and naval bases there, followed by final operations on the Tokyo Plain.43

The "Situation Estimate for the Latter Half of 1945" stated the minority views as secondary possibilities. The Shanghai area and southern Shantung were specified as the likely invasion points in China. Invasions in southern Japan were indicated as possible any time after September if China was not invaded and in late fall if it was invaded. Likely target coasts were Tanega Island, southern Kyushu, and southern Shikoku. This very remarkable situation estimate went on to say that, after the Americans secured the southern bases, they would invade the Tokyo Plain. "This operation may be covered by a diversionary feint at Hokkaido," the estimate noted.44

By the time of the July estimate, the war was almost over, but Japanese planners were still anxious. They could no longer observe American bases with submarines and planes, so they feared the Americans might prepare an invasion of Tokyo before Kyushu, and they would not know it. The planners had access to radio intelligence but felt it was "not completely reliable," especially regarding "direction, time, and strength of the attack." They fretted, moreover, that the Americans might forget the well-walled shores of Kyushu and, instead, invade the Nagoya area. Such an operation would cut Japan in two at slight cost because Nagoya was almost undefended and would prevent reinforcements from reaching the Tokyo Plain from the south. The Japanese high command feared that the Americans would strike Nagoya, whatever their earlier plans or interests, precisely because resources had been concentrated in Kyushu and the Tokyo Plain. Similarly, planners worried that the Americans might land unexpectedly on the Sea of Japan side of Honshu instead of the Tokyo side. The Korea Strait was under American control, and the Sea of Japan coast also was almost undefended. These anxieties were only laid to rest by the end of the war on 15 August.45

In a word, the Japanese by the summer of 1945 believed that there would be attacks on southern Kyushu and the Tokyo Plain, but they also feared attacks on Shikoku, north ern Kyushu, the China coast, and Korea. Besides that, they feared the Americans might divert their attacks to central Japan merely because that area was lightly defended, and in any case, they felt they had to make some defensive arrangements for all of Japan and Korea.

Pastel Versus the Japanese Defense Plans

The Japanese assessments of 1945 are important for gauging the effect of Pastel and Coronet Deception, even though the deception plans were not yet carried out. This is because the Japanese analysis shows both the extent to which deception was needed and what patterns of deception the Japanese were predisposed to believe. For one thing, the Japanese estimates show the remarkable extent to which Japanese strategists could construct future American intentions prior to any observation o� their being implemented. The basic parameters of the American exercise were obvious, and deception was unlikely to persuade the Japanese of fictional operations outside certain limits.

The Japanese plans show that the U.S. Navy's steady insistence that the China coast was a good target for deception was correct. In January, the Japanese ordered Taiwan, Shanghai, and the south Korean coast to be fortified. Even as late as July, the timing the Japanese predicted for the Kyushu landings was contingent on possible landings in China or Korea. The Americans' deception for their Okinawa, Kyushu, and Tokyo invasions targeted Taiwan, Shanghai, and the south Korean coast respectively, and the Japanese considered all these areas real targets. In short, the Japanese predispositions show that the U.S. Navy's round-the-China-Sea deception strategy was on the mark.

The same can be said of the fictional Shikoku landings, which were to be the second phase of Pastel and the first phase of Coronet Deception. As of the July 1945 situation es timate, the Japanese assumed that Shikoku as well as Kyushu would be an invasion target. Therefore, the Japanese would probably have believed the fictional threats prepared for Shikoku. For the Japanese, however, Shikoku was seen as a part of the Kyushu expedition not an alternative to it. So Pastel was extremely unlikely to have persuaded them that Shikoku was the only objective in place of Kyushu. The fictional Hokkaido threats, on the other hand, might have been less successful than the Shikoku project since the July 1945 situation estimate had already anticipated that there would be a distractive feint toward Hokkaido and since Hokkaido had already been reduced to a defense-only status in the Japanese adjustments of June.

As for operational deception, the Japanese anticipated airborne operations against airfields, so Pastel's fictional airborne activities on Okinawa might have worked. On the other hand, unfortunately for Pastel planners, the Japanese had guessed exactly the three Kyushu landings' real target beaches: Miyazaki, Ariake Bay, and Satsuma Peninsula. The designers of Pastel's operational deception might have done well to have arranged some of the elaborate shell-game style feinting, adding a fictional beach or two, especially Fukuoka, that Coronet Deception would.46

Operational deception for Coronet, with fictional landings at Kashima and Sendai, was likely to have been successful, since the Japanese anticipated attacks at Kashima. Moreover, the Japanese believed it was likely that the main assault would come on the Kujukuri (Katakai-Choshi) shore with a secondary one at Sagami or Kashima. That Sagami was only a fictional or secondary target was precisely the story that Coronet Deception was trying to convey, so the Japanese might have been easily convinced.47

Examining the Japanese estimates shows that, all in all, it is remarkable how well the Japanese judged the main outline of American intentions, but it is also remarkable how well American planners chose what the Japanese would view as obvious secondary possibilities and exploited them for deception. It is hard to see what more could have been done to outmaneuver Japanese strategists short of changing the goals of the real operations. Mounting a fictitious threat against Nagoya might actually have caused the nervous Japanese staff to draw troops away from Kyushu, but requiring the American planners to have come up with this in response to an eleventhhour concern within the Japanese Imperial General Headquarters would have been expecting a great deal.

Pastel and Principles of Deception

Having examined the likely effect of Pastel and Coronet Deception in light of the empirical record of Japanese estimates, it is now appropriate to consider the deception plans from a theoretical perspective, since this may provide some additional insights.

In his monograph Strategem: Deception and Surprise in War, Barton Whaley observes that a plan should encompass many possible alternative operations. If only one obvious possibility is developed, the planner "forfeits surprise," and his antagonist can bring all his forces to bear on the single objective. On the other hand, if the planner develops even one additional alternative, this may halve his adversary's effectiveness. Adding further alternatives also helps, but at a diminishing rate. Moreover, if the enemy perceives no additional strategic alternative, small tactical ruses, such as creating dummy artillery or dummy radio traffic, will have no effect.48

Whaley's so-called "Theory of Strategem" effectively illuminates both the extreme importance of Pastel and Coronet Deception and these plans' basically sound approach. Pastel presented several plausible alternatives to the Japanese, thereby obliging them to keep many of their forces dispersed. This was about as much as could be done, since as Whaley also observed, any operation can be dissimulated to some extent, but none "to the full extent."49

The direction taken by the Pastel planners can also be better understood by considering Donald Daniel and Katherine Herbig's theory of types of deception. According to Daniel, there are two types. The first is "ambiguity-increasing deception," which confuses the adversary so he is unsure of what to believe. It tries to keep ambiguity "high enough to protect the secret of the actual operation." Daniel's second type of deception is "misleading deception," which does not increase ambiguity for the target but, rather, "builds up the attractiveness of the wrong alternative" in a positive way by providing information that is clear but false. The adversary thus prepares for the wrong operation. In practice, Daniel points out, ambiguity deception and misleading deception blend into each other. Their outcomes tend to be "utter confusion" and "convinced misdirection" respectively.50

Pastel and Coronet Deception blended these two types of deception. Pastel planners used misleading deception in order to achieve ambiguity deception. They could not have expected the Japanese to accept this misdirection literally and to believe that there would be attacks on south Korea and on Shikoku but that Kyushu would be spared. However, by plainly indicating that Korea and Shikoku were to be the landing targets, the planners created an element of ambiguity in the Japanese analysis so that they could not be certain of the number of landing assaults there would be or where or when they would occur. As a result, the Japanese were obliged to keep their forces at least partially dispersed.

Daniel and Herbig also compare the work of a deception planner to that of a playwright. To convey his story, the planner must use "scenery, props, costumes, principals, extras, dialogue, and sound effects," and control the timing and sequence. The deceiver must use many information routes, or "channels," to transmit his story to the deceived. In this, the Pastel planners did fairly well. They created their stories and then promoted them in many ingenious ways. They did not use props, costumes, extras, dialogue, action, and sound effects exactly, but they did use dummy gliders, shoulder patches of nonexistent corps, fictitious divisions, concocted stories in the press, air reconnaissance in areas not to be attacked, manipulation of radio traffic, and much more besides. They used everything from rumors to dropping life rafts-that is, they used many channels to send their story and actually had a detailed timetable, a script, that coordinated the release of different parts of the story. A good deception planner is like a playwright who carefully orchestrates many small elements of reality to give the impression of a larger, but fictitious, reality. In this capacity, the Pastel strategists were resourceful.51

Daniel and Herbig point out further that good deception partially "meshes" with the reality. If some features of the fictional operation are the same as those of the real operation, the enemy can verify them by observation. This also serves as a kind of cover for the real operation, since its activities can be explained as being for the fictional purposes.52 The Pastel planners, of course, observed this caveat, since the story they devised required major force buildups in Okinawa, Hawaii, and elsewhere, much like those of the real operations. At many points, the planners were wise enough to mesh their fictions with reality.

Theorists like Daniel, Herbig, and Whaley distinguish between cover (concealing reality) and deception (projecting a false reality). The Pastel and Coronet Deception staffs devoted little attention to cover except in radio traffic and radio content. This was probably because they believed, with some reason, that the Japanese could not closely observe the bases where assault preparations were to be conducted. Indeed, the Pastel planners had the opposite problem: their fictional preparations-dummy gliders on Okinawa tarmacs, for example-risked being ineffective because of the Japanese inability to observe them. Under other circumstances, however, Pastel's creators would have had to be criticized for their nearly complete lack of attention to cover.

On the whole, however, Pastel and Coronet Deception were fairly sound plans from a theoretical point of view. They projected plausible alternative assaults and conveyed a story that indicated specific assault landings other than those the Americans were actually preparing for, thereby perplexing the Japanese analysis with an unavoidable element of ambiguity. The two plans employed many methods and routes to transmit their story and coordinated them according to carefully made timetables. Both plans made their fictions compatible with much that might be actually observed.


The deception plans for Olympic and Coronet were comprehensive, involving the coordination of multiple theaters and deception practice at both the strategic and operational levels of war. The plans were developed in a timely way, only shortly after the real operations plans, by cooperating planners in the theaters and in Washington. Japanese perceptions were such that, although the Japanese could not easily have been misled about the main thrust of the invasion plans, the American deception would nevertheless have fanned remaining Japanese doubts. This was in part because the Americans applied deception theory well. They promoted several plausible alternatives and conveyed them with an ingenious variety of methods.

Pastel was painted in subtle strokes, the shadow of a shadow. Little attention has been paid to Pastel in the past because it is only the fictional obverse of events that never took place. But Pastel deserves to be studied. It was prepared in earnest by planners who believed the most important operations of the Pacific war hinged on its quality. Pastel was an outstanding product o� the American deceptionists' art.



AFAEP See note 35.
AFPAC U.S. Army Forces, Pacific; headquarters in Manila
Broadaxe Early comprehensive deception plan for Olympic and Coronet, drafted by JWPC and JSC on 13 June 1945
Bluebird Deception plan to cover the invasion of Okinawa, which prescribed fictional assaults on Taiwan and the south China coast
CINCPAC commander in chief, U.S. Pack Fleet; headquarters on Guam
COMGENALASKA commanding general, Alaska
COMGENCHINA commanding general, China
COMNORPAC commander, North Pacific
Coronet Planned invasion of the Tokyo (Kanto) Plain
Coronet Deception Early deception plan for Coronet, drafted by JWPC on 17 July 1945
FEAF Far East Air Force
JCS Joint Chiefs of Staff
JPS Joint Planning Staff
JSC Joint Security Control
JWPC Joint War Plans Committee
Longtom Planned invasion of Chusan-Shanghai area
Olympic Planned invasion of Kyushu
Pastel Early versions of Olympic deception, drafted by AFPAC and CINCPAC on 13 June 1945 and by AFPAC, CINCPAC, and JWPC on 27 June 1945
Pastel Two Olympic deception plan in its final version, drafted by AFPAC on 30 July 1945
USASTAF U.S. Army Strategic Air Forces
USFCT U.S. Forces, China Theater
X day Planned day of the first Kyushu landings, 1 November 1945
Y day armed day of the first Tokyo Plain landings 1 March 1946



  1. U.S. Army Forces, Pacific, "Staff Study: `PASTEL TWO,' Cover and Deception, OLYMPIC Operations" (Manila, 30 July 1945), 1, 3-5, available at Operational Archive, U.S. Naval Historical Center, Washington, DC. The archives is hereafter cited as OA-NHC. The "author" of this document is hereafter cited as AFPAC.
  2. Ibid., 3.
  3. Ibid., 4-6; and Annex 3c, 1-2.
  4. Ibid., Annex 3c(1)(b), 1; Annex 5a; and Annex 5b(2).
  5. Ibid., 6; and Annex 3c(1) (b), l.
  6. Ibid., Annex 3c.
  7. Ibid., 4-5; and Annex 5b(2), 1
  8. Ibid., Annex 5a, 1-2.
  9. Ibid., Annex 5a, 2-3.
  10. Ibid., Annex 5b(2),1-3.
  11. Ibid., 5-6.
  12. Ibid., 6; and Annex 3c(1)(b), 1.
  13. Ibid., 5-6; and Annex 3c, 1.
  14. Ibid., Annex 3c, l.
  15. U.S. Joint Chiefs of Staff, Joint War Plans Committee, "Staff Study: Cover and Deception Plan for `CORONET,' JWPC 190/16" (Washington, DC, 17 July 1945), Enclosure, 2-4; and Annex D, available at OA-NHC. The "author" of this document is hereafter cited as JWPC.
  16. Ibid., Annex B, 12-16.
  17. Ibid., Annex A, 10; Annex B, 17; and Annex C, 23.
  18. Ibid., Annex A, 6-10.
  19. Ibid., Annex C, 19-23.
  20. Ibid., Annex D, 24-27.
  21. Ray S. Cline, Washington Command Past. The Operations Division, United States Army in World War 11 (1951; reprint, Washington, DC: Office of the Chief of Military History, U.S. Army, 1970), 335-38. Cline cites JPS 476, "Operations Against Japan Subsequent to Formosa [Taiwan]," 6 June 1944.
  22. Supreme Commander for the Allied Powers, Reports of General MacArthur. The Campaigns of MacArthur in the Pacific, 2 vols. in 3 (Washington, DC: U.S. Government Printing Office, 1966, 1950), 1:395-99.
  23. U.S. Commander in Chief, Pacific Fleet, "Brief of OLYMPIC: Naval and Amphibious Operations, CINCPAA Staff Study" (Guam, 13 May 1945), Appendix F: Cover and Deception, 98-102, available at the Military Reference Branch, National Archives and Records Administration, Washington, DC, hereafter cited as MRB-NARA. The "author" of this document is hereafter cited as CINCPAC.
  24. For the reference to Bluebird, see ibid., 98. For awareness of the Navy's planning of China invasions up to the spring o� 1945, I am indebted to Roy Skates who, in his unpublished article, "The Decision to Invade Japan" (N.p., n.d.), refers his reader to "Various messages in Nimitz' Command Summary, Book 6 (1 January-1 July 1945), Naval Operational Archive, Washington, D.C."
  25. AFPAC and CINCPAC, "Joint Staff Study: OLYMPIC, Naval and Amphibious Operations" (Guam, 18 June 1945), Appendix H, H5, available at MRB-NARA. This text sets out the resolutions on "Decep tion Planning" reached at the AFPAC-CINCPAC meeting in Manila, 16 May 1945.
  26. AFPAC, "Cover Plan `PASTEL,' Cover and Deception Plan, OLYMPIC Operation" (Manila, 13 June 1945), 5--6, available at MRB-NARA.
  27. Cline, Washington Command Post, 239-41.
  28. JWPC, "General Directive for Deception Measures Against Japan, CCS 284/16/D" (Washington, DC, 16 June 1945), 1-7, especially 4-5, available at OA-NHC.
  29. For an itemization of the language changes agreed to on 27 June 1945, see JWPC and CINCPAC, Special Planning Group, "Suggested Revision of Cover Plan PASTEL, Memorandum for Joint War Plans Committee, J.S.C. Serial 6117" (Washington, DC, 29 June 1945), 1-5, available at OA-NHC. For the text of the new Pastel of 27 June, see "Enclosure A to J.S.C. Serial 6117, Cover Plan `PASTEL,"' 1-7, available at OA-NHC.
  30. U.S. Joint Chiefs of Staff, Joint Planning Staff, "Plan `PASTEL,' J.C.S. 1410" (Washington, DC, 6 July 1945), available at OA-NHC.
  31. D. E. Fairbanks Jr., "Comments on 'Over-All Cover and Deception Plan for the OLYMPIC Operation,' Memorandum for F-1" (Washington, DC, 22 June 1945), available at OA-NHC.
  32. Major General Clayton Bissell, Joint Security Control, U.S. Joint Chiefs of Staff, "Memorandum for CINCAFPAC re PASTEL Meeting of 27 June" (Washington, DC, 31 July 1945), available at OA-NHC.
  33. AFPAC, "Staff Study: `CORONET,' Operations in the Kanto Plain of Honshu" (Manila, 15 August 1945), available at MRB-NARA.
  34. U.S. Joint Chiefs of Staff, Joint Security Control, "Memorandum for the Joint Staff Planners re Staff Study of Cover and Deception Objectives for CORONET" (Washington, DC, 16 July 1945), available at OA-NHC.
  35. [U.S. Army Air Forces, Assistant Chief of Air Staff, Plans], "AFAEP Notes [on] JWPC 190/16" (Washington, DC, early August 1945), 1-4, available at OA-NHC. The AFAEP Notes and responses to them ap pear in files dealing with Coronet in the Navy Yard's Operational Archive. There is no indication on this document as to what the acronym AFAEP meant. AF was a common abbreviation for the Army Air Forces in the World War 11 period. The U.S. Air Force Historical Research Center at Maxwell Air Force Base, Alabama, checked numerous records for me and supplied the name of the subsidiary Army Air Forces office that I have used here in brackets as the most probable author of the document. This, however, is only an educated guess.
  36. Ibid., 3.
  37. Captain H. R. Thurber, U.S. Navy, "Memorandum for Admiral King re AFAEP Notes on JWPC 190/16" (Washington, DC, 7 August 1945), cover page and enclosure, 1-4, available at OA-NHC.
  38. Supreme Commander, Reports of General MacArthur, vol. 2, pt. 2, 577.
  39. Ibid., 583, 585, 601-5; and U.S. Army Forces Far East, Military History Section, Japanese Monograph no. 17: Homeland Operations Record (Washington, DC: Distributed by the Office of the Chief of Military History, 1959), 201-4, hereafter cited as JM 17.
  40. Supreme Commander, Reports of General MacArthur, vol. 2, pt. 2, 62123.
  41. Ibid., 631-32; and JM 17, 126.
  42. Supreme Commander, Reports of General MacArthur, vol. 2, pt. 2, 635.
  43. Ibid., 636.
  44. Ibid., 639-40.
  45. Ibid., 644, 646.
  46. JM 17, 126-30.
  47. Ibid., 70-72, 85.
  48. Barton Whaley, Strategem: Deception and Surprise in War (Cambridge, MA: Center for International Studies, Massachusetts Institute of Technology, 1969), 129-32, 134-35, 138.
  49. See Barton Whaley, "Toward a General Theory of Deception," in Military Deception and Strategic Surprise, edited by John Gooch and Amos Perlmutter (London: Frank Cass, 1982), 190.
  50. Donald Daniel and Katherine Herbig, "Propositions on Military Deception," 157-78, in Military Deception and Strategic Surprise.
  51. Ibid., 159, 161.
  52. Ibid., 170.


Bissell, Clayton, Major General, Joint Security Control, U.S. Joint Chiefs of Staff. "Memorandum for CINCAFPAC re PASTEL Meeting of 27 June." Washington, DC, 31 July 1945. Available at the Operational Archive, U.S. Naval Historical Center, Washington, DC.

Cline, Ray S. Washington Command Post. The Operations Division. U.S. Army in World War II. 1951. Reprint. Washington, DC: Office of the Chief of Military History, U.S. Army, 1970.

Daniel, Donald, and Katherine Herbig. "Propositions on Military Deception." In Military Deception and Strategic Surprise, edited by John Gooch and Amos Perlmutter, 153-77. London: Frank Cass, 1982.

Fairbanks, D. E., Jr., "Comments on 'Over-All Cover and Deception Plan for the OLYMPIC Operations,' Memorandum for F-l:' Washington, DC, 22 June 1945. Available at the Operational Archive, U.S. Naval Historical Center, Washington, DC.

Skates, Roy. "The Decision to Invade Japan.," Unpublished article. N.p., n.d.

Supreme Commander for the Allied Powers. Reports of General MacArthur: The Campaigns of MacArthur in the Pacific. 2 vols. in 3. Washington, DC: U.S. Government Printing Office, 1966, 1950.

Thurber, H. R., Captain, U.S. Navy. "Memorandum for Admiral King re AFAEP Notes on JWPC 190/16." Washington, DC, 7 August 1945.

[U.S. Army Air Forces. Assistant Chief of Air Staff, Plans]. "AFAEP Notes [on] JWPC 190(16." Washington, DC, early August 1945. - Available at the Operational Archive, U.S. Naval Historical Center, Washington, DC.

U.S. Army Forces, Far East. Military History Section. Japanese Monograph no. 17: Homeland Operations Record. Washington, DC: Distributed by the Office of the Chief of Military History, 1959.

U.S. Army Forces, Pacific. "Cover Plan `PASTEL,' Cover and Deception Plan, OLYMPIC Operation." Manila, 13 June 1945. Available at the Military Reference Branch, National Archives and Records Administration, Washington, DC.

______"Staff Study: 'CORONET,' Operations in the Kanto Plain of Honshu." Manila, 15 August 1945. Available at the Military Reference Branch, National Archives and Records Administration, Washington, DC.

______"Staff Study: `PASTEL TWO,' Cover and Deception, OLYMPIC Operations," Manila, 30 July 1945. Available at the Operational Archive, U.S. Naval Historical Center, Washington, DC.

U.S. Army Forces, Pacific, and U.S. Commander in Chief, Pacific Fleet. "Joint Staff Study: OLYMPIC, Naval and Amphibious Operations." Guam, 18 June 1945. Available at the Military Reference Branch, National Archives and Records Administration, Washington, DC.

U.S. Commander in Chief, Pacific Fleet. "Brief of OLYMPIC: Naval and Amphibious Operations, CINCPOA Staff Study." Guam, 13 May 1945. Available at the Military Reference Branch, National Archives and Records Administration, Washington, DC.

U.S. Joint Chiefs of Staff. Joint Security Control. "Memorandum for the Joint Staff Planners re Staff Study of Cover and Deception Objectives for CORONET." Washington, DC, 16 July 1945. Available at the Operational Archive, U.S. Naval Historical Center, Washington, DC.

U.S. Joint Chiefs of Staff. Joint Planning Staff. "Plan `PASTEL,' J.C.S. 1410:' Washington, DC, 6 July 1945. Available at the Operational Archive, U.S. Naval Historical Center, Washington, DC.

U.S. Joint Chiefs of Staff. Joint War Plans Committee. "General Directive for Deception Measures Against Japan, CCS 284/16/D." Washington, DC, 16 June 1945. Available at the Operational Archive, U.S. Naval Historical Center, Washington, DC.

U.S. Joint Chiefs of Staff. Joint War Plans Committee. "Staff Study: Cover and Deception Plan for 'CORONET,' JWPC 190/16." Washington, DC, 17 July 1945. Available at the Operational Archive, U.S. Naval Historical Center, Washington, DC.

U.S. Joint Chiefs of Staff. Joint War Plans Committee, and U.S. Commander in Chief, Pacific Fleet. Special Planning Group. "Suggested Revision of Cover Plan PASTEL, Memorandum for Joint War Plans Committee, J.C.S. Serial 6117." Washington, DC, 29 June 1945. Available at the Operational Archive, U.S. Naval Historical Center, Washington, DC.

Whaley, Barton. Strategem: Deception and Surprise in War. Cambridge, MA: Center for International Studies, Massachusetts Institute of Technology, 1969.

_____."Toward a General Theory of Deception." In Military Deception and Strategic Surprise, edited by John Gooch and Amos Perlmutter, 178-92. London: Frank Cass, 1982.

Dr. Thomas Huber

Dr. Thomas M. Huber

Dr. Thomas M. Huber is a research fellow at the Combat Studies Institute. Fort. Leavenworth, Kansas. A graduate of Oberlin College, he received his master's degree in history from the University of Michigan-Ann Arbor and his doctorate in history from the University of Chicago. He has taught in the Duke University, University of California-Berkeley, and Stanford University history departments. Dr, Huber lived and studied in Tokyo for several years and is the author of The Revolutionary Origins of Modern Japan (Stanford, 1981) and numerous articles and reviews.



The Combat Studies Institute was established on 18 June 1979 as a department-level activity within the U.S. Army Command and General Staff College, Fort Lea venworth, Kansas. CSI has the following missions:

  1. Conduct research on historical topics pertinent. to the doctrinal concerns of the Army and publish the results in a variety of formats for the Active Army and Reserve Components.

  2. Prepare and present instruction in military history at USACGSC and assist other USACGSC departments in integrating military history into their instruction.

  3. Serve as the U.S. Army Training and Doctrine Command's executive agent for the development and coordination of an integrated, progressive program of military history instruction in the TRAI)OC service school system.

U.S. Army Command andGeneral Staff College
Fort Leavenworth, Kansas

Return to CGSC/CSI Publications page

  contact the command and general staff college Contact the Command and General Staff College  | Contact Information Contact the Combined Arms Research Library Last modified: Wednesday, March 28, 2007  

Required Links:
FirstGov | Army Homepage | TRADOC | More US Army Links | AKO
Army Links: Army Recruiting Command | Warrant Officer Recruiting | Go Army | Army ROTC
About Our Site: Privacy and Security Notice | Accessibility Help and Information | External Link Disclaimer
Questions about Fort Leavenworth? Contact Public Affairs Office | FOIA Requests
 Victory Starts Here!