CSI 98-10001

This publication is prepared for use of US Government Officials, and the format, coverage, and content are designed to meet their specific requirements. US Goverment officials can obtain additional copies of this document directly or through liaison channels from the Central Intelligence Agency.

Other requesters can obtain subscriptions to, or specific editions of copies of, this and some other unclassified CIA publications by addressing inquiries to:

Documents Expediting Project (DOC EX)
ANA Division-Government Documents Section
Library of Congress
101 Independence Ave., S.E.
Washington, DC 20540-4172
Phone: (202) 707-9527
Fax (202) 707-0380

or

National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161
Phone: (703) 605-6000 or 1-800-553-6847
Fax: (703) 321-8547

Queries on the availability of this publiction may be directed to the DOC EX Project of the NTIS Office of Customer Services at the respective addresses or phone or fax numbers indicated above. Publications are not available to the public from the Central Intelligence Agency.

This document will be available soon on the Internet at www.odci.gov/csi.

Copyright pending. Not for distribution or reproduction without permission of the author, the Center for the Study of Intelligence, and Harvard University.

Opinions expressed in this study are those of the author. They do not necessarily represent the views of the Central Intelligence Agency or any other component of the US Intelligence Community.


Table of Contents


Foreword

This monograph was produced under the auspices of CIA's Center for the Study of Intelligence and the Harvard University program for Studies of Intelligence and Policy. The idea was to examine the role of signals intelligence* in US military planning during the final stages of the war with Japan in 1945--particularly its contribution to planning for an Allied invasion of the Japanese homeland.

This study was not intended as an argument for or against the use of the atomic bomb against Japan. Obviously, the importance of the bomb in concluding the war was of such magnitude that it is not plausible to examine intelligence related to invasion planning without addressing the question of whether and to what extent that same intelligence might have influenced the decision to drop the bomb. It also is not plausible to argue that the military calculus concerning an invasion of Japan does not bear directly on evaluations of the bomb decision. Nonetheless, the debates and historical studies supporting or condemning the use of the bomb involve factors that go well beyond the scope of this monograph.

The study's basic objective is not to pass judgment on the decisions that were made, but rather to examine the intelligence that was available at the time and to weigh the role this intelligence played or might have played in the deliberations on an invasion.

The author wishes to express his appreciation to those who reviewed drafts of this study and provided constructive comments--particularly military historian Edward Drea.

*In modern intelligence parlance, the term signals intelligence, or SIGINT, is often used to refer to a broad range of intercepted communications.


A Note About the Author

Photo: Douglas J. MacEachin

Douglas J. MacEachin, the author of this monograph, was Deputy Director for Intelligence at the Central Intelligence Agency from March 1993 until June 1995. He joined the CIA in 1965 and, for the next 24 years, worked mainly on research and analysis of Soviet and European security affairs. He was Director of the Office of Soviet Analysis from 1984 until March 1989, when he became Special Assistant to the Director of Central Intelligence for Arms Control.

Mr. MacEachin holds baccalaureate and master's degrees in economics from Miami University of Ohio. During the period 1964-65, he was a full-time member of the faculty there.

Before retiring from the CIA in 1997, Mr. MacEachin was a CIA Officer-in-Residence at Harvard University's John F. Kennedy School of Government. He is now a Senior Fellow at the Kennedy School. Comments on this study may be directed to him at his office there (617-495-0816), or to the CIA's Center for the Study of Intelligence, which published the monograph (703-613-1751).


I. Setting the Goals--Debating and Planning for a Ground Invasion

Japan 1945

As World War II progressed in the Pacific, the Joint Chiefs of Staff (JCS) confronted the prospect that getting an unconditional surrender from Japan might require invading the Japanese homeland. A number of key Navy and Army Air Force officers led by Fleet Admiral Ernest King, Chief of Naval Operations, and General H. H. "Hap" Arnold, Chief of the Army Air Force, argued that a combination of sea blockade and aerial bombardment could produce a Japanese surrender without the need for a ground invasion. Army Chief of Staff Gen. George C. Marshall and his Army planners, however, believed that Japan's surrender on the terms being demanded by the Allies could be assured only by invasion of its home territory. Both sides made legitimate arguments, but the debate also appears to have reflected organizational competition. (1)

Examining the Options

By mid-1944 a consensus had begun to develop on the need at least to plan and prepare for an invasion, even though some officials evidently continued to believe there was a good chance it would not have to be carried out. In early July the JCS approved a report by its Joint Planning Staff (JPS) (2) that said unconditional surrender was to be achieved by undermining Japan's ability and will to resist through sea and air blockades, intensive air bombardments, and destruction of Japanese air and naval strength--and ultimately by invading and seizing objectives in the Japanese industrial heartland. The report called for invasion of the Ryukyu island of Okinawa and the "home" island of Kyushu in order to establish bases for a decisive ground invasion of the Tokyo Plain, the region around the Japanese capital on the central island of Honshu. (3)

This report became the basis for an agreed statement at the Roosevelt-Churchill meetings in Quebec during September 1944. That pronouncement defined Allied military objectives in the Pacific as "invading and seizing objectives in the heart of Japan," after "establishing [a] sea and air blockade, conducting intensive air bombardment, and destroying Japanese air and naval strength." (4)

The US military leadership did not treat the situation as an "either-or" choice of invasion versus blockade and bombardment, but rather as a melding of the two strategic concepts. For General Marshall and those on the planning staffs who agreed with his view, the JCS/JPS report and the Quebec statement amounted to a commitment to plan, prepare, and ultimately carry out the actions they believed would be necessary to gain Japan's surrender on the "unconditional" terms demanded by the Allies.

For Admiral King and General Arnold, the Roosevelt-Churchill declaration was a commitment to continue and even intensify their campaign of aerial destruction and naval strangulation. They saw an invasion of Kyushu--if it should prove necessary--as a means of gaining bases from which to launch an even more devastating air and sea campaign and thereby produce a surrender without having to mount a ground invasion of the Tokyo Plain.

The debate nonetheless continued through the rest of 1944 and the first few months of 1945. Admiral King, while nominally sticking to the position that the end-game would be an invasion of the Japanese homeland, advocated various operations to be undertaken between the seizure of Okinawa and the invasion of Kyushu--for example, attacks on small islands and coastal areas of Japanese-occupied China between Formosa (Taiwan) and Japan. Some analysts have postulated--plausibly--that these operations were seen by their advocates as a way of creating more time for the bomb-and-blockade campaign to produce the surrender they believed could be obtained without an invasion of the homeland. (5)

Considerable debate also took place on the question of an amphibious assault on the northern Japanese island of Hokkaido. The strategic outline accepted by the Joint Chiefs in July 1944 had specifically named Kyushu as the site for the initial invasion. But some planners--with support from General Arnold--argued for attacking Hokkaido first.

Although these discussions initially focused on Hokkaido as an interim step between Okinawa and Kyushu, the debate evolved into an examination of Hokkaido as an alternative to Kyushu. Nearly all members of the Joint War Plans Committee (see footnote 2), however, strongly supported targeting Kyushu rather than Hokkaido. They also objected strenuously to any diversion of resources toward an interim operation. (6)

Invasion Preparations Begin

On 3 April 1945 the Joint Chiefs formally directed Gen. Douglas MacArthur, then Commander in Chief of US Army Forces in the Pacific (CINCPAC), and Adm. Chester Nimitz, Commander in Chief of the Pacific Fleet and the Pacific Ocean Area (CINCPOA), to develop plans and begin preparations for an invasion of Kyushu. (7) This was strictly a planning directive, not an order for implementation.

Even at this stage, Admiral King and Adm. William Leahy (who was Chief of Staff for the President and who functioned as ex officio chairman of the JCS) remained reluctant to treat the invasion decision as a fait accompli. While not directly opposing an invasion, they continued to advocate intermediate objectives along the China coast. But with the invasion of Okinawa in early April 1945, US military and civilian leaders clearly felt growing pressure to nail down the next step in the Pacific strategy. By the end of April, agreement was reached that instructions would be formulated for Pacific commanders to proceed with an invasion of Kyushu. (8)

Invasion Date Set, Commander Named

Still to be resolved was the naming of an overall commander for the invasion. The choice was between the Army's General MacArthur and the Navy's Admiral Nimitz. While this was being worked out, MacArthur and Nimitz issued a joint recommendation on one point on which they did agree--that the target date for invading Kyushu should be 1 November 1945.

The planning directive of 3 April had given 1 December as the invasion date for Kyushu and 1 March 1946 for Honshu. Both MacArthur and Nimitz argued that scheduling the invasion of Kyushu as late as 1 December would subject the operation to weather uncertainties that could cause it to be set back to the spring of 1946, creating a significant delay in ending the war. On 25 May 1945 the JCS finally sent a directive assigning MacArthur "primary responsibility for the conduct of Operation OLYMPIC" (the codename for the Kyushu invasion) and setting the date as 1 November. The invasion of Honshu (codenamed CORONET) remained slated for 1 March 1946. (9)

II. Assessing the Opposing Forces

In mid-1944, as the planning process was picking up momentum, the Japanese Army forces that US intelligence had identified on Kyushu consisted of only one combat division and two depot divisions. (A combat division was the principal mixed-weapon ground combat unit of the Japanese Army, with a troop complement of 16,000. Depot divisions essentially were pre-positioned stocks of equipment and weapons, garrisoned by manpower pools and training staffs; their roles included marshalling replacement forces and creating new combat divisions.) The US War Department's Military Intelligence Service (MIS) believed that more than half of the Japanese military personnel then on Kyushu were Navy ground troops and ground support personnel of the Army and Navy air forces. Japan's 16th Area Army Headquarters, located in northwestern Kyushu, exercised overall command on the island. (10)

As part of the planning process in mid-1944, the Joint Intelligence Committee (see footnote 2) projected that by the time of the contemplated invasion in autumn 1945, the Japanese would have augmented their Army ground forces on Kyushu to six combat divisions while also maintaining the two depot divisions. The Committee estimated that once the invasion began, the Japanese might be able to reinforce the six combat divisions with up to four more, but that ten was about the maximum they could sustain because of geography and supply constraints. This projection would remain the agreed forecast until mid-1945. (11)

SIGINT Provides the Window

Knowledge of the strength and disposition of Japanese defenses that would be encountered in an invasion was heavily dependent on intercepted communications. Allied intelligence services had no effective agents or spy networks in the homeland, nor were there Western sympathizers with access to this kind of information in any detail.

Aerial reconnaissance played an important role in detecting force movements and identifying physically definable targets such as aircraft, airbases, and concentrations of weapons and vehicles. Its overall utility, however, was constrained by weather, darkness, and technology. Prisoner-of-war interrogations had been a source of intelligence for Allied forces throughout the Pacific campaign, but prisoners available as of the spring of 1945 had little if any knowledge about measures being undertaken for the defense of the main islands.

US intelligence units had been intercepting and decrypting Japanese diplomatic communications since well before the outbreak of World War II. Intelligence of this kind was obtained on a regular basis (with short interruptions as the Japanese changed their encryption systems) from 1935 to the end of the war. Japanese naval communications had been deciphered from the beginning of the war in the Pacific. But it was not until April 1943 that a major break was achieved in deciphering Japanese Army ground communications, and significant quantities of this high-grade cipher message traffic did not begin to be available to US intelligence staffs until the end of that year. (12)

Evidence of Japanese Preparations

In the weeks following the dispatch of the planning directive to MacArthur and Nimitz on 3 April 1945, intercepted communications already were showing that the Japanese were expecting attempts to invade their homeland. (13) A message sent by the German naval attaché in Japan, for example, described a report by the Japanese concerning their preparations for Allied landings in the homeland. The report said they expected an assault on Okinawa "shortly" and anticipated that the Allies ultimately would mount an attack on the Tokyo Plain.

It quickly became clear that the Japanese had identified Kyushu as a likely invasion site. Messages in early April 1945 dealt with Japanese mining of harbors and coastal areas of Kyushu and the evacuation of civilians from Kyushu's "areas of coastal defense." Other communications dealt with the assignment of suicide aircraft to Kyushu. (14)

Other intercepted messages at that time provided evidence of large-scale Japanese troop movements from the Asian mainland to the Japanese islands. Based on the listing of vessels involved, intelligence analysts calculated that 30,000 to 60,000 troops were being moved. By mid-April, intercepts confirmed that one of the units being shifted was a combat division previously identified in Manchuria, and that components of it were already arriving on Kyushu. At about the same time, other messages noted that units were moving from the Kurils to the main islands, although their specific destinations could not be determined at that point. (15)

An assessment by MacArthur's intelligence staff near the end of April concluded: "It is apparent [that the Japanese] now consider invasion certain if not imminent," and that their troop movements and unit dispositions reflected preparations for an all-out defense of their homeland. The report said new combat formations were being created on the main islands, the flow of reinforcements to outlying areas had ceased, "troops guarding the close approaches to Japan [such as Okinawa] were dying in place in desperate delaying actions," and Japanese forces were being sent from Manchuria to Japan "to provide strength for a final defense of the Empire." (16)

In early-to-mid-May, with the completion of the movement of the combat division from Manchuria to Kyushu, the US Military Intelligence Service estimated the number of Japanese troops on the island to be 246,000, including 128,000 in Army ground force units. The Intelligence Service estimated that the four additional divisions expected by 1 November, along with a requisite increase in support units, would add roughly 100,000 more Army ground troops.

Figure 1
Mined Areas Disclosed in Intercepted Message of 7 April 1945

Table 1
Projected Force Structure Compared to Components
Identified on Kyushu by Mid-May 1945 a
Projected
Invasion date of 1 November 1945
Current
12 May 1945
Northern Kyushu
(Army Headquarters) a 
Combat DivisionCombat Division b
Combat Division 
Combat Division 
Depot DivisionDepot Division
Depot DivisionDepot Division
Miscellaneous Independent Brigades/RegimentsIndependent Tank Regiment
Southern Kyushu
(Army Headquarters) a 
Combat DivisionCombat Division
Combat Division 
Combat Division 
Miscellaneous Independent Brigades/Regiments 

a Although not stated in the 1944 estimates, subsequent Military Intelligence Service assessments made clear that the projections included the expectation that an Army headquarters would be established in both northern and southern Kyushu to take command of the forces being deployed in each of these areas.
b This was the division that had recently been moved from Manchuria.

This table is Unclassified.

Table 2
Japanese Troop Strength on Kyushu--Projected
Versus Current, Mid-May 1945 a
 Projected
(for 1 November 1945)
Current
(as of 12 May 1945)
Total350,000246,000 b
Army Ground230,000128,000
Navy Ground25,00025,000
Air Ground95,00093,000

a On projections, see JIC 284, 30 April 1945, RG 165, ABC 384 Kyushu (4 July 1944), Sec. 1-B, Entry 421, Box 434. On the 12 May figures, see SRH 195, Bulletin Nos. 62 and 66.
b MacArthur's staff gave its own estimate of current (including the 57th Division) and future troop strengths. Its overall current total was slightly less than the MIS figures--228,000 rather than 246,000. This reflected a lower number of troops ascribed to base and service support functions. The staff's projections of Army and Navy ground troops--225,000 and 25,000 respectively--were almost identical to the War Department MIS figures. MacArthur's staff thought Japanese Air Force ground support personnel would be cut by as much as 50,000 as air units were removed to locations on other islands from which they could still provide air support with less risk from the invading force. This difference did not materially affect the ground combat potential in the force projections of the two intelligence services.

This table is Unclassified.

During May and the first half of June, intercepted communications disclosed the movement of two more divisions to Kyushu--one from Hokkaido and the other from the Korean Peninsula. (17) Messages also showed that an Army-level headquarters had been established in southern Kyushu and that another was situated in the northern part of the island. Because normal Japanese organizational practice was to subordinate three combat divisions under an Army headquarters (roughly comparable to a US Corps), US intelligence analysts viewed the discovery of these two headquarters as tending to confirm their long-held projections of six divisions evenly divided between northern and southern Kyushu. (18)

Intercepted communications also continued to reflect preparations for extensive use of suicide tactics. One series of messages indicated that up to 2,000 obsolete planes and trainers were being assigned to equip and train units for kamikazi missions. Instructions were issued to outfit biplanes and other older model aircraft with night operations equipment.

Other intercepted transmissions contained information on the construction of underground hangars and new, concealed dispersal airfields on Kyushu--which the analysts presumed would be used by suicide aircraft. A message in mid-June contained a Japanese naval base commander's description of progress being made on construction of suicide boats, and others dealt with measures to disperse and conceal unmanned "boat bombs." One intercepted order revealed the presence of a base for piloted suicide torpedoes (kaiten) on the southeastern tip of Kyushu. (19)

Also during this period, intercepted messages began revealing the assignment of naval ground support personnel to missions normally performed by Army troops. These missions included operation of antiaircraft sites around key points such as bridges and roads, and static defense of bases and depots. A series of transmissions also indicated that one of the units being pulled from the Kuril Islands--a force below division size that was specially tailored for combatting amphibious assaults--was headed for southern Kyushu. (20)

As a result of these developments, the US War Department's Military Intelligence Service in mid-June increased its estimate of Japanese military manpower on Kyushu to 300,000. This estimate was disseminated just two days before President Truman was to meet with his senior military advisers to discuss planning for an invasion of Japan. The force developments it described were still consistent with the projections made a year earlier regarding Japanese forces likely to be defending Kyushu by the time of the planned invasion on 1 November 1945.

Table 3
Japanese Troop Strength on Kyushu--Projected
Versus Current, Mid-June 1945
 Projected
(for 1 November 1945)
As of
(12 May 1945)
Current a
(16 June 1945)
Total350,000246,000300,000
Army Ground230,000128,000160,000
Navy Ground25,00025,00045,000
Air Ground95,00093,00095,000

Note: The division suspected to be moving from Hokkaido was still in transit on the data this estimate was disseminated, and, pending confirmation, it was therefore not included in the Military Intelligence Service's Kyushu manpower estimate for this date.

a SRH 195, Bulletin No. 67, 16 June 1945.

This table is Unclassified.

Table 4
Projected Force Structure Compared to Components Identified on Kyushu by Mid-June 1945
Projected (Invasion date of 1 November 1945)Identified  
12 May 194516 June 1945
Northern Kyushu  
(Army Headquarters) Army Headquarters
Combat DivisionCombat DivisionCombat Division
Combat Division  
Combat Division  
Depot DivisionDepot DivisionDepot Division
Depot DivisionDepot DivisionDepot Division
Miscellaneous Independent Brigades/RegimentsIndependent Tank RegimentIndependent Tank Regiment
Southern Kyushu  
(Army Headquarters) Army Headquarters
Combat DivisionCombat DivisionCombat Division
Combat Division Combat Division
Combat Division (Combat Division) a
Miscellaneous Independent Brigades/Regiments Amphibious Brigade

a See Note with Table 3 regarding movement of the division from Hokkaido.

This table is Unclassified.

III. President Truman Discusses Invasion Plans With His Military Advisers

After the German surrender on 8 May 1945, arrangements were made for Truman, Churchill, and Stalin to meet in Potsdam, on the outskirts of Berlin, to try to settle the postwar arrangements for Europe and to reach agreement on coordinated Allied military operations against Japan. This Tripartite Conference was scheduled to open on 15 July--just three months after President Truman had taken office.

A month before the conference, Truman met with his senior advisers to go over plans for ending the war with Japan and to prepare himself for Potsdam. In a 14 June memorandum to the service chiefs setting up this meeting, his Chief of Staff, Admiral Leahy, said the President wanted to:

. . . discuss details of our campaign against Japan. He expects at this meeting to be thoroughly informed of our intentions and prospects in preparation for his discussions with Churchill and Stalin. He wants an estimate of the time required and an estimate of the losses in killed and wounded that will result from an invasion of Japan proper. He wants an estimate of the time and the losses that will result from an effort to defeat Japan by isolation, blockade, and bombardment by sea and air forces. It is his intention to make his decision on the campaign with the purpose of economizing to the maximum extent possible in the loss of American lives. Economy in the use of time and money cost is comparatively unimportant. I suggest that a memorandum discussion of the above noted points be prepared in advance for delivery to the President at the time of the meeting. . . . (21)

Leahy's memorandum was forwarded immediately to the Joint Planning Staff and the Joint War Plans Committee. The latter body had the task of preparing the initial draft of the paper Leahy had requested for the President. (22)

Centrality of the Casualty Issue

Leahy's description of Truman's intent to make his decision on the basis of casualty calculations apparently caught some senior planners offguard. The archival files of papers for this meeting include a memorandum of a telephone conversation on 14 June between the senior Navy representative on the Joint Planning Staff, Adm. Donald Duncan, and his Army counterpart, Gen. George A. Lincoln. Focusing specifically on the "decision" sentence in the Leahy memorandum, Duncan said he found it "a little disturbing . . . it is late in the day to be making decisions . . . when there is a firm directive to do certain things." (He was evidently referring to the 25 May directive.)

After Lincoln affirmed that he considered the commitment to the Kyushu operation to be "a matter of fact," Duncan pointed out that while this commitment may have been made without having been reviewed and cleared by the President, "the Heads of State did approve the overall objective which this directive supports." (23) (He was almost certainly referring to the Roosevelt-Churchill meeting in Quebec.) Discussions among various planners over the next few days focused not only on what the proper casualty estimate would be, but also on whether an estimate should even be offered.

On 15 June 1945, the Joint War Plans Committee submitted its draft of the requested paper to the Joint Planning Staff. (24) The paper presented essentially the same case for an invasion of Kyushu that had been made in the earlier debates preceding the operational directive of 25 May. It also incorporated the same forecast of Japanese forces (six combat divisions, two depot divisions, 350,000 men) that had been presented in intelligence estimates going back to mid-1944.

In response to the presidential request for casualty estimates, the Joint War Plans Committee report laid down strong caveats on uncertainty and emphasized that the level of opposition and the time required to complete the operation could result in major variations. The report then offered the following figures as an "educated guess":

Invasion ScenariosKilledWoundedMissingTotal
Southern Kyushu, followed by Tokyo Plain40,000150,0003,500193,500
Southern Kyushu-Northwestern Kyushu (Japan sur-renders)25,000105,0002,500132,500
Southern Kyushu-Northwestern Kyushu-Tokyo Plain46,000170,0004,000220,000

Note: The JWPC assessment did not give a specific breakdown for each area individually, but a nominal breakdown can be derived by comparing the component figures given for each scenario. For example, the differences between the second and third scenarios for total casualties and numbers killed are 87,500 and 21,000, respectively. The operational difference between these two scenarios is the inclusion or absence of an attack on the Tokyo Plain. Thus, an interpretation could be made that the estimated casualty total for the attack on the Tokyo Plain was 87,500, including 21,000 killed. Subtracting these figures from the first scenario would yield figures for southern Kyushu of 106,000 total casualties and 19,000 killed, and a similar calculation shows 26,500 total casualties and 6,000 killed for northwestern Kyushu. These breakdowns have been used by some scholars for analysis of the JWPC estimates. Such calculations, however, need to be read with the caveat that the JWPC figures were scenario-driven. For example, an estimate for an attack directly on northwestern Kyushu not preceded by an attack on southern Kyushu would probably result in figures that were different from those obtained through this derivative process.

A revised version of the 15 June report was circulated the following day to the Joint Chiefs through the Joint Planning Staff, which had made a few changes to the language. Although most of these were little more than minor modifications to the wording, there were two important exceptions: the Joint Planning Staff version deleted both the entire casualty estimate table and the figure that showed total US personnel (766,700) who would be involved in the Kyushu operation. The JCS draft offered no descriptive language or numbers to replace these deletions. (25)

In an apparent effort to close or narrow the gap between presenting no casualty figures at all and presenting numbers that the Joint Planning Staff was unwilling to use with the President, the Army's Director of Operations, Maj. Gen. J. E. Hull, asked his staff for casualty figures for operations on Okinawa, Iwo Jima, Leyte and Luzon (both in the Philippines), and "overall figures on MacArthur's operations to date." He said these could be used as background at the upcoming meeting with the President, whom Hull described as "very much disturbed over losses on Okinawa." Hull then incorporated the following casualty figures into a summary of the longer report that had been prepared for the President's meeting. (26)

Islands Invaded EarlierUS Killed, Wounded, and MIAJapanese Killed, Prisoners (Not including wounded)Ratio (US to Japanese)
Leyte17,00078,0001:4.6
Luzon31,000156,0001:5
Iwo Jima20,00025,0001:1.25
Okinawa34,000(ground)
7,700(Navy)
81,000
(Not a final count)*
1:2

Normandy: First 30 days--42,000 US casualties

MacArthur's operations overall (March 1944-May 1945)--Killed: 13,742 US; 310,165 Japanese. Ratio: 1:22

* Note: The struggle on Okinawa was ongoing, and ground force casualties there continued to mount. The figures were updated in an 11 July report that was made part of the package of background papers for use at the Potsdam Conference, which took place over the latter half of July. By 11 July, according to the report, the numbers for US casualties on Okinawa since it was invaded in early April had risen to "39,000 ground, 7,700 Navy" versus a "Japanese total of 119,000."

At the same time that General Hull was pulling these figures together, Marshall cabled MacArthur asking for the "estimate you are using for planning purposes on battle casualties in OLYMPIC up to D plus 90." The response from MacArthur's staff (received in Washington on 17 June) projected battle casualties of 50,800 for the first 30 days of the invasion and a total of 105,050 for the first 90 days. It also anticipated 12,600 nonbattle casualties over the same 90-day span. These numbers applied only to operations to seize the southern part of Kyushu. (27)

Marshall then cabled MacArthur again, asking if the figures provided by the latter's staff were intended purely for planning medical requirements or were actual estimates of battle results. Marshall prefaced his question by emphasizing Truman's concern about casualties. Many readers of the original text of this cable have interpreted it as intending to convey the message that the figures provided by MacArthur's staff might be viewed as unacceptably high. (28)

The CINCPAC's answer was delivered to Marshall in time for him to quote from it during the meeting with the President on 18 June. In this message, MacArthur downplayed the figures his staff had sent earlier, describing them as simply cautious logistic planning estimates and maintaining that actual battle losses were likely to be far less. He did not, however, include an explicit figure for what he thought the casualty total would be. (29)

MacArthur's disclaimer notwithstanding, the numbers offered by his staff were very close to the figure for southern Kyushu derived from the estimates contained in the Joint War Plans Committee paper of 15 June. (See the Note, earlier in this section, accompanying the Joint War Plans Committee's casualty estimates.) They were also consistent with a 30-day casualty estimate prepared by Admiral Nimitz's staff:

 First 30 DaysTotal
MacArthur's Staff50,800105,050
Joint War Plans Committee 106,000
Nimitz's Staff49,000 

a JCS 1388/1, 20 June 1945, "Memorandum by the
Commander in Chief, US Fleet and the Chief of Naval
Operations," RG 165, ABC 384 Japan (3 May 1944)
Sec. 1-B, Entry 421, Box 428, NARA.

Presenting the Case

The 18 June meeting with the President was attended by General Marshall, Admiral Leahy, Secretary of War Stimson, Secretary of the Navy Forrestal, Assistant Secretary of War McCloy, Admiral King, Lieutenant General Eaker representing General Arnold, and the recorder, Brigadier General McFarland. Marshall presented the summary report that General Hull had prepared, including the casualty figures from various operations in the Pacific. The report also included the judgment that "There is reason to believe that the first 30 days in Kyushu should not exceed the price we have paid for Luzon." (30)

Admiral Leahy challenged the Luzon comparison; he contended that the casualty rate from an invasion of Kyushu would be more likely to resemble the experience on Okinawa. Noting that the rate on Okinawa had been 35 percent, (31) Leahy suggested that applying that percentage to the number of US personnel to be committed to the Kyushu operation would produce a more realistic casualty estimate. He asked Marshall what the resulting number would be. Marshall did not respond directly; he merely stated that the total number of US personnel committed to the Kyushu operation would be 766,700--the same number that the Joint Planning Staff had deleted from the Joint War Plans Committee's draft of 15 June. The minutes do not show Marshall or any of the other participants taking the logical next step--calculating what a 35-percent share of this total would be. (32)

The minutes of the meeting also reflect little discussion of the size of the Japanese forces expected to be encountered in an invasion of Kyushu. The only reference is a one-sentence statement by Marshall, and even that was made in answer to a question from the President. Marshall cited the longstanding estimate that by November the Japanese would have eight divisions (referring to six combat and two depot divisions) and a total of 350,000 military personnel on Kyushu. (33)

Most of the discussion about Japanese forces focused on prospects for US air and naval forces to succeed in constraining Japanese reinforcement of the island. Marshall repeatedly emphasized the assessment of the Joint Planners that air and naval power had already reduced movement of Japanese shipping south of Korea and should in the ensuing few months "cut it to a trickle if not choke it off entirely." (34) According to Marshall, these judgments were shared by MacArthur and Nimitz. Later, still reading from the Hull memo, Marshall said that by 1 November "our sea action and air power will have cut Japanese reinforcement capabilities from the mainland to negligible proportions." He made the same point when responding to Leahy's criticism of using Luzon as a model for casualty predictions, stressing that although Japanese reinforcement from other areas was still possible, it was becoming "increasingly difficult and painful."

President Gives the Okay

Truman's questions and comments during this exchange reflected his own continuing unease over the level of US casualties. He asked about the possibility that reinforcements would be sent to Kyushu from other Japanese islands, rather than from the mainland. Marshall assured him that all avenues for such movement were being cut.

The President also expressed concern that an invasion of the homeland by Americans could carry a racial connotation in the minds of the Japanese that would unite them for a fight to the finish. Stimson said there was every indication that this would be the case. At the meeting's end, Truman said he agreed that the plan presented by the Chiefs was the best choice under the circumstances, but he added that he "had hoped there was a possibility of preventing an Okinawa from one end of Japan to the other." (35)

Truman gave the go-ahead to continue preparations for the Kyushu operation; he said the decision on a follow-on invasion of Honshu could be made later. That had been the stance proposed in the paper prepared in advance of the meeting. The minutes of the meeting indicate that an explicit rationale for this postponement was to enable the President and his advisers to take into account the impact of the Kyushu campaign and the anticipated Soviet entry into the war. The upcoming test of the atomic bomb may have been an unspoken factor in the Honshu postponement. (36)

Figure 2
Estimated Japanese Dispositions on Kyushu, 9 July 1945

IV. Tracking the Japanese Buildup As Allied Leaders Meet at Potsdam

On 7 July, Truman boarded a ship bound for Potsdam, arriving there on the 15th. Up to the time of his departure for the conference, intelligence had shown the buildup of Japanese forces on Kyushu to be generally consistent with the earlier projections. As of 18 June--the day he met with his military chiefs to discuss plans for invading Japan--only three combat divisions had been fully confirmed on Kyushu, with a fourth believed to be on the way. By the time Truman reached Potsdam, the presence of the fourth division on Kyushu had been confirmed and two newly created divisions had been discovered, bringing the number of confirmed divisions there to six. (37)

Intercepted communications in this time period continued to show Japanese preparations for extensive use of suicide weapons and tactics. Messages in late June described additional bases for piloted suicide torpedoes (kaiten) and preparations for using oil and gasoline incendiary devices. Intercepted transmissions in July dealt with the deployment of a flotilla of 940 suicide aircraft to 18 concealed bases on Kyushu, as well as extensive efforts to reconfigure floatplanes for suicide missions. The same communications also showed training for night suicide attacks. It was becoming increasingly clear that Japanese naval air elements had been completely turned over to the suicide mission. (38)

All of the Japanese force buildup and other defensive preparations that had been identified on Kyushu up to the start of the Potsdam conference in mid-July fell within the original projections. But they had been achieved much sooner than expected. In fact, the estimated manpower level for this force was 375,000--25,000 higher than the forecast provided by Marshall for 1 November. (39) US Military Intelligence Service analysts, moreover, would subsequently learn that the fifth and sixth Japanese divisions discovered on Kyushu had actually been there as far back as the first week in May. (40)

This meant that at the time of the President's meeting on 18 June with his senior military advisers, the number of divisions on Kyushu had already reached the level that Marshall gave as the forecast for the situation on an invasion date still more than four months down the road. One can only speculate as to how much of an impact this information might have had on the discussions at the 18 June meeting had it been known at the time and conveyed to the President. In any event, the long-held projections would be completely shattered by the end of the first week of the Potsdam Conference.

A Burst of Discoveries

On 21 July the Military Intelligence Service's daily summary on Japanese forces reported that three entirely new divisions had suddenly been discovered on Kyushu. Another was discovered within the next few days, bringing the confirmed total to ten combat divisions and two depot divisions. Intercepted communications provided tenuous evidence that an eleventh combat division was being moved there from Honshu. (41)

Figure 3
Estimated Japanese Dispositions on Kyushu, 2 August 1945

At about this same time, analysis of a series of messages and their addressees disclosed that an Army headquarters, which had formerly controlled Japanese forces on southwestern Formosa, had recently been moved to southern Kyushu. After this was confirmed, a further review of the messages indicated that this army headquarters (the 40th) probably had been on Kyushu since June. (42) Analysts had been looking for evidence of an additional army-level headquarters because of the number of combat divisions that were showing up in southern Kyushu.

By 2 August, as Truman was beginning his voyage back to the US from Potsdam, the Military Intelligence Service had confirmed the arrival of the eleventh combat division on Kyushu; it also reported evidence that two more were present or en route. (43) The intelligence data on the continuing Japanese buildup on the island also included a significant number of specialized combat units below division size--such as mixed brigades, tank regiments, and artillery brigades--and a substantial increase in the strength of naval ground troops assigned to defense of bases and support facilities. (44)

The MIS report of 2 August showed that estimated military manpower on Kyushu had reached 534,000. (45) As substantial as this increase was, it still did not include the full personnel of the recently confirmed eleventh combat division because analysts believed this division was not yet fully deployed. Nor did the new estimate include any manpower for the two suspected but yet-to-be-confirmed divisions. These forces together represented the potential for another 40,000 troops.

SIGINT Picture Raises Concerns for Invasion Plans

These numbers clearly demonstrated that the previously predicted cutoff of Japanese reinforcements--confidently anticipated by Marshall and others in the 18 June briefing of the President--had not happened. A palpable sense of alarm over the implications of this intelligence was exhibited in a paper circulated by the chief of MacArthur's intelligence staff, Maj. Gen. Charles A. Willoughby, on 29 July:

The rate and probable continuity of Japanese reinforcements into the Kyushu area are changing that tactical and strategic situation sharply.

Table 5
Projected Force Structure Compared With Components Identified on Kyushu by 2 August 1945
Projected (Invasion date of 1 November 1945)Identified
12 May 194516 June 19452 August 1945
Northern Kyushu   
(Army Headquarters) Army HeadquartersArmy Headquarters
Combat DivisionCombat DivisionCombat DivisionCombat Division
Combat Division  Combat Division
Combat Division  Combat Division
Depot DivisionDepot DivisionDepot DivisionCombat Division a
Depot DivisionDepot DivisionDepot DivisionDepot Division
Miscellaneous Independent Brigades/RegimentsIndependent Tank RegimentIndependent Tank RegimentIndependent Tank Regiment
Southern Kyushu   
(Army Headquarters) Army HeadquartersArmy Headquarters
Combat DivisionCombat DivisionCombat DivisionArmy Headquarters
Combat Division Combat DivisionCombat Division
Combat Division (Combat Division) b Combat Division
Miscellaneous Independent Brigades/Regiments Amphibious BrigadeCombat Division
   Combat Division
   Combat Division
   Combat Division
   Combat Division
   Independent Brigade
   Independent Brigade
   Independent Brigade
   Artillery Command
   Independent Tank Brigade
   Independent Tank Brigade
   Independent Tank Brigade

a The 2 August Military Intelligence Service daily summary showed four
combat divisions in the North, but two days later MIS analysts learned
that one, the 206th, had redeployed south. The same day, a new division,
the 351st, was discovered in the north, bringing the total to 12--
four in the north and eight in the south. A few days later, the presence
of two more divisions, the 216th and 303rd, was confirmed--one moving
into the north and one into the south. These were the two for which
there had been earlier but unconfirmed indications. The total of 14
divisions on the island was confirmed after the war. MacArthur's
Intelligence Staff carried the total as 13, with one still in transit.
b The 77th Division from Hokkaido was thought to be in transit at this
time, but this was not confirmed, and the division's manpower was not
added until about a week later.

This table is Unclassified.

Table 6
Japanese Troop Strength on Kyushu: Projected Versus Current as of 2 August 1945
 ProjectedAs ofCurrent
 1 November12 May16 June2 August a
Total350,000246,000300,00534,000 a
Army Ground230,000128,000160,000346,000 a
Navy Ground 25,000 25,000 45,000 90,000
Air Ground 95,000 93,000 95,000 98,000 a

a The figures listed in disseminated official estimates for 2 August (see figure 3) included 9,000 additional personnel in Army ground units and 2,000 more in air-ground components. These forces, however, were not actually on Kyushu; they were stationed on the outer Ryukyus but had been resubordinated to the 16th Area Army headquarters on Kyushu. For purposes of consistency over time, such troops from the outer Ryukyus have been excluded from this study's comparisons of forces on Kyushu at different points in time.

This table is Unclassified.

At least six (6) (46) additional major units have been picked up in June/July; it is obvious that they are coming in from adjacent areas over lines of communication that have apparently not been seriously affected by air strikes.

There is a strong likelihood that additional major units will enter the area before target date; we are engaged in a race against time by which the ratio of attack effort vis-a-vis defense capacity is perilously balanced.

Unless the use of these [Japanese land and sea] routes [to Kyushu] is restricted by air and/or naval action ... enemy forces in southern Kyushu may be still further augmented until our planned local superiority is overcome, and the Japanese will enjoy complete freedom of action in organizing the area and in completing their preparations for defense. (47)

Referring to the fact that the original estimates had said the projected force of six combat divisions might be reinforced with three or four more "after the operation begins," the Willoughby report said "these divisions have since made their appearance and the end is not in sight." The report acknowledged that some of the new units on Kyushu were at that time not yet fully manned or equipped, but it nonetheless went on to state that "this threatening development, if not checked, may grow to a point where we attack on a ratio of one to one, which is not the recipe for victory." This report, moreover, was disseminated before MacArthur's intelligence staff knew of the tenth and eleventh divisions on Kyushu or of the evidence that at least two more were there or en route.

The intercepted communications also made clear that this buildup had taken place primarily in southern Kyushu, where the US was planning to conduct its landings. Seven of the eleven identified Japanese divisions and most of the independent brigades and regiments were deployed there. This meant that the number of Japanese combat divisions and equivalent forces in southern Kyushu was already more than double the number originally forecast for that part of the island; in fact, it exceeded what had been forecast for all of Kyushu by the invasion date. US intelligence analysts believed that some 320,000 troops--about 60 percent of the total estimated to be on the island--were deployed in the south. (48)

This aspect of the buildup was the focus of an assessment by the intelligence staff of the Sixth Army, which had primary responsibility for ground combat operations in the OLYMPIC operation. (49) The assessment concluded that the invasion of Okinawa in April had "convinced the Japanese that an assault on southern Kyushu would in all likelihood follow soon afterwards," and that they consequently had "spared no effort to build up the mobile combat potential in Southern Kyushu." The report went on to emphasize that "as many combat divisions . . . have been disposed in Southern Kyushu alone as earlier estimates had computed would be allotted to the whole of Kyushu by [the] target date."

Focusing on the contrast between the level of reinforcement already accomplished and the earlier optimism regarding the ability to cut off any significant movements between the islands, the US Sixth Army's G-2 concluded: "[The] ever increasing aerial offensive can be expected to hamper the execution of the enemy's movements and redispositions, [but] it cannot . . . prevent such movements from being taken prior to [invasion] day . . . . " (Emphasis added.)The judgments offered in this report are all the more noteworthy in view of the fact that--although it was disseminated on 1 August--the authors were working with data that were more than a week old, and they were unaware of the full extent of the buildup that would be confirmed by the time their report was issued.

Apparently the stark messages being circulated by the various US intelligence staffs received immediate attention, with the implications flagged in memos circulated at senior military planning levels. Records show, for example, that a summary of the Willoughby report landed on the desk of General Lincoln, who in addition to being the Army's senior representative on the Joint Planning Staff was Chief of the Strategy and Policy Group of the War Department Operations Division. (50)

The sharply increased numbers presented in the 2 August Military Intelligence Service report were incorporated two days later in a Joint War Plans Committee memorandum to the Joint Planning Staff, recommending that: "The possible effects on OLYMPIC operations of this buildup and concentration" of Japanese forces should prompt US field commanders "to review their estimates of the situation... and prepare plans for operations against... alternate objectives." (51) Although this memo was cautiously worded, its message was clear: the dimensions of the opposing forces and defensive preparations on Kyushu mandated a fundamental re-examination of US invasion plans.

Even as this memorandum was being disseminated, intercepted messages confirmed another new division and provided further evidence on the two that had been suspected to be moving to Kyushu. Confirmation of all the communications evidence would bring the total number of divisions on Kyushu up to 14, more than twice the original estimate. Nine of these were in or being deployed to the south--three times the number of divisions that US analysts had initially projected for that part of the island, where the US landings were to take place. Allowing for partial deployment of these divisions prompted another hike in the Kyushu manpower estimate, this time to 549,000. Soon the figure was upped again to 600,000. (52)

V. Top US Officials' Views of the SIGINT Picture

This alarming intelligence picture did not really begin to come together until about the time the Potsdam Conference was getting underway. The record on how much of it reached the senior US officials there is fragmentary. It seems likely that at least the basic information on the overall dimensions of the buildup reached the key military advisers and perhaps the President, but even this is an inferential judgment.

What Did They Know and When Did They Know It?

The buildup of Japanese forces on Kyushu, including the appearance of new combat units, was reported in signals intelligence summaries addressed to senior US policy-makers. Special channels were set up to handle such material addressed to Potsdam, and a scaled-down map room, modeled on the setup at the White House in Washington for charting such information, was established. Both Truman and Stimson confirmed in later years that they had received reports in Potsdam on intercepted Japanese diplomatic communications. (53)

In a brief given to the "Tripartite" military chiefs on 24 July, Marshall said troop strength on Kyushu and on the outlying Ryukyu Islands other than Okinawa totaled some 500,000. (54) As noted in the footnote with Table 6, the US Military Intelligence Service was including some forces that were located in the outer Ryukyus in its total of troops "subordinate" to the command on Kyushu. The total cited by Marshall--500,000--would have been a rounded version of the fast-rising US estimate of Japanese troops under this command at that time.

Nonetheless, official records of discussions between the President and his key advisers while at Potsdam make no reference to information on the sharply increasing Japanese forces on Kyushu. Nor do the memoirs and diaries produced by so many of the participants in those discussions. (55) Admiral Leahy said in his memoirs that "military matters occupied a relatively minor role" in the discussions at Potsdam." (56)

Some academic specialists have suggested or implied that the absence of references to the intelligence stems from the secrecy imposed on the signals intelligence source material codenamed "ULTRA"--secrecy that endured well after the war ended: (57)

Their sensitivity to the secrecy of signals intelligence, however, did not constrain Secretaries Stimson and Forrestal from recording in their diaries the fact that they were receiving intercepted Japanese diplomatic communications. These diary entries are dated at the time they received the information. Forrestal's diary material was published openly in 1951. Stimson's was made publicly available shortly thereafter, and excerpts were incorporated into the 1960 unclassified State Department History of Foreign Relations publication on the Potsdam meetings. In January 1956, President Truman stated at an open conference that he had known of Japan's efforts to enlist Soviet help in brokering conditions for ending the war in the Pacific, and that he had been aware of this before Stalin informed him. (60)

Did the SIGINT Picture Affect the Discussions at Potsdam?

Some scholars also have suggested that knowledge of the level of Japanese reinforcement efforts was reflected in a statement Truman described some years later as having been made by Marshall at Potsdam. In a 12 January 1953 letter responding to a query from an Air Force historian, the President noted that, after he had learned of the successful nuclear weapon test, he asked Marshall about casualties that would be incurred in carrying the planned invasions through to the Tokyo Plain. The published version of Truman's letter states that Marshall told him it would cost "at a minimum one quarter of a million casualties and might cost as much as a million." (61)

A plausible inference some observers have made is that the fact that these numbers were dramatically higher than those which Marshall had presented at the 18 June meeting with the President was probably the result of knowledge that the Japanese were positioning a much larger defense than had been forecast. (62) This notion carries a further implication that Marshall's statement was influential in the decision to use the atomic weapon. If true, this would establish a link between the intelligence reporting and the decision to drop the bomb.

The origins and validity of the statement attributed to Marshall, however, have been a matter of considerable debate. The "million" end of the range, in particular, has been widely challenged as being without basis and as a product of an ex post facto campaign to put forth a rationale for having used the atomic bomb.

A review of documents from the Truman Library shows that Truman's initial draft response to the query describes Marshall only as saying "one quarter of a million would be the minimum." The "as much as a million" phrase was added to the final draft by Truman's staff, so as not to appear to contradict an earlier statement given in a published article by Stimson (which has been widely challenged). (63)

A quarter of a million is roughly the level--220,000--that the Joint War Plans Committee, in its paper prepared for Truman's 18 June meeting, had estimated would result if Japan's surrender required seizing all of Kyushu plus the Tokyo Plain. As was noted earlier in this study, those figures, taking into account various scenarios and the duration of the operation for which the casualties were estimated, were consistent with estimates made at the same time by the staffs of MacArthur and Nimitz.

To many observers at the time, the quarter-million figure put forward in June by the Joint War Plans Committee could well have seemed intimidating. By comparison, the combined casualty figure for MacArthur's campaign through the Philippines, Okinawa, and Iwo Jima was 133,000. For Normandy, from D-Day through 48 days of conflict, losses were 63,360. For the Battle of the Bulge they were 59,000. (64)

The casualty estimates for the invasion of Japan had been constructed before the receipt of evidence that defensive forces on Kyushu would be much higher than initially expected. Even so, they were excluded from the presentation to Truman on 18 June, apparently because of concern over how the President might react. If, as the evidence seems to show, Marshall was indeed in possession of the latest Kyushu estimates at the time the detailed report of the Alamogordo test was being read in Potsdam (21 July), he would have known even then (a) that the overall number of Japanese combat divisions on Kyushu already exceeded what had been expected by the invasion date still three months away, and (b) that the number for the south--where the landings were to take place--was at least double what had been forecast. Under those conditions, it is not unreasonable, as has been argued, to postulate that Marshall could have--without stretching--responded to a question on expected casualties by citing estimates that he had known about earlier but had considered higher than he wanted to accept, or higher than he thought the President could accept.

Such an interpretation, while not unreasonable, is nonetheless conjecture. Whether in fact Marshall actually made such a statement remains a matter of some ambiguity. And even if one concludes that the intelligence on the Japanese force buildup was fully digested by the US officials in Potsdam, there remains the question of whether and in what way it affected the actions taken there.

VI. The Decision To Use the Atomic Bomb

On the evening of 16 July--Truman's second day at Potsdam--he received a cryptic notification that the atomic bomb had been successfully tested early that morning at Alamogordo, New Mexico. On 21 July he received via special courier a detailed report on the test results from Gen. Leslie Groves. (65) (This was the same day the Military Intelligence Service reported that the number of identified combat divisions on Kyushu had jumped to nine.) Three days later Gen. Carl Spaatz headed to Guam as the new head of Strategic Air Forces in the Pacific, carrying written instructions for his new command to deliver the first "special bomb" as soon after 3 August as weather permitted. The document said the instructions were issued "by direction and with the approval of" Stimson and Marshall. (66) According to General Arnold, the dispatch that initiated the drafting of these instructions was sent to Washington from Potsdam via courier as early as 22 July. (67)

On the morning of the 31st, President Truman was given a cable from Stimson requesting approval of a draft public statement planned for release immediately after the dropping of the first atomic bomb on Japan. Stimson's message included an apology for his "haste," but it added that "the time schedule on General Groves' project is progressing so rapidly that it is now essential that [a] statement for release by you be available not later than Wednesday, 1 August."

Truman penciled his response on the back of the Stimson cable, and it was sent immediately to Washington. It said: "Suggestion approved. Release when ready but not sooner than August 2" (i.e., after Truman's departure from Potsdam). (68) The President gave this go-ahead two days before the Military Intelligence Service issued its report on "eleven divisions" that was cited in the Joint War Plans Committee's recommendation for studying "Alternates to OLYMPIC."

This sequence of events is consistent with the weight of evidence from archival documents and from statements and memoirs of the participants in the Potsdam discussions indicating that for all practical purposes the decision on whether to use the nuclear weapon against Japan had already been reached by the time the President arrived in Potsdam. On this point virtually all scholars who have studied the issue seem to concur, however much they may disagree on the motives for its use and whether its use was justified. (69)

On 1 June the "Interim Committee"--a group established by Truman and chaired by Stimson that included political advisers in and out of the government, scientists, and industrialists, with Marshall and Groves also involved--had recommended to the President that the bomb be used as soon as possible, against a military-industrial target in Japan, and without prior warning. This was the governing concept during all of the Committee meetings over the next five weeks. The meetings also featured discussions of drafts and re-drafts of Presidential public statements to be made when the bomb was used. (70)

Debate continued over whether to provide a warning and perhaps a demonstration of the weapon's devastating power. Some scientists outside the Interim Committee dissented altogether from the idea of using the new weapon. But the record shows that Truman agreed with the course of action recommended by the Committee and had every intention of implementing it. (71)

The record of documents and memoirs also shows that, from the time that word of the successful test arrived in Potsdam, the internal discussions there focused on (1) how soon it would be possible to use the weapon--including whether it might be ready before the USSR formally entered the war against Japan; (2) what would be the first target or targets (from a short list that had been already drawn up); (3) the wording of what would become known as the "Potsdam Declaration", which warned the Japanese of the consequences of not surrendering "unconditionally" and outlined in general terms what that meant (but did not warn specifically about the atomic bomb); and (4) the public statement the President should release immediately after the bomb was dropped. The news of the test also sparked further discussion on whether the USSR's commitment to the war was still needed and whether and how to inform Stalin of the bomb. There are, however, no explicit references to the Japanese defensive buildup as a factor in any of these discussions, and no indications that it affected any of the actions taken. (72)

Whether any formal decisionmaking meeting took place among the key Potsdam participants on the subject of using the bomb is itself a question for which the evidence is at best tenuous and conflicting. A Truman biographer has pointed to 24 July, when Truman and Churchill met jointly with their military chiefs, as the day of the "critical moment." But there is no evidence that this subject came up at that meeting. (73)

Some sources have suggested that a key meeting took place on 22 July, the day after Truman received the Groves report on the test results, and that this meeting may have been the occasion when the President asked the question of Marshall that resulted in the "quarter of a million" statement. (74) All references to this meeting appear to be based on Truman's statement in his January 1953 letter to the Air Force historian--and also in his memoirs--that after receiving the "report" (presumably a reference to the Groves document), he called together all his advisers. The records and memoirs clearly establish that Truman did have a private meeting with Churchill that day, with Marshall and Leahy in attendance, to discuss use of the bomb.

A detailed review of the Potsdam records, however--along with accounts of the activities that day of the officials Truman claimed were present at a larger and more formal meeting--suggest that while the President may have engaged in separate consultations with individual advisers at different times, there was no gathering of the whole group as has been sometimes claimed. Based on the record, it seems more accurate to describe the events in Potsdam regarding the atomic bomb as a series of ad hoc "consultations," probably stretching over a few days following the President's receipt of news of the successful test. (75)

A conclusion that no such formal gathering took place is not an argument that concern over the casualty cost of an invasion of the Japanese homeland was not a central consideration in the decisions regarding the use of the bomb. The concern over casualties is clearly reflected, for example, in the discussions between the President and his advisers on 18 June, including Admiral Leahy's questioning of the merits of paying such a price for unconditional surrender. (76)

A further example is Secretary Stimson's memorandum to the President on 2 July, which ultimately evolved into the Potsdam Declaration. As initially drafted by Stimson, this memo was much more explicit than the version adopted at Potsdam on conceding to the Japanese the right to maintain the institution of the emperor. Stimson described his intentions as seeking Japan's surrender without incurring the high casualties he feared would result from an invasion. (77)

Nor should the evidence be construed as indicating that concern over casualties was the only factor exerting critical force on the A-bomb decision. What the evidence does indicate is that the view of the bomb as a potential way to end the war quickly--in the hope of (1) avoiding the need for an invasion with resulting casualties that by any standard would be of intimidating proportions, (2) minimizing the USSR's postwar leverage, and (3) not having to confront debate over concessions on the terms of unconditional surrender--was the driving force in the minds of the US leadership team before Potsdam, and before the acquisition of intelligence showing much-larger-than-expected Japanese forces on Kyushu. Nonetheless, it is certainly plausible that the buildup disclosed by early August reinforced the belief that the decision to use the bomb was the path of least resistance.

VII. What If the A-Bomb Had Not Been Ready?

The impact of signals intelligence on the decisions at the end of the war thus falls into the "what if" category. There are of course libraries of "what if" analysis on the dropping of the atomic bomb. In this case, however, the "if" question can be narrowed to the specific issue of the impact of signals intelligence, and for that there is an empirical base.

The recommendation to examine alternative invasion sites, outlined in the 4 August memorandum by the Joint War Plans Committee, was explicitly tied to the intelligence derived through intercepted communications. Enclosed with this memorandum was a draft cable that the committee recommended be sent to MacArthur and Nimitz. The cable referred to highly classified reporting that

Indicated a strengthening of Japanese forces and measures in southern Japan [Kyushu] to an extent considerably in excess of that previously estimated as Japanese capability by OLYMPIC target date. While these measures ... are not yet considered to require change to your current directive it is desired that you... make alternative plans and submit timely recommendations. Operations against extreme northern Honshu, against the Sendai area [in northeastern Honshu] and directly against the Tokyo Plain are under intensive study here. (78) (Emphasis added.)

Although this cable, adhering to bureaucratic rules, would have told the Pacific commanders that it was not yet necessary to change their current operational plans, it also said clearly that they should begin preparing for such a turn of events. The question of making such a change was about to come before the Joint Chiefs.

On 6 August--the morning after the A-bomb had been dropped on Hiroshima, (79) and while the results were still being evaluated-- Marshall was notified that, as a consequence of the concerns expressed by the Joint Planning Staff, the next meeting of the Joint Chiefs probably would address the issue of the Japanese buildup on Kyushu and examine alternatives to the planned invasion. (80) There is every reason to believe that, if the atomic bomb had not been ready and used when it was, this JCS meeting would have been held.

Had the meeting taken place, the earlier arguments for avoiding an invasion altogether, relying instead on sea and air strangulation and destruction, almost certainly would have been revived. Such arguments by Admiral King and General Arnold had never really been overridden; advocates of this position had simply been placated. If the bomb had not been available, King's and Arnold's views would have been buttressed by the SIGINT evidence indicating a potential escalation of the estimated costs of an invasion.

Re-Evaluation of the Casualty Estimates?

There is no record that any revised casualty estimates were actually produced as a result of the dramatically changed SIGINT picture of the opposing forces that an invasion would have encountered. But a meeting held specifically in response to intelligence showing a much- larger-than-expected buildup of opposition forces would not have been able to duck the casualty implications of that information.

The original estimates by the Joint War Plans Committee and by MacArthur's staff had been produced when both groups were forecasting opposing forces only half the size that now awaited them. And even those casualty estimates had been purposely excluded from the briefing of a President who had said he planned to "make his decision...with the purpose of economizing to the maximum extent possible in the loss of American lives."

The most recent US experience bearing on the casualty question in the Pacific war was Okinawa. Japanese regular Army troops and naval ground-based defense forces on that island when it was invaded by US forces in April 1945 totaled some 75,000. Also on hand were about 25,000 civilian-paramilitary defenders and an unknown number of additional civilian volunteers. By the beginning of August the US casualties in the ongoing struggle on Okinawa had reached 49,000. (81)

Attacking Kyushu would have meant invading an island many times larger than Okinawa; southern Kyushu alone is well over twice Okinawa's size in square miles. Kyushu was initially expected to be garrisoned by Japanese ground combat forces roughly three-and-one-half times the size of the forces on Okinawa. (82) Kyushu also had a civilian augmentation potential many times greater than Okinawa's. The initial estimates by the Joint War Plans Committee and MacArthur's staff of casualties that would be incurred in capturing southern Kyushu (105,000-106,000) were a little more than twice the Okinawa total.

By the first week in August, the estimated total of Japanese Army and naval ground combat troops on Kyushu was more than six times what it had been on Okinawa. Intercepted communications had been showing Japanese preparations to employ the same kinds of suicide attacks and other unconventional tactics and devices that had caused so many casualties in the Okinawa operation. The number of US Army and Marine troops to be committed in the landing was about three times the force that had been launched against Okinawa. (83)

These figures would have given some senior officials-- including Admiral Leahy, who had already challenged the earlier presentation--a powerful case that a plausible casualty estimate had to be significantly more than double the Okinawa level. Leahy had supported the blockade-and-bomb strategy in earlier debates, and he was the only participant in the 18 June White House meeting recorded in the minutes as saying that he did not think unconditional surrender was worth a high cost in American casualties.

It is not clear at what stage Secretary of War Stimson might have been drawn into the debate. His concerns about casualties, however, had also been clearly reflected in the 18 June meeting and in his efforts on the Potsdam Declaration. (84)

The Argument for Staying the Course

The opposite pole of the debate would doubtless have been the position supported by MacArthur. He favored going ahead with the Kyushu invasion as planned.

When told that alternatives to Kyushu would be the main issue at a coming JCS meeting, Marshall sent a personal cable to MacArthur soliciting his views. (85) Marshall emphasized the large Japanese land and air buildup on Kyushu that had been reported in intelligence, noting that, if the Japanese were in fact deployed in such numbers there, US landing forces risked heavy losses in their amphibious attacks. Pointing out that the buildup on Kyushu had been carried out at the expense of reductions in other locations, Marshall queried MacArthur about "possible alternative objectives" at less defended sites, pointing to the three that the Joint War Plans Committee's 4 August memo had characterized as "under intensive study here."

MacArthur's response was dismissive of the reported buildup:

MacArthur argued that "there should not, repeat not, be the slightest thought of changing the OLYMPIC operation. Its fundamental purpose is to obtain air bases under cover of which we can deploy forces to the northward into the industrial heart of Japan. The plan is sound and will succeed." He concluded: "Throughout the Southwest Pacific Area campaigns, as we have neared an operation, intelligence has invariably pointed to greatly increased enemy forces. Without exception, this buildup has been found to be erroneous." (86)

It is worth noting that MacArthur did not argue that the buildup--if true--should not be viewed as threatening the success of OLYMPIC. Instead he tried to impeach the accuracy of the reporting. This tactic could be interpreted as an indication that he recognized that if SIGINT reflecting the buildup was accurate--or if it was accepted as accurate by Washington--it would indeed have significant implications for the invasion plan.

MacArthur's practice was to not allow intelligence to interfere with his aims, and his history of complaints about Willoughby's reports resulted mainly from their contradiction of his own estimates and preferred courses of action. His denigration of the reported buildup on Kyushu directly contradicted the performance record of his G-2 under Willoughby. In those instances during MacArthur's Pacific campaign when the ULTRA-derived assessments were not entirely accurate, the errors tended to be on the low side. (87)

In this instance, postwar information would show that there had in fact been 14 Japanese combat divisions on Kyushu--and that intercepted communications had identified all of them. The exact locations of a few of the newest arrivals had not been determined at the time of the war's end, but it is quite likely that, once their existence on Kyushu had been confirmed, finding their locations would have been accomplished within a few weeks at most. Japanese documents obtained after the war showed that at the time the US Military Intelligence Service was estimating 600,000 troops on Kyushu, there were 900,000 soldiers assigned to its defense. (88)

Looking for a Middle-Ground Strategy

A middle ground between an invasion of Kyushu and a blockade-and-bomb strategy would have been the approach proposed by the War Plans Committee, apparently with some support from Marshall. The Committee continued to insist on an invasion of the Japanese homeland but sought a target less well defended than Kyushu.

The views of General Marshall and most of the Joint War Plans Committee on the obstacles to achieving unconditional surrender would have made it difficult for them to abandon their advocacy of an invasion of the Japanese homeland. The unconditional surrender objective was about much more than the status of the Emperor. Indeed, the latter issue was the easiest to resolve. While some Allied governments, especially the Australians and Chinese, remained opposed to retention of the Emperor, many key UK and US officials--including some members of the JCS--were not only willing to allow the Emperor to remain, but actually favored doing so in the belief that this would facilitate the administration of Japan by a postwar occupation force. (89)

The more important Allied objectives of unconditional surrender were the unrestricted occupation of Japanese territory, total authority in the governing of Japan, dismantlement of Japan's military and military-industrial complex ("demobilization"), a restructuring of Japanese society ("demilitarization"), and Allied-run war crimes trials--in effect doing to Japan what was being done to Germany. Abandoning these goals would mean Japan would not suffer the same consequences as Germany. Truman's consciousness of the political side to this issue was indicated in his meeting with his military advisers on 18 June, in which he said that he was deliberately leaving the door open to a modification of the surrender terms but that the initiative would have to come from Congress. (90)

Achieving the surrender and unrestricted occupation of the entire national territory of an opponent steeped in a warrior tradition and a history as a great power, without having captured any portion of that territory, posed an extraordinary challenge. It had not been achieved in Germany without invasion:

Whether the Allies' demands could be achieved without capturing any part of the Japanese homeland was really what the debate between invasion and bomb-and-blockade was all about. By early August the casualty costs of an invasion would have added credibility to the case for bomb-and-blockade. That strategy's downside was time: how much destruction had to be imposed, and for how long, and how many more thousands of Japanese had to be killed by bombing or starvation to achieve unconditional surrender?

Implications of Soviet Entry Into the Pacific War

By this time (early August), the prospect of Soviet entry into the war against Japan would have provided arguments to both sides. For those favoring a bomb-and-blockade approach or even just a postponement of any invasion, Soviet entry could have been cited as an additional reason why surrender could be obtained without invading the main Japanese islands. (92)

On the other hand, if Japan's surrender did not take place until after the Soviets had been in the Pacific war for some length of time, and if there were no US forces on Japanese territory (because no US invasion had occurred), how could the United States and its Allies acquire the control over occupation that they were seeking? Unless the mere entry by the USSR somehow caused an immediate surrender on the unconditional terms being demanded by Washington, the Potsdam experience was likely to reinforce the tendency among at least some US officials to see any gain resulting from Soviet entry as also carrying a serious potential cost--the possible emergence of a Far Eastern version of the Soviet hegemony that was beginning to be imposed on Eastern Europe.

Weighing Alternatives

These considerations supported the idea of searching for an alternative that still involved capturing some Japanese homeland territory. The appreciation that Marshall and many members of the Joint War Plans Committee would have had for the casualty implications of the Japanese buildup probably would have led them--MacArthur's views notwithstanding--to look for alternative invasion sites.

In addition to the choices suggested in Marshall's cable to MacArthur and in the Joint War Plans Committee paper of 4 August, there was the option of keeping Kyushu as the target but postponing the ground invasion so as to allow the increased air power from bases being set up on Okinawa to administer an extended pounding. Such an intensified air bombardment campaign had been slated to begin in mid-September; the Army Air Forces at MacArthur's request had already accelerated this timetable by 30 days because of Willoughby's recommendation based on the buildup that had been observed.

Timing and weather posed potential problems for the option of choosing an alternative invasion site. Such a major change in plans at this time presumably would have forced a delay in launching the invasion. As noted earlier, the date for invading Kyushu had already been moved up from 1 December to 1 November in response to concerns expressed by MacArthur and Nimitz, among others, over the greater chance of adverse weather during an invasion that did not begin until December and the possibility that such conditions could set the invasion back to spring 1946.

From this perspective, there was little difference between seeking new invasion alternatives and opting for a bomb-and-blockade strategy. Each involved putting the invasion on hold and engaging in an intensified air and sea attack; if that did not produce a surrender within the next six months, the invasion issue might or might not be back on the planning board. The military alternative to this course of action was to go ahead with the invasion and risk the high casualties. The political alternative was to relax the terms for surrender.

Japanese Perspectives

This was exactly the dilemma that Japanese military leaders had sought to force the United States and its Allies to face. They wanted to buy time in the hope that war-weariness in the Allied countries, in combination with concerns about high casualties, would produce a softening of the unconditional surrender demands. Even for the Japanese, the issue was not whether they would be forced to surrender, but rather on what terms. The best leverage for Japan's leadership was to raise the cost perceptions--both military and political--for Allied decisionmakers.

The downside for the Japanese from Allied decisions dragging out the war would not only have been the devastation and loss of life that would have resulted from bombing and sea strangulation. The longer the war lasted, the longer and deeper Soviet participation would become. And attaining a satisfactory postwar settlement once the inevitable surrender did take place would probably have been more problematic.

Any attempt to conclude how the debate among US leaders over invading Japan would have come out if the atomic bomb had not been available to end the Pacific war abruptly would be a matter of guesses and probably preferences. The planned JCS meeting that was to examine alternatives to an invasion of Kyushu did not take place because the atomic bomb was dropped at the very time the meeting was being scheduled.

This potentially historic meeting had been proposed in direct response to the picture of an accelerating Japanese buildup portrayed by signals intelligence. Had the bomb not been ready when it was, and had the meeting gone ahead, history may well have judged this critical re-examination of strategic choices as one of the most pivotal contributions of SIGINT to the outcome of the Pacific war.

Appendix A: References

Archival Sources

Documentary History of the Truman Presidency, Vol. I: The Decision to Drop the Bomb on Japan (Dennis Merrill, ed.). University Publications of America, 1995.

Foreign Relations of the United States: Diplomatic Papers--The Conference of Berlin (The Potsdam Conference) 1945, Vols. I and II. Washington, DC: Government Printing Office (GPO), 1960.

Record Group 165, Records of War Department General and Special Staffs. National Archives and Records Administration (NARA), College Park, Maryland:

Record Group 457, Records of the National Security Agency, NARA, College Park, Maryland:

Books

Allen, Thomas B. and Polmar, Norman, Codename Downfall: The Secret Plan to Invade Japan - and Why Truman Dropped the Bomb. New York: Simon and Schuster, 1995.

Alperovitz, Gar, The Decision to Use the Bomb and the Architecture of an American Myth. New York: Alfred A. Knopf, 1995.

Andrew, Christopher, For the President's Eyes Only. New York: Harper Perennial, 1992.

Arnold, H.H., Global Mission. New York: Harper and Bros., 1949.

Byrnes, James, Speaking Frankly. New York, Harper and Bros., 1947.

Churchill, Winston, The Second World War: Vol. VI, Triumph and Tragedy. London: Cassell and Co., 1954.

Cline, Ray S., Washington Command Post: The Operations Division. Washington, DC: GPO, 1951.

Drea, Edward, MacArthur's ULTRA: Code Breaking and the War Against Japan, 1942-1945. Lawrence, Kansas: University Press of Kansas, 1992.

Ferrell, Robert H., Off the Record: Private Papers of Harry S. Truman. New York: Harper and Row, 1980.

Hayes, Grace Pearson, The History of the Joint Chiefs of Staff in World War II: The War Against Japan. Annapolis: Naval Institute Press, 1982.

Leahy, William D., I Was There: The Personal Story of the Chief of Staff to Presidents Roosevelt and Truman, Based on His Notes and Diaries Made At the Time. New York: Whittlesey House, 1950.

Lewin, Ronald, The American MAGIC: Codes, Ciphers and the Defeat of Japan. London: Farrar, Straus, and Giroux, 1982.

McCloy, John J., The Challenge to American Foreign Policy. Cambridge: Harvard University Press, 1953.

McCullough, David, Truman. New York: Simon and Schuster, 1992.

Milles, Walter, ed, The Forrestal Diaries. New York: The Viking Press, 1951.

Skates, John Ray, The Invasion of Japan: Alternatives to the Bomb. Columbia, South Carolina: University of South Carolina Press, 1995.

Stimson, Henry, and Bundy, McGeorge, On Active Service in Peace and War. New York: Harper and Bros., 1947.

Truman, Harry S., Memoirs: Vol. I: Year of Decision. Garden City, New York: Doubleday, 1956.

Walker, J. Samuel, Prompt and Utter Destruction: Truman and the Use of the Atomic Bomb Against Japan. Chapel Hill: University of North Carolina Press, 1997.

Weintraub, Stanley, The Last Great War: The End of World War II. July-August 1945, New York: Penguin, 1996.

Winterbotham, F. W., The ULTRA Secret, New York: Harper and Row, 1974.

Articles

Asada, Sadao, "The Shock of the Atomic Bomb and Japan's Decision to Surrender." Manuscript submitted to National Security.

Bernstein, Barton, "Understanding the Atomic Bomb and the Japanese Surrender." Diplomatic History, Vol. 19, No. 2, Spring 1995, pp. 227-274

Bix, Herbert P., "Japan's Delayed Surrender: A Reinterpretation." Diplomatic History, Vol. 19, No. 2, Spring 1995, pp. 197-226

Elsey, George M, "Some White House Recollections." Diplomatic History, Vol.12, Summer 1988, pp. 357-364.

Pape, Robert A., "Why Japan Surrendered." International Security, Vol. 18, No. 2, Fall 1993, pp. 154-201.

Walker, J. Samuel, "History, Collective Memory, and the Decision to Use the Bomb." Diplomatic History, Vol. 1, No. 2, Spring 1995, pp. 319-328.

Appendix B: Data Annex

This appendix describes the chronology of the signals intelligence discovery path for the Japanese force buildup on the island of Kyushu from mid-April through early August 1945. The detailed source references for the increments to the military estimates are shown in the chronology table that follows the narrative.

Intercepted messages confirmed on 18 April 1945 that the 57th Division, previously identified in Manchuria, was part of a large movement of Japanese forces being shifted from the Asian mainland and that components of this division were already arriving on Kyushu. At about the same time, US military intelligence learned that the 3rd Amphibious Brigade, a unit designed for countering amphibious invasions, was moving from the Kurils to somewhere in Japan's main islands. Analysts began to pick up evidence that this brigade was headed for Kyushu, but its actual presence there would not be confirmed until mid-May.

The movement of the 25th Division from the Korean Peninsula during the latter part of May was disclosed in a series of messages between a Japanese headquarters in Korea and a new communications entity in southern Kyushu. Completion of the move was confirmed in early June. In mid-June the transfer to Kyushu of the 77th Division was disclosed in messages to a unit on Kyushu employing what was known to be that division's communications codename. This division had been previously identified at a location on the northern island of Hokkaido. Its movement would be completed by the end of the month.

Signals analysis disclosed the presence of a Japanese headquarters on Kyushu in mid-May; its identity as 57th Army headquarters was confirmed around the end of the month. This discovery stemmed from several weeks of intercepting and tracking message traffic sent by an unknown entity in southern Kyushu that was using an identification code indicative of Army-level headquarters and interacting with addressees at the echelon of Army headquarters or above.

By the end of the first week in June, intercepted message traffic had exposed the presence of a second new Army-level headquarters--this one in northern Kyushu. The US War Department's Military Intelligence Service suspected that this was the 56th Army headquarters, inasmuch as the Service had recently identified newly created Japanese armies numbered 55th, 57th, and 58th. It would be some time before this suspicion would be confirmed. The analysts also had presumed that their projected deployments of three Japanese divisions in northern Kyushu and three in the south would create a requirement for an Army headquarters in each area. They therefore viewed the discovery of these two headquarters as tending to confirm their long-held estimates.

A fifth division on Kyushu was confirmed in communications a few days after President Truman's meeting with his advisers on 18 June. Its specific identity was not determined for another two weeks, when a message disclosed its designator as the 206th Division. This was a newly created division. It also was one of the first identified as carrying a numerical designator in the "200" series, indicative of the "levy" on existing military units that the Japanese were undertaking for the creation of new divisions. Until then, the divisions discovered on Kyushu had been existing ones that had been transferred intact from previously known locations.

On 13 July, shortly after the identity of the 206th was confirmed, the presence of the 212th Division on Kyushu was disclosed. The addition of these two divisions and completion of the 77th Division's move from Hokkaido prompted the Military Intelligence Service to raise its manpower estimate for Kyushu to 375,000.

On 21 July the Military Intelligence Service's daily summary on Japanese military forces reported that three more divisions had suddenly been discovered on Kyushu. Within the next few days these were identified as the 146th, 154th, and 156th Divisions, and the presence of another new division, the 145th, was disclosed. This brought the total on Kyushu to ten combat divisions and two depot divisions. In addition, intercepted communications provided tenuous evidence that an eleventh combat division was being moved to Kyushu from Honshu.

According to the data compiled by the Military Intelligence Service, two of the identified Japanese combat divisions (the 57th and 145th) were based well to the north. One (the 206th) was carried at that time in north-central Kyushu, positioned for reinforcement of critical areas in the south. The other seven identified divisions and most of the independent brigades were in southern Kyushu. This meant that the number of combat divisions and equivalent forces in southern Kyushu was already more than double the number originally forecast for that part of the island, and in fact exceeded what had been forecast for all of Kyushu by 1 November 1945, the planned invasion date.

By the end of July intercepted communications had apparently indicated the presence of two more divisions, designated the 216th and 303rd. Analysts were uncertain as to whether this information might have reflected some misinterpretation of signals from divisions already detected, so they did not immediately add them to their estimate.

On 1 August intercepted messages unambiguously disclosed that the 312th Division was in the process of moving to a site in northwestern Kyushu. The identification of divisions in a "300" series also represented another layer of the Japanese "levy," and was seen by the analysts as yet another indication of more to come. By this time communications had also identified the 4th, 5th, and 6th Tank Brigades on Kyushu.

On 4 August, the same day that the Joint War Plans Committee circulated a memorandum citing the much-larger-than-expected buildup on Kyushu and recommending examination of alternative invasion sites, intercepted messages confirmed that another new division was moving onto Kyushu-- the 351st. The intercepts also revealed that the 206th Division was moving from its position in the central area to the southern part of the island.

During the next few days the analysts would confirm that the 216th and 303rd divisions were at least in the process of deployment on Kyushu. The 303rd appeared headed for southern Kyushu. This would bring the number of divisions there to nine---three times what had been forecast for that area by invasion day. These deployments resulted in a hike of the Kyushu manpower estimate to 549,000. A short time later, when the full manpower of the 351st, 303rd, and 216th divisions was believed to be in place, the estimate would be raised to nearly 600,000.

The following chronology is constructed mainly from Military Intelligence Service reports on Japanese force deployments and estimated troop strengths on Kyushu. With two exceptions, the figures are taken from NSA Special Reporting Series (SRS) daily reports and Special Research History (SRH) weekly summaries contained in Record Group 457, NARA.

One of the exceptions is the information from the initial JIC report of June 1944, which included the force projection that would remain the conventional view until late July 1945, when SIGINT proved it wrong. The other exception is the information from the 25 April 1945 report by MacArthur's intelligence staff, which is included for purposes of illustrating the extent to which the operational commands were constructing SIGINT-based estimates similar to--albeit with small differences from--the MIS in Washington.

The dates given for specific force estimates should be read with the understanding that the daily reports sometimes gave findings that were too close to the publication date of the weekly report to be incorporated into that week's edition and instead appeared in the following week's report. See, for example, the SRS Daily for 20 July and the SRH Weekly for 21 July.

Chronology of Japanese Buildup


Date*(Source)TotalArmy-GroundNavy-GroundAir-GroundKey Information Reported
24 Jun 1944
(JIC 191/1)
    86th Division, 2 Depot Divisions; Forecast: 6 combat, 2 depot divisions; could be further reinforced by up to 4 combat divisions.
12 Apr 1945
(SRS 388)
    Large numbers moving from mainland.
18 Apr 1945
(SRS 394)
    57th Division moving to Kyushu?
20 Apr 1945
(SRS 396)
    57th Division confirmed.
25 Apr 1945
(SWPA G-2 Rept)
228,250a112,75025,00090,500MacArthur's G-2 includes 86th, 57th Division.
28 Apr 1945
(SRH 195,#s 60, 62, 66)
230,000b(112,000)(25,000)(93,000)MIS includes only 86th Division--not yet 57th.
12 May 1945
(SRH 195,#s 62, 66)
246,000128,000(25,000)(93,000)MIS adds 57th Division (Differs from G-2 report of 25 April).
16 May 1945
(SRS 422)
    57th Army at Takanabe; later learned it was there as of 20 Apr.
18 May 1945
(SRS 424)
    3rd Amphibious Brigade confirmed.
25 May 1945
(SRS 431)
    25th Division confirmed. (Parts still in transit.)
6 Jun 1945
(SRH 195, #66)
281,000144,00045,00092,000MIS adds 25th Division, some naval-ground forces; reassesses air-ground forces.
7 Jun 1945
(SRS 444)
281,000   Evidence of Japanese Army at Izuka; later identified as 56th.
15 Jun 1945
(SRS 452)
    Suspect 77th Division from Hokkaido.
16 Jun 1945
(SRH 195, #67)
300,000160,00045,00095,000MIS adds miscellaneous brigades; air-ground again reassessed.
23 Jun 1945
(SRH 195, #68)
329,000181,00050,00098,000New unidentified Division confirmed; would later be identified as 206th; includes part of 77 th Division, some naval guards; air-ground again reassessed.
30 Jun 1945
(SRH 195, #69)
340,000190,00050,000100,000Resubordination of units from outer Ryukyus.c
9 Jul 1945
(SRS 476)
350,000200,00050,000100,000206th Division identified; MIS adds balance of 77th Division.
13 Jul 1945
(SRS 480; SRH 195, #72)
375,000225,00050,000100,000Increase includes confirmation of 212th Division.
20 Jul 1945
(SRS 487)
380,000   MIS adds 126th Independent Mixed Brigade; also includes manpower from 212th Division.
21 Jul 1945
(SRS 488)
455,000305,00050,000100,000MIS adds 3 unidentified divisions; will learn they are 146th, 154th, 156th, and have been there since May-Jun.
26 Jul 1945
(SRS 490,492,493)
525,000350,00075,000100,000Includes manpower from newly discovered 145th Division and 122th Independent Mixed Brigade.
27 Jul 1945
(SRS 494)
    40th Army Headquarters identified.
28 Jul 1945
(SRH 195, #73)
525,000350,00075,000100,000Weekly; includes 145th, 146th, 154th, 156th Divisions, 126th Independent Mixed Brigade, 56th Army Headquarters, miscellaneous.
30 Jul 1945
(SRS 497)
    Tentative evidence of 216th and 303rd Divisions; 4th, 5th, 6th Tank Regiments confirmed.
2 Aug 1945
(SRS 500)
545,000355,00090,000100,000MIS adds beginnings of 312th Division, more naval guards; still not 216th and 303rd.
4 Aug 1945
(SRS 502)
    351st Division confirmed; still not 216th or 303rd.
8 Aug 1945
(SRH 195, #75)
560,000370,00090,000100,000MIS begins adding parts of 216th, 203rd, along with elements of 351st.

Note: Boldface used in the table above signifies the first confirmation of a unit's existence on Kyushu, even if the unit's title/designator was still unknown at the time. Order-of-Battle dates are "as of" dates taken from maps that appear in the various reports. Often, therefore, they carry a date a few days earlier than the dissemination date of the bulletin itself.

a The reports from MacArthur's G-2 used categories of mobile combat, base support, naval-ground, and air-ground forces. The text and tables indicate that the combined total of mobile combat and base support equated to the category listed in the MIS reports as "Army-ground." The totals of the G-2 and the MIS for estimates prepared in April 1945 appear virtually the same. However, the G-2 figures incorporate the 57th Division, while the MIS did not incorporate this division until a few weeks later, and when it did so its totals were about 18,000 higher than the G-2 estimate. The difference appears to be attributable to different holdings for Army ground force support elements.

b Some of the earlier reporting on Kyushu did not give consistent detailed breakdowns. Thus, putting together the figures in this table required some reconstruction through comparing the figures from several reports. The figures in parentheses are the product of this reconstitution. The total manpower figure in the Daily Report for this date is given in Bulletin #60 (28 April) as 229,000, but without a breakdown. Bulletin #62 (12 May) says the addition of the 57th Division (16,000 men) increases the Army-ground force total to 128,000, and the overall manpower total from 230,000 to 246,000. Because the entire 16,000 increase went to Army-ground forces, the Army-ground number before the addition of the 57th Division would have been 112,000. Neither Bulletin #60 or 62 gives a breakdown other than for Army-ground. However, subtracting the Army-ground total from the new overall total leaves 118,000 for Navy- and air-ground. A later bulletin (#66 of 6 Jun) breaks this down to 25,000 and 93,000. Combining these with the Army-ground figures makes possible a reconstruction of the breakdowns of the Bulletin 60 and 62 figures, but produces a total for 28 April of 230,000 rather than 229,000.

c The higher figures given in this report of 30 June were not a result of increases in the forces on Kyushu, but rather from Japanese resubordination of 11,000 troops (9,000 Army ground and 2,000 air ground) located on the outer Ryukyu Islands to the 16th Area Army Headquarters, located on Kyushu. This was done by the Japanese after it was clear that Okinawa headquarters, to which these units had previously been subordinated, was going to fall to the US. For consistency when comparing the estimates on Kyushu, therefore, the 11,000 should be "netted out" of the MIS figures used in the daily and weekly reports disseminated after 30 June 1945.

Appendix C: Selected Archival Documents

Appendix C is composed of verbatim copies of selected US Government documents that have been declassified and served as key source materials in the preparation of this monograph. This cover note to the Appendix is intended as a summary and guide for readers of these documents.

The first two documents deal with the MacArthur-Nimitz rivalry over command responsibilities. This issue is discussed in section I of the monograph.

Document 1: "Directive for Operation OLYMPIC." JCS 1331/2, 14 May 1945. This is a report prepared by the Joint Staff and sent to the Joint Chiefs of Staff (JCS). It points to issues that still had to be settled in order for agreement to be reached on a directive for the final campaign in the Pacific. The document focuses in particular on the "who is to be in charge" question discussed in section I of the monograph. It and document 2 below are cited in footnote 9 of the monograph.

Document 2: "Directive for Operation OLYMPIC." JCS 1331/3, 25 May 1945. This is the last version of the directive for the final Pacific campaign, as dispatched to the Pacific commanders. It contains the decision on overall command of the operation. As the document indicates, a directive to prepare for the operation had already been issued to MacArthur and Nimitz on 3 April. This document is presented here, out of chronological order, because its contents are so closely related to those of document 1.

Document 3: "Japanese Reaction to An Operation Against Southern Kyushu." JIC 191/7, 16 May 1945. This was the eighth version of a report on anticipated Japanese reaction to an invasion of southern Kyushu. This version is included here because it was the latest iteration prepared prior to the dispatch of the directive discussed in documents 1 and 2 above. The document is discussed in section II of the monograph and is cited in footnote 11. (Footnote 10 cites one of the earlier versions, dated 24 June 1944.)

This document demonstrates the consistency of the "six combat divisions, two depot divisions" projection for Japanese units on Kyushu Island by 1 November 1945. That estimate, first made more than a year earlier, formed the basis for the figures that would be given to President Truman on 18 June 1945. This projection remained in place right up to the eve of the Potsdam conference in July. The projected Japanese manpower figure for 1 November in this document was 390,000, rather than the 350,000 figure that was used in most of the previous versions of the report. (MacArthur's staff, for its part, estimated the number at 300,000.) The differences did not relate to combat strength but rather to numbers of support forces and naval and air-ground troops.

Document 4: "Details of the Campaign Against Japan." JPS 697/D, 14 June 1945. This memorandum from Admiral Leahy set up the 18 June meeting with President Truman that would review plans for bringing the war with Japan to an end. The document is discussed in section III of the monograph and is cited in footnote 21. It was forwarded to the Joint Planning Staff (JPS), which directed the Joint War Plans Committee (JWPC) to draft a response. On the document's distribution list, four of the ten names are those of members of the Joint Planning Staff--two Army and two Navy officers. One of these Army representatives, Brig. Gen. Charles Cabell, was an Army Air Force (the Air Force was then part of the US Army) officer who would later go on to serve as Deputy Director of Central Intelligence. Two other members of the JPS--Rear Adm. B.H. Bieri and Brig. Gen. J.E. Hull, the Army's Chief of Plans--were de facto co-chairmen of the JWPC.

Document 5: Details of the Campaign Against Japan. JWPC 369/1, 15 June 1945. This is the JWPC's response to the JPS request for a draft reply to Admiral Leahy's memo of 14 June. It is discussed in section III of the monograph and is cited in footnote 24. Upon receiving this document, the JPS made some modifications and submitted it to the Chiefs as JCS 1388. The JPS revisions in document 5 included deletions that the author of the monograph has marked with brackets; the most noteworthy of these are marked on pages 7 and 9. This is an especially important document because it demonstrates the military planners' sensitivity about confronting the President over casualty estimates. It does give a total estimate of roughly "a quarter of a million," which was consistent with the casualty data used by the staffs of both MacArthur and Nimitz.

Document 6: Details of the Campaign Against Japan. JCS 1388, 16 June 1945. This is the JPS version mentioned above. It too is discussed in section III of the monograph, and it is cited in footnote 25. In addition to the deletions noted above, the "enclosure" attached at the end of this JCS paper is noteworthy. This enclosure, apparently prepared after the main draft was written, proposes further changes. One of these was language to replace the casualty estimate deleted from the JWPC version; this language offers the figures that General Hull had requested on 16 June for the 18 June meeting with the President. Hull's request presumably was prompted by the deletion of the JWPC figures and a belief on his part that, given the language of Admiral Leahy's memo, some figures had to be offered for the President.

Document 7: Memorandum for the Chief of Staff: Amplifying Details on Planners' Paper for Presentation to the President. Undated--presumably 17 or possibly 18 June 1945. Prepared by General Hull after he received the response to his request for casualty figures from various Pacific operations. (See section III of the monograph.) This is a summary of JCS 1388 (described in document 6 above), for use by General Marshall in briefing the President at the 18 June meeting.

Document 8: Minutes of Meeting Held at the White House on Monday 18 June 1945, at 1530. Document 598 of Foreign Relations of the United States (FRUS). Diplomatic Papers: The Conference of Berlin (Potsdam Conference), 1945. Vol. I (Washington, DC, GPO 1960, pp. 902-911). General Marshall read into the record the summary offered by General Hull (document 7 above). These minutes of the 18 June meeting with the President are discussed in section III of the monograph and in several of that section's footnotes. Tables from JCS 1388 were used at the meeting. Because some disagreement persisted over the specific language of JCS 1388, the document itself was not given to the President at that time. In fact, the agreed version was not completed until 11 July, when it was included in the background papers for the Potsdam Conference.

Document 9: Proposed Changes to Details of the Campaign Against Japan. JCS 1388/1, 20 June 1945. This document shows (a) that agreement still had not been reached on JCS 1388 by the time of the 18 June meeting with Truman (prompting Gen. Hull to prepare his summary for the President) and (b) that the casualty issue continued to be debated, with Nimitz's estimates remaining close to those offered by MacArthur's staff and by the JWPC. (See section III and footnote 33 of the monograph.)

Document 10: Proposed Changes to Details of the Campaign Against Japan. Memorandum For The Assistant Secretary, War Department General Staff, 25 June 1945. This document shows that the Chief of Naval Operations, Admiral King, apparently supported Admiral Nimitz's recommendations. It also shows, however, that the War Department did not accept these recommendations, and that the disagreement continued at least through late June.

Document 11: Map--Estimated Japanese Dispositions on Kyushu, 21 July 1945. From MAGIC Far East Summary of 21 July (the same day the Groves report on the successful atomic bomb test arrived in Potsdam). SRS 488.

Document 12: Cables to Potsdam Regarding the A-Bomb Test of 16 July. FRUS, Vol. II, Documents 1303, 1304, and 1305, pp. 1360-1369. The footnotes in these documents are particularly useful.

Document 13: Instruction to General Carl Spaatz on Use of the Atomic Bomb Against a Japanese Target. 25 July 1945. Discussed in monograph section VI and footnote 66. This document can be found in many sources. Spaatz reportedly carried it with him when he departed for the Far East on 24 July (Far East time] to take over a newly created Air Force command role. He was under instructions to deliver the document personally to MacArthur and Nimitz.

Document 14: Map--Estimated Distribution of Japanese Forces on Kyushu. 25 July 1945. From MAGIC Far East Summary on that date, SRS 492.

Document 15: Map--Estimated Disposition of Japanese Forces on Kyushu. 26 July 1945. Also attached is a table (Document 15A) showing estimated Japanese air strength. Both items were from the MAGIC Far East Summary of 26 July, SRS 493.

Document 16: Cable from Stimson to Truman, AGWAR Washington to Tripartite Conference, Babelsberg, Germany. 30 July 1945. This message underscored the need for President Truman's agreement on the language of a statement that would be released as soon as the atomic bomb was used. The message alerted Truman to a text being dispatched by courier that Stimson wanted the President to approve as quickly as possible.

Truman, however, either misunderstood the request or consciously used the occasion to give his guidance on the dropping of the bomb itself. His handwritten message on the back of the cable said "no sooner than August 2" (i.e., after he had left Potsdam). This note was typed and sent to Stimson immediately, before Truman received--later that same day--the couriered text of the public statement to which Stimson had referred in his cable. The dates of these messages have generated much misinterpretation. The records are in the Truman documents collection compiled by historian Dennis Merrill. (See monograph section VI and footnote 68.)

Document 17: Alternatives to OLYMPIC. JWPC 397, 4 August 1945. This subject is addressed in sections IV and VII of the monograph, and this document is cited in footnote 51. Some interesting comments were handwritten on the document by unidentified readers. One such reader wrote, "Sec'y told we non-concur." Two other handwritten notes that appeared to accompany this document reflect concern at the senior level of the military planning groups. One of these is addressed to General Lincoln, the senior Army representative on the Joint War Plans Committee. The other refers to the views of General Cabell, the Army Air Force representative on the Joint Planning Staff.

Documents: 18 and 18 A Through F: On the President's response to a query from Air Force Historian James Cate regarding Truman's role in the orders to use the atomic bomb. December 1952-January 1953. These documents are from the Truman records compiled by Dennis Merrill. They include Cate's letter of request and Truman's handwritten draft response--a personal note in which he said General Marshall had told him that the invasion would have cost "at a minimum a quarter of a million casualties." (Discussed in section V of the monograph and in footnotes 61 and 63.)


Footnotes

(1) Overviews of the evolution of strategic planning from late 1942 to early 1945 are in John Ray Skates, The Invasion of Japan (Columbia, SC: University of South Carolina Press, 1994), chapters 4 and 11; and Thomas B. Allen and Norman Polmar, Codename Downfall: The Secret Plan to Invade Japan (New York: Simon and Schuster, 1995), chapter 6.

(2) The Joint Planning Staff (JPS) was composed of two senior representatives each from the Army and Navy planning organizations. It reported directly to the Joint Chiefs. The JPS in turn was served by the Joint War Plans Committee (JWPC), which--with assistance from such specialized support groups as the Joint Intelligence Committee and the Joint Logistics Committee--produced basic plans that were reviewed and amended as deemed necessary by the JPS and forwarded to the Chiefs. See Ray S. Cline, Washington Command Post: The Operations Division (Washington DC, GPO, 1951), pp. 103-104.

(3) The background descriptions herein of the planning lineage from mid-1944 to April 1945 are mainly from Grace Pearson Hayes, The History of the Joint Chiefs of Staff in World War II: The War Against Japan (Annapolis: Naval Institute Press, 1982), pp. 627-630, 651-712--especially pp. 627-630, 655-661, and 686-707, which draw heavily on the minutes of the JPS meetings that produced the various documents and operational directives described in Hayes' book. The specific report referred to in this paragraph is "Operations Against Japan...," Report by JPS, JCS 924, 7 July 1944. It was the first of 16 iterations incorporating various modifications produced over the ensuing nine months. Copies of most of the 16, including the final versions in April 1945 that articulated the JCS consensus on an invasion of Kyushu (see footnote 9), are in RG 165, ABC 384, Pacific (1-17-43) Entry 421, Box 457, National Archives and Records Administration (NARA).

(4) Ibid, pp. 701-702.

(5) See Barton Bernstein, "Understanding the Atomic Bomb and the Japanese Surrender," Diplomatic History, Vol. 19, No. 2, Spring 1995, pp. 230-231.

(6) Hayes, History of the Joint Chiefs of Staff, pp. 656-657.

(7) Command and Operational Directives for the Pacific, JCS 1259/4 and 1259/5, 3 April 1945, RG 165, ABC 381 Pacific Ocean Area (1-29-43, Sec. 3), NARA.

(8) Report by JPS, "Pacific Strategy," JCS 924/15, 25 April 1945, RG 165, ABC 384 Pacific (1-17-43) Entry 421, Box 457, NARA.

(9) The MacArthur-Nimitz recommendation regarding the date of a Kyushu invasion is JCS 1331/1, 30 April 1945. The Joint Planning Staff recommendation for resolving the command issue is in JCS 1331/2 of 14 May. JCS 1331/3, "Directive for Operation 'OLYMPIC,' " 25 May 1945, assigns the command to MacArthur and sets the 1 November date. RG 165, ABC 384, Kyushu (4 July 1944) Sec. 1-B, Entry 421, Box 434, NARA. Also see Appendix C of this monograph.

(10) JIC 191/1, "Japanese Reaction to an Operation against Southern Kyushu," 24 June 1944, RG 165, ABC 384 Kyushu (4 July 1944) Sec. 1-B, Entry 421, Box 434.

(11) Ibid. The same projection is found in Annex D of JWPC 235/3, "Plan for Operations Against Kyushu," 4 September 1944, RG 165, ABC 384 Kyushu (4 July 1944) Sec. 1-C; and in JWPC 235/4, 16 November 1944, JIC 284, 30 April 1945, and JIS 191/7, 16 May 1945--all in RG 165, ABC 384 Kyushu (4 July 1944) Sec. 1-B. It is also presented in a report by the intelligence staff under MacArthur's command in the Pacific: GHQ, USAFPAC, MIS, GS, "G2 Estimate of the Enemy Situation With Respect to an Operation Against Southern Kyushu in November 1945," 25 April 1945, RG 165, Entry 418, Box 1842, NARA.

(12) In the parlance of signals intelligence, codes and ciphers are different things. At the risk of oversimplification, a statement that "the dove has landed," utilized to indicate that an action has been carried out, could be considered a code. If this phrase is then transcribed into numbers representing the letters, and the numbers are scrambled according to a system that requires a key to unscramble, it can be defined as enciphered. Exploiting the substantive contents of intercepted Army messages usually required both deciphering and decoding. The contents were then provided to military force analysts to interpret their meaning relative to force deployments, strengths, and activities. A detailed discussion can be found in chapter 1 of Edward J. Drea, MacArthur's ULTRA: Code Breaking and the War Against Japan, 1942 -1945 (Lawrence, KS: University Press of Kansas, 1992).

(13) The communications intelligence reports on Japanese military forces cited throughout this paper are located in Record Group 457, "Record of the National Security Agency," NARA: Entry 9001--MAGIC Far East Summaries": Box 5, folder 14 (SRS 381-410, 5 April through 5 May 1945); Box 6, folder 15 (SRS 411-444, 6 May through 7 June 1945), and folder 16 (SRS 445-490), 8 June through 23 July 1945); and Box 7, folder 17 (SRS 491- 509, 24 July through 11 August 1945). Also, Entry 9002--"War Department (MID) Military Intelligence Service, Japanese Ground Order of Battle Bulletins," SRH 195, Part I (7 April through 2 June 1945), Box 76, and Part II (9 June through 11 August 1945), Box 77.

(14) SRS Nos. 383, 385, and 393 of 7, 9, and 17 April 1945.

(15) SRS Nos. 388, 394, and 396 of 12, 18, and 20 April 1945; SRH 195, Bulletin Nos. 59 and 60, 21 and 28 April 1945. The unit from Manchuria was the 57th Division.

(16) Idem, USAFPAC G-2, 25 April 1945.

(17) SRS 431, 25 May, and SRS 452, 15 June 1945; SRH 195, Bulletin Nos. 66, 68, and 71, 6 and 23 June and 11 July 1945. The division from Korea was the 25th Division and the one from Hokkaido was the 77th.

(18) SRS 409, 410, and 422 of 3, 4, and 16 May 1945; SRS 444 of 7 June 1945. The headquarters in southern Kyushu was identified as the "57th Army." It would be several weeks before the identity of the one in the north would be confirmed as the "56th Army," but analysts assumed this was its designator on the grounds that they had already identified newly created armies in various locations carrying the designators 55th, 57th, and 58th.

(19) SRS 425, 19 May 1945; SRS 454 and 459, 17 and 22 June 1945.

(20) SRS 424 and 437, 18 and 31 May 1945; SRH 195, Bulletin No. 66, 9 June 1945. The unit from the Kurils was the 3rd Amphibious Brigade. After the war it was learned that this unit had been reconstituted into the 125th Independent Mixed Brigade, a regular ground combat unit, but because it continued to use the call number of the amphibious brigade it was carried as such in the MIS force assessments.

(21) "Memorandum for the Joint Chiefs of Staff" from Admiral Leahy, enclosure to JPS 697/D, 14 June 1945, "Details of the Campaign Against Japan," RG 165, ABC 384 Japan (3 May 1944), Sec. 1-B, Entry 421, Box 428, NARA.

(22) Ibid.

(23) "Memorandum for the Record," Telephone Conversation, 14 June 1945, Admiral Duncan and General Lincoln, ABC 384 Japan (3 May 1944) Sec. 1-B, Entry 421, Box 428, NARA.

(24) JWPC 369/1, 15 June 1945, "Details of the Campaign Against Japan," RG 165, ABC 384 Japan (3 May 1944) Entry 421, Box 428.

(25) JCS 1388, 16 June 1945, "Details of the Campaign Against Japan," RG 165, ABC 384 Japan (3 May 1944) Sec.1-B, Entry 421, Box 428, NARA. For deletions, compare JWPC 369/1, Section 3, 2nd paragraph and Section 7, 3rd paragraph, with the corresponding paragraphs in this document.

(26) Memorandum for the Chief of Staff: "Amplifying Comments on Planners' Paper for Presentation to the President," fr. J.E. H[ull], RG 165, ABC 384 Japan (3 May 1944), Sec. 1-B, Entry 421, Box 428, NARA. The 11 July version (JCS 1388/3) is in the same box. For Hull's request see Allen/Polmar, p. 203, citing Hull-Lincoln correspondence: Memo, 16 June 1945, Marshall Library, Lexington, VA.

(27) See Skates, The Invasion of Japan, pp. 80-81. Also Allen/Polmar, Codename Downfall, p.204, quoting "To GHA AFPAC (MacArthur) from Marshall, 16 June 1945," WD1050, and "From CINC AFPAC", 17 June 1945, WD1052, both in RG#4, Folder 4 (War Department 1000-1095) 29 April-2 August 1945, MacArthur Memorial Bureau of Archives, Norfolk, VA.

(28) Ibid., citing 19 June 1945, WD1056. The archival copy of this cable and the one cited below are dated according to Manila time. For comments on motives see Drea, MacArthur's ULTRA, p.210; Allen/Polmar, Codename Downfall, p. 204; and Skates, The Invasion of Japan, p. 81.

(29) Idem, 19 June 1945 WD1057, MMBA RG 4.

(30) A copy of the official minutes of this meeting is contained in Department of State, Foreign Relations of the United States, Diplomatic Papers: The Conference of Berlin (The Potsdam Conference) 1945, Vol. 1 (Washington, DC: GPO, 1960), pp. 902-911. Hereafter referred to as FRUS.

(31) This percentage would increase to 39 with the updated figures that appeared in the 11 July report.

(32) According to Allen/Polmar, Codename Downfall, p.145, citing "Commander in Chief, Pacific Fleet, Operation Plan No. 10-45, Central Pacific Area: OLYMPIC, 8 August 1945," Naval Historical Center, the portion of the total forces to be landed in the invasion were in 14 divisions with at most some 350,000 troops. Skates, The Invasion of Japan, p.71, gives a similar figure ("353,000 soldiers and marines"), citing Sixth Army Field Order #74, Troop List, 28 July 1945, Records of the Strategic Plans Division, Box 187, also at the Naval Historical Center.

(33) These were the figures given in Tabs D and G of the draft of JCS 1388 that had been prepared specifically for the President's meeting. At the time of the meeting, however, the Joint Chiefs had not reached agreement on the wording of the text of JCS 1388, and thus the paper itself was not given to the President; only the digest prepared by Hull was presented. See "Memorandum for the Assistant Secretary, WDGS," 25 June 1945, from G.A.L[incoln], RG 165, ABC 384 Japan (3 May 1944), Sec. 1-B, NARA.

(34) FRUS, Vol. I, Minutes, p. 904 and 907. This statement by Marshall, recorded in the minutes, was read verbatim from the memorandum prepared by Hull.

(35) Some authors have cited as further indication of Truman's concern a comment he made earlier in the meeting that Kyushu would "create another Okinawa closer to Japan." In the context of the discussion at the point at which he inserted this comment, however, the issue was the value of Kyushu as a base for air attacks on the Japanese homeland. This comment by Truman, therefore, can just as plausibly be interpreted as endorsing the value of seizing Kyushu, as opposed to bemoaning the cost in casualties. In any event, any reading of the minutes will indicate there is no need to stretch for indications of Truman's concern over casualties. See FRUS, Vol. I, Minutes... p. 908.

(36) FRUS, Vol. I, p. 910. The minutes state that the President and his advisers then discussed "certain other matters," which the records compilers speculated, and which others have asserted, meant the issue of whether the Japanese should be given a specific warning about the atomic bomb. See also page 889ff of FRUS, Vol. I.; John McCloy, Challenge to American Foreign Policy, (Cambridge: Harvard University Press, 1953), pp.42-43; and Walter Millis, ed., The Forrestal Diaries (New York: The Viking Press, 1951), pp. 70-71. For a contesting view see Bernstein, "Understanding the Atomic Bomb," p. 237.

(37) SRS 476 and 480, 9 and 13 July 1945; SRH 195, Bulletin Nos. 68 and 71, 23 June and 14 July 1945. The two additions were the 206th and 212th Divisions. Unlike the divisions previously discovered on Kyushu, these--and all subsequently discovered divisions on the island--were newly created from a combined process of conscription, consolidation of smaller units, and utilization of "depot division" resources.

(38) "Kaiten...", SRS 463 and 465, 26 and 28 June 1945. "940 aircraft...", SRS 467, 30 June 1945. "Floatplanes...", SRS 483 and 485, 16 and 18 July. "Night tactics...", SRS 483, 16 July 1945. "Naval air focus...", SRS 486, 20 July 1945.

(39) SRS 480, 13 July 1945; SRH 195, Bulletin Nos. 71 and 72 of 14 and 21 July 1945.

(40) This discovery resulted from learning the meaning of a special communications term for 'division'... SRH 195, Bulletin No. 73, 28 July 1945.

(41) SRS 488, 492, and 493 of 21, 25, and 26 July 1945; SRH 195, Bulletin No. 73, 28 July 1945. The four confirmed new divisions were, in rough order of discovery, the 146th, 154th, 156th, and 145th.

(42) SRS 494, 20 July 1945; SRH 195, Bulletin No.73, 28 July 1945.

(43) SRS 500, 2 August 1945. Confirmation of the presence of the 312th Division raised the total to eleven. Other newly detected units included the 4th, 5th, and 6th Tank Brigades. By this time, messages also indicated the presence of the 216th and 303rd Divisions, but analysts initially left open the possibility that these were misinterpretations of signals from divisions already detected. See SRS 497, 30 July 1945.

(44) SRS 487, 490, and 494, of 20, 23 and 27 July 1945. Also SRH 195, Bulletins 72 and 73 of 21 and 28 July 1945.

(45) SRS 500, 2 August 1945. This total excludes personnel subordinate to the Area Army Headquarters but not actually located on Kyushu. See note with Table 6.

(46) MacArthur's G-2 carried one division, the 145th, as being in transit; thus it cited six divisions, bringing the total to nine, whereas on this date the MIS would have said--correctly as postwar information confirmed--seven new divisions, bringing the total to ten. See also Drea, MacArthur's ULTRA, p. 220.

(47) General Headquarters, US Armed Forces Pacific, Military Intelligence Summary, General Staff, "Amendment No. 1 to G-2 Estimate of the Enemy Situation with Respect to Kyushu (dated 25 April 1945), 29 July 1945. The paper is cited in Skates, Invasion of Japan,p.141, noting the source's location as the US Army Military History Institute at Carlisle, PA. It is also quoted in Allen/Polmar, Codename Downfall, p. 223, and p. 238, which identifies the location as the National Archives. Drea, MacArthur's ULTRA, p. 216, cites the MacArthur Memorial Bureau of Archives as his source.

(48) SRS 500, 2 August 1945.

(49) Sixth Army G-2 Estimate of the Enemy Situation; OLYMPIC Operation, RG 165, Records of War Department, General and Special Staffs, Entry 418, Box 1843, NARA.

(50) "Memo to Chief, S&P Group. Subject: 'Amendment No. 1 to G-2 Estimate OLYMPIC,'" RG 165, ABC 384 Kyushu (4 July 1944), Sec. 1-F, Entry 421, Box 435, NARA.

(51) JWPC 397, "Alternates to 'OLYMPIC,'" 4 August 1945, RG 165, ABC 384 Kyushu (4 July 1944) Sec. 1-B, Entry 421, Box 434, NARA. This memo's Enclosure B incorporates the data from the 2 August figures in SRS 500.

(52) SRS 502, 4 August 1945; SRH 195 Bulletin 75, 8 August 1945. The confirmed new arrival was the 351st Division. The earlier suspected and now confirmed divisions were the 216th and the 303rd. The 303rd was headed to southern Kyushu; the 216th went to central Kyushu, replacing the 206th, which also moved south, bringing the number there to nine.

(53) Stimson's and Truman's revelations on their receipt of intercepted Japanese Foreign Ministry cables at Potsdam are described in FRUS, Vol.I, p.873. See also Bernstein, "Understanding the Atomic Bomb," pp. 242-43, footnote 45 on Marshall's access to this channel at Potsdam. On the Map Room, see George M. Elsey, "Some White House Recollections," Diplomatic History, Vol. 12, (Summer 1988), pp. 357-364; Leahy, I Was There, p.386, and McCullough, Truman, p.448.

(54) FRUS, Vol.II, p. 346. The Military Intelligence Service manpower estimate for Kyushu (including the outer Ryukyus) for 21 July 1945 was 455,000; four days later it was 525,000. See SRS 488 and 492, 21 and 25 July 1945.

(55) The search included FRUS; Truman, Vol. I: Year of Decision (Garden City, New York: Doubleday, 1956); Leahy, I Was There (New York: Whittlesey House, 1950); Stimson and Bundy, On Active Service in Peace and War (New York: Harper and Bros., 1947); Byrnes, Speaking Frankly (New York: Harper and Bros., 1947); and Dennis Merrill (ed.), Documentary History of the Truman Presidency, Vol. I, "Decision to Drop the Bomb on Japan." University Publications of America, 1995.

(56) Leahy, I Was There, p. 395.

(57) In the early World War II period, the US used the term MAGIC to mark cryptanalysis of high-level Japanese communications systems. The British used the term ULTRA to mark cryptanalysis of high-level German systems. As US-British cooperation increased during the war, the US also marked high-level Japanese systems as ULTRA. The US, however, continued to issue a daily summary of Japanese decrypts under the heading of MAGIC.

(58) Drea, MacArthur's ULTRA, p.xii.

(59) Christopher Andrew, For The President's Eyes Only (New York: Harper Perennial, 1996), p.151)

(60) FRUS, Vol. II, p.1266--footnote 4 has an excerpt from Stimson's Diary for 16 July; Millis, Forrestal Diaries: pp.74-76 describes his reading of the Japanese messages. Truman's comments are described in FRUS, Vol. I, p. 873.

(61) The Truman letter is reproduced in Wesley Frank Craven and James L. Cate, editors, The Air Forces in World War II, Vol. 5, "The Pacific: Matterhorn to Nagasaki," June 1944 to August 1945, (Chicago, 1955), between pp. 712-713. The query was from Cate, and a copy of it can be found in Merrill, Documentary History of Truman Presidency, p. 511.

(62) See for example Drea, MacArthur's ULTRA, p. 222.

(63) Merrill, ed., Documentary History of the Truman Presidency. Vol. I, pp. 510-526 contains copies of the paper trail of this drafting process, beginning with a copy of the request from Air Force historian James Cate, through the various memos from staff assistants suggesting changes, to the final draft as it appeared in the 1953 publication by Cate. The Stimson article appears in the February 1947 issue of Harpers magazine, and is reproduced almost entirely in Stimson/Bundy, On Active Service, Chapter XXIII, pp. 612-633.

(64) The Pacific campaign figure and the Normandy figures are from Skates, Invasion of Japan, p. 78-79. The Battle of the Bulge figures are from Allen/Polmar, Codename Downfall, p.134.

(65) See FRUS, Vol.II, pp. 1360-1370 for copies of the initial reporting cables, the Groves report, and notes from Stimson's diaries regarding his delivery of these messages to the President at Potsdam.

(66) A copy of Spaatz's instructions can be found in Stanley Weintraub, The Last Great Victory: The End of World War II (NY, 1995), p.192, and Gar Alperovitz, The Decision to Use the Bomb and the Architecture of an American Myth, (New York: Alfred A. Knopf, 1995), p.544. Spaatz received the directive while in Washington on the eve of moving from his position as General Arnold's deputy to take command of air forces in the Pacific. The directive was given to him by Gen. Thomas Handy, Acting Army Chief of Staff while Marshall was in Potsdam. Spaatz carried it with him to Guam on 24 July (the 25th in the Far East),with instructions to pass one copy each to MacArthur and Nimitz. Thus archival listings vary, from the copy addressed "Thomas T. Handy to Carl Spaatz," 25 July 1945, Box 21, Spaatz Papers, Library of Congress (a copy of which is also attached to James L. Cate to HST, 6 Dec. 1952, Box 112, Truman Library) versus Spaatz's transmission to MacArthur and Nimitz, cited by Drea in MacArthur's ULTRA, chapter 8, footnote 27, as COMGENUSASTAF, Eyes Only to MacArthur and Nimitz, 25 July 1945, RG 30, Box 7, Folder 7, MMBA.

(67) Arnold, Global Mission, p.589.

(68) Copies of the Stimson cable and Truman's handwritten note on the back are on pp. 174-75 of Merrill, Documentary History... Vol. I. Stimson's cable was sent late on 30 July and was handed to Truman at about 7:30 a.m. on the 31st. A copy of the cable carrying Truman's reply can be found in the National Archives. ("President to Secretary of War," 31 July 1945, File 74, Roll 6, Harrison-Bundy Files.) Stimson's cable said he was sending a copy of the new text by courier; it arrived later on the 31st. See Merrill, Documentary History..., Vol. I, p. 181-193. The 31 July date on both Truman's response and the courier package can be interpreted to mean that Truman was responding to the couriered text, (e.g., Alperovitz, Decision to Drop the Bomb, p.596), but in fact he gave his okay before the text arrived by courier. The text of the statement ultimately released on 6 August is in FRUS, Vol.II, p.1376.

(69) This statement is based on reading the documents, memoirs, and scholarly presentations listed in Appendix A. A suggestion of late input from developments at Potsdam might be inferred from Alperovitz, Decision to Drop the Bomb (pp. 557-9), which raises questions about the fact that the Truman documentary records show a gap in the President's Potsdam diaries for the dates 21-24 July and 27-29 July 1945. Alperovitz's interpretations as to what might have been withheld from disclosure, however, bear more on motives vis-a-vis the USSR.

(70) Merrill, ed. Documentary History of the Truman Presidency, Vol. I, pp.19-48, 94-101, and 106-110. Includes copies of the minutes of the Interim Committee meetings from mid-May through 6 July. See esp. p. 46, fourth paragraph.

(71) Ibid, pp. 534-540 contains a chronology prepared in 1957 by a Truman archivist, laying out the consultative path to the use of the bomb. It provides many page-specific citations from official documents and memoirs and is a useful research guide into the discussions preceding the Potsdam Conference.

(72) FRUS, pp. 1360-1378 show the exchange of cables between Potsdam and Washington from 16 to 30 July. Also included are several excerpts from Stimson's diary describing the discussions in Potsdam coincident with the cable exchange. Pp. 1265-1289 provide similar detail on the evolution of the final wording of the "Potsdam Declaration." Discussions pertaining to the USSR are described on p.1324, and in Leahy, I Was There, p. 419.

(73) McCullough, Truman, p.442. His comments can be compared to the record of the meeting. FRUS, Vol II, p. 339ff.

(74) Allen/Polmar, Codename Downfall, p. 266-267. In their chapter notes, however, they also point to discrepancies regarding the time and attendance of this alleged meeting.

(75) FRUS, Vol. II, p. 243 gives the date and a list of attendees, but cites as a source Truman's memoirs. See also the letter to James Cate, cited in footnotes 61 and 63. Sources for the Truman-Churchill meeting include FRUS; Truman, Memoirs, Vol I, p.419; and Churchill, The Second World War, Vol. VI, Triumph and Tragedy (London, Cassell and Co.,1954), p. 668. Records of Truman's own schedule, including the meeting with Churchill, and the accounts and schedules of others for that day (e.g., Stimson and Arnold) given in FRUS, Vol II, p. 1373, and of Byrnes' attendance at a Foreign Ministers' meeting at 11:10, FRUS Vol. II, p. 226, seem to rule out a larger meeting with all his advisers, unless it was very brief. No person Truman listed as being at the meeting has described it. Stanley Weintraub, who has researched the day-to-day schedule meticulously, flatly says the meeting did not occur (The Last Great Victory, p. 175). It may be that Truman conflated recollections of separate conversations. That does not mean he did not ask what he said he asked, but rather that if he did, the setting in which this occurred is different from what his memo implied.

(76) FRUS, Vol. I, pp. 909-910.

(77) Stimson's letter and draft statement are in FRUS, Vol. I, pp. 889-894.

(78) Idem JWPC 397, Appendix to Enclosure A.

(79) 6 August (Tokyo time) is usually identified historically as the date for the dropping of the first A-bomb (on Hiroshima). In Washington time, the event occurred on 5 August.

(80) Skates, The Invasion of Japan, p.142.

(81) Allen/Polmar, Codename Downfall, p.97.

(82) To avoid exaggerating the Okinawa-Kyushu comparisons, Air Force ground personnel are excluded. Some 100,000 were estimated to be on Kyushu.

(83) Allen/Polmar, Codename Downfall, p.98 gives total US troops in the initial land assault on Okinawa as 100,000, and about 25,000 were landed as reinforcements. As noted in footnote 32, the US total for Kyushu was 350,000.

(84) Leahy... FRUS, Vol. I, p. 909. Stimson... FRUS, Vol. I, p. 889.

(85) Message, WAR 45369, Marshall to MacArthur, 7 August 1945. Described in Skates, The Invasion of Japan, p.142; Allen/Polmar, Codename Downfall, p.223; and Drea, MacArthur's ULTRA, p. 222. Skates gives the location as the Nimitz Command Summary, Naval Historical Center, Washington, DC. Allen and Polmar list the MacArthur Memorial Bureau of Archives, WD1104. (The date is for Far East time, apparently because the available copies are from files there.)

(86) See Skates, The Invasion of Japan, p. 142-3; Allen/Polmar, Codename Downfall, pp.223-4; and Drea, MacArthur's ULTRA, p. 222-3--all citing Message C31897, MacArthur to Marshall: CINCAFPAC to WARCOS, 9 August 1945. Naval Historical Center and MacArthur Bureau of Memorial Archives. WD1106.

(87) See Drea, MacArthur's ULTRA, p. 223 and Allen/Polmar, Codename Downfall, p. 224.

(88) Drea, MacArthur's ULTRA, p. 222. Drea cites as his sources Boeicho, Boi kenshujo, [Japanese National Institute for Defense Studies], ed., Senshi sosho, Vol. 57: Hondo kessen jumbi: Kyushu no boei [Official Military History, Vol, 57, Preparations for the Decisive Battle of the Homeland: Defense of Kyushu] (Tokyo, 1972); and Hayashi Saburo, Taiheiyo senso rikusen gaishi [An Overview of Army Operations in the Pacific War], Tokyo, 1951).

(89) See for example FRUS, Vol. II, p. 1268-69 for JCS views on this point.

(90) FRUS, Vol. I, p. 909

(91) Asada, Sadao, "The Shock of the Atomic Bomb and Japan's Decision to Surrender." Manuscript, 1997; Bernstein, "Understanding the Atomic Bomb," p. 240.

(92) The USSR declared war against Japan on 8 August 1945.


Publications Page | CSI Homepage