by D. M. Giangreco

Journal of Military History, 61 (July 1997): 521-82

The Journal of Military History is published quarterly for the Society for Military
History by the George C. Marshall Foundation and the Virginia Military Institute.
Giangreco was awarded the Society for Military Historyís 1998 Moncado Prize
for his article.


Copyright 1997 by D. M. Giangreco and the Journal of Military History

This article may be reproduced without additional
permission for nonprofit educational use.

Please cite it as follows:

D. M. Giangreco, "Casualty Projections for the U.S. Invasions of Japan, 1945-1946: Planning and Policy Implications" in Journal of Military History, 61 (July 1997): 521-82, online at



PAGE NUMBERS of the JMH text are contained between greater- and less-than symbols, example: <page 548>.

RED text in Triple brackets denote NEW ANNOTATIONS by the AUTHOR, example: [[[ Stimson later revised the figure upward ]]].

END NOTE numbers in the body copy are preceded by a circumflex, example: ^126.

<page 554> D. M. Giangreco is an editor for the U.S. Armyís professional journal, Military Review, published by the Command and General Staff College at Fort Leavenworth, Kansas. He has authored five books and written on such topics as the Falkland Islandsí sovereignty question, decentralization of the Soviet Air Force command and control structure, and Persian Gulf pipeline construction to circumvent the Strait of Hormuz bottleneck for various Capital Cities Communications publications.

<page 521>



Casualty Projections for the U.S.
Invasions of Japan, 1945-1946:
Planning and Policy Implications

by D. M. Giangreco

IN RECENT YEARS, historians looking into the reasons behind the decision to drop the atomic bombs have been hampered by a lack of understanding of how the casualty projections given to President Harry Truman by the U.S. Army were formed, or even that specific methodologies existed for their creation. Complicating the situation even further for modern researchers is the fact that campaign, medical, and logistics planners used a form of verbal shorthand in their communications with colleagues, who had a common understanding of those methodologies and shared similar data on the relative strengths and weaknesses of the opposing U.S. and Japanese forces. Consequently, World War II planning documents frequently have been misinterpreted and, together with a lack of research below the "top layer" of documents, this has led many historians to the conclusion that President Harry Truman's assertion that he expected huge losses during an invasion^! 1 was fraudulent, and his claim that "General Marshall told me that it might cost half a million American lives to force the enemy's surrender on his home grounds"^2 <page 522> was a "postwar creation" to justify dropping the bombs on a "civilian target" and hide more sinister, calculating reasons for their use. For example, Kai Bird stated in the New York Times: "No scholar of the war has ever found archival evidence to substantiate claims that Truman expected anything close to a million casualties, or even that such large numbers were conceivable." ^3

While misconceptions of how casualty projections are formulated are addressed directly at several points in the text and footnotes, the four things I wish to do--- other than establishing the existence and complete acceptance by the War Department and Army of estimates that battle casualties could surpass one million men--- are: (1) explain how highly tentative estimates were, and still are, created; (2) outline the deep historical roots of casualty projections in the U.S. Army (which form the underpinnings of how planners actually look at and interpret the analyses); (3) describe the impact of the non-battle casualties on the formulations; and (4) explain why non-battle casualties specifically go unstated in strategic-level studies even though they are understood by all involved to be a fundamental part of manpower requirements--- the other side of the casualty projections coin.

From its inception during the Revolutionary War, the U.S. Army had made efforts to estimate probable losses because its leaders had to know approximately how many men would still be fit for duty by the last battle of the campaign season (i.e., before winter) after casualties from accidents, planned and unforeseen clashes, the erratic flow of recruitment, and disease drained it of men. Disease, in fact, felled more soldiers than musket fire during the Revolution, and this situation remained essentially unchanged until World War I. ^4 While serving as the commander of America's young army, George Washington worked unceasingly to strengthen and expand its fledgling medical corps, and pointedly informed the Continental Congress that lack of proper planning for medical facilities and personnel had contributed to the drastic reduction in the size of his forces during the winter of 1776-77, stating that: "the <page 523> dread of unde! rgoing the same Miseries of want of proper care and attention when Sick, has much retarded the new inlistments." ^5

"In a letter from his winter headquarters at Morristown, New Jersey, Washington tried to impress upon Congress the gravity of the situation:

We are now, at an enormous Bounty, and with no small difficulty, recruiting an Army of upwards of one hundred Battalions, [but] the ensuing Campaign may, from the same Causes, prove as sickly as the last. If the Hospitals are in no better condition for the reception of the Sick, our Regiments will be reduced to Companies by the end of the Campaign. . . . I leave you to judge whether we have Men enough to allow such a Consumption of Lives." ^6

After visiting an Army hospital near Philadelphia, John Adams wrote:

"I have spent an hour this morning in the congregation of the dead . . . and was never in my life affected with so much melancholy. The graves of the soldiers who have been buried in this ground from the hospital and bettering house during the course of last summer, fall and winter, dead of the small pox and camp diseases, are enough to make the stone melt away. The sexton told me that upwards of two thousand soldiers have been buried there, and by the appearances of the graves and trenches, it is most probable to me, he speaks within bounds. . . . Disease has destroyed ten men for us, where the sword of the enemy has killed one." ^7

Long literary and field experience had given the contending American, British, and French generals a fairly clear idea of how high their losses might climb whether they were the winners or losers of a particular battle or campaign. Likewise, their small medical and planning staffs had an evolving body of French and British writings that could be used as the basis for making casualty projections for logistics purposes. Although warfare had changed little in the previous hundred years, experienced army surgeons from both these armies had begun to produce treatises on the care of soldiers and management of military hospitals with increasing regularity.^8 A central feature of these writings was the emphasis on early planning for both the types and quantities of casualties anticipated as a percentage of the total body of troops committed.

<page 524> There was a marked increase in the number of such studies after the Seven Years' War, and American surgeons, many of whom could read French, had access to numerous useful works from the Bourbon army. For example, in Army Surgery: A Study of Firearms Injuries, Hugues Ravaton concluded that, at the beginning of an average European campaign, a commander could expect approximately 3 percent of his soldiers to be unfit for combat because of some form of illness or non-battle injury. This number would climb to roughly 5 or 6 percent by the middle of the campaign, and double by the time winter brought the campaign season to an end. Concurrent with this steady attrition, of course, would be battle casualties. The dead did not enter into logistics considerations because they used up no supplies and, after burial, required no additional care, but experience demonstrated that the wounded would amount to approximately 10 percent of the force actually engaging ! in each battle, and that percentage would generally drop as the size of the number of the committed troops approached 100,000.^9 However, unforeseen catastrophes, bad generalship or mauvaise fortune (bad luck) could quickly make a farce of even the most thoughtful estimates.

In spite of the availability of such materials, however, lack of adequate funds, ongoing shortages, and organizational problems tended to force a more reactive approach on the Army's fledgling medical corps, with much of the effort centering on establishing adequate hospital facilities for the existing sick and injured, and bureaucratic fights over the proper organization of medical care. Regimental surgeons made up the bulk of such personnel in the Army, but above that, the law of 27 July 1775 capped the number of hospital surgeons that Congress was willing to pay for at five, in addition to a "chief physician," and twenty surgeon's mates. The law of 17 July 1776 allowed this number to fluctuate with the size of the Army with one surgeon per 5,000 men and surgeon's mates at one per 1,000 men.^10 Congress had also established in the 1775 law that one nurse could be employed for every ten sick. This was the most that could be accompl! ished at that time, and no-one-and-everyone was responsible for an analysis of future combat needs, although legislation passed during the War of 1812 would formalize the requirement that "estimates" be made by the Army's physician and "Surgeon General" as well as the "Apothecary General." ^11 The periodic surge of casualties in both conflicts was handled by existing personnel and whatever local doctors were willing to lend a hand.

<page 525> Generals and their campaign staffs would also make their own estimates that added categories not included by medical men planning for logistical needs, such as dead and missing, soldiers taken prisoner, and virtually anything that would remove soldiers from the battle line. Consequently, the totals they arrived at were always higher. Writing to his brother shortly after the 1862 Battle of Shiloh, William Tecumseh Sherman marveled at the commonly held belief that the Civil War would soon be over and wrote: "The people should know that this war will consume 300,000 men per year [North and South] for a long time." ^12 Two years later, in April 1864, he lamented that, although the South was running out of manpower:

"Full 300,000 of the bravest men of this world must be killed or banished [captured]^13 in the South before they will think of peace, and in killing them we must lose an equal or greater number, for we must be the attacking party. Still, we as a nation have no alternative or choice." ^14

The increased tempo of the fighting, beginning the following month as Grant began his drive on Richmond, took a terrible toll on both sides. The following is a brief compilation of Union casualties from Grant's opening moves, and does not include those from Sherman's imminent "march" through Georgia and the Carolinas, or fighting in the West:

The Wilderness, 5-7 May: 17,666

Spotsylvania, 10-12 May: 14,267

Drewry's Bluff, 12-16 May: 4,160

Cold Harbor, 1-3 June: 13,078

Petersburg, 15-30 June: 16,569.

This totals 65,740 Union combat casualties with an incomplete tally of the missing. Those who fell sick are not factored in. North Anna, Bermuda Hundred, New Market, and other lesser fights from this campaign cost an additional 11,000 men for an average of 1,400 casualties every day for seven weeks. Until General Lee's surrender on 9 April 1865, Union losses in this theater of operations would subsequently average over 3,000 per month but spurted past 12,000 during a bloody series of failed assaults from the end of July through late August 1864. <page 527> The final ten days of the Appomattox campaign cost an additional 11,200 men. ^15

Except for a significant number of its commissioned officers, most of the U.S. Army's 14,663 men had remained with the Union and were joined by an additional 2,672,341 recruits throughout the conflict. Army strength at the war's close stood at 1,000,516 with many of the missing men having left at the end of their initial enlistments, or been invalided out (frequently with amputations) because of the severity of their wounds. The number killed in action or died of wounds made up 114,757 of this disparity, but over twice this many--- 233,789--- died of disease with an additional 10,982 non-battle deaths due to other causes. ^16 These grim statistics and others have been frequently quoted over the last century. What they inadvertently mask, though, is just how debilitating was the effect that disease and non-battle injuries had on the actual combat strength of the Union Army.

While Union Army medical officers examined literally hundreds of thousands of wounds, they also treated more than 7,000,000 cases of disease. The average soldier became ill on multiple occasions during his enlistment, with dysentery replacing the deadly Revolutionary War scourge, smallpox, as the greatest threat to an American soldier's life.^17 Later, many of the survivors' own grandsons would die of malaria during the war with Spain where over seven times more soldiers fell to disease than bullets,^18 an occurrence which could have had a disastrous impact on the Army's ability to carry out its mission if it had found itself faced with a more skillful and tenacious foe.

During the period of peace which followed the multiplicity of "modern" wars and major campaigns extending from the Napoleonic period to the Franco-Prussian War of 1870-71, the historical incidence of disease and wounding was closely studied in both the United States and Europe as medical and campaign planners tried to come to grips with the problem, and plan for future conflicts. Although the U.S. Army did not compile medical records that are considered completely reliable--- by today's standards--- for nearly a third of the years between 1820 and 1868,^19 record keeping in the United States was generally on a par with that in Europe, and trends from the major American wars were studied as closely as the Crimean and Russo-Turkish Wars, Prussia's wars with Austria and France, and the campaigns of Napoleon. Despite the fact that statistics were compiled by separate national authorities who sometimes ignored various categories, applied! different criteria, or had their results skewed because of the vagary of battlefield medical practices,^20 the total body of work provided planners with extremely useful information.^21

U.S. Army officers acted as observers during the Boer War and the Russo-Japanese War of 1904-5. During the latter, the percentage of soldiers dying of their wounds was technically the lowest of all nineteenth- and twentieth-century conflicts, until the low American loss rate of World War II, largely because the soldiers of both sides frequently died before they could be moved to an aid station and officially counted as "wounded." Nine Army officers, including John F. Morrison, Arthur MacArthur (father of Douglas), Charles Lynch of the Medical Department, and John J. Pershing, were distributed among several Japanese field armies and the Imperial General Staff in Tokyo. These men closely examined every aspect of Japanese operations, and later produced the five-volume Reports of Military Observers Attached to the Armies in <page 528> Manchuria During the Russo-Japanese War, as well as numerous articles and lectures.^22

Within just a few years all but one of the observers would play a role in the First World War, where, for the first time, only slightly more U.S. soldiers died from disease and non-battle injuries than died in battle or from wounds: 55,868 versus 51,259, with total wounded topping 224,000 when gas casualties are included.^23 U.S. Army involvement in World War I was both extensive and more prolonged than is generally realized, with division-sized units in contact with the enemy from October 1917 through November 1918. This allowed postwar Army planners to compile detailed analyses on the average costs to manpower, by percentage, from a wide variety of tactical settings; information which could be used as a starting point for corps and division casualty estimates when using appropriate projection parameters factoring in the sizes of the opposing forces, and types of situations American combat units would either plan for, or have thrus! t upon them by an enemy.

In general terms, this included such things as the historical loss rate during "offensive" action against an enemy main body; an organized or partly disorganized delaying force; or attacks against fortified hill positions, while "defensive" actions covered such things as positional warfare; a fluid defense in the face of the main enemy force, and the always costly withdrawal in the face of an attacking enemy. The effects of climate and terrain were documented, where applicable, as were changes in the casualty rates over time for units involved in various types of operations. This information was taught to Army officers in the 1920s and 1930s as part of their advanced instruction. It was updated and expanded after the Second World War to take into account the operation of armored, amphibious, and airborne formations, and received additional revisions after the Korean and Vietnam wars. The wording has changed only slightly over the last six decades and the governing princip! les not at all.^24

All of these casualty projections studies came (and still come) with a caution that Army instructors repeatedly drilled into the heads of young officers at the Command and General Staff School at Fort <page 529> Leavenworth, Kansas, and the Infantry School at Fort Benning, Georgia (which had future Army Chief of Staff George C. Marshall as its hands-on academic department chief from 1927 to 1932): pre-battle estimates are educated guesses.^25 Tables used in field manuals and other instructional materials were "not [to be] viewed as directly applicable to any future conflict, but as the basis from which planning can begin." ^26 Officers were reminded that the more simple tables were "designed for rough, quick estimates only [emphasis in original text] and not as a substitute for factors carefully chosen to fit the specific assumptions and conditions of a particular operation plan."! ^27 Moreover, if a young officer wading through a staff course had not already perceived the ambiguous nature of the subject because of the deceptively specific nature of the baseline percentages, his trusty field manual spelled it out for him that the material he had just read "clearly indicates that the estimation of probable casualty rates in advance is not a simple matter that can be reduced to a general formula," ^28 and offers numerous warnings, such as "losses may be staggering" if a unit is forced into a "retrograde movement [which] becomes disorderly." ^29

"All this points to the dangers in generalizations in the estimations of casualties," and in order to help mitigate catastrophic miscalculations "each situation must be studied, and an estimate made for each major fraction of the command rather than one estimate for the command as a whole." ^30 This last point is particularly important for understanding the formation of casualty estimates for the invasion of Japan, as is the following: "The longer the conflict progresses, the more comprehensive the statistical base will become. Accurate estimates of losses should, therefore, improve over time as the unique aspects of the conflict become readily apparent." ^31

<page 530> World War II

Early in the Pacific war, medical and campaign planners built their casualty estimates as best they could using tables constructed from the U.S. Army's World War I experience when factored with projected troop strength, operational plans, and intelligence estimates of Japanese capabilities, terrain, and relative firepower. By the invasion of the Philippines, planners at various echelons in MacArthur's headquarters were able to realistically replace the World War I baseline figures with data compiled from the hard-won battles on and around New Guinea. In general, several sets of figures would emerge that might or might not be completely synchronized, depending on the individual interpretations of intelligence data and the level of coordination between the different staffs within a command:

_ First, those used for specific logistical purposes by the medical staff charged with preserving the lives of the wounded. This staff's requirements also had an impact well beyond their immediate domain, such as their determination of how many landing craft the naval element would need to supply for transferring casualties to hospital ships during a given operation, and the tactical factors likely to affect the evacuation. For example, LCVP landing craft were plentiful, but provided a hard ride for the wounded back through the surf to hospital ships. Their maximum capacity was 36 stretcher or 17 ambulatory cases. On the other hand, the very large LST landing craft could handle 220 patients on the vehicle deck, plus up to 175 more in the troop quarters, and even some facilities for emergency medical treatment. They were, however, bigger targets.^32

_ Second, figures used as the basis for estimating the number of replacements needed after a short battle, or to maintain the combat strength of the force during a lengthier fight. Such estimates would take into consideration the historical percentage of casualties suffered by the different branches (artillery, corps of engineers, quartermaster corps, etc.) in the same type of combat that was projected for portions of the operation in question. Average casualties among the infantry were by far the highest at over 80 percent, and a unit could appear--- on paper--- to have suffered comparatively few losses when, in fact, the fighting elements of the unit were dangerously depleted.

_ Third, unlike the casualty estimates created to fulfill the above functional needs, long-range estimates of an academic nature were also <page 531> compiled at higher staff levels. This was done to estimate how future operations would fit into overall campaign objectives and what might be the possible costs of the options the staff either proposed or was directed to plan for or comment on. This type of estimate might be built from scratch like the first two, but was more likely to have used, or been influenced by, the other staffsí calculations when it was created.

Thus, at least three sets of casualty projections--- and sometimes more--- would be created by staffs at each command level from division, corps, army, army group and theater on up through the strategic and logistics planners in Washington. These staffs were not producing the estimates for the benefit of historians but to meet the needs of their own unit and chain of command. Moreover, they spoke their own language with all that implies for individuals sifting through documents today. Casualty projections were seldom directly listed as such or carried convenient titles like "Estimated Losses for Operation X," but were obliquely stated in terms of "requirements" for manpower, or have to be extrapolated, using contemporary formula, from stated medical needs. Complicating matters even further was the fact that there were four groups of staff planners in Washington under the War Department, Navy Department, Joint Chiefs of Staff (JCS), and Combined Chiefs of Staff (CCS! ). Only a comparatively small portion of this complicated planning structure, however, was directly relevant to understanding the formulation of casualty estimates.

The JCS was formed in the critical months after Pearl Harbor. While it was not a unified high command, it did provide a mechanism whereby the Army, Navy, and the Army Air Forces could reach clear agreements or acceptable compromises on nearly all military matters . . . [and] present a common recommendation to the President . . . on policies important enough to require his approval as Commander in Chief.^33

This organization also made up the American element of the CCS with Great Britain, and because "the United States assumed principal responsibility for conducting military operations in the entire Pacific area," the JCS could make "minor strategic decisions and direct the conduct of all operations in the area." ^34 The members of the JCS between early 1942 and late 1945 were General Marshall; Chief of Naval Operations Admiral Ernest J. King; the personal representative of President Franklin D. <page 532> Roosevelt and, later, Harry S. Truman, Admiral William D. Leahy; and the Commanding General of the Army Air Forces, Henry H. (Hap) Arnold.

The committee primarily responsible for assisting the JCS was the Joint Planning Staff (JPS), frequently referred to as the Joint Staff Planners. The JPS seldom directly utilized analyses generated by the Army Service Forces' (ASF) medical and logistics planners, preferring instead to use their own, but did draw heavily on the work of the Joint War Plans Committee (JWPC) and Joint Logistics Plans Committee (JLPC), whose Army elements came from the Operations Division (OPD) of the War Department General Staff.^35 The JWPC created joint, or all-services, strategic and operational plans while the JLPC dealt specifically with two categories of problems--- units and personnel, and logistics analysis. The plans submitted by these groups were adjusted to ensure conformity with commands like Douglas MacArthur's Southwest Pacific and Chester Nimitz's Central Pacific theaters.^36

The Operations Division was made up of five planning groups, with the Strategy and Policy Group, whose chief was a member of the JPS, and the Logistics Group producing casualty projections for their own purposes.^37 The OPD worked very closely with the JWPC in studying the requirements for operations in the Pacific and, in turn, members of the JWPC frequently enlisted the Strategy Section of the OPD's Strategy and Policy Group when they wanted to sound out "Army opinion." For example, in June and July 1944, "the Army section of the JWPC prepared a study on the then controversial subject of operations against [the southernmost Japanese Home Island of] Kyushu, working entirely apart from the Navy members of the committee." This JWPC study went directly to General Marshall as a Strategy Section paper composed of a series of memos.^38 Marshall then had it reviewed by his logistics chief, General B! rehon Somervell of the ASF, and Admiral Leahy.

The JWPC and strategic planners on the other joint committees frequently shared neither up-to-date intelligence nor the most current information on strategic plans with the JLPC and OPD's Logistics Group. Army Service Forces' logisticians complained that they needed to know more if they were to conduct proper analyses and short-notice critiques of JWPC studies, but the strategic planners "were very insistent on the distinction between the two ways of using information about needs and <page 533> resources, and were inclined to be jealous of their prerogative to draw up the estimates used in future planning." ^39

Whether strategic or logistics planners, however, staff officers in Washington, setting to work on the specifics of projected operations against the Japanese Home Islands, had to operate within a framework quite different from staffs further down the chain. Data from World War I did not appear to be useful for the projection of casualty estimates on Japan because of recent leaps in mobility and firepower (to include airpower), yet the early Pacific fighting already experienced was equally inappropriate for use as a model. Most of these early operations either had too few troops involved or were perceived to be unique in character. For example, on Guadalcanal the Army fought Japanese forces involved in a retrograde movement. Japanese troop strength in the Gilbert and Marshall Islands had been small, and offensive operations on and around New Guinea had been conducted against enemy concentrations that were exposed, poorly supported, and capable of being outmaneuvered on the! "sea flank" once a degree of naval and air superiority had been achieved. In addition, the staggering losses from disease at Guadalcanal and the beginning of the New Guinea campaign^40 were not likely to be repeated.

The first campaign to present high-level planners with a convincing model for combat against Japanese field armies on Formosa, the Philippines, and Japan itself, was the invasion of the Marianas^41 which involved 125,000 troops taking the islands of Guam, Saipan, and Tinian. Operations on Saipan were particularly relevant in that the joint Army-Marine force conducted both an opposed landing and ground offensive on a corps (multidivision) frontage against sizeable enemy forces defending terrain similar to that on Japan, and all on an island which contained large numbers of enemy civilians. In terms of future casualty projections:

The Saipan operation, as the midphase of the offensive against the Japanese, lends itself well to a portrayal of amphibious medical service, afloat and ashore. There was both furious resistance on the beachhead reminiscent of Tarawa, and prolonged fighting over <page 534> rugged terrain that gave the Medical Department an opportunity to establish itself fully on shore.^42

Well before the invasion of Saipan, the Joint Staff Planners had been at work on a document outlining their concept of the final stages of the Pacific war. "Operations Against Japan Subsequent to Formosa" was released for comment on 6 June 1944,^43 and approved for submission to the JCS on 30 June 1944,^44 two weeks after the landings on Saipan. Consequently, there was no mention of the battle in the original document. It was soon clear, however, that the number of both military and civilian deaths during an invasion of Japan would be high.

The major fighting for Saipan was over in early July, as several hundred frightened Japanese civilians committed suicide by leaping to their deaths from the cliffs above Marpi Point. Before the point was secured, 4th Division Marines had observed numerous instances of armed Japanese soldiers forcing the civilians over the cliffs to the jagged rocks below,^45 and later in the war, MacArthur's intelligence would report on captured documents and prisoner interrogations which detailed Japanese units killing their own wounded if they could not be evacuated from field hospitals.^46 Altogether, this model for the invasion of Japan had cost the U.S. 3,426 dead and 13,099 wounded to kill 23,811 Japanese defenders. Less than 300 Japanese were taken prisoner, most because they <page 535> were too badly wounded to either fight on or commit a form of ritual suicide, hara-kiri.^47

These losses had a sobering effect on the JCS's Joint Strategic Survey Committee presiding over the refinements made to "Operations Against Japan Subsequent to Formosa," which through the spring of 1945 (long after its name, but not content, was rendered obsolete by the rush of events), was used as the primary outline for the series of campaigns culminating on Japan's soil. In its 30 August 1944 annex, the planners noted the number of Japanese troops which could be made available to defend the Home Islands-3,500,000-and extrapolated that number against a not yet complete count of the destroyed Japanese garrison. The JPS committee concluded:

"In our Saipan operation, it cost approximately one American killed and several wounded to exterminate seven Japanese soldiers. On this basis it might cost us half a million American lives and many times that number wounded . . . in the home islands."

This "Saipan ratio" set the standard for strategic-level casualty projections in the Pacific. Together with the experience of combat attrition of line infantry units in Europe, plus the assumption that fighting in Japan could stretch nearly as far as 1947, it provided the basis for the Army and War Department manpower policy for 1945, and, thus, the pace for the big jump in Selective Service inductions and expansion of the training base even as the war in Europe was winding down.^48

A minimum of 104 copies^49 of "Operations Against Japan Subsequent to Formosa" were distributed to the Secretary of War, all four members of the JCS, their deputies, certain OPD group chiefs, and a wide variety of officers and support staff. Like virtually all JCS materials, the document was classified "top secret," but this did not prevent its contents from being widely discussed by senior officers well beyond the <page 536> confines of Washington. For example, upon General Curtis LeMay's arrival in the Marianas to assume command of the XXI Bomber Command on 19 January 1945, Twentieth Air Force Chief of Staff General Lauris Norstad, Arnold's personal watchdog over the buildup and employment of airpower from the islands,^50 briefed LeMay that "General Arnold needed results." Not mincing words, Norstad said:

You go ahead and get results with the B-29. If you don't get results, you'll be fired. . . . If you don't get results, it will mean eventually a mass amphibious invasion of Japan, to cost probably half a million more American lives.^51

The casualty projections being produced were not static, and climbed skyward as intelligence estimates of enemy strength in targeted areas were made available to the strategic planners and the staffs engaged in high-priority logistics planning such as OPD's Logistics Group and the JLPC- almost immediately to the strategic planners, and with varying degrees of delay to the logisticians.^52 At the Civil Affairs Staging Area in Monterey, California, the JCS Working Group's chief of the Agricultural Section, George L. McColm, was working on plans for Operation Olympic, the first--- and smaller--- of the two planned invasion operations. He noted that "in February and March 1945, the figure used in staff meetings [for the projected] number of casualties we were likely to have during the invasion of Kyushu [Operation Olympic] was 100,000," but added "this wasn't a fixed number." McColm said that "every time the Japanese moved more troops i! n, they had to revise the numbers up." Expected losses during Olympic "more than doubled by about June," and McColm related that the numbers were being revised virtually every week by summer--- sometimes making steep jumps. "It was so common that I stopped paying attention after a while, and besides, it wasn't directly related to my subject area." McColm added that "it was likely <page 537> that they were going up even further at higher [planning] levels because, at our level, we always worked with older numbers." ^53

Eventually those numbers would reach what General LeMay described as "well up into the imaginative brackets and then some," because it was clear that American forces would have to fight literally "millions of well-trained men." ^54 And unlike the final death throes in Germany, which saw Soviet troops engage the bulk of German strength and suffer 352,475 casualties (including 78,291 dead) during their final, twenty-three-day assault on Berlin and central Germany,^55 the twin U.S. invasions of the Japanese Home Islands were to be conducted almost exclusively by American forces. Said John J. Maginnis: "It hadn't really dawned on me before, but the boys headed for the Orient [Japan] were going to have to do it all." ^56 Colonel Maginnis had worked closely with numerous First Army staff and civil affairs officers as it fought its way across France and Germany! , and First Army was slated to end the war as part of the second invasion, Operation Coronet, scheduled to take place in the spring of 1946 near Tokyo. On 28 April 1945, he "went to First Army [headquarters] at Weimar to clean up a few matters before it pulled out for the Pacific." ^57 There he heard for the first time that "over a half million casualties were expected during the invasion of Japan." ^58

More specific total casualty figures than those appearing in "Operations Against Japan Subsequent to Formosa" were not created because of the highly speculative nature of trying to project casualties for <page 538> an unknown number of battles over as much as a two-year period,^59 but neither was the estimate deleted or ordered removed from the heavily and frequently revised document. Apparently, however, the estimate that subjugating Japan "might cost us half a million American lives and many times that many wounded" appeared in the long run to be unrealistically high to a broad cross-section of planners and senior leaders. After all, U.S. forces would certainly learn how to better cope with Japanese defensive techniques through hard-learned battle experience.

The implied top-end figure of approximately 1,700,000 to 2,000,000 battle casualties built on the basis of the Saipan ratio was slashed down to a best-case scenario figure that was not so huge as to make the task ahead appear insurmountable, and use of a 500,000 battle casualty figure was "the operative one at the working level"^60 during the spring of 1945. Andrew J. Goodpaster was then with the Strategy Section of the JWPC. He noted that Secretary of War Henry L. Stimson used the number regularly.^61 When Office of Strategic Services (OSS) officer Samuel Halpern was pulled back to Washington from the China-Burma-India Theater in May 1945 to assist in the invasion planning, the estimate was given in his initial briefing, and even Eighth Air Force maintenance crews at Clovis Air Field, New Mexico, transitioning from the B-17 Flying Fortress they serviced in England to the B-29 Super Fortress t! hey would operate against Japan, were told in May that "the invasion could cost a half million men [and that] every `Fort' they could keep in the air would mean more boys could make it home alive."^62 Halpern said forty-five years later that the 500,000 figure "made a deep, indelible impression on a young man, 23 years old. It is something I have never forgotten." ^63

<page 539> This smaller figure, however, was based on the assumption that the U.S. military would learn to counter Japanese tactics, and it neglected the fact that, as evidenced by the casualty ratios then emerging from Iwo Jima and Okinawa, the Imperial Japanese Army was likewise learning from its experiences. Thus, the "low" 500,000 number for total battle casualties, used widely in briefings, was a best-case estimate not accepted for strategic planning purposes, and it had no effect on the greatly increased Selective Service call-up, the expansion of the Army's training base, or the plans of the Transportation Corps, Medical Corps, and other U.S. Army organizations. For example, at the same time that the lower figure came into use, the Army Service Forces was working with an estimate of "approximately" 720,000 for the projected number of replacements needed for "dead and evacuated wounded" through 31 December 1946.^64

While strategic planners were reluctant, for good reason, to commit estimates for the balance of the war to paper, they frequently set down short-term estimates of one to three months as benchmarks for analyzing differing interpretations of factors affecting future manpower losses, and also approached the question by examining loss ratios from the preceding year of combat. The JCS history of its wartime activities notes that planners "pointed out . . . that in seven amphibious campaigns in the Pacific the casualty rate had run 7.45 per thousand per day; whereas, in the protracted land warfare in the European Theater of Operations it had only been 2.16." ^65 Ongoing intelligence estimates, coupled with the 7.45/2.16 comparison, and a total of 64,391 soldiers and Marines killed and wounded to take an amount of land half the size of wartime Detroit--- the Islands of Okinawa and Iwo Jima^66--- were largely responsible for the increase.

<page 540> These numbers, while large, were not viewed as being out of character with direct combat losses experienced by the other major powers in the war, or even the United States itself, which by the summer of 1945 had lost at least 1,252,000 personnel to combat-related causes, and seen its ranks thinned even further by additional losses in the disease-infested overseas theaters. These additional losses were not officially added to overall figures, but were frequently included in the final numbers of specific campaigns if they severely affected the combat capability of the forces engaged. The bulk of the battle casualties and non-battle deaths occurred in the last year of fighting against Germany, with roughly 1,047,115 suffered by the Army and Army Air Forces.^67

Aside from the above factors ratcheting up overall casualty estimates then being discussed, was the increasing distribution to lower staff levels of intelligence data on the current and near-term strength of Japanese military in the invasion areas, as well as population analyses outlining how much Japanese manpower could realistically be drawn upon for the defense of the Home Islands. One breakdown of the total Japanese male population ages seventeen to forty-four by prefecture and military districts showed that there were 11,287,000 civilian men that could yet be called to defend the Home Islands, with 1,500,800 on just the island <page 541> targeted for the first invasion operation, Olympic.^68 Other published analyses contained detailed information on the composition and training of individual units for comparative purposes, discussion of the reserve system, and other manpower related matters.

There was, as one would expect, a wide range of informed opinion on the casualties question. The possibility had been discussed early on that a direct assault on Japan "may well not be found necessary" if a naval and air blockade proved effective,^69 and was still firmly believed by Admiral King and General Arnold who, nevertheless, gave their full support to invasion plans because of the need to proceed in a unified manner and avoid costly delays.^70 King also "objected to the comparison of casualties between the [European and Pacific] theaters." He pointed out that "there were so many differences in the character of the fighting and the enemies . . . [that] the casualty figures proved nothing, and it would be wrong to judge the relative cost of amphibious and land campaigns by them." ^71 Moreover, by May 1945, with approval of the directive to launch t! he invasion of Kyushu imminent, and tactical plans for the invasion of the Tokyo area well under way, some senior Army planners were optimistic. The envisioned firepower and scheme of maneuver for the invasions, coupled with the surprisingly modest character of Japanese defenses on Kyushu at that time, promised to help keep casualties under control, and some officers thought suggestions that the number of casualties might exceed those suffered against the Germans were unrealistic.^72

Casualty estimates coming to Secretary of War Stimson from former President Herbert Hoover and an unnamed economist (which some researchers have speculated was also Hoover) were far beyond even the "imaginative brackets" referred to by General LeMay. [[[Hoover and the economist were indeed one and the same. Stimson received a memorandum directly from Hoover on 15 May and, when submitting the paper to the Operations Division for comment, hid its originís. Hooverís 30 May memorandum was forwarded to the Army staff under the former presidentís own name. See Timothy Walch and Dwight M. Miller, eds., Herbert Hoover and Harry S. Truman: A Documentary History (Worland, Wyo: High Plains Publishing Co., 1992), 37; and Henry L. Stimson diary entry on June 11, 1945, Yale University Library, from microfilm at HSTL. This is expanded on in a letter to JMH at the end of this article.]]] In the first half o! f June, Stimson twice asked the Army planners to comment on outside estimates that the number of Americans killed could extend from 500,000 to 1,000,000, figures that imply total casualties running in the area of 2,000,000 to 4,000,000. While these numbers were not, in themselves, unimaginable if the intent had been to conquer all Japanese <page 542> forces on the Home Islands by force of arms, such had never been the intent or desire of planners who firmly believed that effective "military control" of all Japan could be "obtained by the securing of a relatively few vital coastal areas" on Honshu,^73 and that the opening invasion of Kyushu would only entail seizing enough land to serve as a base to launch the Honshu invasion toward Tokyo. Military leaders and planners believed that the Japanese, isolated and without allies, would surrender after their capital was taken and their cities destroyed, thus rendering a bloody mop-up of the mountainous, California-sized nati! on unnecessary.

The War Department's initial 7 June response by the Chief of Operations, General Thomas T. Handy, with comments by Marshall voicing his general agreement, stated: "It is obvious that peace would save lives and resources, but the estimated loss of 500,000 lives due to carrying the war to conclusion under our present plan of campaign [emphasis in original] is considered to be entirely too high." ^74 The 14 June response to President Hoover's memorandum, by the chief of the OPD's Strategy and Policy Group, George A. Lincoln, echoed Handy's letter, and went back to President Truman (who was the original recipient of the Hoover memorandum) with notes by Stimson, who made it a point to emphasize that an invasion would cost "a large number of lives." ^75 There were no guarantees that the Japanese would respond in a manner that seemed to make sense to Army planners, and no experience in either ! Japan's battles with American forces, or its previous wars, suggested that capitulation was certain.

Stimson's fears aside, the casualty estimates coming from essentially conservative planners^76 were grim enough. Marshall's Assistant Chief of <page 543> Staff, General John E. Hull, later commented: "There were all kinds of estimates as to the cost of it in manpower, and had the Japanese continued fighting and fought as hard for their homeland as you would expect them to . . . it would have been a bloody operation. . . . The casualty estimates ran everything from a few hundred thousand to a million men to do the thing," ^77 and there was a good deal of opinion on how valid the higher numbers actually were. Paul H. Nitze, the director of the U.S. Strategic Bombing Survey in Japan immediately after World War II, believed that "these fellows were going to fight to the last man . . . the estimate of 500,000 casualties was a gross underestimate," ^78 w! hile General Dwight D. Eisenhower, the Supreme Allied Commander in Europe, believed the opposite: "I thought this a tremendous error in calculation." Eisenhower met with Stimson during a break in the Potsdam Conference, and found that "he was still under the influence of a statement from military sources who figured it would cost 1,000,000 men to invade Japan successfully." Eisenhower, who was intimately familiar with the redeployment schedule of First Army units to the Pacific, but had received nothing more than cursory updates of ongoing operations from that theater, voiced his opinion to Stimson at a 28 July meeting at his headquarters in Frankfurt, Germany. The Secretary of War, on the other hand, had been consuming information on this subject for months, and apparently did not feel the European chief knew enough to offer an informed opinion. Eisenhower later recalled that "he got very impatient and I was glad to change the subject." It was "the only time I ever had a d! ifference with Secretary Stimson." ^< /FONT>79

Meeting with the President

On Thursday, 14 June 1945, a memorandum stamped "urgent" went out to the members of the Joint Chiefs of Staff from President Truman's personal representative on that body, Admiral Leahy. The President wanted a meeting the following Monday afternoon, 18 June, where he would be, in effect, "reopen[ing] the question of whether or not to proceed with plans and preparation with Olympic. . . . a campaign that was <page 544> certain to take a large price in American lives and treasure." ^80 Leahy informed them that the Commander in Chief wanted:

"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.

"An estimate of the time and losses that will result from an effort to defeat Japan by isolation, blockade, and bombardment by sea and air forces.

"Exactly what we want the Russians to do . . . [as well as] what useful contribution, if any, could be made by other Allied nations."

Leahy's memo stated unequivocally that: "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 in money cost is comparatively unimportant." ^81 In compliance with the President's request, Marshall and King asked for additional information from their staffs and from subordinate commands in the Pacific in order to obtain the most current data available. Numerous papers were submitted by the OPD, JWPC, and JPS, with one from the JWPC, "Details of the Campaign Against Japan," ^82 supplying much of the language and analysis that would go into Marshall's opening remarks and the Memorandum for the President.^83 The JWPC felt that "it would be difficult to predict whether Jap resistance on Kyushu would more closely resemble the fighting! on Okinawa or whether it would parallel the Battle of Leyte," and placed quotation marks around the term "educated guess" when they presented their "best estimate" for battle casualties from several possible sequences of operations in the Home Islands.^84

The JWPC's figures were based on the ratio of American to Japanese losses from Leyte, which resulted in lower casualty estimates than those produced by Navy planners, who had selected Okinawa as their model.^85 But while Marshall used much of the JWPC's language, he discarded their figures. The reason for this had less to do with the probability that (in light of the terrain and recent experience) the Navy's estimate was more likely to be appropriate, than that Marshall, in spite of the President's specific request, was very uncomfortable with presenting hard numbers. Instead, Marshall preferred to present ratios that could be used to predict a range in which casualties might fall.

<page 545> Unfortunately, the JWPC figures, apparently made in haste, also had two problems that the committee's director was probably made aware of later by a colleague or superior: they did not compute, and they were based on intelligence estimates of current Japanese strength in the targeted areas instead of the force structures likely to be faced when the invasions were scheduled to take place. In the case of the Tokyo-area assault in 1946, the numbers were particularly skewed because they were derived from the composition of the current defending force^86 fighting the projected overwhelming superiority of two U.S. field armies a full nine months in the future. Consequently, the numbers of killed and missing for the Tokyo area were only moderately higher than for Kyushu: 22,500 versus 21,000. Moreover, a simple deconstruction of the numbers also showed that there was a problem with the figures given for estim! ated wounded since that number was computed to be 20,000 higher at Kyushu than at Tokyo, or 85,000 versus 65,000.^87

As noted earlier, information was also requested from MacArthur, and on 16 June (Manila time), MacArthur's headquarters received the following from Marshall: "Request by 17 June (Washington [time]) the estimate you are using for planning purposes of battle casualties in Olympic up to D+90." ^88 The request was sent to MacArthur's headquarters and forwarded to its G-1 (Personnel Division) for action, and for information-only to the G-3 (Operations Division), G-4 (Supply Division), and the Adjutant General of the Operations Division. The G-1 staff responded the next day with the following:

"Estimate of OLYMPIC battle casualties for planning purposes . . . as follows:

"D to D+30: 50,800
D+30 to D+60: 27,150
D+60 to D+90: 27,100

"The foregoing are estimated total battle casualties from which estimated return to duty numbers are deducted. Not included in the foregoing are non-battle casualties estimated at 4200 for each thirty day period." ^89

<page 545> This was not a satisfactory response and actually presented more questions than it answered. To begin with, instead of receiving projections from MacArthur's G-3, General Stephen J. Chamberlain, or Chief of Staff, General Richard K. Sutherland (either of whom would presumably have had the most complete understanding of the correlation of forces, terrain, and other critical factors, and thus, the best idea of how the coming battle would play out), it came from the Personnel section which would lack the overview Marshall needed. Moreover, the G-1 estimate was obviously derived straight from a logistics analysis from G-3 staff medical planners. Sitting in Washington, Marshall did not know that the Operations section had received only an "information" copy, but the response still begged the question: Had the Operations and Personnel staffs not conducted their own analyses? And was the medical planning constructed from prewar tables or, more likely, the result! s of MacArthur's recent campaigns? A glimpse at the one named category, nonbattle casualties, showed that they were glaringly understated but, due to its likely origin, probably did not surprise the well-briefed Marshall. It indicated that AFPAC's medical staff was using, or at least relying heavily on, an analysis constructed even further down the chain of command at the field army slated for Olympic, the Sixth Army.

The inordinately small nonbattle casualty rate of 4,200 per month was well under half that of the Okinawa Campaign, still in progress, on which Marshall was receiving daily updates. Dr. Michael DeBakey, who would later become well known for his work in cardiovascular and heart surgery, was then an Army colonel and the Surgical Branch chief of the Medical Corps' Consultants Division. He and Captain Gilbert W. Beebe, Ph.D., M.D., edited the classified War Department bulletin Health and produced the analyses that were the basis of many of the casualty projections emanating from the OPD.^90 DeBakey and Beebe later stated in Battle Casualties: Incidence, Mortality, and Logistic Considerations: "In the Sixth Army, there was long a tendency to deny the existence of the problem [psychiatric breakdowns popularly referred to as `combat fatigue'] except as it could be understood in terms of such concepts as `straggler,' which wer! e current in the Civil War." Consequently, the high neuropsychiatric portion of Sixth Army nondeath casualties went essentially unacknowledged in medical reports, even though the affected soldiers were also not present for duty in their combat units at the time they were most needed. Neuropsychiatric admissions to field hospitals nearly equaled wounded admissions, and generally ran in tandem, but <page 547> slightly behind their occurrence, with the offset being anything from a day or two to perhaps a week. This deficiency in the Sixth Army medical reports was glaringly evident and was well known to senior planners and Army medical personnel privy to the data, but was looked upon as essentially a "reporting problem" to be borne in mind when working with data from the Southwest Pacific theater as a whole.^91

What this indicated to Marshall was that there were at least three specific reasons why the ninety-day estimate did not present data that could be used at the 18 June meeting: (1) it was obviously produced as a logistics analysis at field army level; (2) it was unknown if the field army medical staff appreciated that the theater command above them had a historically wide divergence between the number of Japanese it killed versus the number of American casualties suffered, and that this divergence was likely to narrow because of both the defensive terrain on Kyushu and the probable length of the campaign; (3) the medical staff certainly understood that the Japanese would fight even more tenaciously when defending their own land, but it was unknown whether this was factored into their projections or they were using compilations based strictly on MacArthur's recent Pacific campaigns.

Marshall shot back a request for clarification, but this time, instead of from "Washington" to Headquarters AFPAC and signed "Marshall," it was from "General Marshall" to "General MacArthur (Personal)."

"The President is very much concerned as to the number of casualties we will receive in the OLYMPIC operation. This will be discussed with the President about 3:30 PM today Washington time. Is the estimate given in your C-19571 of 50,800 for the period of D to D+30 based on plans for medical installations to be established or is it your best estimate of the casualties you anticipate from the operational viewpoint. Please rush answer." ^92

Marshall's reference to medical installations indicates that either he or someone on his staff recognized that the steep drop after the first month of combat, followed by essentially flat numbers on subsequent months, indicated that these figures represented estimates of occupied <page 548> beds in forward hospitals during the initial surge of casualties, followed by the average number of occupied hospital beds once evacuation procedures to facilities further to the rear were in full swing. In general terms, forward hospitals clear out admissions from earlier operations and day-to-day activities to the maximum extent possible in anticipation of upcoming operations. This is done by evacuation of patients to hospitals further to the rear or through discharge.

The "dispersion allowance," or number of beds required to remain available, can differ greatly depending on the type of hospital, its location in relation to the area of combat, and how it fits into the evacuation scheme. But, on average, there should be a minimum of 20 percent more beds available in the theater of operations than casualties predicted.^93 What these three numbers suggested to Marshall was that medical planners at field army level anticipated an initial casualty surge of battle and nonbattle casualties of perhaps 70,000 men during the first month and as many as 50,000 per month after that.

MacArthur responded immediately and tried to put the best possible face on the numbers that his G-1 staff had transmitted:

"Estimate of casualties contained in my C-19571 was a routine report submitted direct by a staff section without higher reference for medical and replacement planning purposes. The estimate was derived from the casualty rates in Normandy and Okinawa, the highest our forces have sustained as 3.8 men per thousand per day. The total force involved was estimated as 681,000 with one half engaged the first 15 days and the entire strength thereafter. The estimate is purely academic and routine and was made for planning alone. It had not come to my prior attention. I do not anticipate such a high rate of loss." ^94

This first portion of MacArthur's communication was a mishmash of buzzwords and conflicting data that could not have been very heartening for the Chief of Staff to read. Marshall, of course, knew that medical estimates are anything but "academic," and are used to make projections for specific requirements. Perhaps because of the rush to send a reply to Marshall, MacArthur inadvertently presented a "total force" number at least three months old, and which had since mushroomed by over 85,000 troops, reaching 766,986 by the middle of May. In addition, the statement that the projections were based on campaigns other than MacArthur's did not track with the data from Normandy and Okinawa, with which Marshall was intimately familiar. MacArthur's use of the 3.8 figure for Okinawa was probably obtained from the War Department <page 549> itself (it was later revised down to 3.5^95), but the G-1 figures derived from Sixth Army data impli! ed that, even after evacuations from theater hospitals are considered, casualties would drop off steeply after the first month when, in the models they were alleged to be based on, casualties had actually remained relatively steady throughout the Normandy Campaign, and subsequent "breakout," ^96 as well as during all but the final weeks on Okinawa.^97 And finally, MacArthur (who was just as loath to give a specific casualty figure as Marshall) had still not answered the President's question: What was the casualty estimate he was using for planning purposes for the first ninety days' fighting on Kyushu?

The AFPAC medical planners may have been influenced by Okinawa and possibly Iwo Jima in their casualty estimates for the first thirty days, but there was nothing in the fighting on Okinawa or elsewhere during the war that suggested that a major land campaign would necessarily see casualty rates drop so precipitously after the first month. Earlier island fighting in the Pacific did not factor in length-versus-casualty ratio studies since, with the qualified exceptions of New Guinea and the Philippines, the Japanese garrisons could not be reinforced. In addition to the medical staff projections, MacArthur's G-2 (intelligence) staff had also decided that Okinawa offered the best over-all model for future projections, and used a simple formula when making tentative projections: approximately "two to two and a half Japanese divisions [could] extract . . . approximately 40,000 American casualties on land." Writing in 1945 for a military audience of field grade and general offic! ers on plans for the invasion of Japan, MacArthur's intelligence chief, General Charles A. Willoughby, maintained that "This [ratio] affords a completely authentic yardstick to forecast what it would have taken in losses had we gone in shooting." ^98

It is virtually impossible that MacArthur did not know the thinking of his intelligence chief and long-time friend on this matter, and he would have known that Marshall could not help but notice his dodging and weaving. But MacArthur was groping to get something down on a piece of paper that could well end up being shown to the <page 550> President of the United States. As early as the summer of 1944, when the 1,000,000 figure began to be discussed in Washington, both his Chief of Staff, Sutherland, on 26 August, and Central Pacific Commander in Chief Admiral Chester W. Nimitz's Chief of Staff, on 22 August, "were instructed to make their personnel requirements for operations as low as possible," ^99 and Marshall's statement that "the President is very much concerned as to the number of casualties" sent the warning lights flashing. President Roosevelt would not have not been sidetracked by the numbers, but exactly who was Harr! y S. Truman? Would the new President, fearing a bloodbath, throw out the invasion plans altogether, and perhaps pursue a plan of encirclement which, in the long run, could lose even more men through a costly series of amphibious operations--- and still not get the job done? The second half of MacArthur's letter did not answer Marshall's question but did give the Chief of Staff something he could work with, and it was incorporated into the Monday briefing with only a half hour to spare:

"I believe the operation presents less hazards of excessive loss than any other that has been suggested, and that its decisive effect will eventually save lives by eliminating wasteful operations of a nondecisive character. I regard the operation as the most economical one in effort and lives that is possible. In this respect it must be remembered that the several preceding months will involve practically no losses in ground troops and that sooner or later a decisive ground attack must be made. The hazard and loss will be greatly lessened if an attack is launched from Siberia sufficiently ahead of our target date to commit the enemy to major combat. I most earnestly recommend no change in OLYMPIC. Additional subsidiary attacks will simply build up our final total casualties." ^100

The focus of MacArthur's worries, President Truman, had only recently retired as a highly respected colonel in the Reserve Officer Corps. After serving with distinction as a battery commander during the Meuse-Argonne Offensive in the First World War, the former Missouri "dirt farmer" had retained his commission and rose through the ranks to become the commander of the 381st Artillery Regiment from 1932 to 1935, and then the 379th Artillery Regiment from 1935 to 1940. During this period, the 379th's sister regiments, the 380th and 381st were commanded by artillery officers who would later serve in Truman's administration: Harry Vaughan, as Military Aide, and John Snyder as Treasury Secretary.^101 The President's cousin was Major General Ralph Truman.

<page 551> Truman took his commission seriously and immersed himself in soldiering in spite of his increasing duties in a series of political positions. He was a familiar figure at the artillery range at Fort Riley, Kansas,^102 and after arriving in Washington as a newly elected U.S. senator, he admitted in a letter to his wife that he "played hooky" from a meeting of the Senate Appropriations Committee to attend a lecture by Douglas Southall Freeman on General Robert E. Lee at the Army War College.^103 On another occasion, when Colonel Snyder paid a visit to him in Washington, the two artillerymen drove over to the Gettysburg battlefield to perform what the Army today calls a "terrain walk," in order to examine the military aspects of the ground that Union and Confederate forces fought over in 1863.^104 The Transportation Act of 1940 (or Wheeler-T! ruman Act) was formulated by Truman and greatly increased America's preparedness for war. He was also an active member of three key armed services committees--- the Military Affairs Committee, the Military Subcommittee of the Appropriations Committee, and Chairman of the Senate Special Committee to Investigate the National Defense Program--- so when he went to General Marshall and offered to rejoin the Regular Army he was politely rebuffed by the Chief of Staff who said: "Senator, you've got a big job to do right up there at the Capitol with your Investigating Committee. Besides, this is a young man's war. We don't need old stiffs like you." ^105

In some ways, Truman's working knowledge of the nuances of military planning and analysis has worked to the detriment of historians, and this can be no more clearly seen than in the well-worn misperceptions of what was actually said by the participants of that Monday afternoon meeting in the White House Cabinet Room. A less knowledgeable--- or astute--- President would have needed the Joint Chiefs' opinions expressed in more basic terms which would have not only been helpful to the President, but also to later researchers poring over the meeting's transcript. Truman, however, used the same form of verbal shorthand, based on a common understanding of both the methodologies and assumptions used to formulate military analyses, as his peers around the table. Marshall's steady guidance was evident throughout the sixty-minute discussion, and it is President Truman himself who made it a point to look beyond the numbers to the impact in human terms. The principal participants in ! this meeting with President <page 552> Truman were Marshall, King, Leahy, Stimson, General Ira C. Eaker (substituting for Arnold), Navy Secretary James Forrestal, and Assistant Secretary of War John J. McCloy. Extended portions of the meeting's minutes are presented below, with additional commentary on the institutional context in which the participants were operating, and the various methodologies used in making casualty projections:


Extracted from minutes of meeting held at the White House 18 June 1945 at 1530.

"THE PRESIDENT stated that he had called the meeting for the purpose of informing himself with respect to the details of the campaign against Japan set out in Admiral Leahy's memorandum to the Joint Chiefs of Staff of 14 June. He asked General Marshall if he would express his opinion.

"GENERAL MARSHALL pointed out that the present situation with respect to operations against Japan was practically identical with the situation which had existed in connection with the operations proposed against Normandy. He then read, as an expression of his views, the following digest of a memorandum prepared by the Joint Chiefs of Staff for presentation to the President, J.C.S. (1388):

"Our air and sea power has already greatly reduced movement of Jap shipping south of Korea and should in the next few months cut it to a trickle if not choke it off entirely. Hence, there is no need for seizing further positions in order to block Japanese communications south of Korea. General MacArthur and Admiral Nimitz are in agreement with the Chiefs of Staff in selecting 1 November as the target date to go into Kyushu because by that time:

a. If we press preparations, we can be ready.

b. Our estimates are that our air action will have smashed practically every industrial target worth hitting in Japan as well as destroying huge areas in the Jap cities.

c. The Japanese Navy, if any still exists, will be completely powerless.

d. Our sea action and air power will have cut Jap reinforcement capabilities from the mainland to negligible proportions.

"Important considerations bearing on the 1 November date rather than a later one are the weather and cutting to a minimum Jap time for preparation <page 553> of defenses. If we delay much after the beginning of November the weather situation in the succeeding months may be such that the invasion of Japan, and hence the end of the war, will be delayed up to 6 months.

"An outstanding military point about attacking Korea is the difficult terrain and beach conditions which appear to make the only acceptable assault areas Fusan in the southeast corner and Keijo, well up the western side.^107 To get to Fusan, which is a strongly fortified area, we must move large and vulnerable assault forces past heavily fortified Japanese areas. The operation appears more difficult and costly than an assault on Kyushu. Keijo appears an equally difficult and costly operation. After we have undertaken either one of them we still will not be as far forward as going into Kyushu.

"The Kyushu operation is essential to a strategy of strangulation and appears to be the least costly worthwhile operation following Okinawa. The basic point is that a lodgement in Kyushu is essential, both to tightening our strangle hold of blockade and bombardment on Japan, and to forcing capitulation by invasion of the Tokyo Plain.

"We are bringing to bear against the Japanese every weapon and all the force we can employ, and there is no reduction in our maximum possible application of bombardment and blockade, while at the same time we are pressing invasion preparations. It seems that if the Japanese are ever willing to capitulate short of complete military defeat in the field, they will do it when faced by the completely hopeless prospect occasioned by (1) destruction already wrought by air bombardment and sea blockade, coupled with (2) a landing on Japan indicating the firmness of our resolution, and also perhaps coupled with (3) the entry or threat of entry of Russia into the war.

"With references to clean up of the Asiatic mainland, our objective should be to get the Russians to deal with the Japs in Manchuria (and Korea if necessary) and to vitalize the Chinese to a point where, with assistance of American air power and some supplies, they can mop out their own country.

"Casualties. Our experience in the Pacific war is so diverse as to casualties that it is considered wrong to give any estimate in numbers. Using various combinations of Pacific experience, the War Department staff reaches the conclusion that the cost of securing a worthwhile position in Korea would almost certainly be greater than the cost of the Kyushu operation. Points on the optimistic side of the Kyushu operation are that: General MacArthur has not yet accepted responsibility for going ashore where there would be disproportionate casualties. The nature of the objective area gives room for maneuver, both on the land and by sea. As to any discussion of specific operations, the following data are pertinent:


U.S. Casualties
Killed, wounded, missing

Jap Casualties
Killed and Prisoners
Not inc. wounded

U.S. to Jap




1: 4.6




1: 5

Iwo Jima



1: 1.25


34,000 [Ground] 7,700 Navy

81,000 [Incomplete count]

1: 2

1st 30 days




"The record of General MacArthur's operations from 1 March 1944 through 1 May 1945 shows 13,742 U.S. killed compared to 310,165 Japanese killed, or a ratio of 22 to 1." ^108

[It is at this point in the discussion, that historians--- military and civilian alike--- invariably lose track of exactly what the participants are saying. The numbers above were used solely as a base to establish ratios of U.S. to Japanese casualties during the most recent campaigns. A ratio which stripped U.S. wounded from the equation and measured only killed to killed (which the Japanese casualties almost invariably were) is presented for comparative purposes, as well as a U.S. casualty total for the first third of the fighting in the Normandy Campaign with no ratios given. After the presentation of casualty ratios, casualty numbers are not used. Instead, there are references to operations or portions of operations. These references do not refer to the baseline figures, which are frequently quoted by authors and historians, but to the ratios they spawned and which only suggest how casualties from the much larger Japanese and American forces involved in ! the first of the two invasion operations might play out.]

"There is reason to believe that the first 30 days in Kyushu should not exceed the price we have paid for Luzon. It is a grim fact that there is not an easy, bloodless way to victory in war and it is the thankless task of the leaders to maintain their firm outward front which holds the resolution of their subordinates. Any irresolution in the leaders may result in costly weakening and indecision in the subordinates. . . ."

[The "price" Marshall refers to is one American battle casualty for <page 555> every five Japanese, and not the specific number of casualties from the smaller operation on Luzon. The limitation of the estimate to "the first 30 days" is made because the ratio could very easily change as U.S. soldiers and Marines started to fight their way into the mountains against additional Japanese divisions moving down from northern Kyushu. Detailed speculation beyond thirty days could come back to haunt the Chief of Staff and even the first thirty-day projection is hedged by the qualifier, "there is reason to believe."]

"GENERAL MARSHALL said that he had asked General MacArthur's opinion on the proposed operation and had received from him the following telegram, which General Marshall then read:

" ĎI believe the operation presents less hazards of excessive loss than any other that has been suggested and that its decisive effect will eventually save lives by eliminating wasteful operations of nondecisive character. I regard the operation as the most economical one in effort and lives that is possible. In this respect it must be remembered that the several preceding months will involve practically no loss in ground troops and that sooner or later a decisive ground attack must be made. The hazard and loss will be greatly lessened if an attack is launched from Siberia sufficiently ahead of our target date to commit the enemy to major combat. I most earnestly recommend no change in OLYMPIC. Additional subsidiary attacks will simply build up our final total casualties.í "

[As noted earlier, the opening portion of MacArthur's statement and numbers supplied by his staff could not be used.]

"GENERAL MARSHALL said that it was his personal view that the operation against Kyushu was the only course to pursue. He felt that air power alone was not sufficient to put the Japanese out of the war. It was unable alone to put the Germans out. General Eaker and General Eisenhower both agreed to this. Against the Japanese, scattered throughout mountainous country, the problem would be much more difficult than it had been in Germany. He felt that this plan offered the only way the Japanese could be forced into a feeling of utter helplessness. The operation would be difficult but no more so than the assault in Normandy. He was convinced that every individual moving to the Pacific should be indoctrinated with a firm determination to see it though."

[The statement that "the problem would be much more difficult than it had been in Germany" is telling, and it is not contradicted by the follow-up statement that "the operation would be difficult but not more so than the assault in Normandy." Use of the words "in Normandy" instead of "on Normandy" indicates that Marshall was not limiting his comment to the initial "D-Day" landings, which suffered extremely low casualties on four of the five beachheads, but to the total campaign which saw incremental advances through the tangled hedgerow country as American forces positioned themselves for a breakout from the Normandy Peninsula. There were 133,316 American and 91,223 British battle <page 556> casualties during the Normandy Campaign, 6 June through 31 August 1944.^109]

"ADMIRAL KING agreed with General Marshall's views and said that the more he studied the matter, the more he was impressed with the strategic location of Kyushu, which he considered the key to the success of any siege operations. He pointed out that within three months the effects of air power based on Okinawa will begin to be felt strongly in Japan. It seemed to him that Kyushu followed logically after Okinawa. It was a natural setup. It was his opinion that we should do Kyushu now, after which there would be time to judge the effect of possible operations by the Russians and the Chinese. The weather constituted quite a factor. ^110 So far as preparation was concerned, we must aim for Tokyo Plain; otherwise we will never be able to accomplish it. If preparations do not go forward now, they cannot be arranged for later. Once started, however, they can always be stopped if desired.

"GENERAL MARSHALL agreed that Kyushu was a necessity and pointed out that it constituted a landing in the Japanese homeland. Kyushu having been arranged for, the decision as to further action could be made later.

"THE PRESIDENT inquired if a later decision would not depend on what the Russians agreed to do. It was agreed that this would have considerable influence.

"THE PRESIDENT then asked Admiral Leahy for his views of the situation.

"ADMIRAL LEAHY recalled that the President had been interested in knowing what the price in casualties for Kyushu would be and whether or not that price could be paid. He pointed out that the troops on Okinawa had lost 35 percent in casualties. If this percentage were applied to the number of troops to be employed in Kyushu, he thought from the similarity of the fighting to be expected, that this would give a good estimate of the casualties to be expected. He was interested therefore in finding out how many troops are to be used in Kyushu."

[Leahy apparently did not believe that the presented figure of 34,000 for ground force battle casualties offered a true picture of losses on Okinawa, and used the total number of casualties to formulate the 35 percent figure (see footnotes 66 and 91). Since Leahy, as well as the other JCS members and Truman, also already knew approximately how many men were to be committed to the Kyushu fight, he was obviously making an effort--- commonly done in such meetings--- to focus the participants' attention on the statistical consequences of the disparity.]

<page 557> "ADMIRAL KING called attention to what he considered an important difference in Okinawa and Kyushu. There had been only one way to go on Okinawa. This meant a straight frontal attack against a highly fortified position. On Kyushu, however, landings would be made on three fronts simultaneously and there would be much more room for maneuver. It was his opinion that a realistic casualty figure for Kyushu would lie somewhere between the number experienced by General MacArthur in the operations on Luzon and the Okinawa casualties."

[As intelligence gathering operations were currently discovering, however, and postwar prisoner interrogations and direct ground observations after the war would confirm, the Japanese had, through a process of elimination, correctly deduced the exact invasion beaches on Kyushu as well as the probable time that the invasion would be launched. The effort to fortify the beach areas and inland positions would begin in earnest in July, and was to continue throughout the months leading up to the invasion. Unknown to Admiral King at the time of the 18 June meeting, the three-pronged landings on Kyushu "with much more room to maneuver" were effectively going to become three frontal attacks.]

"GENERAL MARSHALL pointed out that the total assault troops for the Kyushu campaign were shown in the memorandum prepared for the President as 766,700. He said, in answer to the President's question as to what opposition could be expected on Kyushu, that it was estimated at eight Japanese divisions or about 350,000 troops. He said that divisions were still being raised in Japan and that reinforcement from other areas was possible but it was becoming increasingly difficult and painful."

[Marshall's figure of 766,700 differs only slightly from that of MacArthur's headquarters which gives 766,986 as the number of men to be landed within a month and a half of the invasion.^111 The U.S. portion of all ratios presented was not just the casualties from the units in contact with the enemy, but represented total force versus total force numbers. Participants at the meeting were familiar with ongoing total casualty figures which determined the number of troops that could be fielded on a given day, and thus knew that the actual casualty figures are considerably higher than this. For discussion purposes, however, inclusion of non-battle injuries, including psychiatric <page 558> breakdowns and disease, is uncommon in such meetings because: (1) they are not directly inflicted by the enemy, and (2) a portion of the affected will eventually return to duty. Noncombat casualties tend to only be added to discussions if! the manpower pool for replacements is so low, or the ability to get replacements to units needing reconstitution is so difficult, that the ability to accomplish a mission is directly affected. In the case of Okinawa, the intensity and prolonged nature of the fighting resulted in an extraordinarily large number of the non-battle casualties which were so severe that, excluding the killed in action, 61,471 battle and nonbattle casualties were missing from the ground units' present for duty strength at the end of June, weeks after the heaviest fighting had ended.^112]

"THE PRESIDENT asked about the possibility of reinforcements for Kyushu moving south from the other Japanese islands.

"GENERAL MARSHALL said that it was expected that all communications with Kyushu would be destroyed.

"ADMIRAL KING described in some detail the land communications between the other Japanese islands and Kyushu and stated that as a result of operations already planned, the Japanese would have to depend on sea shipping for any reinforcement.

"ADMIRAL LEAHY stressed the fact that Kyushu was an island. It was crossed by a mountain range, which would be difficult for either the Japanese or the Americans to cross. The Kyushu operation, in effect, contemplated the taking of another island from which to bring increased air power against Japan."

[The assurances of the Joint Chiefs were all true enough as stated, but skirted the fact that the "island" was bigger in size and population than some American states, and that Kyushu was largely self-sufficient militarily. Moreover, it was questionable if the mountains would be a greater barrier for the foot-mobile Japanese or the invading American forces who would have to fight through the southern portion of them and then form a noncontiguous defensive line of comparatively isolated hilltop positions for the duration of the war. Truman, who had been monitoring the rising casualty figures from Okinawa on a daily basis,^113 was very well aware of this and cut to the bottom line.]

THE PRESIDENT expressed the view that it was practically creating another Okinawa closer to Japan, to which the Chiefs of Staff agreed.

"THE PRESIDENT then asked General Eaker for his opinion of the operation as an airman.

"GENERAL EAKER said that he agreed completely with the statements made by General Marshall in his digest of the memorandum prepared for the President. He had just received a cable in which General Arnold also <page 559> expressed complete agreement. He stated that any blockade of Honshu was completely dependent upon airdromes on Kyushu; that the air plan contemplated employment of 40 groups of heavy bombers against Japan and that these could not be deployed without the use of airfields on Kyushu. He said that those who advocated the use against Japan of air power alone overlooked the very impressive fact that air casualties are always much heavier when the air faces the enemy alone and that these casualties never fail to drop as soon as the ground forces come in. Present air casualties are averaging 2 percent per mission, about 30 percent per month. He wished to point out and to emphasize that delay favored only the enemy and he urged that there be no delay.

"THE PRESIDENT said that, as he understood it, the Joint Chiefs of Staff after weighing all the possibilities of the situation and considering all possible alternative plans were still of the unanimous opinion that the Kyushu operation was the best solution under the circumstances.

The Chiefs of Staff agreed that this was so.

"THE PRESIDENT then asked the Secretary of War for his opinion.

"MR. STIMSON agreed with the Chiefs of Staff that there was no other choice. He felt that he was personally responsible to the President more for his political than for military considerations. It was his opinion that there was a large submerged class in Japan who do not favor the present war and whose full opinion and influence had never yet been felt. He felt sure that this submerged class would fight and fight tenaciously if attacked on their own ground. He was concerned that something should be done to arouse them and to develop any possible influence they might have before it became necessary to come to grips with them.

"THE PRESIDENT stated that this possibility was being worked on all the time. He asked if the invasion of Japan by white men would not have the effect of more closely uniting the Japanese."

"MR. STIMSON thought there was every prospect of this. He agreed with the plan proposed by the Joint Chiefs of Staff as being the best thing to do, but he still hoped for some fruitful accomplishment through other means."

[The other means included a range of measures from increased political pressure brought to bear through a display of Allied unanimity at the upcoming conference in Potsdam to the as yet untested atomic weapons then in production that it was hoped would "shock" the Japanese into surrender.]

"THE PRESlDENT then asked for the views of the Secretary of the Navy.

"MR. FORRESTAL pointed out that even if we wished to besiege Japan for a year or a year-and- a-half, the capture of Kyushu would still be essential. Therefore, the sound decision is to proceed with the operation against Kyushu. There will still be time thereafter to consider the main decision in the light of subsequent events. . . ."

[The question of how long it would take an air-sea blockade to force a Japanese surrender had been a hotly debated topic among mid- and senior-level military planners for over two years. Approximately a year and a half from the full establishment of airpower on Okinawa, where <page 560> some bases for medium-range bombers were already in operation, was believed to be the outer limit of how long it would take such a course to work. It was the consensus of the military leadership that such a blockade could not be made fully effective without additional bases established further to the north on Kyushu. The merits of unconditional surrender were briefly touched on at this point.]

"THE PRESIDENT said he considered the Kyushu plan all right from the military standpoint, and so far as he was concerned, the Joint Chiefs of Staff could go ahead with it; that we can do this operation and then decide as to the final action later."

[A short discussion of the British, Chinese, and Portuguese roles in ending the war ensued, and the President moved to wrap up the meeting.]

"THE PRESlDENT reiterated that his main reason for this conference with the Chiefs of Staff was his desire to know definitely how far we could afford to go in the Japanese campaign. He had hoped that there was a possibility of preventing an Okinawa from one end of Japan to the other. He was clear on the situation now and was quite sure that the Joint Chiefs of Staff should proceed with the Kyushu operation. . . ."

[President Truman's comment of the invasion operations representing "an Okinawa from one end of Japan to the other"--- now made twice--- is an unequivocal indication of what he believed would be the magnitude of the fighting. The Japanese Navy was essentially destroyed, but Japanese air power was being preserved for the invasions. Japan's field armies in the Home Islands were swelling rapidly, and there was ample time to train recruits not only for the defense of the Tokyo area in 1946, but for the defense of Kyushu in the coming winter.]

"ADMIRAL KING said he wished to emphasize the point that, regardless of the desirability of the Russians entering the war, they were not indispensable and he did not think we should go so far as to beg them to come in. While the cost of defeating Japan would be greater, there was no question in his mind but that we could handle it alone. He thought that the realization of this fact should greatly strengthen the President's hand in the forthcoming conference.

THE PRESIDENT and the Chiefs of Staff then discussed certain other matters."

The "certain other matters," according to McCloy, was the atomic bomb.^114

Secretary of War Stimson summed up his view of the meeting in a 2 July memo to the President:

"The plans of operation up to and including the first landing have been authorized and the preparations for the operation are now <page 561> actually going on. This situation was accepted by all members of your conference on Monday, 18 June.

"There is reason to believe that the operation for the occupation of Japan following the landing may be a very long, costly and arduous struggle on our part. The terrain, much of which I have visited several times, has left the impression on my memory of being one which would be susceptible to a last ditch defense such as has been made on Iwo Jima and Okinawa and which of course is very much larger than either of those two areas. According to my recollection it will be much more unfavorable with regard to tank maneuvering than either the Philippines or Germany.

"If we once land on one of the main islands and begin a forceful occupation of Japan, we shall probably have cast the die of last ditch resistance. The Japanese are highly patriotic and certainly susceptible to calls for fanatical resistance to repel an invasion. Once started in actual invasion, we shall in my opinion have to go through with an even more bitter finish fight than in Germany. We shall incur the losses incident to such a war and we shall have to leave the Japanese islands even more thoroughly destroyed than was the case with Germany. This would be due both to the difference in the Japanese and German personal character and the differences in the size and character of the terrain through which the operations will take place." ^115

The Secretary of War had been a colonel of artillery during the brutal fighting of World War I, and Truman would not take lightly his appraisal of the targeted terrain gained from direct examination on multiple occasions. Moreover, on the subject of casualties, the President did not need to have it explained to him what Stimson meant by "an even more bitter finish fight than Germany" since he and everyone else who had taken part in the 18 June meeting knew that it had cost roughly a million American casualties to defeat the Nazis, and U.S. casualties were actually small when compared to those of our major allies. Stimson's warning that "we shall incur the losses incident to such a war" was equally clear. For anyone not understanding the reference, Stimson spelled it out in a high-profile magazine article after the defeat of Japan:

"We estimated that if we should be forced to carry this plan to its conclusion, the major fighting would not end until the latter part of 1946, at the earliest. I was informed that such operations might be expected to cost over a million casualties." ^116

<page 562> Potsdam

Subsequent to the 18 June meeting, various changes were proposed to "Details of the Campaign Against Japan" which had to be completed before the President left to attend the Potsdam Conference with Winston Churchill and Joseph Stalin. Admiral King believed "that the [Joint] Chiefs of Staff will have to give an estimate of the casualties expected" during Olympic, and called attention to Nimitz's estimate of "49,000 in the first thirty days." ^117 Marshall disagreed saying, "it seems unnecessary and undesirable for the Joint Chiefs of Staff to make estimates, which can at best be only speculative," but indicated that he "might" be willing to include the tabulations establishing casualty ratios.^118 King reluctantly agreed^119 and the final version of the report contained only the baseline figures (quoted by historians) and the ratios they established (used by planners and decision makers), but with two modifications. Instead of counting just the 76,000 uniformed and tra! ined Japanese defenders on Okinawa, the 24,000 "recently impressed indigenous militia and labor groups" (a third of whom surrendered) were added, as was, apparently--- and inappropriately--- a portion of the civilian dead and wounded accounted for by early July.^120 These additions brought the estimated Japanese casualties up to 119,000 in the tabulation, and thus changed the ratio of U.S. to Japanese battle casualties from 1 to 2 in the original report to 1 to 3. An additional line of text was also added after the sentence recounting MacArthur's killed to killed ratio of 22 to 1: "During this same period the total U.S. [battle] casualties, killed, wounded and missing, were 63,501 or a ratio of approximately 5 to 1." ^121

At the same time that the final wording was being negotiated on "Details of the Campaign Against Japan," a series of disturbing, in-depth studies of the Japanese manpower pool were disseminated in the War Department's Military Research Bulletin. Building on an earlier, more general analysis,^122 the War Department's Intelligence Division concluded that: "By the end of 1945 all of the estimated 700,000 men inducted from 1 Dec 1944 to 1 June 1945, together with an estimated 400,000 of those still to be inducted, should be trained and ready for use <page 563> by the armed forces." In addition, it was also estimated that: "The natural growth of the population, the continued mobilization of women, and the return of men invalided out of the armed services should furnish more than 1,000,000 new workers, and largely offset the expected transfer of fit males of military age to the armed forces in 1945." Concurrent with this, tho! ugh, would also be a large release of workers due to the efforts of the strategic bombing campaign waged from the Marianas and Okinawa:

"because of the disruptive effects of air attack and the necessity of dispersing her industries, Japan may run into serious local labor shortages, particularly in the skilled labor classifications. However, it seems fairly clear that the transfer of as many as 2,000,000 men from the working population would not be a limiting factor on Japan's economic activity in 1945."

From the standpoint of manpower and training, Japan had ample time to beef up its forces for the coming attritional warfare on its home ground because of its two complementary induction and training systems capable of handling a minimum of 100,000 men inducted each and every month. Manpower was not going to become a problem for the Japanese military. Of the estimated 10,597,000 males aged seventeen to forty-four in civilian life on 1 June 1945, approximately 83 percent, or 8,797,000, were fit for military service.^123

A follow-up study, sent to Marshall shortly after his arrival at the Potsdam Conference, dissected the Japanese system of defense call up and provided no silver linings.^124 Whichever way one looked at the numbers, it was apparent that the Japanese military would be able to field an army of approximately 5,000,000 men--- most infantry, and all adequately supplied for foot-mobile warfare in rough terrain against the invasions, with an unknown number in the single-digit millions having little or no training but, nevertheless, available as porters, engineers, and "cannon fodder." Thus, the total number of trained and available Japanese soldiers was likely to be far larger than the 3,500,000-man force that the Saipan ratio had originally been applied to in "Operations Against Japan Subsequent to Formosa." While this would complicate the extended fighting during the first invasion operation, Olympic, it was especially troubling for the o! ne fully three-quarters of a year away in the Tokyo area, Coronet.^125

<page 564> This question of how many Japanese troops would actually be faced "on the ground" had been building for over a year, and long before the probable number of sustainable Imperial soldiers reached the estimated 5,000,000 mark, Stimson and Marshall (the War Department set policy and the Army defined requirements) had moved to ensure that an adequate number of replacements would be available to the three to four American field armies that would fight in the Home Islands. Historians examining the composition of the formations taking part in the invasions almost invariably make a head count of the combat and support units based on those formations' reported table of organization strengths, and neglect the fact that the armed services had to plan for logistics and manpower to sustain the multiple land campaigns into at least late 1946. This entailed everything from the construction of massive, prefabricated components for an artificial harbor to support forces in! vading Honshu (a $50,000,000 affair being built in San Francisco which had a priority second only to the Manhattan Project producing the atom bombs),^126 to an actual expansion of the Army's Replacement Training Centers even as the war in Europe was winding down.^127 When Stimson made the statement, "the total U.S. military and naval force involved in this grand design was of the order of 5,000,000 men," it was no exaggeration, and he correctly pointed out that, "if all those directly concerned are included, it was larger still." ^128

The initial levies of fresh troops for combat units in 1945 (running in the 70- and 80-thousands per month from both the late summer 1944 Selective Service induction and the culling of excess soldiers from service and support units during the Battle of the Bulge) were primarily for <page 565> the replacement of casualties. Soldiers inducted during the final months of 1944 or reassigned from European Theater units demobilized in 1945 would generally be used as replacements for men expected to be released after a specified amount of time in combat under what was called the "points system." ^129 These replacements were the soldiers who would fill out the units making the initial assaults during Olympic and Coronet. Yet more troops would be needed to take the place of battle and nonbattle casualties lost during the fighting in Japan. Even as the Army's monthly Selective Service call-up was increased from 60,000 to 80,000 in Januar! y 1945, it was decided to raise it yet again in the spring to 100,000 per month.^130 Under the direction of Army Ground Forces commander General Joseph Stilwell, several organizational schemes were examined for the training and replacement of casualties in the upcoming invasions.

For a war-weary American public, this was a bitter pill, but it was made more palatable by passage of a law decreeing that eighteen-year-olds <page 566> could not be sent into combat unless they had finished a minimum of six months training, as well as the return home of the first soldiers under the points system even though they were not yet released from service. During the spring and summer of 1945, the Army planned for the expansion of the number of training regiments to thirty-four in order to form a ready pool of replacements, and "the capacity of Army Ground Force replacement training centers reached a wartime peak of 400,000 in June 1945," months after the last shots were fired in Europe.^131

The Army viewed the imposed shortfall of eighteen-year-olds entering their replacement pool, due to lengthened training, as purely transitional.^132 Inductees from the spring and summer of 1945 would begin to be available for the Pacific replacement stream before Olympic, and virtually all would be available by the initiation of Coronet. Inductees from the fall of 1945 would generally become available around the time that Tokyo itself was assaulted, while those from the spring of 1946 would take the place of losses from any subsequent operations like the drive up the Tone Valley and across the Mi Kuni Mountains to split Honshu in half, or assaults on the Osaka or Nagoya areas. Beyond this point, it was unlikely that major offensives would be necessary, even if the Japanese had not actually surrendered and major portions of the country remained unoccupied by American troops.

Unlike the later system of limited service by draftees and those who enlisted, young men inducted during the Second World War were in "for the duration" unless they were either killed, invalided through serious wounding or other incapacitation, or returned to the United States through the points system or one of its predecessors. What this meant in terms of the planned invasions of Japan was essentially this: starting in March 1945, when levies were increased to 100,000 per month for the U.S. Army alone (the Navy and Marine Corps were counted separately and brought the total March call-up to 141,200),^133 nearly every man inducted would enter the "replacement stream" now oriented for a one-front war that was estimated to last at least through the latter part of 1946.^134 Induction levels were unaffected by "best case" casualty <page 567> estimates formulated in the spring of 1945 and wou! ld see no significant tapering off until the winter of 1945-46 and even this was somewhat problematical if the fighting in the Home Islands did not go as planned. The Army did not sugar coat the prospect of a long war for the soldiers in the field and new inductees; it warned that various "major factors--- none of them predictable at this stage of the game--- will decide whether it will take 1 year, 2 years or longer to win the Far East war." ^135

Aside from this massive induction of young men and expansion of training facilities, there is another clear indication that the U.S. Army, as an institution, expected casualties to be on par with those suffered in the European Theater even though an "accurate forecast of combat requirements was particularly difficult." ^136 June and July 1945 saw the long-expected surge of incoming casualties to stateside hospitals peak at over 300,000. The Surgeon General's office had always managed to keep the number of occupied beds well ahead of the casualty flow from overseas theaters, and now they looked forward to freeing up almost half of them by November 1945 through an aggressive policy of furloughs and discharges. It was projected that the downward trend in occupied beds, accentuated by a temporary drop during the Christmas season, would continue. These freed-up beds, however, were not being turned over for civilian dependent use as was p! olicy, but were instead retained until after Japan's surrender. ^137 Until that time, there were two more casualty surges for which Major General Norman Kirk, the Army Surgeon General, had to plan.

Concurrent with the steep increases in Selective Service inductions, the Army's expansion of its training facilities and replacement pool, and the Army Medical Department's plans to clear the decks in anticipation of new casualty surges in 1946, was an increased War Department focus on the related questions of casualties and replacements. In June, Stimson directed two of General Arnold's civilian advisors, Dr. E. P. Learned and Dr. Dan T. Smith, to review the replacement system in order to make recommendations on how it could be made "fully effective in the war against Japan," and the Learned-Smith Committee was formed. <page 568> Meanwhile, Stimson advisor Dr. Edward L. Bowles, had Dr. William B. Shockley from the War Department examine the casualties question independently of the Army, but using classified data gathered by its Military Intelligence Division and Medical Corps. Shockley, who would later win a Nobel Prize for Physics and be instrumental in ushering i! n the age of micro-electronics through his development of the transistor, was on loan to the War Department from the Navy where he served as director of research for the Antisubmarine Warfare Operations Research Group.

When the results of the Learned-Smith Committee's study were made available in late June, Army Ground Forces found themselves in general agreement, and were relieved to find that the committee agreed with the current Army policy of producing replacements "against maximum requirements [author's emphasis] rather than against continually revised estimates of minimum needs." ^138 As for Dr. Shockley's initial report to Dr. Bowles, it was not submitted until after Stimson had left for Potsdam. He proposed that a study be initiated "to determine to what extent the behavior of a nation in war can be predicted from the behavior of her troops in individual battles." Shockley utilized the analyses of Dr. DeBakey and Dr. Beebe, and discussed the matter in depth with Professor Quincy Wright from the University of Chicago, author of the highly-respected A Study of War; and Colonel James McCormack, Jr., a Military Intelligence officer and former Rhodes Scholar who served ! in the OPD's small but influential Strategic Policy Section with another former Rhodes Scholar, Colonel Dean Rusk. Shockley said:

"If the study shows that the behavior of nations in all historical cases comparable to Japan's has in fact been invariably consistent with the behavior of the troops in battle, then it means that the Japanese dead and ineffectives at the time of the defeat will exceed the corresponding number for the Germans. In other words, we shall probably have to kill at least 5 to 10 million Japanese. This might cost us between 1.7 and 4 million casualties including [between] 400,000 and 800,000 killed." ^139

<page 569> The war ended before Shockley's proposal could be considered, and too early to furnish any worthwhile comments on the effectiveness of the Army's replacement system in the now "one-front" war.

This was the situation as Truman prepared to leave for his meeting with Stalin and Churchill. The night before the 18 June meeting, Truman had written in his diary that the decision whether to "invade Japan [or] bomb and blockade" was his "hardest decision to date," ^140 but the functional result of the meeting was that the President had decided to do both--- temporarily. Olympic was given the go-ahead, the massive redeployment of forces from Europe for Coronet would continue, and the already implemented blockade and bombing campaign would intensify (but with the primary objective of supporting land operations instead of forcing a surrender), with the casualties question a continual worry. George M. Elsey was then the "watch officer" for the White House Map Room where the progress of the war was graphically charted and updated daily. He noted the close attention paid to growing Japanese troop strength, and remembers "Admiral Leahy d! iscussing the invasion plans with the President in the Map Room prior to our departure for Potsdam." Elsey emphasized "the concern they both had as to the size of the Japanese forces available to oppose us," and that during the course of many conversations with Truman that fateful summer, Truman made it very clear that he "was deeply worried about the casualties that would inevitably be incurred in an invasion." ^141

President Truman started his long journey to Potsdam, Germany, on 6 July, and the conference opened eleven days later. He met with General Marshall on at least four occasions1^42 and later told Air Force Historian James Lea Cate:

"I asked General Marshall what it would cost in lives to land on the Tokyo [P]lain and other places in Japan. It was his opinion that such an invasion would cost at a minimum one quarter million casualties, and might cost as much as a million, on the American side alone.^143

Truman's recollection has been the focus of much debate by historians even though it is somewhat beside the point, since what Marshall is reputed to have said was in line with current Army thinking and the long-implemented manpower policy of 1945. The documents associated with Truman at the Potsdam Conference do not pass the muster of some <page 570> scholars today, and the lack of specificity in the conference log and notes from the less-formal meetings are often used to buttress the contention that expectations by military and civilian leaders of huge losses during the invasions were a "postwar creation." Truman's own shorthand manner of recounting events certainly adds to the confusion, and, consequently, when he said that Marshall's comments were made during a meeting with his senior advisors--- specifically Stimson, Leahy, Marshall, Eisenhower, King, and the new Secretary of State, James Byrnes--- Truman is quite possibly referring to a meeting that might or mi! ght not have happened in the way he stated.

In the conglomeration of notes making up Truman's "diary," and the great volume of letters he wrote to his wife, family, and friends from the 1930s through the 1950s, Truman repeatedly "simplifies" events in a way that anyone making a close comparison of the material with ongoing events might find somewhat nerve-racking. Unlike the virtual day-by-day accounts of Presidents Dwight Eisenhower and Jimmy Carter, things that occurred are frequently not commented on for days or even weeks later, and when they do make their appearance in Truman's hastily written notes, they are often phrased in a way that gives the impression that they had just happened. Consultations with a variety of individuals, that White House logs clearly demonstrate took place over the space of weeks and even months, can show up as occurring in one session, and events of seemingly great importance many years later are found to be mentioned in little more than a cast-off phrase if at all. Further complicat! ing matters is the fact that such items are interspersed among commentaries of events that follow demonstrably accurate timelines and varying levels of detail. In short, Harry Truman wrote about whatever was important to him at the time of the writing. This highly personal material was written for himself and his family, and it is clear that he frequently used his diary as a vehicle to play with ideas for later use in speeches, or to simply get things off his chest. While such material provides valuable insights into Truman's views and makes fascinating reading,^144 it also opens the door to misinterpretations, misunderstandings, and misuse.

For example, it has been suggested that it is "unlikely" that a "formal" meeting between Truman and his advisors took place on casualties and the atom bomb based on the undeniable fact that "none of the available diaries for Potsdam, including those by Leahy, Arnold, and Truman, <page 571> as well as those by Stimson and McCloy, mentions such a meeting." ^145 Virtually none of the diaries in question, however, are in any way comprehensive, and are frequently quite sketchy. Moreover, there are two very important points to remember: First, the proposition that American casualties for the invasion of Japan could parallel those suffered in the struggle against Germany was not new. It had been discussed at some length for nearly a year and with particular intensity in May and June. Although this subject is of great interest to some scholars today, it was, in many ways, "old news" to Truman and his principal advisors by Potsdam in the sense that it was already accepted th! at casualties would be extremely heavy. Second, the successful test of the first nuclear device was an event of immense proportions, but not only are there no direct references to atomic weapons during Potsdam in these individuals' diaries, there are none, except the occasional cryptic reference to "S-1," ^146 and Truman's own notes from July 25,^147 at any time before Hiroshima either. The matter was, after all, top secret. Although it is sometimes hard for today's researchers to understand, the classification "top secret" was--- and still is--- taken very seriously by individuals responsible for the lives of Americans going into harm's way, and diary references do not start pouring out until after the 6 August public release of information on the 5 August destruction of Hiroshima.^148

An examination of documents from the Potsdam Conference and the Log of the President's Trip to the Berlin Conference shows that the President had numerous formal and informal occasions where he met with Stimson, Leahy, Marshall, Eisenhower, King, and Byrnes, singly and in groups, with Eisenhower most frequently absent because of other duties.^149 In his letter to Professor Cate, Truman said that he called all of these gentlemen together for "a meeting" to discuss the atom bomb. But <page 572> while the Log is silent on the specifics of the meeting, and even who was in attendance, the date and time is easily deduced through a process of elimination.

The President was first made aware of the successful test of an atomic device in the New Mexico desert on the evening of 16 July. There was no period before the morning of 18 July, after additional details of the size and scope of the explosion had become available on the seventeenth, that a meeting with his advisors could take place in a discreet, unobtrusive manner. Truman's diary entry for the eighteenth says only that he had breakfast (customarily at 8 A.M. or a little earlier), and then had lunch with Churchill at 1:30 P.M.^150 And while there are many portions of the President's time that are essentially undocumented in the Log, the morning of 18 July is particularly vague and very extended, with the Log stating only that he had breakfast, and then gives no other listing until 1:15 in the afternoon other than: "The President conferred with the Secretary of State and a number of his advisors during the forenoon." ^151 Similarly, the earlier delivery of the 16 July me! ssage on the success of the nuclear test was noted only as "2200: Mr. Davies returned to the Little White House to deliver an urgent message to the President." ^152 Applying the same logic to this entry of 16 July that some historians have used to dismiss the existence of the much more specific reference to a meeting between "the President [and] his advisors" on the eighteenth, would lead to the erroneous conclusion that Truman was never told of the New Mexico bomb test.

[[[ Portions of this study relating to the Postsdam Conference were used in the subsequent D. M. Giangreco Ė Kathryn Moore book Dear Harry . . .Trumanís Mailroom, 1945-1953: The Truman Administration Through Correspondence with "Everyday Americans." The authors substantially modified the first paragraph on page 572 (above) to read as follows in three paragraphs:

<DH473> "Truman was first made aware of the successful test of an atomic device in the New Mexico desert on the evening of July 16, but no mention can be found in the Log of Stimsonís delivery of the message at approximately 7:30 P.M. There was also no time before the morning of July 18, after additional details of the explosionís size and scope had become available on the 17th, that a meeting with his advisors could have taken place in a discreet manner. And though there are many portions of Trumanís time that are essentially undocumented in the Log, the morning of July 18 is particularly vague and very extended. The document states only that he had breakfast (customarily at 8:00 A.M. or a little earlier), then gives no other timed listing until a reference for 1:15 in the afternoon, when he left for a luncheon appointment with Prime Minister Churchill. Likewise, Trumanís diary entry for the 18th says only that he had breakfast, then had lunch with Ch! urchill at 1:30 P.M. There <DH474> is, however, an important Log entry between breakfast and lunch. It reads: "The President conferred with the Secretary of State and a number of his advisors during the forenoon." ^14 Moreover, it was from this meeting that, according to Churchill, Truman brought "the telegrams about the recent experiment" which were discussed during a lengthy, private session along with other bomb-related matters.^15

"At this point it is useful to remember that the earlier delivery of news on the successful--- and highly secret--- nuclear test was not noted in the Log at all. Applying the same logic that has been used to dismiss Marshallís million-casualties statement during what the Log refers to as a meeting between "the President [and] his advisors" to the absence of virtually any Logreference to Truman learning of the New Mexico test from Stimson, would lead to the obvious--- and erroneous--- conclusion that the President had never been informed of this momentous event.

"Further evidence that the meeting was not concocted for the benefit of Professor Cate in 1953 also comes from a Truman diary entry on December 2, 1950. In the midst of Chinaís entry into the Korean War and the grim conflict with MacArthur over the direction military and foreign policy should take, the president wrote: "Now MacArthur is in serious trouble. We must get him out of it if we can. The conference [with Dean Acheson and Omar Bradley] was the most solemn one Iíve had since the Atomic Bomb conference in Berlin. We continue it in the morning. It looks very bad."

14. Lt. William M. Rigdon, USN, Log of the Presidentís Trip to the Berlin Conference, July 6, 1945 to August 7, 1945 (Washington DC: Office of the President, 1945), 23. Extracts from the Log are printed in U.S. Department of State, Foreign Relations of the United States, Diplomatic Papers: The Conference of Berlin (The Potsdam Conference), 1945, 2 vols. (Washington, DC: GPO, 1960) 2: 4-28, See also Naval Aide Files, HSTL.

15. Martin Gilbert, Winston S. Churchill, vol. 8, "Never Despair," 1945-1965 (Boston: Houghton Mifflin Co., 1988), 66.

Truman diary entries were not footnoted in Dear Harry. END INSERT ]]]

But if one wishes to discard the Log account for 18 July 1945, there is also another entry and comments in Truman's diary that are of interest. On the morning after the 24 July plenary session with Marshall and other senior U.S. and British leaders, Truman met again with Marshall when the general joined a meeting, already in progress, with the British Commander in Chief in Southeast Asia, Lord Louis Mountbatten.^153 In Truman's notes covering 25 July, the President said that he "had Gen. Marshall come in and discuss with me the tactical and political situation," adding "he is a levelheaded man." In a second set of notes covering the twenty-fifth, he referred to the meeting as "a most important session" and wrote extensively on the recent successful atomic test, <page 573> which he indicated that he, Marshall, and Mountbatten discussed at some length.^154

George Elsey recalled that "the war was reaching its climax and we were all on edge." ^155 Truman, meanwhile, had never really received an answer at the 18 June meeting to his question of the number of casualties expected during an invasion of Japan. He understood the planning methodologies and Marshall's insistence on the use of ratios, but by now Truman had exchanged his field cap for a civilian fedora. At some point in his discussions with the general, it may have simply come down to the point where the Commander in Chief asked the same question that any president would ask his senior military advisor: "Well General, what do you really think?" While it is unfortunate that Truman's diary entries do not contain more specificity, when the President states that he and Marshall discussed the "tactical situation," it is useful to remember that there was only one tactical operation in the offing during the summer of 1945, and that was the invasion of Japan.

The question of specifically what was discussed at these different meetings--- and whether these meetings even took place--- did not come up until long after the deaths of the participants. One historian has ventured the opinion that President Truman's letter to Professor Cate, where he recounts Marshall's educated guess that casualties could have reached a million men, is not a "reliable source" because the 1,000,000 figure was not in the first draft, it was added at the suggestion of a White House staffer, and was made "long" after the war in 1952. On the other hand, McGeorge Bundy's vague 1988 statement that "defenders of the use of the bomb, Stimson among them, were not always careful about numbers of casualties expected" is characterized by the same historian as authoritative, and presented as proof of Stimson's and Truman's duplicity.^156

Although the image that many carry of Truman is that of a chief executive who frequently shot from the hip in his oral and written statements, it was not an unusual occurrence for him to have his staff read over his hastily penned drafts and offer their suggestions. Some were taken. Some were not. In the original draft of the Cate letter, Truman recounted only the "minimum"^157 number of expected casualties that Marshall gave him as part of a strategic analysis--- which happened to be 250,000 men--- and made no reference to any maximum. His former Secretary of War, on the other hand, had publicly recounted a maximum <page 574> figure, stating that he (Stimson) had been advised that the casualty figure "might" exceed 1,000,000 men. ^158

Presidential assistants Kenneth Hechler and David Lloyd felt that running a maximum figure along with the minimum was important, and a memo was forwarded to Truman that, among other things, reminded him of Stimson's statement.^159 That Truman received this reminder by young staffers, who had not sat in on any meetings between him and Marshall, is not, as has been proposed,^160 something which can be used to either prove or disprove what Truman and Marshall discussed in a private meeting. Neither does it alter the fact that Truman personally approved the addition to his letter, which credited Marshall as the source, and used the number and attribution in his memoirs as well, albeit in a more exaggerated and rhetorical fashion.^161

It is also important to note that Marshall certainly never refuted Truman's statement even in an oblique way. What he did say was that conquering Japan by invasion would have been "terribly bitter and frightfully expensive in lives and treasure." He said that claims the war would have ended soon, even without the use of atomic weapons "were rather silly," and maintained that "it was quite necessary to drop the bomb to shorten the war," adding "I think it was very wise to use it." ^162

The Run-up

As American leaders met with their counterparts from the Soviet Union and Great Britain at Potsdam, Sixth Army and MacArthur's headquarters prepared for Olympic. Colonel Douglas B. Kendrick (who later became MacArthur's personal physician and the Chief of Walter Reed Hospital's Division of Surgical Physiology) produced revised casualty projections in a logistics analysis based on Sixth Army's direct experience on the large Philippine islands of Luzon and the somewhat more mountainous Leyte (which, although smaller than Kyushu, had very similar terrain), and then factored in the loss rate from the recent fighting on Okinawa and the estimated number of enemy divisions Sixth Army would have to face. Battle casualties for the first thirty days were projected to be 22,576 and to increase to 33,330 during the second thirty days^163 as substantial Japanese reinforcements from the northern part of <page 575> the island were brough! t into the fight. Since the whole purpose of Olympic was to establish a massive support base for the invasion of the Tokyo area on Honshu, Sixth Army would then have to push the defending forces far enough up into the mountains to ensure that Japanese artillery and counterattacks could not threaten the air base construction and naval anchorage supporting Operation Coronet.^164

This movement to a difficult-to-support "stop line" in the mountains was expected to consume an additional 69,029 men over the next two months, bringing total Army and Marine battle casualties on Kyushu to 124,935 by the time that the U.S. First and Eighth Armies launched Coronet, 560 miles to the northeast in March 1946.^165 Once on the defensive, Sixth Army casualties were likely to drop below 10,000 per month for the duration of the war whose end, for logistics purposes, was targeted for 15 November 1946.^166 It was estimated that the full range of nonbattle casualties would pump the total up to 394,859 by the end of the fourth month--- effectively 100,000 casualties per month for the ground forces on Kyushu. Thanks to decisions early in 1945 to increase Selective Service call-ups and expand the Army Ground Forces training base, there would be adequate replacements for this rate of loss, es! pecially when one considers that a portion of these casualties would eventually return to duty. However, the ability to keep units up to strength in the fighting on Honshu was more problematical since this loss rate, applied to not one but two U.S. field armies, would outpace the 100,000-men-per-month replacement stream for at least several months after Coronet was launched in 1946. Naval losses were not a factor in these estimates.^167

In addition to these logistics projections made by MacArthur's staff, which made more sense militarily than the ones Marshall was shown before the 18 June meeting, MacArthur's G-2 (intelligence) chief had decided to use the results of the fighting on Okinawa as the basis for a simple formula, already noted, when making tentative casualty projections for strategic calculations: "two to two and half Japanese divisions [could] extract . . . approximately 40,000 American [battle] casualties on land." General Willoughby called this "the sinister ratio," and used this to extrapolate the potential destructiveness of the Japanese combat divisions and division equivalents (where a number of smaller independent <page 576> formations, taken together, were judged to potentially have the "equivalent" strength of a combat division).^168

MacArthur had originally expected the Sixth Army to face from six to ten enemy divisions on Kyushu by the time the invasion was launched near the end of 1945. By summer, however, it was becoming increasingly clear that, after three years of island fighting, the Japanese understood American requirements and correctly deduced that southern Kyushu would likely be the next stepping stone on the drive toward the heart of their empire. When interrogated after the war, a general staff officer charged with divining U.S. intentions simply noted that an invasion of Kyushu made "strategic common sense." ^169 Decryptions of intercepted radio transmissions, largely confirmed or supported by air reconnaissance, displayed an alarming buildup to intelligence staffs in Washington and Manila even as the Joint Chiefs confidently told Truman that Kyushu could be cut off from significant reinforcements from Honshu once U.S. airpower on Okinawa could be ! built up. The Japanese feared this as well and moved quickly to transfer troops and supplies to the threatened area. By combining reinforcements with the existing garrison and indigenously raised units, troop strength on the island climbed from four-plus division equivalents at the end of April to six one month later. By the end of June it was up to MacArthur's projected maximum of ten, and reached thirteen-plus division equivalents by the third week in July "with no end in sight." ^170

MacArthur's intelligence chief warned:

"The rate and probable continuity of Japanese reinforcements into the Kyushu area are changing the tactical and strategical situation sharply. . . . Massing in present attack sectors is evident. Unless the use of these [communications] routes 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 <page 577> the Japanese will enjoy complete freedom of action in organizing the area and in completing their preparations for defense." ^171

Applying Willoughby's "sinister ratio" to this force supplied a range of possible U.S. battle casualties running from approximately 210,000 to 280,000 during the push to the stop line, but the intelligence chief rounded this down to a conservative estimate of "200,000" battle casualties inflicted by the thirteen to fourteen divisions on Kyushu.^172 This figure implies nearly 500,000 losses to all causes during Olympic, although roughly 50,000 of these men would return to duty after light to moderate care. Unfortunately, Japanese troop strength continued to soar, and when Willoughby referred to these widely disseminated numbers given in a 28 July intelligence analysis, he neglected to note that an additional four-plus divisions had appeared by the time the atom bombs were dropped on Hiroshima and Nagasaki.^173

A preliminary analysis of possible alternative invasion sites was initiated,^174 but the clock was running down. If the war was to have a reasonable chance of being concluded by the end of 1946, a powerful invasion force had to be landed within immediate reach of Tokyo before the spring monsoons set in, and that lodgement could not be made until substantial ground-based airpower could be established within range of the target area because air operations from carriers, while impressive on paper, were not nearly enough to support two American field armies in contact with a determined enemy.^175 On 7 August, Marshall again asked MacArthur for a "personal estimate" of the intelligence information and inquired about alternate invasion sites including striking directly at the Tokyo area.^176 MacArthur replied on 9 August that the buildup might be a "deception"! (it was not) and noted that there were indications (which actually were part of a deception) that Japanese airpower was no longer a threat. He reminded Marshall that the factors weighing against the alternative sites in the Home Islands had not changed and said:

"In my opinion, there should 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 our forces to the northward into the industrial heart of Japan. The plan is sound and will be successful." ^177

<page 578> The following day, 10 August 1945, the Japanese government announced its qualified acceptance of the surrender terms spelled out in the Potsdam Declaration. Five days later, Emperor Hirohito broadcast a speech to his nation announcing complete acceptance of those terms and President Truman immediately ordered the suspension of all offensive operations against the Japanese Empire.

Would the unanticipated buildup of Japanese forces in southern Kyushu have caused the cancellation of Olympic? No. An almost ad hoc assault directly toward Tokyo in the November-December 1945 period was a high-risk proposition. While it presents a tempting subject for speculation after the fact, such an operation could not have been adequately supported by either airpower or assault shipping, and thus, risked a costly stalemate on the ground and the failure of stated war aims. Two possible invasion sites north of Tokyo suffered from similar problems, while other sites between Tokyo and Kyushu carried a series of liabilities running from a lack of appropriate loiter time for attacking U.S. aircraft flying the long haul from their Okinawan bases, to insufficient anchorage space once a hard-won lodgement had been obtained. The fact that alternatives were beginning to be explored should not be construed to mean that Olympic would be abandoned.

Marshall was always willing to examine innovative solutions and frequently prodded his staff and commanders to think big. In fact, he had speculated in the summer of 1944 that a lunge to Kyushu--- bypassing the Philippines and Formosa--- might be possible and championed the idea in the face of Navy and Operations Division skepticism until the realities of the fall 1944 "shipping crisis" intervened.^178 Marshall had also actively pushed for an all-or-nothing <page 579> parachute drop near Paris as the main focus of the D-day invasion in Europe until the senior commanders in that theater, Omar N. Bradley, Sir Bernard L. Montgomery, and Eisenhower, demonstrated that the risks outweighed the possible gains and unanimously voiced their opposition.^179 By the summer of 1945, the file cabinets at the Pentagon were crammed with folders containing unrealized plans and staff studies that went nowh! ere.

For better or worse, Kyushu was indeed the best of a poor set of options. Moreover, none of the four possible invasion sites, in addition to Tokyo and Kyushu, offered an easy victory. Said Willoughby: "Each one of these areas had the potentiality of another Okinawa," and estimates for initial American battle casualties ranged from 30,000 at Sendai to 80,000 on Shikoku when examining only the Japanese units stationed in these possible target areas during the summer of 1945.^180 What the troop strength at these sites would be by November-December was anyone's guess given that the Japanese were more adept at reinforcing threatened areas than was originally anticipated (a disturbing precursor to what U.S. forces would find in future wars on the Asian mainland), and the time available for Japan's mobilization of manpower would give them adequate troops at all but the northernmost option,^181 to slo! w the movement of an invasion force to a crawl as they had on Okinawa and Iwo Jima.

<page 580> Use of poison gas against Japanese defenders was a very real possibility,^182 but it was the existence of the new "big bombs" that seemed to offer the best way to offset the growing imbalance of forces in the invasion area. Shipment to the Pacific of components for a third atom bomb to be used against Kokura or Niigata was halted on 9 August to await further developments, but production in the United States continued unabated.^183 If President Truman was forced to continue using nuclear weapons and the dropping of these bombs on cities failed in its strategic purpose of stampeding the Japanese into an early surrender, Marshall was interested in using the growing stockpile tactically to support the Olympic landings. It was believed that seven bombs would be available in time for this initial invasion operation.^184

A more complete appreciation of the dangers posed by nuclear radiation was still years away, and plans called for the majority of the atom bombs--- approximately three for each corps zone of advance--- to be dropped on beach defenses,^185 with hideous consequences for the hundreds of thousands of U.S. soldiers and Marines who would soon pass directly through the devastated areas after landing, and the tens of thousands more men using the same ground for base and airfield construction. It was also estimated that of the 2,200,000 Japanese civilians in the minimum Olympic target area who were unlikely to be evacuated to the north with the retreating Imperial Army, or to have been killed in the preinvasion bombardments, perhaps as many as 180,000 would live in internment camps^186 that would have been located on or near the blast sites. In all, several million Japanese and Americans would be direc! tly affected by nuclear fallout or residual blast radiation on Kyushu.

All of this, however, is purely in the realm of conjecture. What can be stated as fact, is that the estimate that American casualties could surpass the million mark was set in the summer of 1944 and was never changed. In the spring of 1945 various planners and senior officers quibbled over the estimate, or facets of it relating to specific operations, but the statistical possibility of a million casualties, combined with the <page 581> experience of combat attrition of line infantry units in both Europe and the Pacific, had already prompted the Army and War Department manpower policy for 1945, and thus, the pace for the big jump in Selective Service inductions and expansion of the training base in the U.S. even as the war in Europe was winding down. Japan had lost its navy, and its cities were being essentially destroyed by U.S. airpower, but this was largely irrelevant to their ability to inflict casualties on American forces with the aim of forcing the U.S. into a! negotiated peace.

Researchers look at the forest of documents created over fifty years ago and almost immediately become lost during their hunt for extreme comments and inconsistencies. The fundamental truth, however, was that the Army and War Department manpower policy of 1945--- in all its aspects--- was established in such a way that the Army could sustain an average of 100,000 casualties per month from November 1945 through the fall of 1946 and still retain relatively fully manned and combat-effective units through its use of new Selective Service inductees and reassigned soldiers from demobilized units. That casualties would be massive was so basic an understanding, that it was functionally a "self-evident truth" held by decision makers at virtually all levels. Little or no paper discussion was required or conducted within the Army, and events beyond its purview rendered an invasion unnecessary.

The Army, as an institution, believed its soldiers would suffer extreme losses during an invasion of Japan, and all its actions in 1945 were based on that assumption. When Specialist Martin J. Begosh of the 1st Armored Division was wounded by a land mine in Bosnia on 29 December 1995, he, like every soldier, airman, sailor, and Marine wounded in Korea, Vietnam, and the Persian Gulf War, received a Purple Heart for valor, a medal minted in preparation for the invasion of Japan in 1945.^187



1. Harry S. Truman, Memoirs, vol. 1, Years of Decision (New York: Doubleday and Co., 1955), 314.

2. Ibid., 417. The language in Truman's memoir is conversational in tone with a strong rhetorical flavor. For a more detailed account of what he was told by Marshall, see his 12 January 1953 letter to Air Force historian James Lea Cate at the University of Chicago where he recounts Marshall's estimate that the two-stage invasion would cost a "minimum" of 250,000 casualties, but that the number "might" go as high as 1,000,000, a figure that was roughly the same as the total American losses it took to defeat Germany. Wesley Frank Craven and James Lea Cate, The Army Air Forces in World War II, vol. 5, The Pacific: Matterhorn to Nagasaki, June 1944 to August 1945 (Chicago: University of Chicago Press, 1953), insert between pages 712-13.

3. Barton J. Bernstein, "A Postwar Myth: 500,000 Lives Saved," Bulletin of the Atomic Scientists, June-July 1986, 38-40; Gar Alperovitz, "Why the United States Dropped the Bomb," Technology Review, August 1990, 22-34; Kai Bird, "The Curators Cave In," New York Times, 9 October 1994, 15.

4. Mary C. Gillett, The Army Medical Department, 1775-1818 (Washington: Center for Military History, United States Army, 1981), 1: 3.

5. Letter of 14 February 1777, in Worthington Chauncey Ford, ed. Journals of the Continental Congress, 1744-1789, vol. 7, 1 January 1777-21 May 1777 (Washington: GPO, 1907), 134; John C. Fitzpatrick, ed. The Writings of George Washington from the Original Manuscript Sources, 1745-1779, vol. 7, 13 January 1777 --- 30 April 1777 (Washington: GPO, 1932), 149-50.

6. Ibid.

7. Letter of 13 April 1777, in Charles Francis Adams, ed. Letters of John Adams Addressed to His Wife, vol. 1 (Boston: Charles C. Little and James Brown, 1841), 212-13.

8. Gillett, Army Medical Department, 1775-1818, 1: 8.

9. Hugues Ravaton, Chirurgie d'armāes ou traite des plaies d'armes feu (Paris: Didot le jeune, 1768), 635-36.

10. Ford, ed., Journals of the Continental Congress, 2: 209-11, 5: 568-71.

11. Thomas H. Palmer, ed., Historical Register of the United States, vol. 3 (Philadelphia: n.p., 1816), 7-9.

12. Joseph H. Ewing, "The New Sherman Letters," American Heritage, July-August 1987, 30. An exasperated Sherman also said: "I have 8 companies of 4th Illinois--- 420 on paper, about 350 in camp--- but to save my life I cannot get even 250 in the Saddle. . . . Hurlbut has 8 cos of 5th Ohio with about the same results."

13. Banished, as in captured and removed from one's native country and home. See Noah Webster, Webster's Dictionary: An American Dictionary of the English Language (Springfield, Mass.: G. and C. Merriam, 1878), 106.

14. Ewing, "New Sherman Letters," 37.

15. Compiled from: Vincent J. Esposito, The West Point Atlas of American Wars, vol. 1 (New York: Praeger, 1959); Shelby Foote, The Civil War, A Narrative, vol. 3 (New York: Random House, 1974); Burke Davis, To Appomattox: Nine April Days, 1865 (New York: Rinehart, 1959); R. Ernest Dupuy and Trevor N. Dupuy, The Encyclopedia of Military History from 3500 B.C. to the Present (New York: Harper and Row, 1970).

16. Gilbert W. Beebe and Michael E. DeBakey, Battle Casualties: Incidence, Mortality and Logistic Considerations (Springfield, Ill.: Charles C. Thomas, 1952), 21. Authoritative statistics from the Southern forces are unavailable since figures normally used for compilations are frequently inconsistent, fragmentary, or nonexistent.

17. Mary C. Gillett, The Army Medical Department, 1818-1865, vol. 2 (Washington: GPO, 1987), 275-76. More than 1,600,000 cases of diarrhea or dysentery were logged by Army medical personnel, with 14.31 percent of every 1,000 soldiers dying from them.

18. Beebe and DeBakey, Battle Casualties, 21. Total killed in action or died of wounds was 379 while in-theater deaths from disease stood at 4,795 with additional deaths due to disease at supporting bases, transit camps, etc., not included.

19. Frank A. Reister, ed. Medical Department, United States Army: Medical Statistics in World War II (Washington: Office of the Surgeon General, Department of the Army, 1975), 7.

20. For example, in Battle Casualties, 76, Beebe and DeBakey wrote: "Plainly, the dividing line between KIA [killed in action] and DOW [died of wounds] is an arbitrary one, depending upon reporting practice with respect to men dying before admission to an aid station, but after receiving aid on the battlefield, and upon the speed with which the mortally wounded can be admitted to aid stations. If the terrain, the tactical situation, or shortages of personnel conspire to keep aid men from bringing in the wounded very promptly, correspondingly more will inevitably be classified as KIA."

21. See also Harvey E. Brown, The Medical Department of the United States Army from 1775 to 1873 (Washington: Surgeon General's Office, 1873); T. Longmore, Gunshot Wounds (London: Longmans Green, 1877); Richard H. Coolidge, "The Vital Statistics of War, as Shown by the Official Returns of the British Army During the War with Russia, and by Those of the United States Army During the War With Mexico," American Journal of Medical Sciences 35 (1858): 86-94; L. C. Duncan, The Medical Department of the United States Army in the Civil War (Washington: GPO, 1907); and Medical and Surgical History of the War of the Rebellion, 2 vols. (Washington: GPO, 1870, 1888).

22. John T. Greenwood, "The U.S. Army Military Observers with the Japanese Army during the Russo-Japanese War (1904-1905)," Army History, Winter 1996, 4-5.

23. Beebe and DeBakey, Battle Casualties, 21, 77.

24. FM 101-10-1/2, Staff Officers' Field Manual: Organizational, Technical, and Logistical Data Planning Factors (Washington: Department of the Army, 1987), vol. 2, 4-1 to 5-26. Estimate tables compiled from World War II and Korean War combat actions are still taught to field grade Army officers, and used as the basis for casualty estimates of intense, large-scale ground combat that remains non-nuclear. See also FM 12-6, Personnel Doctrine.

25. In George C. Marshall, ed., Infantry in Battle (Washington: Infantry Journal, 1939), the future Army Chief of Staff titled the opening chapter "Rules," yet pointedly emblazoned the legend "Combat situations cannot be solved by rule" immediately underneath. He stated in the introduction (vii) that "the leader who would become a competent tactician must first close his mind to the alluring formulae that well-meaning people offer in the name of victory." During Marshall's tenure at the Infantry School, his department heads included Omar Bradley and Joseph W. Stilwell.

26. FM 101-10-1/2, chap. 5, sec. 4, 4-7.

27. Ibid., 4-16.

28. Ibid., chap. 5, sec. 5, 5-1.

29. Ibid., 5-2.

30. Ibid.

31. Ibid., chap. 5, sec. 4, 4-7.

32. History of the Medical Department of the United States Navy in World War II, vol. 1 (Washington: GPO, 1953), 96. See also O. B. Morrison, Jr., and Robert Mazet, Jr., "The Functions of Medical Units in Amphibious Operations," Military Surgeon 98 (March 1946): 204-14.

33. Ray S. Cline, Washington Command Post: The Operations Division, in the series United States Army in World War II (Washington: Chief of Military History, Department of the Army, 1957), 98.

34. Ibid., 101.

35. Ibid., 103-4, 259-61, 265; Maurice Matloff, Strategic Planning for Coalition Warfare, 1943-1944, in the series United States Army in World War II (Washington: Chief of Military History, Department of the Army, 1959), 108-9.

36. Cline, Washington Command Post, 103, 265-66.

37. Ibid., 103.

38. Ibid., 244.

39. Ibid., 268.

40. History of the Medical Department of the United States Navy, 73; Samuel Milner, Victory In Papua, in the series United States Army in World War II (Washington: Chief of Military History, Department of the Army, 1957), 371. Both the 1st Marine Division on Guadalcanal and the 32d Infantry Division on New Guinea started combat with roughly 11,000 men each and had received almost no replacements during the fighting. The division on Guadalcanal produced 13,081 hospital admissions for disease over its four months of combat, while the one on New Guinea produced 7,125 in just over two. Nearly half of all admissions were for malaria.

41. Matloff, Strategic Planning for Coalition Warfare, 1943-1944, 488.

42. History of the Medical Department of the U.S. Navy, 174.

43. JPS 476, "Operations Against Japan Subsequent to Formosa," 6 June 1944. CCS, JCS, JIC, and JPS documents and minutes of meetings are located in the Office of the Chief of Staff files, Record Group 165, National Archives and Records Administration, Washington, D.C. (Hereafter, RG 165, NA.). These records are also in RG 218, NA.

44. JCS 924, "Operations Against Japan Subsequent to Formosa," 30 June 1944.

45. Henry I. Shaw, Jr., Bernard C. Natly, and Edwin T. Turnbladh, Central Pacific Drive: History of U.S. Marine Corps Operations in World War II (Washington: Historical Branch G-3 Division Headquarters USMC, 1966), 345; Philip A. Crowl, Campaign in the Marianas, in the series United States Army in World War II (Washington: Department of the Army, 1960), 264-65. "At times," said the Army history, "the waters below the point were so thick with floating bodies of men, women, and children that naval small craft were unable to steer a course without running over them." A similar situation developed during the conquest of Okinawa, but by this time, civil affairs personnel were prepared for this eventuality. A comparatively small number of Japanese civilians committed suicide, and most were induced to surrender without harming themselves or their families.

46. Daily Summary of Enemy Intelligence, no. 1199, 16-17 July 1945, General Headquarters, United States Army Forces, Pacific, Military Intelligence Section, General Staff. Copy in the Combined Arms Research Library, U.S. Army Command and General Staff College, Fort Leavenworth, Kansas. (Hereafter cited as CARL.)

47. Shaw, Natly, and Turnbladh, Central Pacific Drive, 346. The total garrison probably numbered closer to 30,000.

48. JCS 924/2, "Operations Against Japan Subsequent to Formosa," 30 August 1944, 120. Copies of each annex were sent to individuals and organizations on the original distribution lists. The last annex added to this document was JCS 924/15, 25 April 1945. The Saipan ratio was also used to project casualties for other proposed operations. For example, the OPD estimated that an invasion of Formosa would cost 88,600 American casualties, including approximately 16,000 dead, which Marshall rounded out to "approximately 90,000 casualties" in a 1 September 1944 memorandum for the Joint Strategic Survey Committee chief, General Stanley D. Embick. See Larry I. Bland and Sharon Ritenour Stevens, eds., The Papers of George Catlett Marshall, vol. 4 (Baltimore: Johns Hopkins University Press, 1996), 567-69.

49. The distribution of twenty-two numbered copies noted at the end of JCS 924 and each annex is deceptive in that it does not include individuals and organizations on the "Special Distribution" list.

50. Wesley Frank Craven and James Lea Cate, The Army Air Forces in World War II, vol. 5, The Pacific: Matterhorn to Nagasaki, June 1944 to August 1945 (Chicago: University of Chicago Press, 1953) 525-36. XXI Bomber Command was not designated as an Air Force like the Eighth in England or the Fifteenth in the Mediterranean, even though it was organized as one. Instead, it was part of a "peculiar" command arrangement under the Twentieth Air Force, commanded by Arnold in Washington and Norstad on-site in the Marianas, to both give the unit a degree of independence from theater (Navy) directives, and ensure a close measure of operational control by the JCS.

51. Curtis E. LeMay and MacKinlay Kantor, Mission With LeMay: My Story (New York: Doubleday and Co., 1965), 347.

52. Cline, Washington Command Post, 267.

53. From a summary of a 26 March 1996 telephone interview with George L. McColm, U.S. Navy, Retired. Examined and approved by McColm on 17 April 1996. A highly respected agricultural specialist before the war, dubbed by U.S. Agriculture Secretary Henry Wallace the "Weather Bookmaker," McColm's studies were used as the basis for Office of Strategic Services (OSS) theater map "Crop Growing Season Climate Map of the Japanese Home Islands of Honshu - Shikoku - Kyushu" as early as 1942. He served as Chief of Agriculture on the JCS working group planning for the invasion and occupation of Japan.

54. LeMay and Kantor, Mission with LeMay, 381.

55. V. M. Andronikov, P. D. Burikov, V. V. Gurkin, et al., Grief sekretnosti snyat: Poteri Vooryzhennykh Sil SSSR v voynakh, boevykh deystviyakh konfliktakah: Statisticheskoe issledovanie (Moscow: Voyenizdat, 1993), 219-20.

56. Author telephone interview with John J. Maginnis of Worcester, Massachusetts, December 1982. Maginnis headed the first Allied Military Government detachment in Normandy after the invasion of France, was chief of military government operations in the Berlin districts of Schoneberg and Friedenau, and Chief of Staff, Berlin District Headquarters, representing U.S. interests in the Four-Power Kommandantura governing occupied Berlin.

57. John J. Maginnis, Military Government Journal: Normandy to Berlin (Amherst: University of Massachusetts Press, 1971), 244-45.

58. Maginnis interview.

59. Grace Person Hayes, The History of the Joint Chiefs of Staff in World War II: The War Against Japan (Annapolis, Md.: Naval Institute Press, 1982), 702. This source originated as a classified document generated in 1953 by the Historical Section, JCS.

60. Samuel Halpern to Barton J. Bernstein, 16 July 1990, in The Enola Gay Debate, August 1993-May 1995 (Arlington, Va.: U.S. Air Force Association, 1995), chap. 10, no pg.

61. "Strategic Bombing Symposium on the Atomic Bombing of Japan," Smithsonian Institution, Air and Space Museum, 12 July 1990; author's telephone interview with General Goodpaster, 7 May 1997. In addition to serving as the Supreme Allied Commander in Europe, General Goodpaster was Assistant to the Chairman, Joint Chiefs of Staff, Commandant of the National War College, and Superintendent of the U.S. Military Academy.

62. From numerous interviews in 1978 with maintenance personnel of the bombardment groups (very heavy) stationed at Clovis Field, New Mexico, and other Army Air Force bases in the American southwest from May through September 1945.

63. Halpern to Bernstein, 27 August 1990.

64. Summary of Redeployment Forecast, 14 March 1945, p. 6, Demobilization Branch, Plans and Operations Division, Army Service Forces, RG 160, NA. See also pp. 13 and 24.

65. Hayes, History of the Joint Chiefs of Staff: Japan, 702.

66. Figures frequently stated for ground force "losses" on Okinawa run from 65,631 to 72,000 due to the addition of 26,211 to 33,096 non-battle casualties. U.S. Tenth Army Action Report, Ryukyus, 26 March to 30 June 1945, vol. 1, chap. 11, sec. I, 12; Thomas M. Huber, Japan's Battle for Okinawa, April-June 1945 (Fort Leavenworth, Kans.: Combat Studies Institute, U.S. Army Command and General Staff College, 1990), 119-20. Approximate Navy losses were 4,900 dead and 4,800 wounded, according to Samuel Eliot Morrison, The Two Ocean War: A Short History of the United States Navy in the Second World War (Boston: Little, Brown and Co., 1963), 556. Insight into what most military personnel mean by the word "lost" is essential to understanding minutes of mid- and senior-level meetings, as well as oral history interviews conducted after the war. In general, the term "lost" represents the total number of men deducted from a unit's "effectives," or the men unavailable or incapable of! performing their duties for any of several reasons: casualties severe enough to be evacuated, troops missing or captured, and the dead. Men temporarily assigned to other duties, or wounded but able to return to their unit after a modest amount of medical care, do not figure in, but, nevertheless, can have a severe effect on a unit's present-for-duty strength.

67. Army Battle Casualties and Nonbattle Deaths in World War II, Final Report, 7 December 1941-31 December 1946 (Washington: Comptroller of the Army, Office of the Chief of Staff, 1987). The final total of Army and Army Air Force casualties stands at 945,515, but while this figure includes non-battle deaths, it does not include categories which drained the Army of manpower and were closely monitored by senior leaders such as non-battle injuries in combat zones like loading accidents; losses due to disease and combat related psychiatric breakdowns; and 101,600 American prisoners of war who survived their captivities and were counted as casualties during the war. Although Navy and Marine Corps battle casualties were only 159,495 to 162,668 (depending on how one constructs the numbers), Captain Louis H. Roddis, U.S. Navy, M.D., noted in "Naval and Marine Corps Casualties in the Wars of the United States" (Military Surgeon 99 [October 1946]: 305-10) that these figures were mo! re than eight times the number of killed and wounded among our naval personnel in all the other wars of the United States combined. An additional 12,000+ Merchant Marine and Coast Guard battle casualties, primarily from German submarines, were not included in Roddis's study, but do not alter his basic point. The 12,000 figure includes dead and missing but not combat injuries (burns, wounds, etc.), and there were an additional 30,442 Navy and Marine non-battle deaths. Non-battle deaths are frequently included in battle casualty tabulations because, unlike other categories of non-battle casualties, dead personnel are irrevocably removed from the manpower pool, while a percentage of those in other categories returned to full duty status or were able to serve in a reduced capacity within the United States and thus free healthy personnel for overseas duty.

68. Military Research Bulletin, no. 13, 25 April 1945 (Washington: War Department, Military Intelligence Division, 1945), 5-9.

69. CCS 168, "Conduct of War in Pacific Theater in 1943," 22 January 1943.

70. Cline, Washington Command Post, 337, 342; Matloff, Strategic Planning for Coalition Warfare, 1943-1944, 536.

71. Hayes, History of the Joint Chiefs of Staff: Japan, 702.

72. JIC [Joint Intelligence Committee] 266/1, "Defeat of Japan by Blockade and Bomb," 18 April 1945; JWPC 263/4, "An Outline Plan for Invasion of Kanto (Tokyo) Plain," 5 May 1945; JWPC 264/1, "Strategic Positions Selected for Occupation Upon Japanese Withdrawal, Collapse, or Surr[ender]," 16 May 1945; JWPC 264/2, "Forces Required for Occupation of Strategic Positions in Japan Proper," 16 May 1945; JCS 1331/3, "Directive for Op[eratio]n `Olympic,' Kyushu," 25 May 1945. JWPC papers may be found in RG 319, NA.

73. Daily Summary of Enemy Intelligence, no. 1209, 26-27 July 1945, 8, Military Intelligence Section, General Staff, United States Army Forces, Pacific, CARL.

74. Memorandum for the Record, 11 June 1945, OPD 336, TS, sect. 4, case 119, 1945, box 143, RG 165, NA.

75. Material in Safe File: "Japan," Record Group 107, NA. Truman had met with Hoover on 1 June 1945 and requested his thoughts on ending the war with Japan. Hoover submitted "Memorandum for ending the Japanese War" which Truman sent to Stimson for comment on 9 June. Stimson forwarded the memorandum to the War Department, which responded on 14 June, and returned all materials to the President the following day with his own comments stressing that casualties would be very heavy. For an interesting analysis of this material and the previous War Department analysis of the 1 June request by Stimson, see the Halpern-Bernstein letters cited in footnotes 60 and 63. Truman also sent copies of the Hoover letter for comment to Fred M. Vinson, the Director of the Office of War Mobilization and Reconversion; Acting Secretary of State Joseph C. Grew; and former Secretary of State Cordell Hull (who called Hoover's memorandum "his appeasement proposal"). Their analyses are at the Harry S! . Truman Library (HSTL), Independence, Missouri.

76. Cline, Washington Command Post, 259. In a 9 August 1943 memorandum from General Marshall to General Handy, the Chief of Staff recounted that President Roosevelt told him: "Planners were always conservative and saw all the difficulties, and that more could usually be done than they were willing to admit."

77. Oral history interview with General John T. Hull, 8 April 1974, 22, Senior Officer Oral History Program, vol. 2, sect. 7, tape 1, United States Army Military History Institute, Carlisle Barracks, Pennsylvania.

78. Robert P. Newman, Truman and the Hiroshima Cult (East Lansing: Michigan State University Press, 1995), 37.

79. Dwight D. Eisenhower to John J. McCloy, 18 June 1965, copy provided by Robert James Maddox, Penn State University, University Park, Pennsylvania.

80. Hayes, History of the Joint Chiefs of Staff: Japan, 706.

81. Leahy to JCS members Marshall, Arnold, and King; Secretary of War Stimson; and Secretary of the Navy James Forrestal, Memo SM-2141, 14 June 1945, in U.S. Department of Defense, The Entry of the Soviet Union into the War against Japan: Military Plans, 1941-1945 (Washington: GPO, 1955), 76.

82. JWPC 369/1, "Details of the Campaign Against Japan," 15 June 1945.

83. JCS 1388, "Details of the Campaign Against Japan," 16 June 1945.

84. JWPC 369/1, "Details of the Campaign Against Japan," 5-7.

85. "G-2 Estimate of the Enemy Situation with Respect to an Operation Against the Tokyo (Kwanto) Plain of Honshu," 31 May 1945, CINCAFPAC, 15-18.

86. "Joint Staff Study: Olympic, Naval and Amphibious Operations," Commander-in-Chief, U.S. Pacific Fleet and Pacific Ocean Areas (CINCPOA), 18 June 1945, 70, CARL. The JWPC estimates for total ground force battle casualties from both operations Olympic and Coronet were 193,500 including 43,500 killed and missing, roughly that of the first four months of American battle casualties in France. The Navy's estimate, supplied by CINCPOA, only went out for one month and totaled 49,000 battle casualties. Also see D. M. Giangreco, "Operation Downfall: The Devil Was in the Details," Joint Force Quarterly, Autumn 1995, 86-94. Because of a combination of climate and logistic factors, both Olympic and Coronet would have been delayed by approximately a month each, thus complicating the manpower picture.

87. JWPC 369/1, "Details of the Campaign Against Japan," 7.

88. War Department transmission W-17477, MacArthur Memorial Archive, Norfolk, Virginia. (Hereafter MMA.)

89. Commander in Chief, Army Forces, Pacific (CINCAFPAC) transmission C-19571, MMA.

90. Dr. DeBakey became the Director of the Surgical Consultants Division of the Surgeon General's Office after World War II, and his work led directly to the formation of mobile Army surgical hospital (MASH) units prior to the Korean War. He later invented numerous important medical machines and surgical techniques and became internationally recognized.

91. Beebe and DeBakey, Battle Casualties, 27-28. The severe losses due to combat-induced psychiatric breakdowns prompted many articles by Beebe and DeBakey's contemporaries after the war. Probably the two best articles detailing efforts to lessen the effect on field units are Perry C. Talkington, "Combat Psychiatry: Third United States Army Experience," Military Surgeon 98 (May 1946): 401-4; and Alfred O. Ludwig and Stephen W. Ranson, "A Statistical Follow-Up of Effectiveness of Treatment of Combat-Induced Psychiatric Casualties: Returns to Full Combat Duty," Military Surgeon 100 (January 1947): 51-62. Postwar editions of the Military Surgeon also contain numerous articles on the treatment of neuropsychiatric patients once they had entered Army hospitals.

92. War Department transmission W-18528, MMA.

93. FM 101-10-1/2, chap. 5, sec. 5, 5-9 to 5-11. See also Beebe and DeBakey, Battle Casualties, 27, 266-68.

94. ClNCAFPAC transmission C-19848, MMA.

95. Beebe and DeBakey, Battle Casualties, 61.

96. Army Battle Casualties (Final Report), 32. U.S. Army battle casualties for June (3+ weeks) and July-September 1944 (4+ weeks each) were 39,367; 51,424; 42,525; and 42,183, respectively. The monthly European Theater average from November 1944 through January 1945 was 65,760, and 44,660 from February through April 1945. U.S. Army average monthly casualties in the Mediterranean Theater through this time frame were 7,884, p. 34.

97. Army Battle Casualties (Final Report), 42. April-June 1945: 19,650, 14,981, and 7,427, respectively. This does not include Marine ground forces casualties.

98. Charles A. Willoughby, "Occupation of Japan and Japanese Reaction," Military Review, June 1946, 3-4.

99. Matloff, Strategic Planning for Coalition Warfare, 1943-1944, 485.

100. ClNCAFPAC transmission C-19848, MMA.

101. Francis B. Kish, "Citizen-Soldier Harry S. Truman, 1884-1972," Military Review, February 1973, 30-44.

102. Ibid.

103. Letter of 11 February 1937 to Bess Truman, in Robert H. Ferrell, ed., Dear Bess: The Letters from Harry to Bess Truman, 1910-1959 (New York: W. W. Norton, 1983), 391.

104. David McCullough, Truman (New York: Simon and Schuster, 1992), 229.

105. Alfred Steinberg, The Man from Missouri: The Life and Times of Harry S. Truman (New York: G. P. Putnam's Sons, 1962), 190.

106. 18 June 1945 Joint Chiefs of Staff Corrigendum to JCS 1388, "Details of the Campaign Against Japan," 16 June 1945. It was suggested in Barton J. Bernstein's "Understanding the Bomb and the Japanese Surrender: Missed Opportunities, Little Known Near Disasters, and Modern Memory," Diplomatic History 19 (Spring 1995): 233, that Marshall's opening statement was "virtually a lecture to Truman, the former army captain," and that the "insecure" president "received counsel that few men, and certainly not a newcomer, could have resisted." Marshall's statement is characterized as "clear and pungent: Be tough and decisive, endorse the Kyushu operation, and be a true leader." In reality, Marshall's introductory remarks, while certainly to the point, were of a style and format that the President would find extremely familiar after a nearly thirty-year career as an Army officer.

107. Keijo and Fusan are the Japanese names for Seoul and Pusan on the Korean Peninsula.

108. The AFPAC kill ratio for this period was actually closer to 22.6 to 1. In a statistical freak, the kill ratios of the Leyte and Luzon operations also played out at 22 to 1, although the daily tally of known American and estimated Japanese deaths varied considerably within these campaigns. Aside from the 22 to 1 ratio, taken from the 31 May 1945 Health, 15, were ratios of 8, 9, and 4 to 1 from Saipan, Guam, and Iwo Jima respectively, which were published in that edition but not used by Marshall at the 18 June meeting.

109. For monthly U.S. combat losses in Northwest Europe, see footnote 96. British total (ground force including Canadian and Polish elements) is from War Diary, 21st Army Group, `A' Section, SITREP, 29 August 1944 and (Royal Air Force) from L. F. Ellis and A. E. Warhurst, Victory in the West, in the United Kingdom Military Series (London: HMSO, 1962), 1: 488. See also United Kingdom, Central Statistical Office, Statistical Digest of the War (London: HMSO and Longmans Green, 1951), 11, 13.

110. For an analysis of the weather's critical impact on the timing of the two invasion operations, see Giangreco, "Operation Downfall."

111. "Staff Study, `Olympic,' Operations in Southern Kyushu," 28 May 1944, Appendix B, Annex 4, CINCAFPAC, CARL. What this figure does not include is the land, sea, and air forces operating in direct combat support of the Kyushu invasion from the Philippines, Okinawa, Iwo Jima, the Marianas, Ulithi, or the sea, nor does it include combat divisions and other units involved in garrison duty, formations specifically earmarked for employment in the follow-up invasion on Honshu, or the support infrastructure at bases further to the rear such as those in Hawaii, the ports of embarkation, etc. It also does not count the combat division in "strategic reserve" on Okinawa, but does include the one held in "floating reserve" during the landings since it was slated to be committed at the discretion of the Sixth Army commander.

112. Huber, Japan's Battle for Okinawa, 119-20. Extrapolated from 10th Army, G-1 Periodic Reports, no. 13, 1-3.

113. Harry S. Truman diary, 1 June 1945, HSTL.

114. John J. McCloy, The Challenge to American Foreign Policy (Cambridge, Mass.: Harvard University Press, 1953), 42-43.

115. For the full text of the 2 July 1945 memo, see Henry L. Stimson, "The Decision to Use the Atomic Bomb," Harper's Magazine 194 (February 1947): 97-107.

116. Ibid., 102.

117. JCS 1388/1, "Details of the Campaign Against Japan," 20 June 1945, 21. Nimitz estimate in Joint Staff Study: Olympic, Naval and Amphibious Operations, 70.

118. JCS 1388/2, "Details of the Campaign Against Japan," 26 June 1945, 24.

119. JCS 1388/3, "Details of the Campaign Against Japan," 2 July 1945, 25.

120. Huber, Japan's Battle for Okinawa, 119.

121. JCS 1388/4, "Details of the Campaign Against Japan," 11 July 1945, 34.

122. "Japanese Military Manpower," Military Research Bulletin, no. 13, 25 April 1945, 5-9.

123. "Japanese Military Manpower Potential for 1945," Military Research Bulletin, no. 18, 4 July 1945, 1-4.

124. "The Japanese System of Defense Call-up," Military Research Bulletin, no. 19, 18 July 1945, 1-3.

125. The total number of Japanese armed forces personnel demobalized by the U.S. Military Government after the surrender was 6,465,435, including 784,047 on Kyushu. MacArthur in Japan: The Occupation: Military Phase, vol. 1, supplement, p. 266, from the Reports of General MacArthur (Tokyo: SCAP, 1950).

126. Cline, Washington Command Post, 348; Martin Halliwell, "The Projected Assault on Japan," Military Review, January 1949, 77-80. See also C. A. Bartholomew, Mud, Muscle, and Miracles: Marine Salvage in the United States Navy (Washington: Department of the Navy, 190), 173-79, 449-52. Bartholomew recounts the Japanese "thoroughly professional job blocking [Manila] harbor and rendering the port useless" through the sinking of "an estimated 750 ships, barges, and associated craft." To answer the challenge, the U.S. Navy produced seven "unique" salvage ships, in two classes, especially for Operations Coronet and Olympic.

127. Robert R. Palmer, Bell I. Wiley, and William R. Keast, The Procurement and Training of Ground Combat Troops, in the series United States Army in World War II (Washington: Department of the Army, 1948), 216-25, tables on 172 and 177.

128. Stimson, "Decision to Use the Atomic Bomb," 102. Army Service Forces estimated that deployed U.S. Army strength in Japan and its environs by the end of 1946 would stand at 3,243,000 after deductions of approximately 720,000 dead and evacuated wounded. Navy and Marine personnel were not included in the figure, nor were Army personnel in supporting theaters. Summary of Redeployment Forecast, p. 6, Demobilization Branch, Planning Division, Army Service Forces, 14 March 1945. Copy in CARL.

129. It was initially feared that the 16 December 1944 German counterattack in the Ardennes Forest would force a large and unforeseen expansion of Selective Service call-ups at precisely the time that the manpower barrel was practically scraped clean, but it quickly became apparent that the short-term personnel crisis brought on by the counterattack was not as great as it might have been, and would largely be solved by the culling of excess Zone of the Interior personnel in the European Theater. Unfortunately this expedient exhausted the European Theater's surplus manpower; thus, to ensure adequate replacements in case the war in Europe extended into the summer, a modest increase of 20,000 men per month (80,000 total per month) was made in the January and February 1945 call-ups. These increases, however, represented only a troublesome blip in the Army's manpower projections for 1945, and the 3 January telegram from Selective Service Director Lieutenant General Louis B. He! rshey to the State Selective Service Directors tied proposed and directed changes in various deferments to long-term needs rather than a passing crisis. The message read in part: "The following letter from the Director of the Office of War Mobilization [James F. Byrnes] has been received by the Director of Selective Service: `The Secretaries of War and Navy have advised me jointly that the calls from the Army and Navy to be met in the coming year will exhaust the eligibles in the 18 through 25 age group at an early date. The Army and Navy believe it is essential to the effective prosecution of the war to induct more men in this age group.'"

After the Ardennes battle, the general slowdown in fighting allowed infantry strength to be brought up to appropriate levels before the spring offensive, and line units were essentially at full strength by the end of March 1945. This was accomplished not only through the December-January culling but also through the normal replacement stream from the July-August (and, conceivably, a portion of the September) 1944 inductions as well as the acceleration of some troops already in the transit pipeline. A shortage of shipping combined with transport bottlenecks prevented the acceleration from being as extensive as commanders desired. Recruit training at Army Ground Force facilities in the U.S. was not cut short to meet the short-term crisis of the Battle of the Bulge.

130. Selective Service and Victory: The 4th Report of the Director of Selective Service, July 1, 1944 to December 31, 1945 (Washington: GPO, 1946), 591, 596.

131. Palmer, Wiley, and Keast, Procurement and Training of Ground Combat Troops, 221-25.

132. Ibid., 224-25.

133. Selective Service and Victory, 591, 596. Call-ups from the United States and its territories in March 1945 totaled 141,200, but manpower shortages limited actual inductions to only 115,631.

134. A ruvulet of replacements would continue to go to Europe in 1946, but combat forces projected to make up the U.S. share of occupation forces in Europe, even if Nazi guerrilla activities persisted, was to be a maximum of only 1+ functioning division equivalents in Germany by the summer of 1946, with the same amount in the Austria northern Italy area. Combat units in Europe thus made up only a tiny sliver of the projected 1946 manpower pie which included three to four field armies operating in Japan through the end of the year (the Army's thirty-six division equivalents shown in the chart in Giangreco, "Operation Downfall," plus three others from Europe and up to three more from elsewhere in the Pacific). The six Marine divisions slated for Olympic and Coronet are not included.

135. "How Long Will We Have to Fight the Jap War?" Yank 3 (8 June 1945): 2.

136. Palmer, Wiley, and Keast, Procurement and Training of Ground Combat Troops, 224.

137. McKitterick Clarence Smith, The Medical Department: Hospitalization and Evacuation, Zone of Interior, in the series United States Army in World War II (Washington: Office of the Chief of Military History, Department of the Army, 1956), 200-213.

138. Palmer, Wiley, and Keast, Procurement and Training of Ground Combat Troops, 234-37.

139. W. B. Shockley to Edward L. Bowles, 21 July 1945, "Proposal for Increasing the Scope of Casualties Studies," Edward L. Bowles Papers, box 34, Library of Congress. Attached: "Historical Study of Casualties," by Quincy Wright. Dr. Shockley extrapolated these numbers from earlier analyses summarized in Health (see footnote 108), and others compiled by the Military Intelligence Division from field reports. No accurate total of German military and civilian deaths was available at the time he prepared his report, but the number was eventually set at roughly 11,000,000. Japan was not invaded and finished the war with just over 7,000,000 casualties, most of them from its armed services on the Asian mainland in fighting from September 1931 to September 1945.

140. Truman diary, 17 June 1945, HSTL.

141. George M. Elsey to author, 30 March 1996. See also his lengthy introduction to "Blueprints for Victory," National Geographic, May 1995, 55-77, for information on the Map Room.

142. Truman diary, 25 July 1945; William M. Rigdon, Log of the President's Trip to the Berlin Conference, 6 July 1945 to 7 August 1945 (Washington: Office of the President, 1945).

143. Craven and Cate, Pacific: Matterhorn to Nagasaki, insert between 712-13.

144. Ferrell, Dear Bess; Off the Record: The Private Papers of Harry S. Truman (New York: Harper and Row, 1980); The Autobiography of Harry S. Truman (Boulder: Colorado Associated University Press, 1980). See also D. M. Giangreco, Dear Mr. President: Letters to Harry Truman from "Everyday Americans," Stackpole Books, scheduled for release in the fall of 1997.

145. Barton J. Bernstein, letter: "To Be Among Those Numbers," Joint Force Quarterly, Spring 1996, 6-7.

146. Henry L. Stimson diary, 19 June 1945, 51: 183, Yale University Library.

147. Truman diary, 25 July 1945.

148. When the atom bomb exploded, it was 8:15 A.M. on Monday, 6 August in Hiroshima (9:15 A.M. on Tinian where the bomber was launched), and 7:15 P.M. on Sunday, 5 August in Washington, D.C. A report was issued to newsmen at the White House the following morning, and Monday evening papers throughout the country carried the stunning news under banner headlines on 6 August, but the story had come too late to make the Late City Edition of the New York Times. Instead, the Times ran the story in its Tuesday, 7 August editions, and made it a point to make it clear to readers that the atom bomb was dropped "about the time that citizens on the Eastern seaboard were sitting down to their Sunday suppers."

149. Rigdon, Log of the President's Trip to the Berlin Conference, 17-37. Extracts from the Log are printed in U.S. Department of State, Foreign Relations of the United States, Diplomatic Papers: The Conference of Berlin (The Potsdam Conference), 1945, 2 vols. (Washington: GPO, 1960), 2: 4-28. See also Naval Aide Files, HSTL.

150. Truman diary, 18 July 1945.

151. Rigdon, Log, 23.

152. Ibid., 20. Joseph E. Davies was the U.S. Ambassador to the Soviet Union when Joseph Stalin and Adolf Hitler signed the Nazi-Soviet nonaggression pact, was the Ambassador to Belgium when Hitler invaded that country, and was a special advisor to Truman on the Soviet Union.

153. Ibid., 30.

154. Truman diary, 25 July 1945.

155. Elsey letter to author, 12 May 1997.

156. Bernstein letter to Joint Force Quarterly. The quote is from McGeorge Bundy in Danger and Survival: Choices About the Bomb in the First Years (New York: Random House, 1988), 647.

157. Draft of President's letter to James Lea Cate in President's Secretary's File (PSF), HSTL.

158. Kenneth Hechler to David Lloyd, 2 January 1953, memo, PSF, HSTL.

159. David Lloyd to President Truman, 6 January 1953, memo, PSF, HSTL.

160. Bernstein letter to Joint Force Quarterly.

161. Truman, Memoirs, 1: 417.

162. Forrest C. Pogue, George C. Marshall: Statesman, 1945-1959 (New York: Viking Penguin, 1987), 23.

163. "Medical Service in the Asiatic and Pacific Theaters," chap. 15, p. 18, unpublished monograph, Center of Military History, Washington, D.C.

164. Engineers of the Southwest Pacific: Amphibian Engineer Operations (Washington: GPO, 1959), 4: 694.

165. "Medical Service in the Asiatic and Pacific Theaters."

166. Cline, Washington Command Post, 346.

167. "Medical Service in the Asiatic and Pacific Theaters."

168. Willoughby, "Occupation of Japan and Japanese Reaction," 3-4.

169. Alvin D. Coox, "Japanese Military Intelligence in the Pacific: Its Non-Revolutionary Nature," in The Intelligence Revolution: A Historical Perspective (Washington: Office of Air Force History, 1991), 200.

170. "Amendment No. 1 to G-2 Estimate of the Enemy Situation with Respect to Kyushu," G-2, AFPAC, 29 July 1945, CARL. In hindsight, one could argue that the trend of the buildup should have been clear by the 18 June meeting between the JCS members and President Truman, but allowing for the lag time in both the formulation and dissemination of the intelligence information, the troop-strength figures available to the individual JCS members would have been anywhere from about five days to two weeks old at the time of the meeting. Once the matter of Japanese reinforcements had been flagged as an area of great concern (as it almost certainly was by July), extraction, analysis, and dissemination of the information could be accomplished within a two-day cycle or even less.

171. Ibid.

172. Willoughby, "Occupation of Japan and Japanese Reaction."

173. Edward J. Drea, "Previews of Hell: Intelligence, the Bomb, and the Invasion of Japan," Military History Quarterly 7 (Spring 1995): 74-81.

174. JWPC 397, "Alternatives to OLYMPIC," 4 August 1945.

175. Giangreco, "Operation Downfall."

176. War Department transmission WAR-45369, MMA.

177. CINCAFPAC transmission C-31897, MMA.

178. Bland and Stevens, eds., Papers of George Catlett Marshall, 4: 492-94, 567-68, 690-92; Cline, Washington Command Post, 444-45; Robert W. Coakley and Richard M. Leighton, Global Logistics and Strategy, 1943-1945, in the series United States Army in World War II (Washington: Office of the Chief of Military History, Department of the Army, 1968), 551-62; M. Hamilton Cannon, Leyte: The Return to the Philippines, in the series United States Army in World War II (Washington: Office of the Chief of Military History, Department of the Army, 1954), 6-7, 34, 37-39; Michael D. Pearlman, Unconditional Surrender, Demobilization, and the Atomic Bomb (Fort Leavenworth, Kans.: Combat Studies Institute, 1996), 10-11. The Chief of Staff's suggestions for a quick thrust to Kyushu (no formal proposal was put forward) met with Navy and OPD reticence, and his armchair admiraling was disregarded by King and Nimitz who believed that such an operation--- in the fall of 1944--- was not sustai! nable because of the distance from existing air bases and logistic centers. So distant, in fact, that substantial carrier assets would be required to provide ongoing air cover for the lodgment, yet only be able to support it in task group shifts after the initial assault because of the Navy's own logistic problems. Meanwhile, there was also an inadequate number of Army engineer units available for building airfields on Kyushu--- a shortage which did, in fact, contribute significantly to Leyte not becoming the major center for offensive air power it was envisioned to be (see Giangreco, "Operation Downfall"). Thus, the Navy would be forced into a position where it would be unable to mass its airpower for an unknown, but lengthy, period at the beginning of the land campaign. Such a situation, where substantial carrier assets would be tied up protecting the lines of communications, risked the destruction in detail of the separate task groups operating close to the Japanese coas! t, and those same dispersed task groups could not be depended on to ef fectively support land operations, which would be completely dependent upon the naval air arm to prevent Japanese reinforcements from arriving (unlike the Philippines, Kyushu was immediately adjacent to centers of Japanese manpower). Marshall acknowledged the "question of security of our lines of communications [and] the logistical complications," but he believed that it was worthwhile to examine a direct thrust to Kyushu as an option that might, in the long run, be less costly than a lengthy, set-piece approach to Japan.

179. Roger Hand, "Overlord and Operational Art," Military Review, May-June 1995, 86-91.

180. Willoughby, "Occupation of Japan and Japanese Reaction."

181. The general dearth of population in the Aomori Prefecture would work to the advantage of the Sixth Army landing in the Ominato (Mutsu Bay) area, but its extreme distance from U.S. air bases on Okinawa and Iwo Jima meant that an invasion there would have to be supported almost exclusively from carriers operating in gale-swept northern waters. The area had also not been thoroughly examined to see if it would make an appropriate base from which to launch an assault toward Tokyo, and it was too far removed from the Tsushima Straits to act as an effective barrier against Japanese reinforcements being sent from Korea and China to Japan.

182. "Preparation for a Decision to Conduct Gas Warfare," AFPAC Staff Study, 5 June 1945.

183. Vincent C. Jones, Manhattan: The Army and the Atomic Bomb, a volume in the series United States Army in World War II (Washington: Center of Military History, 1985), 529-30, 541.

184. Marc Gallicchio, "After Nagasaki: General Marshall's Plan for Tactical Nuclear Weapons in Japan," Prologue 23 (Winter 1991): 396-404; John R. Sutherland, "The Story General Marshall Told Me," U.S. News and World Report, 2 November 1959, 53.

185. Ibid.

186. "Logistic Study for Projected Operations: Kyushu," Planning Division, Army Service Forces, March 1945, CARL.

187. The Enola Gay Debate, August 1993-May 1995, chap. 4, no p.n. Confirmed by Defense Personnel Support Center (Army Support, Medals and Assembly), Philadelphia, Pennsylvania. Portions of the half-century-old Purple Heart medals are periodically shipped to a private contractor for cleaning and to have new ribbons attached. Since 1945, more than 370,000 have been awarded to wounded U.S. military personnel and the families of those killed in action. The Defense Personnel Support Center's stock of Purple Hearts stood at approximately 35,000 after the Korean and Vietnam wars, and remains almost unchanged today.

End of article

D. M. Giangreco letter in the Journal of Military History, 63 (October 1999): 1068-1070

To the Editor:

I am in complete agreement with Barton J. Bernsteinís statement in the July 1998 JMH that the casualty dispute [Bernsteinís italics] "is fundamentally about what top US officials--- not lower and middle-level people--- believed." Pursuant to his query for "any high-level supporting archival documents from the Truman months before Hiroshima that, in unalloyed form, provides even an explicit estimate of 500,000 casualties let alone a million or more," I offer the documents cited in footnote 75 of my "Casualty Projections for the US Invasions of Japan, 1945-1946: Planning and Policy Implications" (JMH, July 1997).

The documents in question were not examined in any detail in my casualty projections article because, quite frankly, its focus was on the US Armyís estimates and not President Harry S. Truman. These documents focus on Herbert Hooverís "Memorandum on Ending the Japanese War," a paper submitted to Truman after their 28 May 1945 meeting which now reside in the White House Confidential File at the Harry S. Truman Library. The memorandum itself is well-known and has been mentioned in a wide variety of publications. Inexplicably, what has virtually never been discussed is Trumanís reaction to the Hoover paper. I touch briefly on this matter in my new book Dear Harry . . .Trumanís Mailroom, 1945-1953: The Truman Administration Through Correspondence with "Everyday Americans":

"Some historians have maintained that the huge casualty estimates later quoted by Truman were a "postwar creation" designed to justify the use of nuclear weapons, and that such numbers were never even contemplated outside of strictly military circles. However, recently discovered documents at the Harry S. Truman Library tell a different story. Soon after Stimson circulated [Oswald C.] Brewsterís letter, former president Herbert Hoover submitted his "Memorandum on Ending the Japanese War"--- at Trumanís request--- and Truman was so struck by its estimate that the invasion could cost as many as 500,000 to 1,000,000 lives that he ordered his senior advisors to personally examine the memorandum. Trumanís manpower czar Fred M. Vinson, director of the Office of War Mobilization and Reconversion, received it first, and then former Secretary of State Cordell Hull as well as [Secretary of War Henry] Stimson and Acting Secretary of State Joseph C. Grew who were both instructed ! to prepare a written analysis before coming in for a face-to-face with the president. None of these civilian advisors batted an eye at the casualty estimate, and Truman promptly ordered a meeting [to be held on 18 June 1945] with Stimson and the Joint Chiefs of Staff to discuss "the losses in killed and wounded that will result from an invasion of Japan proper."

Grewís 13 June response to the Hoover memo that the Japanese "are prepared for prolonged resistance" and that the war could "cost a large number of human lives" has remained essentially unknown. In hindsight, however, Grew's concurrence would not have been a surprise to Truman. On 28 May 1945 he told the president: "The Japanese are a fanatical people capable of fighting to the last ditch and the last man. If they do this, the cost in American lives will be unpredictable"(Turbulent Era, A Diplomatic Record of Forty Years: Joseph C. Grew [1952]: 1429). Curiously, historians interested in what Truman was told of the upcoming invasion of Japan have had little interest in this statement.

Stimson meanwhile had forwarded his copy of the memorandum to Army Chief of Staff George C. Marshall who subsequently sent it to a senior planner on his staff for comment. Hooverís estimate of up to a million US dead was fully double the estimate used in JCS 924 and far exceeded the Army manpower policy of 1945 which was currently generating a 100,000-men-per-month replacement stream for the Japanese war then projected to last until nearly 1947. Consequently, Marshall endorsed the comments by the senior Army planner which stated that Hooverís estimate "appears to deserve little consideration." The memorandum with comments was sent back to Stimson who stated in his 2 July 1945 memo to Truman that: "We shall in my opinion have to go through a more bitter finish fight than in Germany [and] we shall incur the losses incident to such a war. . . . This would be due to both the difference in the Japanese and German personal character and the differences in the size and ch! aracter of the terrain through which the operations will take place" (July 1997 JMH, 560-61).

The Hoover paper and subsequent memos between Truman and his most senior advisors unquestionably meet Bernsteinís most recent criteria.

As for JCS 924, it and its annexes represent a single, comprehensive whole--- a document which, notes Douglas J. MacEachin, the CIAís former Deputy Director for Intelligence, "incorporat[ed] various modifications [and] articulated the JCS consensus on an invasion of Kyushu" (The Final Months of the War With Japan, unclassified CIA monograph, 1999, 1-2). Planning documents of this type are working documents. Nothing is modified or removed unless there is a written directive to do so, and such orders become part of the document itself. All were classified. All were numbered. The Saipan ratio of JCS 924/2 which resulted in an estimate that the invasion Japan "might cost us half a million American lives and many times that number wounded" did not, as Bernstein contends "quickly disappear," and it was one of the principal factors behind the steep jump in Selective Service inductions even as the war in Europe was winding down as well as manpower-related decisions ! on training and other military matters by Stimson, Army Chief of Staff George C. Marshall and Director of the Office of War Mobilization James F. Byrnes who would soon be appointed Secretary of State by Truman (July 1997 JMH, 564-67).

For more on this subject readers may consult:

D. M. Giangreco

US Army Command and General Staff College

Fort Leavenworth, Kansas 

Other links to invasion & atomic bomb:

Annotated bibliography:

Giangreco, Dennis. "To Bomb Or Not To Bomb." Naval War College Review, Spring 1998, 140-145. Online at

review essay of the most recent Alperovitz, Ferrell and Wainstock books

Debate on this article

More discussion


Copyright 1997 by D. M. Giangreco and the Journal of Military History

This article may be reproduced without additional permission for nonprofit educational use.

Please cite it as follows:

D. M. Giangreco, "Casualty Projections for the U.S. Invasions of Japan, 1945-1946: Planning and Policy Implications" in Journal of Military History, 61 (July 1997): 521-82, online at