December 2003 Newsletter

CASUALTY PROJECTIONS FOR THE INVASION OF JAPAN, PHANTOM ESTIMATES, AND THE MATH OF BARTON BERNSTEIN
by Michael Kort

Over the past thirty years, Barton Bernstein has been a prominent participant in scholarly discussion about the bombing of Hiroshima and the surrender of Japan. He is widely recognized as a leading authority on various aspects of that subject, including the question of how many U.S. casualties Harry Truman and his advisors expected would result from the invasion of Japan planned to begin in the fall of 1945.1 He played an important role, as a defender of low-end casualty estimates, in the 1994-95 debate over the script of the ill-fated Enola Gay exhibit that the Smithsonian Institution had planned to mark the fiftieth anniversary of the bombing of Hiroshima. His long list of articles examining the bombing of Hiroshima is clearly the product of extensive archival work. He seems to have examined every important document in English related to the main debates about the atomic bombing of Japan, and he has chastised other historians, including those whose conclusions seem to concur with his, for being less than thorough in their research on the casualties issue.2 It therefore seems reasonable to apply Bernstein’s standards to his own work on that same subject.

First we must go over some familiar ground regarding terminology, in particular the terms “casualties” and “battle casualties.” At times some historians seem to use them interchangeably, causing serious confusion. “Casualties” is obviously an umbrella term that includes battle and non-battle casualties; thus it always exceeds battle casualties alone, often by a large number. As will be demonstrated below, Bernstein has mangled this distinction when dealing with comments made by Admiral William D. Leahy, at the time President Truman’s chief of staff, at the crucial White House meeting of June 18, 1945, where Truman and several of his top advisors reviewed Operation Olympic, the plan to invade Kyushu, Japan’s southernmost home island. In addition, Bernstein has substituted his own suppositions about what George C. Marshall said at the meeting, using words like “apparently,” “probably,” and “strongly suggests” for what is missing in the documented historical record to create a casualty estimate that in fact does not exist.

Bernstein has always given a range for the casualties U.S. leaders expected in an invasion of Japan, while maintaining that the numbers were much lower than those Truman later quoted and those many historians, who usually support Truman’s decision to use the bomb, have used. For example, in his 1986 article “A Postwar Myth: 500,000 U.S. Lives Saved,” Bernstein states that military planners placed expected losses “at 46,000 and sometimes as low as about 20,000 lives.”3 In the1990s he tried to become more precise, at least with regard to what Marshall said and Leahy implied at the meeting of June 18, 1945. In addition, in an article published in 1999, Bernstein tries to reconcile their estimates from that day to the point where Marshall and Leahy “were only mildly disagreeing—63,000 versus 66,500—U.S. casualties.” The insignificance of this difference, Bernstein adds, may explain why four participants in the meeting—Leahy, Secretary of War Henry Stimson, Secretary of the Navy James Forrestal, and Assistant Secretary of War John McCloy—did not mention any disagreement on casualty estimates in their diaries.4

Bernstein uses a lot of qualifiers in writing about this subject, which is not a bad idea given the problems involved, but it seems fair to say he concludes that as of June 18, 1945, Marshall believed that the invasion of Kyushu would entail 63,000 U.S. battle casualties. For example, in an article published in 1998 Bernstein writes that “General George Marshall, the army chief of staff, had apparently told the 18 June 1945 White House meeting, including Truman, that American casualties would not exceed 63,000 among the 190,000 U.S. combat troops in the forthcoming operation on southern Kyushu.”5 In 1999 Bernstein asserts that the “63,000 estimate . . . strongly suggests that Marshall was thinking in terms of somewhat under 100,000 battle casualties for Olympic, and possibly no more than 63,000 total battle casualties in that American operation,” and on the same page he refers to “Marshall’s own stated estimate of 63,000 ground-force casualties.”6 In another 1999 article he reaffirms that “General Marshall stated (in Leahy’s paraphrased summary) that Olympic ‘will not cost us in casualties more than 63,000 of the 190,000 combatant troop estimated as necessary for the operation.’ ”7

The problem with all of this is that nowhere in the minutes of the June 18 meeting is it recorded that Marshall said these things. That is not because the secretary who took the minutes was shy about taking down numbers. On the contrary, the June 18 minutes record that Marshall cited combat casualties for the battles of Leyte, Luzon, Iwo Jima, and Okinawa, the last of which was not yet over. He provided statistics on American and Japanese soldiers killed in General MacArthur’s operations between March 1944 and May 1945. He said that the first thirty days of the Kyushu invasion would not result in more casualties than did Luzon (31,000), but that comment, coming in the sanguinary wake of Iwo Jima and Okinawa, was unlikely to provide Truman with much comfort, since the battle for Kyushu was expected to last much longer than thirty days. Marshall also mentioned the size of the U.S. force that would invade Kyushu--766,700--and the number of Japanese troops expected to defend the island--350,000. These numbers are all clearly and precisely recorded in the minutes, as are the statistics on air casualties (2 percent per mission; about 30 percent per month) provided by General Ira Eaker.8 But where is Bernstein’s 63,000? Given the reason the meeting was called in the first place and what was at stake, is it possible that if Marshall had mentioned such a critical number the meeting’s official secretary, JCS secretary Brigadier General A. J. McFarland, would not have made absolutely sure it was recorded? Would an army general, who surely would have understood its importance, leave out such a statement by, of all people, General George C. Marshall? Such assumptions are so unlikely they can be ruled out. Since the 63,000 casualty estimate is not recorded in the detailed minutes of the meeting, which occupy more than seven single-spaced pages for a meeting that lasted about an hour, historians must decide which document is more reliable: the carefully prepared and reviewed official minutes of the June 18 meeting taken on the spot, none of the drafts of which mentions the 63,000 figure, or Leahy’s haphazardly organized diary entry written down some time later in the day.

Marshall was clearly reluctant to give a specific casualty figure at this meeting. His first comment about casualties was, “Our experience in the Pacific War is so diverse as to casualties that it is considered wrong to give any estimate in numbers.”9 Given this reluctance and the lack of any recorded references to the 63,000 number outside Leahy’s diary, it would be reasonable to conclude that the evidence decisively favors the minutes.10 Bernstein, who insists that other historians writing about this subject provide archival evidence to back up their arguments, should do the same to support his own claims.11 Leahy’s hearsay diary entry does not come close to meeting his standard, since it disagrees with the official minutes of the June 18 meeting and is not verified by any other contemporary document. It cannot in any way establish that Marshall used the figure 63,000 that afternoon. It seems fair to say that Bernstein, who demands “unalloyed” evidence from others, has created his estimate through alchemy by transmuting Leahy’s nighttime diary entry into an afternoon statement by Marshall at the June 18 meeting.

Indeed, the documentary record of the June 18 meeting suggests casualty estimates several times higher than 63,000. The minutes record that Marshall made his presentation and included the reference to 31,000 casualties in the first thirty days. Then Leahy, who opposed the invasion and, as a Navy man, favored a strategy of blockade (the Navy’s job) and bombardment (the job of the Army Air Force), commented that if the president wanted to know how many casualties there would be on Kyushu he should look at Okinawa, where, he said, the casualty rate was 35 percent. Next, Admiral King suggested that casualties on Kyushu would fall somewhere between the losses suffered on Luzon and Okinawa. Finally, Marshall said that the assault force for Kyushu would comprise 766,700 troops. Here we have a link in black and white—evidence Bernstein would call “archival”—that we do not have to create with supposition or speculation: Leahy says the expected casualty rate is likely to be 35 percent of the assault force, and then Marshall gives the size of that force as 766,700. Marshall does not use the number 190,000, as Bernstein claims he did. The next number, in the following sentence, is the expected number of Japanese defenders, 350,000.12 As many commentators have pointed out, some quick math gives a total well in excess of 250,000 casualties, and this only for Olympic, the first half of a two-stage invasion (the other, Coronet, would be launched on March 1, 1946, against Tokyo). It is eminently reasonable to assume that every person at the meeting made this straightforward mental calculation.

Then comes a twist. Having first created a casualty estimate for George Marshall through speculation, Bernstein then tries to reconcile conflicting remarks by Marshall and Leahy actually recorded in the June 18 minutes. His does this by maintaining that when Leahy made his 35 percent statement in the June 18 meeting, he had in mind the 190,000 combat troops he mentions in his diary. And 35 percent of 190,000 is 66,500. This last number, which thus far has appeared only in Bernstein’s suppositions as opposed to the archival documents he demands of others, is most convenient since, Bernstein notes, it is rather close to 63,000, leaving Marshall and Leahy nearly in agreement (“only mildly disagreeing”) and explaining, Bernstein says, why no diary kept by a meeting participant pointed out a disagreement on this point. By implication, this calculation, made decades after the fact by an historian who was not there, proves that Leahy suggested, and Marshall concurred with, a casualty estimate that detailed minutes taken on the spot by a brigadier general charged with keeping an accurate record did not record. Meanwhile, two military historians, Thomas Allen and Norman Polmar, are criticized for “entirely omitting Marshall’s estimate”—which appears nowhere in the minutes—from their discussion of the meeting. It is hard to blame them for this alleged oversight. There is, in fact, absolutely no documentary basis for Bernstein’s claim in “Truman and the A-Bomb” that Leahy “probably” was using 190,000 as the base number for his 35 percent casualty rate suggestion.13

In addition, Leahy, the source of the number in question, is unclear about when he heard it. His diary for June 18 simply states, “General Marshall is of the opinion that such an effort will not cost us in casualties more than 63,000 of the 190,000 combatant troops estimated as necessary for the operation.” What is interesting about the statement is that it is written in the present tense. So are a number of other entries for June 18, all concerned with Leahy’s or someone else’s view about the invasion, surrender, or occupation of Japan, but none fixed in time--that is, clearly noted as having actually taken place on June 18. (“The Army seems determined to occupy and govern Japan,” “I am unable to see any justification from a national defense point of view for a prolonged occupation,” “It is my opinion at the present time that a surrender of Japan can be arranged,” and, of course, “General Marshall is of the opinion . . . .”) These entries differ from every entry that, in contrast, is clearly a statement of fact regarding an event that took place that day. Those entries are all written in the past tense. Thus “Eisenhower arrived,” he “made a very well prepared address” (“not delivered with particular skill”); after the address “we proceeded to the Statler Hotel,” and “the President conferred with the Joint Chiefs of Staff, the Secretary of War . . . .” Indeed, there is a clear reference to what General Marshall did say at the June 18 meeting, and it, like the other statements of events that occurred that day, also is in the past tense. (“General Marshall and Admiral King both strongly advocated an invasion of Kyushu.”) In addition, two of the present tense opinion entries—the reference to the army’s intent to occupy and govern Japan and Leahy’s opposition to a long occupation—deal with a subject that we know was not mentioned at the June 18 meeting.14

In short, a careful reading of Leahy’s entries for June 18 reveals that there is no reason to think the reference to Marshall’s "opinion" about casualties refers to estimates was made that day or at the meeting in question. Given the context and lack of supporting evidence in the minutes, that is a far more reasonable conclusion than an assumption that Marshall used the 63,000 number at the June 18 meeting but somehow that vitally important statistic was overlooked by those responsible for keeping a record of the proceedings. As with the June 18 minutes themselves (see below), Bernstein overlooks or ignores Leahy’s language in order to reach a conclusion that the language simply does not support.

The problem is that Bernstein puts words in Leahy’s mouth, as he does with Marshall. As far as we know from the minutes, indisputably the most reliable source for the event, Leahy did not say 66,500 casualties at the June 18 meeting, nor did he correct Marshall when the latter, immediately after Leahy mentioned the 35 percent figure for Okinawa casualties, said 766,700 troops would invade Kyushu. It also strains the imagination to think that Leahy would have made his comment about percentage if his estimate (66,500) was so close to Marshall's (63,000). After all, 3500 casualties in this context is hardly a major difference.

While accepting without corroboration Leahy’s diary numbers regarding Marshall’s alleged estimates, Bernstein argues that Leahy got his percentage wrong. As far as I am aware, he does this for the first time in 1986 in “A Postwar Myth,” where he claims the correct casualty rate on Okinawa was 29 percent. Later he also corrects Allan and Polmar, who in their 1995 article “Invasion Most Costly” and in their book Downfall: The Secret Plan to Invade Japan and Why Truman Dropped the Bomb also correct Leahy, saying the actual percentage of casualties suffered on Okinawa by U.S. forces was 39 per cent.15 Yet it turns out that the percentage figures suggested by Leahy on the one hand and Allen and Polmar on the other were both correct, and Bernstein’s wrong.

How is this possible? For that matter, how can two numbers, 35 and 39 percent, be correct, and only Bernstein’s 29 percent (or, as he later wrote, 26 percent) be wrong? In “Reconsidering ‘Invasion Most Costly’,” Bernstein makes a series of calculations and suppositions based on an assortment of casualty and troop numbers. There seem to be two goals here: to establish that both Leahy and Allen/Polmar are incorrect and to have Marshall (63,000 casualties) and Leahy (66,500 casualties) agree with each other. Bernstein accomplishes this by putting Leahy’s diary statement through a filter of “interpretation” within an “understanding of the context.” He correctly points out that the number of U.S. troops on Okinawa varied during the course of the battle, which is not a surprise, since this is true of any such battle. He writes incorrectly (see below) that “U.S. troops on Okinawa numbered about 154,000 in April 1945,” which he calculates, using the ground force battle casualty number of 39,420, would yield a rate of 26 percent “or even a few points lower.”16 The actual math is close enough, but the numbers used are wrong and so, therefore, are the results.

The minutes make it clear that Leahy did not say “battle casualties”; he said “casualties.” The minutes have Leahy using the word three times. By definition “casualties” means total casualties.17 And what were the ground force total casualties, battle and non-battle, on Okinawa? For these answers we can turn to one of Bernstein’s sources in “Reconsidering ‘Invasion Most Costly’.” Okinawa: The Last Battle is an official study produced shortly after the war by the United States Army. The authors of the study report “a total of 183,000 troops was made available for the assault phases of the operation.” Given that its source is the United States Army, this figure is authoritative and therefore must be the one used to calculate the casualty percentage figures. (The breakdown is as follows: 183,000 total ground force troops, of which 154,000 were in seven combat divisions.) The 183,000 figure is also the one I have found most commonly cited in reliable secondary works as the assault force for Okinawa. The next task is to determine the total battle and non-battle casualties suffered by the ground forces. Navy casualties should not be included; the issue under discussion at the June 18 meeting was the assault force--that is, the ground force that was going ashore. Thus, from the total battle casualties of 49,151 suffered by all U.S. forces in the battle of Okinawa, one must first subtract the navy total of 9,731 dead and wounded. That leaves 39,420 army and marine battle casualties. The army (15,613) and marine (10,598) non-battle casualties add up to 26,211, for a total casualty figure of 65,631. Divide 65,631 by 183,000, the number of troops the U.S. Army reports was in the assault force, and the answer is 35.8 percent, almost exactly Leahy’s figure.18 It should not be surprising that Leahy came that close, even though the battle for Okinawa was not over. As D. M. Giangreco points out, Leahy and the other participants in the June 18 meeting knew the 34,000 casualty figure read by Marshall was too low, and Leahy clearly did not limit himself to battle casualties. In fact, Giangreco even provides the total casualties for Okinawa and points out that this explains how Leahy got his 35 percent.19

But what about Allen and Polmar’s 39 percent figure? First of all, they used the wrong numbers to get it. It is possible that they encountered the 39 percent figure in an authoritative source, took the battle casualty figure for the army and marines—39,420—and made the numbers work by postulating about 100,000 assault troops, an incorrect figure.20 Still, 39 percent is a valid figure, for as Giangreco notes, an alternate method for calculating non-battle casualties used by the government pushed the total on Okinawa to 33,096, giving a total casualty figure of 72,516. Divide 72,516 by 183,000; the answer is 39.6 percent. As Douglas J. MacEachin points out, these revised figures became available about a month after the June 18 meeting.21

In one of his recent articles (“Reconsidering ‘Invasion Most Costly’ ”), Bernstein minimizes the importance of non-battle casualties, insisting that “aside from neuropsychiatric casualties” they were “frequently comparatively minor, as indicated by the fact that under one-third of one percent (.003) of nonbattle casualties died on Luzon” compared to a quarter of battle casualties there and in the southern Philippines. Therefore, mixing battle and nonbattle casualties “can be more distorting than illuminating.”22 That depends on how widely one focuses the lens and where one shines the spotlight. It turns out that 28 percent of all U.S. World War II military deaths, more than 115,000 soldiers, were non-battle casualties, a grim reality that indisputably puts into perspective and certainly illuminates the importance of those casualties. To be sure, some military statistics do not include non-battle wounded (as opposed to non-battle dead), but some do. A key issue for military leaders was the extent to which an injury affected a unit’s effectiveness.23 An official army study that deals with Luzon, Triumph in the Philippines, points out the large number of non-battle losses on Luzon and observes that “an infantryman who was hospitalized for pneumonia contracted in the mountains of northern Luzon was as much a loss as an infantryman who was hospitalized with a wound inflicted from a Japanese bullet. Combat fatigue casualties, permanent or temporary, fit into the same category.”24 In short, to military planners responsible for keeping units at full strength, non-battle casualties mattered a great deal: that is why Leahy included them when he reported the 35 percent casualty rate.

Bernstein’s later articles, especially “The Alarming Japanese Buildup on Southern Kyushu,” are packed with statistics about casualties and troop deployments, as well as an assortment of computations, none of which do anything to correct the fundamental errors he made regarding Leahy’s percentage statement at the June 18 meeting or add any validity to his effort to have Marshall utter the number 63,000 at that meeting. One issue he raises is the precise deployment of Japanese troops on Kyushu as of late July 1945. Bernstein points out that any of the troops were in the north, not in the south where the United States was planning to come ashore.25 He seems to think these mid-summer deployments had military significance, but the American military commanders whose job it was to evaluate the situation and prepare for the expected battle in the fall did not. On July 29, 1945, General MacArthur’s intelligence staff, noting the rising numbers of Japanese troops on Kyushu, concluded, “The assumption that enemy strength will remain divided in North and South (Kyushu) compartments is no longer tenable.” The staff added, “The trend of reinforcements from North to South (Kyushu) is unmistakable.” Making matters still worse, the Japanese were massing exactly where the Americans were planning to attack and unless something were done “enemy forces in Southern Kyushu may be still further augmented until our planned local superiority is overcome.”26 Nor in the impending battle were Japanese forces going to be limited to the soldiers already on Kyushu. On August 1 MacArthur’s intelligence staff reported that

"the Japanese have ample reserves in rear areas . . . .it is probable that 3 divisions or division-equivalents will be available in NORTHERN KYUSHU on x-day for reinforcement purposes. Moreover, higher headquarters indicated that a sufficient number of divisions are in western HONSHU, in KOREA, and in western SHIKOKU to enable the enemy to reinforce the KYUSHU garrison on a scale of its own choosing, without unduly jeopardizing local security or long range strategic requirements for the defense of HONSHU . . . .
Movement of enemy reserves to NORTHERN KYUSHU will present no insurmountable obstacles . . . .the enemy still will possess a sufficient supply of small craft to move, over an extended period of time and taking advantage of darkness and periods of low visibility, almost any desired number of troops across the narrow waters separating the northern tip of KYUSHU from its northern, western, and eastern neighbors."27

Events on the ground confirmed these assessments. On August 2, Ultra intercepts of Japanese military radio transmissions placed the Japanese 206th division on northern Kyushu; two days later Ultra found that division had redeployed to the south. It was joined there a few days later by the 303rd division, which had just arrived on the island with the 216th division, which went to central Kyushu.28 Regardless of what the Americans did, the Japanese were successfully moving their troops from north to south on Kyushu and at the same time moving additional divisions to the island.

It is worth noting that applying Bernstein’s analysis of Japanese troop deployment to the disposition of American troops on July 29, three months before the planned invasion, would lead to absurd conclusions about how many troops the United States would be able to land on southern Kyushu on X-day. After all, American troops were hundreds of miles farther away from southern Kyushu than the Japanese troops on the northern part of the island. Bernstein also tends to minimize the military importance of the Japanese buildup, which reached 900,000 troops on Kyushu by August, pointing out that many Japanese troops were “hastily raised, ill-equipped and ill-trained.”29 This is to a certain extent true, but it is far less significant than Bernstein implies. For example, the Japanese situation on Iwo Jima was hardly ideal in June of 1944 when General Tadamichi Kuribayashi arrived to take command. In July, after reviewing his forces, he told an aide that he had “no soldiers, just poor recruits who don’t know anything. The officers are either fools or superannuated scarecrows. We cannot fight the Americans with them.”30 But six months later, reinforced and worked into shape despite long, tenuous supply lines and daunting obstacles, the Japanese troops on Iwo Jima fought a battle that sent a collective shudder through the American military. Japanese soldiers on Okinawa, outnumbered, outgunned, and cut off from reinforcements, fought no less effectively. As for the situation on Kyushu in July 1945, the Japanese still had three months to prepare for an American invasion and were working intensively to do just that. Meanwhile, American forces assigned to the second stage of the projected invasion of Japan’s home islands were far from ready; in fact, given the tight schedule, many of the troops being redeployed from Europe for Coronet would have landed on the beaches of Honshu without proper amphibious training.31

In summary, Bernstein’s assessments about what Marshall estimated and Leahy calculated are based on invalid assumptions and faulty math. His assertions about what Marshall said at the June 18 meeting are contradicted by the only reliable archival evidence available. He has misread Leahy’s diary. Because he has used the wrong numbers, he has been wrong about Leahy’s percentage statement since 1986 and also wrong about Allen and Polmar (even though they did indeed err in citing the number of U.S. troops on Okinawa) since 1999. This review of the casualty estimates Marshall and Leahy used may not settle the ongoing debate about how many casualties Truman and his advisors were expecting, but it certainly further weakens the already tottering argument for low numbers.

1. The book that increasingly is being recognized as the standard work on the bombing of Hiroshima and Japan’s surrender is Richard B. Frank, Downfall: The End of the Imperial Japanese Empire (New York, 1999). Two other recent excellent overviews are Robert P. Newman, Truman and the Hiroshima Cult (East Lansing, MI, 1995) and Robert James Maddox, Weapons for Victory: The Hiroshima Decision Fifty Years Later (Columbia, MO, 1995). The most comprehensive study of casualty projections for the expected invasion of Japan is D. M. Giangreco, “Casualty Projections for the U.S. Invasion of Japan, 1945-1946: Planning and Policy Implications,” The Journal of Military History 61 (July 1997), 521-81.
2. See Barton Bernstein, “Reconsidering ‘Invasion Most Costly’: Popular-History Scholarship, Publishing Standards, and the claim of High U.S. Casualty Estimates to Help Legitimize the Atomic Bombings,” Peace and Change 24 (April 1999), 220-48. For a critique of a like-minded historian (Rufus Miles), see Bernstein, “A Postwar Myth: 500,000 U.S. Lives Saved,” Bulletin of the Atomic Scientists (June/July 1986), 40, n. 2. Bernstein’s criticism of Miles “for arguing substantially on the basis of inference” is ironic in light of his own argument made some years later regarding Marshall’s casualty estimates at the June 18, 1945 White House meeting (see below). See also Bernstein,“Truman and the A-Bomb: Targeting Noncombatants, Using the Bomb, and Defending the ‘Decision’,” Journal of Military History 62 (July 1998), 552, n. 10.
3. Bernstein, “A Postwar Myth,” 38.
4. Bernstein, “Reconsidering ‘Invasion Most Costly’,” 229-30.
5. Bernstein, “Truman and the A-Bomb,” 551.
6. Bernstein, “Reconsidering ‘Invasion Most Costly’,” 229.
7. Bernstein, “The Alarming Japanese Buildup on Southern Kyushu, Growing U.S. Fears, and Counterfactual Analysis: Would the Planned November 1945 Invasion of Southern Kyushu Have Occurred?” Pacific Historical Review 99 (November 1999), 573. Bernstein made the 63,000 number a heated point of contention during the 1994-95 debate over the Smithsonian Institution’s Enola Gay exhibit. At a meeting in November 1994 of representatives of the Organization of American Historians and the Smithsonian staff, Bernstein strongly criticized the section of a revised version of the exhibit script, which went through several revisions during the long controversy, dealing with Leahy’s June 18 comments. He insisted that, based on Leahy’s diary entry for that day, 63,000 was what Truman’s chief of staff had meant when he spoke of casualties at the meeting. That contention stunned Martin Harwit, director of the Smithsonian’s National Air and Space Museum, as it contradicted an assessment Bernstein himself had made in 1986, when he had written that at the June 18 meeting Leahy “suggested that total casualties might run as high as 230,000.” See Martin Harwit, An Exhibit Denied: Lobbying and the History of the Enola Gay (New York, 1996), 380-82, 345-46, and Bernstein, “Reconsidering ‘Invasion Most Costly,’” 242, n.15. For Bernstein’s 1986 estimate see “A Postwar Myth: 500,000 U.S. Lives Saved,” 40.
8.”Minutes of Meeting held at the White House on Monday, 18 June 1945 at 1530," Dennis Merrill, ed., Documentary History of the Truman Presidency, Volume 1: The Decision to Drop the Bomb on Japan (University Publications of America, 1995), 2,4,5. (Hereafter “Minutes of Meeting . . .18 June.” Pages cited are from the minutes themselves.) Along with the final official version, the Merrill collection also includes two drafts of the minutes.
9.“Minutes of Meeting . . .18 June,” 2. Elsewhere Marshall commented that he did not believe the JCS should make estimates “which can at best be only speculative.” See D. M. Giangreco, “‘A Score of Bloody Okinawas and Iwo Jimas’: President Truman and Casualty Estimates for the Invasion of Japan,” Pacific Historical Review 72 (February 2002), 119, n.73. See also Frank, Downfall, 144-45.
10. “Minutes of Meeting . . .18 June,” 1-8. Marshall’s numbers are on pages 2 and 4. Leahy’s percentage statement (see below) is on page 4.
11. For example, Bernstein criticizes D. M. Giangreco for not providing “unalloyed 1945 archival evidence” about the influence of an August 1944 calculation, known as the Saipan ratio, for estimating potential U.S. casualties. Bernstein, Letter to the Editor, The Journal of Military History 63 (January 1999), 249.
12.“Minutes of Meeting . . .18 June,” 2-4.
13. Bernstein, “Reconsidering Invasion Most Costly’,” 229-30; idem, “Truman and the A-Bomb,” 551, n. 7.
14. William D. Leahy, The Leahy Diaries, June 18, 1945. http://www.historians.org/archive/hiroshima/180645.html The comment on Marshall’s opinion follows a discussion of the June 18 meeting and is in turn followed by a one-sentence reference to Truman’s decision to approve the invasion. That in turn is followed by the opinion statements about the occupation and surrender of Japan (four sentences). Leahy then returns to the events of the day, all recorded in the past tense.
15. Bernstein, “”A Postwar Myth,” 40; idem, “Reconsidering ‘Invasion Most Costly’,” 229. See also Thomas B. Allen and Norman Polmar, “Invasion Most Costly,” Proceedings, U.S. Naval Institute (August 1995), 56; and idem, Code Name Downfall: The Secret Plan to Invade Japan and Why Truman Dropped the Bomb (New York, 1995), 211.
16. Bernstein, “Reconsidering ‘Invasion Most Costly’,” 228-29. Here is where Bernstein runs into trouble with his terminology. In the paragraph on Leahy (line 28 of page 229) he writes, “So, Leahy should have said about 26 percent, or even a few points lower.” In the preceding paragraph he refers to Leahy’s statement about the “casualty rate for U.S. ground forces in the Okinawa campaign.” In note 17 (on page 243), where he provides the numbers that produced his 26 percent, Bernstein speaks about “battle casualties.” The problem is that Leahy and Bernstein (“casualties” versus “battle casualties”) are not referring to the same thing. See also Bernstein, “The Alarming Japanese Buildup on Southern Kyushu,” 571-72, n.15. Here Bernstein refers to the “U.S. ground force battle-casualty rate” on Okinawa, which, of course, is not the same thing as the total casualty rate. The ground force battle casualty rate, “even using the unduly low figure of 154,000 for the total number of troops,” is placed at “slightly under 26 percent.” None of this has any relevance to Leahy’s estimate, as neither number—the battle casualty rate or the 154,000 figure—should be used if one is calculating the total casualty rate suffered by U.S. ground forces on Okinawa, which is what Leahy was doing (see the next paragraph in the text.) At the same time, by lumping ground forces with naval forces and counting only battle casualties, Bernstein reduces the casualty rate “overall for Okinawa” to “slightly under 10 percent.” Similar accounting for the “entire operation” at Iwo Jima yields “slightly over 10 percent.” The point of these calculations is unclear, since these figures are artificially low because of the inclusion of naval forces and therefore lack any relevance whatsoever for understanding what might have happened to American ground forces in the Kyushu operation.
17. See Giangreco, “Casualty Projections,” 556, where he points out that Leahy “used the total number of casualties to formulate the 35 percent figure.”
18. Roy E. Appleman et al., Okinawa: The Last Battle (Washington, DC, 1948), 26, 36, 473. Appleman adds that the “garrison force” on Okinawa eventually reached 270,000. For a secondary source that uses the 183,000 figure (and the 154,000 figure for combat divisions) see George Feifer, Tennozan: The Battle of Okinawa and the Atomic Bomb (New York, 1992), 146. For a Japanese commentator who correctly cites the 183,000 figure, see Takashi Itokazu, “The Battle of Okinawa.”
www.okinawatimes.co.jp/summit/english/2000/20000722_6.html accessed 4/1/2003.
19. Giangreco, “President Truman and Casualty Estimates,”124. See also idem, “Casualty Projections,” 539. It seems surprising that Bernstein, who apparently went over Giangreco’s article with a fine-tooth comb, ignored statistics that were right in front of him and clearly demonstrated that Leahy’s 35 percent figure was correct. In fact, in “The Alarming Buildup On Southern Kyushu” (page 596, n. 55), Bernstein points out that Allen and Polmar “misread the nonbattle casualties as battle casualties.”
20. Allen and Polmar, Code-Name Downfall, 211.
21. Giangreco, “President Truman and Casualty Estimates,” 124; idem, “Casualty Projections,” 539. The report that raised the casualty percentage to 39 percent was dated 11 July 1945. See Douglas J. MacEachin, The Final Months of the War With Japan: Signals Intelligence, U.S. Invasion Planning, and the A-Bomb (Washington, DC, 1998), 14, n. 31. Also see Frank, Downfall, 71. Frank gives the figure 72,358 for the final Okinawa total. He adds that this approximated the total of 76,000 Japanese trained defenders, “an extremely ominous indicator amid preparations for an invasion of Japan.”
22. Bernstein, “Reconsidering ‘Invasion Most Costly’,” 230-31.
23. See Giangreco, “Casualty Projections,” 539-40, footnotes 66 and 67.
24. Robert Ross Smith, Triumph in the Philippines (Washington, DC, 1963), 652-53.
25. Bernstein, “The Alarming Japanese Buildup on Kyushu,” 570-71, 577-78, n.30, 599. See also idem, “Truman and the A-Bomb,” 551, 553.
26. G-2 GHQ, AFPAC, "Amendment No. 1 to G-2 Estimates of the Enemy Situation with Respect to Kyushu (Dated 25 April 1945)," 29 July 1945, Reports of General MacArthur: The Campaigns of MacArhtur in the Pacific, Vol 1. (Washington, D.C.: U.S. Government Printing Office, 1994), 418.
27. G-2 Estimate of the Enemy Situation with respect to OPERATION OLYMPIC (Southern Kyushu), 1 August 1945, Record Group165, Box 1843, National Archives and Records Administration, College Park, MD.
28. MacEachin, 20, 23 (n. 52).
29. Bernstein, “The Alarming Japanese Buildup on Southern Kyushu,” 577-78, n.30. See also idem, “Reconsidering ‘Invasion Most Costly’,” 247, n.32.
30.Quoted in Richard F. Newcomb, Iwo Jima (New York, 1965), 12. The actual source for this statement was one of the general’s aides, Major Yoshikata Horie. He arrived on Iwo Jima in July and later was sent to the island of Chichi Jima, about 140 miles to the north, where he was responsible for the control of arms and supply traffic to Iwo Jima. Horie closely monitored the battle for Iwo Jima until the last radio report from General Kuribayashi’s radioman on March 23. He survived the war and became the chief chronicler of what happened on Iwo Jima from the Japanese perspective.
31. Robert M. Palmer, Bell I. Wiley, and William R. Keast, The Procurement and Training of Ground Combat Troops (Washington, DC, 1948), 647.