The Manhattan Engineer District surveyIn 1946, the Manhattan Engineer District published a study that concluded that 66,000 people were killed at Hiroshima out of a population of 255,000. Of that number, 45,000 died on the first day and 19,000 during the next four months. In addition, "several hundred" survivors were expected to die from radiation-induced cancers and lukemia over the next 30 years. (This report is also known as the Oughterson Commission study.) This is the low-ball estimate, evidently because it was based on a census of households in Hiroshima and therefore did not account for the deaths of soldiers and Korean forced laborers, who are generally numbered at 20,000--though I can't find any solid justification for that figure. If they all died, which is very unlikely, and if we add a thousand deaths instead of the several hundred estimated by Oughterson's group, then we seem to be talking 87,000 fatalities directly attributable to the explosion.
The American researchers did an extensive random sampling of the surviving population, asking how large their family was and how many had been killed. From the results it was calculated that 25.5% of the civilian population had been killed. The great unknown, of course, is how large the population was at the time of the explosion. Where the Manhattan Engineer District gave a figure of 255,000--a figure based on the June 1945 rice-ration records, which survived the blast--others have posited 300,000 or even 400,000 including military and "day workers" (the eumphemism of choice for the Korean slave laborers). These populations would not have been shown on the rice-rationing records.
But even if 400,000 people were present in Hiroshima on August 6, the death toll ought not to exceed 102,000, if the American methodology was sound.
The Hiroshima police studyAlso in 1946, the Hiroshima police estimated the dead at 78,150 and the missing at 13,983, for a total of about 92,000 if all the missing are presumed dead (again, a very unlikely hypothesis). So this estimate is not radically different from the American estimate.
Perhaps significantly, the police study gave a figure of 129,558 for total casualties, including those with minor as well as major injuries. (These figures are suspiciously precise, but never mind that.) Today's "consensus" figure--that is, the one you see most often where the writer is not trying to prove a point one way or another--seems to be 130,000 dead. Writing for Air & Space magazine in the 1990s, I discovered to my horror that at least one editor didn't know the difference between a casualty and a fatality. Is this simply another case of counting all the wounded as dead?
The Japanese Reconstruction SurveyOne possible source of confusion is where to stop counting the deaths of survivors. In 1978, the Japanese Reconstruction Survey compiled the times of death for 16,007 people known to have been present in Hiroshima. This survey found that 73.4% had died by 1 November 1945, and that an additional 5.6% had died between then and the October 1950 census. Interestingly, the latter death rate is 1.1% a year--almost exactly the normal mortality rate for the Japanese population. From this I conclude that the methodology of the Manhattan Engineer District report was sound. Counting deaths as of the end of 1945 must have captured essentially all of them.
The Radiation Effects Research Foundation website gives a range of 90,000-140,000 1945 deaths at Hiroshima out of a population of 310,000. The Hiroshima Peace Site website gives a figure of 140,000 deaths by December 1945, out of a population of 350,000. And the Guinness Book of Records gives a suspiciously precise figure of 155,200 killed by Little Boy, including deaths from radiation within one year.
In all three cases above, there is no information whatever on where the figures come from.
The Committee for the Compilation of Materials on Damage Caused by the Atomic Bombs in Hiroshima and Nagasaki estimated in 1978 that 346,000-356,000 people were present in Hiroshima at the time of the bombings, with fatalities of "some 200,000". This seems to be a bit of a stretch, since the last census conducted by the Japanese government prior to the bombings, in February 1944, showed a population of 343,034, The Committee is thus claiming a net gain in population during the final year of the war, when widespread evacuations were going on during the fire bombings and other cities were rapidly losing people.
In 1998, a Japanese delegation in India presented this version: "At that time, Hiroshima's population was 400,000, of which 140,000 died by the end of 1945, 90 per cent of them within a week of the explosion." So far, so good--that tracks other recent Japanese estimates. But the statement continues: "People continue to die even today, from the after-effects of radiation.... As of , there were 202,118 registered deaths due to the Hiroshima bombing." So here we have 62,000 deaths added to the total, with the count continuing at least into 1998. Clearly we are in an entirely different field by now. A 21-year-old in 1945 would have been 74 in 1998, and therefore have already lived past his normal life expectancy!
It's true that lives were shortened by the blast--but then, they were shortened by the war itself, and especially by the malnutrition that was general in Japan in 1945. Even if that hypothetical 21-year-old, laid to rest in 1998, would have otherwise lived into his eighties or even nineties, can we fairly attribute his death to Little Boy? After all, nobody is counting the American prisoners of war who have died in the past ten years, and calling them fatalities of the Japanese PW system.
Updated: In refreshing contrast to the accelerating figures published above, the City of Hiroshima has a project called Actual Status Survey of Atomic Bomb Survivors. The survey from 1979 to 1999 accumulated the names of 88,800 individuals present in Hiroshima in August 1945 who died before the end of the year. Certainly some of these died from other causes; just as certainly, some died who will never be known.
ConclusionFrom all that I have read, the 1946 consensus figure of 90,000 dead seems about right to me. Deaths after December 1945 evidently were not very numerous, and they seem to have been adequately accounted for in the 1946 studies. Even the Radiation Effects Research Foundation (cited above) seems to confirm this. The foundation's website concludes that the number of excess deaths among 50,000 survivors who got a severe dose of radiation comes to only a few hundred, and certainly not as many as a thousand.
Three things seem to be going on here. First, there is the confusion between fatalities and casualties--that may well be how the original jump from 90,000 to 130,000 took place. Secondly, there is the problem that once a figure has been widely circulated, it ceases to amaze, and there is a very human tendency (especially among journalists) to hype it a bit: you want the reader to say wow! (I first encountered this phenomenon when I followed the growth of the Flying Tigers' victory claims over the years.) Thirdly, there is a strong constituency for anything that serves to demonize the United States in world affairs--a constitutency that exists not only in Japan, as the victim of the bomb; and in Europe, resenting America's dominance in world affairs; but also in American universities and journals of opinion.
Take them all together, and they seem to have exaggerated the death toll at Hiroshima by more than 100 percent.