The survivors of the A-bomb explosions over Hiroshima and Nagasaki were exposed to a mixed neutron and gamma radiation field. To validate the high-energy portion of the neutron field and thus the neutron dose to the survivors, a method is described that allows retrospective assessment of the fast neutrons from the A-bombs. This is accomplished by the extraction of the noble gas argon from biotites separated from Hiroshima granite samples, and then the detection of the Ar-39 activity that was produced by the capture of the fast neutrons on potassium. Adjusted to the year 1945, activities measured in the first samples taken at distances of 94, 818, 992, and 1,173 m from the hypocenter were 6.9 +/- 0.2, 0.32 +/- 0.01, 0.14 +/- 0.02, and 0.09 +/- 0.01 mBq/g K, respectively. All signals were significantly above detector background and show low uncertainties. Considering their uncertainties they agree with the calculated Ar-39 activation in the samples, based on the most recent dosimetry system DS02. It is concluded that this method can be used to investigate samples obtained from large distances in Hiroshima, where previous data on fast neutrons are characterized by considerable uncertainties. Additionally, the method can be used to reconstruct the fast neutron fluence in Nagasaki, where no experimental data exist.