Enrico Fermi estimated the yield of the Trinity test to be about 10 kt by dropping small pieces of paper and observing their motion in the blast wave. This is about 40% of the radiochemically derived value of  kt that necessarily includes thermal and nuclear radiation that do not contribute to the blast. Although this story is classic, there appears to be no account of how he related his observation to the yield. This paper attempts to reconstruct how he might have done so.