Du Pont, which ran the whole proposition, was careful about keeping big pile-ups of people. And the start-up had Fermi and a few assistants of Fermi around, and Crawford Greenewalt, who was the Du Pont head man, later became President of the Du Pont company, but loved to calculate, with a slide rule, some of the formulas. At any rate, I wasn't out there at the start-up, I was in at the laboratory center. But then I heard about the difficulties pulling the control rod out, the pile started up, fine, and got going better and better and so on. And then it died the death, and then came back to life. And I concluded that all these strange features could only be understood if there were two radioactive nuclei involved, one the mother, innocent, two, the daughter, poisonous to the reaction. The mother was formed in quantity by the operation of the pilot power, but it took several hours for the mother to transform into the poisonous daughter, and then that absorption of neutrons produced by the poisonous daughter slowed down the reaction and caused reactivity to disappear. But wait a few hours, and the poisonous daughter disappeared and the reactivity came back. But the lifetime for transformation from mother to daughter and from daughter poisonous to daughter dead, those two lifetimes have to be each one of several hours duration. I could look at the chart of nuclei on the board and see that there were only two candidates that would fit the bill. One was xenon, for the poisonous daughter, and the other, iodine, for the innocent mother. Did that analysis all take place in the course of a single day? Right. Fermi was always a great skeptic on everything, but when he came into the lab, and saw these calculations, he had to agree that that made sense.