"What emerged, of course, was that the magnitude scale presupposed that all earthquakes were alike except for a constant scaling factor. And this proved to be closer to the truth than we expected."