One of the chief downsides of getting older—I am now seventy-two—is that one’s friends die. On a Friday evening in early August, near midnight, it was the turn of Donald Kagan, aged eighty-nine. I am still having trouble accepting that he is gone.

I first met Don in the spring of 1968. I was a freshman, then, at Cornell. He was teaching an introductory course in Roman history—as it happens, for the very last time. I was enrolled in the class and assigned to his section, and I was mesmerized. Don was an entertaining and provocative lecturer. He knew when and how to introduce the ham, and what he had to say was invariably interesting and informative.

The section meetings in his course were organized around historical puzzles. We were asked to read the evidence and to try to make sense of it; thanks to its paucity, we could spread it all out in front of us. We were in the position of intelligence analysts at Langley. We knew odds and ends—in our case, the flotsam of fragmentary information carried down the ages by time—and we were called upon to make sense of it all. There was rarely an obvious right or wrong. There were, instead, the plausible, the remotely possible, and the completely absurd. We were not simply memorizing the facts. We were doing what all historians do. We were trying to describe what must have happened on the basis of a documentary record limited in