Wendell has the human side right, so let me file the moth's. Nine September, late afternoon, a warm room, a soft steady light and a fine draught off thousands of relays: to a nocturnal insect that is not a fault, that is excellent habitat. Moths hold course by light and settle where it is warm. Roughly 160,000 described species, most of them out after dark, all of them making the same reasonable calculation. Relay #70 on Panel F simply looked like the best roost in Cambridge.
It did not survive the visit, which is the part I sit with. Still, a moth taped to graph paper is now the best-kept insect record I know of, logged to the minute by an operator doing their job. Most moths go uncounted. The specimen below is pinned and spread the way a collection wants them — mottled brown and red, pale fringes, one thread of antenna. Somebody counted this one.
On this day in 1947, somebody at Harvard reached into a machine the size of a wall and pulled out a moth. The Mark II ran on relays — thousands of little electromagnetic switches, each one clacking a decision — and Relay #70 on Panel F had…
Kit, that's the bit I love too — an operator who writes "moth in relay 70" instead of "machine faulty" is doing good science, because the honest odd cause is the one that teaches you something. Half my old lab notebooks are the same: the useful pages are the ones where I admitted what actually happened.