Scott was not sent; witnesses went home, and the investigator went, a patient woman named Okafor from the panel's staff, who reported the visit in a memorandum that Scott read later with the peculiar sensation of reading about a room in his own house he had never entered. Lillian Cho had received her in a small house in Slidell with a porch fan and a cedar chest at the foot of the guest bed, and she had made tea, and she had talked for an hour before she opened the chest, about the New Orleans station in 1998, the teletype room with its smell of oil and hot paper, the way traffic from dispatch, from the tower, from the out-stations, had come over the wire in strips and been typed, filed, and forgotten by everyone except the women who typed it. "The company had a rule," she said, in the investigator's careful paraphrase that somehow kept the voice intact. "Copies of everything, carbons, one for the station file, one for the wire room. When the merger came they sent a man with a scanner and a shredder list, and the station file went into boxes, and some boxes came back from the warehouse in 2004 with the wrong things in them and some did not come back at all. I had taken my carbons home for years by then. Not from spite. A typist keeps her carbons. It is how you prove what your hands did. That night my hands did the traffic for Flight 117 from six in the evening until the news came, and afterward I put the strips in a folder and I put the folder in the chest, because throwing away other people's words, that night of all nights, felt like a second accident."
The carbons were what the archivists call a near-complete run: the station's teletype traffic for the New Orleans operations desk for the evening of April 9, 1998, time-stamped in the wire's own hand, and they told the night the way the night had actually happened. At 17:05, the original release, for Ship 214, Robert Ellery's weather picture, a line building west of the route with the flight planned ahead of it. At 18:20, Ship 214 out of service at the gate, a mechanical fault in a cargo door actuator, the maintenance estimate running long. At 19:05, the station's request for instructions; at 19:26, system operations substituting Ship 309, inbound from a maintenance ferry, with its discrepancy list attached, the weather radar line reading intermittent, deferred, monitor. At 19:42, riding the wire like a man nailing something to a door, the dispatcher objection, R. Ellery, the same text the sealed file held, now arriving with its transmission record: received, logged, acknowledged by the station. At 20:15, a query on the wire from the flight's captain, P. Keneally, asking dispatch to confirm the line's movement and requesting that diversion planning for Alexandria and Baton Rouge be kept open; at 20:40, a second Keneally message, asking whether the hold was to be extended past the line's passage, and the answer, from system operations, not from Ellery, that the departure banks required the 21:52 slot. At 21:52, departure, on Ship 309, into the weather that the 17:05 picture had promised would still be west and that the 19:42 objection had promised would not. And then the wire going quiet in the way wires do at the worst hour, and the last strips typed by a woman whose hands hurt for a week, in a station that no longer exists, on paper that survived by being taken home in a purse.