The salvage inventory was the first voice. When Flight 117 was recovered from the marsh, the investigation had catalogued the wreckage by component and serial number with the thoroughness that investigators bring to the mechanical evidence even in a weather accident, and the catalogue had been filed with the safety board's docket and forgotten with it, a public document that had sat in the open for the entire twenty-eight years, unconsulted by the podcast, the families, and the company alike, because everyone believed they already knew what it contained. It contained the weather radar transceiver, recovered from the nose section, its manufacturer's serial plate legible, the number entered in a clerk's hand in 1998 and scanned into the docket in 2007. The review's technical member took the number to the manufacturer, whose service registry kept, for every radar unit it had ever built, the biography of the unit: manufacture, sale, installation, removal, overhaul, the tail numbers it had flown under like a man's addresses in a census. The transceiver in the marsh had been installed, the registry said, in the airframe the company called Ship 309, in a scheduled change-out the previous autumn. It had never been installed in Ship 214 at all. Ship 214's own radar unit, a different serial, had been removed from that airframe in a hangar in Houston in 1999, overhauled, and fitted to a third airplane, which had flown it without incident for another decade. The component the investigators had lifted out of the Louisiana marsh with their own hands, and catalogued, and filed, was proof in metal that the airplane at the bottom of the report's first page was not the airplane at the bottom of the marsh. Scott heard this the way the families heard everything that season, read aloud in a room by a person without a history in it, and what he felt was not triumph but a species of vertigo he had come to recognize as the season's signature: the truth had been in a public docket since he was seventeen. Anyone could have read it. The report's authority had simply made reading it unnecessary, and that was the mechanism, the whole mechanism, by which a reconstructed archive had governed two families' lives: not secrecy, exactly, but the settled belief that the record had already been read.
Ship 309's own paper completed the indictment of the word reconciliation. Her maintenance trail survived outside the company's operational archive, in the manufacturer's records and in the files of a contract maintenance provider the merger had discarded, and it showed the radar discrepancy in the plain handwriting of line mechanics: intermittent returns, loss of picture in precipitation, three entries across five weeks, each deferred under the minimum equipment list with the notation monitor and report. The last entry was dated April 8, 1998, the day before the accident, initialed by a mechanic whose statement the review took in due course, an old man in Baton Rouge who remembered the airplane and remembered the notation and said, with the care of a man who had had twenty-eight years to understand what he was remembering, that deferred radar faults were the common cold of that fleet, that he had written monitor on fifty of them, and that no mechanic of his acquaintance had ever believed a deferred set could not see weather at all, only that in heavy rain it painted a liar's picture, generous in some places and blind in others. The review was careful with him, and careful in its drafts, and the care became one of the things Scott respected most about Hutchinson's panel: they refused, at every stage, to let the radar become the story's new villain in place of the old ones. What the evidence showed was an airplane with a known, documented, imperfect weather radar, dispatched on a night when the release it carried had been written for other weather and another machine, over the written objection of the dispatcher and the twice-entered misgivings of its captain. Whether a perfect radar would have saved it, the panel wrote, no honest man could say; the line that night had contained cells that had killed better-equipped aircraft in better decades, and the crash remained, in its physics, what it had always been: weather, and a tired system, and men working at the edge of what the tools would tell them. What was not uncertain was the record. The record had assigned the loss to the wrong airplane, and in doing so had converted a system failure into two men's disgrace, because a disgrace can be filed, and a system cannot. Scott, reading the technical memorandum at his counter in the by-now customary posture, found himself thinking of the transceiver's serial number, entered in a clerk's hand and waiting twenty-eight years in the open, and of his father's sentence about witnesses: the record was the only witness that could not be frightened. It had not been frightened. It had been mislabeled, which was worse, because a mislabeled witness testifies forever, confidently, to the wrong man's name.