For thirty-eight years, Dora Kessler had been at the corner of Elm and Fourth every school day from seven-forty to eight-fifteen and again from two-fifty to three-thirty. She had seen the neighborhood change seven times by her count, measuring change by which businesses stayed and which names on the mailboxes changed. She kept this information in her head because it was the kind of information no one else was keeping. The children were the constant, which surprised people when she told them. Not the same children, obviously, but children in aggregate behaved the same way they had in 1986: the ones who ran ahead, the ones who dragged behind, the ones who stopped to show her things they had found on the ground. She made a point of looking at whatever they showed her. A bottle cap. A feather. Once a very small frog that had somehow made it three blocks from the nearest pond. When the city sent her a retirement notice, she showed up anyway. The school principal came out to ask her to stop. She explained that she had a crosswalk and children and a stop sign and she was not sure how the retirement of a person canceled any of those things. The principal went back inside. She stayed until spring, when the city quietly reassigned a replacement without making it an official transfer.

Ellen had spent twelve years calculating load tolerances for highway bridges, which meant she had spent twelve years thinking about the point at which a structure fails. Not collapses, which was the dramatic end of a longer process, but the earlier moment when the margin between holding and not holding became thin enough to matter. She knew this moment was invisible to everyone crossing the bridge. She knew it was not invisible to the bridge. On a Wednesday in November she drove to the Harlan Street overpass because she had read in the inspection report that its load rating had been downgraded and she wanted to see it with her own eyes. She stood on the sidewalk above the traffic deck and put both hands on the railing and felt the vibration of vehicles passing below. The math she did in her head was the same math she did at work. The conclusion was the same: holding, for now. She drove home and filed a formal recommendation for accelerated maintenance through the standard channel, which she had done twice before for other structures. She did not tell anyone about standing on the bridge. She did not entirely understand why she had gone. What she wrote in the report was technical and accurate. What she knew that she did not write was that there is a difference between a structure that is safe and a structure that will not fail, and she spent her career working in the gap between them.