Movement Overhaul Quoted at Two Months; Collector Reevaluates Entire Concept of Linear Time

"Two months," Silas Tilling said, and wrote it in the remarks box of the service ticket, where other shops write "regulating."

The watch went in on a Thursday: a Vaudrey manual-wind, bought used thirteen years ago, last measured at gaining eleven seconds a day, dial up. It stopped mid-wind, which the collector reported as a fault in his own technique. Tilling, who has handled worse confessions, told him to leave it.

The report came back Monday. "Main spring has let go," Tilling said. "Lubricant is past use. Escapement shows wear consistent with age, which is the polite way of saying it has worked." The collector asked whether it might be the beat error. "Beat error is what we look at after," Tilling said. "Two months, because the barrel is coming from a dealer in Ghent, and he is slow in August."

Nine years he has worn that watch. He did not expect the wrist to be the part that complained. At traffic lights he lifts an arm and reads nothing. He catches himself starting sentences with "since the watch went in," as though it had left willingly. The first day he described himself to a colleague as being "in a service state," and the colleague asked whether that was a diagnosis.

At the meeting he reported the situation in the vocabulary he had: amplitude, endshake, the words he had picked up standing too close to other people's conversations. Inez Bellamy, who left the group in protest and has not missed a meeting since, said she has had three watches in service at once. "You learn to tell time by other people's obligations," she said. "Someone is always five minutes out. It's a system."

Basil Keene offered the club's loaner, a quartz diver kept in a drawer for exactly this purpose, accurate to two seconds a month. The collector wore it home and opened a notebook. For eleven days he logged its deviation by position and temperature, carefully, before admitting he had never once measured his own watch without a phone app held at arm's length.

He also began attending other people's timing claims with new interest. A man at the next table said his watch gained eight seconds a day. The collector asked about winding state, ambient temperature, whether the eight seconds were dial up or crown down. The man had not known there were positions. He did not know it either, three weeks ago, when he owned a watch that needed him to know it.

He is precisely on time now. He arrives early to things he used to arrive late to, holding the phone face down, saying he is waiting on a part. There is a second watch in the drawer, a microbrand automatic he ordered in week two and described to Tilling as redundancy rather than accumulation. Tilling said nothing, which is a form of service advice.

At week six the call came. The barrel is wrong by a tenth of a millimeter. Three more weeks. The collector said fine, and asked how long the whole thing takes, and Tilling said two months, still, as though the first two months had been a deposit.

That night the collector set an alarm for the pick-up. Not 9:40. 9:39:50, because the watch had always been wound to run five seconds fast and he wanted the first reading on the wrist to match the first reading on the phone. He has never in thirteen years set an alarm for a wrist. The watch, in its box under a bench light in a closed shop, is not responsible for any of this.