Chapter 82: The Part That Came Later

Mara started the timer at 03:09:12.

The digits advanced without regard for preparation or argument. She placed the display where all three of them could see it without turning from the reconstruction area.

Every item had been logged, photographed, and positioned within the authorization. Kell waited at the stock threshold with an inert relay matched to ES-17-44’s dimensions, mass, and uneven balance. Vale stood beside C-8 with the logged tools. Inside the housing, a removable analogue reproduced the original CR-6 geometry, contact placement, retainers, and seating resistance.

Nothing authenticated would be cut, forced, or unsealed.

The compartment smelled of warm insulation and old seal compound. Heat from the working lights collected under Mara’s collar. Her support strip pressed against her bad knee, and her wrist brace prevented her from safely carrying either analogue.

Vale had raised that before the test.

Mara had entered it as a limitation, not an excuse. A run performed by her would measure an impaired operator. Kell and Vale would establish the fastest safely repeatable sequence; her task was to keep speed from becoming invention.

“Fastest permitted route,” she said. “Not fastest imaginable.”

“No bypasses,” Kell said.

Vale glanced at the tools. “No unlogged handling.”

“And no assuming the result.”

She released them.

Kell registered the withdrawal and took the maintenance passage at a controlled run. Sprinting might have saved seconds, but not while carrying a relay of the specified mass through blind turns. He held the inert body against his torso, rotated it through the narrow section, and slowed where the passage folded around a pressure trunk.

The withdrawal acknowledgement was quick, not instantaneous. Relay and carrier had to pass fully beyond the stock threshold before the event counted as complete.

Kell reached C-8 at two minutes eleven seconds.

Vale moved when he crossed the marked station boundary. The first tool seated with a dry metallic click. The cover came free, releasing a breath of hotter air and the bitter odor of aged contact resin.

The CR-6 analogue resisted at the older seating points.

Vale’s first pull produced no useful movement. Rather than drag metal across the edge, he changed grip, supported the mass from beneath, and released the rear contact. The older geometry required a lift and partial rotation instead of the direct withdrawal used by later units.

Seconds accumulated in necessary motions.

Once clear, the heavier analogue had to be carried to the preservation tray and lowered into its supports. Vale neither dropped it nor leaned it against the station.

Four minutes thirty-eight.

Kell presented the replacement. Vale received it with both hands, turned it, and approached the open housing. One guide seated shallow. He withdrew before it could bind, corrected the angle, and engaged the contact block.

The timer passed six minutes.

Vale restored the cover. One edge settled; the other remained fractionally high. He shifted pressure, checked the contact line, and brought it flush.

The alarm sounded before the first closure check.

Seven minutes.

Its brief tone left the compartment abruptly quiet.

Vale kept his hands beside the unsecured cover. Kell remained outside the working boundary. Mara recorded the state at expiration: withdrawal complete, transport complete, original removed, replacement fitted, cover restored but unsecured, closure unverified, seal unverified.

“I could close it faster on another attempt,” Vale said.

“You may. The first clean run remains the first clean run.”

“The stock step is carrying too much of the interval,” Kell said.

“It is part of the claim.”

They restored the starting configuration. Vale reseated the CR-6 analogue and re-engaged its contact. Kell returned the inert relay to the stock side and isolated the test withdrawal from the evidentiary log.

Once every item was accounted for, he came back to C-8.

“Maintenance variation,” Kell said. “Remove withdrawal registration. Assume the replacement was already outside stock.”

“Recorded as an assumption.”

Mara entered it before resetting the timer. The variation could not amend the authenticated access record. It tested whether the work at C-8 fit if one required step had occurred earlier or elsewhere.

Vale opened the housing as the second run began. Familiarity with the analogue was unavoidable, so Mara marked the attempt as practiced.

The same seating points resisted. Vale released the rear contact, supported the greater mass, and rotated the analogue clear.

Removal took two minutes twenty-four.

Kell presented the staged replacement. Vale aligned the lower guides on the first approach, but the upper seat remained shallow. He withdrew rather than force it, corrected, and seated it under controlled pressure.

Fitting took two minutes fifty-one.

After engaging the contact block, Vale replaced the cover and secured each point in sequence. The closures answered with four separate clicks. He traced the edge, found no proud corner, and verified the seal position without treating appearance as proof.

Total time: seven minutes thirty-six.

The housing was complete, but the permitted interval had expired thirty-six seconds earlier.

Vale opened the sequence breakdown. Delay had not come from one dramatic obstruction. It lay across acts that could not occupy the same second: supporting the original, clearing its rear geometry, correcting the replacement seat, securing the cover, checking the seal.

“Assume the original was loosened earlier,” he said.

Kell looked toward C-8. “Retainers released?”