Chapter 35: The Shape of an Absence

The six couplers matched their individual certifications.

That was the problem.

Orra removed the last unit from its transport sleeve and placed it beneath the evidence hood. Six foam recesses filled. The seventh remained empty. Above the work surface, the turnover clock showed thirty-eight minutes in red numerals.

Warm polymer and old machine oil moved through the filtered exhaust. The couplers were externally alike: compact bodies, reinforced noses, indexed pins protected by guide shoulders. Orra aligned their identification plates without allowing any two units to touch.

CR-2/01 through CR-2/06 were present.

CR-2/07 was not.

Vale set a portable supply beside CR-2/04. The leads remained coiled, but the supply altered the room merely by being available. A powered test might produce an answer. It might also erase the conditions under which the answer could be trusted.

“A functional pulse would settle compatibility,” he said.

“No.” Mara opened the joint record. “Energizing a suspect coupler may alter charge history, contact residue, or maintenance memory. I object to any powered test. Dimensional, residue, and mass comparisons only.”

Vale left his hand beside the case for a moment. He did not offer reduced voltage, limited duration, or a sacrificial lead. Those arguments concerned control of the test. Mara’s objection concerned survival of the evidence.

Orra pushed the supply beyond the marked examination boundary. “Good. I dislike findings that depend on evidence surviving our curiosity.”

She began with the cast taken from SJ-4C-12’s emergency service face. It preserved the shallow access geometry, contact alignment, and parallel pressure marks left by guide shoulders. The cast did not identify whoever had made the marks. It retained only what pressure had done to metal.

CR-2/04 came near the cast without touching it. Orra lowered the examination light until the broad reflection narrowed along the edges.

The guide shoulders matched the pressure marks in width and spacing. The indexed asymmetry appeared in both the cast and the coupler. When Orra lowered the measuring arm, the pins reached the established service contacts without requiring the controller housing to be opened.

Vale leaned toward the hood glass. “So the housing remained sealed.”

“The coupler did not require the controller housing to be opened,” Mara said. “That is what the geometry proves.”

Orra entered the narrower statement.

The service face could be reached directly through its established emergency access point. No remote route, impossible bypass, or opening through the controller body was necessary.

The physical fit also had to survive the timing.

Mara placed the raw junction record beside the cage sampler. Their summaries had treated different thresholds as if they described the same boundary. Orra discarded the summaries and aligned the underlying traces.

Contact and external feed lasted eight seconds.

The junction monitor blank occupied the middle six seconds, beginning after current rose and ending before it fell. The cage alarm remained above threshold for four seconds inside that blank, while the raw sampler retained the lower rise and decay around it.

Orra overlaid the intervals.

Eight seconds of physical contact contained six seconds of reporting loss. Those six contained the four seconds above the cage alarm threshold. No revised clock or invented delay was required.

“Eight connected,” Mara dictated. “Six unreported. Four above threshold. Distinct measurements of the same event.”

Orra read the finding back and locked the timing panel against casual amendment.

Vale studied the traces. “The four-second figure never contradicted the contact interval.”

“It contradicted calling the threshold interval the whole event.”

He nodded once. The clock passed thirty-one minutes.

The removed and reseated seal rested on a separate evidence pad. Under magnification, its adhesive pattern showed controlled handling rather than random failure. Someone had lifted it carefully enough to preserve apparent continuity and then placed it back.

No authenticated feature dated the second seating.

Vale glanced toward it. “It establishes access.”

“It establishes removal and reseating,” Mara said. “Not when either occurred. We do not lend a dated event confidence borrowed from an undated one.”

Orra left the seal separately illuminated and separately logged. Exclusion from the intervention finding did not erase it. It prevented an uncertain time from being made precise by proximity.

She turned to the coupler noses.

Five carried flat storage oxidation. CR-2/04 bore a narrow paired arc where contact had broken through the dull surface. Gray residue had collected beside the marks and within a shallow recess.

Orra sampled the minimum permitted area. The residue was chemically consistent with material from SJ-4C-12’s service face.

Consistent was not unique. The pattern did not identify a hand or a time. It did not prove that CR-2/04 had been inserted during the eight-second intervention.

“Compatible unit present,” Mara dictated. “Actual use unproven.”

Vale did not ask her to strengthen it.

The turnover clock showed twenty-seven minutes when Orra began the mass comparison.

Each CR-2 had a certified individual mass within a narrow service tolerance. She verified the evidence balance, set the reference, and weighed the six disconnected units one at a time.

Every reading matched its certification.

Their combined mass also matched the tray’s amended aggregate figure.

It did not match the original aggregate for seven units.

Orra opened CR-2/07’s certification. “One point eight four kilograms.”

The difference between the original tray total and the amended total was exactly 1.84 kilograms.

No rounding correction produced that subtraction. Packaging tolerance could not explain it. The amendment had removed from the aggregate record precisely the certified mass of the missing unit.

Vale opened the visible inventory view. It continued to call for seven active CR-2 couplers, while the amended tray total described the mass of six.

“That proves the amendment normalized the reduced tray,” he said.