Chapter 74: A Battery Nobody Logged

Rain ticked across the service roof while Bell locked the six returned carts into a numbered line. Gray Flood District mud remained in their seams. The empty service lane beyond them showed what the examination already cost: one evening delivery delayed, every returned cart held until the control run was complete.

Eli stayed behind the yellow cable. Bell controlled every connection and primary display inside the evidence boundary. His authorized mirror showed measurements after a short delay and allowed no terminal input, annotation access, or custodial signature.

Ortiz entered the cost before Bell touched the load bank.

“Twenty-four charger minutes,” he read. “Six carts temporarily held. Five available for release if the control clears them. One delivery delayed.”

He set a charged coalition pack beside the custody terminal. Its stencil read COAL-B-2086.

“Continued powered custody also requires this pack,” Ortiz said. “Transferred from the South Loop mobile reserve. That reserve opened with three charged packs and falls to two.”

Voss looked up from the entry. “Threshold?”

“Two keeps the Jardine medical route provisioned against one vehicle-pack failure. Another draw weakens or suspends that route.”

Nothing in the record called the expenditure routine. Voss signed beneath the threshold warning.

Bell began with C1. Eighteen amps for forty seconds, using the same leads, meter, and temperature correction. She repeated the procedure on C2, C4, C5, and C6. Their voltage drops and recoveries remained within their respective service histories.

C3 currently held R4-B-6634. Bell read the serial to the camera before attaching the load bank.

B-6634 sagged under load and recovered in shallow steps. Bell repeated the run after the prescribed rest interval, producing the same general response. She then tested the separately cradled R4-B-7719 under identical conditions. Its recovery also rose by shallow steps.

Voss studied the paired traces. “Does that establish B-7719 occupied C3?”

“No,” Eli said from behind the cable.

Bell waited, leaving the curves displayed.

“Those are two different packs,” Eli continued. “The resemblance shows the slot and test conditions can produce a comparable demand-and-recovery pattern. It validates C3 as a control. It cannot tell us which pack occupied C3 before B-6634.”

The stronger claim would have been useful. It would also have been wrong.

Voss nodded once. “Record the rejection.”

Ortiz entered: curve comparison not probative of historical serial occupancy.

Bell closed the live-test file and opened C3’s preserved powered slot history. Unlike the unrestricted mirror, the source display remained inside the boundary. Ortiz independently read each relevant field before Bell released a bounded extract to Eli’s screen.

The slot ledger identified R4-B-7719 by serial. It recorded continuous powered recognition through C3’s pre-departure cycle and a service release at 01:51:47. No intervening serial appeared before the release. At 01:52:14, charger three accepted R4-B-7719.

“That establishes the first movement,” Eli said. “C3 to charger three.”

“Through powered serial records,” Voss said.

“Yes. Not through the curves.”

Bell opened C3’s compartment under camera. Fresh brass showed through corrosion on the right latch. Blue serial lacquer remained inside a recent gouge, and magnification resolved the surviving digits as 7719. The damage geometry aligned with a fresh scrape on B-7719’s lower casing.

Below the fresh mark, older parallel scars crossed the cradle. Bell mapped them, changed the lamp angle twice, and photographed their edges.

“Mechanically separate earlier movement,” she said. “Not datable from the scars and not attributable.”

Ortiz preserved that limit with the finding.

Compressed paper fibers also clung beneath the cradle. A double pressure ridge ran through them at the same spacing and angle as the preserved reference image from R4-RI-22.

Bell collected the C3 material as sample C3-R-1. She removed the previously authorized reference lift RI22-R-1 from the counterfoil’s nested evidence sleeve without opening the primary counterfoil seal. A contact lift from B-7719 became sample B7719-R-1.

Under magnification, C3-R-1 and RI22-R-1 shared the same blue-gray paper fiber mix and compressed coating residue. B7719-R-1 carried matching coating particles along the casing edge corresponding to C3’s pressure ridge.

Bell divided each sample. One half received the depot seal, the other the coalition seal. Ortiz verified all six numbers before the samples entered separate custody envelopes. R4-RI-22 itself remained sealed. No second counterfoil had been found.

“The recovered counterfoil traveled under pressure with the pack while B-7719 occupied C3,” Bell said. “The residue does not establish who placed it there.”

The search for R4-B-4412 came next.

Ortiz ran the powered battery inventory while Bell and a depot mechanic independently checked physical locations. B-4412 did not appear in the powered inventory. The return rack held seven packs; every serial was read aloud and photographed. None was B-4412. Bell then checked active charger bays one through four against their live serial returns. All four were occupied by identified packs, none matching B-4412.

“Powered inventory excluded,” Ortiz said. “Return rack excluded. Active charger bays excluded.”

The pack remained missing.

Charger three supplied the middle of the transfer. It had accepted B-7719 at 01:52:14 and recorded manual removal at 02:04:32. C7’s preserved history showed B-4412 disengaging at 01:59:06.

A separate event appeared before B-7719’s acceptance.