They Were Watching for Danger. She Was Watching for Cause.

Kilo was the most complex. He had stopped eating because the feeding schedule the handlers were using had been inadvertently synchronized with the ventilation unit's on-cycle, which meant every meal arrived with the high-frequency sound, and he had begun to associate food with the aversive stimulus rather than with the handler. The behavioral deterioration was not aggression. It was learned avoidance expressed as withdrawal.

I found all of this in the first three hours.

The documentation took another two hours and was, in the way of documentation that is going to change several subsequent decisions, more precise than what the reporting officer receiving it initially expected from a source she had dismissed as a civilian ER nurse on a temporary pass.

Mercer read it at 0700 the following morning.

She came to the kennel at 0730, where I was working with Rex through the fence in the methodical way that behavioral rehabilitation works — not dramatically, not quickly, but with the specific consistency that allows an animal to update its threat assessment about a particular presence over multiple exposures. Rex had reduced his pacing by approximately sixty percent in the preceding six hours. Kilo had eaten, without handlers present, at 0430.

Mercer stood outside the outer fence and watched for eleven minutes before she spoke.

"The ventilation unit is being replaced today," she said. "Facilities confirmed the oscillation range this morning. The explosive residue is being cleared by chemical decontamination team."

"Good," I said.

She was quiet for a moment. "The feeding schedule adjustment for Kilo. How did you—"

"The incident report timestamps," I said. "His refusal events clustered in the nine minutes following the ventilation unit's on-cycle. Once you see the pattern, the mechanism is straightforward."