Tyrell refused.
Castellano continued questioning him for several minutes before telling him to empty the bag “so everybody could go home.”
Textbooks.
A laptop.
A water bottle.
Two granola bars.
Nothing illegal.
Kowalski’s finding had read:
Officer conducted reasonable field inquiry. Complainant likely misunderstood officer safety procedure.
Captain Avery read the line twice.
“What officer safety procedure?”
Lieutenant Cho shook his head.
“There isn’t one that says a citizen walking home empties a backpack because an officer wants reassurance.”
They reopened all nine complaints.
Eight complainants agreed to be interviewed.
The ninth had moved overseas and could not be reached immediately.
The interviews were not identical.
That mattered.
Investigators were not looking for a script.
Some complainants admitted Castellano had been polite.
Others remembered him as confrontational.
In two encounters, officers had received legitimate service calls before Castellano arrived.
But the same decision-making pattern appeared repeatedly.
Ambiguous circumstances.
Person of color.
Assumption of suspicion.
Demand for proof.
Escalation when questioned.
Little or no enforcement result afterward.
Then the data analysts looked at Castellano’s activity.
The raw numbers alone could not prove discrimination.
Everyone involved understood that.
Patrol zones differ.
Calls for service differ.
Time of day matters.
Deployment patterns matter.
A responsible review could not simply count racial categories and declare bias.
So they controlled for district, call type, shift, location, and officer assignment.
That was when the disparity became harder to explain.
Over four years, Castellano had taken suspicious-person calls substantially more often than peers working comparable shifts in comparable zones.
His field contacts following vague quality-of-life or suspicious-person complaints disproportionately involved Black and Latino residents.
More importantly, those contacts produced unusually few arrests, citations, recovered stolen items, or documented offenses.
He was frequently finding suspicion.
He was rarely finding crime.
Captain Avery wrote one sentence on her legal pad.
High intervention, low evidentiary yield.
The department hired an outside statistician to verify the analysis before placing it in the final report.
The conclusion was cautious.
The pattern did not, by itself, prove discriminatory intent in every stop.
Combined with repeated similar complaints, body-camera footage, and the Whitmore incident, however, it raised a serious and credible concern that race was influencing Castellano’s assessment of who appeared suspicious or out of place.
Then Internal Affairs turned toward Lieutenant Vernon Kowalski.
Kowalski had retired six months earlier after twenty-seven years.
His retirement party had filled an entire banquet room.
Officers told stories about how fiercely he defended his people.
He had been called an officer’s officer.
Darnell had heard the phrase throughout his career.
Sometimes it meant a supervisor who stood beside good officers when political pressure became unfair.
Sometimes it meant something else.
Kowalski’s complaint files showed what it had meant in East Division.
There were 127 citizen complaints he had overseen during his final five years.
Ninety-four involved discrimination, improper detention, disrespect, unnecessary escalation, or force.
Only three had resulted in sustained findings.
That number alone was not proof of misconduct.
So the department sampled thirty files for a deeper audit.
Investigators compared reports with body-camera footage, dispatch logs, GPS records, complainant interviews, and available third-party video.
In eighteen of the thirty, the original investigation had significant shortcomings.
Not always because the officer had clearly violated policy.
Sometimes the problem was simpler.
Kowalski had never done enough work to know.
He frequently accepted officer reports without interviewing officers separately.
He failed to review body-camera footage unless a complainant explicitly referenced it.
He sometimes characterized disputes as “misunderstandings” before interviewing independent witnesses.
In three Castellano cases, he had closed allegations without questioning Castellano at all.
Captain Avery interviewed Kowalski at his home as a retired member.
He agreed voluntarily.
He was sixty-two, gray-haired, broad-shouldered, still carrying himself like a lieutenant.
“I protected good cops from nonsense,” he said.
Avery placed several files on the table.