Cold

The memo is four pages long and carries two dates. The first is the date it was written. The second, handwritten in the upper corner and circled, is the date of the original incident. Fourteen years separate them. Between those dates the file passed through the hands of at least six detectives, two supervisors, one records clerk who was trying to clear shelf space, and a property room that moved buildings once.

The memo you are about to read is a composite. I built it from the structure and language of real cold-case review memos, and it is labeled that way here and everywhere it appears in this book. No single family's case is being described. The headings it uses, the order of its sections, and the kinds of things it finds are drawn from the documents themselves. So is the tone, which is flat, procedural, and very careful.

It begins:

Case reassigned to Cold Case Unit for review. Original lead investigator retired. Secondary investigator deceased. Case last actively worked approximately eleven years prior. Most recent file entry: family contact, telephone, logged by desk officer. No action noted.

Read that last line twice. A parent called, someone wrote down that they called, and nothing followed. The memo does not say what the parent asked. It does not need to. You can probably guess, and so could the reviewer, who underlined the entry in pencil.

Ask the three questions. The memo says the case was inactive. It misses the reason, because no form in the file has a field for why work stopped. Nobody ever writes we stopped here. Entries simply get further apart until they end. The question it forces is the one the reviewer wrote in the margin of her own copy: Stalled or stopped?

The two are different. A case stalls when the leads run out. A case stops when the people run out.

Most readers have heard that a case "went cold" and pictured a wall: investigators exhausted every avenue and nothing remained. Sometimes that is true. Often the file tells a quieter story. The lead detective was carrying eleven open homicides when this one came in. He worked it hard for four months, according to the supplemental reports, and then the reports thin to one a month, then one a quarter. In the second year a new case arrived, a double homicide with a suspect in custody and a trial date, and the child's file went into a drawer. Nobody decided to abandon it. A calendar did.

This is ordinary. Homicide clearance rates in the United States, as reported in the FBI's Uniform Crime Reporting data, have hovered around half to sixty percent in recent decades. The Murder Accountability Project, a nonprofit that compiles federal and state homicide records, estimates that more than a quarter of a million homicides committed since 1965 remain unsolved. Those numbers come from many causes, and child homicides are cleared at different rates from adult ones. But behind any figure that large is a shelf, and on the shelf are boxes, and each box was once somebody's whole working day.

The reviewer does not start with the evidence. She starts with the paper, and the order she reads it in reveals something.

She reads the first-responder report first, before anything that came after. She reads it slowly, looking at what was written down before anyone had a theory. Every later document was shaped by what investigators came to believe. The first report was shaped only by what the officer saw. She notes a field on the first page that was left blank and writes its line number on a sticky note. She does not yet know whether it matters. She marks it anyway.

Then she reads the victimology, the background on the child herself. In this file it runs three pages. The child was nine and had started piano that fall. She walked the same route to a friend's house most afternoons. The reviewer reads these pages more than once, because they describe the life the case interrupted, and the people who moved through it.

Then she reads the transcript from the interview room, all of it, and especially the pages the original detectives read once and set aside. That is where the last chapter left her, on page seventeen with a car door after two in the morning and a blank where the follow-up question should be.

She writes the page number on a second sticky note.

What she does not do yet is form a view. Review units that work well tend to enforce this as a discipline. Some assign a reviewer who has never met the original investigators. Some require the review memo to list every assumption the original file rested on before it is allowed to list a single new lead. The discipline exists because a cold file is a kind of argument. It has been arguing the same conclusion for years, in the same handwriting, and it is persuasive in the way familiar things are.

The memo's second section is headed File Integrity. It is the shortest section and the hardest to read.

Tip log: present, incomplete. Entries from months 3 through 7 not located. Interview recordings: audio for two of five interviews located; remainder not recovered. Photographs: scene set complete; autopsy set present. Property: see Section 4.

Months three through seven are missing. The memo does not speculate about where they went. Paper gets misfiled. Departments migrate from one records system to another, and the migration project has a budget, a deadline, and a contractor who scans what he is handed. A detective takes a working folder home to review over a weekend and dies nine years later with it in a closet. There is a dry absurdity in how many cold files have been reconstructed from the photocopies one investigator kept because he did not trust the originals to survive. There is nothing amusing about how often he was right.

The reviewer cannot recover those months, so she changes course. She pulls the original detectives' personnel files to find out who else worked the case in that period. She finds a name, a patrol officer temporarily detailed to the squad and long since transferred to another county. She calls him. He remembers the case, as officers usually do when the victim was a child. He does not remember the tips. He remembers that there were boxes of them, and that a sergeant told him to sort them by whether they mentioned the father.

She writes that down exactly as he said it and does not comment on it in the memo. The memo records what was said. Its readers can see where the sorting pointed.

Meanwhile there is the family.

Families of murdered children describe the years of waiting in remarkably consistent language, in court statements, in testimony before legislatures, and in the interviews collected by victim advocacy groups. They describe calling and being transferred. They describe learning that the detective who knew their child's name had retired, and that they found out from a stranger at the front desk. They describe anniversaries spent wondering whether to call, because calling and hearing nothing is worse than not calling. Many describe a single question they asked early on that no one ever answered. They kept asking it for years, until asking became a kind of ritual and the answer mattered less than the hope that someone was still listening.

In the composite file, the mother called on the child's birthday for six straight years. The desk logs show it. Six entries, six dates, one week apart in the calendar each year, each followed by No action noted. In the seventh year there is no entry.

The reviewer reads that silence twice, then makes a call of her own, before she has anything to tell. She tells the mother her name and says the file is open on her desk. She says she cannot promise anything. The mother asks whether the people who worked the case before are still around. The reviewer tells her the truth. The mother is quiet for a long time and then asks the question she asked fourteen years ago. The reviewer writes it down word for word. She does not answer it, because she cannot yet, and she says so.

Section 4 of the memo is headed Property and Evidence. It is where the file stops being about paper.

Every item collected at a scene or an autopsy is supposed to be logged, sealed, and stored under a property number. Each transfer is recorded on a chain-of-custody form, a running signature log that shows every hand that held it. In theory the property room is a place where time stops. A swab sealed in 1990 should come out in 2004 exactly as it went in. That is the whole promise of the room. In practice, rooms flood, freezers fail, and retention schedules are written by people thinking about shelf space and not about technology that has not yet been invented. In 2013, a federal working group convened by the National Institute of Standards and Technology and the National Institute of Justice published a handbook on biological evidence preservation. It was written largely because practices across jurisdictions varied so widely, and items that could have mattered were being lost or disposed of.

The reviewer requests the property inventory. It comes back as a printout with nineteen line items.

She goes down to the room herself. She does not have to. Most reviewers would send the request and wait for boxes to be brought up. She goes because the inventory lists item 7, clothing, victim, outer garment, with a disposition code she does not recognize. The property clerk looks it up in a binder. The code means released for destruction per retention schedule. It was applied during the move between buildings, along with forty other items from cases more than ten years old with no recent activity. The case had no recent activity because no one was working it. No one was working it because the detective had a double homicide with a trial date.

She stands there for a while. Then she asks for everything else.

Items 1 through 6 are present. Item 8 is a set of fingernail scrapings, collected at autopsy, sealed in an envelope inside a box inside a bag. The original seal is intact and signed with the pathologist's initials. Items 9 through 14 are present. Item 15 is listed as swabs (2) and is not on the shelf where the inventory says it should be. The clerk searches for forty minutes. It is in a different aisle under a transposed number. The seal is intact.

Items 16 through 19 were sent to the state lab in the second year, tested with the methods of the time, and returned. The lab report from that testing, which the next chapter will open properly, says the samples were insufficient for comparison. The reviewer reads that phrase and asks what it meant then and what it would mean now. She does not know the answer. She knows that insufficient described the tools as much as the samples.

Look at what the inventory leaves out. It does not record why item 7 was chosen for destruction and item 8 was not. It does not record whether anyone asked the original detectives before the purge. They had retired by then. It records outcomes and does not record decisions. Whatever happens next in this file depends on two envelopes that survived by the luck of a shelf and a transposed number.

The memo's final section is headed Recommendations. In real review memos this section is usually brief, and in this one it runs to six lines.

1. Submit items 8 and 15 for re-examination under current laboratory methods.

2. Locate and re-interview the witnesses identified in transcript pages 12–19, including any neighbors able to address a vehicle in the area after 0200.

3. Attempt reconstruction of tip log months 3–7 from the officer recollection and any duplicate records.

4. Review first-responder report, page 1, unfilled field (see note).

5. Request audit of retention-schedule destructions for all homicide property, 1995–2002.

6. Assign single point of contact for family. Contact monthly regardless of developments.

The sixth recommendation costs nothing and had never been made before.

The fifth will take a year and will not be popular. An audit of destroyed evidence is a list of cases that can no longer be fully reviewed, and someone will have to tell some of those families. The reviewer knows this when she writes it. She writes it anyway.

The first recommendation goes out on a lab submission form two days later. The form asks for the case number, the item numbers, the examinations requested, and the priority. There are boxes for Active Trial, Active Investigation, and Other. There is no box for a case fourteen years old. She checks Other and writes in the margin. The lab's posted turnaround time for non-priority biological submissions is several months, and the backlog notice taped to the intake counter is dated the previous spring.

She calls the mother that afternoon to say that something has been sent, without saying what, and to give her a direct number. The mother writes it down and reads it back. Then the reviewer goes back to her desk, where two other files are waiting. Each has a sticky note on the first page, and each holds an inventory she has not yet had time to request.