The report is two pages long. The first page is a form, and the second is a narrative typed into a box that was never quite large enough for what it had to hold. Someone signed it at 6:52 in the morning, a little more than four hours after the call came in. The ink of the signature is heavier than the rest of the page, as though the pen was pressed down hard.

Before you read it, you need to know one thing about what is in front of you. The report reproduced in this chapter is a composite. I assembled it from the standard fields of American patrol incident reports and from the patterns that appear again and again in the first pages of documented child homicide files. It is not a transcription of any real case, and the child in it is not a real child. I built it this way because the first page of a real file belongs to a real family. On that page, the child is still almost entirely unknown to the people writing about her. Every document after this one in the book comes from the public record of cases that have been resolved in court. This one is a teaching instrument, and I will not pretend otherwise.

Here is the top of the form.

INCIDENT TYPE: Missing Person / Juvenile — amended Death Investigation

DATE/TIME RECEIVED: 0247

DATE/TIME ARRIVED: 0301

LOCATION: [residential address], rear yard and alley

REPORTING PARTY: [name], relationship: parent

VICTIM: V-1, female, W, age 7

WITNESSES: (1) [name], parent. (2)

OFFICER: [badge number]

Read it slowly. Fourteen minutes passed between the call and the arrival. The incident type was changed partway through the night, and the amendment is written above the original rather than in place of it, so you can still see what the officer believed when the car pulled up. The child is listed as V-1. The report gives her sex, race, and age.

She was seven. The report says nothing else about her.

Then there is the line marked WITNESSES. One name follows the first numeral. After the second numeral there is nothing at all. The form is not torn or smudged there, and nothing was crossed out. Someone typed the numeral, meaning to add a name, and never came back to it.

Turn to the narrative.

R/O responded to above location ref. missing juvenile. Upon arrival R/O met with R/P who stated V-1 was last seen in her bedroom at approx. 2100 hrs. R/P stated V-1's window was found open. R/O conducted search of residence and yard with R/P. A neighbor advised R/O she had heard a vehicle in the alley earlier in the night. At approx. 0338 R/O located V-1 in the alley behind the detached garage. V-1 was unresponsive. R/O requested EMS and supervisor. EMS pronounced at 0351. R/O secured scene and maintained log until relieved by Det. [name] at 0420.

The narrative takes nine sentences to cover the worst hour of a family's life.

Start with what the page tells you, because it tells you a good deal. "R/O" means reporting officer, and "R/P" means reporting party, the person who placed the call. "EMS" is emergency medical services, and "pronounced" means a medic formally declared death and recorded the time. The scene log the officer "maintained" is a separate sheet that records every person who crossed the tape and when. It is one of the most underrated documents in any file, and we will come back to it. From this page you have a timeline with four fixed points: last seen at 2100, call at 0247, discovery at 0338, and pronouncement at 0351. You have a location. You know the search began inside the house and moved outward. You know that an open window was the first physical fact anyone wrote down.

You also know something about the officer. This report is competent. The times are precise, the handoff is recorded, and the scene was secured. Nothing here is careless in the way television teaches us to expect careless police to be.

Now read it again for what it leaves out.

The neighbor is the obvious gap. She is the only person in this narrative, apart from the parent, who offers information about the night. She reports a vehicle in the alley. She has no name. The report gives no time for the vehicle beyond "earlier in the night," no direction, no description, and no note of whether she saw it or only heard it. She is not listed in the witness field, and she is very probably the person that empty second line was meant for.

Picture how this happens. It is three in the morning, and a parent is screaming a child's name into the dark. A woman in a bathrobe comes to her back gate and says something to the officer as he passes. He is moving toward the alley, toward the garage, toward what he is about to find. He keeps the sentence and loses the woman.

The page misses quieter things too. It doesn't record which lights were on in the house when the officer arrived, or who else was home. It doesn't say who opened the window, or whether it was already open at 2100. It doesn't record the weather, though the weather will matter to every scientist who touches this case later. Nor does it record whether the gate to the alley was latched, or how many people walked through it during the search before anyone thought of it as a crime scene. A responding officer searching for a living child does not walk like a forensic examiner, and should not. That search was right. The cost of it is simply not on the page.

A first report records what the responders noticed. It does not record what happened. Those are two different documents, and only one of them exists.

So ask the third question: what does this page force you to find out?

Every gap points somewhere, but they do not point with equal weight. The weather can be recovered from public records. The household can be reconstructed from interviews. The unnamed neighbor is different. If she is not found in the first days, she becomes a sentence with no source. Defense attorneys can dismiss it, investigators can't confirm it, and it will drift through the file for years, repeated in summaries until nobody remembers it was hearsay from a woman no one identified.

So the page forces a task more than a theory: go back to the alley and find her.

In a working case, this is where pursuit begins. It happens in the dullest possible form. A detective takes the report and a street map and knocks on every door with a sightline to the alley. He writes down who answers, who doesn't, and who has a back gate. He asks each person the same question in the same words, because varied wording produces varied memory. He notes which houses had no one home and returns to them in the evening. If the canvass is done well, there is a second sheet in the file, a canvass log, with every address on the block and an entry beside each. If it is done badly, or late, or not at all, the blank line on the first report stays blank. Every later reader of the file inherits it as a hole they can see but cannot fill.

I want to be careful with you here, because this is the moment where a certain kind of reader leans forward. A vehicle in an alley. An open window. A household. A neighbor who spoke and vanished. The mind wants to assemble these into a story with a person at the center of it. That instinct has destroyed innocent lives. It has steered real investigations toward the wrong door for years. It has turned grieving parents into suspects in the comment sections of strangers.

This book will not do that, and it will ask you not to. Every case examined in detail in these pages has been resolved in a courtroom or documented as an exoneration. Where a case remains open, or where a living person has never been charged, I will not speculate about who did it, and I will not hint. That is no squeamishness about hard facts. A file is evidence, and an unsupported guess placed beside evidence contaminates it as surely as an unbagged glove at a scene.

What I will do is teach you to read. The question at the center of this book is plain and brutal: when a child is murdered and the killer is not found, what actually went wrong, and what would it take to find them? The answer is rarely that nothing was there. More often, something was there in the first days, and it was written down partially, or written down and not followed, or never written down at all. Sometimes it was the third kind of thing, and the only trace it left is an absence on a form.

Silence on a page is data. It tells you where to dig, even before it tells you what is buried.

The file will grow from here. A medical examiner will write a document far longer and colder than this one. Hundreds of tips will come in, most of them useless, and someone will have to log every one. There will be interviews, and some of them will go wrong in ways that take years to undo. Each of those documents will be built on top of this first page. Each will carry its assumptions forward unless somebody stops and checks.

So before you turn to the next one, do something practical. Take a pencil and look again at the witness field. Write, in the margin beside that empty second line, the four things a later investigator would need in order to fill it: a name, an address, a time, and the exact words she said. Then look at the narrative and count how many of those four the report gives you.

It gives you none of them. That is where the work starts.