Chapter 3: Physical Evidence: Types, Collection, Preservation, and What It Can (and Can't) Tell You

"Every contact leaves a trace." — the principle attributed to Edmond Locard (1877–1966), the French criminalist who built one of the first police laboratories. He never wrote those five words in exactly that form, but they distill his life's argument — and, read carelessly, they have promised juries more than physical evidence can ever deliver.

Overview

A scene is a haystack, and physical evidence is the field's faith that there is a needle in it. The faith has a name and a hundred-year pedigree: Locard's exchange principle, the idea that a person cannot enter a place, touch an object, or struggle with another human being without leaving something of themselves behind and carrying something of the place away. It is the closest thing forensic science has to a founding law, and it is true. But here is the part television never tells you: Locard's principle guarantees that the trace exists. It does not guarantee that anyone will find it, that the lab can read it, that it survived the fire or the rain or the well-meaning paramedic who moved the body, or — most importantly — that the trace means what the detective hopes it means. A trace can exist and be undetectable. It can be detectable and be uninterpretable. It can be interpretable and still not answer the question anyone cares about.

This chapter is about the raw material of the entire field: physical evidence itself. Before we can analyze a fingerprint (Chapter 14), type a DNA profile (Chapter 7), or confirm an accelerant (Chapter 23), we have to recognize that something is evidence at all, decide what kind it is, collect it without destroying it, package it so it survives the trip to the lab, and — crucially — understand in advance what category of question it could possibly answer. Most failures in forensic science do not happen at the bench. They happen here, in the first hours, when fragile evidence is missed, robust evidence is contaminated, the wrong controls are collected (or none at all), and a trace that could have excluded a suspect is instead rendered meaningless before any instrument ever touches it.

We will sort physical evidence into the categories practitioners actually use, because category determines handling and category determines what an item can prove. We will draw the single most important — and most neglected — distinction in evidence work: the difference between the questioned sample, the known/reference sample you compare it against, and the control sample that proves your test and your collection are trustworthy. And we will keep faith with the book's spine: every type of evidence in this chapter gets its limits stated plainly, because the most dangerous physical evidence is the kind that looks more conclusive than it is.

In this chapter, you will learn to:

• State Locard's exchange principle correctly — as a guarantee that traces exist, not a promise we will find or correctly read them — and name its real-world limits.
• Classify any item as transient, conditional, pattern, transfer, or associative evidence, and say what each category can and cannot establish.
• Tell a known/reference sample from a control sample from a questioned sample, and explain why the right controls are the difference between an interpretable result and a worthless one.
• Hold the line between what physical evidence can do (associate, exclude, sequence, establish presence) and what it cannot (identify an actor, prove intent, fix the moment of contact, or establish guilt).
• Explain how degradation, contamination, and bad packaging destroy evidence, and why the most fragile evidence demands the most urgent attention.
• Describe how a scene is reconstructed from physical evidence — and why a reconstruction names a sequence, not a suspect.

Learning Paths

This chapter is foundational for everyone; physical evidence is the substrate every later technique operates on. Here is how the four paths run through it:

🔎 Investigator/CSI: This is your chapter. Sections 3.2, 3.3, and 3.6 are your daily craft — recognizing evidence types, collecting the right controls, and packaging so the lab gets something usable. The collection decisions you make in the first hour set a hard ceiling on everything the lab can do later. 🧪 Lab analyst: Weight §3.3 and §3.6. You will live or die by whether the scene team collected the control samples you need and packaged the questioned items so they were not contaminated or degraded en route. An item with no control is often an item you cannot interpret. ⚖️ Law/courtroom: §3.4 and §3.5 are where the cross-examination lives — the gap between "this evidence associates two things" and "this evidence proves my client did it." Learn to hear when a witness has let physical evidence speak to identity, intent, or timing it cannot address. 👥 General reader/juror: All of it, but especially §3.1 and §3.4. The phrase "his DNA was at the scene" sounds like a conviction and is often nearly meaningless without the questions this chapter teaches you to ask.

3.1 Locard's exchange principle

Start with the question, the way we always will. A burglar climbs through a window, crosses a room, pries open a desk, and leaves. Is there, somewhere in that room, proof he was there? The hopeful answer — the one that built forensic science — is yes: he could not have done any of that without exchanging material with his surroundings. He left skin cells on the windowsill, fibers from his jacket on the carpet, a partial shoe impression in the dust, tool marks on the desk. And he carried away, on himself, fibers from the carpet, splinters from the desk, soil from the lawn outside. This is Locard's exchange principle: whenever two objects come into contact, there is a mutual transfer of material between them. Every contact leaves a trace — and, just as importantly, every contact takes one.

The principle is named for Edmond Locard, the French criminalist who, as we saw in Chapter 1, founded one of the world's first forensic laboratories in Lyon around 1910. The famous five-word slogan — "every contact leaves a trace" — is a later distillation; Locard's own writing made the point at greater length and with more care. But the idea is genuinely foundational. It is the reason there is anything to collect at all. If contact left no trace, there would be no fingerprints, no transferred fibers, no gunshot residue, no soil on a boot, no DNA under a fingernail. The entire enterprise of criminalistics rests on the premise that human activity leaves physical residue, and that residue, recovered and read, can connect a person to a place, an object, or another person.

🔬 At the Bench Locard's principle is bidirectional, and the second direction is the one investigators forget. The scene receives material from the perpetrator — but the perpetrator also leaves carrying material from the scene. This is why a suspect's clothing, shoes, vehicle, and hands are themselves scenes to be processed: the carpet fiber that fell from his jacket onto the victim's floor has a twin still clinging to the jacket; the soil from the cabin lot is in the treads of his boots. Some of the strongest associations in a case come not from what the perpetrator left behind but from what the scene left on him — which is exactly why getting to a suspect's clothing before he launders or discards it can matter more than another hour at the scene.

Now the discipline. Read carelessly, "every contact leaves a trace" becomes a promise that a clever enough examiner will always find the proof. That is precisely the Sherlock Holmes fantasy Chapter 1 warned against, and it is wrong in four specific, teachable ways. Locard's principle guarantees that a trace is deposited. It guarantees nothing about the four things that have to go right afterward.

First, the trace may be too small or too sparse to detect. A transfer can be a few cells, a single fiber, a smear below the sensitivity of any available test. "It must be here somewhere" is not a finding; an undetected trace and an absent trace are operationally identical. The principle promises deposition, not detectability.

Second, the trace may not persist. Transfer and persistence are different things (a distinction Chapter 19 develops for trace evidence). A fiber shed onto a moving victim's coat may fall off within hours; skin cells on an outdoor railing wash away in rain; a fingerprint in dust is destroyed by a breeze. By the time anyone processes the scene, much of what Locard deposited may simply be gone. The fire on Mill Creek Road is a brutal example: heat, water from suppression, and the boots of firefighters destroy and redistribute trace evidence wholesale.

Third, the trace may be undetectable against its background. A cotton fiber from a suspect's white T-shirt, shed onto a white cotton bedsheet, is real, present, and effectively invisible — there is no way to distinguish it from the millions of identical fibers already there. Locard guarantees the fiber is there; he does not guarantee you can ever pick it out. The most useful transfers are of rare materials onto dissimilar surfaces, which is the exception, not the rule.

Fourth — and most important — even a recovered, readable trace rarely answers the question the case is asking. Suppose you do find the burglar's skin cells on the windowsill and recover a usable DNA profile. You have established that his cells are on that sill. You have not established when they got there, how (he may have leaned on that public windowsill last week for an innocent reason), or that he committed the burglary. Locard's principle delivers an association. It does not deliver an actor, a time, or an intent. The whole rest of this chapter — and much of the book — is about that gap.

🧠 Cognitive-Bias Watch Locard's principle has a dangerous psychological side effect: it licenses the expectation that evidence connecting the suspect must exist, which makes examiners more likely to find it. If you are certain the burglar was here, you will search harder, interpret an ambiguous smear as "consistent," and treat the absence of contradicting evidence as confirmation. The principle is true, but "every contact leaves a trace" can quietly become "I will keep looking until I find the trace that fits my suspect." The safeguard, developed in Chapter 31, is to search and interpret without knowing — or without letting yourself be steered by — whom the detective favors.

So hold Locard correctly. It is the reason physical evidence exists, and it is a real and powerful idea. But it is a statement about the world (contact deposits material), not a statement about our knowledge (we will recover and correctly read that material). Forensic science begins with Locard's optimism and survives on the discipline of remembering its four limits.

3.2 Types of physical evidence

Before you can collect, package, or interpret an item, you have to know what kind of evidence it is — because the kind determines how fast it will vanish, how you must handle it, and what category of question it can answer. Physical evidence is any tangible object or material that can establish, through scientific examination, that a crime occurred or that connects (or excludes) a person, place, or thing in relation to the crime. That is a deliberately broad definition: a fingerprint, a bullet, a hair, a bloodstain, a charred document, a soil sample, a tire impression, a stomach's contents, and a deleted file are all physical evidence, even though some are objects you can bag and others are patterns or conditions you can only photograph.

Practitioners sort physical evidence in several overlapping ways. The most useful working scheme for this chapter divides it into five categories by how it behaves and what it can show: transient, conditional, pattern, transfer, and associative. These are not airtight boxes — a single bloodstain can be transfer evidence (it links a person to the scene), pattern evidence (its shape encodes how it was deposited), and associative evidence (it is associated with one source) all at once — but the categories train you to ask the right questions about handling and meaning.

Transient evidence

Transient evidence is evidence that is temporary by nature and will change or disappear if not recorded immediately. An odor — gasoline, decomposition, a smell of bleach — is transient: it tells you something now and is gone in an hour. So is temperature (the warmth of a car hood, the body temperature used to estimate the postmortem interval in Chapter 11), the color of a flame still burning, a wet versus dry stain, the position of an unmelted ice cube, or a dying victim's statement. Transient evidence cannot, in most cases, be collected at all. It can only be documented — measured, photographed, smelled and noted, timed — and the documentation becomes the evidence. The lesson is brutal and simple: if no one records it in the first minutes, it never existed as far as the case is concerned. This is why Chapter 2 insisted that observation and documentation precede collection. Much of the most decisive evidence at a scene is the kind you can only write down.

Conditional evidence

Conditional evidence is evidence produced by an event or action and dependent on the conditions at a specific moment — the state of things at the scene. Whether a light was on or off, a door locked or unlocked, a window open or closed; the position of a body; the direction a victim's blood flowed (which depends on body position and therefore encodes it); the state of a fire when first responders arrived; whether the television was warm. Conditional evidence is fragile in a different sense than transient evidence: it does not evaporate, but it is destroyed the instant anyone changes the conditions — turns off the light to "see better," closes the door, moves the body, drives the fire out. Like transient evidence, it is usually preserved through documentation rather than collection, and like transient evidence, its great enemy is the well-meaning first responder who tidies, rescues, or investigates before anyone records the original state.

🔬 Read the Evidence

text FIGURE 3.1 — "Two kinds of evidence that vanish" [constructed teaching example] THE ITEM At a house fire: (A) a strong smell of gasoline near the front door, noticed by the first firefighter inside; and (B) the front door, found standing open with the deadbolt in the extended (locked) position, its frame splintered near the strike. THE CONTEXT A small structure fire, knocked down in minutes. The first-in firefighter is not a CSI; the scene is wet and dark. No photographs are taken until an investigator arrives forty minutes later, by which time the smell has dissipated and the door has been pushed fully open and stepped through repeatedly. WHAT IT SHOWS (A) The gasoline odor is TRANSIENT — it was real evidence of a possible accelerant, but it now exists only in the firefighter's memory and (if he wrote it down) his notes. (B) The door's locked-deadbolt-with-splintered-frame is CONDITIONAL — it recorded the state of entry at the moment of the fire, but that state has been altered by everyone who has since used the door. WHAT IT DOESN'T Neither, by itself, proves arson or forced entry by an attacker. The odor could have an innocent source; the splintering could predate the fire or be from firefighter forcible entry. THE INFERENCE Both are worth pursuing — an accelerant smell plus a deadbolt thrown against a splintered frame are threads — but their evidentiary value now depends almost entirely on whether anyone documented them before they changed. THE LESSON The most fragile evidence is often the most decisive, and it is preserved only by recording it. A photograph taken in the first five minutes is worth more than any instrument run on day three.

This figure is not idle: a deadbolt thrown into a splintered frame is exactly the kind of conditional evidence that will matter at the Mill Creek cabin, where the door carries pry marks. Hold the example; we return to it in the Case File.

Pattern evidence

Pattern evidence is evidence in which the arrangement — the shape, distribution, or geometry — carries the information, not (or not only) the substance. A bloodstain pattern (Chapter 10) encodes the physics of how blood traveled; a fingerprint (Chapter 14) is a pattern of ridges; a shoe or tire impression (Chapter 16) is the pattern of a sole or tread; glass-fracture patterns reveal the direction of an impact; tool marks (Chapter 16) are patterns scored by a tool's edge; even a pattern of burning across a floor is read (cautiously — see Chapter 22) for a fire's origin. Pattern evidence is overwhelmingly preserved by photography with a scale, because the geometry is the evidence and the object often cannot be removed (you cannot bag a bloodstain pattern on a wall). Pattern evidence is also where some of the field's worst overreach lives: reading a pattern is interpretive, examiner-dependent, and — for several pattern disciplines — sits low on the validity spectrum precisely because the inference from pattern to event is so much softer than it looks.

Transfer evidence

Transfer evidence is material moved from one place or person to another by contact — Locard's principle made tangible. Hairs, fibers, glass fragments, soil, paint, gunshot residue, skin cells, blood transferred by touch: all are transfer evidence. This is the classic "trace evidence" of the crime lab (Chapter 19 takes it up in depth, explicitly as Locard in practice), and its value depends on two things the categories above don't: how rare the transferred material is, and how well the transfer persisted. A single red cashmere fiber transferred onto a victim is far more telling than a white cotton fiber, because cashmere is rarer; a glass fragment embedded in a suspect's shoe sole persists for days, while one resting on his sleeve is gone by lunch. Transfer evidence is collected physically — picked, taped, scraped, swabbed, or seized whole — and it is exquisitely vulnerable to contamination, because the entire point is that material moved by contact, and the investigator's own hands, tools, and clothing can move material too (more in §3.6).

Associative evidence

Associative evidence is the category defined by what it does rather than what it is: any evidence that tends to associate a person, object, or place with the crime or with each other. A wallet dropped at the scene, a weapon traced to an owner, a vehicle placed at a location by its tire impressions, a DNA profile matching a suspect — these associate. Most transfer evidence is also associative; so is much pattern evidence. The reason to name the category separately is to keep front-of-mind the central honest question of the whole chapter: associate is not the same as identify, and certainly not the same as prove. An item can strongly associate a suspect with a scene and tell you nothing about whether he committed the crime, when, or why. The word "associative" is a built-in reminder to keep the verb honest.

⚖️ In the Courtroom Categories matter in court because they map onto what a witness may honestly claim. Transfer and associative evidence speak to association ("these fibers are consistent with that sweater"); they do not, by themselves, speak to identity of an actor or to intent. Conditional evidence speaks to the state of the scene at a moment, not to who created that state. Pattern evidence speaks to how something happened, not who did it. A careful cross-examination often consists of moving an item up the ladder it does not belong on — getting an examiner who can defend "associated with" to imply "committed by." Knowing the category an item belongs to is knowing the highest rung it can honestly stand on.

A final scheme worth knowing, because you will hear it constantly: evidence is also sorted into class and individual characteristics, the distinction Chapter 1 made central. Most physical evidence carries class characteristics — it associates an item with a group (this fiber is nylon of this type; this soil has this mineral profile; this shoe is a size-11 of this brand). A smaller amount carries individual characteristics that can point toward a single source (the accidental wear on one shoe, the ridge detail of one finger, the full STR profile of one person's DNA). Every category above can be either, depending on the item, and the whole game of §3.4 is keeping straight which one you actually have.

3.3 Known, reference, and control samples

Here is a section that television omits entirely and that working analysts will tell you is the difference between evidence and noise. A questioned item, by itself, usually means nothing. It acquires meaning only by comparison — against something whose source you already know, and against something that tells you your collection and your test are trustworthy. Get the comparison samples wrong, or fail to collect them, and you can render the best questioned evidence in the world uninterpretable. This is not a refinement. It is the foundation.

Three terms must be kept rigorously distinct.

A questioned sample (also called an unknown or evidence sample) is the item recovered from the scene or a person whose source is unknown and in question — the bloodstain of unknown origin, the latent print, the fiber on the victim, the white powder. It is what you are trying to identify or source.

A known sample (or reference sample) is material from a verified, documented source, collected specifically so the questioned sample can be compared against it. If you recover an unknown bloodstain (questioned), you collect a reference DNA sample — a buccal (cheek) swab — from the victim and from each suspect, so the unknown can be compared to known people. If you recover an unknown fiber, you collect reference fibers from the suspect's sweater. The defining feature of a known/reference sample is that you are certain where it came from, which is exactly what the questioned sample lacks. Comparison evidence is meaningless without the right knowns: an unidentified bloodstain matched to nobody is just a stain.

A control sample is something different and more often forgotten — material collected to establish a baseline and to prove that a result is real rather than an artifact of the substrate, the environment, or the collection process itself. The classic example: you swab a bloodstain off a painted windowsill for DNA. You must also swab an adjacent, unstained area of the same windowsill — a substrate control. Why? Because if the paint itself, or some contaminant on the sill, interferes with the test or carries DNA, the substrate control will reveal it. Without the control, a positive result might come from the stain or from the sill, and you cannot tell which. The control turns "we got a result" into "we got a result attributable to the stain."

🔬 At the Bench Controls come in several flavors, and a good analyst (and a good cross-examiner) knows which were run: - Substrate / background control — unstained material from the same surface as the questioned stain, to show the surface itself is not the source of the signal. - Reagent / negative control — the test run with no sample (just the chemicals), to show the reagents are not contaminated and not producing a false positive on their own. - Positive control — the test run on a sample known to contain the target, to show the test is actually working and would have detected the target if present. (A negative result is only meaningful if you've shown the test can go positive.) - Elimination samples — reference samples from people who had legitimate access (the homeowner, the responding officers, the lab staff), so their DNA, prints, or fibers can be recognized and set aside rather than mistaken for a suspect's. A result reported without its controls is not wrong, exactly — it is unverifiable, which in forensic terms is nearly as bad.

The logic of controls is the logic of the experiment, imported into the messy world of the crime scene. In a laboratory experiment, you run a control condition so that any difference you see can be attributed to the variable you changed rather than to something else. Forensic controls do the same job under far worse conditions: they let you attribute a result to the evidence rather than to the surface, the chemicals, the air in the lab, or the gloves of the person who collected it. The deepest reason the cold case's contamination near-miss (Chapter 4) matters is that contamination is exactly what controls are designed to catch — and a scene team that skips controls has blinded the lab to its own errors.

🧠 Cognitive-Bias Watch Elimination samples are also a bias safeguard, not just a bookkeeping tool. Without the homeowner's reference prints on file, an examiner under pressure to "find the intruder" may seize on a print that is, in fact, the homeowner's. Knowns and elimination samples don't only enable comparison; they constrain the examiner, making it harder to mistake innocent presence for guilty presence. We will see in Chapter 14 that the absence of elimination prints is a recurring ingredient in fingerprint error.

One more discipline about reference samples: they must be representative and adequately documented. A single reference fiber from one corner of a large, multicolored garment may not represent the fibers shed from the part that actually contacted the victim; one soil sample from the center of a lot may miss the distinctive soil at its edge where a vehicle parked (a point that will matter for the cabin in Chapter 24). Collecting reference samples is itself a skill: take enough, from the right places, and document precisely where each came from, because a reference sample whose own provenance is uncertain has quietly become a second questioned sample.

🔍 Check Your Understanding 1. You recover an unknown reddish stain from a tile floor and swab it for DNA. Name the two other samples you should collect for the result to be interpretable, and say what each one rules out. 2. A lab reports that a presumptive blood test on a swab was negative. What single control would you want to know was run before trusting that "negative," and why? 3. Why is an unidentified DNA profile from a crime scene — matched to no known person — of limited immediate use, even though it is a complete, high-quality profile?

3.4 What physical evidence can and cannot establish

We now reach the heart of the chapter, and the place where the book's first theme — forensic science excludes; it rarely proves — does its hardest work. Physical evidence is powerful, but its power is specific and bounded, and most miscarriages of justice involving forensic evidence come from asking an item to answer a question outside its competence. So let us be exact about the four things physical evidence can genuinely do, and the four things it cannot.

What physical evidence CAN establish:

It can establish that a crime occurred — or that one did not. A body with a skull fracture and no soot in the airways (Chapter 11) establishes a homicide where an "accidental fire" was assumed. Conversely, evidence can show that a reported crime did not happen — that a "burglary" was staged, that a "fire" was electrical. Physical evidence often answers the threshold question before it answers any question about a person.

It can establish association — or exclusion. This is the central thing. Evidence can associate a person, object, or place with the crime (the soil on the boots matches the lot; the DNA is consistent with the suspect) or — more powerfully and cleanly — exclude them (the profiles differ; the shoe is the wrong size; the tool that made these marks is not this tool). As Chapter 1 argued, exclusion is forensic science's surest voice. A single robust mismatch can eliminate a suspect; an association only narrows.

It can establish a sequence of events. Physical evidence is unusually good at order. Blood beneath a char line was deposited before the fire (Chapter 10); a bullet hole through a window from the outside in; a footwear impression over a tire track means the person walked after the vehicle passed. Reconstruction (§3.5) is largely the science of sequence — establishing what happened in what order — and sequence is often more defensible than any claim about who or when by the clock.

It can establish the presence of a substance or condition. Instruments can confirm that a debris sample contains gasoline (Chapter 23), that blood is human (Chapter 10), that a powder is a controlled substance (Chapter 21). "This substance is X" is, when done with validated instrumentation, among the most reliable claims in all of forensics.

What physical evidence usually CANNOT establish:

It generally cannot identify the actor. Evidence associates objects and materials, not necessarily the person who used them. The recovered DNA on the gas can identifies whose cells are on the handle, not who carried it; the fingerprint identifies whose finger touched the surface, not who committed the act; the boot print identifies a shoe, not the foot inside it at the relevant moment. Closing the gap between "this object is associated with this person" and "this person did this thing" is the job of the whole case, not of any single item.

It cannot establish intent or state of mind. No physical test reveals whether an act was deliberate, reckless, accidental, or in self-defense. Evidence can show a fire was incendiary (set deliberately, on the physical facts); it cannot show who intended what. Intent is a legal inference drawn from the totality of evidence, never a finding an instrument produces.

It usually cannot fix the time of contact. This is the most underappreciated limit. DNA on a doorknob does not say when the cells were deposited; a fingerprint cannot be reliably dated; a fiber transfer carries no timestamp. The "transfer problem" (Chapter 7, Chapter 8) is precisely this: a trace consistent with the suspect may have been deposited innocently, days before, or even moved there secondarily by another person or object. "His DNA was on the weapon" and "his DNA was on the weapon at the time of the killing" are different claims, and physical evidence usually supports only the first.

It cannot establish guilt. Guilt is a legal conclusion, reached by a jury weighing all the evidence against a legal standard. No item, however strong, "proves the defendant guilty"; at most it strongly supports an element of the case. An expert who says evidence "proves he did it" has stepped off the science and into the jury's job — the ultimate-issue problem we take up in Chapter 30.

⚠️ Junk-Science Alert Watch for evidence asked to leap its category. The leap from association to identity ("his fibers were there, so he was there, so he did it") drops two unstated assumptions: that the fibers got there at the relevant time, and that being there means committing the crime. The leap from presence to action ("gunshot residue on his hands, so he fired the gun") ignores that GSR transfers by handling, proximity, and contaminated surfaces (Chapter 24). The leap from sequence to intent ("the fire had multiple origins, so he is a murderer") confuses what happened (incendiary fire) with who did it and why. None of these leaps is supported by the physical evidence alone; each requires argument the evidence cannot supply. The overstatement is almost never in the data — it is in the sentence built on top of the data.

The honest framing, then, is a hierarchy of verbs you should now recognize from Chapter 1, applied to physical evidence:

• Excludes — the evidence and the source do not share features they would have to share; the suspect (or object) is eliminated. Strongest and cleanest.
• Cannot exclude / consistent with — they share features; the suspect cannot be ruled out. True and useful, routinely overstated.
• Strongly supports — the shared features are numerous, specific, or quantifiably rare enough that the evidence supports the association far over a random alternative — ideally with a number, as DNA gives.
• Proves — reserved for almost nothing. A physical fit (the torn tape that jigsaws back together) or a quantified DNA match approaches it for association; nothing physical "proves" the human story of guilt.

Memorize the ladder, and climb it only as far as the evidence carries you. The cross-examinations and the wrongful convictions ahead are, almost without exception, stories of someone who climbed one rung too high.

3.5 Reconstruction: reading the scene from the evidence (preview)

Individual items of physical evidence are sentences; reconstruction is the attempt to assemble them into a paragraph — a coherent account of what happened, in what order, at a scene. It is one of the most intellectually demanding tasks in forensic science, and we only preview it here (Chapter 10 develops it through bloodstain pattern analysis, Chapter 22 through fire, and Chapter 39 assembles the whole cold case), but the logic belongs in this foundational chapter because it shows physical evidence working at its highest level — and reveals, sharply, the line it must not cross.

Reconstruction reasons from the physical facts to a sequence of events. The position and shape of bloodstains, the location of cartridge cases, the direction of glass fracture, the layering of footwear and tire impressions, the char patterns and the depth of burning, the state of doors and locks — each constrains what could have happened. An impression over a bloodstain was made after the blood was deposited; a bloodstain interrupted by a body's outline means the body was there when the blood landed and was later moved. By chaining such constraints, an analyst can sometimes establish a remarkably detailed order of events: the victim was struck here, fell there, the fire was set after the body was already on the floor, the entry was forced before any of it.

🔬 Read the Evidence

```text FIGURE 3.2 — "What the layering tells you (and what it doesn't)" [constructed teaching example]

PLAN VIEW (not to scale):

    ▣▣▣▣▣▣▣▣▣▣▣  (wall)
    ▣                       N
    ▣   × stain S1          ↑
    ▣   ╱shoe impr. I1 over S1
    ▣  ═══ tire track T1 (drive)
    ▣        × stain S2 (on top of T1)
    door▢ (pry marks)

THE ITEM A scene sketch: a wall stain (S1) with a shoe impression (I1) pressed into it; a tire track (T1) across the dirt; a second stain (S2) lying on top of the tire track; and a door with pry marks. THE CONTEXT All four documented in place with scales before anything was collected. WHAT IT SHOWS I1 sits ON TOP OF S1 → the shoe contacted that spot AFTER S1 was deposited. S2 sits ON TOP OF T1 → that stain landed AFTER the vehicle passed. The pry marks on the door are a class match to a pry-bar tip. WHAT IT DOESN'T It does not say WHO wore the shoe, WHO drove the vehicle, WHEN (by the clock) any of it happened, or that the same person did all of it. The pry marks show a TYPE of tool, not a specific bar, and not who wielded it. THE INFERENCE A defensible SEQUENCE: vehicle present → S2 deposited after → (separately) S1 deposited before a shoe pressed into it → entry forced with a pry-bar-type tool. The order is physical fact; the actor and the clock-time are not. THE LESSON Reconstruction establishes ORDER, which is real and powerful — and stops there. The instant a sequence acquires a name ("the defendant did this, then that"), the science has handed off to inference the evidence has not earned. ```

The figure encodes the discipline of reconstruction in one rule: reconstruction names a sequence, not a suspect. "The blood was deposited before the fire" is a physical fact the evidence can carry. "The victim was beaten by the defendant and then the defendant set the fire" is a story — possibly the right story, but one that adds an actor and a motive the bloodstains and char patterns cannot themselves supply. Good reconstruction is scrupulous about the seam between the two: it states the sequence with confidence and flags the actor and the timing as inferences that require other evidence.

⚖️ In the Courtroom Reconstruction is most dangerous when it is most persuasive. A vivid narrated sequence — "he stood here, struck her there, the blood flew this way" — is enormously compelling to a jury and can outrun the evidence badly, especially in pattern disciplines that sit low on the validity spectrum. Several contested cases (the bloodstain disputes you will meet in Chapter 10) turn on a reconstruction that was presented as physical certainty when it was, in truth, one interpretation among several the same stains could support. The honest reconstructionist offers the sequence the evidence constrains, names the alternatives it does not rule out, and leaves the actor to be established by other means.

Reconstruction, then, is physical evidence at its most ambitious and its most disciplined. It can transform a pile of items into a constrained account of events — and its integrity depends entirely on refusing to let that account quietly acquire a name before the rest of the case has earned it. We will practice exactly this restraint, piece by piece, across the cold case.

3.6 Evidence integrity: degradation, contamination, packaging

Everything in this chapter assumes the evidence reaches the lab in a condition that lets it answer questions. That assumption is heroic. Physical evidence is under attack from the moment it is deposited — by the environment, by time, by the very people sent to recover it — and evidence integrity (the principle that an item must be collected, packaged, stored, and transported so that it is neither altered nor contaminated, and so that its connection to the scene is documented and unbroken) is what stands between a usable result and a worthless one. Three threats matter most: degradation, contamination, and bad packaging.

Degradation is the natural deterioration of evidence over time and under environmental stress. Biological evidence is the most vulnerable: DNA breaks down with heat, moisture, sunlight, and microbial growth, which is why a bloodstain sealed wet in a plastic bag — where bacteria flourish — can be destroyed within days, while the same stain air-dried and packaged in paper may yield a profile years later. Degradation is not contamination; nothing foreign is added. The sample simply decays until it can no longer be read. The cold case sharpens this: the gas-can DNA is heat-degraded (Chapter 8) by the fire, which is why it yields a difficult, partial result rather than a clean profile. Heat, water, and time are the enemies, and the countermeasures are speed (collect fragile evidence first), drying (moisture is the accelerant of biological decay), and cold or controlled storage.

Contamination is the introduction of foreign material — or the transfer of material between items — that did not belong to the evidence as deposited. This is Locard's principle turned against the investigator: because contact transfers material, the collector's own hands, breath, tools, and clothing can deposit DNA, fibers, or other trace onto the evidence, and a single tool used on two items can carry material from one to the other. Contamination is uniquely corrosive because it can manufacture an association that never existed — a real but irrelevant DNA profile (the analyst's own) on a swab, a fiber carried from the squad car onto the victim. It is also, unlike degradation, frequently invisible and unrecoverable: once a foreign profile is on the swab, no later care removes it.

🔬 At the Bench The anti-contamination toolkit is mostly mundane and mostly about separation: - Personal protective equipment — gloves (changed between items, not just between scenes), masks, and clean suits — keeps the collector's own DNA and fibers off the evidence. (PPE protects the evidence from you as much as you from the evidence.) - One tool, one item, or clean between — forceps, swabs, and scissors are single-use or cleaned to the point of sterility between items, so nothing rides from one to the next. - Package each item separately — never let two evidence items touch inside one container; cross-transfer between bagged items is a classic, avoidable contamination. - Collect knowns and questioned items with separate equipment, and ideally separate personnel — so a suspect's reference sample cannot contaminate the crime-scene swab. A reference buccal swab brimming with one person's DNA is the last thing you want near a low-template questioned sample. - Controls (again) — substrate and reagent controls (§3.3) are the detector for contamination that slipped through. This is why controls are not bureaucratic box-checking; they are how the lab catches the contamination it could not see.

Packaging is where degradation and contamination are won or lost, and it follows rules that are specific, evidence-type-dependent, and frequently violated. The single most important rule — and the one Chapter 2 flagged — is that biological evidence is packaged in breathable paper (envelopes, bags, boxes), never in airtight plastic. Plastic traps moisture; moisture breeds the bacteria and mold that destroy DNA. Bloodstained clothing goes in a paper bag, air-dried first; a swab is air-dried and boxed. Plastic is correct only for items where moisture loss is the threat — a liquid sample, or volatile fire debris that would evaporate through paper, which is why fire-debris samples for accelerant analysis go into airtight metal cans or special nylon bags (Chapter 22) to trap the volatile gasoline vapor. The rule is not "always paper" but "match the package to the threat": paper to let biological evidence breathe, airtight to keep volatiles in. Get it backward — biological evidence in plastic, fire debris in paper — and the packaging itself destroys the result.

🔬 Read the Evidence

text FIGURE 3.3 — "Same stain, two packages, two fates" [constructed teaching example] THE ITEM A bloodstained cotton work glove recovered from a scene, divided (hypothetically) so half is packaged each way. THE CONTEXT Recovered damp. Half (A) is sealed immediately, still damp, in a zip-top PLASTIC bag "to keep it contained." Half (B) is air-dried on clean paper, then folded into a labeled PAPER bag. WHAT IT SHOWS Within days, (A) shows mold growth and a musty odor; a later DNA attempt yields a degraded, partial, hard-to-interpret result. (B) is dry and intact; a later DNA attempt yields a full, clean profile. WHAT IT DOESN'T The plastic did not add anyone's DNA — this is DEGRADATION, not contamination. And neither result tells you WHOSE glove it is until compared to a known sample (§3.3) or WHEN the blood was deposited. THE INFERENCE Identical evidence, opposite outcomes, decided entirely by the package. The plastic bag did not lose the evidence — it actively destroyed it. THE LESSON Packaging is not clerical. For biological evidence, breathable paper after drying; airtight plastic or cans only for items (like fire debris) where escaping volatiles, not trapped moisture, are the threat.

Finally, integrity has a documentary dimension that ties this chapter back to Chapter 2's chain of custody: an item is only as good as the unbroken, recorded trail from scene to courtroom. Each transfer is logged; each container is sealed across its opening with tamper-evident tape, signed and dated across the seal so any later opening is visible; each item is uniquely labeled. A perfectly collected, perfectly packaged sample with a broken chain of custody — an unexplained gap, an unsealed bag, a missing signature — is evidence whose integrity can no longer be vouched for, and a defense attorney will (rightly) attack it. Physical integrity and documentary integrity are two halves of the same principle: the evidence must arrive unchanged and be provably unchanged.

🧠 Cognitive-Bias Watch Packaging and labeling errors interact with bias in a quiet, dangerous way. A mislabeled or commingled sample can put one person's profile under another person's name — and once that error is in the file, every later analyst reads the evidence through it, "confirming" an association that is an artifact of a switched label. The defense against this is the same dull rigor that prevents contamination: separate handling, immediate labeling, sealed packaging, and controls that would expose the mix-up. Integrity is not glamorous, and it is where most cases are actually lost.

🗂️ The Case File

The Mill Creek cabin — building the evidence inventory. You have the scene as Chapter 2 left it: processed late, partly compromised, the fire treated from the start as an accident. Now you walk the inventory of physical evidence recovered from the cabin and its grounds, and your single task this chapter is to log each item and state honestly what category it is and what it could — at most — establish. You are not interpreting yet. You are building the list every later chapter will draw from.

The recovered items: - A red metal gas can, found in the front room near the area of heaviest burning, its handle relatively intact. (Transfer/associative potential: it may carry touch DNA and a latent print — questioned samples whose source is unknown. Establishes nothing yet about who handled it or when.) - Hand tools (a hammer, a pry bar, miscellaneous renovation tools) consistent with an active job site. (Associative and pattern potential: a tool may have made marks elsewhere — but a contractor's own tools at his own renovation are innocently present by default.) - Charred documents — partially burned papers recovered from the front room, including what appear to be partnership and insurance materials. (Conditional/associative; some may bear indented writing. Fragile, fire-damaged; meaning unknown.) - The cabin door, bearing pry marks near the strike and frame. (Conditional + pattern evidence: it records the state of entry and a toolmark pattern. Documented in place. Could support forced entry; cannot yet say by whom, or whether the marks predate the fire.) - The victim's phone, recovered from the scene. (Associative/digital; a sealed container of potential evidence — see Chapter 25. Source-attributable only after forensic imaging.) - A single stray cartridge case, found on the floor. (Associative potential — but a cartridge case with no corresponding wound or weapon is, so far, an item of unknown relevance.)

What this chapter lets you say, and no more: every one of these is logged and categorized; not one is yet interpreted. Each is, at this stage, a questioned item awaiting its known/reference and control samples and its turn at the appropriate bench. Note the discipline the chapter taught: the gas can's potential DNA establishes whose cells are present, not who set a fire; the pry marks establish a class of tool, not a culprit; the cartridge case establishes only that a cartridge case is present. Recall Figure 3.1 — the pry-marked door is exactly the conditional evidence whose value depends on its having been documented before it changed.

Honest status after Chapter 3: Items logged; meaning unknown. The inventory exists; the science has not yet spoken. Resist, still, the urge to assemble these into a story. We are cataloguing the haystack, not yet claiming we have found the needle.

Conclusion

Physical evidence is the raw material of every method in this book, and this chapter built the frame the rest of the book hangs on. We began with Locard's exchange principle — the true and foundational idea that every contact leaves (and takes) a trace — and immediately disciplined it: the principle guarantees a trace exists, never that we will detect it, that it persisted, that it stands out from its background, or that it answers the question being asked. We sorted physical evidence into the categories practitioners use — transient, conditional, pattern, transfer, associative — because category dictates both how fragile an item is and what it can show. We drew the distinction television never mentions and analysts cannot work without: questioned versus known/reference versus control samples, and saw that the right controls are the difference between an interpretable result and a meaningless one. We held the honest line on what physical evidence can establish — association, exclusion, sequence, presence — and what it cannot — the actor, the intent, the time of contact, or guilt — and met the verb ladder (excludes → consistent with → strongly supports → proves) again, this time for physical evidence. We previewed reconstruction and its one inviolable rule, that it names a sequence and not a suspect. And we faced the unglamorous truth that most cases are won or lost not at the bench but in the handling: degradation, contamination, and packaging decide whether the lab gets evidence or noise.

Two of the book's themes ran through all of it. Forensic science excludes; it rarely proves — physical evidence's surest power is association and exclusion, not identification of an actor. And cognitive bias is the chief threat — Locard's optimism, missing elimination samples, and mislabeled packages each open a door through which expectation contaminates interpretation, and controls and blind handling are the locks. In the next chapter we follow the evidence inside the building it travels to: the crime laboratory, with its accreditation, its backlogs, its proficiency tests, and its scandals — because everything physical evidence can become is capped by the quality of the lab that processes it, and by whether that lab caught the contamination this chapter taught you to fear.

Key Terms

• Physical evidence — any tangible object or material that can establish, through scientific examination, that a crime occurred or that connects (or excludes) a person, place, or thing in relation to it.
• Locard's exchange principle — the principle that whenever two objects come into contact there is a mutual transfer of material; "every contact leaves a trace" (and takes one). It guarantees a trace exists, not that it will be found, persist, or be interpretable.
• Transient evidence — evidence that is temporary by nature (odors, temperature, a wet stain, a dying statement) and is preserved only by immediate documentation, not collection.
• Conditional evidence — evidence produced by an event and dependent on the conditions at a moment (a door's lock state, a body's position, a light on or off), destroyed when those conditions are changed.
• Known sample / reference sample — material from a verified, documented source, collected so a questioned sample can be compared against it; its defining feature is that its origin is certain.
• Control sample — material collected to establish a baseline and prove a result is attributable to the evidence rather than to the substrate, the reagents, or the collection process (substrate, reagent, positive, and elimination controls).

(Transfer, pattern, associative, and questioned-sample distinctions are used throughout; reconstruction and evidence integrity are previewed here and developed in Chapters 10, 22, and 39.)

Spaced Review

1. (This chapter, §3.1) State Locard's exchange principle in one sentence, then name two of its four real limits — reasons a deposited trace may still fail to help the case.
2. (This chapter, §3.3) You swab an unknown stain from a windowsill for DNA. Name the known sample and the control sample you also need, and say exactly what each one makes possible (the known) or rules out (the control).
3. (Chapter 2 → Chapter 3) Chapter 2 stressed documentation before collection. Using the category of conditional evidence from this chapter, explain why a body's position must be photographed before anyone moves it — what is destroyed if they don't?
4. (Chapter 1 → Chapter 3) Chapter 1 introduced the verbs excludes, consistent with, strongly supports, proves. A detective says a suspect's DNA on the gas-can handle "proves he set the fire." Identify two distinct overstatements in that sentence using this chapter's limits on what physical evidence can establish.
5. (Validity-spectrum preview) Of the evidence categories in §3.2, which one — pattern evidence — contains several disciplines that sit low on the NAS/PCAST validity spectrum, and what is it about reasoning from a pattern to an event that makes such methods more contestable than confirming that a substance is present? (We test this idea fully from Chapter 6 onward.)