DH Kaye: NCFDNA Report

A Report to the Working Group
on Legal Issues,
National Commission
on the Future of DNA Evidence

This report is current as of February 2, 2000. There have been significant legal developments since then. The report was to have been included in the publications of the Commission available at http://www.ojp.usdoj.gov/nij/dna/pubs.htm, but as of May 22, 2003, there was no reference to it there.

* Regents' Professor, Arizona State University College of Law; Fellow, Center for the Study of Law, Science, and Technology, 480 966-2922; k@asu.edu.

     1. Novel scientific methods
     2. Proficiency test records
          a. Proficiency testing as a prerequisite to admission
          b. The admissibility of errors on proficiency tests
          c. The use of proficiency tests to modify random-match probabilities
          d. The opportunity to retest as a response to defense arguments about proficiency testing

     1. Methods of extracting samples
     2. Offenses included
     3. Permissible uses
          a. Criminal investigations
          b. Civil litigation
          c. Research
     4. Security and sample retention

DNA has been used with great success to investigate crimes in which traces of blood, saliva, semen, hair, or other biological material are available, to convict the guilty and to exonerate the innocent. This report examines selected legal issues that arise when DNA is employed in the investigative and adjudicatory phases of the criminal justice system.

Part I discusses constitutional issues posed when the government uses DNA evidence in investigating a crime. It focuses primarily on methods of acquiring DNA from an individual suspected of committing the crime. Comparing such DNA with traces found at the scene of crime might exonerate the suspect, or it might associate the suspect with the crime.⁽¹⁾ Part I discusses the constitutionality of compelling suspects to submit to DNA sampling and of acquiring stored samples of a suspect's DNA or medical records relating to these samples from private medical providers or laboratories.⁽²⁾ It also considers the constitutionality of gathering DNA from large numbers of people to see whether any have genotypes that match those seen in the trace evidence. It shows that although the Fourth Amendment usually requires the police to have probable cause and a warrant to compel a person to provide a DNA sample, there are many situations in which police may be able to secure DNA samples without these protections.

Part I also considers a second investigative use of DNA--"profiling" physical or other characteristics of an individual whose DNA is found at the scene of crime. Genetic typing will permit inferences as to ancestry, physiognomy, or inherited disorders, and Part I concludes that investigators can use genetic data in this way without infringing any constitutional rights.

Part II concerns legal issues that arise at a later stage in the justice system, when DNA analysis is used as evidence in the prosecution of persons charged with crimes. It discusses the admissibility of new DNA tests and the results of proficiency tests at particular laboratories.

Part III deals with law enforcement DNA databases. Querying a database derived by analyzing DNA from large numbers of people and from many crime scenes offers law enforcement authorities a powerful tool with which to identify the perpetrators of certain types of crimes. Part III explores constitutional and ethical questions that must be confronted in establishing and operating DNA databases for law enforcement purposes. It also indicates the array of policy choices that must be made in developing these systems--from the determination of which individuals are subject to having their genotypes placed in the databank, to the specification of how the DNA samples are obtained, to the decision as to how long the samples should be retained, and to the enumeration and definition of the uses to which the information may be put.

Traditionally, DNA has been used to link a suspect to a crime. When a suspect's DNA matches the DNA left at a crime scene, for example, one inference is that the suspect is the source of the crime-scene DNA.⁽³⁾ Inversely, when the DNA does not match, the suspect can be excluded as the source of the crime-scene DNA. If trace evidence is to be used to exclude and include individuals as viable suspects, the police must secure samples of DNA from individuals who might have committed the crime under investigation.⁽⁴⁾ Officials can secure such samples in many ways. They can seek a court order to compel an individual to submit to sampling, they can turn to a preexisting collection of DNA samples, or they can take a sample with the consent of the individual. As a matter of constitutional law, the principal constraint⁽⁵⁾ on such government action is the search and seizure clause of the Fourth Amendment to the United States Constitution,⁽⁶⁾ which reads:

The right of the people to be secure in their persons, houses, papers, and effects, against unreasonable searches and seizures, shall not be violated, and no Warrant shall issue, but upon probable cause, supported by Oath or affirmation, and particularly describing the place to be searched, and the persons or things to be seized.

Because the search and seizure clause protects "[t]he right of the people to be secure in their persons," it applies both to restraining a person and to searching the inside or outside of a person's body.⁽⁷⁾ As a general rule, police will need a warrant or comparable court order issued by a judge or magistrate who is persuaded that there is probable cause to believe that the DNA sample will produce evidence linking the suspect to the crime.⁽⁸⁾ With such authorization, police can use necessary force to extract the biological material.⁽⁹⁾

In some circumstances, however, either a warrant or probable cause might not be essential. For instance, if a person is legitimately under arrest, the seizure of the person is justified,⁽¹⁰⁾ and routine, non-invasive DNA sampling of all arrestees for the purpose of creating a record of the true identity of the individual is probably constitutional.⁽¹¹⁾ Furthermore, once the government has acquired the sample consistently with the Fourth Amendment, the search and seizure clause does not bar its use for another purpose.⁽¹²⁾ Exigent circumstances also could justify the absence of a warrant.⁽¹³⁾

It also is possible that an order compelling a person to give a sample could be issued on something less than probable cause. In dictum in Davis v. Mississippi,⁽¹⁴⁾ the Supreme Court suggested such a procedure. A woman in Meridian, Mississippi, reported that "a Negro youth" broke into her home and raped her. Police, "without warrants, took at least 24 Negro youths," including Davis, "to police headquarters where they were questioned briefly, fingerprinted, and released without charge."⁽¹⁵⁾ After Davis's fingerprints were discovered to match those lifted from the window sill, he was indicted, tried, and convicted. His objection to the admission of the fingerprint evidence was overruled, and the Mississippi Supreme Court affirmed the conviction on the theory that fingerprint evidence is so reliable that the Fourth Amendment exclusionary rule does not apply to this evidence. The United States Supreme Court reversed. It held that the Fourth Amendment requires the exclusion of evidence that is the fruit of an unreasonable search or seizure, regardless of how reliable that evidence may be. Reasoning that Davis was detained without a warrant and without probable cause, and that he was not merely fingerprinted but interrogated, the Court concluded that the resulting fingerprints were inadmissible. However, the Court's response to the state's argument that an arrest made solely for the purpose of obtaining fingerprints should be allowed without probable cause was less definitive. Although Justice Brennan, writing for the majority of the Court, emphasized that "[d]etentions for the sole purpose of obtaining fingerprints are . . . subject to the constraints of the Fourth Amendment,"⁽¹⁶⁾ he proceeded to write that:

It is arguable, however, that, because of the unique nature of the fingerprinting process, such detentions might, under narrowly defined circumstances, be found to comply with the Fourth Amendment even though there is no probable cause in the traditional sense . . . . Detention for fingerprinting may constitute a much less serious intrusion upon personal security than other types of police searches and detentions. Fingerprinting involves none of the probing into an individual's private life and thoughts that marks an interrogation or search. Nor can fingerprint detention be employed repeatedly to harass an individual, since the police need only one set of each person's prints. Furthermore, fingerprinting is an inherently more reliable and effective crime solving tool than eyewitness identifications or confessions . . . . Finally, because there is no danger of destruction of fingerprints, the limited detention need not come unexpectedly or at an inconvenient time.⁽¹⁷⁾

Thus, the Court opened the door to the possibility that "the requirements of the Fourth Amendment could be met by narrowly circumscribed procedures for obtaining, during the course of a criminal investigation, the fingerprints of individuals for whom there is no probable cause to arrest."⁽¹⁸⁾

Many states seized on this invitation by adopting statutes or court rules permitting the police to obtain evidence of identifying physical characteristics after a showing of founded or reasonable suspicion.⁽¹⁹⁾ For instance, an Arizona statute authorizes magistrates to issue "an order authorizing . . . temporary detention, for the purpose of obtaining evidence of identifying physical characteristics"on a showing of "[r]easonable cause for belief that a felony has been committed" and proof that the "physical characteristics . . . may contribute to the identification of the individual who committed such offense."⁽²⁰⁾ As in this instance,⁽²¹⁾ the language of many of these statutes and court rules is broad enough to apply to DNA samples.⁽²²⁾ Although the Supreme Court has not ruled authoritatively on the constitutionality of these procedures, in Hayes v. Florida,⁽²³⁾ the Court referred approvingly to its previous statement in Davis.⁽²⁴⁾

The conclusion that a court order based on probable cause (or perhaps reasonable suspicion) normally is required applies even if the police do not themselves demand or collect the DNA sample, but direct or request private citizens to acquire the sample. Suppose that shortly after an incident, a suspect goes to or is taken to a private hospital. While the suspect is still at the hospital, the police learn of the suspect's location. At that point, the police contact the hospital staff and request them to obtain a DNA sample from the suspect for law enforcement use. The private hospital would be acting as a government agent in making the intrusion, and the Fourth Amendment would apply.⁽²⁵⁾

The police also might obtain a suspect's DNA sample surreptitiously, without detaining the person. Saliva deposited on a coffee cup at a restaurant, for example, could be collected without probable cause or a warrant if one accepts the theory that the individual, having exposed the material to the public, retains no reasonable expectation of privacy in it.⁽²⁶⁾

Rather than compel a person to submit to DNA sampling, police might obtain DNA data on a suspect from preexisting samples or databases. As of 1998, it was estimated that there were more than 282 million specimens of human biological materials stored in the United States, with samples from another 20 million individuals accumulating each year.⁽²⁷⁾ Such samples are stored in academically based repositories of scientists studying genetic disorders, commercially based repositories that offer DNA banking as a service to researchers and individuals, teaching and other hospitals that have acquired samples in the course of clinical diagnostic or surgical procedures, laboratories that screen blood samples of newborns for metabolic or other diseases pursuant to state public health laws, and Armed Forces repositories of pathology specimens and samples collected to permit identification of human remains.⁽²⁸⁾ Although the Fourth Amendment plainly applies to police efforts to obtain samples directly from suspects,⁽²⁹⁾ the prohibition of unreasonable searches applies only to government action.⁽³⁰⁾ As the Supreme Court commented in United States v. Jacobsen,⁽³¹⁾

This Court has . . . consistently construed this protection as proscribing only governmental action; it is wholly inapplicable "to a search or seizure, even an unreasonable one, effected by a private individual not acting as an agent of the Government or with the participation or knowledge of any governmental official."⁽³²⁾

This state action doctrine raises the possibility that police may be able to acquire preexisting information from cooperative private hospitals or laboratories without a court order and without probable cause or reasonable suspicion.⁽³³⁾ This practice would fall outside the constraints of the Fourth Amendment if (1) the government did not instigate the original acquisition of the data, and (2) in acquiring the data that the suspect already has provided private entities, the state is not engaging in any search or seizure. The first condition is relatively straightforward. It often will be satisfied for medical records and tissue samples.⁽³⁴⁾ Pathology specimens at private hospitals, for example, would fall into this category, but newborn screening samples compelled under state law would not. The second condition is more subtle, for it depends on the meaning given to the phrase "search or seizure."

The basic framework for determining whether a form of data collection amounts to a search or seizure for Fourth Amendment purposes can be found in Katz v. United States,⁽³⁵⁾ decided in 1967. In Katz, the government acquired key evidence to convict the defendant of interstate gambling by attaching an electronic listening and recording device to the outside of a public telephone booth. The government argued that the interception was not a search because there was no physical trespass and the telephone booth was a public place. The Supreme Court held that neither entry onto private property nor inspection of tangible items is an essential feature of a search, for "the Fourth Amendment protects people, not places."⁽³⁶⁾ It protected the defendant, the Court explained, because "a person in a telephone booth . . . who occupies it, shuts the door behind him, and pays the toll that permits him to place a call is surely entitled to assume that the words he utters into the mouthpiece will not be broadcast to the world."⁽³⁷⁾ Because the federal agents had no warrant authorizing the interception, the majority held that the search violated the Fourth Amendment. In a concurring opinion, Justice Harlan elaborated on the majority's remarks. In perhaps the most famous passage in the opinions to emanate from the Justices in Katz, he wrote:

[T]here is a twofold requirement, first that a person have exhibited an actual (subjective) expectation of privacy and, second, that the expectation be one that society is prepared to recognize as "reasonable."⁽³⁸⁾

Applying this standard, he explained that "[t]he point is not that the booth is 'accessible to the public' at other times, but that it is a temporarily private place whose momentary occupants' expectations of freedom from intrusion are recognized as reasonable."⁽³⁹⁾

Although the courts have applied the Katz "test" in many contexts, for present purposes, the most important is the Supreme Court's 1976 decision in United States v. Miller.⁽⁴⁰⁾ In Miller, the accused was charged with possessing an unregistered still, carrying on a distillery business without paying the whiskey tax, and possessing whiskey on which no taxes had been paid. Prior to trial, the government served subpoenas on two banks at which the defendant had accounts. The banks surrendered copies of the defendant's checks and deposit slips as well as the bank's own records of the defendant's accounts. Before trial, the defendant moved to suppress the documents. The trial judge denied the motion and admitted the evidence at trial. The defendant was convicted and later appealed. The Court of Appeals for the First Circuit reversed, holding that the banks' surrender of the records violated the defendant's Fourth Amendment rights.

On appeal, the Supreme Court reversed the First Circuit's decision and reinstated the defendant's conviction. Writing for the majority, Justice Powell relied heavily on Katz. The Court upheld the denial of the suppression motion because "there was no intrusion into any area in which [the defendant] had a protected Fourth Amendment interest . . . ."⁽⁴¹⁾ Since the defendant had transferred the checks and deposit slips to the bank, the defendant could not assert "ownership []or possession" as to any of the subpoenaed records,⁽⁴²⁾ and "[a]ll of the documents obtained, including financial statements and deposit slips, contain only information voluntarily conveyed to the banks and exposed to their employees in the ordinary course of business."⁽⁴³⁾

The defendant had argued that he retained an expectation of privacy because he had made the information available to the bank only "for a limited purpose."⁽⁴⁴⁾ However, Justice Powell made short shrift of that argument:

The depositor takes the risk, in revealing his affairs to another, that the information will be conveyed by that person to the Government. . . . This Court has held repeatedly that the Fourth Amendment does not prohibit the obtaining of information revealed to a third party, even if the information is revealed on the assumption that it will be used only for a limited purpose and the confidence placed in the third party will not be betrayed.⁽⁴⁵⁾

Justice Powell concluded that even if the banks were "acting solely as Government agents in" transcribing and surrendering the information, there was "no intrusion upon the [defendant's] Fourth Amendment rights."⁽⁴⁶⁾

The logic of Miller is easily applied to medical samples or records. Like financial information, the patient or research subject neither owns nor possesses tissue samples or data that have been "voluntarily conveyed" to health care providers or medical researchers. Thus, in many of the cases challenging police requests for medical samples and records, the lower courts invoke Miller to defeat the defendant's claim.⁽⁴⁷⁾ Nevertheless, Miller might be distinguished on at least three bases. First, the Miller Court stated that the bank records in question concerned "commercial transactions."⁽⁴⁸⁾ In contrast, a medical record can relate to far more intimate aspects of a person's life.⁽⁴⁹⁾ Recent polls indicate that the vast majority of Americans are concerned about the privacy of their medical records.⁽⁵⁰⁾

Although it seems clear that a patient has a greater privacy interest in medical records than a business or individual has in bank records,⁽⁵¹⁾ it is not clear that the interest is so much deeper that the two types of records are distinguishable when the prosecution seeks the records to advance a criminal investigation.⁽⁵²⁾ Most jurisdictions have enacted a general physician-patient privilege,⁽⁵³⁾ and the privilege typically encompasses not only communications in a conventional sense, but also any information that the physician gains by virtue of the examination.⁽⁵⁴⁾ However, most jurisdictions do not recognize the privilege in criminal cases,⁽⁵⁵⁾and several courts have pointed to that limitation as evidence that society is unprepared to recognize a constitutionally enforceable privacy expectation in medical records or samples. In Thurman v. State,⁽⁵⁶⁾ for instance, the Texas Court of Appeals pointed out that the medical privilege is inapplicable to criminal cases;⁽⁵⁷⁾ hence, even if the citizen has a subjective expectation of privacy, society does not recognize that expectation as reasonable in a criminal setting.⁽⁵⁸⁾ For that matter, many jurisdictions not only carve out an exception to the privilege for criminal proceedings; they go farther and require physicians to report certain types of events such as violent injuries and child abuse to the public authorities.⁽⁵⁹⁾ A few jurisdictions have enacted even more sweeping legislation, requiring hospitals to furnish blood test results to the prosecutor on request in a criminal investigation.⁽⁶⁰⁾ The limited scope of the medical privilege and reporting requirements strongly suggest that the person who is the subject of the medical record lacks a "societally-recognized," constitutionally protected privacy expectation.⁽⁶¹⁾

Second, Miller might be distinguished by arguing that even if it permits the government to obtain medical records from a private source such as a hospital, it does not apply to a tissue sample. Again, it seems plausible to argue that there is a more intense privacy expectation in the sample than in the records reflecting the results of a test of the sample. The sample represents a greater threat to privacy in that its existence would permit further testing and the revelation of information in addition to that already contained in the record. Although noting that "[t]he precise question as to who owns the blood upon extraction from an individual raises a novel point without apparent judicial precedent," a New York trial court suggested that the "defendant had a [property] interest in the blood specimen in . . . vial containers" retained by a private hospital.⁽⁶²⁾ Also, in upholding subpoenas, courts occasionally stress that the records in question were "made [and] kept . . . by the hospital"⁽⁶³⁾ or that the records were not "personal papers created or kept by" the defendant,"⁽⁶⁴⁾ suggesting that the outcome might be different if the defendant personally had produced the subpoenaed items.

Yet, most courts construe Miller to apply to biological samples as well as to mere records.⁽⁶⁵⁾ In Miller itself, the prosecution sought not only financial statements that the bank had generated, but also checks and deposit slips from the defendant. Miller expressly rejected the argument that there was a significant difference between the documents generated by the bank and those prepared by the depositor.⁽⁶⁶⁾ Financial statements prepared by a bank are like medical records prepared by a hospital or laboratory, and checks and deposit slips from the depositor are like samples from the defendant. To this extent, Miller appears pertinent whether the government seeks the original genetic samples or merely records documenting the results of tests on those samples.

Finally, in Miller, the defendant "voluntarily conveyed" the information to the banks,⁽⁶⁷⁾ but the voluntariness of providing tissue samples might be questioned. In People v. Perlos,⁽⁶⁸⁾ a leading case involving blood alcohol testing, a dissenting justice of the Michigan Supreme Court argued that although under the implied consent statute a driver agrees in advance to a government test of his or her alcohol concentration, the driver does not consent to a search of medical records reflecting a test conducted by a private entity. As the dissent observed, "[i]n today's society, a person has little choice but to undergo medical treatment at a medical facility, generally licensed by and authorized to operate by the state. Few persons have the ability to obtain medical treatment in their homes . . . .⁽⁶⁹⁾ Nevertheless, the majority relied on Miller to upheld a state statute mandating that hospital personnel disclose to the prosecution the results of any blood alcohol test of a driver involved in an accident.⁽⁷⁰⁾

When a conscious person in need of medical treatment is admitted to a hospital, the patient consents to treatment⁽⁷¹⁾ and explicitly or implicitly agrees to medical testing incident to the treatment. The patient therefore voluntarily conveys the data disclosed by the test results to the institution's staff in the same manner that the depositor in Miller "voluntarily conveyed [information] to the banks [to be] exposed to their employees in the ordinary course of business."⁽⁷²⁾ Given the rules of medical ethics, the patient might have a stronger subjective expectation that the hospital will keep the information in question confidential. However, Miller states that when a person voluntarily reveals information to a third party, for Fourth Amendment purposes the person "takes the risk" that the third party will disclose the information "to Government authorities, even if the information is revealed only for a limited purpose and the [person expects that] confidence placed in the third party will not be betrayed."⁽⁷³⁾

In sum, the argument that Miller governs both medical records and samples is strong, but not conclusive. The statutory patterns, restricting the medical privilege and imposing reporting duties on physicians, lend powerful support to the view that there is no reasonable expectation of privacy in medical records or samples. Although a minority of cases reject this conclusion,⁽⁷⁴⁾ sometimes on dubious grounds,⁽⁷⁵⁾ and although Miller itself has been the target of much criticism,⁽⁷⁶⁾ the majority view is that if the private hospital or laboratory obtains a biological sample on its own initiative for lawful medical reasons, its subsequent surrender of the sample to the authorities does not violate any constitutionally protected expectation of privacy.⁽⁷⁷⁾ The same is true, even more clearly, of laboratory findings or medical records involving the samples.⁽⁷⁸⁾

In addition to compelling individuals to submit DNA samples or to acquiring samples indirectly from medical care providers or researchers, the authorities might simply ask a suspect to provide them with a sample. Even if the acquisition of the sample constitutes a seizure under the Fourth Amendment and the authorities do not obtain a warrant or court order, a suspect's consent is an adequate justification for a search or seizure.⁽⁷⁹⁾ What, then, must authorities do to obtain legally effective consent? When that question arose with respect to waivers of the privilege against self-incrimination embodied in the Fifth Amendment, early Supreme Court cases applied a general voluntariness test.⁽⁸⁰⁾ However, in 1966, the Court mandated in Miranda v. Arizona⁽⁸¹⁾ that the police administer specific warnings to a suspect in custody to ensure that any consent to interrogation was voluntary in the specific sense that the suspect's waiver was intelligent and knowing. In particular, police must advise a suspect of the right not to remain silent.

The Court has taken a different approach to waivers of the Fourth Amendment right to be free from unreasonable searches and seizures. In Schneckloth v. Bustamonte,⁽⁸²⁾ the Court confirmed the continued applicability of a general voluntariness test and expressly held that police need not warn a suspect of the right to refuse to consent to a search. The existence of such a warning is simply one factor in the totality of the circumstances that must be considered in determining the voluntariness of the consent. In Schneckloth, the Court sharply distinguished between the Fourth and Fifth Amendment settings. The Court emphasized that while Miranda warnings helped to ensure the reliability of any confession by a suspect, the Fourth Amendment exclusionary rule has little or nothing to do with the reliability of the seized evidence. According to the Court, concerns about the reliability of the evidence and the integrity of the factfinding process justify a more rigorous standard under the Fifth Amendment than under the Fourth.

Although the Schneckloth standard is relatively lax,⁽⁸³⁾ in some cases Fourth Amendment consent has been found involuntary. For example, in Bumper v. North Carolina,⁽⁸⁴⁾ four police officers went to the house of "a 66-year-old Negro widow . . . located in a rural area at the end of an isolated, mile-long dirt road."⁽⁸⁵⁾ She met the officers at the front door. One of them announced, "I have a search warrant to search your house." She responded, "Go ahead," and opened the door. In the kitchen the officers found the rifle that was later introduced in evidence at the trial of her grandson for rape. The Court reversed the conviction because the officers had no search warrant. The Court explained that:

When a law enforcement officer claims authority to search a home under a warrant, he announces in effect that the occupant has no right to resist the search. The situation is instinct with coercion--albeit colorably lawful coercion. Where there is coercion there cannot be consent.⁽⁸⁶⁾

Under Bumper, consent to submitting a DNA sample would be involuntary if, for instance, the police gave the suspect the impression that he or she had no alternative other than to provide the sample. However, on balance, consent should be deemed sufficient when the police make it clear that they seek a sample "for criminal investigation purposes"⁽⁸⁷⁾ and avoid statements that could mislead the suspect into believing that there is a legal duty to furnish the sample when there is not.⁽⁸⁸⁾

European countries have solved difficult murder cases by appealing to all local residents to submit to DNA testing.⁽⁸⁹⁾ In perhaps the largest such mass screening, 16,400 men in western Germany were tested in the hunt for an 11-year-old girl's killer. A 30-year-old man arrested after his DNA was found to match, confessed to raping, stabbing and killing the girl, as well as to raping another 11-year-old girl.⁽⁹⁰⁾ In the United Kingdom alone, police have conducted 118 such "mass screens," resulting in 48 hits and seven convictions.⁽⁹¹⁾

In principle, mass screening could be used in the United States.⁽⁹²⁾ Nothing in the Fourth Amendment prevents the police from approaching everyone in a community and asking for their cooperation. The "dragnet" nature of the inquiry is no obstacle.⁽⁹³⁾ A consensual contact between a police officer and a citizen is neither a search nor a seizure under the Fourth Amendment.⁽⁹⁴⁾ As a legal matter, police may ask anyone and everyone to give DNA; as long as they do not engage in coercion or misrepresentation, they may collect samples for analysis.⁽⁹⁵⁾ The practice seems less likely to be effective in this country, however, since the number of residents who would choose not to cooperate could be larger. In addition, the drain on police resources in creating, what is in effect, an ad hoc database, usually would be excessive.⁽⁹⁶⁾

To help trace the flow of human populations, geneticists and anthropologists have located genetic markers that help distinguish among ancestral populations, and various genes are known to have alleles that occur predominantly in certain racial or ethnic groups.⁽⁹⁷⁾ In addition, genetic typing permits inferences as to inherited disorders and may offer clues to facial or other bodily features.⁽⁹⁸⁾ Learning that the person whose DNA was found at a crime scene might have such physical characteristics could well be useful in some criminal investigations.

DNA analysis to conduct such "physical profiling" poses few constitutional problems. The principal issue arises under the equal protection clause of the Fourteenth Amendment.⁽⁹⁹⁾ Normally, the government is free to draw whatever reasonable lines it wishes in adopting and enforcing the law. The Internal Revenue Service, for instance, can choose to concentrate its investigations of tax evasion on higher-income taxpayers. Some classifications, however, are suspect.⁽¹⁰⁰⁾ Race is the prototypical example.⁽¹⁰¹⁾ Imposing the death sentence on the killers of whites but not blacks would be impermissible;⁽¹⁰²⁾ likewise, a police officer who adopted a policy of arresting only African-Americans would be depriving those citizens of the equal protection of the law.⁽¹⁰³⁾ Does the fact that race is a suspect classification prohibit the government from conducting or funding research to develop or refine genetic markers for racial identification? Or, when it appears from such markers that the source of the crime-scene DNA is likely to belong to a particular racial or ethnic group, does the equal protection clause prohibit the police from using that fact as an investigative lead and focusing on members of that group?

These questions require us to examine the purpose and impact of the racial classification. That race is a suspect classification does not mean that the government never can inquire into race. To the contrary, the collection and analysis of information about race are commonplace in enforcing the law and in criminological research undertaken or funded by the government. More generally, a great deal of social science and medical research supported or conducted by the government involves the collection of data on race and the analysis of race as a variable of interest. Likewise, if using physical evidence of race to focus an investigation were impermissible, police could not rely on an eyewitness's report that a person fleeing the scene of a crime was Hispanic, on a victim's report that a rapist was white, or on a linguist's analysis of accent or word choice in a recorded death threat that suggested that the caller was African-American. These reports could be in error in any given case, but if they are generally accurate, paying attention to them is not unconstitutional.⁽¹⁰⁴⁾

The government can rely on racial information in a criminal investigation because the practice does not unfairly burden any racial group.⁽¹⁰⁵⁾ In cases in which racial classifications have been struck down, the explicit purpose or actual use of the racial classification was to burden or stigmatize a racial group. Yick Wo v. Hopkins⁽¹⁰⁶⁾ offers an early illustration. In 1880, San Francisco passed an ordinance requiring that persons obtain a permit before operating laundries in wooden structures. Yick Wo was convicted of operating such a laundry without a permit. The Supreme Court set aside the conviction because it concluded that city officials had issued permits with "an evil eye and an uneven hand."⁽¹⁰⁷⁾ Almost without exception, permits were denied to Chinese and granted to non-Chinese. Thus, the permit requirement, although not explicitly racial, was used to exclude the Chinese from the laundry business. It burdened this group for no legitimate reason.

A different question is presented when racial information is collected and used evenhandedly to advance legitimate state interests and in ways that are not designed to disadvantage any individual because of race. In these situations, it has been held that the government can record the racial information. For instance, Hamm v. Virginia State Board of Elections⁽¹⁰⁸⁾ involved an equal protection challenge to a Virginia law that required every decree of divorce to recite the race of the spouses. A three-judge district court upheld this record-keeping provision because the racial information served the valid purpose of collecting social statistics and did not single out or burden any racial group.⁽¹⁰⁹⁾ The Supreme Court affirmed without discussion.

Under these principles, governmental sponsorship of research on the variations of particular alleles across races and the investigative use of alleles that are reasonably accurate indicators of race should pass constitutional muster. Two factors are crucial: no group is singled out for special treatment, and no one is penalized because of hostility toward race.⁽¹¹⁰⁾ If the police make investigative use of racial information whenever that information is useful, then all racial groups are treated alike, and none is stigmatized or disadvantaged in the enforcement of laws that apply with equal force to members of every race.

It is true that the information could have a disparate impact and lead to the apprehension of more criminals from one race than another--but not because of official (or even unofficial) hostility toward particular races or individual prejudices about those races. Recent years have witnessed outcries against "racial profiling" in policing.⁽¹¹¹⁾ The constitutional defect in this practice is that authorities unfairly target minorities for traffic stops or arrests. As in Yick Wo, laws that are neutral on their face--that do not explicitly classify people by race--can be applied disproportionately to racial minorities.⁽¹¹²⁾ With DNA samples from crime-scenes, however, statistically valid inferences as to race cannot lead the authorities to target minorities because of subjective racial stereotypes or prejudices.⁽¹¹³⁾ If anything, by focusing the investigation on the pertinent physical characteristics--whatever they may be--reliance on genetic information in crime-scene samples could correct any tendency to pursue one racial group exclusively or disproportionately. For example, if DNA analysis indicated that the source of a sample was more likely to be Caucasian than African-American, it might help overcome a stereotypical assumption that only blacks need be considered as prime suspects. By providing objective information, DNA analysis could serve as an antidote to the objectionable form of "racial profiling" in police work.

Of course, this is not to say that the government should institute a research program to develop more precise DNA markers for racial identity or that police should use existing markers that are demonstrated to be informative as to race. Our conclusion is simply that these are policy choices to be made about a developing technology--these options are not foreclosed by the Constitution as it been construed to date.

In courts in the United States, two major standards exist for deciding whether scientific findings will be admitted into evidence: the "general-acceptance" test and the "sound-methodology" standard.⁽¹¹⁴⁾ If a timely objection is raised, the judge must determine whether the applicable standard has been met. The general-acceptance standard was first articulated in an influential 1923 federal court of appeals case, Frye v. United States.⁽¹¹⁵⁾ In jurisdictions that follow Frye, the proponent of the scientific evidence typically must establish that the underlying theory and methodology are generally accepted within the relevant portions of the scientific community.

The sound-methodology standard is derived from phrases in the Federal Rules of Evidence. In Daubert v. Merrell Dow Pharmaceuticals, Inc.,⁽¹¹⁶⁾ the Supreme Court held that these rules implicitly jettison general acceptance as an absolute prerequisite to the admissibility of scientific evidence. Instead of the Frye test, the court prescribed a broader framework for deciding whether proposed testimony has sufficient scientific validity and reliability to be admitted as relevant "scientific knowledge" that would "assist the trier of fact." In that framework, the lack of general acceptance weighs against admissibility but is not invariably fatal. It is circumstantial evidence that the technology has not been studied widely or that the bulk of the specialists who have studied the technique have found it wanting. The court discussed other factors that might be considered. Its nonexhaustive list includes the extent to which the theory and technology have been tested, the existence of a body of peer-reviewed studies, and the known error rates of the procedure.

Labels like "general acceptance" and "sound methodology" are just that--labels. Cases decided in each jurisdiction help to define the scientific community in which the degree of scientific acceptance is to be ascertained, the extent of disagreement that can be tolerated, the information that may be used to gauge the extent of consensus, and the specific factors other than general acceptance that bear on relevance and helpfulness. The degree of scientific consensus is important to the admissibility of scientific evidence in all jurisdictions, and pretrial hearings in hotly contested cases have lasted months and generated thousands of pages of testimony probing the opinions of experts on various aspects of DNA profiling. The courts have examined affidavits or testimony from scientists selected by the parties, specific papers in scientific periodicals, the writings of science journalists, the body of court opinions, and other scientific and legal literature.

In particular, the judicial reception of DNA evidence can be divided into at least five phases.⁽¹¹⁷⁾ The first phase was one of rapid acceptance. Initial praise for RFLP testing in homicide, rape, paternity, and other cases was effusive. Indeed, one judge proclaimed "DNA fingerprinting" to be "the single greatest advance in the 'search for truth' . . . since the advent of cross-examination."⁽¹¹⁸⁾ In this first wave of cases, expert testimony for the prosecution rarely was countered, and courts readily admitted RFLP findings.⁽¹¹⁹⁾

In a second wave of cases, however, defendants pointed to problems at two levels--controlling the experimental conditions of the analysis and interpreting the results.⁽¹²⁰⁾ Some scientists questioned certain features of the procedures for extracting and analyzing DNA employed in forensic laboratories, and it became apparent that determining whether RFLPs in VNTR loci in two samples actually match can be complicated by measurement variability or missing or spurious bands.⁽¹²¹⁾ Despite these concerns, most cases continued to find forensic RFLP analyses to be generally accepted,⁽¹²²⁾ and a number of states provided for admissibility of DNA tests by legislation.⁽¹²³⁾ Concerted attacks by defense experts of impeccable credentials, however, produced a few cases rejecting specific proffers on the ground that the testing was not sufficiently rigorous.⁽¹²⁴⁾ Moreover, a minority of courts, perhaps concerned that DNA evidence might well be conclusive in the minds of jurors, added a "third prong" to the general acceptance standard.⁽¹²⁵⁾ This augmented Frye test requires not only proof of the general acceptance of the ability of science to produce the type of results offered in court, but also of the proper application of an approved method on the particular occasion.⁽¹²⁶⁾ Whether this inquiry is properly part of the special screening of scientific methodology, however, is debatable.⁽¹²⁷⁾

A different attack on DNA profiling that began in cases during this period proved far more successful and led to a third wave of cases in which many courts held that estimates of the probability of a coincidentally matching VNTR profile were inadmissible.⁽¹²⁸⁾ These estimates relied on a simplified population-genetics model for the frequencies of VNTR profiles that treats each race as a large, randomly mating population. Some prominent scientists claimed that the applicability of the model had not been adequately verified.⁽¹²⁹⁾ A heated debate on this point spilled over from courthouses to scientific journals and convinced the supreme courts of several states that general acceptance was lacking.⁽¹³⁰⁾ A 1992 report of the National Academy of Sciences proposed a more "conservative" computational method as a compromise,⁽¹³¹⁾ and this seemed to undermine the claim of scientific acceptance of the less conservative procedure that was in general use.⁽¹³²⁾

In response to the population-genetics criticism and the 1992 NAS report came an outpouring of critiques of the report and new studies of the distribution of VNTR alleles in many population groups. Relying on the burgeoning literature, a second National Academy panel concluded in 1996 that the usual method of estimating frequencies of VNTR profiles in broad racial groups was sound.⁽¹³³⁾ In the corresponding fourth phase of judicial scrutiny of DNA evidence, the courts almost invariably returned to the earlier view that the statistics associated with VNTR profiling are generally accepted and scientifically valid.⁽¹³⁴⁾

The fifth phase of the judicial evaluation of DNA evidence is well underway. As results obtained with the new PCR-based methods enter the courtroom, it becomes necessary to ask whether each such method rests on a solid scientific foundation or is generally accepted in the scientific community.⁽¹³⁵⁾ Sometimes, the answer will be obvious even without an extensive pretrial hearing,⁽¹³⁶⁾ and the opinions are practically unanimous in holding that the more commonly used PCR-based procedures satisfy these standards.⁽¹³⁷⁾

In sum, in little more than a decade, DNA typing has made the transition from a novel set of methods for identification to a relatively mature and well studied forensic technology. However, one should not lump all forms of DNA identification together. New techniques and applications continue to emerge. These range from the use of new genetic systems and new analytical procedures to the typing of DNA from plants and animals.⁽¹³⁸⁾ Before admitting such evidence, it will be necessary to inquire into the biological principles and knowledge that would justify inferences from these new technologies or applications.⁽¹³⁹⁾

Having discussed the importance of studies validating new DNA technologies in determining the admissibility of evidence obtained through these technologies, we turn to a different type of study--namely, a study of the proficiency of the laboratory conducting the test. In a validation study, the researchers empirically verify the ability of the technology to identify features of DNA molecules. In a proficiency study, the focus is on how competently the laboratory's analysts apply a technology that already has been validated.⁽¹⁴⁰⁾ The purpose of proficiency testing is to uncover difficulties that a particular technician or a particular laboratory might be encountering in applying established methods.

Proficiency testing raises a variety of legal issues. It has been suggested that participation in a program of proficiency testing should be a prerequisite to the admission of evidence from a forensic laboratory,⁽¹⁴¹⁾ that proficiency test results should be admissible to show how likely it is that the laboratory erred in the test at bar,⁽¹⁴²⁾ and that random match probabilities should be inadmissible unless they are combined with proficiency test results to estimate the probability of a false match. And, if the second suggestion is followed, and the defense is allowed to introduce evidence of proficiency tests to suggest that the laboratory is prone to err, it could be argued that the prosecution should be permitted to present testimony that the defense has not retested or even requested the opportunity to retest the samples.⁽¹⁴³⁾

The first suggestion, that courts condition admissibility on proficiency testing, is a departure from the usual practice. As indicated in the previous section, the scientific validity and general acceptance standards relate to the capacity of an analytical procedure to generate accurate results when properly applied, and not to whether the individual or institution using a valid or generally accepted method is skilled and careful or is instead careless and prone to error.⁽¹⁴⁴⁾ Of course, the latter issue can be of paramount importance, but usually it is said to be a matter affecting the weight of the evidence rather than its admissibility.⁽¹⁴⁵⁾

The second suggestion, that testimony about proficiency test results be used to reveal the chance of error in the case at bar, presupposes that such evidence is admissible at trial. In its 1992 report, a committee of the National Academy of Sciences took the position that "laboratory error rates must be continually estimated in blind proficiency testing and must be disclosed to juries."⁽¹⁴⁶⁾ Some courts then held that when the prosecution introduces testimony about the probability of a coincidentally matching profile, the defendant is entitled to introduce testimony about the laboratory's proficiency tests.⁽¹⁴⁷⁾ Indeed, it has been held that the opponent must be allowed to cross-examine one laboratory representative about errors committed by other analysts at the laboratory.⁽¹⁴⁸⁾

In contrast, in a report published in 1996, a second committee of the National Academy declined to take a position on whether evidence of laboratory error rates, as estimated from proficiency studies, should be admissible at trial.⁽¹⁴⁹⁾ However, the report's discussion of proficiency testing raises questions about the probative value of such evidence. For example, the report notes that "[t]he pooling of proficiency-test results across laboratories" could mislead a jury and "penalize the better laboratories."⁽¹⁵⁰⁾ It adds that even a test of the same laboratory might be outdated, since the laboratory may have taken corrective action.⁽¹⁵¹⁾ In these circumstances, the testimony could be vulnerable to an objection under Federal Rule of Evidence 403, which requires the exclusion of evidence whose probative value is substantially outweighed by the dangers of prejudice, confusion of the issues, or undue consumption of time.⁽¹⁵²⁾

A further objection is that the testimony represents inadmissible character evidence.⁽¹⁵³⁾ If the theory of logical relevance is that the laboratory's past commission of errors increases the probability that the laboratory erred on the occasion in question, then the theory amounts to forbidden character reasoning.⁽¹⁵⁴⁾ It is precisely the theory of logical relevance generally banned by Federal Rule of Evidence 404.⁽¹⁵⁵⁾ Moreover, to the extent that proficiency test results constitute evidence of specific acts introduced to show a general tendency to make mistakes, they seem to run afoul of Rule 405, which forbids this form of proof of character.⁽¹⁵⁶⁾

This issue is rarely recognized in the trial court,⁽¹⁵⁷⁾ but it is difficult to see how a trial judge could justify overruling a properly phrased character-evidence objection when the theory of relevance is nothing more than a general tendency of the laboratory to make mistakes. If there is a consensus that the jury sometimes needs the proficiency test results as an antidote to overwhelmingly small random match probabilities, then the federal and state rules governing character evidence would need to be altered to give the trial court the discretion to admit the evidence.

In some circumstances, however, proficiency tests of the laboratory involved in the case will be admissible without relaxing the ban on character evidence. The ban applies only when the sole theory of logical relevance is that the existence of errors in the past suggests a tendency to err that might affect the result in the case at bar. There might be situations in which the defense can use the test data at trial on an entirely different theory of logical relevance. Assume, for instance, that the experts in a case disagree over whether a peak or a band observed in a DNA test is due to an allele or is an artifact. Evidence that spurious peaks or bands have occurred under similar circumstances in proficiency tests on known samples would lend support to the defense theory that the band in the pending case is an artifact. In this situation, proficiency test data are relevant because they provide information about the operating characteristics of the DNA test at that particular laboratory.⁽¹⁵⁸⁾

The third suggestion relating to proficiency testing is that random-match probabilities should be inadmissible unless accompanied by or blended with the laboratory's error rate.⁽¹⁵⁹⁾ The 1996 committee observed that combining the two figures "would deprive the trier of fact of the opportunity to separately evaluate the possibility that the profiles match by coincidence as opposed to the possibility that they are reported to match by reason of laboratory or handling error."⁽¹⁶⁰⁾ The committee took the position that "a calculation which combines error rates with match probabilities is inappropriate."⁽¹⁶¹⁾ The reasoning supporting the committee's position essentially sounds under Federal Rule of Evidence 403.⁽¹⁶²⁾ If anything, the Rule 403 objection is more substantial here than when it is urged as a basis for excluding testimony offered to impeach the laboratory's competence. In this situation, the questions about the validity of industry-wide error rates and the staleness of even the laboratory's own tests are equally applicable and call into question the probative worth of the testimony. Moreover, there is a heightened risk that the jury will be confused. Error rates and random match probabilities relate to distinct hypotheses, and a lay juror may find it difficult to understand the significance of a computation which merges the rates and the probability. That mode of computation could place even greater strain on the jurors' ability to comprehend the body of evidence submitted to them.⁽¹⁶³⁾ The few courts that have addressed the argument that error rates should be used to the exclusion of random-match probabilities have not been persuaded.⁽¹⁶⁴⁾

d. The opportunity to retest as a response to defense arguments about proficiency testing

The fourth suggestion related to proficiency testing has been made by prosecutors. This suggestion is that when the defense is allowed to introduce evidence of proficiency tests of the laboratory employing the prosecution expert to suggest that the laboratory is prone to err, the prosecution should be permitted to present testimony that the defense has not retested or even requested the opportunity to retest the samples analyzed by the prosecution expert.

The testimony would be logically relevant on several theories. To begin with, if a defense expert testifies that the laboratory result is untrustworthy, it would be relevant to impeach the defense expert's credibility on the ground that a scientist who truly doubted the accuracy of the analysis normally would have retested the samples to resolve the matter.⁽¹⁶⁵⁾ Inasmuch as replication is a crucial and common feature of scientific inquiry,⁽¹⁶⁶⁾ it could be argued that neglecting to retest is prior inconsistent conduct. On this theory, the defense would be entitled to a limiting instruction to the effect that the expert's failure to retest is not offered to show that the test result is correct, but only to demonstrate that the defense expert is not sincere in asserting that it is flawed.⁽¹⁶⁷⁾

The value of a failure to retest in showing an expert's insincerity, however, is open to question. It is not uncommon for scientists to question in print or otherwise the adequacy of another researcher's experiment before undertaking to replicate it. And even if such opinions were unheard of in the course of ordinary science, the expert may have been retained for the limited purpose of giving an opinion on the adequacy of the testing that was done rather than redoing that testing. Nevertheless, the inference of insincerity need not be particularly strong for the "inconsistent" conduct to be a proper subject for cross-examination.⁽¹⁶⁸⁾

Second, if the defense expert offers an opinion that the laboratory's results may be in error, the expert's failure to request or conduct an independent test would be relevant to suggest that the jury should give less weight to that opinion.⁽¹⁶⁹⁾ The prosecution could argue to the jury that an expert who fails to use a more definitive and readily available procedure for ascertaining whether the initial test results are correct has not been thorough in evaluating those results, and that such experts deserve little credence. Again, the inference may be debatable, but the standard of relevance, particularly on cross-examination, is lenient.

Third, whether or not a defense expert discusses proficiency tests, the prosecution could argue that the defense failure to retest (or to request a retest) amounts to an admission by conduct by the defendant.⁽¹⁷⁰⁾ The courts have applied the admission-by-conduct theory to a litigant's failure to present evidence when "it would be natural" for the litigant to introduce such testimony.⁽¹⁷¹⁾ The prosecution might urge that it would be natural for a defendant affected by a false match to seek retesting and that it would be natural for a DNA expert who entertained serious doubts about the accuracy of a prior test to retest the samples.⁽¹⁷²⁾

In short, there are reasonable arguments for permitting the prosecution to raise the issue of retesting when a defendant questions the laboratory's ability to type DNA samples correctly. But even if the inquiry is probative of the insincerity or lack of thoroughness of the expert, or an admission by the defendant, there are potential objections to this counterthrust by the prosecution. One objection is that the inquiry is inconsistent with the prosecution's burden of proof.⁽¹⁷³⁾ To reinforce the allocation of the burden to the government, some courts generally forbid prosecution comment on the defense failure to produce evidence.⁽¹⁷⁴⁾ The argument runs that the defense is entitled to rely on the burden and has no obligation to present any evidence at trial. According to this line of argument, it is improper to convert the defense's failure to present testimony into prosecution evidence.

A further objection is that the admission of the testimony is inconsistent with the defendant's attorney-client privilege. A number of jurisdictions apply the attorney-client privilege when, as part of trial preparation, defense counsel hires an expert to evaluate private information from the defendant, such as the defendant's mental or physical condition.⁽¹⁷⁵⁾ The Advisory Committee Note to draft Federal Rule of Evidence 503 endorsed the application of the attorney-client privilege to experts,⁽¹⁷⁶⁾ and some courts have gone to the length of invoking the theory even when the expert did not evaluate information realistically originating from the defendant.⁽¹⁷⁷⁾ Based on these authorities, the defense might contend that the attorney-client privilege applies to a defense expert's retest of a DNA sample. The gist of the objection would be that if the result of a retest would be privileged, it is wrong-minded to penalize the defense for failing to retest.

As with the other suggestions related to proficiency testing, the case law offers little guidance. In principle, it would seem that once the defense has sharpened the issue of the prosecution expert's use of proper test procedures, the prosecution should be allowed to elicit testimony about the defense's failure to retest at least to probe the basis for the expert's opinion and as circumstantial evidence of defendant's belief that retesting would not yield a different result. The fact that the prosecution has the burden of persuasion does not make such inferences impermissible.⁽¹⁷⁸⁾ In appropriate circumstances, the majority of courts permit prosecutors to comment on a defendant's failure to produce evidence such as an available witness who would presumably corroborate the defendant's testimony.⁽¹⁷⁹⁾

Neither should the attorney-client privilege pose an insurmountable barrier. Certainly, the prosecution cannot comment on a defendant's decision to exercise a constitutional privilege,⁽¹⁸⁰⁾ and comment on a defendant's failure to produce a witness is often forbidden when the defendant stands in a privileged relationship with the witness.⁽¹⁸¹⁾ Consequently, it might be justifiable to apply the attorney-client privilege to a defense expert's analysis of material that has become available because of the defendant's exercise of the right to prepare a defense with the assistance of counsel. Perhaps material that both emanates from the defendant and is still confidential would fall into this category. However, these conditions do not seem to be satisfied in this setting. The DNA sample that the defendant suggests has been misanalyzed might be crime-scene material that was not obtained from the defendant, or it could be a sample that the prosecution lawfully acquired from the defendant. In these situations, the attorney-client privilege should not preclude adverse comment on the defense failure to retest.

The power of DNA evidence has promoted proposals to create an exception to the statute of limitations for sexual assault when DNA profiling links the suspect to the assault.⁽¹⁸²⁾ However, devising a workable "DNA exception" that would respect the interests of defendants and society in defining a point after which litigation no longer can be commenced is a formidable challenge.⁽¹⁸³⁾

Statutes of limitations serve a variety of purposes. Most obviously, they protect individuals against the risk that they will unable to assemble adequate evidence for a defense because too much time has passed since the alleged crime was committed. With time, memories fade, evidence is misplaced, witnesses become harder to locate, and the accused's ability to defend himself is reduced.⁽¹⁸⁴⁾ Thus, the Supreme Court has described statutes of limitations as "the primary guarantee against bringing overly stale criminal charges."⁽¹⁸⁵⁾ In addition, they give innocent (as well as guilty individuals) a certain peace of mind and encourage the police to move on to more recent cases that are more likely to be solved and for which punishment would be more effective.⁽¹⁸⁶⁾

A DNA exception attends only to the first justification for statutes of limitations. If a comparison of the defendant's DNA with the trace evidence DNA collected many years ago were to establish conclusively that the defendant is guilty,⁽¹⁸⁷⁾ then it could be argued that any degradation in the defendant's ability to mount a defense would be harmless because it could not affect the outcome of the trial. For example, even if a defendant's alibi witness had died after the statutory period, if no reasonable jury could have believed the alibi in the face of the DNA proof, the availability of the witness could not have resulted in an acquittal.

The premise that DNA evidence is dispositive, however, is not always true. First, there are cases in which a defendant might succeed in raising a reasonable doubt about the reported results of the DNA tests. This situation would arise, for instance, when there is reason to think that samples were switched or cross-contaminated in the laboratory or in the collecting and handling of the trace evidence before it reached the laboratory. Many years later, the police officers and laboratory personnel involved could be impossible to locate, and the written records remaining might be inadequate to resolve these claims.

Second, even if one were to conclude that such cases are too rare to be an obstacle to creating an exception to the statute, DNA evidence can be conclusive only as to one factual issue--whether the DNA in the trace evidence somehow originated from the defendant--and proof of that fact ordinarily falls far short of demonstrating guilt for every type of sexual assault. Thus, a defendant's semen might be present on an alleged victim's clothing or a bedsheet even if there had been no penetration, and it would be expected to be found in a vaginal swab if the sex had been consensual.

The legislature might try to respond to these concerns by confining the DNA exception to cases in which identity is the only issue that needs to be resolved.⁽¹⁸⁸⁾ But which cases are these? Can a defendant avoid the extension of the period of limitations by conceding his identity as the source of the trace evidence but alleging that he reasonably believed that the woman invited his actions, that he was acting under duress, or the like? Should the court be required to find that these defenses have no basis in fact for the prosecution to proceed after the statute has run? It might be possible to draft a suitably sensitive DNA exception to the statute of limitations, but the task is not so simple as, initially, it might appear to be.

The situation is much simpler when a defendant brings forward DNA evidence in a case involving a single rapist that shows that the DNA in the trace evidence is not his. If that evidence is believed, then he is not the guilty party. But while DNA evidence can be conclusive of innocence, DNA evidence is not logically sufficient to prove guilt. Because identity is not the only element of the offense and because there are affirmative defenses that can be pled, even when the state brings forward uncontestable DNA evidence of identity, the defendant might not be guilty of sexual assault. Consequently, it would be consistent to advocate an extension of the period in which post-conviction relief can be sought while opposing an extension of the statute of limitations. The time limitation on post-conviction appeals of convictions and the statute of limitations for prosecutions are not symmetrical with respect to DNA evidence.⁽¹⁸⁹⁾

Few statutes prescribe the method of collecting DNA samples.⁽¹⁹⁰⁾ Most statutes provide for flexibility as to the origin of the DNA, referring to such materials as "a blood or tissue sample,"⁽¹⁹¹⁾ "a blood, tissue, or hair follicle sample,"⁽¹⁹²⁾ "a body fluid or tissue sample,"⁽¹⁹³⁾ or simply "[a]ny biological sample containing DNA."⁽¹⁹⁴⁾ However, a few states use narrower phrasing. The California law, for example, insists on "two specimens of blood [and] a saliva sample."⁽¹⁹⁵⁾ The table below lists the number of state statutes referring to various types of sample material:

SAMPLE MATERIAL	NUMBER OF STATES
blood	39
saliva	6
tissue	13
oral	1
body fluid	5
other samples necessary	1
other biological samples	2
any biological evidence	2
human biological specimen	2
scientific biological sample	1
hair follicle	2
buccal cells	1
body fluids	1

Table 1. Types of Biological Samples
for Collecting DNA for State Databanks⁽¹⁹⁶⁾

To encourage the use of the least invasive and most cost-effective collection methods, a legislature might wish to avoid being overly restrictive in defining the types of DNA-bearing biological samples that may be used.

Perhaps the most obvious question in the design of a compulsory databank is whose DNA should be taken. The trend is toward increased coverage. All states require convicted sex offenders to give samples, 36 include murder as a qualifying offense, 27 reach assault and battery, and a minority include other felonies such as kidnapping, burglary, and robbery.⁽¹⁹⁷⁾ At least five states demand samples from all felons.⁽¹⁹⁸⁾ In these jurisdictions, activities such as sodomy, prostitution, and racketeering can result in inclusion in the state databank.⁽¹⁹⁹⁾

Deciding whose DNA should be included in a database is not a simple matter. There are two possible rationales for limiting DNA databanking to offenders in the first place. One is that convicts do not deserve the same level of privacy protection as other citizens.⁽²⁰⁰⁾ This rationale applies to all felons, but the premise that conviction or even imprisonment works a loss of all privacy rights is dubious.⁽²⁰¹⁾ Prisoners do not have the same privacy as individuals who are not serving a criminal sentence: their mail, their telephone conversations, their living quarters, and even their bodies are subject to government inspection.⁽²⁰²⁾ But such searches are justified by the need to run a penal institution, not by the theory that a convict automatically forfeits all civil rights. A more convincing justification for targeting offenders as opposed to all citizens, and for choosing among the classes of offenders, therefore is needed.

Thus, the second rationale for focusing on offenders looks to a pragmatic consideration. The argument is that prior offenders are significantly more likely than non-offenders to commit some crimes that often produce biological trace evidence.⁽²⁰³⁾ This recidivism argument is more complex than folk wisdom like "Once a thief, always a thief," would suggest. To decide which crimes meet this condition, we must consider not merely whether a particular offense tends to be repeated, but whether the commission of any given offense is associated with the commission of any of the offenses for which biological evidence tends to be available. Some notation is helpful in formulating the recidivism argument clearly. Let C stand for a "collection crime" -- one for which DNA can be collected from the offender. Let T stand for a target, or traceable, offense -- one for which biological trace evidence might be found. The predictivist argument requires that the probability that a person guilty of one instance of C will commit (or has committed)⁽²⁰⁴⁾ a target crime T must exceed the probability that a person who has not committed any instance of C will commit (or has not committed) T. In symbols,

where P is a probability, the vertical line means "given," and the hyphen stands for "not." In other words, C must be a risk factor for T: individuals who possess C (a conviction for a collection offense) must be at greater risk for committing a target offense T than people who are not guilty of C.

Not long ago, it was thought that the only plausible target offense was forcible rape. In 1992, a committee of the National Academy of Sciences thought that:

[I]t is clear that crimes of most types will not afford the opportunity to recover relevant biological that will allow the police to identify an unknown suspect--i.e., the perpetrator's own body fluids. They include larcenies, burglaries, and assaults, for which ordinary fingerprints for which ordinary fingerprints are frequently found. They major exception is rape, for which semen samples can be recovered in many cases . . . . [ś] A DNA profile databank would thus be valuable primarily in investigating forcible rape . . . .⁽²⁰⁵⁾

Moreover, the committee apparently assumed that rapists rarely commit other crimes, for it suggested that the only collection crimes for rape should be sex offenses.⁽²⁰⁶⁾

This narrow view of the target and collection crimes is untenable. First, the rapid development of and experience with forensic DNA technology has shown that the notion that rape is the only realistic target crime to have been shortsighted. By 1999, the United Kingdom's Forensic Science Service was reporting a 5% chance of developing a DNA profile for property crimes.⁽²⁰⁷⁾ Although one in twenty might not seem like a large fraction, property crimes are more common than other crimes. Consequently, most of the hits in the database for England and Wales now come from burglary and vehicle theft cases rather than from rapes or murders.⁽²⁰⁸⁾

As for collection crimes C, the assumption that only convictions for sex crimes are predictive of forcible rape always has been questionable. Studies show that convicted rapists released from prison are more likely to be re-arrested for crimes other than rape than they are for another rape. The Bureau of Justice Statistics examined the arrest records of a representative sample of all prisoners released in 1983 from prisons in eleven states.⁽²⁰⁹⁾ During the three-year follow-up period, 7.7% of the 2,214 released rapists were re-arrested for another rape, but over five times this number (42.8%) were rearrested on other charges.⁽²¹⁰⁾ Apparently, rapists do not confine their criminal activities to rape.

Even so, if rapists were an insignificant proportion of the criminal population, it would not follow that a large proportion of individuals who are convicted of other crimes than rape have committed or will commit rapes. Although this situation is conceivable, rapists seem to constitute a substantial fraction of all incarcerated felons.⁽²¹¹⁾ Thus, there is reason to think that including non-sex offenders in databases would yield a meaningful number of additional hits in rape investigations.

Other data confirm this surmise. The Bureau of Justice Statistics examined the conviction histories of offenders serving time in state prisons for rape or sexual assault as of 1994. Its report suggests that persons serving time for sexual offenses were some four times more likely to have been convicted previously of non-sex crimes than of other sex crimes.⁽²¹²⁾ Furthermore, jurisdictions that have expanded their databases report matches between trace evidence samples in rape cases and databank samples from men convicted of burglary⁽²¹³⁾ and aggravated assault.⁽²¹⁴⁾

In sum, even with respect to the single target crime of rape, many offenses are plausible candidates for collection crimes. And, as we have seen, rape is hardly the only justifiable target crime. Trace samples of DNA can assist in solving nonviolent property crimes as well as crimes like assault and murder that can involve transfers of body cells from a victim to an attacker, or vice versa.

Of course, with respect to all the target offenses, the question of whether a prior conviction for a collection offense is indeed a risk factor remains. Data showing that individuals with a history of collection crimes C are more likely to commit a new target crime T than are individuals with no such history to commit such a crime are not readily available. But recidivism rates for many offenses are known to be high--much higher than one would expect the rate of new offenses to be in the general population. In the Bureau of Justice Statistics follow-up study, for example, the rates of re-arrests and re-convictions for rapists were 52% and 36%, respectively. For all violent offenders, the rates were still higher--60% and 42%, respectively.⁽²¹⁵⁾ But the most direct evidence of the value of expanding the range of collection crimes well beyond sexual assault comes from states that have been collecting DNA from many types of criminals since the early 1990s. In Virginia, 60% of all database hits have come from convicted burglars; in Florida, the figure is about 50%.⁽²¹⁶⁾

This is not to say that every state should extend its collection offenses to reach or all, or even most, crimes. As stated at the outset, the coverage issue is not simple. That question requires some evaluation of the marginal costs and benefits of creating and maintaining larger databases, and there is a paucity of data on these matters. We are dealing in probabilities both as to the recurrence of crime and the depositing of DNA, and with a criminal justice system in which the multiplicity of crimes and the practice of plea bargaining can result in very different types of convictions for the same conduct. In addition, the deterrent impact of DNA databases is exceedingly difficult to quantify.

Since the purpose of creating offender databases is to identify the perpetrators of new or previously unsolved crimes, law enforcement personnel must have access to it for the purpose of ascertaining whether trace evidence samples match any records in the database. But the data or samples might be used in other ways as well. For example, retesting certain samples might be useful for quality control and assurance purposes. Statistical research with offender data could produce more accurate estimates of the frequencies of the alleles and genotypes used for identification.⁽²¹⁷⁾ These uses are closely tied to the raison d'ętre of the databanking system--identification of the perpetrators of crimes. But there are other possible uses as well. The data might assist in the identification of missing persons or the victims of major accidents or disasters. They might be employed in certain civil cases, such as paternity actions, in which the genotype of an individual who is not available for DNA testing is on file. Finally, it is conceivable--though far from clear--that a large enough collection of samples might have some value in studying the genetics of some behavioral traits.⁽²¹⁸⁾

Given the variety of uses to which DNA samples might be put, database statutes should enumerate clearly the uses that are permissible. This section surveys the possible uses to which law enforcement databanks or databases might be dedicated.

Naturally, all offender database statutes authorize the DNA samples and records in the databank to be used in criminal investigations to identify the perpetrators of crimes.⁽²¹⁹⁾ As a Texas statute explains, "[t]he principal purpose of the DNA database is to assist federal, state, or local criminal justice or law enforcement agencies in the investigation or prosecution of . . . offenses in which biological evidence is recovered."⁽²²⁰⁾

In performing this core task, it is possible that a database search might implicate an offender's relatives rather the previously convicted offender himself. For instance, if the closest thing to a complete match is a match at 12 out 13 loci, then everyone in the database is excluded, but it is exceedingly likely that the trace evidence DNA comes from a full sibling of the individual who matches at the 12 loci.⁽²²¹⁾ The possibility that a relative might be incriminated as a result of a partial match, it could be argued, invades the relative's genetic privacy and thus makes DNA databanking improper or unreasonable under the Fourth Amendment.⁽²²²⁾

The argument dissolves under analysis.⁽²²³⁾ It is true that the relative might not have been suspected but for the database search, but the search does not violate any cognizable privacy right of the relative. Relatives, like other individuals, have no right to be free from legitimate investigation just because suspicion is triggered by information obtained from another person. Suppose a rape victim studying mugshots states that the face in one of the photographs is very similar to the perpetrator's, but the rapist was "somewhat older." The police happen to know that the person depicted in that photograph has an older brother. Could the older brother object to his becoming a target of the investigation because he did nothing that caused his younger brother's picture to be in the police files? As with the mugshot, if a DNA sample is properly obtained, there is no reason to prohibit its use as an investigative lead to a relative.

Although there is no direct case law on this question, a well-settled Fourth Amendment doctrine supports the conclusion that it would not violate the relative's constitutional rights for the police to follow up on the partial match.⁽²²⁴⁾ The doctrine is that of "plain view."⁽²²⁵⁾ When evidence comes into the plain view of police who are lawfully in place, the state may use that evidence.⁽²²⁶⁾ Here, the police are conducting a lawful DNA database search, and that search yields a partial match.⁽²²⁷⁾

A variation on the claim that the privacy of relatives is invaded by examining the identifying loci of samples in a database posits a prime suspect who has fled the jurisdiction or is otherwise unavailable for DNA testing but who has close relatives in the database. "In this kind of situation," it has been said that "law enforcement officials may argue that it is necessary to compare the DNA of this prime suspect's close biological relatives to that of the DNA artifact [sic] from the crime scene to determine whether to continue or cease pursuit of the suspect."⁽²²⁸⁾ However, the relatives who are legitimately in the database can hardly complain that the state should not be able to compare the crime-scene DNA with all the profiles stored in the database. And, for the reasons already given, neither can the individual who already is "the prime suspect."

In sum, the constitution does not compel any limitations on database searches that might implicate relatives on the ground that the privacy of the relatives is threatened. Neither it is apparent that any sound policy would be served by creating such limitations.

Most statutes do not limit the databanks to the investigation of criminal cases.⁽²²⁹⁾ Another commonly authorized use is the identification of human remains⁽²³⁰⁾ or missing persons.⁽²³¹⁾ Of course, in many circumstances, these investigations can be considered criminal in that they are used to determine whether criminal misconduct is involved, but in cases of mass disasters and certain other situations, the identifications may not be part of the criminal process. Nevertheless, many jurisdictions denominate them an allowable, "humanitarian" use.⁽²³²⁾ Indeed, a few states permit data from "close biological relatives" of missing persons to be included in the state system.⁽²³³⁾ Some states specifically allow the information in the database to be used, pursuant to a court order, in proceedings establishing parent-child relationships,⁽²³⁴⁾ if not in any civil case.⁽²³⁵⁾

The use in missing-person investigations or parentage proceedings has been characterized, somewhat disparagingly, as "familial surveillance."⁽²³⁶⁾ However, where civil discovery mechanisms could be employed to compel the person whose DNA is desired to provide a sample, obtaining a pre-existing sample of DNA would not seem to violate any overriding privacy interest.

DNA databanking statutes typically allow the data on the genotypes of convicted offenders or the associated samples to be used in certain kinds of research. Arguably, merely authorizing sample collection for "law enforcement identification purposes"⁽²³⁷⁾ would permit the information to be used for research into the forensic allele and genotype frequencies, and it would allow the samples to be used for quality control or assurance. Many statutes also explicitly provide for this, usually on the condition that the samples or data be "anonymous"⁽²³⁸⁾ or stripped of "identifying information."⁽²³⁹⁾ Still other statutes authorize "identification research and protocol development of DNA forensic methods"⁽²⁴⁰⁾ It also could be argued, albeit with some difficulty, that the provisions speaking of "law enforcement purposes" or "humanitarian purposes" would countenance releasing samples for research into possible genetic influences on crime or violence.⁽²⁴¹⁾

A few observers have contended that any research use of any person's DNA samples without the person's consent is unethical or illegal.⁽²⁴²⁾ But even in the usual medical research context, the propriety of the research use of archived tissue samples is a more complex matter than these uncompromising pronouncements seem to recognize. The traditional understanding of the ethical limits on research with human subjects is considerably more open to the use of archival data and specimens in medical, epidemiological, and health services studies--even without consent.⁽²⁴³⁾ In any event, DNA samples in convicted-offender databases are different from tissue samples stored in hospitals or research laboratories. For law enforcement samples, there is no issue of consent in the first place. A patient has a right, grounded in personal autonomy, to decline to donate tissues to medical research.⁽²⁴⁴⁾ But where a patient can refuse to have the tissues removed at all, a prisoner convicted of a collection crime has no choice--the law demands that he supply a sample of DNA. The samples can be acquired legitimately for a public purpose without regard to the desires or wishes of the "donors," because there is no moral or legal obligation to honor their wishes over the benefits that DNA databanking offers society. Unless a further research use of the samples poses some risk of harm to the individuals who are compelled to donate their DNA, the usual foundation for the supposed need for consent is weakened.⁽²⁴⁵⁾ Consequently, if anonymizing the data or samples is sufficient to protect the individual sources of the DNA from stigmatization or any other harm from the research, then securing the consent of the offenders may not be essential.

If there is no legal categorical imperative that precludes the use of offender DNA samples for research, the issue remains one that legislators should resolve in light of the balance of individual and public interests. Statistical research into the distributions of biometrically useful alleles or genotypes in the convicted offender populations would seem to pose no risk to anyone and at least some benefit to forensic science. Studies of the quality of laboratory performance also can be beneficial and cannot harm individual interests. Although a state might choose to conduct such research with other DNA sources, the use of offender data or samples is directly connected to operating and administering an offender DNA databanking system, and it is not surprising that many statutes permit the use of a convenient source of DNA for this research.

Although nonoperational research lacks the same nexus to the databanking system, it is not necessarily impermissible for that reason. If medicine or science stood to make major gains by reason of these samples and if there could be no harm to the individuals from whom the samples originated, then the case for a broad research-use provision would be clear. But neither any great public gain nor any grave private loss is apparent. Tissue samples linked to comprehensive health records are valuable for biomedical research; DNA samples from convicted offenders are of little importance in themselves. It is possible that some researchers would be interested in acquiring the samples to trawl for associations between offenses and alleles. But the utility of convicted offender samples for such research in behavioral genetics is unclear for two major reasons. First, unlike most diseases, the behaviors associated with these samples are not well-defined. Certainly, the notion that there is a gene or set of genes for "crime" seems naive at best. Researchers in this field speak of genetic predispositions toward broadly defined behaviors such as "novelty seeking," which "can be expressed in many different ways" and are not strictly determined by a single gene or even a set of interacting genes.⁽²⁴⁶⁾ Second, even if specific categories, such as pedophilia, might merit investigation, offender databanks, which merely contain samples from unrelated individuals, would be of limited value in genetic linkage studies.⁽²⁴⁷⁾ Thus, the scientific case for access to the samples often will be weak.

On the other hand, to dismiss all possible research that might benefit from these samples as scientifically worthless would be dogmatic or doctrinaire. Research that can advance knowledge should be encouraged unless it poses actual dangers to the research subjects or to the public. Public concern with studies of the genetics of behavior--particularly criminal behavior--is acute. There is palpable fear that data will be misused, that tentative findings will be misrepresented as certain knowledge, and that weak correlations will be transformed into powerful predictions that cannot bear the weight that will be placed upon them. The risk of misuse of knowledge is not fanciful.⁽²⁴⁸⁾ Inasmuch as the lay perception of the extent to which we are puppets of our genes far outstrips any scientific reality, research findings about genetic influences on behavior could assume a social importance that they do not merit.⁽²⁴⁹⁾

One response is to avoid knowledge that might be misused. But another is to encourage the scientific community to pursue all methodologically sound research into reasonable hypotheses and to rely on public discourse and debate to place the research results in their proper context. The argument that research into the bases of behavior should be impeded because findings about genes might be misused comes perilously close to a policy of censorship of unwanted discoveries of genetic components of complex human behavior.

Of course, not every possible research use deserves accommodation, and legislatures that choose to allow nonoperational research uses would do well to consider procedures for ensuring that the proposed research respect the rights of the subjects and the interests of the public. For example, one might require that no DNA samples or data be provided to genetics researchers unless their research has been approved by an institutional review board charged with assessing its scientific merit and its risks to human subjects.⁽²⁵⁰⁾ Since no existing, let alone immutable, rule of law altogether precludes this type of research, such matters have to be resolved as policy choices by American society.

Delineating the permissible uses for the DNA used to create offender databanks is one thing; ensuring that the actual uses will correspond to the permitted ones is another. Both public policy and constitutional doctrine militate in favor of meaningful protections against unauthorized uses or disclosures of samples or records, and most statutes do strive to curtail dissemination and use.⁽²⁵¹⁾ Although even the best security systems and the threat of criminal penalties cannot assure that misuse never will occur,⁽²⁵²⁾ the level of security is considerably higher than that sometimes surrounding DNA samples in the private sector.

Combined with a physically secure system, most statutory protections probably satisfy constitutional requirements. In Whalen v. Roe,⁽²⁵³⁾ the Supreme Court rebuffed a constitutional attack on a database containing information that was indisputably "personal in character and potentially embarrassing or harmful."⁽²⁵⁴⁾ The database, maintained by the New York Department of Health, consisted of copies of all prescriptions for certain dangerous drugs. The medical information, including the name and address of each patient, was entered into a computerized data base. The Court deemed the combination of (1) a "statutory or regulatory duty to avoid unwarranted disclosures,"⁽²⁵⁵⁾ (2) physical measures to ensure security, and (3) a history of operation that had not been marked by breaches of confidentiality, adequate to satisfy the interest of patients in the privacy of their prescriptions.⁽²⁵⁶⁾ No less should be required of a government databank of DNA samples. With sufficient safeguards--but not without them--a system for collecting DNA on arrest, analyzing it for appropriate genotypes, and storing those data for law enforcement purposes should be constitutional.⁽²⁵⁷⁾

Security beyond the constitutional minimum could be provided, however, by departing from the current procedure of retaining most samples indefinitely and by insisting that authorities analyze only those genotypes that have no more social significance than other identifying features such as skin color, eye color, fingerprint patterns, and blood and tissue types.⁽²⁵⁸⁾ These genotypes are far less sensitive or revealing than the vast array of non-genetic information that is the traditional subject of privacy protection.⁽²⁵⁹⁾ With such genotypes, the most effective form of protection against unauthorized disclosure of samples would be the automatic destruction of the samples once the identifying alleles are recorded. The result would be a database of computer-searchable--but socially trivial--numerically encoded genotypic identifiers. DNA databanking with personal identifiers would not be practiced, although anonymized samples might be retained for quality control or research purposes. This, in turn, would help put to rest any fears that the government would access the DNA samples to discover disease-related information.⁽²⁶⁰⁾

An earlier report to the working group analyzed the constitutional protections implicated by laws requiring individuals to submit to DNA sampling at the time of an arrest. It concluded that a minimally invasive method of sampling combined with a very secure system of storing identifying information probably is constitutional. However, it did not discuss the possible need to expunge records of individuals who are not convicted of the crimes for which they are arrested. There are many reasons that a conviction might not be obtained, ranging from an unlawful arrest⁽²⁶¹⁾ to a verdict of not guilty.⁽²⁶²⁾ A state might well choose to remove the DNA sample from its databank as well as the record of the individual's alleles;⁽²⁶³⁾ or it might provide for the destruction of these materials if requested by the individual. But must it provide for such expungement?

To put this issue in perspective , we should distinguish both a different issue and a different body of law. To begin with, we must differentiate an action for expungement from a case in which the prosecution brings a citizen to trial on the charge that was the subject of the arrest and proffers an analysis of the DNA sample as evidence. In this case, the government is the moving party. The government is asking the court to receive the DNA analysis into evidence. If the arrest was unlawful, the DNA sample is inadmissible, derivative evidence. In Wong Sun v. United States,⁽²⁶⁴⁾ the Supreme Court announced that the Fourth Amendment exclusionary rule requires the suppression not only of evidence of "the primary illegality" but also of evidence "come at by exploitation of that illegality."⁽²⁶⁵⁾ When the government is the moving party and seeks the introduction of the DNA analysis, the citizen has a full-fledged federal constitutional right to have the derivative evidence suppressed. In contrast, in the fact situation now of interest, the citizen sues for the return or destruction of the DNA sample taken and included in the databank.

Moreover, we must distinguish the federal law from the corresponding body of state law. Even if the arrest is perfectly lawful, a state legislature or court has the power to authorize or require expunction of the arrest records, possibly going to the length of destroying a DNA sample taken on arrest.⁽²⁶⁶⁾ A state could do so pursuant to its own constitution, statute,⁽²⁶⁷⁾ or a court's inherent equitable power.⁽²⁶⁸⁾ When the courts resort to this equitable power, they typically balance the competing interests to determine whether to decree expungement. A court may weigh the likelihood that the retention of the arrest record will disadvantage the citizen in seeking employment or obtaining professional licensing against the extent to which record might enhance the ability of the police to identify persons with a criminal propensity.⁽²⁶⁹⁾

The states vary widely in their approach to the question of when a citizen is entitled to have his or her arrest records expunged. In a few jurisdictions, the courts are receptive to requests for expungement and absent a compelling reason to the contrary, will afford relief even when the prosecution has merely dropped charges.⁽²⁷⁰⁾ However, most states tend to take a more restrictive approach.⁽²⁷¹⁾ In some states, the citizen must show that substantial time has elapsed with no further arrests or prosecutions.⁽²⁷²⁾ In other jurisdictions, individuals who have been acquitted at trial are generally entitled to expungement.⁽²⁷³⁾ In still other jurisdictions the court may expunge the records only if it makes a "determination that the person arrested is factually innocent."⁽²⁷⁴⁾ In the most restrictive jurisdictions, the tendency is to grant relief only when the arrest itself was illegal or accompanied by another violation of constitutional rights.⁽²⁷⁵⁾ The states disagree not only over when relief is permissible but also over the form of the relief. Some jurisdiction merely segregate and seal "expunged" records; others physically destroy them.⁽²⁷⁶⁾ If the state grants the former type of relief, it might still be possible for the police, prosecutors, or courts to obtain access to the records for purposes of an investigation or prosecution.⁽²⁷⁷⁾

The issue of interest in this report, though, is whether the federal constitution would warrant granting relief to a citizen seeking expungement.⁽²⁷⁸⁾ In Davidson v. Dill,⁽²⁷⁹⁾ the Colorado Supreme Court discerned a limited constitutional right to expungement. The citizen in question had been fingerprinted and photographed on arrest. After she was acquitted, she filed suit for an order expunging her arrest records and, in the alternative, compelling their return to her. The lower court dismissed the complaint for failure to state a claim on which relief can be granted. However, the appellate court reversed. Citing Griswold v. Connecticut,⁽²⁸⁰⁾ the state supreme court invoked a "fundamental" constitutional "right of privacy."⁽²⁸¹⁾ The court declared:

We have now reached the point where our experience demands the existence of a right of privacy in the fingerprints and photographs of an accused who has been acquitted, to be at least placed in the balance, against the claim of the state for a need for their retention.⁽²⁸²⁾

However, Dill has little precedential value today. The decision, rendered in 1972, antedated the United States Supreme Court's 1976 decision in Paul v. Davis.⁽²⁸³⁾ In this case, Davis was arrested for shoplifting, but he was never brought to trial on the charge. When he was arrested, his photograph was taken. Davis alleged that the police later included the photograph in a flyer that purported to identify "active shoplifters" in the metropolitan area. Davis further alleged that the police distributed the flyer to merchants throughout the area. Davis sued the police for defamation and violation of his federal civil rights.

The Supreme Court held that the police had not deprived Davis of any constitutional right. The majority opinion treats that the right of privacy conferred by the Due Process Clause as limited to a right to autonomy in certain areas:

[O]ur "right to privacy" cases, while defying categorical description, deal generally with substantive aspects of the Fourteenth Amendment. In Roe [v. Wade, 410 U.S. 113 (1973)], the Court pointed out that the personal rights found in this guarantee of pesonal privacy must be limited to those which are "fundamental" or "implicit in the concept of ordered liberty" . . . . The activities detailed as being within this definition were ones very different from that for which respondent claims constitutional protection--matters relating to marriage, procreation, contraception, family relationships, and child rearing and education. Respondent's claim is far afield from this line of decisions. He claims constitutional protection against the disclosure of the fact of his arrest upon a shoplifting charge. None of our substantive privacy decisions hold this or anything like this, and we decline to enlarge them in this manner.⁽²⁸⁴⁾

It would be possible to limit Paul to its facts. The Paul Court noted that the case did not involve "an unreasonable [and unconstitutional] search."⁽²⁸⁵⁾ However, the language of the opinion is so sweeping and forceful that the lower federal courts have concluded that there is no constitutional privacy right to have arrest records expunged or destroyed.⁽²⁸⁶⁾ Instead, in expungement cases the federal courts rely on their inherent equity power.⁽²⁸⁷⁾ Some insist that such requests must be analyzed on a case-by-case basis.⁽²⁸⁸⁾ Other federal courts, though, have developed a general list of the situations in which they will grant expungement relief:

These lists typically include situations where the arrest was unlawful, where mass arrests rendered judicial determination of probable cause impossible, [and] where the arrest, although proper, was based upon a statute that was later declared unconstitutional . . . .⁽²⁸⁹⁾

Whatever approach the court takes, relief is usually granted only in "extreme circumstances."⁽²⁹⁰⁾ The federal courts ordinarily find such circumstances only when the arrest itself was illegal or accompanied by a constitutional violation.⁽²⁹¹⁾ Thus, in all probability a federal court would not brand a policy of retaining DNA information to be unconstitutional on its face,⁽²⁹²⁾ and it would order the destruction or return of a DNA sample taken on arrest only if the court concluded that the arrest itself was unconstitutional.⁽²⁹³⁾

Thus far, we have considered DNA databanks and databases that are designed to include only those individuals who, by reason of other contact with the criminal justice system, seem more likely than the typical citizen to be linkable through a database to past or future crimes. Such systems are "exclusive" in the sense that they seek to exclude most of the population, reserving places in the database for the individuals of most interest to the criminal justice system. A radically different approach is conceivable. "Inclusive databanking" would seek to include more--perhaps much more--of the population, without regard to the risk that any given individual ever will commit a crime. This section discusses certain legal and policy aspects of inclusive databanking.

Part I examined the Fourth Amendment law applicable to acquiring a DNA sample of a particular suspect from a private repository of tissue samples. Here, we consider the possibility of acquiring the entire private databank to expand the jurisdiction's law enforcement database.⁽²⁹⁴⁾ Would acquiring the samples for testing invade the Fourth Amendment rights of the individuals whose samples are in the private database?

For the reasons indicated in Part I, it is likely that the courts would refuse to find a constitutionally protected interest in the database samples. As we saw in that section, United States v. Miller⁽²⁹⁵⁾ holds that a person lacks a "legitimate expectation of privacy"⁽²⁹⁶⁾ in commercial, financial records provided to a bank. The expectation is diminished because the information "conveyed to the banks" is "exposed to their employees in the ordinary course of business."⁽²⁹⁷⁾ The same reasoning applies to most subpoenaed medical or laboratory samples.⁽²⁹⁸⁾ Typically, a patient or research subject who provides tissue samples agrees not only that biomedical workers may handle and analyze these samples for the patient's own diagnosis or treatment, but also for many other uses, giving access to the samples to a still larger circle of individuals.

A system that placed the identifying DNA genotypes of all newborns, immigrants, and visitors to the United States in a central location could serve to identify virtually anyone whose DNA is deposited at the scene of a crime anywhere in the United States. The case in favor of such a comprehensive DNA database for identification is straightforward.⁽²⁹⁹⁾ First, the deterrent effect of DNA databanking is greatest for a universal database. Convicted-offender databases can deter only those offenders who have been caught and convicted for previous crimes.⁽³⁰⁰⁾ By increasing the probability of detection of new and old offenders alike, a comprehensive database can do much more to reduce the rate of certain crimes. And, making apprehension more certain permits the same level of deterrence with less Draconian (and costly) periods of imprisonment.⁽³⁰¹⁾

Second, a comprehensive database avoids many problems or issues associated with offender or arrestee databanking. It obviates the need to draw some line between those offenses for which databanking is permitted and those for which it is not. It avoids any risk that police will make pretextual arrests merely to secure DNA samples. It makes it unnecessary to infer physical traits or racial or ethnic identity from trace evidence samples. Perhaps most important, it avoids stigmatizing any person or group. A comprehensive database imposes the same obligation on all racial and ethnic groups. There is a widespread perception that minorities are overrepresented in the criminal justice system in part because they are wrongfully arrested and convicted to a greater degree than whites. A universal database would help prevent wrongful convictions and arrests of minorities. When an eyewitness mistakenly concludes that the criminal was a minority member, a wrongful arrest (and conviction) can ensue. A universal, readily accessible national database would aid in preventing such arrests in subsequent miscarriages of justice.⁽³⁰²⁾

Third, a single, national database would be more efficient than a system of over 50 separate databases of offenders or suspected offenders. From this perspective, the current system of multiple, overlapping databases represents unnecessary duplication and a waste of scarce resources. For all these reasons, a single, secure, national DNA identification database is attractive.

To a greater extent than any other step the federal government could take, however, the creation of such a database raises the spectre of Big Government.⁽³⁰³⁾ It would depart from a tradition that has eschewed national identity cards and compulsory fingerprinting. Numbers, photographs, and fingerprints seem tame in comparison to DNA samples, which could reveal predispositions to certain diseases. In short, even if comprehensive DNA databanking for law enforcement use were economically and technologically feasible--a prospect that may soon be realized--this country would hesitate before demanding its citizens to "donate" their DNA to a massive, centralized databank.

On the other hand, a single, national database might well pose less of a threat to privacy than the many databases and databanks already in existence. As noted earlier, there are already databases or databanks that include a significant percentage of the American population.⁽³⁰⁴⁾ Even if the federal government takes no additional action, the percentage will grow steadily.⁽³⁰⁵⁾ The proliferation of large databases with varying degrees of security means that someone intent on invading a given person's genetic privacy might have a choice of databases to target for penetration. If a single, national database could replace most of these diverse repositories, it could reduce the threat to privacy.⁽³⁰⁶⁾

While the balance of advantage seems to favor creating a comprehensive database, it is important to note the relationship between this issue and the question of sample retention. The argument for a comprehensive database is at its apogee when only a database is contemplated--samples are taken, genetic data is extracted and recorded for the database, but the samples themselves are discarded. Destruction of the samples minimizes the danger that, for example, unauthorized persons (such as employers or insurers) will access the sample and extract other, medically relevant information for improper reasons. That danger is greatest when the proposal takes the form of a recommendation to retain samples and maintain a databank rather than a mere database. Indeed, a database derived from markers that have no interest except for identification seems no more dangerous than traditional fingerprint databases.

Yet, there is one unavoidable difference. Even the non-coding STR loci currently in place for CODIS could provide information about family relationships. With a comprehensive database, the alleles of all members of every family are on record. Officials could compare the alleles of family members to ascertain genetic relationships. For instance, comparing the alleles of the parents to a child (as indicated on a birth certificate) could indicate that the child is the offspring of the mother but not the father.⁽³⁰⁷⁾ Revelations that a child is illegitimate, that siblings are only half-siblings, and so on, could well be highly distressing to family members. For this reason, a comprehensive database, although less worrisome than a repository of samples, implicates meaningful privacy interests. This does not necessarily mean that such a database should not be established, but it does underscore the importance of protecting security and confidentiality in the genetic records.

1. About a third of those named as the primary suspect in rape cases are excluded. FBI, The Application of DNA Testing to Solve Violent Crimes (1993).

2. Legal issues associated with collecting DNA from individuals who are not specifically suspected of having committed the crime that is under investigation arise with regard to DNA databases. Issues involved in building and administering DNA databases for law enforcement are treated in Part III.

3. Some other possible inferences are that the match is a laboratory artifact, that the match is coincidental in that an unrelated person is the source of the trace evidence, or that the match is the result of kinship in that a close relative of the defendant is the source. Suitable investigative and testing procedures can serve to eliminate such alternative hypotheses. See David Kaye & George Sensabaugh, Reference Guide to DNA Evidence, in Reference Manual on Scientific Evidence (Federal Judicial Center, 2d ed. forthcoming 2000).

4. Of course, the police also must obtain the trace evidence legally. The Fourth Amendment, for example, imposes strict limits on entering private residences to search for evidence. However, DNA evidence raises no special difficulties or considerations in this regard.

5. As explained in a separate report, the Self-Incrimination Clause of the Fifth Amendment and the Due Process Clauses of the Fifth and Fourteenth Amendments have little force in this context. See D.H. Kaye, The Constitutionality of DNA Sampling on Arrest (1999).

6. Some state constitutions provide enhanced protection from searches and seizures and other police practices.

8. See, e.g., People v. Marshall, 244 N.W.2d 451, 457 (Mich. Ct. App. 1976); In re J.W.K., 583 N.W.2d 752, 755 (Minn. 1998); Commonwealth v. Riedel, 651 A.2d 135, 139 (Pa. 1994); State v. Evans, 338 N.W.2d 788, 794 (Neb. 1983); In re Death of Abe A., 437 N.E.2d 265, 266 (N.Y. 1982).

10. A valid arrest requires probable cause to believe that the individual has committed an offense.

12. But see Harold Krent, Of Diaries and Data Banks: Use Restrictions Under the Fourth Amendment, 74 Tex. L. Rev. 49 (1995) (arguing that the current understanding of the Fourth Amendment should be altered so that "each government use of information about an individual constitutes a separate seizure of that person's effects" and hence "all such uses must satisfy the reasonableness requirement," id. at 53 n.25). A narrower version of this view that might prove more persuasive holds that when the police rely on a special justification to deviate from the normal

Fourth Amendment requirements such as probable cause and warrant, their utilization of the evidence seized must be limited to uses that promote that special justification. Thus, if as in Schmerber v. California, 384 U.S. 757 (1966), the police justify a warrantless seizure of evidence on an exigency theory, they could not use the evidence seized for any non-exigent purposes.

The conventional view that evidence legitimately acquired for one purpose can be used for another has implications for DNA samples that are obtained with probable cause and a warrant. Authorities might wish to compare the identifying profile to those on file in a database derived from unrelated, unsolved crime-scene stains. Because the invasion of bodily integrity and privacy in the information contained in the suspect's DNA was justified by probable cause and a warrant for the seizure of the DNA and the search of its structure, the conventional additional-use theory would allow the further comparisons. An analogy can be drawn to the situation in which police searching a dwelling for specific stolen items record the serial number of an item not enumerated in the warrant and check this number against a list of serial numbers of other stolen items. Recording the serial number would not be considered a search if the number was in plain view. Cf. Arizona v. Hicks, 480 U.S. 321, 324-25 (1987) ("the mere recording of the serial numbers did not constitute a seizure," but " taking action, unrelated to the objectives of the authorized intrusion, which exposed to view concealed portions of the apartment or its contents, did produce a new invasion of respondent's privacy unjustified by the exigent circumstance that validated the entry"). Under the additional-use theory, checking the number against the list is not an additional search of the person or his property and therefore would be allowed. Cf. State v. Wamre, 599 N.W.2d 268 (N.D. 1999) (holding that the Fourth Amendment was satisfied when the police used a serial number that was in plain view during a search to secure a warrant by telephone that allowed them to seize the additional items).

13. In Schmerber v. California, 384 U.S. 757 (1966), the Supreme Court sustained the warrantless acquisition of a blood sample for intoxication testing. The Court reasoned that there was an exigency because blood alcohol concentrations decline rapidly. Id. at 771-72. However, DNA cases are readily distinguishable. In the typical case, the police will desire a DNA sample to test for permanent, identifying markers--markers that will not evaporate or disappear with the mere passage of time. Consequently, it will be more difficult for the authorities to justify a warrantless seizure of a DNA sample than it would be to justify a similar acquisition of a blood sample for intoxication testing. In re J.W.K., 583 N.W.2d 752, 757 (Minn. 1998). In the absence of exigent circumstances, the police would be obliged to obtain the functional equivalent of a warrant, that is, a court order that the suspect furnish a DNA sample. Thurman v. State, 861 S.W.2d 96, 100 (Tex. Ct. App. 1993).

19. Jerold Israel, Legislative Regulations of Searches and Seizures, 73 Mich. L. Rev. 222, 238-41 (1974); P. Michael Drake, Comment, Detention for Taking Physical Evidence Without Probable Cause, 14 Ariz. L. Rev. 132 (1972); Mark P. Asselta, Comment, The Constitutionality of Compulsory Identification Procedures on Less Than Probable Cause: Reassessing the Davis Dictum, 89 Dick. L. Rev. 501 (1985); Note, Detention to Obtain Physical Evidence Without Probable Cause: Proposed Rule 41.1 of the Federal Rules of Criminal Procedure, 72 Colum. L. Rev. 712 (1972). One might argue that blood, urine, or hair samples should be treated differently because they have the potential to reveal information that is more significant than the pattern of whorls and ridges that constitute a fingerprint. That this difference requires a different outcome is not obvious. All the factors listed in the Davis dictum apply to these other samples. But regardless of how this broad issue might be resolved, it is sufficient to note here that for the specific practice of DNA typing conducted at loci that are of strictly biometric interest, the analogy is compelling.

21. The Arizona statute specifies that "'identifying physical characteristics' includes, but is not limited to, the fingerprints, palm prints, footprints, measurements, handwriting, handprinting, sound of voice, blood samples, urine samples, saliva samples, hair samples, comparative personal appearance or photographs of an individual." Id. at § 13-3905(G).

22. Clyde M. Tande, Note, DNA Typing: A New Investigatory Tool, 1989 Duke L.J. 474. For example, Alaska Rule of Criminal Procedure 16(c) allows a court to order detention to "[p]ermit the taking of samples of blood, hair and other materials of the person's body which involve no unreasonable intrusion thereof" on the basis of an affidavit or testimony establishing "probable cause to believe that: (i) An offense has been committed by one of several persons comprising a narrow focal group that includes the subject person; (ii) The evidence sought may be of material aid in identifying who committed the offense; and (iii) The evidence sought cannot practicably be obtained from other sources." Alas. R. Crim. P. 16(c)(2)(vii) & 16(c)(1).

25. There is sufficient state action to trigger the Fourth Amendment if "the drawing of blood is instigated by the government," State v. Hardy, 963 S.W.2d 516, 523 (Tex. Crim. App. 1997), by, for instance, requesting that an emergency room doctor take a sample. Commonwealth v. Kohl, 615 A.2d 308, 310 (Pa. 1992). When the authorities intervene even before the private entity take a sample and request a sample, the otherwise private entity is acting as a government agent in seizing the sample.

26. See The Constitutionality of DNA Sampling on Arrest, supra note 5. New York City police used this tactic to acquire DNA from a suspected serial killer and rapist. The Crier Report: Mandatory DNA Testing (Fox television broadcast, Mar. 11, 1999), available in 1999 WL 18330169.

27. 1 National Bioethics Advisory Commission, Research Involving Human Biological Materials: Ethical Issues and Policy Guidance 13 (1999) [hereinafter NBAC].

28. See id. at 13-15; Lawrence O. Gostin, Health Information Privacy, 80 Cornell L. Rev. 451, 464 & 467-68 (1995); M. Therese Lysaught et al., A Pilot Test of DNA-based Analysis Using Anonymized Newborn Screening Cards in Iowa, in Stored Tissue Samples: Ethical, Legal, and Public Policy Considerations 3, 17 (Robert F. Weir ed., 1998); Jean E. McEwen, DNA Databanks, in Genetic Secrets: Protecting Privacy and Confidentiality in the Genetic Age 231 (Mark Rothstein ed., 1997).

29. When the police obtain a sample directly from a private citizen, there is undeniably sufficient state action to bring the Fourth Amendment into play. In re J.W.K., 583 N.W.2d 752 (Minn. 1998); State v. Binner, 886 P.2d 1056 (Or. Ct. App. 1994). There is sufficient state action whether the motivation of the police is to obtain the sample in order to add the sample to a database or to acquire an evidential sample to be compared to a database.

30. United States v. Pervaz, 118 F.3d 1, 6 (1st Cir. 1997); United States v. Reed, 15 F.3d 928, 931 (9th Cir. 1994); State v. Grant, 620 N.E.2d 50, 60 (Ohio 1993); State v. Maxfield, 886 P.2d 123, 127 (Wash. Ct. App. 1994), rev'd on other grounds, 945 P.2d 196 (Wash. 1997).

32. Id. at 113-14 (1984) (quoting Walter v. United States, 447 U.S. 649, 662 (1980)). See also Coolidge v. New Hampshire, 403 U.S. 443, 488 (1971); Burdeau v. McDowell, 256 U.S. 465 (1921); Tims v. State, 711 So.2d 1118, 1122 n. 2 (Ala. Crim. App. 1997) ("no state involvement"); People v. Perlos, 462 N.W.2d 310, 315 (Mich. 1990) ("no state involvement"); State v. Nelson, 941 P.2d 441, 445 (Mont. 1997) ("no State involvement"); State v. Enoch, 536 P.2d 460, 461 (Or. Ct. App. 1975) ("if an independent private citizen finds evidence and turns it over to the police, the evidence is legitimate"); State v. Guido, 698 A.2d 729, 733 (R.I. 1997) ("no state action").

33. If the medical provider or researcher were uncooperative, the authorities could resort to normal compulsory process such as a subpoena. See, e.g., United States v. Dionisio, 410 U.S. 1, 10-11 (1973) (stating that "a grand jury subpoena to testify is not that kind of governmental intrusion on privacy against which the Fourth Amendment affords protection once the Fifth Amendment is satisfied," but recognizing that "[t]he Fourth Amendment provides protection against a grand jury subpoena duces tecum too sweeping in its terms "to be regarded as reasonable.'"); State v. Fears, 659 S.W.2d 370 (Tenn. Crim. App. 1983) (requiring only a showing that the object of the subpoena is logically relevant to the subject-matter of a legitimate criminal investigation).

34. It will be fulfilled if, prior to the government's request, the private parties acted on their "own initiative" and out of an "independent" medical or research "motivation." State v. Comeaux, 818 S.W.2d 46, 50 (Tex. Crim. App. 1991).

47. E.g., Tims v. State, 711 So. 2d 1118, 1123-24 (Ala. Crim. App. 1997); People v. Perlos, 462 N.W.2d 310, 316 (Mich. 1990); State v. Guido, 698 A.2d 729, 733 (R.I. 1997); State v. Fears, 659 S.W.2d 370, 375-76 (Tenn. Crim. App. 1983); Thurman v. State, 861 S.W.2d 96, 98 (Tex. Ct. App. 1993); see Robert M. Gellman, Prescribing Privacy: The Uncertain Role of the Physician in the Protection of Patient Privacy, 62 N.C. L. Rev. 255, 290-91 (1984) ("As a practical matter, in the absence of a statute or a definitive court decision, the Miller decision is effectively being applied when medical records are subpoenaed"). However, few cases dealing with the specific question of obtaining DNA samples from private entities. See In re J.W.K., 583 N.W.2d 752 (Minn. 1998). The vast majority concern blood samples used for intoxication testing. E.g., People v. Dolan, 408 N.Y.S.2d 249 (Sup. Ct. 1978); State v. Enoch, 536 P.2d 460 (Or. Ct. App. 1975); Commonwealth v. Kohl, 615 A.2d 308 (Pa. 1992); State v. Guido, 698 A.2d 729 (R.I. 1997); State v. Comeaux, 818 S.W.2d 46 (Tex. Crim. App. 1991). To an extent, that area of law is sui generis. In the United States, drunk driving exacts a huge toll in death and economic loss. For that reason, every state has enacted an implied consent law. Andre A. Moenssens et al., Scientific Evidence in Civil and Criminal Cases Ch. 3 (4th ed. 1995). By the terms of the implied consent statute, when a citizen obtains a driver's license, he or she impliedly consents to later tests of his or her blood, breath, or urine alcohol concentration. The existence of these statutes in every jurisdiction reflects legislative recognition of an important public interest in the prevention of drunk driving, and the provisions of the statutes also weaken the argument that the citizen has a protected expectation of privacy in either the sample or a medical record reflecting the results of a test of the sample. In the drunk driving area, the implied consent laws enable the government to contend both that the statutes reduce the citizen's privacy expectations and that any expectations are ones that society is unprepared to recognize as reasonable.

49. People v. Perlos, 462 N.W.2d 310, 324 (Mich. 1990) (Levin, J., dissenting) ("A person's medical records are an intensely personal matter. Few persons would willingly share their medical records with the state"); State v. Binner, 886 P.2d 1056, 1059 (Or. Ct. App. 1994) ("Such testing could reveal the most personal of the medical details of our private lives that would not be known to the public in general"); Peter H. W. van der Goes, Jr., Comment, Opportunity Lost: Why and How to Improve the HHS-Proposed Legislation Governing Law Enforcement Access to Medical Records, 147 U. Pa. L. Rev. 1009, 1037-40 (1999).

50. Comment, supra note 49, at 1011 n.10 ("A 1996 CNN/Time poll indicates that 87% of Americans want to be asked permission every time their medical records are accessed for any reason").

52. State v. Fears, 659 S.W.2d 370, 376 (Tenn. Crim. App. 1983) (after discussing Miller, the court asserts that "the same principle applies to medical records as to bank records and that the above holding of the Miller case is applicable").

53. State v. Hardy, 963 S.W.2d 516, 525 (Tex. Crim. App. 1997) ("most states"). A minority of jurisdictions do not recognize the privilege at all. Tims v. State, 711 So. 2d 1118, 1122 (Ala. Crim. App. 1997); State v. Fears, 659 S.W.2d 370, 375 (Tenn. Crim. App. 1983).

54. Hale v. Superior Court (DeFelice), 34 Cal. Rptr. 2d 279 (Ct. App. 1994); People v. Maltbia, 653 N.E.2d 402 (Ill. Ct. App. 1995); State v. Comeaux, 818 S.W.2d 46, 54-56 (Tex. Crim. App. 1991) (Campbell, J., concurring); Cal. Evid. Code § 992 (1995); Christopher B. Mueller & Laird C. Kirkpatrick, Modern Evidence: Doctrine and Practice § 5.36, at 621-22 (1995); Comment, The Physician-Patient Privilege, 56 Nw. U. L. Rev. 263 (1961).

55. Cal. Evid. Code § 998 (1995); Green v. Cooper Medical Hosp., 968 F. Supp. 249, 252 (E.D. Pa. 1997) ("Under Pennsylvania law, the physician-patient privilege does not extend to criminal matters"), aff'd, 151 F.3d 1025 (3d Cir. 1998); Alexander v. Commonwealth, 708 A.2d 1251 (Pa. 1998); State v. Hardy, 963 S.W.2d 516 (Tex. Crim. App. 1997).

57. Id. at 100 ("Society can afford the physician-patient privilege in certain civil cases to protect personal privacy, but the need to protect the public from crime requires disclosure of the same information in criminal cases.").

59. State v. Fears, 659 S.W.2d 370, 376 (Tenn. Crim. App. 1983); Gellman, supra note 47, at 274.

60. People v. Perlos, 462 N.W.2d 310 (Mich. 1990) (discussing Mich. Comp. L. Ann. § 257.625(a)(9)).

61. State v. Hardy, 963 S.W.2d 516, 524 (Tex. Crim. App. 1997). In other contexts, the Supreme Court has treated widespread state legislative patterns as persuasive evidence of societal expectations. E.g., Jaffee v. Redmond, 518 U.S. 1 (1996) (the Court was impressed by the number of states that had codified a psychotherapist privilege).

62. People v. Dolan, 408 N.Y.S.2d 249, 252 (Sup. Ct. 1978). The usual view is that body parts and tissues are not property, but that they are subject to the legal protections afforded to a person's body. See Moore v. Regents of the Univ. of California, 793 P.2d 479 (Cal. 1990) (holding, in response to a claim for conversion, that a commercially valuable cell line is not the property of the patient from which the original cells were taken); Moe M. Litman, The Legal Status of Genetic Material, in Human DNA: Law and Policy, International and Comparative Perspectives 17, 25-27 (Bartha Maria Knoppers ed., 1997) (reviewing cases and advocating a sui generis classification in which genetic material is regarded as a hybrid of private property, common property, person, and information).

65. People v. Perlos, 462 N.W.2d 310, 317 (Mich. 1990) (stating, immediately after discussing Miller, that this line of authority deals with "privacy rights in medical records [and] blood samples").

69. Id. at 324 (Levin, J., dissenting); see also State v. Hardy, 963 S.W.2d 516, 524 (Tex. Crim. App. 1997) ("release of medical information to hospitals is less optional than the release of financial information to banks. A person can choose not to maintain a bank account, but it hardly seems reasonable to expect someone to forego medical attention").

71. If the person in need of emergency medical treatment is unconscious, the hospital will still administer treatment on the legal theory of implied intent: The hospital personnel may ordinarily assume that if the person were conscious and understood his condition, he would want to be treated. See People v. Woodson, 630 N.Y.S.2d 670 (Sup. Ct. 1995) (when he was originally admitted to the hospital, the defendant was unconscious and comatose); State v. Guido, 698 A.2d 729 (R.I. 1997) (the defendant was unconscious when he was found, and the hospital staff withdrew a blood sample during emergency room treatment); State v. Hardy, 963 S.W.2d 516, 526-27 (Tex. Crim. App. 1997) ("Many statutes also permit officers to conduct a chemical test (without a warrant) on an unconscious person . . . . Some statutes even permit obtaining a sample for chemical testing (without a warrant) from an unconscious person even when that person is not under arrest"); State v. Jenkins, 259 N.W.2d 109 (Wis. 1977) (the defendant was in and out of consciousness after an accident).

74. See State v. Hardy, 963 S.W.2d 516, 525 (Tex. Crim. App. 1997) ("No consensus exists in courts decisions on whether an expectation of privacy exists in medical records"), id. at 524 (after acknowledging a division of sentiment in the Texas cases, the court states that "[f]our out of five [Texas courts] held that society does not recognize as reasonable an expectation of privacy in those kinds of records"); State v. Nelson, 941 P.2d 441, 448 (Mont. 1997) ("medical records have not been historically protected by the Fourth Amendment's prohibition against unreasonable searches and seizures"); 1 Wayne R. LaFave, Search and Seizure: A Treatise on the Fourth Amendment § 2.7(d), at 640 (3d ed. 1996).

The minority view that the suspect retains a constitutionally protected privacy expectation in the sample would not necessarily prevent the state from obtaining the medical information or sample. It usually would have to establish probable cause or founded suspicion to obtain a court order that the hospital or laboratory deliver the record or the sample.

75. For example, in a People v. Woodson, 630 N.Y.S.2d 670 (Sup. Ct. 1995), an unconscious patient was taken to a private hospital. The hospital personnel prepared vials of blood and urine specimens. The grand jury subpoenaed the samples, the hospital turned over both vials, and the authorities conducted toxicological tests. The defendant argued that the authorities' acquisition of the two vials violated the Fourth Amendment and moved to suppress the vials and testimony as to their contents. A New York Supreme Court granted the motion. This trial court declared that "[b]y now, it is settled law that a hospitalized defendant retains a Fourth Amendment right in blood and urine samples." Id. at 671. The court also ruled that the suspect had an objectively reasonable expectation of privacy in his hospital records. However, the reasoning in Woodson is dubious. The judge cited People v. Natal, 553 N.E.2d 239 (N.Y. 1990) to support the exclusion of the evidence. The rub is that Natal has nothing to do with blood or urine samples. Natal involved clothing and personal effects taken from a defendant when he was placed in pretrial confinement. Worse still, although the Natal court granted relief on the ground that the district attorney had abused the subpoena power, the court ruled that "inspection of personal effects previously exposed to police view . . . does not invade any substantial privacy interest." Id. at 241. Moreover, the Woodson court does not so much as mention, much less evaluate, the possibility that Miller is pertinent. Similarly ignoring Miller, the Pennsylvania Supreme Court recognized a privacy interest in medical records in Commonwealth v. Riedel, 651 A.2d 135 (Pa. 1994).

77. The courts have employed a variety of rationales to justify this view. Some reason that the suspect had bailed the sample with the private hospital and that as bailee, the hospital "may consent to . . . permit the warrantless search and seizure." People v. Dolan, 408 N.Y.S.2d 249, 252 (Sup. Ct. 1978). Other courts simply assert that if a private hospital properly obtains a sample for medical reasons, as "an independent private citizen" it may "turn[] it over to the police." State v. Enoch, 536 P.2d 460, 461 (Or. Ct. App. 1975). But most courts invoke Miller. E.g., Thurman v. State, 861 S.W.2d 96, 98-101 (Tex. Ct. App. 1993) (although the case involved a subpoena for medical records, the court used broad language, expansive enough to apply to samples as well, and noted that "[c]ourts in other states have reached the same conclusion.").

78. See State v. Gonzalez, 852 P.2d 851 (Or. Ct. App. 1993); State v. Fears, 659 S.W.2d 370 (Tenn. Crim. App. 1983); Conrad v. Texas, 1997 WL 764527, 1997 Tex. Ct. App. LEXIS 6330, cert. denied, 119 S.Ct. 1462 (1999); Thurman v. State, 861 S.W.2d 96 (Tex. Ct. App. 1993). Many of these courts not only cite Miller but also treat it as dispositive. See People v. Perlos, 462 N.W.2d 310, 316-17 (Mich. 1990); State v. Fears, supra, at 375-76; Gellman, supra note 47, at 290-91. In State v. Nelson, 941 P.2d 441 (Mont. 1997), the Montana Supreme Court found a protected expectation of privacy in medical records under its state constitution. However, the Montana court pointed out that its cases construing its constitution protect privacy "more strict[ly] than" the cases interpreting the Fourth Amendment to the national constitution. Id. at 447. Indeed, the Montana court conceded that "medical records have not been historically protected by the Fourth Amendment's prohibition against reasonable searches and seizures." Id. at 448.

80. Ward v. Texas, 316 U.S. 547 (1942); Vernon v. Alabama, 313 U.S. 547 (1941); Lomax v. Texas, 313 U.S. 544 (1941); White v. Texas, 310 U.S. 530 (1940); Brown v. Mississippi, 297 U.S. 278 (1936); see also Colorado v. Connelly, 479 U.S. 157 (1986).

83. Under its own constitution, a state court can adopt a higher standard and require warnings about the search be administered for consent to be valid. Cooper v. California, 386 U.S. 58 (1967); Commonwealth v. Walsh, 460 A.2d 767, 771 (Pa. Super. Ct. 1983).

87. Commonwealth v. Walsh, 460 A.2d 767, 773 (Pa. Super. Ct. 1983); see also State v. Bailey, 417 A.2d 915, 919 (R.I. 1980) ("The notion of a free and voluntary consent necessarily implies that the person knows what it is he is allowing the police to do"). It has been held that it is not necessary to inform a suspect of the specific investigations for which the sample is desired. Pace v. State, No. S99P0647, 1999 WL 1087018 (Ga. Dec. 3, 1999).

88. People v. Cardenas, 604 N.E.2d 953 (Ill. Ct. App. 1992) (holding consent to be involuntary when the officer created the false impression that there could a search without the suspect's consent).

89. See, e.g., Fred Barbash, Crime-Solving by DNA Dragnet; Britain Makes Arrests in Rape Cases After Thousands of "Voluntary" Neighborhood Tests, Wash. Post, Feb, 2, 1996, at A21, available in 1996 WL 3061978; Alex Smith, Village Takes DNA Tests: Breton Men Give Saliva Samples in Hunt for Cornish Girl's Murderer, The Guardian (London), Oct. 11, 1997, available in 1998 WL 7775322.

90. DNA Evidence Brings Confession, Ariz. Republic, May 31, 1998, at A20, available in 1998 WL 7775322.

91. David Werrett, The Strategic Use of DNA Profiling, Address to the 18th International Congress on Forensic Haemogenetics, San Francisco, Cal., Aug. 19, 1999. In these investigations, explicit permission was obtained from the individuals who provided DNA samples, and the samples from people who are eliminated as possible suspects were destroyed. Id. In late 1999, however, § 64 of the Police and Criminal Evidence Act was amended to permit the retention and use of DNA samples and data with a volunteer's written permission. Home Office, Proposals for Revising Legislative Measures on Fingerprints, Footprints and DNA Samples (July 1999).

92. Indeed, in 1994, police in Ann Arbor investigating a series of rapes reportedly collected blood samples from 160 African-American men in the area. Alice Robinson, DNA of Innocent Rape Suspects Will Not be Kept: Ann Arbor Resident Filed Civil Lawsuit that Spurred Ruling, Mich. Daily Online, Nov. 21, 1997, http://www.pub.umich.edu/daily/1997/nov/11-21-97/news/news12.html, visited Dec. 11, 1998.

93. Cf. Davis v. Mississippi, 394 U.S. 721 (1969) (no suggestion that the compulsory fingerprinting of 24 young men was unconstitutional merely because there was no reason other than their fitting the very general description provided by the victim); United States v. Dionisio, 410 U.S. 1 (1973) (no infirmity in issuing, without probable cause, grand jury subpoenas to 20 men to secure voice exemplars from them).

94. Cf. Davis, 394 U.S. at 727 n.6 (1969) (referring to "the settled principle that while the police have the right to request citizens to answer voluntarily questions concerning unsolved crimes they have no right to compel them to answer").

96. It would be possible to retain the samples of volunteers for possible later use, given adequate consent. The prospects for establishing more permanent, population-wide databases are addressed infra Part III.

97. Kelly Owens & Mary-Claire King, Genomic Views of Human History, 286 Science 451 (1999); M. Klintschar et al., Is It Possible to Determine the Ethnic Origin of Caucasian Individuals Using Short Tandem Repeat Loci?, in Abstracts, Eighteenth International Congress, International Society for Forensic Haemogenetics 63 (1999) ("it is possible to differentiate between major ethnic groups using forensically relevant short tandem repeat loci").

98. See, e.g., M. McCulley et al., Genes and Faces: Classification of Midline Features, in Abstracts of the Eighteenth International Congress of the International Society for Forensic Haemogenetics 109 (1999)

99. We are grateful to James Weinstein for comments on a draft of this discussion of the equal protection clause.

100. E.g., Korematsu v. United States, 323 U.S. 214, 216 (1944). Laws or other state action that distributes burdens or benefits according to "suspect" classifications are usually, but not invariably unconstitutional. Id. (finding a compelling government interest in confining Japanese-Americans during World War II). The government bears an extremely heavy burden of showing that the classification is necessary to achieve a compelling state interest. In re Griffiths, 413 U.S. 717, 721 (1973).

101. Certain other classifications, such as sex, are "quasi-suspect." Their use can be justified by a showing of a close enough connection to an important state interest. The sex of the source of a biological sample from a crime scene can be ascertained with far greater confidence than can the race of the source. Although we shall not discuss the use of genetic tests for inferring gender in criminal investigations, the analysis of the constitutionality of testing for race is applicable to testing for sex as well.

102. See McCleskey v. Kemp, 481 U.S. 279, 292 n.8 (1987) ("It would violate the Equal Protection Clause for a State to base enforcement of its criminal laws on an unjustifiable standard such as race, religion, or other arbitrary classification.") (internal quotation marks and citation omitted).

103. Cf. Hall v. Pennsylvania State Police, 570 F.2d 86 (3d Cir. 1978) (reasoning that allegations that the state police directed banks to photograph only blacks entering the banks stated a violation of equal protection).

104. There are few cases on the constitutionality of using race as identifying feature in criminal investigations, presumably because the validity of the practice always has been taken for granted. In Hall v. Pennsylvania State Police, 570 F.2d 86 (3d Cir. 1978), the Court of Appeals for the Third Circuit held that the Fourteenth Amendment prohibited the police from asking banks to photograph suspicious-looking Blacks who entered their establishments. The constitutional infirmity in this practice, the court reasoned, was that "[t]he police simply instituted a general photographic survey limited to one race." Id. at 91. The court indicated that the result would have been otherwise had the practice applied to all races, or had the request been part of an investigation of a crime by a Black suspect: "This is not a situation where suspects are being sought on the basis of descriptions which include race as well as other physical characteristics. No crime was under investigation nor was there any information that a robbery was planned." Id.

105. Cf. Strauder v. West Virginia, 100 U.S. 303, 304 (1880) (stating that the Fourteenth Amendment confers "the right to exemption from unfriendly legislation against [a group] distinctively" on account of race). Consider also the use of racial information by government doctors on the staff of Veterans Administration hospitals. Suppose that a physician who is considering administering isoniazid to an Asian patient who is either has pulmonary sarcoidosis or pulmonary tuberculosis. See Richard I. Kopelman et al., Clinical Problem Solving: A Little Math Helps the Medicine Go Down, 341 New Eng. J. Med. 435 (1999). Although this therapy improves the chance of survival for tuberculosis (but not sarcoidosis), it can lead to hepatitis. Id. at 437. The risk of this complication is about twice as large in Asian men as in other men. Id. In weighing the therapeutic benefits against the risks it would be medically inadvisable for the physician to disregard the fact that the patient is Asian. See id. at 437-38. Legally, it is permissible for the physician consistently to use race as a factor in diagnosis, prognosis, or treatment.

108. 230 F.Supp. 156 (E.D. Va.), aff'd sub nom., Tancil v. Woolls, 379 U.S. 19 (1964) (per curiam).

[T]he designation of race, just as sex or religious denomination, may incertain records serve a useful purpose, and the procurement and compilation of such information by State authorities cannot be outlawed per se. For example, the securing and chronicling of racial data for identification or statistical use violates no constitutional privilege. If the purpose is legitimate, the reason justifiable, then no infringement results. . . .

Vital statistics, obviously, are aided by denotation in the divorce decrees of the race of the parties. This provision . . . is not objectionable in law. Of course, the advertence must be made in every case, not just in suits involving Negroes.

Id. Other laws required lists of qualified voters and property owners to be maintained so as to reveal the race of the individuals. The court struck down these requirements because "they serve no other purpose than to classify and distinguish official records on the basis of race or color." Id. at 158.

110. Where the government publicizes the race of individuals in a way calculated to disadvantage members of certain racial groups, it offends the Fourteenth Amendment. Thus, in Anderson v. Martin, 375 U.S. 399 (1964), individuals wishing to become candidates in an election to the county school board of East Baton, Louisiana, brought an action to enjoin the state from enforcing a statute that required nomination papers and ballots to designate the race of candidates for elective office. On its face, the law applied to all races, and it was applied to evenhandedly. Nonetheless, its only plausible purpose was to invite racial cohesion in elections. As the Court explained:

[B]y directing the citizen's attention to the single consideration of race or color, the State indicates that a candidate's race or color is an important--perhaps paramount--consideration in the citizen's choice, which may decisively influence the citizen to cast his ballot along racial lines. Hence in a State or voting district where Negroes predominate, that race is likely to be favored by a racial designation on the ballot, while in those communities where other races are in the majority, they may be preferred. The vice lies not in the resulting injury but in the placing of the power of the State behind a racial classification that induces racial prejudice at the polls.

The use of race as an identifying feature in criminal investigations is quite different. It is not designed to induce or capitalize on racial prejudice, but rather serves the legitimate purpose of enabling authorities to identify and apprehend the perpetrators of crimes. Moreover, as discussed in the text below, the use of DNA markers could serve to reduce the operation of racial prejudice by providing an objective indication of the race of the perpetrator.

These characteristics of inferences as to the race of the source of a crime-scene sample apply even if the DNA genotypes associated with race distinguish among but a handful of races. For example, suppose (unrealistically) that only two groups could be distinguished with reasonable accuracy--individuals of African descent and those of European descent. Deciding that the crime-samples in one set of cases probably originated from one group, while those in another set of cases probably came from the other group, would represent an evenhanded application of the available information to all people. It is comparable to using black-and-white surveillance videos of bank robberies. As long as the police consistently focus investigations on individuals with the complexion of those whose depicted on film, they do not purposefully disadvantage any one group, and the practice would not constitute the invidious discrimination forbidden by the equal protection clause.

111. See Jeffrey Goldberg, The Color of Suspicion, N.Y. Times Mag., June 20 1999; Steven A. Holmes, Clinton Orders Investigation on Possible Racial Profiling, N.Y. Times, June 10, 1999, available in 1999 WL 30527392; Jerry Gray, Unwelcome Addition to a Legacy, N.Y. Times, May 1, 1999, available in 1999 WL 9882423.

112. This risk can be intolerably high when race is itself the factor on which officers rely. Consequently, the Equal Protection Clause may require the use of factors--gang membership, for example--that are not explicitly racial, even though these factors can be highly correlated with race. (It is well settled that the use of valid decision-making factors does not violate the Equal Protection Clause even when they have a disparate impact on one race. Hunt v. Crowmartie, 526 U.S. 541 (1999) ("A facially neutral law, on the other hand, warrants strict scrutiny only if it can be proved that the law was motivated by a racial purpose or object, or if it is unexplainable on grounds other than race.") (internal quotation marks and citations omitted); Washington v. Davis, 426 U.S. 229 (1976).) When the underlying, non-racial factors are available, they are more accurate indicators than the racial surrogate for them.

113. By "racial information," we simply mean that the genetic markers are in fact far more common within one "racial group" than another.

114. The description of these standards that follows is drawn from Committee on the

DNA Technology in Forensic Science: An Update, National Research Council, The Evaluation of Forensic DNA Evidence (1996) [hereinafter NRC II].

117. This history of the judicial reception of DNA evidence is adapted from 1 McCormick on Evidence § 205 (John Strong ed., 5th ed. 1999).

119. Andrews v. State, 533 So.2d 841 (Fla. Ct. App.1988); People v. Wesley, 533 N.Y.S.2d 643 (Albany County Ct.1988); Spencer v. Commonwealth, 384 S.E.2d 775 (Va. 1989) (early version of DQ test properly admitted where "[t]he record is replete with uncontradicted expert testimony that no 'dissent whatsoever (exists) in the scientific community'"); State v. Woodall, 385 S.E.2d 253 (W. Va. 1989) (taking judicial notice of general scientific acceptance where there was no expert testimony, but holding that inconclusive results were properly excluded as irrelevant); Thomas M. Fleming, Annotation, Admissibility of DNA Identification Evidence, 84 A.L.R.4th 313 (1991).

120. For a comprehensive survey of possible sources of error and ambiguity in VNTR profiling, see William Thompson & Simon Ford, The Meaning of a Match: Sources of Ambiguity in the Interpretation of DNA Prints, in Forensic DNA Technology 93 (Farley & Harrington eds., 1990).

121. See United States v. Yee, 134 F.R.D. 161 (N.D. Ohio 1991), aff'd sub nom. United States v. Bonds, 12 F.3d 540 (6th Cir. 1993); Anderson, DNA Fingerprinting on Trial, 342 Nature 844 (1989); Thompson & Ford, Is DNA Fingerprinting Ready for the Courts?, New Scientist, Mar. 31, 1990, at 38; Gina Kolata, Some Scientists Doubt the Value of "Genetic Fingerprint" Evidence, N.Y. Times, Jan. 29, 1990, at A1 col. 1, available in 1990 WL 2053257 (reporting that "[l]eading molecular biologists say a technique promoted by the nation's top law-enforcement agency for identifying suspects in criminal trials through the analysis of genetic material is too unreliable to be used in court"; but the accuracy of this report is seriously questioned in Moenssens, DNA Evidence and Its Critics--How Valid Are the Challenges?, 31 Jurimetrics J. 87 (1990)).

122. E.g., Yee, 134 F.R.D. 161; State v. Pennington, 393 S.E.2d 847 (N.C.1990) (uncontradicted expert testimony that false positives are impossible); Glover v. State, 787 S.W.2d 544 (Tex. Ct. App. 1990) (admissible in light of other decisions where "[a]ppellant did not produce any expert testimony").

123. Md. Code Ann. Cts. & Jud. Proc. Code Ann. § 10-915 (Michie Supp. 1992) ("In any criminal proceeding, the evidence of a DNA profile is admissible to prove or disprove the identity of any person"); Minn. Stat. Ann. § 634.25 (West Supp. 1992) ("In a criminal trial or hearing, the results of DNA analysis . . . are admissible in evidence without antecedent expert testimony that DNA analysis provides a trustworthy and reliable method of identifying characteristics in an individual's genetic material upon a showing that the offered testimony meets the standards for admissibility set forth in the Rules of Evidence."); Kenneth E. Melson, Legal and Ethical Considerations, in DNA Fingerprinting: An Introduction 189, 199-200 (Lorne T. Kirby ed., 1990).

124. People v. Castro, 545 N.Y.S.2d 985, 995 (Sup. Ct. 1989) (principles of DNA testing generally accepted, but "[i]n a piercing attack upon each molecule of evidence presented, the defense was successful in demonstrating to this court that the testing laboratory failed in its responsibility to perform the accepted scientific techniques and experiments"); State v. Schwartz, 447 N.W.2d 422, 428 (Minn. 1989) ("DNA typing has gained general acceptance in the scientific community," but "the laboratory in this case did not comport" with "appropriate standards"); Colin Norman, Maine Case Deals a Blow to DNA Fingerprinting, 246 Science 1556 (1989); Rorie Sherman, DNA Tests Unravel?, Nat'l L.J., Dec. 18, 1989, at 1, 24-25.

125. This innovation was introduced in People v. Castro, 545 N.Y.S. 2d 985 (Sup. Ct. 1989). It soon spread. See United States v. Two Bulls, 918 F.2d 56, 61 (8th Cir. 1990) ("it was error for the trial court to determine the admissibility of the DNA evidence without determining whether the testing procedures . . . were conducted properly"), vacated for rehearing en banc, app. dismissed due to death of defendant, 925 F.2d 1127 (8th Cir. 1991); Ex parte Perry, 586 So. 2d 242 (Ala. 1991). For cases declining to graft a "third prong" onto Frye, see, for example, State v. Bible, 858 P.2d 1152 (Ariz. 1993); Hopkins v. State, 579 N.E. 2d 1297 (Ind. 1991); State v. Vandebogart, 616 A.2d 483 (N.H. 1992); State v. Cauthron, 846 P.2d 502 (Wash. 1993).

126. Later, some courts insisted on such a showing as part of the demonstration of scientific soundness required under Daubert. E.g., United States v. Martinez, 3 F.3d 1191 (8th Cir. 1993).

127. For an analysis concluding that such matters are better handled not as part of the special test for scientific evidence, but as aspects of the balancing of probative value and prejudice, see Margaret A. Berger, Laboratory Error Seen Through the Lens of Science and Policy, 30 U.C. Davis L. Rev. 1081 (1997).

128. See NRC II, supra note 114, at 205-11 (exhaustively tabulating cases); David H. Kaye, DNA Evidence: Probability, Population Genetics, and the Courts, 7 Harv. J. L. & Tech. 101 (1993).

129. See Kaye, supra note 128; William C. Thompson, Evaluating the Admissibility of New Genetic Identification Tests: Lessons from the "DNA War," 84 J. Crim. L. & Criminology 22 (1993). In the light of the totality of information on the distribution of various genes in populations, the criticism of the simple random-mating model may have been overblown. See Bernard Devlin & Kathryn Roeder, DNA Profiling: Statistics and Population Genetics, in 1 Modern Scientific Evidence: The Law and Science of Expert Testimony 710 (David Faigman et al. eds., 1997).

131. National Research Council Committee on DNA Technology in Forensic Science, DNA Technology in Forensic Science (1992) [hereinafter NRC I].

132. See Kaye, The Forensic Debut of the National Research Council's DNA Report: Population Structure, Ceiling Frequencies and the Need for Numbers, 96 Genetica 99 (1995), published in slightly different form in 34 Jurimetrics J. 369 (1994).

133. NRC II, supra note 114. The 1996 report provides more refined methods for estimating allele frequencies in ethnic subpopulations.

134. People v. Soto, 981 P.2d 958, 974 (Cal. 1999) ("Several developments since the filing of Barney indicate the controversy over population substructuring and use of the unmodified product rule has dissipated."); People v. Miller, 670 N.E.2d 721, 731-32 (Ill. 1996) ("while there has been some controversy over the use of the product rule in calculating the frequency of a DNA match, that controversy appears to be dissipating"); Armstead v. State, 673 A.2d 221, 238 (Md. 1996) ("the debate over the product rule essentially ended in 1993"); Commonwealth v. Fowler, 685 N.E.2d 746 (Mass. 1997) (product rule with and without ceilings for VNTRs now meets test of scientific reliability in light of 1996 NRC Report), departing from Commonwealth v. Lanigan, 596 N.E.2d 311 (Mass. 1992) (dispute over population structure evinces lack of general acceptance).

135. Harrison v. State, 644 N.E.2d 1243 (Ind. 1995) (error not to hold Frye hearing on PCR-based method).

136. For example, a procedure may be so similar to accepted protocols that acceptance or validity can be inferred from previous cases. See United States v. Johnson, 56 F.3d 947 (8th Cir. 1995) (in response to a defense expert's testimony that a police department's variation on the FBI protocol for RFLP-VNTR testing had not been validated, the court held that the variation did not preclude admission where an FBI analyst testified that the difference was of no significance); People v. Oliver, 713 N.E.2d 727, 734 (Ill. Ct. App. 1999) ("the minor variations . . . in the . . . second RFLP test did not render it a new scientific technique for the purposes of Frye."). In general, trial courts have considerable "latitude in deciding how to test an expert's reliability, and to decide whether or when special briefing or other proceedings are needed to investigate reliability." Kumho Tire Co., Ltd., v. Carmichael, 119 S.Ct. 1167, 1176 (1999).

137. Commonwealth v. Rosier, 685 N.E.2d 739 (Mass. 1997) (STR testing and product-rule calculations admissible); Commonwealth v. Vao Sok, 683 N.E.2d 671, 672-73 (Mass. 1997) (PCR-based DQ, Polymarker, and D1S80 analysis "meet the test of scientific reliability"); State v. Jackson, 582 N.W.2d 317 (Neb. 1998) (STR testing); State v. Harvey, 699 A.2d 596 (N.J. 1997) (DQ and Polymarker tests generally accepted); State v. Lyons, 924 P.2d 802 (Or. 1996) (DQ admissible under relevancy standard); State v. Moeller, 548 N.W.2d 465 (S.D. 1996) (DQ admissible under Daubert standard).

138. George Sensabaugh & D.H. Kaye, Non-human DNA Evidence, 39 Jurimetrics J. 1 (1998).

140. Proficiency testing in forensic genetic testing is designed to ascertain whether an analyst can correctly determine genetic types in a sample the origin of which is unknown to the analyst but is known to a tester. Proficiency is demonstrated by making correct genetic typing determinations in repeated trials, and not by opining on whether the sample originated from a particular individual. Proficiency tests also require laboratories to report random match probabilities to determine if proper calculations are being made.

141. See, e.g., Barry C. Scheck, DNA and Daubert, 15 Cardozo L. Rev. 1959, 1979-87 (1994); William C. Thompson, Accepting Lower Standards: The National Research Council's Second Report on Forensic DNA Evidence, 37 Jurimetrics J. 405, 417 (1997).

142. See, e.g., Jonathan J. Koehler, Error and Exaggeration in the Presentation of DNA Evidence at Trial, 34 Jurimetrics J. 21, 37-38 (1993); Scheck, supra note 141, at 1984 n.93.

143. Cf. James Wooley & Rockne P. Harmon, The Forensic DNA Brouhaha: Science or Debate?, 51 Am. J. Hum. Genetics 1164 (1992) (letter urging defense experts who criticize laboratory procedures to do their own tests).

145. See Edward J. Imwinkelried, The Case Against Evidentiary Admissibility Standards that Attempt to "Freeze" the State of a Scientific Technique, 67 U. Colo. L. Rev. 887 (1996). In extreme cases, where the laboratory departs so grossly from accepted practices that the reliability of its findings are in serious doubt, the court may well exclude the evidence on the ground that its probative value is too slight to warrant its admission.

149. The report stated that the committee had chosen to limit its remarks to the question of "what aspects of the procedures used in connection with forensic DNA testing are scientifically valid . . . ." NRC II, supra note 114, at 185.

151. Id. (asserting that "[a] laboratory is not likely to make the same error again"); see also NRC I, supra note 131, at 120 (recognizing that "errors on proficiency tests do not necessarily reflect permanent probabilities of false-positive or false-negative results").

152. Cf. United States v. Lowe, 954 F. Supp. 401, 415 (D. Mass. 1997), aff'd, 145 F.3d 45 (1st Cir. 1998) (rejecting the argument that an expert must present an error rate from proficiency tests along with the random-match probability).

153. Edward J. Imwinkelried, Coming to Grips with Scientific Research in Daubert's "Brave New World": The Courts' Need to Appreciate the Evidentiary Differences between Validity and Proficiency Studies, 61 Brooklyn L. Rev. 1247, 1273-78 (1995).

154. See, e.g., Moorhead v. Mitsubishi Aircraft Int'l, 828 F.2d 278 (5th Cir. 1987) (error to admit pilot's low marks at flight school refresher course); see generally 1 McCormick on Evidence, supra note 117, § 186.

155. Federal Rule 404(a) provides that "[e]vidence of a person's character or a trait of his character is not admissible for the purpose of proving that he acted in conformity therewith on a particular occasion . . . ." Section (b) of the rule recognizes certain exceptions to this blanket rule of exclusion, but none are apposite here. There also is an exception permitting a witness's opponent to impeach the witness by questioning the witness about prior untruthful acts. Fed. R. Evid. 608(b). However, those acts relate to the witness's character trait for untruthfulness, rather the trait of competence or proficiency. The Federal Rules expressly carve out the exception for untruthfulness, but there is no comparable exception for the character trait of competence or proficiency.

156. Federal Rule 405(a) provides that "In all cases in which evidence of character or a trait of character of a person is admissible, proof may be made by testimony as to reputation or by testimony in the form of an opinion. On cross-examination, inquiry is allowable into relevant specific instances of conduct." Section (b) permits specific act evidence only when character is "in issue"--a term of art that has no application to the tendency of laboratory personnel to make mistakes in performing DNA tests. See 1 McCormick on Evidence, supra note 117, § 187.

157. But see United States v. Shea, 957 F. Supp. 331, 344 n.42 (D.N.H. 1997) ("The parties assume that error rate information is admissible at trial. This assumption may well be incorrect. Even though a laboratory or industry error rate may be logically relevant, a strong argument can be made that such evidence is barred by Fed. R. Evid. 404 because it is inadmissible propensity evidence."); Unmack v. Deaconess Medical Center, 967 P.2d 783 (Mont. 1997) (going to the brink of explicitly holding that the character evidence prohibition bars this type of testimony).

158. Of course, even when the proficiency test data would be admissible and the defense has a legitimate need to discover this type of information, there might be means of satisfying the need other than by furnishing proficiency test results. By way of example, a sampling of the laboratory's case work could meet the need. However, in most cases permitting discovery of proficiency test data may be preferable. It will likely be more convenient for the laboratory to reveal the proficiency test data, since that data has already been compiled and giving the defense access to actual case work could compromise the privacy of the persons involved in those cases. When the defense needs to discover information about the operating characteristics of a laboratory's test for a purpose other than merely establishing the laboratory's general error rate, the data could prove to be admissible at trial; hence, the courts would not be justified in denying discovery of proficiency test results on the ground that such discovery cannot lead to the production of admissible evidence at trial.

159. Combining the random-match probability with the probability of a false-positive laboratory error according the rules governing conditional probabilities would give the jury an estimate of the probability that the laboratory would find a match if the source of the crime-scene DNA were neither the defendant nor a close relative. When random-match probabilities are orders of magnitude smaller than estimates of the chance of a laboratory error of some kind, the possible error rate derived from proficiency testing dominates the combined error risk. As explained in William C. Thompson, Accepting Lower Standards: The National Research Councils Second Report on Forensic DNA Evidence, 37 Jurimetrics J. 405, 421 n.59 (1997):

The overall probability that a match will be declared if the samples are from different people is approximately (although not precisely) the sum of the probability of an erroneous match and the probability of a random match. Let S designate that two samples have the same source and S that they do not; let M designate that two samples have matching DNA profiles and M that they do not; and let D designate that a match is declared by a DNA analyst following testing. The overall probability of a false match, P(D|S), is not simply the sum of the probability of a random match, P(M|S), and the probability of an erroneous match, P(D|M), but rather, P(D|S) = P(D|M)P(M|S) + P(D|M)P(M|S). Because P(D|M) and P(M|S) will usually be close to one, however, the sum of the probability of a random match and an erroneous match is close to the overall probability of a false match.

162. See generally Margaret A. Berger, Laboratory Error Seen Through the Lens of Science and Policy, 30 U.C. Davis L. Rev. 1081 (1997).

163. See Jason Schklar & Shari Seidman Diamond, Juror Reactions to DNA Evidence: Errors and Expectancies, 23 Law & Hum. Behav. 159, 179 (1999) (concluding that separate figures are desirable in that "[j]urors . . . may need to know the disaggregated elements that influence the aggregated estimate as well as how they were combined in order to evaluate the DNA test results in the context of their background beliefs and the other evidence introduced at trial.").

164. E.g., Armstead v. State, 673 A.2d 221 (Md. 1996) (rejecting the argument that the introduction of a random match probability deprives the defendant of due process because the error rate on proficiency tests is many orders of magnitude greater than the match probability).

165. This theory does not apply if the defense introduces the proficiency test data by cross-examining the prosecution's experts rather than producing its own expert.

166. See, e.g., Bert Black et al., Science and the Law in the Wake of Daubert: A New Search for Scientific Knowledge, 72 Tex. L. Rev. 716 (1994); George Rice et al., Male Homosexuality: Absence of Linkage to Microsatellite Markers at Xq28, 284 Science 665 (1999).

168. The impeaching statement or conduct "need only bend in a different direction." John M. Mcnaught & Harold Flannery, Massachusetts Evidence: A Courtroom Reference 13-5 (1988).

169. Thus, in People v. Oliver, 713 N.E.2d 727 (Ill. Ct. App. 1999), the state was allowed to show that a defense expert who questioned the results of DNA tests had done no testing of his own. See id. at 736 ("it was proper for the prosecution to bring out on cross-examination that the defense criticisms of the prosecution's expert witnesses were not based on any independent testing that it had done.").

172. On this theory, the defense is not entitled to a limiting instruction; an admission by conduct qualifies as substantive evidence.

175. Miller v. District Court, 737 P.2d 834, 838 (Colo. 1987); State v. Pratt, 398 A.2d 421 (Md. 1979); White v. State, __P.2d__, 1999 W.L. 124310 (Okla. Crim. App. 1999); State v. Riddle, 964 P.2d 1056, 1063 (Or. Ct. App.), modified, 969 P.2d 1032 (Or. Ct. App. 1998); Casenote, Disclosures by Criminal Defendant to Defense-Retained Psychiatirst Held Within Scope of Attorney-Client Privilege Which Defendant Does Not Waive by Pleading Insanity, 9 U. Balt. L. Rev. 99, 111 (1979).

176. As enacted, Rule 503 leaves the recognition and development of privileges under federal law to the courts. The original draft would have codified and defined the privileges. Its description of the attorney-client privilege remains useful to courts as they continue to define and refine that privilege.

177. State v. Riddle, 964 P.2d 1056 (Or. Ct. App.), modified, 969 P.2d 1032 (Or. Ct. App. 1968) (accident reconstruction expert).

178. Thus, in Fluellen v. Campbell, 683 F. Supp. 186 (M.D. Tenn. 1987), defense counsel argued that the state's case was weakened by the fact that it failed to have blood tests performed, and the prosecutor remarked in rebuttal "if he thinks that is such good evidence, why didn't he request that it be done?" Id. at 89. The federal district court found that "this comment in no way imposed upon the jury a presumption which conflicted 'with the overriding presumption of innocence with which the law endows the accused and which extends to every element of the crime.'" Id.

179. Alan Stephens, Annotation, Adverse Presumption or Inference Based on Party's Failure to Produce or Examine Family Members Other than Spouse--Modern Cases, 80 A.LR.4th 337, 344 (1990); Alan Stephens, Annotation, Adverse Presumption or Inference Based on Party's Failure to Produce or Examine Friend--Modern Cases, 79 A.L.R.4th 779, 785-86 (1990).

180. Most, but not all jurisdictions also forbid comment on the invocation of a statutory or common-law privilege. 1 McCormick, supra note 117, § 74.1.

181. See, e.g., Alan Stephens, Annotation, Adverse Presumption or Inference Based on Party's Failure to Produce or Examine Spouse-Modern Cases, 79 A.L.R.4th 694 (1990).

182. In May 1998, for example, the Illinois Senate approved House Bill No. 2257, which provides:

[I]f the identity of the accused is unknown and at the time of the offense physical evidence is collected that is capable of being tested for its DNA characteristics which would identify the accused, a prosecution for predatory criminal sexual assault of a child, criminal sexual assault or aggravated criminal sexual assault may be commenced at any time.

1997 IL H.B. 2257 (SN); see David Heckelman, Senate Oks Bill Ending Limits on Prosecuting Some Sex Crimes, Chi. Daily L. Bull., May 8, 1998. Currently most felony prosecutions in Illinois must be commenced within three years after the crime was committed. 720 Ill. Comp. Stat. 5/3-5 (West 1993). If the victim reported the assault to the police within six months of the assault, prosecution may be commenced within five years of the assault. 720 Ill. Comp. Stat. 5/3-6 (West Supp. 1998). When the victim is a minor, an action for sexual assault may be brought within one year after the victim reaches majority. 720 Ill. Comp. Stat. 5/3-6 (West Supp. 1998).

183. See Jonathan W. Diehl, Note, Drafting a Fair DNA Exception to the Statute of Limitations in Sexual Assault Cases, 39 Jurimetrics J. 431 (1999). Portions of our discussion of statutes of limitations are adapted, without further attribution, from this note.

185. United States v. Ewell, 383 U.S. 116, 122 (1966); see also United States v. Marion, 404 U.S. 307, 321 (1971); Model Penal Code at 86; Tyler T. Ochoa & Andrew J. Wistrich, The Puzzling Purposes of Statutes of Limitation, 28 Pac. L.J. 453, 458 (1997). The Due Process Clause of the Fifth Amendment also provides protection, but only when the defendant establishes that the delay not only substantially prejudiced the defense, but also that it was a tactical ploy by the prosecutor. See Marion, supra, 404 U.S. at 324; United States v. Lovasco, 431 U.S. 783, 790 (1977). The Speedy Trial Clause of the Sixth Amendment offers the defendant no protection against preindictment delay. Id. at 313.

186. The need for specific deterrence, incapacitation, and retribution usually fades with time. See Model Penal Code § 1.06 cmt. 1, at 86 (1985).

187. The DNA evidence does not grow stale with the passage of time. Even ancient DNA sometimes can be analyzed successfully. See, e.g., Ancient DNA: Recovery and Analysis of Genetic Material from Paleontological, Archaeological, Museum, Medical, and Forensic Specimens (B. Hermann & S. Hummel eds., 1994). However, the ability of fragments of the DNA molecule to remain intact virtually indefinitely under certain conditions is not dispositive. Indeed, if the prosecution relies on DNA profiling of a stain shortly after a fairly fresh crime-scene sample first was collected, it is not even pertinent.

188. Some of the wording of the Illinois House Bill 2257 points in this direction. See supra note 182 (requiring that "the identity of the accused is unknown and at the time of the offense physical evidence is collected that is capable of being tested for its DNA characteristics").

189. For recommended procedures of handling requests for DNA testing after conviction, see National Commission on the Future of DNA Evidence, Report and Recommendations Postconviction Relief (1999).

190. The agencies that administer the program may be required to promulgate regulations. E.g., Md. Ann. Code art. 88B § 12A(b) (1998); Mass. Ann. Laws ch. 22E § 6 (Law. Co-op. 1999) ("The director shall promulgate regulations governing the collection, submission, receipt, identification, storage, and disposal of DNA samples.").

196. The table was prepared from research conducted by Judith Schelling and completed in April 1999.

197. M. Dawn Herkenham, U.S. Dep't of Justice, State DNA Database Statutes: Summary of Provisions, at unnumbered page 4 (1999) (table of "State DNA Database Laws Qualifying Offenses").

199. Statutes often include juvenile offenders as well. See, e.g., In re Appeal in Maricopa County Juvenile Action Nos. JV-512600 and JV-512797, 930 P.2d 496 (Ariz. Ct. App. 1996) (juvenile sex offenders can be compelled to give blood samples because the fact of a conviction substitutes for probable cause and a warrant and the need for apprehension and deterrence outweigh minor intrusion of blood sampling). This is reasonable because some juvenile offenders have committed or soon will commit other target crimes. However, indefinite retention of the samples and records of juvenile offenders who have matured without further incident is more difficult to justify. See Nicholas deB. Katzenbach & Richard W. Tomc, Crime Data Centers: The Use of Computers in Crime Detection and Prevention, in Surveillance, Dataveillance, and Personal Freedoms 59, 63-65 (Columbia Human Rights Law Review ed., 1972)

200. Cf. Rise v. Oregon, 59 F.3d 1556, 1560 (9th Cir. 1995) ("convicted felons . . . do not have the same expectations of privacy in their identifying genetic information that 'free persons' have."); Jones v. Murray, 962 F.2d 302, 306 (4th Cir. 1992) ("With the person's loss of liberty upon arrest comes the loss of at least some, if not all, rights to personal privacy otherwise protected by the Fourth Amendment.").

201. The Courts of Appeals in Rise and Jones were not asserting so broad a claim. Rather, they were speaking of a diminution in the legitimate expectation of privacy that has accompanied the practice of recording identifying features of individuals who are imprisoned.

202. See Bell v. Wolfish, 441 U.S. 520, 559-60 (1979) (pretrial detainees); Hudson v. Palmer, 468 U.S. 517, 530 (1984) (convicted prisoners); cf. Griffin v. Wisconsin, 483 U.S. 868, 880 (1987) (probationers).

203. For example, in Jones v. Murray, 962 F.2d 302 (4th Cir. 1992), the Court of Appeals for the Fourth Circuit observed that:

Commonwealth officials say that the program attempts to address the problem of felony recidivism in Virginia by identifying and increasing the likelihood of convicting repeat offenders and by deterring those who might otherwise commit a second felony. According to a study of violent felons convicted in Virginia between 1985 and 1987, 36.4% had at least one prior conviction for a felony. Only 26% had no prior criminal record and just over 19% had previously been convicted of nonviolent felonies. A United States Department of Justice survey of more than half of those persons released from the prisons of eleven states in 1983 revealed that an estimated 62.5% were arrested again for a felony or serious misdemeanor within three years after release.

Id. at 304. Likewise, in Rise v. Oregon, 59 F.3d 1556 (9th Cir. 1995), the Ninth Circuit referred to "uncontroverted evidence documenting the high rates of recidivism among certain types of murderers and sexual offenders." Id. at 1561.

204. The past tense is necessary because finding that an individual who has committed a crime today also has committed previously unsolved crimes is important.

205. NRC I, supra note 131, at 120; see also Philip L. Bereano, The Impact of DNA-based Identification Systems on Civil Liberties, in DNA on Trial: Genetic Identification and Criminal Justice 119, 121 (Paul R. Billings ed., 1992) ("much crime leaves no identifying tissue behind (e.g., burglary)."). The committee noted, however, that "the databank would be useful for some other investigations," apparently because "[i]n a small minority of homicides, blood, hair, or tissue samples from the perpetrator are left at the scene of the crime . . . ." NRC I, supra, at 120.

206. NRC I, supra note 131, at 120 ("State legislatures setting up . . . databanks should weigh the benefits in terms of solved rape cases and the costs in terms of collecting samples from people likely to commit rapes (primarily, it seems, convicted sex offenders).").

209. See Lawrence A. Greenfeld, U.S. Dep't of Justice, Sex Offenses and Offenders: An analysis of Data on Rape and Sexual Assault 26 (1997).

210. Id. The rapists' re-arrest rates for other violent crimes were 2.8% for murder, 8.5% for robbery, and 10.7% for assault. The percentages relate to a base of 2,214 rapists who were studied. Id.

211. In 1994, 429,000 offenders were serving time in state penitentiaries for violent offenses; 88,100 of these prisoners were serving time for rape or sexual assault. Id. at 22.

212. Of the 88,000 or so prisoners serving time for rape and sexual assault, some 42,000 had prior convictions for crimes other than rape or sexual assault, while less than 11,000 had prior convictions for rape or sexual assault. See id. at 22 figure 23. Because the study does not indicate how prisoners with prior convictions for both types of crimes were classified, however, the ratio of 4 to 1 is only an approximation.

213. See, e.g., 60 Minutes: DNA Data Banks Keep Track of Criminals, Cause Controversy (CBS television broadcast, Apr. 18, 1999), available in 1999 WL 16209028 (according to the director of Virginia's DNA database, "[i]n about 60 percent of these [rape and murder] cases, the sample that we had in the data bank came from an earlier burglary or breaking-and-entering conviction."); Richard Willing, States Adding Burglars to DNA Databases: Police Re-examine Criminals Once Thought of as Relatively Innocuous, USA Today, Dec. 7, 1998, at 1A, available in 1998 WL 5743996 (reporting that Virginia's DNA database has matched burglars to 18 rapes).

214. David Coffman, Address at the Fourth Annual National Conference on the Future of DNA: Implications for the Criminal Justice System, Albuquerque, N.M., May 3, 1999 (reporting that hits in rape investigations doubled after aggravated battery became a "collection crime" in Florida, perhaps because aggravated battery is the lesser plea that prosecutors in Florida most commonly accept in plea bargains with rapists).

215. Greenfeld, supra note 209, at 26. As these numbers suggest, sex offenders are not more prone to repeat their offenses than are other classes of major felons. See Katherine K. Baker, Once a Rapist? Motivational Evidence and Relevancy in Rape Law, 110 Harv. L. Rev. 563, 578-80 (1997); Thomas J. Reed, Reading Gaol Revisited: Admission of Uncharged Misconduct Evidence in Sex Offender Cases, 21 Am. J. Crim. L. 127, 149, 154-55 (1993); Paul R. Rice, The Evidence Project: Proposed Revisions to the Federal Rules of Evidence with Supporting Commentary, 171 F.R.D. 330, 479 (1997).

216. Willing, supra note 213. Nearly two-thirds of the 27,099 hits in Britain's database from mid-1997 to the end of 1998 were matches with burglars. Of course, the value of collecting DNA from burglars (and other offenders) could decline as criminals learn to take precautions against depositing their DNA at crime scenes.

217. The frequency with which the genotype in a trace evidence sample occurs among individuals who are not associated with the crime is pertinent to assessing the probative value of a match. See, e.g., D.H. Kaye, The Relevance of "Matching" DNA: Is the Window Half Open or Half Shut?, 85 J. Crim. L. & Criminology 676 (1995). Given the nature of the loci used in forensic testing, no systematic differences would be expected in the frequencies among convicted offenders and people who are not guilty of the crime under investigation. Therefore, the offender data can be used to increase the precision of frequency estimates derived from smaller samples provided by blood banks or other convenient sources. They also offer a source of information on the distribution of alleles in individual states, permitting one to verify the assumption that estimates from national data or data in other locales apply to each state. Cf. NRC II, supra note 114 (pointing out that the geographic hetereogeneity in frequencies is likely to be the result of statistical fluctuations in random mating and arguing that national samples are likely to give better estimates of local frequencies because of their larger size).

219. E.g., Cal. Penal Code § 295(c) (West 1999) ("the expeditious detection and prosecution of individuals responsible for sex offenses and other violent crimes, the exclusion of suspects who are being investigated for these crimes"); Md. Ann. Code art. 88B § 12A(g)(1)(ii) (Michie 1999) ("In the furtherance of an official investigation into a criminal offense"); S.C. Code Ann. § 23-3-640(B)(1) (Law. Co-op. 1998) ("to identify suspects in otherwise nonsuspect cases"); id. at § 23-3-640(B)(4) ""to generate investigative leads"); Tex. Gov't Code Ann. § 411.143(e)(1) (West 1999) ("generate an investigative lead or exclusion"); Va. Code Ann. § 19.2-310.5 (Michie 1995 & Supp. 1999) (making searches available for use in any criminal offense investigation).

221. In one reported case, DNA obtained with a search warrant served on a suspected rapist excluded the suspect, but the unusual similarity between this suspect's genotype and the trace evidence led investigators to his brother, whose DNA did match. Flowers v. State, 654 N.E.2d 1124, 1124 (Ind. 1995).

222. See Michelle Hibbert, DNA Databanks: Law Enforcement's Greatest Surveillance Tool?, 34 Wake Forest L. Rev. 767 (1999); Jean E. McEwen, Storing Genes to Solve Crimes: Legal, Ethical, and Public Policy Considerations, in Stored Tissue Samples: Ethical, Legal, and Public Policy Considerations 311, 316-18 (Robert F. Weir ed., 1998) (describing this possibility as having "potentially important civil liberties ramifications" for "the harassment of whole families or racial groups by a sort of 'genetic guilt by association'"); Brief of Amicus Curiae, Institute for Science, Law and Technology, Illinois Institute of Technology [IIT Brief], Landry v. Attorney General, 709 N.E.2d 1085 (Mass. 1999), at 34 ("[W]here a law enforcement agency either purposefully or incidentally gathers information about a non-banked individual by comparing a DNA artifact [sic] to their [sic] sibling's profile digitized in the system, the state has intruded on the privacy of an individual who has not committed any act warranting this level of genomic intrusion."). Of course, even if this concern were valid, it would not invalidate DNA databanking as such. At most, it would warrant a prohibition on noting the presence of any nearly matching genotypes in the database (and perhaps the collection of DNA from individuals with monozygotic twins).

223. In contrast, when genetic testing for disease-related alleles indicates a substantial chance that a relative is at risk for a preventable or curable disease, difficult problems of confidentiality and possible liability for a failure to warn the relatives arise. See, e.g., Ellen Wright Clayton, What Should the Law Say about Disclosure of Genetic Information to Relatives?, 1 J. Health Care L. & Pol'y 373 (1998); Sonia M. Suter, Whose Genes Are These Anyway? Familial Conflict over Access to Genetic Information, 91 Mich. L. Rev. 1854 (1993).

224. The doctrine of "standing" also comes into play here. At one time, some state supreme courts recognized a vicarious standing doctrine, permitting the defendant to move to suppress on the basis of the violation of another person's Fourth Amendment rights. These courts reasoned that even if the defendant did not allege a violation of his own constitutional rights, he should be permitted to assert the rights of others "on the ground that the government must not be allowed to profit by its own wrong and thus encouraged in the lawless enforcement of the law." People v. Martin, 290 P.2d 855, 857 (Cal. 1955). However, both before and after Katz, the Supreme Court made it clear that it does not embrace the doctrine. Wong Sun v. United States, 371 U.S. 471 (1963); Alderman v. United States, 394 U.S. 165, 174-75 (1969). The Court demands that the person moving to suppress evidence establish that the evidence is the product of an intrusion that violated his or her own reasonable expectation of privacy. Minnesota v. Carter, 119 S.Ct. 469 (1998); Rakas v. Illinois, 439 U.S. 128 (1978). It is not enough that the intrusion yields evidence that happens to incriminate that person. Rawlings v. Kentucky, 448 U.S. 98 (1980).

By the same token, the relative who becomes the target of an investigation and a criminal defendant may not challenge a DNA database search simply because the search gives the police information that incriminates the relative. Even if conducting the search constitutes a cognizable Fourth Amendment intrusion, the search does not intrude upon the defendant's privacy; the search entails an analysis of the sample furnished by the defendant's relative. The defendant therefore may not complain on Fourth Amendment grounds.

A more difficult question is whether a convicted offender who somehow could demonstrate that the statute requiring him to provide a DNA sample is an unreasonable infringement of the privacy of relatives has vicarious standing to assert the relatives' rights. One plausible view is that all interests that would be affected by databanking should considered in ascertaining the reasonableness of the statutory scheme. See Krent, supra note 12. The barrier, however, is purely procedural. Including relatives as well as offenders as plaintiffs in an action for an injunction or a declaratory judgment should sidestep this possible obstacle, for even if the offender lacks vicarious standing in this context, the relatives who might be affected should be permitted to argue that the use of the DNA would infringe their rights.

227. In the true plain view cases, the police need at least a founded suspicion and perhaps probable cause to believe that the object seized is evidence of a crime. Texas v. Brown, 460 U.S. 730, 742 n. 7 (1983) ("We need not address whether in some circumstances, a degree of suspicion lower than probable cause would be sufficient basis for seizure in certain cases"). However, it must be remembered that in these cases, the police are seizing an object covered by the Fourth Amendment such as physical evidence. In the DNA case, the police are not seizing a sample or medical record of the relative. The DNA database search puts the police in possession of information about the relative, but that is the extent of the intrusion into the relative's privacy.

229. For a possible exception, see R.I. Gen. Laws § 12-1.5-11 (1998) ("Access to the DNA databank shall be for identification analysis only and shall be limited upon a showing of need to . . . law enforcement agencies . . . and by court order to the defendant . . . in a criminal case"). A less ambiguous expression of legislative hostility toward using a database in civil cases is Okla. Stat. tit. 74 § 150.27a (C)(1998), which declares that "[t]he information . . . is privileged from discovery and inadmissible as evidence in any civil court proceeding. The information . . . is confidential and shall not be released to the public."

230. E.g., Md. Ann. Code art. 88B § 12A(g)(1)(iii) (1999) ("identification of human remains"); S.C. Code Ann. § 23-3-640(B)(6) (Law. Co-op 1998) ("identification of human remains from mass disasters"); cf. N.H. Rev. Stat. Ann. § 632-A:22 (1998) ("law enforcement officers shall use such information only for the purposes of criminal investigations and prosecutions, or as necessary to the functions of an office of chief medical examiner") (emphasis added).

231. E.g., Cal. Penal Code § 295(c) (West 1999) ("the identification of missing and unidentified persons, particularly abducted children"); Md. Ann. Code art. 88B § 12A(g)(1)(iv) (1999) ("the identification of missing persons"); Tex. Gov't Code Ann. § 411.143(c)(2) (West 1999) ("identification of living or deceased missing persons"); S.C. Code Ann. § 23-3-640(B)(7) (Law. Co-op 1998) ("for other humanitarian purposes including identification of missing persons").

232. E.g., La. Rev. Stat. Ann. 15:602 (West 1997) ("it is the policy of this state to assist federal, state, and local criminal justice and law enforcement agencies in the identification and detection of individuals in criminal investigations and in the identification of missing persons, to assist in the recovery or identification of human remains from disasters, and to assist with other humanitarian identification purposes."); Mo. Stat. Ann. § 650.052(1)(5) (West 1998); S. C. Code Ann. §§ 23-3-610 & 23-3-640(B)(7) (Law. Co-op 1998).

233. See Ky. Rev. Stat. Ann. § 17-175(1) (Michie 1998); Mont. Code Ann. § 44-6-102(2)(b) (1997); Ohio Rev. Code Ann. § 109.573 (Anderson 1999); Tex. Gov't Code Ann. § 411.142(g)(3) (West 1999); Wyo. Stat. Ann. § 7-19-401 (Michie 1999).

234. Ohio Rev. Code Ann. § 109.573(B)(2)(b) (Anderson 1999) ("pursuant to a court order . . . to determine the existence of a parent and child relationship").

235. See, e.g., Ala. Code. § 36-18-27 (b) (1994) ("[i]n judicial proceedings, if otherwise admissible"); N.J. Stat. Ann. § 53:1-20.21(f) (West 1999) ("judicial proceedings, by order of court, if otherwise admissible pursuant to applicable statutes or rules"); cf. DNA Identification Act of 1994, 42 U.S.C. § 14133(b)(1)(B) (results of DNA tests performed by the FBI may be disclosed "in judicial proceedings, if otherwise admissible pursuant to applicable statues or rules"); Violent Offender DNA Identification Act of 1999 § 1502(b)(2)(B) (conditioning federal funding to state plans that limit uses to, among others, "judicial proceedings, if otherwise admissible pursuant to applicable statues or rules").

237. E.g., Violent Offender DNA Identification Act of 1999 § 1502(B)(2)(a); Ky. Rev. Stat. Ann. § 17.170(1) (Michie 1998); Tex. Gov't Code Ann. § 411.147(c)(1) (West 1999).

238. E.g., Cal. Penal Code § 295.1(c)(5) (West 1999) ("Anonymous DNA records used for training, research, statistical analysis of populations, or quality control."); Idaho Code §19-5505(2)(d) (1998) ("[a]nonymous DNA records used for research or quality control"). In the medical research context, the term "anonymous" sometimes is used quite strictly, to require that no one have information linking a sample to its source. See, e.g., Clayton, supra note 243, at 1787; NBAC, supra note 27, at i (table 1). As an interpretation of existing administrative regulations pertaining to federally funded research involving "human subjects" this definition is debatable, and it is not clear whether the statutory requirement of anonymous databases for research is intended to suggest that no one retain a key that could link the records in a research database to the individuals from whom they came. Indeed, to someone who has access to the original offender database, the genotypes themselves could link the records to those individuals.

239. Md. Ann. Code art. 88B § 12A(g)(1)(v)(1) (1999) ("Development of a population data base after identifying information is removed"). In some statutes, these provisions apply only to the data generated from the samples, and not to the samples. See Landry v. Attorney General, 709 N.E.2d 1085, 1096 (Mass. 1999) ("Further, even if the loci tested could reveal genetic information, the Act confines the use of the samples to a DNA analysis which generates on 'numerical identification information' . . . . In addition, the Act limits the purposes for which the DNA records may be distributed, and does not permit dissemination of the DNA samples themselves."). Records supplied to CODIS can include those from states that use samples "for a population statistics database, for identification research and protocol development purposes, or for quality control purposes," as long as "personally identifiable information is removed" for these uses. DNA Identification Act of 1994, 42 U.S.C. § 14133(b)(2).

240. E.g., Ala. Code § 36-18-24(b) (1994) (not referring to anonymity); Tex. Gov't Code Ann. § 411.147(c)(4)(B) (West 1999) (requiring anonymity).

241. At least one statute simply refers to "research and administrative purposes." Md. Ann. Code art. 88B § 12A(g)(1)(v) (1999). These purposes include "[d]evelopment of a population data base after personal identifying information is removed"; "[s]upport of identification research and protocol development of forensic DNA analysis methods;" and "[q]uality control purposes." Id. Another statute uses especially sweeping language, referring to "humanitarian endeavors including, but not limited to, educational research or medical research or development," but only in the context of "a DNA population statistical database which shall not include therein individually identifiable information." Ala. Code § 36-18-31 (1994).

242. See, e.g., IIT Brief, supra note 222, at 30 ("even if the sample has been anonymized, basic ethical principles governing research, such as the Nuremberg Code, mandate that research not be undertaken without the subject's advance, voluntary, informed consent"). The analogy to the type of human experimentation condemned by the Nuremburg tribunal is too strained to require extended discussion. Cf. Arthur I. Caplan & George J. Annas, Letter, 285 Science 48, 48-49 (1999) ("Analogies must be generated with caution. Sloppy analogies to historical events . . . abound. Caution and accuracy are crucial so as not to demean or deprecate the horrific moral abuses to which human beings were subjected in the past in the name of medical progress.").

243. See, e.g., David Korn, Genetic Privacy, Medical Information Privacy, and the Use of Human Tissue Specimens in Research, in Genetic Testing and the Use of Information 16, 30 (Clarisa Long ed., 1999); Ellen Wright Clayton et al., Informed Consent for Genetic Research on Stored Tissue Samples, 274 JAMA 1786 (1995) ("[I]t is widely accepted that informed consent must be obtained for the many projects that involve the direct prospective involvement of individual subjects. The role of informed consent has been much less clear for research that does not require such personal involvement but rather can be performed using tissue samples.").

244. See, e.g., Gerald Dworkin, The Theory and Practice of Autonomy (1988); R.J. Levine, Ethics and Regulations of Clinical Research (2d ed. 1988). But see John Harris, Ethical Genetic Research on Human Subjects, 40 Jurimetrics J. 77 (1999).

245. Most of the rationales for informed consent simply do not apply to research resulting from involuntary takings that cannot harm the sources of the tissue samples. To support a demand for consent, commentators sometimes cite the following interests: (1) "[d]emonstrating respect for persons and avoiding harm"; (2) "promot[ing] the willingness of people to seek medical care because patients can be reassured that they do not give up their right to decide whether to participate in research when they enter the health care system"; (3) "appris[ing] prospective subjects about the nature of the project and about the risks and benefits that accompany participation so that they can decide whether to participate. If they choose to take part, they will know what to expect and may have the opportunity to take steps to avail themselves of the benefits or to avoid or ameliorate the risks." Clayton et al., supra note 243.

246. Dean Hamer & Peter Copeland, Living with Our Genes: Why They Matter More than You Think 30 & 47 (1998).

248. The problem of overreliance on preliminary or insecure biomedical knowledge is hardly confined to genetics. Reports of new cures, toxic substances, miracle diets, and dangerous side-effects usually appear and influence public opinion well before an adequate scientific basis for such conclusions exists. For reports demonstrating the difficulty of locating genetic markers for behavioral characteristics, see, e.g., Peter Aldhous, The Promise and Pitfalls of Molecular Genetics, 257 Science 164 (1992); John C. Crabbe et al., Genetics of Mouse Behavior: Interactions with Laboratory Environment, 284 Science 1670 (1999); John R. Kelsoe et al., Re-evaluation of the Linkage Relationship Between Chromosome 11p Loci and the Gene for Bipolar Affective Disorder in the Old Order Amish, 342 Nature 238 (1989); Charles C. Mann, Behavioral Genetics in Transition, 264 Science 1686 (1994); George Rice et al., Male Homosexuality: Absence of Linkage to Microsatellite Markers at Xq28, 284 Science 665 (1999). For a more enthusiastic account of the promise of behavior genetics, see, e.g., Hamer & Copeland, supra note 246, at 302-03.

249. See generally R.C. Lewontin, The Dream of the Human Genome, N.Y. Rev. Books, May 28, 1992, at 31; Hugh Miller, III, DNA Blueprints, Personhood, and Genetic Privacy, 8 Health Matrix 179 (1998).

251. Forty-six states expressly restrict disclosure, and 32 penalize authorized disclosure. Herkenham, supra note 197, at unnumbered page 7 (1999) (table of "State DNA Database Laws"). For examples, see DNA Identification Act of 1994, 42 U.S.C. § 14133(c) (establishing a criminal penalty of up to $100,000 for knowing disclosure of "individually identifiable DNA information indexed in a database created or maintained by any Federal law enforcement agency" or for knowing receipt of "DNA samples or individually identifiable DNA information" in a federal database); Md. Ann. Code, Art. 88B, § 12(p) (199) (misdemeanor penalty of up to $1,000 fine and three years imprisonment).

252. Two blatant abuses of law enforcement and military intelligence files are described in Sam J. Ervin, Jr., The First Amendment: A Living Thought in the Computer Age, in Surveillance, Dataveillance, and Personal Freedoms 23, 38-40 (Columbia Human Rights Law Review ed., 1972).

256. Although the Whalen Court refused to elevate the interest in informational privacy to a level that would trigger strict scrutiny, it examined the system for maintaining the confidentiality of the prescription records in some detail:

The receiving room is surrounded by a locked wire fence and protected by an alarm system. The computer tapes containing the prescription data are kept in a locked cabinet. When the tapes are used, the computer is run "off-line," which means that no terminal outside of the computer room can read or record any information. Public disclosure of the identity of patients is expressly prohibited by the statute and by a Department of Health regulation. Willful violation of these prohibitions is a crime punishable by up to one year in prison and a $2,000 fine. At the time of trial there were 17 Department of Health employees with access to the files; in addition, there were 24 investigators with authority to investigate cases of overdispensing which might be identified by the computer.

257. For a brief discussion of some safeguards, see Katzenbach & Tomc, supra note 199, at 63-65.

258. Rhode Island's law is even more restrictive. It provides that "DNA samples and DNA records collected under this chapter shall never be used under the provisions of this chapter for the purpose of obtaining information about physical characteristics, traits or predispositions for disease." R.I. Gen. Laws § 12-1.5-10(4) (1998).

259. Records of credit card purchases, bank transactions, Internet use, and public library borrowing, for example, are much more invasive of personal privacy. But see United States v. Miller, 425 U.S. 435, 440 (1976) (reasoning that subpoenas to a bank for checks and deposit slips did not intrude "into any area in which [the defendant] had a protected Fourth Amendment interest" because the defendant had voluntarily relinquished these materials to the bank).

260. Likewise, one would hope that the destruction of identifiable samples would eliminate any public suspicion--implausible as it might be--that insurance companies or employers would entice personnel at state crime laboratories to violate state law by releasing DNA samples so that these organizations could use them to deny insurance coverage or employment to applicants. In many states, not only would the release of the information violate the DNA databanking laws, but the acquisition or use of it by insurers or employers would violate laws adopted to protect "genetic privacy" or to deter "genetic discrimination." See, e.g., Helen R. Davis & Janice V. Mitrius, Note, Recent Legislation on Genetics and Insurance, 37 Jurimetrics J. 69 (1996) (summarizing and tabulating the legislation); William F. Mullholland & Ami S. Jaeger, Comment, Genetic Privacy and Discrimination: A Survey of State Legislation, 39 Jurimetrics J. 317 (1999). Efforts to detect such uses of genetic information, both before and after the enactment of the state legislation, have not been notably successful. See Mark A. Hall, Legal Rules and Industry Norms: The Impact of Laws Restricting Health Insurers' Use of Genetic Information, 40 Jurimetrics J. 93 (1999); Philip R. Reilly, Genetic Discrimination, in Genetic Testing and the Use of Information 106 (Clarisa Long ed., 1999).

261. If an arrest is made in violation of the Fourth Amendment, the court may suppress evidence obtained as a result of the arrest. This could leave the prosecution with insufficient evidence with which to proceed.

262. Even if there is ample evidence to convict, the prosecution and defendant may have entered into a bargain which entailed the dismissal of the charges. Alternatively, the prosecution might have unilaterally dismissed the charges because it did not believe that it had enough evidence to warrant a trial. Even if the prosecution believed it had adequate evidence, the judge can grant a motion for a judgment of acquittal (see Fed. R. Crim. P. 29), or a jury can return a verdict of not guilty.

263. The United Kingdom takes DNA samples on arrests for "recordable" offenses, but destroys them if there is no subsequent conviction. Werrett, supra note 91.

266. A state expungement does not necessarily bind a federal agency or court. G.R. Dickerson v. New Banner Inst., Inc., 460 U.S. 103 (1983) (even though a state court had expunged a citizen's prior conviction, the firearms disabilities imposed by the 1968 Gun Control Act applied to the citizen).

267. E.g., Ark. Code Ann. § 16-90-906 (Michie 1997); Cal. Penal Code § 851.8 (West 1998); Fla. Stat. Ann. § 943.0585(1) (West 1999); La. Rev. Stat. Ann. § 614(A) (West 1998); Minn. Stat. Ann. § 609A.03 (West 1998); Michael D. Mayfield, Comment, Revisiting Expungement: Concealing Information in the Information Age, 1997 Utah L. Rev. 1057.

268. James W. Diehm, Federal Expungement: A Concept in Need of a Definition, 66 St. John's L. Rev. 73, 81 (1992).

271. See, e.g., Minn. Stat. Ann. § 609A.03 (West 1998) (denominating expungement "an extraordinary remedy to be granted only upon clear and convincing evidence that it would yield a benefit to the petitioner commensurate with the disadvantages to the public and public safety.").

272. Comment, supra note 267, at 1059-60 (citing Fla. Stat. Ann. § 943.0585(2)(h) (ten year waiting period), Pa. Cons. Stat. Ann. § 9122(b)(1) (the citizen "has been free of arrest or prosecution for ten years"), and Utah Code Ann. 77-18-12(2) (minimum time requirements)).

273. State v. Anonymous, 680 A.2d 956 (Conn. 1996); Commonwealth v. D.M., 695 A.2d 770 (Pa. 1997).

276. Doe v. Commander, Wheaton Police Dep't., 329 A.2d 35 (Md. 1974); Diehm, supra note 268, at 93; Comment, supra note 267, at 1057 n. 6.

278. There is no federal statute on point. 28 U.S.C. § 534(a) authorizes the Attorney General to acquire, preserve, and disseminate crime records. See United States v. Schnitzer, 567 F.2d 536, 539 (2d Cir. 1977).

288. Id. at 84 (citing United States v. Bagley, 899 F.2d 707, 708 (8th Cir. 1990), and Diamond v. United States, 649 F.2d 496, 497 (7th Cir. 1981)).

291. E.g., Maurer v. Pitchess, 691 F.2d 434, 437 (9th Cir. 1982) ("federal courts have inherent equitable power to order 'the expungement of local arrest records as an appropriate remedy in the wake of police action in violation of constitutional rights'"); Sullivan v. Murphy, 478 F.2d 938 (D.C. Cir. 1973) (if proven, the plaintiffs' allegation would demonstrate a violation of Fourth Amendment rights).

292. Cf. Hodge v. Jones, 31 F.3d 157 (4th Cir. 1994) (holding that, given the state's interest in maintaining a computerized database of investigations of child abuse, the constitution does not require the files of parents who had been investigated and cleared of child abuse charges to be removed from the database).

293. If the arrest passed constitutional muster, in an extraordinary case a federal court might nevertheless exercise its broad equitable discretion to order expungement. Diehm, supra note 268, at 81-84.

294. In some instances, law enforcement officials also might want these samples for a specific investigation. For example, they would be interest to the police if a patient in a hospital were raped and blood samples of all employees were stored in the hospital. Cf. Shirley E. Perlman, DNA Links Aide To Coma Rape, Newsday, Mar. 21, 1996, at A22, available in 1996 WL 2516248 (reporting that "a preliminary DNA test has linked a fired nurse's aide to a baby born this week to a 29-year-old brain-damaged nursing-home resident who was raped" and bore a premature infant).

298. E.g., Tims v. State, 711 So.2d 1118, 1124 (Ala. Crim. App. 1997) ("Under the Miller analysis, once the hospitals obtained the results for medical purposes, it would have been unreasonable for defendants to assume that the results would necessarily remain private. As the very least, various hospital employees become aware of the test results in the normal course of their work.").

299. See generally Amitai Etzioni, The Limits of Privacy (1999); Joseph W. Eaton, Card-Carrying Americans (1986).

300. A database that includes DNA data on arrestees as well as convicts is superior to a pure convicted-offender database Suppose that a rapist knows that the database includes only convicts' data. The rapist commits a rape and is arrested and charged with the rape, but the charges are dismissed or there is an acquittal. The rapist has escaped unpunished and realizes that if he commits another rape, the limited database will be of no assistance to the police in identifying him. As a mere arrestee, he will not be included in the database. By parallel reasoning, a universal database is more effective than one that merely includes DNA data on convicts, arrestees, or both.

302. Some critics of DNA databanking have complained that there is a significant risk of false identifications. Although we believe that this complaint is unfounded, to the extent that such a risk exists, universal databanking ensures that it does not fall disproportionately on any racial or other group. No group would be overrepresented in a universal database; every group would be represented to the same extent as its numbers in the general population.

303. See, e.g., George Annas, Privacy Rules for DNA Databanks: Protecting Coded 'Future Diaries,' 270 JAMA 2346 (1993) (opining that the "central problem" with comprehensive DNA databanking is that "this treats everyone in the United States (whose DNA is on file) as a crime suspect, making us a 'nation of suspects,' and radically alters the relationship between the citizen and government."); McEwen, supra note 28, at 236 ("[A] population-wide DNA data bank could fundamentally alter the relationship between individuals and the state, essentially turning us into a nation of suspects.").

306. Cf. Victor Walter Weedn, Stored Biologic Specimens for Military Identification: The Department of Defense DNA Registry, in Stored Tissue Samples: Ethical, Legal, and Public Policy Implications 345, 357 (Robert A. Weir ed., 1998):

There are innumerable scattered and diverse stores of biologic material in existence, all of which contain DNA and some of which are larger than the Armed Forces Repository . . . . Examples . . . include hospital pathology and microscope slides and tissue blocks, blood banks, sperm banks, organ and tissue banks, and research tissue stores. DNA is far more widespread and available than is commonly believed . . . . The protections and safeguards of these other sources of DNA are far less than those of the DoD program.

307. See, e.g., Dan L. Burk & Jennifer A. Hess, Genetic Privacy: Constitutional Considerations in Forensic DNA Testing, 5 Geo. Mason U. Civ. Rts. L.J. 1, 12 (1994).