OPINION

REINHARDT, Circuit Judge.

Each leap forward in forensic science promises ever more efficient and swift resolution of criminal investigations. At the same time, technological advances frequently raise new constitutional concerns and threaten our basic liberties. Here, we confront the challenge compulsory DNA collection poses to one of the most fundamental and traditional preserves of individual privacy, the human body.

We decide, in this case of first impression, whether the forced extraction of blood from parolees pursuant to the federal DNA Analysis Backlog Elimination Act of 2000, 42 U.S.C. § 14135a (the "DNA Act" or the "Act"), violates the Fourth Amendment. First, we must consider whether, under general Fourth Amendment principles, blood may be extracted from parolees without their consent, simply because of their status as parolees. We conclude that, as a matter of general Fourth Amendment law, forced blood extraction from parolees requires individualized suspicion. Second, we must determine whether forced blood extraction under the DNA Act falls within the exception of the Supreme Court's "special needs" doctrine. We hold that, because the DNA Act primarily serves a law enforcement purpose, the compulsory collection of blood samples pursuant to the Act does not fall within the special needs exception. Accordingly, we reverse the judgment of the district court (1) upholding the Probation Department's order requiring Thomas Kincade to submit to the extraction of blood for the purpose of providing a DNA sample, and (2) sentencing him to a term of imprisonment and increasing the period of his supervised release for his refusal to comply.

I. BACKGROUND

A. The Act

The DNA Act requires those in federal custody, on parole, on probation, or on supervised release to provide a DNA sample. As we explain later, for practical purposes, this requires all such persons to submit to the non-consensual withdrawal of blood by governmental authorities or their designees. No suspicion that an individual will commit or has committed another offense is required. Nor is there any requirement that the sample be taken in order to aid in the investigation of a particular crime. Once taken, the DNA sample is turned over to the Federal Bureau of Investigation, which carries out an analysis of it and includes the results in the "Combined DNA Index System" (CODIS), a DNA information bank. The DNA evidence is then permanently available for future use in connection with the investigation and prosecution of crimes. Federal, state, and local law enforcement officials who conduct such investigations are able to compare CODIS information with DNA evidence obtained from crime scenes and, thereby, to identify the perpetrator, and subject him to prosecution.

In 1994, Congress passed the Violent Crime Control and Law Enforcement Act, which authorized the FBI to establish a national index — CODIS — of DNA samples from convicted offenders, crime scenes, crime victims, and unidentified human remains. See DNA Act House Report, at 8; see also United States v. Reynard, 220 F.Supp.2d 1142, 1154 (S.D.Cal. 2002). Before the DNA Act was passed, all fifty states had adopted some form of legislation mandating DNA collection for inclusion into CODIS. See Nancy Beatty Gregoire, Federal Probation Joins the World of DNA Collection, 66 FED. PROBATION 30, 30 (2002). Between 1994 and 1996, however, no samples were collected from any persons convicted of federal crimes because the language of the 1994 act authorized only the creation of CODIS and not the collection of samples from convicted federal offenders. DNA Act House Report, at 8. In 1996, as part of the Anti-Terrorism and Effective Death Penalty Act ("AEDPA"), Pub.L. 104, 132, 110 Stat. 1214 (1996), Congress directed the FBI to "expand CODIS to include Federal crimes." DNA Act House Report, at 9. After AEDPA's passage, however, the Department of Justice concluded that it still did not confer upon federal law enforcement officials the requisite legal authority to collect DNA samples from federal offenders. Consequently, the DNA Act of 2000 was enacted, and now serves as the statutory basis for the forced extraction of blood samples from federal parolees, probationers, and prisoners. CODIS is currently a part of at least 137 laboratories throughout the country, and as of 2002, had provided forensic assistance in more than 1,900 investigations in 31 states. Gregoire, supra, at 30.

B. Factual and Procedural History

On September 1, 1993, Kincade pled guilty to one count of armed bank robbery and the use of a firearm. 18 U.S.C. §§ 2113(a), 924(c). On January 4, 1994, the district court sentenced him to 97 months, to be followed by a three-year term of supervised release. Included in the standard conditions of Kincade's supervised release is the obligation to "follow the instructions of the probation officer," and to refrain from committing "another Federal, state or local crime."

Kincade was released from prison on August 4, 2000. In March 2002, pursuant to the Act, the Probation Office ordered him to submit to a blood extraction for DNA analysis. He was subject to the DNA Act because the substantive offense to which he pled guilty, armed bank robbery, is one of the specified covered offenses. 42 U.S.C. § 14135a(d)(1)(E). Kincade refused to comply with the order, and his refusal was the basis for the Probation Office's recommendation to the district court that he be found in violation of his supervised release. The failure "to cooperate in the collection of[a] sample" under the Act is a class A misdemeanor. 42 U.S.C. § 14135a(a)(5)(A).

After conducting a hearing, the district court rejected Kincade's constitutional challenge and found that his refusal to submit to the compulsory blood extraction ordered by the Probation Office constituted a violation of the terms of his supervised release. The court sentenced Kincade to four months in custody for the violation and ordered that supervised release continue for a two-year term following release. The court stayed the order of custody pending this expedited appeal.

II. DISCUSSION

A. Blood Extraction Constitutes a Fourth Amendment Search

The DNA Act simply states that "[t]he probation office responsible for the supervision under Federal law of an individual on probation, parole, or supervised release shall collect a DNA sample from each such individual who is, or has been, convicted of a qualifying Federal offense." 42 U.S.C. § 14135a(a)(2). It does not prescribe any particular method for collecting the samples. Collection, however, is accomplished pursuant to a nationwide policy of compulsory blood extractions. The "DNA Collection Letter of Instruction," sent by the Central District of California Probation Office to Kincade, states that "[t]he Federal Bureau of Investigation (FBI) requires that DNA be obtained from blood samples." See also Gregoire, supra, at 31 (describing FBI requirement of blood sampling); Reynard, 220 F.Supp.2d at 1146 (describing procedures to collect and identify blood samples). Thus, Kincade challenges the Act on the basis of its standard method of implementation.

Blood extractions are searches for purposes of the Fourth Amendment, and are subject to the normal Fourth Amendment requirements. See Skinner, 489 U.S. at 616, 109 S.Ct. 1402 ("We have long recognized that a 'compelled intrusio[n] into the body for blood' . . . must be deemed a Fourth Amendment search."); Schmerber v. California, 384 U.S. 757, 767, 86 S.Ct. 1826, 16 L.Ed.2d 908 (1966) ("[Blood] testing procedures plainly constitute searches of 'persons' . . . within the meaning of [the Fourth] Amendment"); United States v. Wright, 215 F.3d 1020, 1025 (9th Cir. 2000) ("Intrusions into the human body, including the taking of blood, are searches subject to the restrictions of the Fourth Amendment."). In virtually every culture around the world, human blood possesses great symbolic power, and its spillage — whether in a drop or in a torrent — has carried enormous cultural significance. Throughout history, we have waged war, organized societies and religions, and created myths based upon the substance. See Dorothy Nelkin, Cultural Perspectives on Blood, in BLOOD FEUDS: AIDS, BLOOD, AND THE POLITICS OF MEDICAL DISASTER 273 (Eric A. Feldman Ronald Bayer eds., 1999). The formal policy, pursuant to the DNA Act, that all those covered by the legislation must submit to the compulsory extraction of blood samples unquestionably calls for a "search" within the meaning of the Fourth Amendment.

The government contends, however, that compulsory blood extraction under the Act is permissible because the method of data collection employed — the taking of blood — is no more intrusive than fingerprinting. We reject this false analogy.

First, the elision of two very different kinds of evidence obscures the constitutional difference between invasive procedures of the body that necessitate penetrating the skin, and an examination or recording of physical attributes that are generally exposed to public view. An individual cannot hold the same expectation of privacy for this latter category of information that he does for his internal properties, including blood. Cf. United States v. Dionisio, 410 U.S. 1, 5-6, 93 S.Ct. 764, 35 L.Ed.2d 67 (1973) ("It has long been held that the compelled display of identifiable physical characteristics infringes no interest protected by the privilege against compulsory self incrimination."). Although, as Pudd'nhead Wilson remarked, fingerprints do contain unique identifying information, they, like the human voice and the features of the face, are external to the individual. To obtain this identifying information requires no intrusive invasion of bodily privacy. By contrast, while DNA, like fingerprints, identifies an individual, DNA identification results from a forced intrusion into an individual's body. Fingerprinting, involving aspects of an individual's identity routinely exposed to public view, "represents a much less serious intrusion upon personal security than other types of searches and detentions." Hayes v. Florida, 470 U.S. 811, 814, 105 S.Ct. 1643, 84 L.Ed.2d 705 (1985).

Second, we note that even fingerprinting is not entirely free from the kind of Fourth Amendment concerns at stake here. Fingerprints taken pursuant to an arrest are part of so-called "booking" procedures, designed to ensure that the person who is arrested is in fact the person law enforcement officials believe they have in custody. See, e.g., Smith v. United States, 324 F.2d 879, 882 (D.C. Cir. 1963) ("[I]t is elementary that a person in lawful custody may be required to submit to photographing and fingerprinting as part of routine identification processes." (internal citations omitted)); Napolitano v. United States, 340 F.2d 313, 314 (1st Cir. 1965) (stating that taking of fingerprints upon admission to bail is "universally standard procedure, and no violation of constitutional rights"). This administrative procedure affirms that law enforcement has the right person in its custody. When law enforcement officials detain individuals for the purpose of obtaining fingerprints in furtherance of a criminal investigation, however, that detention violates the Fourth Amendment unless supported by probable cause or a warrant. Hayes, 470 U.S. at 815, 105 S.Ct. 1643. The Supreme Court so held even while acknowledging that fingerprints "involve neither repeated harassment nor any of the probing into private life and thoughts that often marks interrogation and search." Id. at 814, 105 S.Ct. 1643 (quoting Davis v. Mississippi, 394 U.S. 721, 727, 89 S.Ct. 1394, 22 L.Ed.2d 676 (1969) (holding detention for sole purpose of obtaining fingerprints without probable cause or warrant violates Fourth Amendment)). Here, of course, law enforcement does not question Kincade's true identity; it merely seeks to obtain evidence for future criminal investigations.

Thus, there can be no question that taking blood from a parolee against his will constitutes a search for Fourth Amendment purposes. We next turn to the question whether such a search, when conducted upon the body of a parolee, requires individualized suspicion.

B. Individualized Suspicion is Required for Searches of Parolees' Bodies

Although forced blood extractions constitute searches, Kincade can successfully challenge the requirement that he submit to that process only if his rights under the Fourth Amendment are violated. As a general rule, a search — even one that may lawfully be conducted without a warrant — must be based upon probable cause. New Jersey v. T.L.O., 469 U.S. 325, 340, 105 S.Ct. 733, 83 L.Ed.2d 720 (1985). In certain limited circumstances, however, the Fourth Amendment's demands of reasonableness have been held to be satisfied by a degree of suspicion less than probable cause, i.e., by reasonable suspicion. See id. at 340-41, 105 S.Ct. 733 (citing cases). Finally, in a discrete and limited category of cases not involving law enforcement purposes, see infra part II.C, searches without any degree of suspicion at all have been held to comport with Fourth Amendment reasonableness.

In United States v. Knights, 534 U.S. 112, 122 S.Ct. 587, 151 L.Ed.2d 497 (2001), the Supreme Court considered whether the search of the home of a probationer was reasonable under the Fourth Amendment. Police conducted a warrantless search of the probationer's home, pursuant to a probation condition requiring Knights to submit to searches without a warrant or "reasonable cause." 122 S.Ct. at 589. Using a "totality of the circumstances" analysis, the Supreme Court upheld the search as reasonable, with the probation condition being a "salient circumstance." Id. at 591. Noting that the search was in fact supported by reasonable suspicion, the Knights court held that "no more than" that degree of suspicion was required to search the probationer's home. Id. at 592. See also United States v. Stokes, 292 F.3d 964 (9th Cir. 2002) (determining that no more than reasonable suspicion is required to uphold warrantless search of probationer's car). The Knights Court did not reach the question whether searches of parolees and probationers could lawfully be conducted in the absence of reasonable suspicion.

Applying Knights' totality of the circumstances test to the case before us, we conclude that reasonable suspicion must exist before the government may compel parolees to submit to the extraction of blood from their bodies contrary to their wishes. The reasonableness of a parole search is determined by "'assessing, on the one hand, the degree to which it intrudes upon an individual's privacy and, on the other, the degree to which it is needed for the promotion of legitimate governmental interests.'" Knights, 122 S.Ct. at 591 (quoting Wyoming v. Houghton, 526 U.S. 295, 300, 119 S.Ct. 1297, 143 L.Ed.2d 408 (1999)).

We determine that the search in question constitutes a substantial intrusion on Kincade's legitimate expectation of privacy. As we explained in part II.A., supra, compulsory blood extractions generally are searches subject to customary Fourth Amendment probable cause requirements. The government asserts that the Fourth Amendment rights of parolees are "minimal," and, that they have no Fourth Amendment protection at all with respect to forced blood extractions. Whatever the accuracy or inaccuracy of the characterization of parolees' rights generally, the latter proposition is simply not correct. It is true that Kincade's parole status reduces the expectation of privacy that would otherwise be considered reasonable; however, while parolees enjoy lesser Fourth Amendment rights than other citizens, their rights are not extinguished. Even parolees maintain a reasonable expectation of privacy in their own bodies. The "integrity of an individual's person is a cherished value of our society," and a preeminent zone of constitutionally recognized privacy. Schmerber, 384 U.S. at 772, 86 S.Ct. 1826. Thus, although parole may reduce the degree of constitutional protection afforded an individual's body, it does not eradicate it.

In balancing the government's interests against the intrusion on Kincade's interests in privacy and bodily integrity, we consider the government's expressed interests in the search: to prevent, solve, and prosecute future crimes, and to complete the CODIS database. The government argues that its interests in the parole search must be evaluated in light of the statement in Griffin that probation is "a 'special need' of the State, permitting a degree of impingement upon privacy that would not be constitutional if applied to the public at large." 483 U.S. at 875, 107 S.Ct. 3164. The impingement that the Court referred to in Griffin, however, was a reduction in the degree of individualized justification required for the search, not the elimination of all need for such a basis. Specifically, the Court stated that the "special needs" associated with probation may "justify departures from the usual warrant and probable-cause requirements." Id. at 874, 107 S.Ct. 3164 (emphasis added). It did not suggest that the latter requirement could be eliminated entirely, or that the lesser standard of reasonable suspicion need not be met. To the contrary, the probation search that the Court upheld in Griffin was a search that the state justified on the basis of "reasonable grounds." 483 U.S. at 871, 107 S.Ct. 3164. Thus, Griffin did not authorize suspicionless searches.

More important for purposes of the present case, in Griffin the government's interest was directly related to the administration of parole itself. The purpose of the search, as in Knights, was to supervise parolees during the finite term of their parole period. It related to Griffin's conduct while in his status as a parolee. By contrast, the purpose of obtaining DNA samples is to obtain material for future use in a permanent DNA data bank to help solve future crimes, no matter how long after the end of a parole term they may be committed. Any use of the information to solve crimes committed during a parolee's term of supervision is fortuitous and incidental to the primary purpose of the Act. That purpose, according to the government, is to further "the overwhelming public interest in creating a comprehensive nationwide DNA bank that will improve the accuracy of criminal prosecutions" for generations to come. It is undoubtedly true that, were we to maintain DNA files on all persons living in this country, we would even more effectively further the public interest in having efficient and orderly criminal prosecutions, just as we would were we willing to sacrifice all of our interests in privacy and personal liberty. We chose, however, not to follow that course when we adopted the Fourth Amendment. Thus, it is the constitutional concept of Fourth Amendment reasonableness that governs our inquiry here, and not simply the governmental interest in solving and prosecuting crimes more efficiently. In short, the Government's desire to complete a comprehensive data bank does not outweigh Kincade's reasonable expectation of privacy in his body.

Finally, we weigh all of the above interests in light of the fact that, as discussed more fully in Section II.C., infra, under controlling Supreme Court authority we are not free to approve suspicionless searches conducted for law enforcement purposes. While weighing these interests could affect the degree of suspicion or cause required to conduct such searches, it could not serve to eliminate the requirement of individualized suspicion entirely.

Balancing the factors present here, and considering the applicable constitutional limitations, we conclude that, under general Fourth Amendment principles, individualized suspicion is required in searches of parolees conducted pursuant to the DNA Act. We hold that, under those principles, a search of a parolee's body to obtain DNA — the compulsory extraction of blood for a law enforcement purpose — is reasonable only if the search is supported by individualized reasonable suspicion.

Ordinarily, this would conclude our analysis of the Act, for there is no assertion that searches under the Act are conducted with any individualized suspicion whatsoever. However, the government next argues that, aside from the need to supervise the conduct of parolees during their parole period, the compulsory collection of DNA samples falls within the category of special needs cases in which the governmental conduct is exempted from the Fourth Amendment principles that are ordinarily applicable to searches and seizures. We thus now consider whether the special needs doctrine serves to exempt from the customary limitations of the Fourth Amendment the involuntary extraction of blood pursuant to the DNA Act.

C. The Special Needs Doctrine Does Not Exempt the Extraction of Blood from Parolees' Bodies from the Ordinary Requirements of the Fourth Amendment

The government argues that searches pursuant to the DNA Act, even if conducted without individualized suspicion, are constitutionally permissible because they fall within the "special needs" doctrine. In light of the Supreme Court's recent cases establishing a barrier between "special needs" searches and searches with law enforcement objectives, we reject that claim. We hold that searches pursuant to the DNA Act primarily serve a law enforcement purpose, and are not exempt from the ordinary requirements of the Fourth Amendment.

1. A Law Enforcement Purpose Is Not a Special Need

In "certain well-defined circumstances," Skinner, 489 U.S. at 619, 109 S.Ct. 1402, the Supreme Court has carved out a categorical exception even to the lesser Fourth Amendment requirement of individualized suspicion. This exception applies to a set of cases known as the "special needs" cases. In these special needs cases, searches that would otherwise violate the Fourth Amendment for lack of probable cause or individualized suspicion are deemed constitutionally permissible because they serve "special needs, beyond the normal need for law enforcement." City of Indianapolis v. Edmond, 531 U.S. 32, 37, 121 S.Ct. 447, 148 L.Ed.2d 333 (2000) (emphasis added). They are conducted for purposes other than the solving and punishing of crime. See, e.g., Bd. of Educ. v. Earls, 536 U.S. 822, 122 S.Ct. 2559, 153 L.Ed.2d 735 (2002) (urine testing of students for extracurricular activities to prevent health and safety risks from drug use); see id. at 833, 122 S.Ct. 2559 (emphasizing that "the test results are not turned over to any law enforcement authority"); Vernonia Sch. Dist. 47J v. Acton, 515 U.S. 646, 115 S.Ct. 2386, 132 L.Ed.2d 564 (1995) (random urine testing of student athletes to prevent injury and drug dependency); Skinner, 489 U.S. 602, 109 S.Ct. 1402, 103 L.Ed.2d 639 (blood and urine tests of railroad employees to prevent railway accidents); Nat'l Treasury Employees Union v. Von Raab, 489 U.S. 656, 109 S.Ct. 1384, 103 L.Ed.2d 685 (1989) (urine tests of U.S. Customs Service employees seeking transfer or promotion to insure the officials' fitness to interdict drugs and handle firearms); United States v. Gonzalez, 300 F.3d 1048 (9th Cir. 2002) (searches of employee backpacks to prevent inventory loss). The Court also has permitted suspicionless searches in certain roadway checkpoint programs, Mich. Dep't of State Police v. Sitz, 496 U.S. 444, 110 S.Ct. 2481, 110 L.Ed.2d 412 (1990) (highway sobriety checkpoints for public safety), and has authorized routine searches absent individualized suspicion at the national border. See United States v. Montoya de Hernandez, 473 U.S. 531, 105 S.Ct. 3304, 87 L.Ed.2d 381 (1985) (routine border searches to prevent entry of contraband); United States v. Martinez-Fuerte, 428 U.S. 543, 96 S.Ct. 3074, 49 L.Ed.2d 1116 (1976) (fixed checkpoint routine border search to deter illegal immigration). Finally, the Court has approved limited searches for administrative purposes without individualized suspicion. See, e.g., New York v. Burger, 482 U.S. 691, 107 S.Ct. 2636, 96 L.Ed.2d 601 (1987) (purely administrative search of "closely regulated" business); Camara v. Municipal Court, 387 U.S. 523, 87 S.Ct. 1727, 18 L.Ed.2d 930 (1967) (residential building code inspections to prevent hazardous conditions); see also McMorris v. Alioto, 567 F.2d 897 (9th Cir. 1978) (purely administrative search in public buildings).

a. The Effect of Edmond and Ferguson

In City of Indianapolis v. Edmond, 531 U.S. 32, 121 S.Ct. 447, 148 L.Ed.2d 333 (2000), the Court considered a Fourth Amendment challenge to an Indianapolis highway checkpoint program that authorized suspicionless searches for the discovery and interdiction of illegal narcotics. Rejecting the contention that the Indianapolis program fell within "the limited circumstances in which the usual rule [of the Fourth Amendment] does not apply," 531 U.S. at 37, 121 S.Ct. 447, the Court stated that:

[T]he gravity of the threat alone cannot be dispositive of questions concerning what means law enforcement officers may employ to pursue a given purpose. Rather, in determining whether individualized suspicion is required, we must consider the nature of the interests threatened and their connection to the particular law enforcement practices at issue. We are particularly reluctant to recognize exceptions to the general rule of individualized suspicion where governmental authorities primarily pursue their general crime control ends.

Id. at 42-43, 121 S.Ct. 447 (emphasis added). The Court also explained that Mich. Dep't of State Police v. Sitz, 496 U.S. 444, 110 S.Ct. 2481, 110 L.Ed.2d 412 (1990), in which it upheld a state's use of highway sobriety checkpoints, did not stand for the proposition that law enforcement needs could render legitimate a program of suspicionless searches. Id. at 39, 110 S.Ct. 2481. Rather, Edmond made it plain that the checkpoint program at issue in Sitz was designed to reduce "the immediate hazard posed by the presence of drunk drivers on the highways." Id. The Court in Edmond then cautioned that it had "never approved [a general program of suspicionless seizures] whose primary purpose was to detect evidence of ordinary criminal wrongdoing." Id. at 41, 121 S.Ct. 447 (emphasis added).

Thus, in Edmond, the Court refused to permit a "'general interest in crime control' as justification for a regime of suspicionless stops." Id. (quoting Delaware v. Prouse, 440 U.S. 648, 659, 99 S.Ct. 1391, 59 L.Ed.2d 660 (1979)). Allowing use of the special needs doctrine in cases in which the justification for the search was "the general interest in crime control" would "do little to prevent such intrusions from becoming a routine part of American life." Id. at 42, 121 S.Ct. 447. Where "law enforcement authorities pursue primarily general crime control purposes," the Court cautioned, searches and seizures "can only be justified by some quantum of individualized suspicion." Id. at 47, 121 S.Ct. 447.

The Supreme Court reaffirmed its Edmond holding a year later in Ferguson v. City of Charleston, 532 U.S. 67, 121 S.Ct. 1281, 149 L.Ed.2d 205 (2001). In that case, petitioners challenged a state hospital program that tested pregnant women for drug use and then made available to the police the results of these tests if a woman tested positive twice. 532 U.S. at 72, 121 S.Ct. 1281. The Court held that while a significant goal of the program was to aid women with drug abuse problems, "the immediate objective of the [suspicionless] searches was to generate evidence for law enforcement purposes," id. at 83, 121 S.Ct. 1281 (emphasis in original), and, accordingly, tests conducted for purposes of the program violated the Fourth Amendment. Prosecutors and police were heavily involved in both the program's conception and implementation. Id. at 88, 121 S.Ct. 1281. As the Court emphasized, "[i]n none of our previous special needs cases have we upheld the collection of evidence for criminal law enforcement purposes. . . . [T]he extensive entanglement of law enforcement cannot be justified by reference to legitimate needs." Id. at 84 n. 20, 121 S.Ct. 1281. Justice Kennedy, concurring in the judgment, agreed that the

special needs cases we have decided do not sustain the active use of law enforcement, including arrest and prosecutions, as an integral part of a program which seeks to achieve legitimate, civil objectives. The traditional warrant and probable-cause requirements are waived in our previous cases on the explicit assumption that the evidence obtained in the search is not intended to be used for law enforcement purposes.

Id. at 88, 121 S.Ct. 1281 (Kennedy, J., concurring in the judgment).

Thus, both Ferguson and Edmond reaffirm that the special needs doctrine applies to a narrow category of cases to which traditional Fourth Amendment requirements are inapplicable. In those two decisions, the Supreme Court made explicit its underlying premise that special needs cases qualify for an exception to the Fourth Amendment's customary requirements only because they involve programs or activities "not designed to serve the ordinary needs of law enforcement." Von Raab, 489 U.S. at 679, 109 S.Ct. 1384; see also Earls, 536 U.S. at 833, 122 S.Ct. 2559 (noting that "the [urine] test results are not turned over to any law enforcement authority"); Vernonia, 515 U.S. at 658, 115 S.Ct. 2386 (stating that "the results of the tests are disclosed only to a limited class of school personnel who have a need to know; and they are not turned over to law enforcement authorities or used for any internal disciplinary function"); Skinner, 489 U.S. at 620-621, 109 S.Ct. 1402 (noting that toxicological tests intended "to prevent accidents," "not to assist in the prosecution of employees"); Von Raab, 489 U.S. at 666, 109 S.Ct. 1384 ("Test results may not be used in a criminal prosecution of the employee without the employee's consent."). Forcing a parolee like Kincade to submit to blood extraction against his will, and without any constitutionally cognizable level of suspicion for the law enforcement purpose here involved, is not analogous to testing, for example, persons who voluntarily participate in a highly regulated endeavor, such as students who engage in organized sports or extracurricular activities, Earls, 536 U.S. at 830-31, 122 S.Ct. 2559, or federal Customs employees who seek promotion to certain sensitive positions, Von Raab, 489 U.S. at 671, 109 S.Ct. 1384. See also Vernonia, 515 U.S. at 657, 115 S.Ct. 2386 ("By choosing to 'go out for the team,' [these students] voluntarily subject themselves to a degree of regulation even higher than that imposed on students generally." (emphasis added)).

The government's primary response to Edmond and Ferguson is to argue that Rise v. Oregon, 59 F.3d 1556 (9th Cir. 1995), resolves the case before us. In Rise, we held that a state DNA collection statute, similar in implementation to the one we consider here, comported with Fourth Amendment reasonableness. Although a panel of this court cannot ordinarily overrule Ninth Circuit precedent, United States v. Gay, 967 F.2d 322, 327 (9th Cir. 1992), we nevertheless in appropriate circumstances may, indeed must, reconsider past cases in light of an intervening decision of the Supreme Court that undermines that earlier precedent. "[T]he issues decided by the higher court need not be identical in order to be controlling. Rather, the relevant court of last resort must have undercut the theory or reasoning underlying the prior circuit precedent in such a way that the cases are clearly irreconcilable." Miller v. Gammie, 335 F.3d 889, 900 (9th Cir. 2003) (en banc). At that point, "a three-judge panel should consider itself bound by the later and controlling authority, and should reject the prior circuit opinion as having been effectively overruled." Id. at 893. See also Galbraith v. County of Santa Clara, 307 F.3d 1119, 1123 (9th Cir. 2002); United States v. Lancellotti, 761 F.2d 1363, 1366 (9th Cir. 1985); Spinelli v. Gaughan, 12 F.3d 853, 855 n. 1 (9th Cir. 1993).

Rise relied on an interpretation of Supreme Court precedent that has since been wholly discredited by Edmond, 531 U.S. at 44, 121 S.Ct. 447, and further undermined by Ferguson, 532 U.S. at 84, 121 S.Ct. 1281; Rise and Edmond/Ferguson are clearly irreconcilable. Rise rejected a Fourth Amendment challenge to a state DNA collection statute; we stated at the outset, citing Sitz, that "[e]ven in the law enforcement context," suspicionless searches could be conducted if the intrusion was "minimal" and was "justified by law enforcement purposes." 59 F.3d at 1559. Indeed, Rise was premised on the understanding that the searches it endorsed were law enforcement searches conducted for the primary purpose of controlling crime; we assumed that Sitz also involved such searches, and followed what we believed to be Sitz's example, concluding that searches for law enforcement purposes could proceed even absent individualized suspicion. As we have explained, however, Edmond expressly repudiated this interpretation of Sitz. The Court stated that the suspicionless searches in Sitz were justified because they were searches to prevent an immediate threat to highway safety by removing drunk drivers from the road, not searches to further "a general purpose of investigating crime." Edmond, 531 U.S. at 39, 121 S.Ct. 447. Furthermore, Edmond — and then Ferguson — made clear that the Court would "decline to suspend the usual requirement of individualized suspicion where the police seek to employ a [search] primarily for the ordinary enterprise of investigating crimes." Id. at 44, 121 S.Ct. 447. Rise, however, did precisely that. It approved suspicionless searches designed to further law enforcement objectives. Because Rise depended on reasoning and reached a result that the Supreme Court has since expressly disavowed, "its precedential effect has been dissipated by the Supreme Court's intervening ruling[s]," and thus, in deciding Kincade's challenge to searches pursuant to the Act, we do not apply it. Grunwald v. San Bernardino City Unified Sch. Dist., 994 F.2d 1370, 1375 n. 4 (9th Cir. 1993). In fact, were we to follow Rise, we would be acting in disregard of controlling Supreme Court doctrine.

The government suggests, however, that Edmond and Ferguson do not apply to searches pursuant to the Act, because both Supreme Court cases dealt with programs affecting "law abiding citizens," while the Act addresses "probationers and felons." In neither case did the Court suggest that its exclusion of law enforcement purposes from the special needs context applies only in the case of law-abiding persons. Rather, these two decisions stand for the proposition that a program of suspicionless searches, conducted for law enforcement purposes, violates the Fourth Amendment, whether the person searched is a model citizen (and thus the search will produce no useful evidence), or whether he is not law-abiding (and as a result evidence of a crime may sometimes be obtained). Indeed, it would be a unique construction of the Constitution to hold that Fourth Amendment protections exist only for the benefit of the innocent, and not for all persons in our society, regardless of one's propensity to engage in criminal activity.

b. Other Decisions

No circuit court has heretofore given full consideration to the question of the constitutionality of the DNA Act. The government cites decisions from other circuits rejecting Fourth Amendment challenges to searches pursuant to state statutes similar to the Act. Nearly all of these cases involving state statutes were decided before Edmond and Ferguson, and a number of them rely on the now-repudiated Rise. See, e.g., Velasquez v. Woods, 329 F.3d 420 (5th Cir. 2003) (relying in part on Rise to affirm the dismissal, as frivolous, of a Fourth Amendment challenge to Texas DNA collection statute); Roe v. Marcotte, 193 F.3d 72 (2d Cir. 1999) (relying in part on Rise to reject Fourth Amendment challenge to Connecticut DNA collection statute); Shaffer v. Saffle, 148 F.3d 1180 (10th Cir. 1998) (relying on Boling, which relied in part on Rise, to dismiss Fourth Amendment challenge to Oklahoma DNA collection statute); Boling v. Romer, 101 F.3d 1336 (10th Cir. 1996) (relying in part on Rise to reject Fourth Amendment challenge to Colorado DNA collection statute); Schlicher v. Peters, 103 F.3d 940 (10th Cir. 1996) (relying on Rise to deny Fourth Amendment challenge to Kansas DNA statute); Jones v. Murray, 962 F.2d 302 (4th Cir. 1992) (finding Virginia DNA statute reasonable and rejecting Fourth Amendment challenge). Significantly, while the two other district court decisions in this circuit of which we are aware did address the constitutionality of the DNA Act, and arrived at directly opposite conclusions, both stated that Rise's rationale does not survive Edmond and Ferguson. See Reynard, 220 F.Supp.2d at 1166 n. 29 (stating that Rise "potentially conflicts with recent Supreme Court decisions"); United States v. Miles, 228 F.Supp.2d 1130, 1135 (E.D.Cal. 2002) (stating that Edmond and Ferguson have "effectively overruled Rise").

2. The DNA Act Serves a Law Enforcement Purpose

Next, the government contends, rather disingenuously, that the forced extraction of blood samples from Kincade and others similarly situated does not serve a law enforcement purpose. As we stated earlier, Edmond and Ferguson make plain that a regime of suspicionless searches is constitutionally impermissible if it is designed to serve such a purpose.

As an initial matter, the government, citing Knights, 122 S.Ct. at 593, argues that in reviewing the purpose of searches pursuant to the Act, we may not question the "official purpose" of the search. The government misreads Knights and misapplies it to the Act. In Knights, the Supreme Court, citing Whren v. United States, 517 U.S. 806, 116 S.Ct. 1769, 135 L.Ed.2d 89 (1996), expressed its unwillingness to delve into the subjective motivations of the detective who searched Knights's apartment. 517 U.S. at 593, 116 S.Ct. 1589. In Edmond, however, the Supreme Court explicitly distinguished inquiries into the motivations of a single law enforcement official from those regarding official purpose underlying a general policy mandating suspicionless searches. "[O]ur cases dealing with intrusions that occur pursuant to a general scheme absent individualized suspicion have often required an inquiry into purpose at the programmatic level." 531 U.S. at 46, 121 S.Ct. 447. The portion of the Act before us here mandates DNA collection of all persons convicted of a qualifying offense. 42 U.S.C. § 14135a(2). Kincade makes no argument about the subjective intent of any individual parole officer who would forcibly extract a blood sample from him. Thus, Whren and its progeny have no application here.

a. The Act's Enforcement and Legislative History Show a Clear Law Enforcement Purpose

In determining the "programmatic purpose" of searches pursuant to the Act, "we consider all the available evidence in order to determine the relevant primary purpose." Ferguson, 532 U.S. at 81, 121 S.Ct. 1281. The government argues that two purposes of searches pursuant to the Act are to "help law enforcement solve unresolved and future cases," and to increase accuracy in the criminal justice system. The government's own argument establishes that prototypical law enforcement purposes underlie the DNA searches in question. Under the government's own theory, the searches are conducted in order to collect DNA evidence samples for CODIS, so that those samples may be used in criminal investigations, to help solve crimes and prosecute the culprits, and to enable law enforcement agencies to be more accurate and effective in achieving their law enforcement objectives.

Moreover, a review of the Act's enforcement and its legislative history makes it equally plain that searches pursuant to the Act serve a law enforcement purpose. First, while probation officers initially collect the samples, law enforcement officers, as in the South Carolina pre-natal drug testing program at issue in Ferguson, are "extensively involved" thereafter. Ferguson, 532 U.S. at 82, 121 S.Ct. 1281. Like the program at issue in Ferguson, suspicionless searches conducted pursuant to the Act evince "a penal character with a far greater connection to law enforcement than other searches sustained under [the] special needs rationale." Id. at 88-89, 121 S.Ct. 1281 (Kennedy, J., concurring in the judgment). Under the Act, once Kincade's compulsorily extracted blood sample is included in CODIS, it has no other purpose than to aid law enforcement in matching his DNA to that which may be found at crime scene investigations, and thus to help determine whether Kincade is the person who has committed certain crimes and, if so, to facilitate his prosecution and conviction.

Next, it is clear that in passing the Act mandating these suspicionless searches, Congress was primarily concerned with the swift and accurate solution and prosecution of crimes generally; the legislative history is replete with references to the utility of DNA evidence in prosecuting crimes. See DNA Act House Report, at 8-11, 23-27, 32-36 (2000). For example, in addressing whether or not the DNA information collected would be used for insurance or medical purposes, the Department of Justice assured the House Committee that "existing legal rules for the DNA identification system generally ensure that DNA samples and indexed information will be used solely for law enforcement identification purposes." Id. at 25 (emphasis added); see also DNA Analysis Backlog Elimination Act of 2000, 146 Cong. Rec. S11645-02, at S11647 (2000) ("Each day that DNA evidence goes uncollected and untested, solvable crimes remain unsolved, and people across the country are needlessly victimized." (quoted in Reynard, 220 F.Supp.2d at 1156 n. 15)).

It is also apparent that the executive branch understands law enforcement to be the primary objective of searches pursuant to the Act. See, e.g., Dep't of Justice, Using DNA to Solve Cold Cases 4 (July 2002) (stating that the DNA database system is a "powerful tool for law enforcement"); Dep't of Justice, No Suspect Casework DNA Backlog Reduction Program (FY 2001), at 1 (August 2001) ("DNA evidence used in conjunction with the Combined DNA Index System (CODIS) is a powerful investigative tool beginning at the crime scene with the collection of evidence and ending with a judicial conclusion."); see also Justice Dep't. Acts to Clear DNA Backlog, MIAMI HERALD, Aug. 2, 2001, at 19A (quoting Attorney General Ashcroft as saying "DNA technology can operate as a kind of truth machine, ensuring justice by identifying the guilty and clearing the innocent."). In sum, the record overwhelmingly demonstrates that the Act mandating these suspicionless searches was enacted, is enforced, and is understood by all concerned to serve the purpose of law enforcement, and to further its objectives.

b. The "Immediate Purpose" of the Act is Law Enforcement

The government also asserts that even if the "ultimate objective" of searches pursuant to the Act is law enforcement, their "immediate purpose" is to "fill a gap in the CODIS database." In Ferguson, the Court held that the hospital's testing policy was not justified by a special need because its immediate purpose was law enforcement, even if its ultimate purpose was to help those mothers who were substance abusers and their children. 532 U.S. at 82-84, 121 S.Ct. 1281. Here, the government's explanation of the searches' immediate purpose is in no way supported by Ferguson's distinction. First, Ferguson said that the government could not disclaim a law enforcement purpose by pointing to an ultimate purpose and ignoring an immediate purpose. It did not suggest that the reverse would be any less objectionable — that the courts could ignore the ultimate law enforcement purpose if the government could point to an immediate other purpose. To the contrary, if either the immediate or the ultimate objective serves a law enforcement purpose, the special needs doctrine is inapplicable. Second, the immediate purpose of searches pursuant to the Act, as we have understood the "intent" of legislation in other cases, is to gather evidence for criminal investigation, not to put evidence into a database. The purpose of these searches is no more to put samples into CODIS than is the purpose of finger-printing to place cards into index files. Accordingly, we reject the government's suggested distinction. Both the "immediate purpose" and the "ultimate objective" of searches under the Act are to further law enforcement ends.

c. The Possibility of Exoneration

Finally, the government suggests that searches pursuant to the Act not only help to convict the guilty but also serve the commendable purpose of ensuring that the innocent will not be wrongly convicted. We would hope so. However, even if we were to assume that the clearing of the innocent is not a function of law enforcement — and that would be a troubling assumption, indeed — exoneration would still not serve to supplant the primary law enforcement objective of these searches — the solving of crimes and the prosecution of those responsible. The presence of a "benign" motive cannot "justify a departure from Fourth Amendment protections, given the pervasive involvement of law enforcement." Ferguson, 532 U.S. at 85, 121 S.Ct. 1281. Otherwise, a regime of suspicionless searches for law enforcement purposes would nearly always be permissible. See id. at 84 n. 22, 121 S.Ct. 1281 ("[U]nder respondents' approach, any search to generate evidence for use by the police in enforcing general criminal laws would be justified by reference to the broad social benefits that those laws might bring about (or, put another way, the social harms that they might prevent)."). The Supreme Court has emphasized that the special needs cases provide a narrow exception to the ordinary Fourth Amendment requirements, not a convenient means by which to avoid the strictures of the Constitution.

Recent experience has proven the efficacy of DNA testing to exonerate the wrongfully convicted, and we do not doubt the importance of DNA collection for this worthwhile purpose. Those who claim wrongful conviction or even wrongful accusation, however, may volunteer their DNA for this purpose; equally important, an act that required the states and federal government to collect and analyze the DNA when requested by such persons would not offend the Fourth Amendment. To the contrary, such an act would well serve some of the objectives the Department of Justice endorses here. The DNA Act, as presently constituted, however, provides no choice to those from whom it requires that DNA samples be collected, and no option to others not covered by the Act who might be able to benefit greatly from its provisions. Accordingly, it is difficult to accept the government's representation of its concerns regarding the innocent. See also Miles, 228 F.Supp.2d at 1139 ("It is disingenuous for the government to state that it needs to exonerate people who do not want to be exonerated.").

Whatever benign secondary purposes these searches may happen to serve, the primary purpose is to provide law enforcement officials, both at the state and federal level, with information about individuals that can be used to identify them as criminals and to prosecute them for their crimes. Kincade, should he be subjected to such a search, in effect will have been compelled to provide evidence with respect to any and all crimes of which he may be accused, for the rest of his life. Kincade has completed his sentence, and the court has ordered a term of supervised release for three years. The forced extraction of his blood with the subsequent categorization of his DNA would affect him for the rest of his life. Thus, he is entitled to the protection afforded by the Fourth Amendment.

III. CONCLUSION

We hold that forced blood extractions pursuant to the DNA Analysis Backlog Elimination Act of 2000 violate the Fourth Amendment because they constitute suspicionless searches with the objective of furthering law enforcement purposes. Compulsory searches of the bodies of parolees such as Kincade require, at a minimum, reasonable suspicion. Accordingly, we reverse the judgment of the district court, with instructions to vacate the order revoking Kincade's supervised release.

However intermingled with good intentions, DNA statutes, like the thermal imaging procedure struck down in Kyllo, 533 U.S. at 34-41, 121 S.Ct. 2038, represent an

alarming trend whereby the privacy and dignity of our citizens [are] being whittled away by imperceptible steps. Taken individually, each step may be of little consequence. But when viewed as a whole, there begins to emerge a society quite unlike any we have seen — a society in which government may intrude into the secret regions of man's life at will.

Osborn v. United States, 385 U.S. 323, 343, 87 S.Ct. 429, 17 L.Ed.2d 394 (1966) (Douglas, J., dissenting). The fact that these statutes currently affect only those individuals most susceptible to state supervision renders this threat no less important. Under the government's view of the "special needs" doctrine, the rest of us might not be far behind. Privacy erodes first at the margins, but once eliminated, its protections are lost for good, and the resultant damage cannot be undone.

REVERSED and REMANDED.