STEVENS, J., delivered the opinion of the Court, in which SCALIA, SOUTER, THOMAS, and GINSBURG, JJ., joined:
After Rodney Gant was arrested for driving with a suspended license, handcuffed, and locked in the back of a patrol car, police officers searched his car and discovered cocaine in the pocket of a jacket on the backseat. Because Gant could not have accessed his car to retrieve weapons or evidence at the time of the search, the Arizona Supreme Court held that the search-incident-to-arrest exception to the Fourth Amendment’s warrant requirement, as defined in Chimel v. California, 395 U.S. 752 (1969), and applied to vehicle searches in New York v. Belton, 453 U.S. 454 (1981), did not justify the search in this case. We agree with that conclusion.
Under Chimel, police may search incident to arrest only the space within an arrestee’s “‘immediate control,’” meaning “the area from within which he might gain possession of a weapon or destructible evidence.” 395 U.S., at 763. The safety and evidentiary justifications underlying Chimel‘s reaching-distance rule determine Belton‘s scope. Accordingly, we hold that Belton does not authorize a vehicle search incident to a recent occupant’s arrest after the arrestee has been secured and cannot access the interior of the vehicle. Consistent with the holding in Thornton v. United States, 541 U.S. 615 (2004), and following the suggestion in JUSTICE SCALIA’s opinion concurring in the judgment in that case, id., at 632, we also conclude that circumstances unique to the automobile context justify a search incident to arrest when it is reasonable to believe that evidence of the offense of arrest might be found in the vehicle.
. . . .
In Chimel, we held that a search incident to arrest may only include “the arrestee’s person and the area ‘within his immediate control’ — construing that phrase to mean the area from within which he might gain possession of a weapon or destructible evidence.” Ibid. That limitation, which continues to define the boundaries of the exception, ensures that the scope of a search incident to arrest is commensurate with its purposes of protecting arresting officers and safeguarding any evidence of the offense of arrest that an arrestee might conceal or destroy. See ibid. (noting that searches incident to arrest are reasonable “in order to remove any weapons [the arrestee] might seek to use” and “in order to prevent [the] concealment or destruction” of evidence (emphasis added)). If there is no possibility that an arrestee could reach into the area that law enforcement officers seek to search, both justifications for the search-incident-to-arrest exception are absent and the rule does not apply. E.g., Preston v. United States, 376 U.S. 364, 367-368 (1964).
[In Belton] we held that when an officer lawfully arrests “the occupant of an automobile, he may, as a contemporaneous incident of that arrest, search the passenger compartment of the automobile” and any containers therein. Belton, 453 U.S., at 460 (footnote omitted). That holding was based in large part on our assumption “that articles inside the relatively narrow compass of the passenger compartment of an automobile are in fact generally, even if not inevitably, within ‘the area into which an arrestee might reach.’” Ibid.
. . . .
. . . [O]ur opinion [in Belton] has been widely understood to allow a vehicle search incident to the arrest of a recent occupant even if there is no possibility the arrestee could gain access to the vehicle at the time of the search. This reading may be attributable to Justice Brennan’s dissent in Belton, in which he characterized the Court’s holding as resting on the “fiction . . . that the interior of a car is always within the immediate control of an arrestee who has recently been in the car.” 453 U.S., at 466. Under the majority’s approach, he argued, “the result would presumably be the same even if [the officer] had handcuffed Belton and his companions in the patrol car” before conducting the search. Id., at 468.
Since we decided Belton, Courts of Appeals have given different answers to the question whether a vehicle must be within an arrestee’s reach to justify a vehicle search incident to arrest, [footnote omitted] but Justice Brennan’s reading of the Court’s opinion has predominated. As Justice O’Connor observed, “lower court decisions seem now to treat the ability to search a vehicle incident to the arrest of a recent occupant as a police entitlement rather than as an exception justified by the twin rationales of Chimel.” Thornton, 541 U.S., at 624 (opinion concurring in part). JUSTICE SCALIA has similarly noted that, although it is improbable that an arrestee could gain access to weapons stored in his vehicle after he has been handcuffed and secured in the backseat of a patrol car, cases allowing a search in “this precise factual scenario . . . are legion.” Id., at 628 (opinion concurring in judgment) (collecting cases). [Footnote omitted.] Indeed, some courts have upheld searches under Belton “even when . . . the handcuffed arrestee has already left the scene.” 541 U.S., at 628 (same).
Under this broad reading of Belton, a vehicle search would be authorized incident to every arrest of a recent occupant notwithstanding that in most cases the vehicle’s passenger compartment will not be within the arrestee’s reach at the time of the search. To read Belton as authorizing a vehicle search incident to every recent occupant’s arrest would thus untether the rule from the justifications underlying the Chimel exception — a result clearly incompatible with our statement in Belton that it “in no way alters the fundamental principles established in the Chimel case regarding the basic scope of searches incident to lawful custodial arrests.” 453 U.S., at 460, n. 3. Accordingly, we reject this reading of Belton and hold that the Chimel rationale authorizes police to search a vehicle incident to a recent occupant’s arrest only when the arrestee is unsecured and within reaching distance of the passenger compartment at the time of the search.Â [Footnote omitted.]
Although it does not follow from Chimel, we also conclude that circumstances unique to the vehicle context justify a search incident to a lawful arrest when it is “reasonable to believe evidence relevant to the crime of arrest might be found in the vehicle.” Thornton, 541 U.S., at 632 (SCALIA, J., concurring in judgment). In many cases, as when a recent occupant is arrested for a traffic violation, there will be no reasonable basis to believe the vehicle contains relevant evidence. See, e.g., Atwater v. Lago Vista, 532 U.S. 318, 324 (2001); Knowles v. Iowa, 525 U.S. 113, 118 (1998). But in others, including Belton and Thornton, the offense of arrest will supply a basis for searching the passenger compartment of an arrestee’s vehicle and any containers therein.
Neither the possibility of access nor the likelihood of discovering offense-related evidence authorized the search in this case. Unlike in Belton, which involved a single officer confronted with four unsecured arrestees, the five officers in this case outnumbered the three arrestees, all of whom had been handcuffed and secured in separate patrol cars before the officers searched Gant’s car. Under those circumstances, Gant clearly was not within reaching distance of his car at the time of the search. An evidentiary basis for the search was also lacking in this case. Whereas Belton and Thornton were arrested for drug offenses, Gant was arrested for driving with a suspended license — an offense for which police could not expect to find evidence in the passenger compartment of Gant’s car. Cf. Knowles, 525 U.S., at 118. Because police could not reasonably have believed either that Gant could have accessed his car at the time of the search or that evidence of the offense for which he was arrested might have been found therein, the search in this case was unreasonable.
We do not agree with the contention in JUSTICE ALITO’s dissent . . . that consideration of police reliance interests requires a different result. Although it appears that the State’s reading of Belton has been widely taught in police academies and that law enforcement officers have relied on the rule in conducting vehicle searches during the past 28 years, many of these searches were not justified by the reasons underlying the Chimel exception. Countless individuals guilty of nothing more serious than a traffic violation have had their constitutional right to the security of their private effects violated as a result. The fact that the law enforcement community may view the State’s version of the Belton rule as an entitlement does not establish the sort of reliance interest that could outweigh the countervailing interest that all individuals share in having their constitutional rights fully protected. If it is clear that a practice is unlawful, individuals’ interest in its discontinuance clearly outweighs any law enforcement “entitlement” to its persistence.Â . . . .
JUSTICE SCALIA, concurring:
JUSTICE STEVENS acknowledges that an officer-safety rationale cannot justify all vehicle searches incident to arrest, but asserts that that is not the rule Belton and Thornton adopted. . . . JUSTICE STEVENS would therefore retain the application of Chimel v. California, 395 U.S. 752 (1969), in the car-search context but would apply in the future what he believes our cases held in the past: that officers making a roadside stop may search the vehicle so long as the “arrestee is within reaching distance of the passenger compartment at the time of the search.” Ante, at 18. I believe that this standard fails to provide the needed guidance to arresting officers and also leaves much room for manipulation, inviting officers to leave the scene unsecured (at least where dangerous suspects are not involved) in order to conduct a vehicle search. In my view we should simply abandon the Belton-Thornton charade of officer safety and overrule those cases. I would hold that a vehicle search incident to arrest is ipso facto “reasonable” only when the object of the search is evidence of the crime for which the arrest was made, or of another crime that the officer has probable cause to believe occurred. Because respondent was arrested for driving without a license (a crime for which no evidence could be expected to be found in the vehicle), I would hold in the present case that the search was unlawful.
. . . .
No other Justice . . . shares my view that application of Chimel in this context should be entirely abandoned. It seems to me unacceptable for the Court to come forth with a 4-to-1-to-4 opinion that leaves the governing rule uncertain. I am therefore confronted with the choice of either leaving the current understanding of Belton and Thornton in effect, or acceding to what seems to me the artificial narrowing of those cases adopted by JUSTICE STEVENS. The latter, as I have said, does not provide the degree of certainty I think desirable in this field; but the former opens the field to what I think are plainly unconstitutional searches — which is the greater evil. I therefore join the opinion of the Court.
JUSTICE BREYER, dissenting:
I agree with JUSTICE ALITO that New York v. Belton, 453 U.S. 454 (1981), is best read as setting forth a bright-line rule that permits a warrantless search of the passenger compartment of an automobile incident to the lawful arrest of an occupant — regardless of the danger the arrested individual in fact poses. I also agree with JUSTICE STEVENS, however, that the rule can produce results divorced from its underlying Fourth Amendment rationale. Compare Belton, supra, with Chimel v. California, 395 U.S. 752, 764 (1969)(explaining that the rule allowing contemporaneous searches is justified by the need to prevent harm to a police officer or destruction of evidence of the crime). For that reason I would look for a better rule — were the question before us one of first impression.
The matter, however, is not one of first impression, and that fact makes a substantial difference. The Belton rule has been followed not only by this Court in Thornton v. United States, 541 U.S. 615 (2004), but also by numerous other courts. Principles of stare decisis must apply, and those who wish this Court to change a well-established legal precedent — where, as here, there has been considerable reliance on the legal rule in question — bear a heavy burden.Â . . . I have not found that burden met. Nor do I believe that the other considerations ordinarily relevant when determining whether to overrule a case are satisfied. I consequently join JUSTICE ALITO’s dissenting opinion with the exception of Part II-E.
JUSTICE ALITO, with whom THE CHIEF JUSTICE and JUSTICE KENNEDY join, and with whom JUSTICE BREYER joins except as to Part II-E, dissenting:
Twenty-eight years ago, in New York v. Belton, 453 U.S. 454, 460 (1981), this Court held that “when a policeman has made a lawful custodial arrest of the occupant of an automobile, he may, as a contemporaneous incident of that arrest, search the passenger compartment of that automobile.” (Footnote omitted.) Five years ago, in Thornton v. United States, 541 U.S. 615 (2004) — a case involving a situation not materially distinguishable from the situation here — the Court not only reaffirmed but extended the holding of Belton, making it applicable to recent occupants. Today’s decision effectively overrules those important decisions, even though respondent Gant has not asked us to do so.
To take the place of the overruled precedents, the Court adopts a new two-part rule under which a police officer who arrests a vehicle occupant or recent occupant may search the passenger compartment if (1) the arrestee is within reaching distance of the vehicle at the time of the search or (2) the officer has reason to believe that the vehicle contains evidence of the offense of arrest. Ante, at 18. The first part of this new rule may endanger arresting officers and is truly endorsed by only four Justices; JUSTICE SCALIA joins solely for the purpose of avoiding a “4-to-1-to 4 opinion.” Ante, at 4 (concurring opinion). The second part of the new rule is taken from JUSTICE SCALIA’s separate opinion in Thornton without any independent explanation of its origin or justification and is virtually certain to confuse law enforcement officers and judges for some time to come. The Court’s decision will cause the suppression of evidence gathered in many searches carried out in good-faith reliance on well-settled case law, and although the Court purports to base its analysis on the landmark decision in Chimel v. California, 395 U.S. 752 (1969), the Court’s reasoning undermines Chimel. I would follow Belton, and I therefore respectfully dissent.
 Because a broad reading of Belton has been widely accepted, the doctrine of qualified immunity will shield officers from liability for searches [*32] conducted in reasonable reliance on that understanding.