THORNTON v. UNITED STATES.

In New York v. Belton, 453 U. S. 454 (1981), we held that when a police officer has made a lawful custodial arrest of an occupant of an automobile, the Fourth Amendment allows the officer to search the passenger compartment of that vehicle as a contemporaneous incident of arrest. We have granted certiorari twice before to determine whether Belton's rule is limited to situations where the officer makes contact with the occupant while the occupant is inside the vehicle, or whether it applies as well when the officer first makes contact with the arrestee after the latter has stepped out of his vehicle. We did not reach the merits in either of those two cases. Arizona v. Gant, 540 U. S. 963 (2003) (vacating and remanding for reconsideration in light of State v. Dean, 206 Ariz. 158, 76 P. 3d 429 (2003) (en banc)); Florida v. Thomas, 532 U. S. 774 (2001) (dismissing for lack of jurisdiction). We now reach that question and conclude that Belton governs even when an officer does not make contact until the person arrested has left the vehicle.

Officer Deion Nichols of the Norfolk, Virginia, Police Department, who was in uniform but driving an unmarked police car, first noticed petitioner Marcus Thornton when petitioner slowed down so as to avoid driving next to him. Nichols suspected that petitioner knew he was a police officer and for some reason did not want to pull next to him. His suspicions aroused, Nichols pulled off onto a side street and petitioner passed him. After petitioner passed him, Nichols ran a check on petitioner's license tags, which revealed that the tags had been issued to a 1982 Chevy two-door and not to a Lincoln Town Car, the model of car petitioner was driving. Before Nichols had an opportunity to pull him over, petitioner drove into a parking lot, parked, and got out of the vehicle. Nichols saw petitioner leave his vehicle as he pulled in behind him. He parked the patrol car, accosted petitioner, and asked him for his driver's license. He also told him that his license tags did not match the vehicle that he was driving.

Petitioner appeared nervous. He began rambling and licking his lips; he was sweating. Concerned for his safety, Nichols asked petitioner if he had any narcotics or weapons on him or in his vehicle. Petitioner said no. Nichols then asked petitioner if he could pat him down, to which petitioner agreed. Nichols felt a bulge in petitioner's left front pocket and again asked him if he had any illegal narcotics on him. This time petitioner stated that he did, and he reached into his pocket and pulled out two individual bags, one containing three bags of marijuana and the other containing a large amount of crack cocaine. Nichols handcuffed petitioner, informed him that he was under arrest, and placed him in the back seat of the patrol car. He then searched petitioner's vehicle and found a BryCo 9-millimeter handgun under the driver's seat.

A grand jury charged petitioner with possession with intent to distribute cocaine base, 84 Stat. 1260, 21 U. S. C. § 841(a)(1), possession of a firearm after having been previously convicted of a crime punishable by a term of imprisonment exceeding one year, 18 U. S. C. § 922(g)(1), and possession of a firearm in furtherance of a drug trafficking crime, § 924(c)(1). Petitioner sought to suppress, inter alia, the firearm as the fruit of an unconstitutional search. After a hearing, the District Court denied petitioner's motion to suppress, holding that the automobile search was valid under New York v. Belton, supra, and alternatively that Nichols could have conducted an inventory search of the automobile. A jury convicted petitioner on all three counts; he was sentenced to 180 months' imprisonment and 8 years of supervised release.

Petitioner appealed, challenging only the District Court's denial of the suppression motion. He argued that Belton was limited to situations where the officer initiated contact with an arrestee while he was still an occupant of the car. The United States Court of Appeals for the Fourth Circuit affirmed. 325 F. 3d 189 (2003). It held that "the historical rationales for the search incident to arrest doctrine—`the need to disarm the suspect in order to take him into custody' and `the need to preserve evidence for later use at trial,'" id., at 195 (quoting Knowles v. Iowa, 525 U. S. 113, 116 (1998)), did not require Belton to be limited solely to situations in which suspects were still in their vehicles when approached by the police. Noting that petitioner conceded that he was in "close proximity, both temporally and spatially," to his vehicle, the court concluded that the car was within petitioner's immediate control, and thus Nichols' search was reasonable under Belton.1 325 F. 3d, at 196. We granted certiorari, 540 U. S. 980 (2003), and now affirm.

In Belton, an officer overtook a speeding vehicle on the New York Thruway and ordered its driver to pull over. 453 U. S., at 455. Suspecting that the occupants possessed marijuana, the officer directed them to get out of the car and arrested them for unlawful possession. Id., at 454-455. He searched them and then searched the passenger compartment of the car. Id., at 455. We considered the constitutionally permissible scope of a search in these circumstances and sought to lay down a workable rule governing that situation.

We first referred to Chimel v. California, 395 U. S. 752 (1969), a case where the arrestee was arrested in his home, and we had described the scope of a search incident to a lawful arrest as the person of the arrestee and the area immediately surrounding him. 453 U. S., at 457 (citing Chimel, supra, at 763). This rule was justified by the need to remove any weapon the arrestee might seek to use to resist arrest or to escape, and the need to prevent the concealment or destruction of evidence. 453 U. S., at 457. Although easily stated, the Chimel principle had proved difficult to apply in specific cases. We pointed out that in United States v. Robinson, 414 U. S. 218 (1973), a case dealing with the scope of the search of the arrestee's person, we had rejected a suggestion that "`there must be litigated in each case the issue of whether or not there was present one of the reasons supporting the authority'" to conduct such a search. 453 U. S., at 459 (quoting Robinson, supra, at 235). Similarly, because "courts ha[d] found no workable definition of `the area within the immediate control of the arrestee' when that area arguably include[d] the interior of an automobile and the arrestee [wa]s its recent occupant," 453 U. S., at 460, we sought to set forth a clear rule for police officers and citizens alike. We therefore 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." Ibid. (footnote omitted).

In so holding, we placed no reliance on the fact that the officer in Belton ordered the occupants out of the vehicle, or initiated contact with them while they remained within it. Nor do we find such a factor persuasive in distinguishing the current situation, as it bears no logical relationship to Belton's rationale. There is simply no basis to conclude that the span of the area generally within the arrestee's immediate control is determined by whether the arrestee exited the vehicle at the officer's direction, or whether the officer initiated contact with him while he remained in the car. We recognized as much, albeit in dicta, in Michigan v. Long, 463 U. S. 1032 (1983), where officers observed a speeding car swerve into a ditch. The driver exited and the officers met him at the rear of his car. Although there was no indication that the officers initiated contact with the driver while he was still in the vehicle, we observed that "[i]t is clear . . . that if the officers had arrested [respondent] . . . they could have searched the passenger compartment under New York v. Belton." Id., at 1035-1036, and n. 1.

In all relevant aspects, the arrest of a suspect who is next to a vehicle presents identical concerns regarding officer safety and the destruction of evidence as the arrest of one who is inside the vehicle. An officer may search a suspect's vehicle under Belton only if the suspect is arrested. See Knowles, supra, at 117-118. A custodial arrest is fluid and "[t]he danger to the police officer flows from the fact of the arrest, and its attendant proximity, stress, and uncertainty," Robinson, supra, at 234-235, and n. 5 (emphasis added). See Washington v. Chrisman, 455 U. S. 1, 7 (1982) ("Every arrest must be presumed to present a risk of danger to the arresting officer"). The stress is no less merely because the arrestee exited his car before the officer initiated contact, nor is an arrestee less likely to attempt to lunge for a weapon or to destroy evidence if he is outside of, but still in control of, the vehicle. In either case, the officer faces a highly volatile situation. It would make little sense to apply two different rules to what is, at bottom, the same situation.

In some circumstances it may be safer and more effective for officers to conceal their presence from a suspect until he has left his vehicle. Certainly that is a judgment officers should be free to make. But under the strictures of petitioner's proposed "contact initiation" rule, officers who do so would be unable to search the car's passenger compartment in the event of a custodial arrest, potentially compromising their safety and placing incriminating evidence at risk of concealment or destruction. The Fourth Amendment does not require such a gamble.

Petitioner argues, however, that Belton will fail to provide a "bright-line" rule if it applies to more than vehicle "occupants." Brief for Petitioner 29-34. But Belton allows police to search the passenger compartment of a vehicle incident to a lawful custodial arrest of both "occupant[s]" and "recent occupant[s]." 453 U. S., at 460. Indeed, the respondent in Belton was not inside the car at the time of the arrest and search; he was standing on the highway. In any event, while an arrestee's status as a "recent occupant" may turn on his temporal or spatial relationship to the car at the time of the arrest and search,2 it certainly does not turn on whether he was inside or outside the car at the moment that the officer first initiated contact with him.

To be sure, not all contraband in the passenger compartment is likely to be readily accessible to a "recent occupant." It is unlikely in this case that petitioner could have reached under the driver's seat for his gun once he was outside of his automobile. But the firearm and the passenger compartment in general were no more inaccessible than were the contraband and the passenger compartment in Belton. The need for a clear rule, readily understood by police officers and not depending on differing estimates of what items were or were not within reach of an arrestee at any particular moment, justifies the sort of generalization which Belton enunciated.3 Once an officer determines that there is probable cause to make an arrest, it is reasonable to allow officers to ensure their safety and to preserve evidence by searching the entire passenger compartment.

Rather than clarifying the constitutional limits of a Belton search, petitioner's "contact initiation" rule would obfuscate them. Under petitioner's proposed rule, an officer approaching a suspect who has just alighted from his vehicle would have to determine whether he actually confronted or signaled confrontation with the suspect while he remained in the car, or whether the suspect exited his vehicle unaware of, and for reasons unrelated to, the officer's presence. This determination would be inherently subjective and highly fact specific, and would require precisely the sort of ad hoc determinations on the part of officers in the field and reviewing courts that Belton sought to avoid. Id., at 459-460. Experience has shown that such a rule is impracticable, and we refuse to adopt it. So long as an arrestee is the sort of "recent occupant" of a vehicle such as petitioner was here, officers may search that vehicle incident to the arrest.4

The judgment of the Court of Appeals is affirmed.

It is so ordered.

JUSTICE O'CONNOR, concurring in part.

I join all but footnote 4 of the Court's opinion. Although the opinion is a logical extension of the holding of New York v. Belton, 453 U. S. 454 (1981), I write separately to express my dissatisfaction with the state of the law in this area. As JUSTICE SCALIA forcefully argues, post, at 627-629 (opinion concurring in judgment), 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 v. California, 395 U. S. 752 (1969). That erosion is a direct consequence of Belton's shaky foundation. While the approach JUSTICE SCALIA proposes appears to be built on firmer ground, I am reluctant to adopt it in the context of a case in which neither the Government nor the petitioner has had a chance to speak to its merit.

JUSTICE SCALIA, with whom JUSTICE GINSBURG joins, concurring in the judgment.

In Chimel v. California, 395 U. S. 752, 762-763 (1969), we held that a search incident to arrest was justified only as a means to find weapons the arrestee might use or evidence he might conceal or destroy. We accordingly limited such searches to the area within the suspect's "`immediate control'"—i. e., "the area into which an arrestee might reach in order to grab a weapon or evidentiary ite[m]." Id., at 763. In New York v. Belton, 453 U. S. 454, 460 (1981), we set forth a bright-line rule for arrests of automobile occupants, holding that, because the vehicle's entire passenger compartment is "in fact generally, even if not inevitably," within the arrestee's immediate control, a search of the whole compartment is justified in every case.

When petitioner's car was searched in this case, he was neither in, nor anywhere near, the passenger compartment of his vehicle. Rather, he was handcuffed and secured in the back of the officer's squad car. The risk that he would nevertheless "grab a weapon or evidentiary ite[m]" from his car was remote in the extreme. The Court's effort to apply our current doctrine to this search stretches it beyond its breaking point, and for that reason I cannot join the Court's opinion.

* I see three reasons why the search in this case might have been justified to protect officer safety or prevent concealment or destruction of evidence. None ultimately persuades me.

The first is that, despite being handcuffed and secured in the back of a squad car, petitioner might have escaped and retrieved a weapon or evidence from his vehicle—a theory that calls to mind Judge Goldberg's reference to the mythical arrestee "possessed of the skill of Houdini and the strength of Hercules." United States v. Frick, 490 F. 2d 666, 673 (CA5 1973) (opinion concurring in part and dissenting in part). The United States, endeavoring to ground this seemingly speculative fear in reality, points to a total of seven instances over the past 13 years in which state or federal officers were attacked with weapons by handcuffed or formerly handcuffed arrestees. Brief for United States 38-39, and n. 12. These instances do not, however, justify the search authority claimed. Three involved arrestees who retrieved weapons concealed on their own person. See United States v. Sanders, 994 F. 2d 200, 210, n. 60 (CA5 1993) (two instances); U. S. Dept. of Justice, Federal Bureau of Investigation, Uniform Crime Reports: Law Enforcement Officers Killed and Assaulted 49 (2001). Three more involved arrestees who seized a weapon from the arresting officer. See Sanders, supra, at 210, n. 60 (two instances); U. S. Dept. of Justice, Federal Bureau of Investigation, Uniform Crime Reports: Law Enforcement Officers Killed and Assaulted 49 (1998). Authority to search the arrestee's own person is beyond question; and of course no search could prevent seizure of the officer's gun. Only one of the seven instances involved a handcuffed arrestee who escaped from a squad car to retrieve a weapon from somewhere else: In Plakas v. Drinski, 19 F. 3d 1143, 1144-1146 (CA7 1994), the suspect jumped out of the squad car and ran through a forest to a house, where (still in handcuffs) he struck an officer on the wrist with a fireplace poker before ultimately being shot dead.

Of course, the Government need not document specific instances in order to justify measures that avoid obvious risks. But the risk here is far from obvious, and in a context as frequently recurring as roadside arrests, the Government's inability to come up with even a single example of a handcuffed arrestee's retrieval of arms or evidence from his vehicle undermines its claims. The risk that a suspect handcuffed in the back of a squad car might escape and recover a weapon from his vehicle is surely no greater than the risk that a suspect handcuffed in his residence might escape and recover a weapon from the next room—a danger we held insufficient to justify a search in Chimel, supra, at 763.

The second defense of the search in this case is that, since the officer could have conducted the search at the time of arrest (when the suspect was still near the car), he should not be penalized for having taken the sensible precaution of securing the suspect in the squad car first. As one Court of Appeals put it: "`[I]t does not make sense to prescribe a constitutional test that is entirely at odds with safe and sensible police procedures.'" United States v. Mitchell, 82 F. 3d 146, 152 (CA7 1996) (quoting United States v. Karlin, 852 F. 2d 968, 971 (CA7 1988)); see also United States v. Wesley, 293 F. 3d 541, 548-549 (CADC 2002). The weakness of this argument is that it assumes that, one way or another, the search must take place. But conducting a Chimel search is not the Government's right; it is an exception—justified by necessity—to a rule that would otherwise render the search unlawful. If "sensible police procedures" require that suspects be handcuffed and put in squad cars, then police should handcuff suspects, put them in squad cars, and not conduct the search. Indeed, if an officer leaves a suspect unrestrained nearby just to manufacture authority to search, one could argue that the search is unreasonable precisely because the dangerous conditions justifying it existed only by virtue of the officer's failure to follow sensible procedures.

The third defense of the search is that, even though the arrestee posed no risk here, Belton searches in general are reasonable, and the benefits of a bright-line rule justify upholding that small minority of searches that, on their particular facts, are not reasonable. The validity of this argument rests on the accuracy of Belton's claim that the passenger compartment is "in fact generally, even if not inevitably," within the suspect's immediate control. 453 U. S., at 460. By the United States' own admission, however, "[t]he practice of restraining an arrestee on the scene before searching a car that he just occupied is so prevalent that holding that Belton does not apply in that setting would . . . `largely render Belton a dead letter.'" Brief for United States 36-37 (quoting Wesley, supra, at 548). Reported cases involving this precise factual scenario—a motorist handcuffed and secured in the back of a squad car when the search takes place—are legion. See, e. g., United States v. Doward, 41 F. 3d 789, 791 (CA1 1994); United States v. White, 871 F. 2d 41, 44 (CA6 1989); Mitchell, supra, at 152; United States v. Snook, 88 F. 3d 605, 606 (CA8 1996); United States v. McLaughlin, 170 F. 3d 889, 890 (CA9 1999); United States v. Humphrey, 208 F. 3d 1190, 1202 (CA10 2000); Wesley, supra, at 544; see also 3 W. LaFave, Search and Seizure § 7.1(c), pp. 448-449, n. 79 (3d ed. 1996 and Supp. 2004) (citing cases). Some courts uphold such searches even when the squad car carrying the handcuffed arrestee has already left the scene. See, e. g., McLaughlin, supra, at 890-891 (upholding search because only five minutes had elapsed since squad car left).

The popularity of the practice is not hard to fathom. If Belton entitles an officer to search a vehicle upon arresting the driver despite having taken measures that eliminate any danger, what rational officer would not take those measures? Cf. Moskovitz, A Rule in Search of a Reason: An Empirical Reexamination of Chimel and Belton, 2002 Wis. L. Rev. 657, 665-666 (citing police training materials). If it was ever true that the passenger compartment is "in fact generally, even if not inevitably," within the arrestee's immediate control at the time of the search, 453 U. S., at 460, it certainly is not true today. As one judge has put it: "[I]n our search for clarity, we have now abandoned our constitutional moorings and floated to a place where the law approves of purely exploratory searches of vehicles during which officers with no definite objective or reason for the search are allowed to rummage around in a car to see what they might find." McLaughlin, supra, at 894 (Trott, J., concurring). I agree entirely with that assessment.

If Belton searches are justifiable, it is not because the arrestee might grab a weapon or evidentiary item from his car, but simply because the car might contain evidence relevant to the crime for which he was arrested. This more general sort of evidence-gathering search is not without antecedent. For example, in United States v. Rabinowitz, 339 U. S. 56 (1950), we upheld a search of the suspect's place of business after he was arrested there. We did not restrict the officers' search authority to "the area into which [the] arrestee might reach in order to grab a weapon or evidentiary ite[m]," Chimel, 395 U. S., at 763, and we did not justify the search as a means to prevent concealment or destruction of evidence.1 Rather, we relied on a more general interest in gathering evidence relevant to the crime for which the suspect had been arrested. See 339 U. S., at 60-64; see also Harris v. United States, 331 U. S. 145, 151-152 (1947); Marron v. United States, 275 U. S. 192, 199 (1927); Agnello v. United States, 269 U. S. 20, 30 (1925); cf. Weeks v. United States, 232 U. S. 383, 392 (1914).

Numerous earlier authorities support this approach, referring to the general interest in gathering evidence related to the crime of arrest with no mention of the more specific interest in preventing its concealment or destruction. See United States v. Wilson, 163 F. 338, 340, 343 (CC SDNY 1908); Smith v. Jerome, 47 Misc. 22, 23-24, 93 N. Y. S. 202, 202-203 (Sup. Ct. 1905); Thornton v. State, 117 Wis. 338, 346-347, 93 N. W. 1107, 1110 (1903); Ex parte Hurn, 92 Ala. 102, 112, 9 So. 515, 519-520 (1891); Thatcher v. Weeks, 79 Me. 547, 548-549, 11 A. 599, 599-600 (1887); 1 F. Wharton, Criminal Procedure § 97, pp. 136-137 (J. Kerr 10th ed. 1918); 1 J. Bishop, Criminal Procedure § 211, p. 127 (2d ed. 1872); cf. Spalding v. Preston, 21 Vt. 9, 15 (1848) (seizure authority); Queen v. Frost, 9 Car. & P. 129, 131-134 (1839) (same); King v. Kinsey, 7 Car. & P. 447 (1836) (same); King v. O'Donnell, 7 Car. & P. 138 (1835) (same); King v. Barnett, 3 Car. & P. 600, 601 (1829) (same). Bishop's 1872 articulation is typical:

"The officer who arrests a man on a criminal charge should consider the nature of the charge; and, if he finds about the prisoner's person, or otherwise in his possession, either goods or moneys which there is reason to believe are connected with the supposed crime as its fruits, or as the instruments with which it was committed, or as directly furnishing evidence relating to the transaction, he may take the same, and hold them to be disposed of as the court may direct." Bishop, supra, § 211, at 127.

Only in the years leading up to Chimel did we start consistently referring to the narrower interest in frustrating concealment or destruction of evidence. See Sibron v. New York, 392 U. S. 40, 67 (1968); Preston v. United States, 376 U. S. 364, 367 (1964).

There is nothing irrational about broader police authority to search for evidence when and where the perpetrator of a crime is lawfully arrested. The fact of prior lawful arrest distinguishes the arrestee from society at large, and distinguishes a search for evidence of his crime from general rummaging. Moreover, it is not illogical to assume that evidence of a crime is most likely to be found where the suspect was apprehended.

Nevertheless, Chimel's narrower focus on concealment or destruction of evidence also has historical support. See Holker v. Hennessey, 141 Mo. 527, 539-540, 42 S. W. 1090, 1093 (1897); Dillon v. O'Brien, 16 Cox C. C. 245, 250 (Exch. Div. Ir. 1887); Reifsnyder v. Lee, 44 Iowa 101, 103 (1876); S. Welch, Essay on the Office of Constable 17 (1758).2 And some of the authorities supporting the broader rule address only searches of the arrestee's person, as to which Chimel's limitation might fairly be implicit. Moreover, carried to its logical end, the broader rule is hard to reconcile with the influential case of Entick v. Carrington, 19 How. St. Tr. 1029, 1031, 1063-1074 (C. P. 1765) (disapproving search of plaintiff's private papers under general warrant, despite arrest). But cf. Dillon, supra, at 250-251 (distinguishing Entick); Warden, Md. Penitentiary v. Hayden, 387 U.S. 294, 303-304 (1967).

In short, both Rabinowitz and Chimel are plausible accounts of what the Constitution requires, and neither is so persuasive as to justify departing from settled law. But if we are going to continue to allow Belton searches on stare decisis grounds, we should at least be honest about why we are doing so. Belton cannot reasonably be explained as a mere application of Chimel. Rather, it is a return to the broader sort of search incident to arrest that we allowed before Chimel—limited, of course, to searches of motor vehicles, a category of "effects" which give rise to a reduced expectation of privacy, see Wyoming v. Houghton, 526 U.S. 295, 303 (1999), and heightened law enforcement needs, see id., at 304; Rabinowitz, supra, at 73 (Frankfurter, J., dissenting).

Recasting Belton in these terms would have at least one important practical consequence. In United States v. Robinson, 414 U.S. 218, 235 (1973), we held that authority to search an arrestee's person does not depend on the actual presence of one of Chimel's two rationales in the particular case; rather, the fact of arrest alone justifies the search. That holding stands in contrast to Rabinowitz, where we did not treat the fact of arrest alone as sufficient, but upheld the search only after noting that it was "not general or exploratory for whatever might be turned up" but reflected a reasonable belief that evidence would be found. 339 U. S., at 62-63; see also Smith, 47 Misc., at 24, 93 N. Y. S., at 203 ("This right and duty of search and seizure extend, however, only to articles which furnish evidence against the accused"); cf. Barnett, supra, at 601 (seizure authority limited to relevant evidence); Bishop, supra, § 211, at 127 (officer should "consider the nature of the charge" before searching). The two different rules make sense: When officer safety or imminent evidence concealment or destruction is at issue, officers should not have to make fine judgments in the heat of the moment. But in the context of a general evidence-gathering search, the state interests that might justify any over-breadth are far less compelling. A motorist may be arrested for a wide variety of offenses; in many cases, there is no reasonable basis to believe relevant evidence might be found in the car. See Atwater v. Lago Vista, 532 U. S. 318, 323-324 (2001); cf. Knowles v. Iowa, 525 U. S. 113, 118 (1998). I would therefore limit Belton searches to cases where it is reasonable to believe evidence relevant to the crime of arrest might be found in the vehicle.

In this case, as in Belton, petitioner was lawfully arrested for a drug offense. It was reasonable for Officer Nichols to believe that further contraband or similar evidence relevant to the crime for which he had been arrested might be found in the vehicle from which he had just alighted and which was still within his vicinity at the time of arrest. I would affirm the decision below on that ground.3