with whom Justice Marshall joins, dissenting.
In Chimel v. California, 395 U. S. 752 (1969), this Court carefully analyzed more than 50 years of conflicting precedent governing the permissible scope of warrantless searches incident to custodial arrest. The Court today turns its back on the product of that analysis, formulating an arbitrary “bright-fine” rule applicable to “recent” occupants of automobiles that fails to reflect Chimel’s underlying policy justifications. While the Court claims to leave Chimel intact, see ante, at 460, n. 3,1 fear that its unwarranted abandonment of *464the principles underlying that decision may signal a wholesale retreat from our carefully developed search-incident-to-arrest analysis. I dissent.
I
It has long been a fundamental principle of Fourth Amendment analysis that exceptions to the warrant requirement are to be narrowly construed. Arkansas v. Sanders, 442 U. S. 753, 759-760 (1979); Mincey v. Arizona, 437 U. S. 385, 393-394 (1978); Coolidge v. New Hampshire, 403 U. S. 443, 454-455 (1971); Vale v. Louisiana, 399 U. S. 30, 34 (1970); Katz v. United States, 389 U. S. 347, 357 (1967); Jones v. United States, 357 U. S. 493, 499 (1958). Predicated on the Fourth Amendment’s essential purpose of “shield[ing] the citizen from unwarranted intrusions into his privacy,” Jones v. United States, supra, at 498, this principle carries with it two corollaries. First, for a search to be valid under the Fourth Amendment, it must be “ ‘strictly tied to and justified by’ the circumstances which rendered its initiation permissible.” Terry v. Ohio, 392 U. S. 1, 19 (1968), quoting Warden v. Hayden, 387 U. S. 294, 310 (1967) (Fortas, J., concurring). See Chimel v. California, supra, at 762; Cupp v. Murphy, 412 U. S. 291, 295 (1973). Second, in determining whether to grant an exception to the warrant requirement, courts should carefully consider the facts and circumstances of each search and seizure, focusing on the reasons supporting the exception rather than on any bright-line rule of general application. See Sibron v. New York, 392 U. S. 40, 59 (1968); Preston v. United States, 376 U. S. 364, 367 (1964).1
The Chimel exception to the warrant requirement was designed with two principal concerns in mind: the safety of the arresting officer and the preservation of easily concealed or destructible evidence. Recognizing that a suspect might have *465access to weapons or contraband at the time of arrest, the Court declared:
"When an arrest is made, it is reasonable for the arresting officer to search the person arrested in order to remove any weapons that the latter might seek to use in order to resist arrest or effect his escape. Otherwise, the officer’s safety might well be endangered, and the arrest itself frustrated. In addition, it is entirely reasonable for the arresting officer to search for and seize any evidence on the arrestee’s person in order to prevent its concealment or destruction. And the area into which an arrestee might reach in order to grab a weapon or evidentiary items must, of course, be governed by a like rule.” 395 IT. S., at 762-763.
The Chimel standard was narrowly tailored to address these concerns: it permits police officers who have effected a custodial arrest to conduct a warrantless search “of the ar-restee’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.” Id., at 763. It thus places a temporal and a spatial limitation on searches incident to arrest, excusing compliance with the warrant requirement only when the search “ ‘is substantially contemporaneous with the arrest and is confined to the immediate vicinity of the arrest.’ ” Shipley v. California, 395 U. S. 818, 819 (1969), quoting Stoner v. California, 376 U. S. 483, 486 (1964). See United States v. Chadwick, 433 U. S. 1, 14-15 (1977); Dyke v. Taylor Implement Mfg. Co., 391 U. S. 216, 220 (1968); Preston v. United States, supra, at 367; United States v. Edwards, 415 U. S. 800, 810 (1974) (Stewart, J., dissenting).2 When the arrest has been *466consummated and the arrestee safely taken into custody, the justifications underlying Chimel's limited exception to the warrant requirement cease to apply: at that point there is no possibility that the arrestee could reach weapons or contraband. See Chimel v. California, supra, at 764.
In its attempt to formulate a “ 'single, familiar standard . . . to guide police officers, who have only limited time and expertise to reflect on and balance the social and individual interests involved in the specific circumstances they confront/ ” ante, at 458, quoting Dunaway v. New York, 442 U. S. 200, 213-214 (1979), the Court today disregards these principles, and instead adopts a fiction — that the interior of a car is always within the immediate control of an arrestee who has recently been in the car. The Court thus holds:
“[W]hen 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 . . . [and] may also examine the contents of any containers found within the passenger compartment . . . Ante, at 460.
In so holding, the Court ignores both precedent and principle and fails to achieve its objective of providing police officers with a more workable standard for determining the permissible scope of searches incident to arrest.
II
As the facts of this case make clear, the Court today substantially expands the permissible scope of searches incident to arrest by permitting police officers to search areas and containers the arrestee could not possibly reach at the time of arrest. These facts demonstrate that at the time Belton and his three companions were placed under custodial arrest— which was after they had been removed from the car, patted down, and separated — none of them could have reached the jackets that had been left on the back seat of the car. The *467New York Court of Appeals described the sequence of events as follows:
“On April 9, 1978, defendant and three companions were traveling on the New York State Thruway in Ontario County when their car was stopped by a State trooper for speeding. Upon approaching the vehicle, the officer smelled the distinct odor of marihuana emanating from within and observed on the floor an envelope which he recognized as a type that is.commonly used to sell the substance. At that point the officer ordered the occupants out of the vehicle, patted each down, removed the envelope from the floor and ascertained that it contained a small amount of marihuana.
“After the marihuana was found, the individuals, still standing outside the car, were placed under arrest. The officer then re-entered the vehicle, searched the passenger compartment and seized the marihuana cigarette butts lying in the ashtrays. He also rifled through the pockets of five jackets on the hack seat. Upon opening the zippered pocket of one of them, he discovered a small amount of cocaine and defendant’s identification.” 50 N. Y. 2d 447, 449, 407 N. E. 2d 420, 421 (1980) (emphasis added).3
Concluding that a “warrantless search of the zippered pockets of an unaccessible jacket may not be upheld as a search incident to a lawful arrest where there is no longer any danger that the arrestee or a confederate might gain access to the article,” ibid, (emphasis added), the court further stated:
“One searches the record in vain for support of the dissenter’s claim that at the time of the arrest — the point from which the predicate for the warrantless search is measured — 'the jackets were within reach of the four sus*468pects and had not yet been reduced to the exclusive control of the officer.’ ” Id., at 452, n. 2, 407 N. E. 2d, at 423, n. 2, quoting id., at 454, 407 N. E. 2d, at 424 (dissenting opinion).
By approving the constitutionality of the warrantless search in this case, the Court carves out a dangerous precedent that is not justified by the concerns underlying Chimel. Disregarding the principle “that the scope of a warrantless search must be commensurate with the rationale that excepts the search from the warrant requirement,” Cupp v. Murphy, 412 U. S., at 295, the Court for the first time grants police officers authority to conduct a warrantless “area” search under circumstances where there is no chance that the arrestee “might gain possession of a weapon or destructible evidence.” Chimel v. California, 395 U. S., at 763. Under the approach taken today, the result would presumably be the same even if Officer Nicot had handcuffed Belton and his companions in the patrol car before placing them under arrest, and even if his search had extended to locked luggage or other inaccessible containers located in the back seat of the car.
This expansion of the Chimel exception is both analytically unsound and inconsistent with every significant search-incident-to-arrest case we have decided in which the issue was whether the police could lawfully conduct a warrantless search of the area surrounding the arrestee. See, e. g., United States v. Chadwick, 433 U. S., at 15 (search of footlocker “conducted more than an hour after federal agents had gained exclusive control of the footlocker and long after respondents were securely in custody” not incident to arrest); Coolidge v. New Hampshire, 403 U. S., at 456-457, and n. 11 (search of car in driveway not incident to arrest in house); Chambers v. Maroney, 399 U. S. 42, 47 (1970) (warrantless search of car invalid once arrestee has been placed in police custody); Vale v. Louisiana, 399 U. S., at 35 (area of immediate control does not extend to inside of house when suspect is arrested on front step); Dyke v. Taylor Implement Mfg. Co., 391 U. S., at 220 *469(search of car after occupant placed in custody and taken to courthouse not valid as incident to arrest); Preston v. United States, 376 U. S., at 368 (search of car not valid as incident to arrest: although suspects were in car when arrested, they were in custody at police station when car was searched). These cases demonstrate that the crucial question under Chimel is not whether the arrestee could ever have reached the area that was searched, but whether he could have reached it at the time of arrest and search. If not, the officer’s failure to obtain a warrant may not be excused.4 By disregarding this settled doctrine, the Court does a great disservice not only to stare decisis, but to the policies underlying the Fourth Amendment as well.
111
The Court seeks to justify its departure from the principles underlying Chimel by proclaiming the need for a new “bright-line” rule to guide the officer in the field. As we pointed out in Mincey v. Arizona, 437 U. S., at 393, however, “the mere fact that law enforcement may be made more efficient can never by itself justify disregard of the Fourth Amendment.” Moreover, the Court’s attempt to forge a “bright-line” rule fails on its own terms. While the “interior/trunk” distinction may provide a workable guide in certain routine cases— for example, where the officer arrests the driver of a car and then immediately searches the seats and floor — in the long run, I suspect it will create far more problems than it solves. The Court’s new approach leaves open too many questions and, more important, it provides the police and the courts with too few tools with which to find the answers.
Thus, although the Court concludes that a warrantless search of a car may take place even though the suspect was *470arrested outside the car, it does not indicate how long after the suspect’s arrest that search may validly be conducted. Would a warrantless search incident to arrest be valid if conducted five minutes after the suspect left his car? Thirty minutes? Three hours? Does it matter whether the suspect is standing in close proximity to the car when the search is conducted? Does it matter whether the police formed probable cause to arrest before or after the suspect left his car? And why is the rule announced today necessarily limited to searches of cars? What if a suspect is seen walking out of a house where the police, peering in from outside, had formed probable cause to believe a crime was being committed? Could the police then arrest that suspect and enter the house to conduct a search incident to arrest? Even assuming today’s rule is limited to searches of the “interior” of cars — an assumption not demanded by logic — what is meant by “interior”? Does it include locked glove compartments, the interior of door panels, or the area under the floorboards? Are special rules necessary for station wagons and hatchbacks, where the luggage compartment may be reached through the interior, or taxicabs, where a glass panel might separate the driver’s compartment from the rest of the car? Are the only containers that may be searched those that are large enough to be “capable of holding another object”? Or does the new rule apply to any container, even if it “could hold neither a weapon nor evidence of the criminal conduct for which the suspect was arrested”? Compare ante, at 460-461, n. 4, with ante, at 461.
The Court does not give the police any “bright-line” answers to these questions. More important, because the Court’s new rule abandons the justifications underlying Chimel, it offers no guidance to the police officer seeking to work out these answers for himself. As we warned in Chimel: “No consideration relevant to the Fourth Amendment suggests any point of rational limitation, once the search is allowed to go beyond the area from which the person arrested *471might obtain weapons or evidentiary items.” 395 U. S., at 766. See also Mincey v. Arizona, supra, at 393. By failing to heed this warning, the Court has undermined rather than furthered the goal of consistent law enforcement: it has failed to offer any principles to guide the police and the courts in their application of the new rule to nonroutine situations.
The standard announced in Chimel is not nearly as difficult to apply as the Court suggests. To the contrary, I continue to believe that Chimel provides a sound, workable rule for determining the constitutionality of a warrantless search incident to arrest. Under Chimel, searches incident to arrest may be conducted without a warrant only if limited to the person of the arrestee, see United States v. Robinson, 414 U. S. 218 (1973), or to the area within the arrestee’s “immediate control.” While it may be difficult in some cases to measure the exact scope of the arrestee’s immediate control, relevant factors would surely include the relative number of police officers and arrestees, the manner of restraint placed on the arrestee, and the ability of the arrestee to gain access to a particular area or container.5 Certainly there will be some close cases, but when in doubt the police can always turn to the rationale underlying Chimel — the need to prevent the arrestee from reaching weapons or contraband — before *472exercising their judgment. A rule based on that rationale should provide more guidance than the rule announced by the Court today. Moreover, unlike the Court’s rule, it would be faithful to the Fourth Amendment.
As we noted in Go-Bart Importing Co. v. United States, 282 U. S. 344, 357 (1931): “There is no formula for the determination of reasonableness. Each case is to be decided on its own facts and circumstances.”
“ 'Once an accused is under arrest and in custody, then a search made at another place, without a warrant, is simply not incident to the arrest.’ ” Chambers v. Maroney, 399 U. S. 42, 47 (1970), quoting Preston v. United States, 376 U. S., at 367.
See also 50 N. Y. 2d, at 454, n. 2, 407 N. E. 2d, at 423, n. 2; Tr. of Oral Arg. 4-5; App. A-36.
“ ‘We cannot be true to [the Fourth Amendment] and excuse the absence of a search warrant without a showing by those who seek exemption from the constitutional mandate that the exigencies of the situation make that course imperative.’ ” Chimel v. California, 395 U. S., at 761, quoting McDonald v. United States, 335 U. S. 451, 456 (1948).
The Court sets up a strawman when it claims that under the “exclusive control” approach taken by the Court of Appeals, “no search or seizure incident to a lawful custodial arrest would ever be valid; by seizing an article even on the arrestee’s person, an officer may be said to have reduced that article to his ‘exclusive control.’ ” Ante, at 461-462, n. 5. If a police officer could obtain exclusive control of an article by simply holding it in his hand, I would certainly agree with the Court. But as we recognized in United States v. Chadwick, 433 U. S. 1, 14-15 (1977), exclusive control means more than that. It means sufficient control such that there is no significant risk that the arrestee or his confederates “might gain possession of a weapon or destructible evidence.” Chimel v. California, 395 U. S., at 763. The issue of exclusive control presents a question of fact to be decided under the circumstances of each case, just as the New York Court of Appeals has decided it here.