with whom Justice Stevens and Justice Ginsburg join, dissenting.
Anyone who travels by air today submits to searches of the person and luggage as a condition of boarding the aircraft. It is universally accepted that such intrusions are necessary to hedge against risks that, nowadays, even small children understand. The commonplace precautions of air travel have not, thus far, been justified for ground transportation, however, and no such conditions have been placed on passengers getting on trains or buses. There is therefore an air of unreality about the Court’s explanation that bus passengers consent to searches of their luggage to “enhanc[e] their own safety and the safety of those around them.” Ante, at 205. Nor are the other factual assessments underlying the Court’s conclusion in favor of the Government more convincing.
The issue we took to review is whether the police’s examination of the bus passengers, including respondents, amounted to a suspicionless seizure under the Fourth Amendment.1 If it did, any consent to search was plainly *209invalid as a product of the illegal seizure. See Florida v. Royer, 460 U. S. 491, 507-508 (1983) (plurality opinion) (“[T]he consent was tainted by the illegality and ... ineffective to justify the search”); id., at 509 (Powell, J., concurring); id., at 509 (Brennan, J., concurring in result).
Florida v. Bostick, 501 U. S. 429 (1991), established the framework for determining whether the bus passengers were seized in the constitutional sense. In that case, we rejected the position that police questioning of bus passengers was a per se seizure, and held instead that the issue of seizure was to be resolved under an objective test considering all circumstances: whether a reasonable passenger would have felt “free to decline the officers’ requests or otherwise terminate the encounter,” id., at 436. We thus applied to a bus passenger the more general criterion, whether the person questioned was free “to ignore the police presence and go about his business,” id., at 437 (quoting Michigan v. Chesternut, 486 U. S. 567, 569 (1988)).
Before applying the standard in this ease, it may be worth getting some perspective from different sets of facts. A perfect example of police conduct that supports no colorable claim of seizure is the act of an officer who simply goes up to a pedestrian on the street and asks him a question. See Royer, 460 U. S., at 497; see id., at 523, n. 3 (Rehnquist, J., dissenting). A pair of officers questioning a pedestrian, *210without more, would presumably support the same conclusion. Now consider three officers, one of whom stands behind the pedestrian, another at his side toward the open sidewalk, with the third addressing questions to the pedestrian a foot or two from his face. Finally, consider the same scene in a narrow alley. On such barebones facts, one may not be able to say a seizure occurred, even in the last case, but one can say without qualification that the atmosphere of the encounters differed significantly from the first to the last examples. In the final instance there is every reason to believe that the pedestrian would have understood, to his considerable discomfort, what Justice Stewart described as the “threatening presence of several officers,” United States v. Mendenhall, 446 U. S. 544, 554 (1980) (opinion of Stewart, J.). The police not only carry legitimate authority but also exercise power free from immediate check, and when the attention of several officers is brought to bear on one civilian the imbalance of immediate power is unmistakable. We all understand this, as well as we understand that a display of power rising to Justice Stewart’s “threatening” level may overbear a normal person’s ability to act freely, even in the absence of explicit commands or the formalities of detention. As common as this understanding is, however, there is little sign of it in the Court’s opinion. My own understanding of the relevant facts and their significance follows.
When the bus in question made its scheduled stop in Tallahassee, the passengers were required to disembark while the vehicle was cleaned and refueled. App. 104. When the passengers returned, they gave their tickets to the driver, who kept them and then left himself, after giving three police officers permission to board the bus in his absence. Id., at 77-78. Although they were not in uniform, the officers displayed badges and identified themselves as police. One stationed himself in the driver’s seat by the door at the front, facing back to observe the passengers. The two others went to the rear, from which they worked their way for*211ward, with one of them speaking to passengers, the other backing him up. Id., at 47-48. They necessarily addressed the passengers at very close range; the aisle was only 15 inches wide, and each seat only 18.2 The quarters were cramped further by the overhead rack, 19 inches above the top of the passenger seats. The passenger by the window could not have stood up straight, id., at 55, and the face of the nearest officer was only a foot or 18 inches from the face of the nearest passenger being addressed, id., at 57. During the exchanges, the officers looked down, and the passengers, had to look up if they were to face the police. The officer asking the questions spoke quietly. He prefaced his requests for permission to search luggage and do a body pat-down by identifying himself by name as a police investigator “conducting bus interdiction” and saying, “‘We would like for your cooperation. Do you have any luggage on the bus?’ ” Id., at 82.
Thus, for reasons unexplained, the driver with the tickets entitling the passengers to travel had yielded his custody of the bus and its seated travelers to three police officers, whose authority apparently superseded the driver’s own. The officers took control of the entire passenger compartment, one stationed at the door keeping surveillance of all the occupants, the others working forward from the back. With one officer right behind him and the other one forward, a third officer accosted each passenger at quarters extremely close and so cramped that as many as half the passengers could not even have stood to face the speaker. None was asked whether he was willing to converse with the police or to take part in the enquiry. Instead the officer said the police were “conducting bus interdiction,” in the course of which they “would like . . . cooperation.” Ibid. The reasonable inference was that the “interdiction” was not a consensual exercise, but one the police would carry o.ut what*212ever the circumstances; that they would prefer “cooperation” but would not let the lack of it stand in their way. There was no contrary indication that day, since no passenger had refused the cooperation requested, and there was no reason for any passenger to believe that the driver would return and the trip resume until the police were satisfied. The scene was set and an atmosphere of obligatory participation was established by this introduction. Later requests to search prefaced with “Do you mind ...” would naturally have been understood in the terms with which the encounter began.
It is very hard to imagine that either Brown or Drayton would have believed that he stood to lose nothing if he refused to cooperate with the police, or that he had any free choice to ignore the police altogether. No reasonable passenger could have believed that, only an uncomprehending one. It is neither here nor there that the interdiction was conducted by three officers, not one, as a safety precaution. See id., at 47. The fact was that there were three, and when Brown and Drayton were called upon to respond, each one was presumably conscious of an officer in front watching, one at his side questioning him, and one behind for cover, in case he became unruly, perhaps, or “cooperation” was not forthcoming. The situation is much like the one in the alley, with civilians in close quarters, unable to move effectively, being told their cooperation is expected. While I am not prepared to say that no bus interrogation and. search can pass the Bostick test without a warning that passengers are free to say no, the facts here surely required more from the officers than a quiet tone of voice. A police officer who is certain to get his way has no need to shout.
It is true of course that the police testified that a bus passenger sometimes says no, App. 81, but that evidence does nothing to cast the facts here in a different light. We have no way of knowing the circumstances in which a passenger elsewhere refused a request; maybe that has happened only *213when the police have told passengers they had a right to refuse (as the officers sometimes advised them), id., at 81-82. Nor is it fairly possible to see the facts of this case differently by recalling INS v. Delgado, 466 U. S. 210 (1984), as precedent. In that case, a majority of this Court found no seizure when a factory force was questioned by immigration officers, with an officer posted at every door leading from the workplace. Id., at 219. Whether that opinion was well reasoned or not, the facts as the Court viewed them differed from the case here. Delgado considered an order granting summary judgment in favor of respondents, with the consequence that the Court was required to construe the record and all issues of fact favorably to the Immigration and Naturalization Service. See id., at 214; id., at 221 (Stevens, J., concurring). The Court therefore emphasized that even after “th[e] surveys were initiated, the employees were about their ordinary business, operating machinery and performing other job assignments.” Id., at 218. In this case, however, Brown and Drayton were seemingly pinned-in by the officers and the customary course of events was stopped flat. The bus was going nowhere, and with one officer in the driver’s seat, it was reasonable to suppose no passenger would tend to his own business until the officers were ready to let him.
In any event, I am less concerned to parse this case against Delgado than to apply Bostick’s totality of circumstances test, and to ask whether a passenger would reasonably have felt free to end his encounter with the three officers by saying no and ignoring them thereafter. In my view the answer is clear. The Court’s contrary conclusion tells me that the majority cannot see what Justice Stewart saw, and I respectfully dissent.
The Court proceeds to resolve the voluntariness issue on the heels of its seizure enquiry, but the voluntariness of respondents’ consent was not within the question the Court accepted for review. Accord, Reply Brief for United States 20, n. 7 (stating that the consent issue “is not presented by this case; the question here is whether there was an illegal seizure *209in the first place”). While it is true that the Eleventh Circuit purported to address the question “whether the consent given by each defendant for the search was ‘uncoerced and legally voluntary,”’ 231 F. 3d 787, 788 (2000), elsewhere the court made it clear that it was applying the test in Florida, v. Bostick, 501 U. S. 429 (1991), which is relevant to the issue of seizure, 231 F. 3d, at 791, n. 6. There is thus no occasion here to reach any issue of consent untainted by seizure. If there were, the consent would have to satisfy the voluntariness test of Schneckloth v. Bustamonte, 412 U. S. 218 (1973), which focuses on “the nature of a person’s subjective understanding,” id., at 230, and requires consideration of “the characteristics of the accused [in addition to] the details of the interrogation,” id., at 226.
The figures are from a Lodging filed by respondents (available in Clerk of Court’s case file). The Government does not dispute their accuracy.