concurring in the result.
I fully agree with the results the Court has reached in these cases. They are, I think, consonant with and dictated by the decision in Terry v. Ohio, ante, p. 1. For reasons I do not understand, however, the Court has declined to rest the judgments here upon the principles *71of Terry. In doing so it has, in at least one particular, made serious inroads upon the protection' afforded by the Fourth and Fourteenth Amendments.
The Court is of course entirely correct in concluding that we should not pass upon the constitutionality of the New York stop-and-frisk law “on its face.” The statute is certainly not unconstitutional on its face: that is, it does not plainly purport to authorize unconstitutional activities by policemen. Nor is it “constitutional on its face” if that expression means that any action now or later thought to fall within the terms of the statute is, ipso facto, within constitutional limits as well. No statute, state or federal, receives any such imprimatur from this Court.
This does not mean, however, that the statute should be ignored here. The State of New York has made a deliberate effort to deal with the complex problem of on-the-street policework. Without giving carte blanche to any particular verbal formulation, we should, I think, where relevant, indicate the extent to which that effort has been constitutionally successful. The core of the New York statute is the permission to stop any person reasonably suspected of crime. Under the decision in Terry a right to stop may indeed be premised on rea-sonablé suspicion and does not require probable cause, and hence the New York formulation is to that extent constitutional. This does not mean that suspicion need not be “reasonable” in the constitutional as well as the statutory sense. Nor does it mean that this Court has approved more than a momentary stop or has indicated what questioning may constitutionally occur during a stop, for the cases before us do not raise these questions.1
*72Turning to the individual cases, I agree that the conviction in No. 63, Sibron, should be reversed, and would do so upon the premises of Terry, At the outset, I agree that sufficient collateral legal consequences of Sibron’s conviction have been shown to prevent this case from being moot, and I agree that the case should not be reversed simply on the State’s confession of error.
The considerable confusion that has surrounded the “search” or “frisk” of Sibron that led to the actual recovery of the heroin seems to me irrelevant for our purposes. Officer Martin repudiated his first statement, which might conceivably have indicated a theory of “abandonment,” see ante, at 45-46. No matter which of the other theories is adopted, it is clear that there was at least a forcible frisk, comparable to that which occurred in Terry, which requires constitutional justification.
Since carrying heroin is a crime in New York, probable cause to believe Sibron was carrying heroin would also have been probable cause to arrest him. As the Court says, Officer Martin clearly had neither. Although Sibron had had conversations with several known addicts, he had done nothing, during the several hours he was under surveillance, that made it “probable” that he was either carrying heroin himself or engaging in transactions with these acquaintances.
Nor were there here reasonable grounds for a Terry-type “stop” short of an arrest. I would accept, as an adequate general formula, the New York requirement that the officer must “reasonably suspect” that the person he stops “is committing, has committed or is about to commit a felony.” N. Y. Code Crim. Proc. § 180-a. “On its face,” this requirement is, if anything, more stringent than the requirement stated by the Court in Terry: “where a police officer observes unusual conduct which leads him reasonably to conclude in light of his experience that criminal activity may be afoot . . . .” *73Ante, at 30. The interpretation of the New York statute is of course a matter for the New York courts, but any particular stop must meet the Terry standard as well.
The forcible encounter between Officer Martin and Sibron did not meet the Terry reasonableness standard. In the first place, although association with known criminals may, I think, properly be a factor contributing to the suspiciousness of circumstances, it does not, entirely by itself, create suspicion adequate to support a stop. There must be something at least in the activities of the person being observed or in his surroundings that affirmatively suggests particular criminal activity, completed, current, or intended. That was the case in Terry, but it palpably was not the case here. For eight continuous hours, up to the point when he interrupted Sibron eating a piece of pie, Officer Martin apparently observed not a single suspicious action and heard not a single suspicious word on the part of Sibron himself or any person with whom he associated. If anything, that period of surveillance pointed away from suspicion.
Furthermore, in Terry, the police officer judged that his suspect was about to commit a violent crime and that he had to assert himself in order to prevent it. Here there was no reason for Officer Martin to think that an incipient crime, or flight, or the destruction of evidence would occur if he stayed his hand; indeed, there was no more reason for him to intrude upon Sibron at the moment when he did than there had been four hours earlier, and no reason to think the situation would have changed four hours later. While no hard-and-fast rule can be drawn, I would suggest that one important factor, missing here, that should be taken into account in determining whether there are reasonable grounds for a forcible intrusion is whether there is any need for immediate action.
*74For these reasons I would hold that Officer Martin lacked reasonable grounds to intrude forcibly upon Sibron. In consequence, the essential premise for the right to conduct a self-protective frisk was lacking. See my concurring opinion in Terry, ante, p. 31. I therefore find it unnecessary to reach two further troublesome questions. First, although I think that, as in Terry, the right to frisk is automatic when an officer lawfully stops a person suspected of a crime whose nature creates a substantial likelihood that he is armed, it is not clear that suspected possession of narcotics falls into this category. If the nature of the suspected offense creates no reasonable apprehension for the officer’s safety, I would not permit him to frisk unless other circumstances did so. Second, I agree with the Court that even where a self-protective frisk is proper, its scope should be limited to what is adequate for its purposes. I see no need here to resolve the question whether this frisk exceeded those bounds.
Turning now to No. 74, Peters, I agree that the conviction should be upheld, but here I would differ strongly and fundamentally with the Court’s approach. The Court holds that the burglar’s tools were recovered from Peters in a search incident to a lawful arrest. I do not think that Officer Lasky had anything close to probable cause to arrest Peters before he recovered the burglar’s tools. Indeed, if probable cause existed here, I find it difficult to see why a different rationale was necessary to support the stop and frisk in Terry and why States such as New York have had to devote so much thought to the constitutional problems of field interrogation. This case will be the latest in an exceedingly small number of cases in this Court indicating what suffices for probable cause. While, as the Court noted in Terry, the influence of this Court on police tactics “in *75the field” is necessarily limited, the influence of a decision here on hundreds of courts and magistrates who have to decide whether there is probable cause for a real arrest or a full search will be large.
Officer Lasky testified that at 1 o’clock in the afternoon he heard a noise at the door to his apartment. He did not testify, nor did any state court conclude, that this “led him to believe that someone sought to force entry.” Ante, at 66. He looked out into the public hallway and saw two men whom he did not recognize, surely not a strange occurrence in a large apartment building. One of them appeared' to be tip-toeing. Lasky did not testify that the other man was tiptoeing or that either of them was behaving “furtively.” Ibid. Lasky left his apartment and ran to them, gun in hand. He did not testify that there was any “flight,” ante, at 66,2 though flight at the approach of a gun-carrying stranger (Lasky was apparently not in uniform) is hardly indicative of mens rea.
Probable cause to arrest means evidence that would warrant a prudent and reasonable man (such as a magistrate, actual or hypothetical) in believing that a particular person has committed or is committing a crime.3 *76Officer Lasky had no extrinsic reason to think that a crime had been or was being committed, so whether it would have been proper to issue a warrant depends entirely on his statements of his observations of the men. Apart from his conclusory statement that he thought the men were burglars, he offered very little specific evidence. I find it hard to believe that if Peters had made good his escape and there were no report of a burglary in the neighborhood, this Court would hold it proper for a prudent neutral magistrate to issue a warrant for his arrest.4
In the course of upholding Peters’ conviction, the Court makes two other points that may lead to future confusion. The first concerns the “moment of arrest.” If there is an escalating encounter between a policeman and a citizen, beginning perhaps with a friendly conversation but ending in imprisonment, and if evidence is developing during that encounter, it may be important to identify the moment of arrest, i. e., the moment when the policeman was not permitted to proceed further unless he by then had probable cause. This moment-of-arrest problem is not, on the Court’s premises, in any way involved in this case: the Court holds that Officer Lasky had probable cause to arrest at the moment he caught Peters, and hence probable cause clearly preceded anything that might be thought an arrest. The Court implies, however, that although there is no problem about whether the arrest of Peters occurred *77late enough, i. e., after probable cause developed, there might be a problem about whether it occurred early enough, i. e., before Peters was searched. This seems to me a false problem. Of course, the fruits of a search may not be used to justify an arrest to which it is incident, but this means only that probable cause to arrest must precede the search. If the prosecution shows probable cause to arrest prior to a search of a man’s person, it has met its total burden. There is no case in which a defendant may validly say, “Although the officer had a right to arrest me at the moment when he seized me and searched my person, the search is invalid because he did not in fact arrest me until afterwards.”
This fact is important because, as demonstrated by Terry, not every curtailment of freedom of movement is an “arrest” requiring antecedent probable cause. At the same time, an officer who does have probable cause may of course seize and search immediately. Hence while certain police actions will undoubtedly turn an encounter into an arrest requiring antecedent probable cause, the prosecution must be able to date the arrest as early as it chooses following the development of probable cause.
The second possible source of confusion is the Court’s statement that “Officer Lasky did not engage in an unrestrained and thorough-going examination of Peters and his personal effects.” Ante, at 67. Since the Court found probable cause to arrest Peters, and since an officer arresting on probable cause is entitled to make a very full incident search,5 I assume that this is merely a factual observation. As a factual matter, I agree with it.
Although the articulable circumstances are somewhat less suspicious here than they were in Terry, I would affirm on the Terry ground that Officer Lasky had reason*78able cause to make a forced stop. Unlike probable cause to arrest, reasonable grounds to stop do not depend on any degree of likelihood that a crime has been committed. An officer may forcibly intrude upon an incipient crime even where he could not make an arrest for the simple reason that there is nothing to arrest anyone for. Hence although Officer Lasky had small reason to believe that a crime had been committed, his right to stop Peters can be justified if he had a reasonable suspicion that Peters was about to attempt burglary.
It was clear that the officer had to act quickly if he was going to act at all, and, as stated above, it seems to me that where immediate action is obviously required, a police officer is justified in acting on rather less objectively articulable evidence than when there is more time for consideration of alternative courses of action. Perhaps more important, the Court’s opinion in Terry emphasized the special qualifications of an experienced police officer. While “probable cause” to arrest or search has always depended on the existence of hard evidence that would persuade a “reasonable man,” in judging on-the-street encounters it seems to me proper to take into account a police officer’s trained instinctive judgment operating on a multitude of small gestures and actions impossible to reconstruct. Thus the statement by an officer that “he looked like a burglar to me” adds little to an affidavit filed with a magistrate in an effort to obtain a warrant. When the question is whether it was reasonable to take limited but forcible steps in a situation requiring immediate action, however, such a statement looms larger. A court is of course entitled to disbelieve the officer (who is subject to cross-examination), but when it believes him and when there are some articu-lable supporting facts, it is entitled to find action taken under fire to be reasonable.
*79Given Officer Lasky’s statement of the circumstances, and crediting his experienced judgment as he watched the two men, the state courts were entitled to conclude, as they did, that Lasky forcibly stopped Peters on “reasonable suspicion.” The frisk made incident to that stop was a limited one, which turned up burglar’s tools. Although the frisk is constitutionally permitted only in order to- protect the officer, if it is lawful the State is of course entitled to the use of any other contraband that appears.
For the foregoing reasons I concur in the result in these cases.
For a thoughtful study of many of these points, see ALI Model Code of Pre-Arraignment Procedure, Tentative Draft No. 1, §§ 2.01, 2.02, and the commentary on these sections appearing at 87-105.
It is true, as the Court states, that the New York courts attributed such a statement to him. The attribution seems to me unwarranted by the record.
E. g., Beck v. Ohio, 379 U. S. 89; Rios v. United States, 364 U. S. 253; Henry v. United States, 361 U. S. 98. In Henry, supra, at 100, the Court said that 18 U. S. C. § 3052 “states the constitutional standard” for felony arrests by FBI agents without warrant. That section authorized agents to “make arrests without warrant for any offense against the United States committed in their presence, or for any felony cognizable under the laws of the United States if they have reasonable grounds to believe that the person to be arrested has committed or is committing such felony.” Under Ker v. California, 374 U. S. 23, a parallel standard is applicable to warrantless arrests by state and local police.
Compare Henry v. United States, 361 U. S. 98, in which the Court said there was “far from enough evidence ... to justify a magistrate in issuing a warrant." Id., at 103. Agents knew that a federal crime, theft of whisky from an interstate shipment, had been committed “in the neighborhood.” Petitioner was observed driving into an alley, picking up packages, and driving away. I agree that these facts did not constitute probable cause, but find it hard to see that the evidence here was more impressive.
The leading case is United States v. Rabinowitz, 339 U. S. 56.