concurring in part and dissenting in part.
I find myself in disagreement with the Court’s disposition of this case in two respects.
I
I cannot join the Court’s casual treatment of the issue that has been presented by both parties as the major issue in this case: petitioner’s claim that he received ineffective assistance of counsel at his trial. As the Court acknowledges, petitioner met Mr. Tamburo, his trial counsel, for the first time en route to the courtroom on the morning of trial. Although a different Legal Aid Society attorney had represented petitioner at his first trial, apparently neither he nor anyone else from the society had conferred with petitioner in the interval between trials. Because the District Court did not hold an evidentiary hearing on the habeas petition, there is no indication in the record of the extent to which Mr. Tamburo may have consulted petitioner’s previous attorney, the attorneys for the other defendants, or the files of the Legal Aid Society. What the record does disclose on this claim is essentially a combination of two factors: the entry of counsel into the case immediately *56before trial, and his handling of the issues that arose during the trial.1
As respondent must concede, counsel’s last-minute entry into the case precluded his compliance with the state rule requiring that motions to suppress evidence be made before trial, even assuming that he had sufficient acquaintance with the case to know what arguments were worth making. Furthermore, the record suggests that he may have had virtually no such acquaintance.
In the first place, he made no objection to the admission in evidence of the objects found during the search of the car at the station house after the arrest of its occupants, although that search was of questionable validity under Fourth Amendment standards, see infra.
Second, when the prosecution offered in evidence the bullets found in the search of petitioner’s home, which had been excluded on defense objection at the first trial, Mr. Tamburo objected to their admission, but in a manner that suggested that he was a stranger to the facts of the case. While he indicated that he did know of the earlier exclusion, he apparently did not know on what ground the bullets had been excluded, and based his *57objection only on their asserted irrelevance.2 Later in the trial he renewed his objection on the basis of the inadequacy of the warrant, stating, “I didn’t know a thing about the search Warrant until this morning.” App. 130.3
Third, when prosecution witness Havicon made an in-court identification of petitioner as the man who had *58threatened him with a gun during one of the robberies, Mr. Tamburo asked questions in cross-examination that suggested that he had not had time to settle upon a trial strategy or even to consider whether petitioner would take the stand. Mr. Tamburo asked whether, at a pretrial lineup, a detective had not told Havicon that petitioner “was the man with the gun.” After Havicon’s negative answer, this colloquy ensued:
“THE COURT: I take it you will be able to disprove that, will you?
“MR. TAMBURO: What?
“THE COURT: You shouldn’t ask that question unless you are prepared to disprove that, contradict him.
“MR. TAMBURO: I have the defendant’s testimony.
“THE COURT: Disprove it in any way at all.
“MR. MEANS [the prosecutor]: I don’t understand how the defendant would know what the detectives told him.
“THE COURT: He said he is going to disprove it by the defendant, that’s all right, go ahead.” App. 34.
The next witness was a police officer who had been present at the lineup, and who testified that no one had told Havicon whom to pick out. Petitioner’s counsel did not cross-examine, and petitioner never took the stand.
On this state of the record the Court of Appeals ruled that, although the late appointment of counsel necessitated close scrutiny into the effectiveness of his representation, petitioner “was not prejudiced by the late appointment of counsel” because neither of the Fourth Amendment claims belatedly raised justified reversal of *59the conviction. 408 F. 2d 1186, 1196. I agree that the strength of the search and seizure claims is an element to be considered in the assessment of whether counsel was adequately prepared to make an effective defense, but I cannot agree that the relevance of those claims in this regard disappears upon a conclusion by an appellate court that they do not invalidate the conviction.
This Court recognized long ago that the duty to provide counsel “is not discharged by an assignment at such a time or under such circumstances as to preclude the giving of effective aid in the preparation and trial of the case.” Powell v. Alabama, 287 U. S. 45, 71 (1932); Hawk v. Olson, 326 U. S. 271, 278 (1945). While “the Constitution nowhere specifies any period which must intervene between the required appointment of counsel and trial,” the Court has recognized that
“the denial of opportunity for appointed counsel to confer, to consult with the accused and to prepare his defense, could convert the appointment of counsel into a sham and nothing more than a formal compliance with the Constitution's requirement that an accused be given the assistance of counsel.” Avery v. Alabama, 308 U. S. 444, 446 (1940).
Where counsel has no acquaintance with the facts of the ease and no opportunity to plan a defense, the result is that the defendant is effectively denied his constitutional right to assistance of counsel.
It seems to me that what this record reveals about counsel’s handling of the search and seizure claims and about the tenor of his cross-examination of the government witness Havicon, when coupled with his late entry into the case, called for more exploration by the District Court before petitioner’s ineffective assistance of counsel claim could be dismissed. Such an exploration should *60have been directed to ascertaining whether the circumstances under which Mr. Tamburo was required to undertake petitioner’s defense at the second trial were such as to send him into the courtroom with so little knowledge of the case as to render him incapable of affording his client adequate representation. The event of that exploration would turn, not on a mere assessment of particular missteps or omissions of counsel, whether or not caused by negligence, cf. McMann v. Richardson, 397 U. S. 759 (1970), but on the District Court’s evaluation of the total picture, with the objective of determining whether petitioner was deprived of rudimentary legal assistance. See Williams v. Beto, 354 F. 2d 698 (C. A. 5th Cir. 1965). And, of course, such an exploration would not be confined to the three episodes that, in my opinion, triggered the necessity for a hearing.
It is not an answer to petitioner’s claim for a reviewing court simply to conclude that he has failed after the fact to show that, with adequate assistance, he would have prevailed at trial. Glasser v. United States, 315 U. S. 60, 75-76 (1942); cf. White v. Maryland, 373 U. S. 59 (1963); Reynolds v. Cochran, 365 U. S. 525, 530-533 (1961). Further inquiry might show, of course, that counsel’s opportunity for preparation was adequate to protect petitioner’s interests,4 but petitioner did, in my view, raise a sufficient doubt on that score to be entitled to an evidentiary hearing.5
*61II
In sustaining the search of the automobile I believe the Court ignores the framework of our past decisions circumscribing the scope of permissible search without a warrant. The Court has long read the Fourth Amendment's proscription of “unreasonable” searches as imposing a general principle that a search without a warrant is not justified by the mere knowledge by the searching officers of facts showing probable cause. The “general requirement that a search warrant be obtained” is basic to the Amendment’s protection of privacy, and “ 'the burden is on those seeking [an] exemption ... to show the need for it.’ ” E. g., Chimel v. California, 395 U. S. 752, 762 (1969); Katz v. United States, 389 U. S. 347, 356-358 (1967); Warden v. Hayden, 387 U. S. 294, 299 (1967); Preston v. United States, 376 U. S. 364, 367 (1964); United States v. Jeffers, 342 U. S. 48, 51 (1951); McDonald v. United States, 335 U. S. 451, 455-456 (1948); Agnello v. United States, 269 U. S. 20, 33 (1925).
Fidelity to this established principle requires that, where exceptions are made to accommodate the exigencies of particular situations, those exceptions be no broader than necessitated by the circumstances presented. For example, the Court has recognized that an arrest creates an emergency situation justifying a war-rantless search of the arrestee’s person and of “the area from within which he might gain possession of a weapon or destructible evidence”; however, because the exigency giving rise to this exception extends only that far, the search may go no further. Chimel v. California, 395 U. S., at 763; Trupiano v. United States, 334 U. S. 699, 705, 708 (1948). Similarly we held in Terry v. Ohio, 392 U. S. 1 (1968), that a warrantless search in a “stop and frisk” situation must “be strictly circum*62scribed by the exigencies which justify its initiation.” Id., at 26. Any intrusion beyond what is necessary for the personal safety of the officer or others nearby is forbidden.
Where officers have probable cause to search a vehicle on a public way, a further limited exception to the warrant requirement is reasonable because “the vehicle can be quickly moved out of the locality or jurisdiction in which the warrant must be sought.” Carroll v. United States, 267 U. S. 132, 153 (1925). Because the officers might be deprived of valuable evidence if required to obtain a warrant before effecting any search or seizure, I agree with the Court that they should be permitted to take the steps necessary to preserve evidence and to make a search possible.6 Cf. ALI, Model Code of Pre-Arraignment Procedure § 6.03 (Tent. Draft No. 3, 1970). The Court holds that those steps include making a warrantless search of the entire vehicle on the highway — a conclusion reached by the Court in Carroll without discussion — and indeed appears to go further and to condone the removal of the car to the police station for a warrantless search there at the convenience of the police.7 I cannot agree that this result is con*63sistent with our insistence in other areas that departures from the warrant requirement strictly conform to the exigency presented.
The Court concedes that the police could prevent removal of the evidence by temporarily seizing the car for the time necessary to obtain a warrant. It does not dispute that such a course would fully protect the interests of effective law enforcement; rather it states that whether temporary seizure is a “lesser” intrusion than warrantless search “is itself a debatable question and the answer may depend on a variety of circumstances.” Ante, at 51-52.8 I believe it clear that a war-rantless search involves the greater sacrifice of Fourth Amendment values.
The Fourth Amendment proscribes, to be sure, unreasonable “seizures” as well as “searches.” However, in the circumstances in which this problem is likely to occur, the lesser intrusion will almost always be the simple seizure of the car for the period — perhaps a day— necessary to enable the officers to obtain a search warrant. In the first place, as this case shows, the very facts establishing probable cause to search will often *64also justify arrest of the occupants of the vehicle. Since the occupants themselves are to be taken into custody, they will suffer minimal further inconvenience from the temporary immobilization of their vehicle. Even where no arrests are made, persons who wish to avoid a search— either to protect their privacy or to conceal incriminating evidence- — -will almost certainly prefer a brief loss of the use of the vehicle in exchange for the opportunity to have a magistrate pass upon the justification for the search. To be sure, one can conceive of instances in which the occupant, having nothing to hide and lacking concern for the privacy of the automobile, would be more deeply offended by a temporary immobilization of his vehicle than by a prompt search of it. However, such a person always remains free to consent to an immediate search, thus avoiding any delay. Where consent is not forthcoming, the occupants of the car have an interest in privacy that is protected by the Fourth Amendment even where the circumstances justify a temporary seizure. Terry v. Ohio, supra. The Court’s endorsement of a warrantless invasion of that privacy where another course would suffice is simply inconsistent with our repeated stress on the Fourth Amendment’s mandate of “ ‘adherence to judicial processes.’ ” E. g., Katz v. United States, 389 U. S., at 357.9
Indeed, I believe this conclusion is implicit in the opinion of the unanimous Court in Preston v. United *65States, 376 U. S. 364 (1964). The Court there purported to decide whether a factual situation virtually identical to the one now before us was “such as to fall within any of the exceptions to the constitutional rule that a search warrant must be had before a search may be made.” Id., at 367 (emphasis added). The Court concluded that no exception was available, stating that “since the men were under arrest at the police station and the car was in police custody at a garage, [there was no] danger that the car would be moved out of the locality or jurisdiction.” Id., at 368. The Court’s reliance on the police custody of the car as its reason for holding “that the search of the car without a warrant failed to meet the test of reasonableness under the Fourth Amendment,” ibid., can only have been based on the premise that the more reasonable course was for the police to retain custody of the car for the short time necessary to obtain a warrant. The Court expressly did not rely, as suggested today, on the fact that an arrest for vagrancy provided “no cause to believe that evidence of crime was concealed in the auto.” Ante, at 47; see 376 U. S., at 368; Wood v. Crouse, 417 F. 2d 394, 397-398 (C. A. 10th Cir. 1969). The Court now discards the approach taken in Preston, and creates a special rule for automobile searches that is seriously at odds with generally applied Fourth Amendment principles.
Ill
The Court accepts the conclusion of the two courts below that the introduction of the bullets found in petitioner’s home, if error, was harmless. Although, as explained above, I do not agree that this destroys the relevance of the issue to the ineffectiveness of counsel claim, I agree that the record supports the lower courts’ conclusion that this item of evidence, taken alone, was harmless beyond a reasonable doubt.
Respondent concedes in this Court that “no other facts are available to determine the amount and the quality of the preparation for trial pursued by Mr. Tamburo or the amount of evidentiary material known by and available to him in determining what, if any, evidentiary objections were mandated or what, if any, defenses were available to petitioner.” Brief for Respondent 13. The Court of Appeals stated: “We do not know what preparation, if any, counsel was able to accomplish prior to the date of the trial as he did not testify in the state habeas corpus proceeding and there was no evidentiary hearing in the district court. From the lower court opinion, as will appear later, we are led to believe that counsel was not wholly familiar with all aspects of the case before trial.” 408 F. 2d 1186, 1191.
Mr. Tamburo stated to the trial court:
“Your Honor, at the first trial, the District Attorney attempted to introduce into evidence some .38 calibre bullets that were found at the Chambers’ home after his arrest. ... At that trial, it was objected to and the objection was sustained, and I would also like to object to it now, I don’t think it is good for the Jury to hear it. I don’t feel there is any relevancy or connection between the fact there were .38 calibre bullets at his home and the fact that a .38 calibre gun was found, not on the person of Chambers, but in the group.” App. 82.
This was the only instance in which Mr. Tamburo expressed any knowledge of what had transpired at the first trial, and it does not appear whether he learned of the exclusion from his brief talk with petitioner en route to the courtroom or from sources within the Legal Aid Society. The record does not disclose the reason for the exclusion of the bullets at the first trial.
This colloquy followed the renewed objection:
“THE COURT: Well, of course, you have known about this from the other trial three weeks ago.
“MR. TAMBURO: I wasn’t the attorney at the other trial.
“THE COURT: But, you knew about it?
“MR. TAMBURO: I didn’t know a thing about the search Warrant until this morning.
“THE COURT: You knew about the evidence about to be introduced, you told me about it.
“MR. TAMBURO: It wasn’t admitted.
“THE COURT: That doesn’t mean I have to exclude it now.” Id., at 130.
The court proceeded to overrule the objection on the ground that it had not been made in a pretrial motion, adding that “I think there is reasonable ground for making a search here, even without a Warrant.” Id., at 130-131.
In Avery, this Court concluded on the basis of a hearing: “That the examination and preparation of the case, in the time permitted by the trial judge, had been adequate for counsel to exhaust its every angle is illuminated by the absence of any indication, on the motion and hearing for new trial, that they could have done more had additional time been granted.” 308 U. S., at 452.
The absence of any request by counsel for a continuance of the trial should not, in my opinion, serve to vitiate petitioner’s claim at this juncture.
Where a suspect is lawfully arrested in the automobile, the officers may, of course, perform a search within the limits prescribed by Chimel as an incident to the lawful arrest. However, as the Court recognizes, the search here exceeded those limits. Nor was the search here within the limits imposed by pre-Chimel law for searches incident to arrest; therefore, the retroactivity of Chimel is not drawn into question in this case. See Preston v. United States, 376 U. S. 364 (1964).
The Court disregards the fact that Carroll, and each of this Court's decisions upholding a warrantless vehicle search on its authority, involved a search for contraband. Brinegar v. United States, 338 U. S. 160 (1949); Scher v. United States, 305 U. S. 251 (1938); Husty v. United States, 282 U. S. 694 (1931); see United States v. Di Re, 332 U. S. 581, 584-586 (1948). Although subsequent dicta have omitted this limitation, see Dyke v. Taylor Imple*63ment Mfg. Co., 391 U. S. 216, 221 (1968); United States v. Ventresca, 380 U. S. 102, 107 n. 2 (1965); United States v. Rabinowitz, 339 U. S. 56, 61 (1950), id., at 73 (Frankfurter, J., dissenting), the Carroll decision has not until today been held to authorize a general search of a vehicle for evidence of crime, without a warrant, in every case where probable cause exists.
The Court, unable to decide whether search or temporary seizure is the “lesser” intrusion, in this case authorizes both. The Court concludes that it was reasonable for the police to take the car to the station, where they searched it once to no avail. The searching officers then entered the station, interrogated petitioner and the car’s owner, and returned later for another search of the car — this one successful. At all times the car and its contents were secure against removal or destruction. Nevertheless, the Court approves the searches without even an inquiry into the officers’ ability promptly to take their case before a magistrate.
Circumstances might arise in which it would be impracticable to immobilize the ear for the time required to obtain a warrant — for example, where a single police officer must take arrested suspects to the station, and has no way of protecting the suspects’ car during his absence. In such situations it might be wholly reasonable to perform an on-the-spot search based on probable cause. However, where nothing in the situation makes impracticable the obtaining of a warrant, I cannot join the Court in shunting aside that vital Fourth Amendment safeguard.