NORRIS, Circuit Judge:
Appellant Steven Winsor was found guilty by a jury of possessing forty dollars taken in a bank robbery.1 See 18 U.S.C. § 2113(c). On appeal he argues that his conviction rests on evidence obtained as a result of an unconstitutional search of the hotel room where he lived. A panel of this court affirmed the conviction, holding that although the police searched the room with only reasonable suspicion rather than probable cause, the search did not violate the Fourth Amendment because it was a minimal intrusion justified by important law enforcement interests. United States v. Winsor, 816 F.2d 1394 (9th Cir.1987). A [1571]*1571limited en banc panel was convened to review the case following a vote of the majority of our active judges. See United States v. Winsor, 822 F.2d 1466 (9th Cir.1987).
We now hold that Winsor’s motion to suppress must be granted because the search of Winsor’s room without probable cause violated the Fourth Amendment. The panel opinion and the judgment of conviction are, accordingly, vacated and the case is remanded to the district court for further proceedings.
I
In January 1986, Dennis Winsor (appellant’s brother, who is not a party to this appeal) robbed a bank in Hollywood and fled on foot. Los Angeles Police Department (LAPD) Officer Bowser saw him run out of the bank and pursued him to the nearby Chesterfield Hotel, a two-story low-rent residential hotel. In response to Bow-ser’s call for assistance, LAPD and FBI officers surrounded the hotel and a police helicopter circled above. Before entering the hotel, the officers were informed that no weapon had been seen during the robbery, although the robber had suggested he had a gun by holding his hand in his pocket.
The police decided to enter the hotel and to go from room to room looking for the suspect. At each room, two LAPD officers and an FBI agent, with their guns drawn, knocked on the door and announced: “Police. Open the door.” Three residents of the Chesterfield who somewhat fit the description of the bank robber were ordered to leave their rooms to go to the hotel manager’s office for identification. After checking all the rooms on the first floor and some of the rooms on the second floor (approximately fifteen to twenty-five rooms) the officers arrived at the room where Steven and Dennis Winsor were living. When the police knocked on the door and demanded that it be opened, Dennis Winsor obeyed. The police recognized him as the robber, pointed their guns at him, and told him to put his hands up, which he did. At this point, the police of course had probable cause to enter and to search the room, which they did. Inside, they found appellant Steven Winsor and evidence of the bank robbery. Both Winsors were arrested. While in custody, Steven made incriminating statements which he later moved to suppress along with the evidence found during the search of the room.
II
The district court denied Winsor’s motion to suppress. on the basis of the government’s argument that “hot pursuit” of the bank robber into the hotel provided sufficient cause to search each room in the hotel, including Winsor’s. We agree with the original panel that this was error. As the panel stated, “[h]ot pursuit may excuse police from the Fourth Amendment’s warrant requirement, but never does it excuse the absence of the requisite degree of suspicion before effecting a search.” 816 F.2d at 1396; see United States v. Scott, 520 F.2d 697, 700 (9th Cir.1975), cert. denied, 423 U.S. 1056, 96 S.Ct. 788, 46 L.Ed.2d 645 (1976).
The district court also ruled that the police conduct was immune from constitutional attack because the police operated with the consent of the hotel manager. Again, we agree with the original panel that this was error because a “hotel proprietor cannot waive his guests’ Fourth Amendment rights.” 816 F.2d at 1397 n. 3; see Stoner v. California, 376 U.S. 483, 489-90, 84 S.Ct. 889, 893, 11 L.Ed.2d 856 (1964).
The original panel nonetheless affirmed the district court’s denial of Winsor’s suppression motion. The panel decided that the search passed constitutional muster even though the police had only reasonable suspicion rather than probable cause to believe that the suspect would be in any one of the rooms that had not yet been searched, because the minimal intrusion on Winsor’s privacy rights was outweighed by important law enforcement interests. 816 F.2d at 1398-99.
[1572]*1572III
We now turn to the issues as defined by the parties during en banc briefing and argument. Before doing so, however, we find it helpful to clarify two points that are not in dispute. First, at the time the police knocked on Winsor’s door, they had reasonable suspicion to believe that the suspected bank robber was inside, but did not have probable cause to believe so. Second, each room of the hotel, including Winsor’s, enjoyed its own zone of Fourth Amendment protection, and Winsor’s expectation of privacy was not reduced simply because he lived in a single room in a low-rent hotel rather than in a single-family house or apartment.2
Winsor argues that the police effected a nonconsensual search of the room when they knocked on the door and commanded that it be opened under claim of lawful authority. He argues further that under Arizona v. Hicks, — U.S. -, 107 S.Ct. 1149, 94 L.Ed.2d 347 (1987), the search of his one-room residence could not be based on a level of suspicion less than probable cause.
The government argues that the police did not effect a search when they first viewed the interior of the room because they had not yet physically entered it. The government further argues that even if a search without probable cause was effected when the police looked through the open door into the interior of the room, that search was not unconstitutional because it “was not a full blown search for evidence” requiring probable cause, but was at most a “limited intrusion” requiring only reasonable suspicion. Appellee’s Supplemental Brief at 5. Such limited intrusions, the government contends, may be conducted on less than probable cause if the governmental interest in conducting the search outweighs the intrusion on the privacy interests of the subject of the search. Finally, the government appears to argue that even if the search of Winsor’s room constituted more than a limited intrusion, the overriding governmental interest in apprehending a suspected bank robber outweighed the intrusion on Winsor’s privacy interests and thus, under the circumstances, the search was constitutional.
IV
The threshold question we must decide is whether the police conducted a search within the purview of the Fourth Amendment when they looked into Winsor’s room through the open door while standing in the hotel corridor.
In essence, the government maintains that police do not effect a search of a home when they force open the front door and look inside without crossing the threshold. According to the government, “the cases are clear that without some form of ‘entry’ into the room, either physical or with the aid of electronic or sophisticated visual enhancement, a mere command to open the door does not transform a legitimate police procedure into a search.” Appellee’s Supplemental Brief at 2.
We agree with Winsor that this assertion “flies in the face of both precedent and common sense.” Appellant’s Supplemental Brief at 1. As the Supreme Court made clear in Katz v. United States, 389 U.S. 347, 353, 88 S.Ct. 507, 512, 19 L.Ed.2d 576 (1967), “the reach of [the Fourth] Amendment cannot turn upon the presence or absence of a physical intrusion into any given enclosure.” To draw a distinction based upon whether there had been a physical entry into the premises would enable police officers to evade the reach of the Fourth Amendment simply by forcing a door open and visually examining the interior without crossing the threshold. That the officers gained visual access to the interior of a dwelling without physically entering it is irrelevant to the question whether a search was effected. We find support for our position not only in Katz but also in United States v. Johnson, 626 F.2d 753 (9th Cir.1980), aff'd on other grounds, 457 U.S. 537, 102 S.Ct. 2579, 73 L.Ed.2d 202 (1982). In Johnson, police [1573]*1573standing outside the front door of a home arrested a person standing inside the door. We rejected the government’s argument that the arrest was not inside the home because the police had not crossed the threshold, reasoning that “it is the location of the arrested person, and not the arresting agents, that determines whether an arrest occurs within a home. Otherwise, arresting officers could avoid illegal ‘entry’ into a home simply by remaining outside the doorway and controlling the movements of suspects within_” 626 F.2d at 757. See also United States v. Karo, 468 U.S. 705, 714-15, 104 S.Ct. 3296, 3303, 82 L.Ed.2d 530 (1984) (holding that a warrant-less “search” of a home conducted by monitoring a radio transmitter violated the Fourth Amendment, noting, “[a]t the risk of belaboring the obvious, private residences are places in which the individual normally expects privacy free of governmental intrusion not authorized by a warrant, and that expectation is plainly one that society is prepared to recognize as justifiable”).
Neither of the cases relied upon by the government undermines our position. In United States v. Hersh, 464 F.2d 228, 229-30 (9th Cir.), cert. denied, 409 U.S. 1008, 93 S.Ct. 442, 34 L.Ed.2d 301 (1972), the police, while standing on the front porch, looked through a window and saw incriminating evidence inside the residence. We held no search was effected because police merely did what any member of the public was free to do — walk onto the front porch and observe whatever was in plain view through an unobstructed window. Similarly, in Davis v. United States, 327 F.2d 301, 303 (9th Cir.1964), the police did what any person could do — they knocked on the front door of a residence, but did not use their authority as police officers to command the occupants to open the door. When the occupant opened the door, he did so voluntarily, not, as Dennis Winsor did, in response to a claim of lawful authority.3
In sum, we hold that the police did effect a “search” when they gained visual entry into the room through the door that was opened at their command.
V
Having decided that the police searched Winsor’s room, we now consider the appropriate level of suspicion constitutionally required to justify the search. Winsor argues that Arizona v. Hicks, — U.S. -, 107 S.Ct. 1149, 94 L.Ed.2d 347 (1987), controls this case. In Hicks, the police were lawfully inside an apartment to investigate the source of gunfire reported by neighbors. While inside, one officer noticed an expensive stereo that seemed out of place in the poorly furnished apartment. Suspecting that the stereo was stolen, the officer moved the turntable slightly to note its serial number. The Supreme Court held that in moving the turntable, the officer effected a search of the apartment which was unrelated to the lawful objective of the [1574]*1574police entry, id. 107 S.Ct. at 1152, and that “[a] dwelling-place search, no less than a dwelling-place seizure, requires probable cause.” Id. at 1154. Winsor reads Hicks as establishing a rule that the police violate the Fourth Amendment whenever they effect a search of a dwelling without probable cause.
The government, in arguing that Hicks does not control this case, again relies on the fact that the police viewed the interior of the room without physically entering it. Although conceding that “[actual physical entry by the police into a home to conduct a search would be so intrusive that probable cause might be required,” the government contends that “where ... physical entry is lacking, the nature of the intrusion changes as a matter of kind, not degree, and the ‘dwelling place’ bright line test articulated by Hicks is inapplicable.” Appellee’s Supplemental Brief at 9 n. 4. In other words, the government argues that although Hicks established a bright-line rule requiring probable cause for dwelling-place searches, it is not controlling authority here because this search “was not a full blown search for evidence” requiring probable cause, but at most was “a limited intrusion” requiring only reasonable suspicion. Appellee’s Supplemental Brief at 5.
We agree with the government that Hicks apparently adopted a bright-line rule requiring probable cause to support a search of a dwelling.4 Unlike the government, however, we find nothing in Hicks to suggest that the Court, in establishing this bright-line rule, intended to apply it only to what the government labels “full-blown searches for evidence.” Since Hicks involved the least full-blown search imaginable — moving a phonograph to see its serial number — we cannot read it as suggesting that the bright-line rule applies only to dwelling-place searches that are broad in scope. Indeed, Justice Scalia, writing for the Court, expressly refused to make any such distinction between limited and full searches:
Justice O’Connor’s dissent suggests that we uphold the action here on the ground that it was a ‘cursory inspection’ rather than a ‘full-blown search,’ and could therefore be justified by reasonable suspicion instead of probable cause. [...] We are unwilling to send police and judges into a new thicket of Fourth Amendment law, to seek a creature of uncertain description that is neither a plain-view inspection nor yet a ‘full-blown search.’ Nothing in the prior opinions of this Court supports such a distinction. ...
107 S.Ct. at 1154.
The government cites no authority, and we know of none, that a search of a residence may constitute such a limited intrusion on Fourth Amendment interests that it may be justified by a degree of suspicion less than probable cause. The intrusiveness of a search is measured not by its scope, but by the expectation of privacy upon which the search intrudes. See infra Part VI. Because the expectation of privacy in one’s home is that most jealously guarded by the Fourth Amendment,5 the Supreme Court has never suggested that a search of a home, however limited in scope, could ever be considered less than a major intrusion. In the absence of such authority, we follow Hicks in holding that no search of a dwelling may be deemed a minor intrusion on Fourth Amendment rights.6
[1575]*1575VI
In the last analysis, the government’s position is that the search, however intrusive, passes constitutional muster because Winsor’s privacy interests were outweighed by the law enforcement interests at stake. This argument, premised on the bold assertion that “[n]o court has ever held that probable cause is a necessary requirement of a valid search in any given context,” Appellee’s Supplemental Brief at 7, is a radical one. Construing the two clauses of the Fourth Amendment — the proscription of “unreasonable searches and seizures” and the probable cause requirement — the government would apparently have us dispense with the probable cause requirement and judge every search by its reasonableness.7 The reasonableness of each search would be assessed by balancing the competing interests at stake. Thus, the government argues that “the appropriate constitutional test by which to assess [the] reasonableness [of the search of Winsor’s room] involves a ‘careful balancing of governmental and private interests.’ ” Appellee’s Supplemental Brief at 10 (quoting New Jersey v. T.L.O., 469 U.S. 325, 341, 105 S.Ct. 733, 742, 83 L.Ed.2d 720 (1985)). The government argues that the Fourth Amendment interests sacrificed in the search were outweighed by the law enforcement interests in apprehending the fleeing bank robber. Specifically, the government argues that “it is clear both from the circumstances requiring this exigent procedure and from the manner in which the search was actually conducted, that the search was a carefully limited one, and that it was the least intrusive means possible to accomplish the essential state interest of protecting the public from a potentially dangerous felon.” Appellee’s Supplemental Brief at 5-6.
In urging us to apply such a balancing analysis in this case, the government relies on the following passage from Justice White’s opinion in New Jersey v. T.L.O., 469 U.S. 325, 337, 105 S.Ct. 733, 740, 83 L.Ed.2d 720 (1985):
To hold that the Fourth Amendment applies to searches conducted by school authorities is only to begin the inquiry into the standards governing such searches. Although the underlying command of the Fourth Amendment is always that searches and seizures be reasonable, what is reasonable depends on the context within which a search takes place. The determination of the standard of reasonableness governing any specific class of searches requires “balancing the need to search against the invasion which the search entails.”
(quoting Camara v. Municipal Court, 387 U.S. 523, 537, 87 S.Ct. 1727, 1735, 18 L.Ed.2d 930 (1967)). The government further contends that Hicks implicitly contemplates the general application of such a balancing analysis, citing the following dicta:
We do not say, of course, that a seizure can never be justified on less than probable cause. We have held that it can— where, for example, the seizure is minimally intrusive and operational necessities render it the only practicable means of detecting certain types of crime.
Hicks, 107 S.Ct. at 1154. The government’s reading of T.L.O. and Hicks would expand the role of the balancing test in Fourth Amendment jurisprudence well beyond the limited place the Supreme Court has given it to date.
As we read T.L.O., Hicks, and other Supreme Court cases, the Court has developed a two-tier approach to the Fourth Amendment. The general rule is that seizures and searches must be supported by probable cause. At the same time, the Court has recognized a narrowly defined exception to this general rule. A level of suspicion less than probable cause may justify a search or seizure if the intrusion on Fourth Amendment interests is minimal, and if the minimal intrusion is outweighed [1576]*1576by the governmental interests served by the police action. Generally speaking, the Supreme Court has defined a minimally intrusive seizure as one that occurs in public and is brief. See, e.g., Terry v. Ohio, 392 U.S. 1, 88 S.Ct. 1868, 20 L.Ed.2d 889 (1968); United States v. Hensley, 469 U.S. 221, 229, 105 S.Ct. 675, 680, 83 L.Ed.2d 604 (1985). The Supreme Court has assessed the intrusiveness of a search, on the other hand, by considering whether it invades an “expectation of privacy ... that society is prepared to recognize as ‘reasonable.’ ” Katz, 389 U.S. at 361, 88 S.Ct. at 516 (Harlan, J., concurring). Specifically, the Court has upheld searches conducted on less than probable cause when they occur in certain clearly defined places which by their public nature give rise to reduced expectations of privacy. See infra pages 1577.
The genesis of the Court’s two-tier approach is, of course, Terry v. Ohio, 392 U.S. 1, 88 S.Ct. 1868, 20 L.Ed.2d 889 (1968). In Terry, a seizure case, the Court upheld the brief detention of a person in a public place and a “frisk” for weapons based on reasonable suspicion rather than probable cause. Since Terry, the Court has continued to refine its two-tier approach. For instance, writing for the Court in United States v. Place, 462 U.S. 696, 103 S.Ct. 2637, 77 L.Ed.2d 110 (1983), Justice O’Con-nor explained:
The exception to the probable-cause requirement for limited seizures of the person recognized in Terry and its progeny rests on a balancing of the competing interests to determine the reasonableness of the type of seizure involved within the meaning of “the Fourth Amendment’s general proscription against unreasonable searches and seizures.” We must balance the nature and quality of the intrusion on the individual’s Fourth Amendment interests against the importance of the governmental interests alleged to justify the intrusion. When the nature and extent of the detention are minimally intrusive of the individual’s Fourth Amendment interests, the opposing law enforcement interests can support a seizure based on less than probable cause.
Id. at 703, 103 S.Ct. at 2642 (quoting Terry, 392 U.S. at 20, 88 S.Ct. at 1879). In Place, the Court refused to apply the Terry reasonable suspicion standard because the 90-minute seizure of luggage in an airport was too lengthy to qualify as the sort of minimal intrusion contemplated by Terry. Id. at 709-10, 103 S.Ct. at 2645-46.
Similarly, in Dunaway v. New York, 442 U.S. 200, 99 S.Ct. 2248, 60 L.Ed.2d 824 (1979), the Court rejected the state’s argument that police may take a suspect to the police station without probable cause because the government’s interests in crime prevention and detection outweigh the intrusion on the suspect’s Fourth Amendment interests. Writing for the Court, Justice Brennan observed:
The narrow intrusions involved in [Terry and its progeny] were judged by a balancing test rather than by the general principle that Fourth Amendment seizures must be supported by the “long-prevailing standards” of probable cause ... only because these intrusions fell far short of the kind of intrusion associated with an arrest.
Id. at 212, 99 S.Ct. at 2256 (quoting Brinegar v. United States, 338 U.S. 160, 176, 69 S.Ct. 1302, 1311, 93 L.Ed. 1879 (1949)). The Court, expressed grave concern about the prospect of extending the balancing test to more intrusive police conduct:
[T]he protections intended by the Framers could all too easily disappear in the consideration and balancing of multifarious circumstances presented by different cases, especially when that balancing may be done in the first instance by police officers engaged in the “often competitive enterprise of ferreting out crime.” ... Indeed, our recognition of these dangers, and our consequent reluctance to depart from the proved protections afforded by the general rule, are reflected in the narrow limitations emphasized in the cases employing the balancing test. For all but those narrowly defined intrusions, the requisite ‘balancing’ has been performed in centuries [1577]*1577of precedent and embodied in the principle that seizures are ‘reasonable’ only if supported by probable cause.
Id. at 213-14, 99 S.Ct. at 2257 (quoting Johnson v. United States, 333 U.S. 10, 14, 68 S.Ct. 367, 369, 92 L.Ed. 436 (1948)). See also Michigan v. Summers, 452 U.S. 692, 697-98, 101 S.Ct. 2587, 2591, 69 L.Ed.2d 340 (1981) (“[S]ome seizures significantly less intrusive than an arrest have withstood scrutiny under the reasonableness standard embodied in the Fourth Amendment. In these cases the intrusion on the citizen’s privacy ‘was so much less severe’ than that involved in a traditional arrest that ‘the opposing interests in crime prevention and detection and in the police officer’s safety’ could support the seizure as reasonable.”) (quoting Dunaway, 442 U.S. at 209, 99 S.Ct. at 2254).8
The Court’s two-tier approach also becomes apparent from a comparison of cases in which the Supreme Court has and has not upheld a search on less than probable cause. As Hicks and T.L.O. nicely illustrate, in the search context, the level of suspicion required to justify police action turns on the expectation of privacy that society will recognize in the place in which the search occurred. In Hicks, the Court applied the general rule requiring probable cause because the place searched — a residence — is a place where societally sanctioned expectations of privacy are at their strongest. 107 S.Ct. at 1154. In T.L.O., by contrast, the balancing test was applied and the search was sustained on the basis of reasonable suspicion because the search took place in a public school, where privacy interests are necessarily reduced. 469 U.S. at 338, 105 S.Ct. at 741.
The cases in which the Court has applied a balancing test to determine whether a search could be sustained on less than probable cause have involved, without exception, searches that occurred in places where expectations of privacy are necessarily reduced. Thus, the Court authorized a governmental employer to search, for work-related reasons and upon reasonable suspicion, an employee’s office, noting that “the privacy interests of government employees in their places of work ..., while not insubstantial, are far less than those found at home or in some other contexts.” O’Connor v. Ortega, — U.S. -, 107 S.Ct. 1492, 1502, 94 L.Ed.2d 714 (1987) (plurality opinion). Similarly, the Court has held that authorities may search, on less than probable cause, the place of business of an organization in a pervasively regulated industry because the regulation reduces the expectation of privacy. See New York v. Burger, — U.S. -, 107 S.Ct. 2636, 2644, 96 L.Ed.2d 601 (1987); United States v. Biswell, 406 U.S. 311, 316, 92 S.Ct. 1593, 1596, 32 L.Ed.2d 87 (1972). See also Cady v. Dombrowski, 413 U.S. 433, 448, 93 S.Ct. 2523, 2531, 37 L.Ed.2d 706 (1973) (upholding search of a car trunk upon reasonable suspicion after owner no longer had possession of car); South Dakota v. Opperman, 428 U.S. 364, 367, 96 S.Ct. 3092, 3096, 49 L.Ed.2d 1000 (1976) (upholding an inventory search of an impounded automobile on less than probable cause); United States v. Villamonte-Marquez, 462 U.S. 579, 593, 103 S.Ct. 2573, 2582, 77 L.Ed.2d 22 (1983) (customs officials may, on less than probable cause, stop and board vessels on inland waterways with access to the open seas to examine documentation because the brief detention of the vessel and inspection of [1578]*1578the public deck area are minimally intrusive); Bell v. Wolfish, 441 U.S. 520, 559-60, 99 S.Ct. 1861, 1884-85, 60 L.Ed.2d 447 (1979) (upholding routine visual inspections of body cavities of pretrial detainees because incarcerated persons of necessity relinquish expectations of privacy that others enjoy); United States v. Ramsey, 431 U.S. 606, 619, 97 S.Ct. 1972, 1980, 52 L.Ed.2d 617 (1977) (upholding routine searches of people and things crossing the border, based on Congress’s plenary power to control the borders); Balelo v. Baldridge, 724 F.2d 753, 764-66 (9th Cir.) (en banc) (government observers may board regulated fishing boats on a scheduled basis without probable cause, but may not search persons, personal effects, or living quarters of the crew), cert. denied, 467 U.S. 1252, 104 S.Ct. 3536, 82 L.Ed.2d 841 (1984).
The Supreme Court has consistently refused, however, to engage in a balancing of competing interests or to sustain a search on less than probable cause when the search occurred in a place where society recognizes a strong interest in privacy. As noted above, the Court in Hicks required probable cause to move a phonograph because the search occurred in an apartment. Similarly, in United States v. Ortiz, 422 U.S. 891, 896, 95 S.Ct. 2585, 2588, 45 L.Ed.2d 623 (1975), the Court, noting that a full search of a car “is a substantial invasion of privacy,” held that probable cause is required to search an automobile at a traffic checkpoint even though probable cause is not required to stop the vehicle. In the same vein, the Court held in Almeida-Sanchez v. United States, 413 U.S. 266, 273, 93 S.Ct. 2535, 2539, 37 L.Ed.2d 596 (1973), that a roving unit of the border patrol must have probable cause to search vehicles in the vicinity of the border. See also Ybarra v. Illinois, 444 U.S. 85, 94-96, 100 S.Ct. 338, 343-45, 62 L.Ed.2d 238 (1979) (rejecting argument that reasonable suspicion justifies search of a person detained on premises during execution of warrant to search premises); Sibron v. New York, 392 U.S. 40, 65-66, 88 S.Ct. 1889, 1904, 20 L.Ed. 2d 917 (1968) (reaching into suspect’s pocket in search of evidence rather than a weapon is more intrusive than a Terry “frisk” for weapons and cannot be sustained on less than probable cause); Berger v. New York, 388 U.S. 41, 58-59, 87 S.Ct. 1873, 1883-84, 18 L.Ed.2d 1040 (1967) (statute authorizing electronic eavesdropping on less than probable cause violates the Fourth Amendment).
That the Supreme Court has limited the balancing approach to discrete places where expectations of privacy are necessarily reduced is not surprising. As Justice Scalia, the crucial fifth vote in O’Connor v. Ortega, — U.S. -, 107 S.Ct. 1492, 94 L.Ed.2d 714 (1987), admonished the four-Justice plurality, a fact-specific case-by-case approach would plunge courts into a neverending and essentially standardless assessment of every search:
The plurality opinion instructs the lower courts that existence of Fourth Amendment protection for a public employee’s business office is to be assessed “on a case-by-case basis,” in light of whether the office is “so open to fellow employees or the public that no expectation of privacy is reasonable.” No clue is provided as to how open “so open” must be; much less is it suggested how police officers are to gather the facts necessary for this refined inquiry. As we observed in Oliver v. United States, 466 U.S. 170, 181 [104 S.Ct. 1735, 1743, 80 L.Ed.2d 214] (1984), “[t]his Court repeatedly has acknowledged the difficulties created for courts, police, and citizens by an ad hoc, case-by-case definition of Fourth Amendment standards to be applied in differing factual circumstances.” Even if I did not disagree with the plurality as to what result the proper legal standard should produce in the case before us, I would object to the formulation of a standard so devoid of content that it produces rather than eliminates uncertainty in this field.
107 S.Ct. at 1504-05 (Scalia, J., concurring).9
[1579]*1579We appreciate the government’s concern about excessive constitutional restrictions on police conduct. We must echo, however, the Supreme Court’s recent reminder in Hicks:
[TJhere is nothing new in the realization that the Constitution sometimes insulates the criminality of a few in order to protect the privacy of us all.... [W]e choose to adhere to the textual and traditional standard of probable cause.
Hicks, 107 S.Ct. at 1155. With this admonition in mind, we refuse the government’s invitation to decide this case by balancing the competing interests at stake. Instead, we adhere to the bright-line rule that Hicks appears to have announced: The Fourth Amendment prohibits searches of dwellings without probable cause.
We also reject the government’s argument, raised for the first time in its supplemental brief filed after en banc argument, that the fruits of the warrantless search should be admissible in evidence because the police conducted the search in good faith. See United States v. Leon, 468 U.S. 897, 104 S.Ct. 3405, 82 L.Ed.2d 677 (1984). The Supreme Court has applied the so-called “good faith” exception to the exclusionary rule only to searches conducted in good faith reliance on a warrant or a statute later declared to be unconstitutional. See Illinois v. Krull, — U.S. -, 107 S.Ct. 1160, 94 L.Ed.2d 364 (1987). We decline to extend Leon’s, good faith exception to searches not conducted in reliance on a warrant or a statute. See United States v. Whiting, 781 F.2d 692, 698-99 (9th Cir.1986) (refusing to apply Leon exception to search conducted in reliance on regulations later determined not to authorize the search at issue).
The district court’s order denying Win-sor’s motion to suppress is REVERSED. The panel opinion and the judgment of conviction are VACATED and the case is REMANDED to the district court for further proceedings consistent with this opinion.