ERVIN, Chief Judge:

Defendant Eldon Han was convicted in federal district court of conspiracy to distribute heroin. He argues that the district court improperly admitted evidence discovered in a warrantless search of a bag that was in his possession when he was arrested. The district court found that probable cause existed for the arrest, the search was incident to the arrest, and Han abandoned his privacy interest in the bag. We agree, and therefore affirm.

I.

We must decide at the outset what facts we may consider. Han contends that, in our probable-cause inquiry, we should consider only the evidence that was before the district court at the suppression hearing. See 4 Wayne R. LaFave, Search and Seizure Section(s) 11.7(c), at 520 (2d ed. 1987) (noting that "appellate court may be accepting as true certain testimony which conceivably neither the trial judge nor the jury believed or which neither judge or jury had any reason to assess."). His position is supported by a few state-court decisions, see Glover v. State, 287 A.2d 333, 336 (Md.App. 1972), overruled by Goode v. State, 398 A.2d 801 (Md.App. 1979); People v. Williams, 118 N.W.2d 391, 409 (Mich. 1962), cert. denied, 373 U.S. 909 (1963), but federal courts have held uniformly that an appellate tribunal may consider evidence adduced at trial that supports the district judge's ruling. See Carroll v. United States, 267 U.S. 132, 162 (1925); United States v. Villabona-Garnica, 63 F.3d 1051, 1055 (11th Cir. 1995); United States v. Rios, 611 F.2d 1335, 1344 n. 14 (10th Cir. 1995); Washington v. United States, 401 F.2d 915, 919 n. 19 (D.C. Cir. 1968); United States v. Longmire, 761 F.2d 411, 418; United States v. McKinney, 379 F.2d 259, 264 (6th Cir. 1967); Rent v. United States, 209 F.2d 893, 896 (5th Cir. 1954). Thus we consider testimony from both the hearing and the trial.

On March 28, 1994, in Washington, D.C., DEA agents arrested Reece Whiting on heroin charges. Whiting agreed to cooperate in the DEA's continuing investigation. He identified Tom Riley of San Dimas, California, as one of his sources. At the DEA's behest he called Riley, purportedly to arrange a meeting between Riley and a prospective buyer. On April 19, Whiting and the "prospective buyer" — DEA Special Agent Lisa Somers — met Riley at the Dulles Airport Marriott in Loundoun County, Virginia. Riley told Somers that he could supply two pounds of "white" heroin for $144,000. He said that, after seeing her money, he would call two Asian suppliers in Honolulu who would bring the heroin to the Washington area. Later that day, by telephone, Riley told Somers that his suppliers would not come to Washington, and he suggested that she travel to California.

On May 3, 1994, Somers and Special Agent Clyde Shelley went to California. Riley agreed to complete the transaction at the Hilton Hotel in Ontario, California, which is just outside Los Angeles. Somers, Shelley, and Riley met at the Hilton at about 5:15 on May 4. Riley said that his source was in Los Angeles, but would be able to supply only one and one-half pounds. He said the source wanted to conduct the sale in half-pound increments: Riley would pick up the source at his hotel and take him to Riley's residence; Riley then would bring one-half pound of heroin to the hotel, exchange it, and return the money to his source; he would repeat the process twice. Somers agreed to the arrangement, and Riley left the hotel.

Special Agent Michael Orton followed Riley from the Hilton. Riley stopped briefly at his residence, and then drove to the LAX Holiday Inn. Orton lost surveillance of Riley at the Holiday Inn; knowing by radio of Riley's plans, he returned to Riley's residence. Riley arrived before Orton did, and none of the agents conducting the surveillance actually saw whether anyone accompanied Riley into his residence.

At approximately 9 p.m., Somers called Riley at his residence. Riley told her that the source was with him and that they were ready to complete the deal. He said he would meet Somers at the Hilton at 9:45. At about 9:05, Special Agent James Burns saw Riley speaking with an Asian male in front of his residence. He conveyed his observations, including a description of the man's clothing, to Orton. Riley arrived at the Hilton on time, handed one-half pound of heroin to Shelley, and was arrested immediately. He agreed to cooperate, and gave the agents permission to search his residence.

Accompanied by other agents, Orton knocked on the door of Riley's home. Riley's roommate, Laurie Bennett, answered and allowed the agents to enter. Orton saw Han seated on the living room couch, with a travel bag next to his feet. Orton asked if he could move the bag for safety purposes, and Han agreed; a sheriff's deputy moved it out of Han's reach. A sweep of the residence revealed that Bennett and Han were the only people inside. At that point, Orton testified, his safety concerns were alleviated.

Orton then interviewed Bennett in a back room. She told him that Han had brought the bag into the residence and it never had left his side. Orton returned to the living room, sat down next to Han, and picked up the bag. He asked Han if he could look at his bag. Han responded that it was not his bag, and Orton asked if Han had a problem with Orton looking inside. Han said that he did not. Inside the bag, Orton found heroin and a wallet containing Han's driver's license and other identification.

Han was charged with conspiracy to distribute heroin in violation of 21 U.S.C. § 846. He was tried and convicted in the Eastern District of Virginia, and he timely appealed to this court under 28 U.S.C. § 1291. Han raises two primary arguments to support his contention that the search was not a proper search incident to arrest. He argues that the arrest was invalid because the officers did not have probable cause before the search; and even if the arrest was valid, he contends, the search was not "incident" to the arrest because the search occurred before the arrest, the bag had been moved away from him, and the agents already had alleviated their safety concerns. The government responds that the officers did have probable cause before the search, and that the search was incident to the arrest. Alternatively, it asserts that Han abandoned his expectation of privacy in the bag, so Han had no standing to challenge the search.

II.

This court reviews pure questions of law de novo and pure questions of fact for clear error. United States v. McDonald, 61 F.3d 248, 254 (4th Cir. 1995). Mixed questions of law and fact are evaluated under a hybrid standard. The court reviews the trial judge's ultimate conclusions de novo, but in reaching its independent resolutions construes the evidence in the manner most favorable to the government. See United States v. Elwood, 993 F.2d 1146, 1151 (5th Cir. 1993).

A.

Han argues that the arrest was invalid because it was prompted by the evidence Orton found in the bag. Because a search incident to arrest is permitted only when there is a valid arrest, the validity of the arrest cannot depend on evidence found during the search. This was not a search incident to arrest, Han asserts, but an "arrest incident to illegal search." Han is correct that the legality of the arrest cannot depend on the search, but it does not follow that the arrest must precede the search. A search may be incident to a subsequent arrest if the officers have probable cause to arrest before the search. Rawlings v. Kentucky, 448 U.S. 98, 111 (1980), cited in United States v. Miller, 925 F.2d 695, 698 (4th Cir.), cert. denied, 502 U.S. 833 (1991).

This court, following the Supreme Court, has defined probable cause to arrest as

facts and circumstances within the officer's knowledge that are sufficient to warrant a prudent person, or one of reasonable caution, in believing, in the circumstances shown, that the suspect has committed, is committing, or is about to commit an offense. The evidence needed to establish probable cause is more than a mere suspicion, rumor, or strong reason to suspect but less than evidence sufficient to convict.

United States v. Williams, 10 F.3d 1070, 1073-74 (4th Cir.) (internal quotations and citations omitted), cert. denied, 115 S.Ct. 313 (1994). Examined in the light most favorable to the government, the evidence reveals that the agents knew the following facts before the search:

* Riley described his source as Asian;

* Riley said that he picked up his source and took him to his residence;

* Riley said that, when he delivered the first half-pound of heroin to the Hilton, his source would remain at his residence with the remaining pound;

* Shortly before Riley left for the Hilton, an agent saw him talking outside his residence with a man who appeared to be Asian. The agent conveyed the information and described the Asian man's clothing to Orton;

* Upon his arrest at the Hilton, Riley said that his source was still at his residence;

* The agents' search of the residence revealed that Han was the only person there other than Riley's roommate and the officers.

Applying the law to that evidence de novo, we conclude that the officers had sufficient information to support a reasonable belief that Han was Riley's source.

Han argues alternatively that, even if the evidence was sufficient to support a reasonable belief that Han was the source, Orton did not actually harbor such a belief until he searched the bag. Thus, Han contends, the arrest was invalid. That argument fails. The Supreme Court's definition of probable cause asks not whether the arresting officer reasonably believed that the arrestee had committed a crime, but whether the evidence was sufficient to support such a reasonable belief. DeFillippo, 443 U.S. at 37. The test is objective, so Orton's subjective state of mind is irrelevant. The evidence created probable cause, so the arrest was valid.

B.

Even if the arrest was valid, Han argues, the search was not incident to the arrest. The leading authority on incidence to arrest is United States v. Chimel. 395 U.S. 752 (1969). The Supreme Court held in Chimel that the need to protect officers' safety and prevent destruction of evidence justifies "a search of the arrestee'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. Han contends that, at the time of the search, the bag was out of Han's area of immediate control and the officers' safety concerns already had been alleviated.

Since Chimel, the Supreme Court has interpreted broadly both the area under "immediate control" and the likelihood of danger or destruction of evidence. In New York v. Belton it admitted cocaine found in the zipped pocket of a jacket on the back seat of a car, even though the officer removed the occupants from the car and separated them from one another before he searched the car. 453 U.S. 454, 456 (1981) (Stewart, J., plurality opinion). We followed the Belton Court's lead in United States v. Litman, approving the admission of evidence found in a shoulderbag even though the officer took the bag away from the suspect before he searched it. 739 F.2d 137, 138 (4th Cir. 1984) (en banc). Thus it is well-settled that officers may separate the suspect from the container to be searched, thereby alleviating their safety concerns, before they conduct the search.

That does not end our inquiry, however. Unlike the officers in Belton and Litman, the agents in this case did not perform the search immediately after separating the container from the suspect and securing the scene. Instead, Orton left the room with Bennett for a few minutes before returning to search the bag. The determinative issue, then, is whether a search is incident to arrest when there is a delay after elimination of the safety concerns.

The Supreme Court addressed delayed searches in United States v. Edwards, 415 U.S. 800 (1974), and United States v. Chadwick, 433 U.S. 1 (1977). The Edwards Court held that, once the justification for a search incident to arrest is established, a delay before the search does not render the search invalid:

[O]nce the defendant is lawfully arrested and is in custody, the effects in his possession at the place of detention that were subject to search at the time and place of his arrest may lawfully be searched and seized without a warrant even though a substantial period of time has elapsed . . . .

415 U.S. at 807. In Edwards's wake, this and other courts of appeals consistently upheld container searches even if they occurred after delays and the suspects no longer could reach the containers' contents. 2 LaFave, supra, Section(s) 5.5(a), at 531 (citing, e.g., United States v. Johnson, 495 F.2d 378, 381-82 (4th Cir.), cert. denied, 419 U.S. 860 (1974); United States v. Schleis, 543 F.2d 59, 62 (8th Cir. 1976) (holding that delay was reasonable under the circumstances), vacated, 433 U.S. 905 (1977)). But three years after Edwards, in Chadwick, the Court appeared to change course. Writing for the majority, Chief Justice Burger stated the Court's holding broadly:

Once law enforcement officers have reduced luggage or other personal property not immediately associated with the person of the arrestee to their exclusive control, and there is no longer any danger that the arrestee might gain access to the property to seize a weapon or destroy evidence, a search of that property is no longer an incident of the arrest.

433 U.S. at 15. In a footnote, he sought to distinguish Edwards: "Unlike searches of the person, searches of possessions within an arrestee's immediate control cannot be justified by any reduced expectations of privacy caused by the arrest." Id. at 16 n. 10. Chadwick's language was expansive, but its facts were extreme. The container in question was a locked footlocker in the open trunk of the suspects' car. Agents arrested the defendants and took them, the car, and the footlocker to the local federal building. There, an hour and a half after the initial encounter, the agents unlocked and searched the footlocker. The court found that the footlocker was not in the suspects' immediate control when the agents initially approached them. Id. at 14. Moreover, the Justices did not agree that the search would have been valid had it occurred immediately at the scene of the arrests. Id. at 16-17 (Brennan, J., concurring). Four years after Chadwick, in Belton, the Court confirmed that Chadwick's import was not as broad as its language. The Belton officer, by moving the arrestees away from the jacket, defused the danger before he performed the search, but the Court upheld his actions. See 453 U.S. at 456. We recognized Belton's narrowing effect in Litman, holding in direct conflict with Chadwick's language that an officer's "exclusive control" of a bag did not invalidate his search. 739 F.2d at 138.

Belton established that incidence to arrest continues to justify a search even after the likelihood of danger or destruction of evidence has been eliminated, but Chadwick indicates that the justification does not continue indefinitely. The determinative question appears to be whether the time and distance between elimination of the danger and performance of the search were reasonable. Chadwick is the Supreme Court's example of an unreasonable delay, and the delay in this case is not comparable to Chadwick: Han's bag unquestionably was in his immediate control at the beginning of the encounter; the delay lasted just a few minutes instead of an hour and a half; and the search occurred at the scene of the arrest, not at the police station. More important, the delay in this case was objectively reasonable. Although Special Agent Orton already had sufficient evidence to arrest Han and search the bag, he chose to confirm his information by speaking with Bennett. He may have caused minimal additional infringement of Han's rights by waiting a few minutes to conduct a search that would have been justified had he performed it immediately. But the degree to which his conversation with Bennett reduced the likelihood of arresting the wrong individual outweighed that marginal infringement. To deem this search unreasonable would encourage officers either (1) to proceed more hastily than necessary, risking unnecessary infringement on rights, or (2) to allow the dangerous condition to continue during their deliberate investigation. Either result would violate the basic purpose of the search-incident-to-arrest exception — to deter interference with suspects' Fourth Amendment rights while ensuring officers' safety. Thus we hold that when a container is within the immediate control of a suspect at the beginning of an encounter with law enforcement officers; and when the officers search the container at the scene of the arrest; the Fourth Amendment does not prohibit a reasonable delay, such as the one in this case, between the elimination of danger and the search.

C.

We recently stated in United States v. Leshuk that

a person who voluntarily abandons property loses any reasonable expectation of privacy in the property and is consequently precluded from seeking to suppress evidence seized from the property.

65 F.3d 1105, 1111 (4th Cir. 1995). The Leshuk defendants denied owning a bag when sheriff's deputies approached. Id. at 1107. Denial of ownership, we held, constitutes abandonment. Id. at 1110-11; accord United States v. Zapata, 18 F.3d 971, 978 (1st Cir. 1994); United States v. Torres, 949 F.2d 606, 608 (2d Cir. 1991); United States v. Springer, 946 F.2d 1012, 1017 (2d Cir. 1991).

The government asserts that Han abandoned his privacy interest in his bag in two ways: by disclaiming his ownership of the bag and by consenting to the search. Han characterizes the government's argument as an "effort at avoiding the issue in this case." The officers knew the bag was his, he says, because he had said so himself and Bennett had corroborated his statement. They asked only because they hoped to obtain a disclaimer or consent and thereby obviate the need for a warrant. Thus, he contends, his denial of ownership did not constitute abandonment of his privacy interest.

1.

Han argues that his denial of ownership was ineffective because he thought that claiming ownership would incriminate him. He quotes dicta from a district court opinion that in turn quotes Professor LaFave: "[A] mere disclaimer of ownership in an effort to avoid making an incriminating statement in response to police questioning should not alone be deemed to constitute abandonment." 4 LaFave, supra, Section(s) 11.3(f), at 343 (1987), quoted in United States v. Perea, 848 F. Supp. 1101, 1103 (E.D.N.Y. 1994). But LaFave himself acknowledges that the circuits disagree on the subject, and his example of the opposing view is a decision of this court. See 4 LaFave, supra, Section(s) 11.3(a) at 288 n. 43 (citing United States v. Williams, 538 F.2d 549 (4th Cir. 1976)).

The facts in Williams were very similar to this case. FBI agents entered the Williams's hotel room with his permission. He denied ownership of a briefcase and a typewriter case that were in the room, and allowed the agents to open them. They found tools used to alter and counterfeit securities. He argued on appeal that his consent was ineffective, but this court disagreed:

The record, however, shows that defendant voluntarily admitted the agents into his motel room, disclaimed ownership of the brief case and the typewriter case and stated that he had no objection to a search of the cases. His disclaimer is analogous to abandonment and made the cases subject to seizure.

538 F.2d at 550-51. Like Han, Williams and the Leshuk defendants denied owning their bags because the bags were inculpatory. Thus we follow Williams and Leshuk in holding that a disclaimer of ownership is not rendered ineffective merely because the defendant was trying to avoid incriminating himself.

2.

Han also notes that he initially implied that he owned the bag — by responding affirmatively when the officers asked if they could move "his" bag — and that Orton confirmed Han's ownership by talking to Bennett. Even if disclaimer generally constitutes abandonment, he contends, it does not when the police know who owns the property but ask repeatedly in order to circumvent the warrant requirement. In fact, the agents did not repeatedly question Han. When they first arrived they asked whether they might move his bag, and he simply responded affirmatively. Later, and only once, they asked him directly whether the bag was his, and he said that it was not; they then asked if they could search it, and he did not object. Truly repetitive questioning might be coercive in particular circumstances, but there was no coercion in this case. Moreover, whether the officers knew that Han owned the bag is irrelevant. The constitutional property right belonged to Han, and his abandonment of that right did not depend on whether the officers knew that it existed. Cf. United States v. Canada, 527 F.2d 1374, 1378 (9th Cir. 1975) (holding that the "state of mind of the searcher regarding the possession or ownership of the item searched is irrelevant to the issue of standing."), cert. denied, 429 U.S. 867 (1976), quoted in 4 LaFave, supra, Section(s) 11.3(f) at 341 n. 275.

III.

Because the agents had probable cause before they searched the bag, the search was incident to the arrest, and Han abandoned his interest in the bag, we hold that the search was reasonable. Thus we affirm the conviction.

AFFIRMED