United States v. Casellas-Toro

Pablo Casellas-Toro appeals from a final judgment of conviction, assigning as error the district court’s denials of his motions to change venue and to suppress evidence. Having jurisdiction under 28 U.S.C. § 1291, this court reverses and remands.

I.

On June 17, 2012, Casellas reported he was a victim of an armed carjacking. The next day, he spoke with an FBI agent. He claimed he was driving to the shooting range when he heard gunshots break his back window. He saw two people running from the car. A third person forced him to stop his car and ordered him to move to *383 the passenger’s seat. Casellas said he escaped, climbing out the car window after being shot in the arm. Police found the car nearby. Casellas reported two guns missing from the car.

On June 25, Casellas gave the FBI written consent to search his car for evidence about the carjacking. The FBI took custody of the car. On July 9, the FBI scheduled the search for July 16. Casellas called the FBI four times, asking, “Have you done the search, can I have my car back?” After the search on July 16, the FBI obtained a warrant for a more detailed search of the car, which took place August 13.

On July 14, Casellas’s wife was murdered. He was the prime suspect. His murder trial began December 10, 2013. The Commonwealth alleged he staged the carjacking to make the murder weapon “stolen”. A jury convicted him on January 22, 2014.

Eight days later, a federal grand jury indicted him on three counts of making false statements to a federal officer, based on his account of the carjacking. A week later, the Commonwealth court sentenced Casellas to 109 years’ imprisonment for the murder. The next day, he made his first appearance in federal court.

Immediately after Casellas’s wife was murdered in July 2012, the media began extensively covering the case. Casellas moved to transfer the federal trial to another venue, arguing the pretrial publicity about his murder conviction prevented a fair and impartial jury in Puerto Rico. He submitted to the district court an analysis of the publicity as well as a sampling of newspaper articles, videos, and online blogs. The district court described the publicity:

Hours after the discovery of [Casel-las’s wife’s] body, “just about every” news media outlet in Puerto Rico descended upon Mr. Casellas’s home and remained there for the day. Several tabloid news programs immediately made the murder investigation the main focus of their programming. Television, radio, internet, and print media outlets in Puerto Rico “have continuously, intensely and uninterruptedly covered the Casellas case virtually on a daily basis.”

Many facts about the murder investigation were leaked to the media, including the substance of Mr. Casellas’s interview with police and the condition of the victim’s body at the crime scene. The media published and broadcast a number of allegedly false rumors about Mr. Casellas, including that he was a drug user, that he threatened people with firearms, that he was involved in a hit-and-run vehicle accident, and that he drunkenly bragged about assassinating the then-governor of Puerto Rico.

Although local authorities summoned Mr. Casellas to the Bayamon courthouse for the filing of charges, he was intercepted outside the courthouse, arrested, and Mirandized in public in view of media personnel who broadcast the event live. Members of the media “covered every minute of every day” of the commonwealth trial which ran from December 10, 2013, to January 22, 2014. Many reporters tweeted the trial testimony verbatim. Cameras followed the defendant, his family, and his lawyers during breaks.

Citizens celebrated outside the courthouse and an entire stadium of people attending a baseball game erupted into cheers upon news of the guilty verdict in the commonwealth case. Television coverage of the Casellas verdict received the top Nielson rating for that month. The Supreme Court of Puerto Rico permitted the media to broadcast Mr. Ca- *384 sellas’s sentencing live on television, internet, and radio.

Adding to the sensational nature of the Casellas murder case is the fact that the defendant’s father is a United States District Judge. The media scrutinized Judge Casellas for appearing at the scene of the crime on the morning of the murder, and some local attorneys called for Judge Casellas’s resignation.

(internal citations omitted).

The United States Attorney did not oppose the transfer, agreeing Casellas made “a prima facie showing about the pervasive nature of the coverage” of the murder case. The government did urge the court to begin voir dire and “see what happens.” The district court noted, “I can’t think that you could get any further on the prejudicial publicity continuum than we are.” The court added that “the rules ... provide specifically for change of venue in circumstances, if not like this, so near this that I’m having considerable difficulty in making the call.” Since the court could not “confidently presume” “this far in advance of trial” that it would be “virtually impossible” for Mr. Casellas to obtain a fair trial”, it reserved ruling.

Voir dire began April 7, 2014 — two months after the sentencing for murder. The court asked the venire, totaling about 160, if anyone had heard of Casellas. There was, according to the court, “almost an unanimous show of hands.” Those who had heard of Casellas were individually questioned in a separate room. These interviews followed a similar format: The court asked the potential juror to tell it what he or she knew about Casellas; the sources of the information, including whether he or she watched Casellas’s Commonwealth sentencing and discussed it with anyone; whether he or she had any opinions about Casellas; and whether he or she could put aside any knowledge and be fair. The court permitted counsel to recommend follow-up questions.

After interviewing 20 potential jurors, the court heard arguments on the change-of-venue motion. It asked the government:

Why strain to find a jury here which simply on paper says it can be fair but has such extensive knowledge of wrongdoing by the defendant that no one can say with certainty that they won’t be heavily influenced by that bias when they make the evidentiary connection between the two cases, and why not go somewhere else where nobody ever heard of [Casellas]?

The government responded, “The case against him for murder was pervasive here on the island. That’s not an arguable fact” and “[Y]ou’re not going to find many people who don’t know something about the case.” 'It also noted that a number of the interviewed potential jurors indicated they could put aside any opinion and be fair. The court again asked, “Why not take it somewhere else?” The government replied, “Well, that certainly would be easier.”

The court nevertheless overruled the motion to change venue, stating “I certainly agree that we don’t know yet if we can get a jury” but “there is a sufficient possibility we can get a jury.” It noted Casel-las could renew the motion if necessary. The court continued to individually interview potential jurors, following the same format.- After two days of interviewing 114 potential jurors 1 , the court qualified 35 and ended voir dire.

*385 On April 28, after peremptory challenges, the court empaneled 12 jurors and 2 alternates. The jury convicted Casellas of all three false-statement counts, but the court granted a motion of acquittal on two counts. He was sentenced to 21 months’ imprisonment on the final count, to run concurrently with his Commonwealth sentence. Casellas appeals.

II.

Casellas claims that, by not changing the venue, the district court violated the Sixth Amendment and Federal Rule of Criminal Procedure 21. He argues he could not — and did not — receive a fair trial in Puerto Rico due to prejudicial pretrial publicity. This court reviews the denial of a motion for change of venue for an abuse of discretion. United States v. Quiles-Olivo, 684 F.3d 177, 181 (1st Cir.2012). “A trial court’s findings of juror impartiality may be overturned only for manifest error.” Mu’Min v. Virginia, 500 U.S. 415, 428, 111 S.Ct. 1899, 114 L.Ed.2d 493 (1991) (internal quotation marks omitted). “[T]he deference due to district courts is at its pinnacle.” Skilling v. United States, 561 U.S. 358, 396, 130 S.Ct. 2896, 177 L.Ed.2d 619 (2010).

The Sixth Amendment guarantees criminal defendants the right to trial by an impartial jury. Quiles-Olivo, 684 F.3d at 181, citing U.S. Const. amend. VI, and Skilling, 561 U.S. at 377, 130 S.Ct. 2896. If “extraordinary local prejudice will prevent a fair trial,” the trial must be transferred to another district. Skilling, 561 U.S. at 378, 130 S.Ct. 2896 (“The theory of our [trial] system is that the conclusions to be reached in a case will be induced only by evidence and argument in open court, and not by any outside influence, whether of private talk or public print.”) (alteration in original) (quoting Patterson v. Colorado ex rel. Att’y Gen. of Colo., 205 U.S. 454, 462, 27 S.Ct. 556, 51 L.Ed. 879 (1907) (opinion for the Court by Holmes, J.)).

Rule 21 authorizes a change of venue if “the court is satisfied that so great a prejudice against the defendant exists in the transferring district that the defendant cannot obtain a fair and impartial trial there.” Fed.R.Crim.P. 21(a). 2 See also Quiles-Olivo, 684 F.3d at 182 (“Venue change on grounds of prejudice will be deemed appropriate where there is an ever-prevalent risk that the level of prejudice permeating the trial setting is so dense that a defendant cannot possibly receive an impartial trial.”).

A fair-trial claim based on venue encompasses two questions: “first, whether the district court erred by failing to move the trial to a different venue based on a presumption of prejudice and, second, • whether actual prejudice contaminated the jury which convicted him.” In re Tsarnaev, 780 F.3d 14, 20 (1st Cir.2015), discussing Skilling, 561 U.S. at 358, 130 S.Ct. 2896. Casellas claims both presumed and actual prejudice.

*386 A.

A presumption of prejudice is generally “reserved for those extreme cases where publicity is both extensive and sensational in nature.” Quiles-Olivo, 684 F.3d at 182 (internal quotation marks omitted). Prejudice is presumed when a “degree of inflammatory publicity had so saturated the community such as to make it virtually impossible to obtain an impartial jury.” See United States v. Misla-Aldarondo, 478 F.3d 52, 58 (1st Cir.2007) (internal quotation-marks omitted). 3

The “foundation precedent” for presumed-prejudice analysis is Rideau v. Louisiana, 373 U.S. 723, 83 S.Ct. 1417, 10 L.Ed.2d 663 (1963). 4 See Skilling, 561 U.S. at 379, 130 S.Ct. 2896. The defendant’s confession in that case — obtained without counsel present and filmed without his knowledge — was broadcast to the community three times shortly before trial. See Rideau, 373 U.S. at 724, 83 S.Ct. 1417. The community where the crime occurred had about 150,000 people; about 24,000, 53,000, and 20,000 saw and heard each broadcast, respectively. Id. Three jurors had seen and heard the televised confession. Id. at 725, 83 S.Ct. 1417. The Supreme Court noted that “the people of Calcasieu Parish had been exposed repeatedly and in depth to the spectacle of Rideau personally confessing in detail to the crimes with which he was later to be charged.” Id. at 726, 83 S.Ct. 1417. Further, “to the tens of thousands of people who saw and heard it, in a very real sense [the confession] was Rideau’s trial — at which he pleaded guilty to murder.” Id. The failure to change venue violated the Constitution’s guarantee of due process. Id. at 726-27, 83 S.Ct. 1417.

In its most recent case on this subject, the Supreme Court identifies four factors relevant to presuming prejudice: the size and characteristics of the community, the nature of the publicity, the time between the media attention and the trial, and whether the jury’s decision indicated bias. See Skilling, 561 U.S. at 379, 382-84, 130 S.Ct. 2896.

By the Skilling factors, prejudice should be presumed in this case. Examining the size and characteristics of the community, the district court noted “more than 3 million people live in Puerto Rico, mitigating the potential for prejudice among the jurors ultimately selected.” The district court did acknowledge that Puerto Rico is “a compact, insular community” that is “highly susceptible to the impact of local media.” United States v. *387 Moreno Morales, 815 F.2d 725, 734 (1st Cir.1987). Compare Tsarnaev, 780 F.3d at 21 (noting Boston is a “large, diverse metropolitan area” with residents that “obtain their news from a vast array of sources”); Skilling, 561 U.S. at 382, 130 S.Ct. 2896 (noting Houston is the fourth largest city in the United States, with 4.5 million eligible for jury duty at the time of trial). And during voir dire the district court agreed with defense counsel that Puerto Rico seemed to be a “small” island.

The government agreed the media coverage was “massive” and “sensational.” See Quiles-Olivo, 684 F.3d at 182. Cf. United States v. Angiulo, 897 F.2d 1169, 1181 (1st Cir.1990) (“If the media coverage is factual as opposed to inflammatory or sensational, this undermines any claim for a presumption of prejudice.”). Nor did it oppose Casellas’s change of venue motion, explaining, “The case against [Casellas] for murder was pervasive here on the island. That’s not an arguable fact.” The district court, denying the motion, questioned: “Why strain to find a jury here which simply on paper says it can be fair but has such extensive knowledge of wrongdoing by the defendant that no one can say with certainty that they won’t be heavily influenced by that bias when they make the evidentiary connection between the two cases, and why not go somewhere else where nobody ever heard of [Casellas]?”

Like the broadcasts of Rideau’s confession, the media here publicized “blatantly prejudicial information of the type readers or viewers could not reasonably be expected to shut from sight.” See Skilling, 561 U.S. at 382, 130 S.Ct. 2896. The media reported rumors about Casellas’s character — that he was a drug user, threatened people with firearms, was involved in a hit- and-run vehicle accident, and bragged about assassinating the then-governor of Puerto Rico. The public took to Facebook and Twitter to publicly discuss Casellas’s case. Most importantly, the media extensively and sensationally covered Casellas’s Commonwealth trial, conviction, and sentencing in a just-concluded case intertwined with this one. The Commonwealth claimed Casellas lied about the carjacking — the crime in this case. The Commonwealth used the “fake” carjacking to show premeditation for the murder. In this case, in a notice of intent to use evidence, the government stated that the motive for Casellas’s false statements was to set up a defense to murdering his wife. The government announced its intent to introduce the “stolen” guns and projectiles later found in Casellas’s home on the day of the murder. The district court excluded evidence that the defendant murdered his wife, but allowed the government to introduce evidence that the “stolen” guns, as well as discharged bullets and casings matching the pistol, were found at Casel-las’s house on July 14. Since virtually the entire venire knew of the murder, the government told the court during voir dire, “So we’re saying, you know, they’re related, and then asking [the jurors] later to pretend that it’s not.”

A jury may be able to disbelieve unfounded opinions of the media or other people. However, it may have difficulty disbelieving or forgetting the opinion of another jury, twelve fellow citizens, that a defendant is guilty in an intertwined, just-concluded case. See Skilling, 561 U.S. at 383, 130 S.Ct. 2896 (“A jury may have difficulty in disbelieving or forgetting a defendant’s opinion of his own guilt but have no difficulty in rejecting the opinions of others because they may not be well-founded.” (quoting United States v. Chagra, 669 F.2d 241, 251-52, n. 11 (5th Cir.1982), overruled on other grounds by Garrett v. United States, 471 U.S. 773, 794, 105 S.Ct. 2407, 85 L.Ed.2d 764 (1985))). When some jurors knew of the defendant’s past *388 crimes but no juror “betrayed any belief in the relevance of [defendant’s] past to the present case,” there is no presumption of prejudice. Murphy v. Florida, 421 U.S. 794, 800, 95 S.Ct. 2031, 44 L.Ed.2d 589 (1975). See also Angiulo, 897 F.2d at 1182 (“Mere knowledge or awareness of a defendant’s past ... is not sufficient to presume prejudice. More must be shown, such as the actual existence of a present predisposition against defendants for the crimes currently charged.”). Here, because the just-concluded murder case and this case are intertwined, the murder conviction “invited prejudgment of his culpability.” See Skilling, 561 U.S. at 383, 130 S.Ct. 2896.

Voir dire in this case occurred two months after Casellas’s televised sentencing in the murder case. See Tsarnaev, 780 F.3d at 22 (“The nearly two years that have passed since the Marathon bombings has allowed the decibel level of publicity about the crimes themselves to drop and community passions to diminish.”). The government does not dispute that .sensational publicity continued through the start of federal voir dire. Compare Irvin v.Dowd, 366 U.S. 717, 725, 81 S.Ct. 1639, 6 L.Ed.2d 751 (1961) (finding prejudice when publicity was extensive in six months before trial), with Skilling, 561 U.S. at 383, 130 S.Ct. 2896 (no prejudice when four years passed between peak of publicity and trial), Patton v. Yount, 467 U.S. 1025, 1032, 104 S.Ct. 2885, 81 L.Ed.2d 847 (1984) (same), and Murphy, 421 U.S. at 803, 95 S.Ct. 2031 (finding no prejudice when news about the defendant largely ended seven months before trial).

Finally, in Skilling it was “of prime significance” that the jury acquitted the defendant on nine counts. See Skilling, 561 U.S. at 383, 130 S.Ct. 2896. Here, the jury’s verdict supports a presumption of juror bias. The jury convicted Casellas of all three counts — and the court then acquitted him of two, finding the government did not prove each element of Counts One and Two beyond a reasonable doubt.

The Skilling factors reveal this to be an extreme case. See Quiles-Olivo, 684 F.3d at 182. The government cites no case denying a presumption of prejudice in a case like this: “Massive” and “sensational” publicity blanketing the community for two years before trial; extensive reporting on the defendant’s conviction by a jury, of an intertwined, heinous crime; televised sentencing only two months before voir dire. And the government did not oppose a transfer.

Casellas would be relatively unknown outside Puerto Rico. Cf Tsarnaev, 780 F.3d at 22 (noting national coverage of the case). Instead, he was tried in Puerto Rico, in an atmosphere that prejudiced the trial’s fundamental fairness. This court presumes that the pretrial publicity prejudiced Casellas’s ability to be judged by a fair and impartial jury.

B.

Finding a presumption of prejudice does not resolve the case because the parties dispute whether it is rebuttable. Finding no presumption of prejudice, the Supreme Court did not reach this question in Skilling. Skilling, 561 U.S. at 385 n. 18, 130 S.Ct. 2896. Compare Patton, 467 U.S. at 1035, 104 S.Ct. 2885 (noting passage of time before second trial .“clearly rebuts any presumption or partiality” that existed at the time of the initial trial), with Rideau, 373 U.S. at 727, 83 S.Ct. 1417 (finding prejudice “without pausing, to examine a particularized transcript of the voir dire”). The only circuit directly addressing this issue found the presumption rebuttable. See, e.g., Chagra, 669 F.2d 241. See also Coleman v. Kemp, 778 F.2d 1487, 1541 n. 25 (11th Cir.1985) (assuming *389 without deciding that presumption is re-buttable), cited favorably by Moreno Morales, 815 F.2d at 739 n. 18.

Assuming the presumption is rebuttable, the government argues that voir dire was sufficient to seat an impartial jury, pointing to the court’s individual questioning and excusing potential jurors whose statements of impartiality it found unbelievable. 5 However, rather than reducing concerns of bias, voir dire revealed the depth of community knowledge of, and hostility to, Casellas. See Misla-Aldarondo, 478 F.3d at 58 (“A court may judge the partiality of the community by looking to the ‘length to which the trial court must go in order to select jurors who appear to be impartial.’ ” (quoting Murphy, 421 U.S. at 802-03, 95 S.Ct. 2031)). Of the prospective jurors, 96.6 percent knew about Casel-las’s murder of his wife. Of the 93 potential jurors individually interviewed,. 48 knew of the carjacking. The court excused 60 potential jurors (65 percent) for cause, which is much higher than almost all the cases that reject presumed prejudice. See Murphy, 421 U.S. at 803, 95 S.Ct. 2031 (20 of 78 potential jurors— 26% — excused for cause); Misla-Aldarondo, 478 F.3d at 59 (13 of 84 potential jurors — 15%—excused for cause). Cf. Skilling, 561 U.S. at 382 n. 15, 130 S.Ct. 2896 (12.3% of Houstonians believed Skill-ing guilty of crimes); Moreno Morales, 815 F.2d at 735 (finding that about 25% of potential jurors admitting to disqualifying prejudice is below threshold to presume bias of rest of venire). Although the Supreme Court in Patton rejected a presumption of prejudice when 77 percent of the venire had formed opinion on guilt, the Court emphasized that the trial “did not occur until four years later, at a time when prejudicial publicity was greatly diminished and community sentiment had softened.” Patton, 467 U.S. at 1029, 1032, 1034-35, 104 S.Ct. 2885 (noting time “soothes and erases” and reduces the fix-edness of jurors’ opinions).

Casellas’s case is like Irvin, where after extensive publicity in the months before the trial, 62 percent of the venire was dismissed for cause. See Irvin, 366 U.S. at 727, 81 S.Ct. 1639. (“[T]he ‘pattern of deep and bitter prejudice’ shown to be present throughout the community” was “clearly reflected in the sum total of the voir dire examination of a majority of the jurors finally placed in the jury box.”). The Supreme Court did not doubt that “each juror was sincere when he said that he would be fair and impartial to petitioner.” Id. at 728, 81 S.Ct. 1639. However, where “so many, so many times, admitted prejudice, such a statement of impartiality can be given little weight.” Id.

Of the 14 empaneled jurors, Casellas challenged 11 for cause. Two of the three not challenged were never individually interviewed. Compare Skilling, 561 U.S. at 376, 130 S.Ct. 2896 (prejudice rejected when 1 juror challenged for cause), Patton, 467 U.S. at 1036, 104 S.Ct. 2885 (1 juror and 2 alternates challenged for cause), and Misla-Aldarondo, 478 F.3d at 58 (1 juror challenged for cause). All of the challenged jurors knew about the murder conviction, and at least two knew of *390 the carjacking. Compare United States v. Drougas, 748 F.2d 8, 30 (1st Cir.1984) (“[0]nly one juror who recalled hearing anything about the case or its participants was seated and no defendant specifically challenge[d] his impanelment.”).

The government emphasizes the empaneled jurors’ avowals of impartiality. True, “juror impartiality ... does not require ignorance.” Tsarnaev, 780 F.3d at 28, quoting Skilling, 561 U.S. at 381, 130 S.Ct. 2896. But, “[w]here a high percentage of the venire admits to a disqualifying prejudice, a court may properly question the remaining jurors’ avowals of impartiality, and choose to presume prejudice.” See Angiulo, 897 F.2d at 1181-82. The murder conviction — combined with knowledge of the carjacking — is “blatantly prejudicial information of the type readers or viewers could not reasonably be expected to shut from sight.” See Skilling, 561 U.S. at 383, 130 S.Ct. 2896. Due to the disqualifying opinions of two-thirds of the venire and the specific knowledge of the murder conviction by nearly all jurors and the carjacking by at least two jurors, the government has not met its burden to rebut the presumption of prejudice.

The government has not met its burden to show Casellas was tried by an impartial jury. The voir dire here confirms “an ever-prevalent risk that the level of prejudice permeating the trial setting [was] so dense that a defendant [could not] possibly receive an impartial trial.” Quiles-Olivo, 684 F.3d at 182 (emphasis added). The district court abused its discretion by denying Casellas’s motion to change venue. 6

III.

Casellas argues that the district court erred in denying his motion to suppress evidence from the two searches of his car. He claims that his phone calls to the agents before the first search implicitly revoked his written consent. Because this issue is likely to recur at re-trial, if there is one, 7 this court addresses it. See Acosta-Ramirez v. Banco Popular de Puerto Rico, 712 F.3d 14, 15-16 (1st Cir.2013).

This court reviews de novo any legal conclusions in the denial of a motion to suppress. United States v. Fermin, 771 F.3d 71, 77 (1st Cir.2014). This court reviews findings of fact for clear error, in light most favorable to the ruling. Id. at 76. The district court’s determination on consent is factual, and this court reviews the decision for clear error. See United States v. $304, 980.00 in U.S. Currency, 732 F.3d 812, 820 (7th Cir.2013) (“Like the question whether consent was given at all, the question whether the suspect subse *391 quently withdrew or limited the scope of his consent is a question of fact that we review for clear error.”). Cf. United States v. Forbes, 181 F.3d 1, 5-6 (1st Cir.1999) (reviewing voluntariness of the failure to withdraw consent for clear error).

Consent is an “established exception[ ]” to the Fourth Amendment warrant requirement. Schneckloth v. Bustamonte, 412 U.S. 218, 219, 93 S.Ct. 2041, 36 L.Ed.2d 854 (1973). “Warrantless searches may not exceed the scope of the consent given. The scope of consent is measured by a test of objective reasonableness: “what would the typical reasonable person have understood by the exchange between the officer and subject?’ ” United States v. Marshall, 348 F.3d 281, 286 (1st Cir.2003) (quoting Florida v. Jimeno, 500 U.S. 248, 251, 111 S.Ct. 1801, 114 L.Ed.2d 297 (1991)). Cf. United States v. Brown, 345 F.3d 574, 580 (8th Cir.2003) (“An expression of impatience does not establish an intent to revoke consent”).

Casellas gave written consent to search his car on June 25, without any time limit or other restriction. The FBI first searched the car on July 16. In the intervening three weeks, Casellas called the FBI four times. His first call, Casellas asked if the FBI could return the car because insurance adjusters needed to inspect it. The next three calls, Casellas asked, “Have you done the search, can I have my car back?”

After the first search, the FBI believed that any bullets fired at Casellas may be lodged behind the dashboard or in hard-to-reach places. On August 6, it obtained a search warrant for the car- — still in police custody — and executed a second search. At trial, Casellas moved to suppress evidence from both searches. The district court denied his motion, finding that the FBI conducted the search within a reasonable time, that Casellas’s calls “reaffirmed” his consent, and that there was probable cause for the warrant-authorized search.

First, there is no precise time-frame to complete a warrantless search. Cf Fed.R.Crim.P. 41(e)(2)(A)® (stating search warrant must command the officer “execute the warrant within a specified time no longer than 14 days”). ' The car remained in custody, unsearched, for 21 days. The government claimed it “could not search the vehicle any sooner because other matters had precedence.” The district court found that a reasonable person “would have known such an endeavor would not be conducted momentarily, but would take some time, especially when the alleged assailants of the car and the defendant were at large.” While 21 days approaches the outer limit of a reasonable time to complete a consent search, the district court did not clearly err in finding the officers “searched the car within a reasonable time for a carjacking.”

Next, a typical person would understand Casellas’s calls as inquiries about when the search would be complete. Although Casellas asked for his car back, he never told the agents not to search it. He never said his previous consent was no longer valid. There is no evidence that Casellas’s consent was involuntary or that he simply acquiesced to legal authority. See Bumper v. North Carolina, 391 U.S. 543, 548-49, 88 S.Ct. 1788, 20 L.Ed.2d 797 (1968) (finding consent was not voluntary when person acquiesced in a search after an officer asserted having a search warrant).

Casellas argues that the agents could not reasonably believe his consent was still valid when they conducted the search two days after he was a suspect in his wife’s murder. Casellas, however, does not dispute that the government scheduled the search before the murder and executed it *392 as planned. The district court did not clearly err in finding that Casellas did not withdraw or revoke his consent.

Casellas argues that “absent information gleaned during the July 16, 2012, search” the affidavit and warrant for the second search lack probable cause. Since the first search was valid, Casellas has waived any other challenge to the second search warrant. See Sleeper Farms v. Agway, Inc., 506 F.3d 98, 104 (1st Cir.2007) (“[T]his court will only consider arguments made before this court; everything else is deemed forfeited.”).

IV.

The judgment is reversed. The case is remanded for proceedings consistent with this opinion, including any retrial. See Irvin, 366 U.S. at 728, 81 S.Ct. 1639 (vacating conviction due to pretrial publicity and noting defendant “is still subject to custody ... and may be tried on this or another indictment”).