The Michigan v. Long standard is clear: a protective area search of the car is constitutional as long as the police officers can point to “specific and articulable facts which ... reasonably warrant [them] in believing that the suspect is dangerous and the suspect may gain immediate control of the weapons.” 463 U.S. 1032, 1049, 103 S.Ct. 3469, 77 L.Ed.2d 1201 (1983) (quoting Terry v. Ohio, 392 U.S. 1, 21, 88 S.Ct. 1868, 20 L.Ed.2d 889 (1968)). This Court has held that the Long test is subjective, rather than objective. United States v. Lott, 870 F.2d 778, 783-84 (1st Cir.1989) (highlighting the requirements that the officers “possess” a reasonable fear that their safety has been compromised and that they “believe” and “suspect” that the detainee is armed).1 Consequently, under First Circuit law, a Long search passes Fourth Amendment scrutiny only if the officers in the field were actually concerned for their safety. Lott, 870 F.2d at 784. Because I believe the district court did not analyze the facts using this standard, I would remand this case for further factual findings.
At the suppression hearing, Officer Cel-lucci stated that he entered the car for the sole purpose of retrieving Nee’s license. Cellucci testified that while he was reaching to the center console, the passenger seat gave way, causing him to “inadvertently” discover the weapons inside the knapsack. The district court discounted this testimony, finding Cellucci’s account “simply not credible.” Instead, it found that Cellucci intended to search the car when he entered it. The district court made no further finding concerning the purpose of the search. A protective area search under Long is consistent with these findings. That is, even if the officers lied about entering the car to retrieve Nee’s license and were in fact conducting an intentional search for weapons, Long permits such an intentional search if the officers believed that the suspects might have access to weapons located therein. The question for the court, then, was whether the officers did in fact possess a reason*88able fear for their safety.2
The majority, however, infers from the district court’s opinion that the search was for contraband, not weapons, thereby rendering a Long analysis irrelevant. See Lott, 870 F.2d at 785 (concluding that search was improper because it “was not a search for weapons only”). Although the opinion contains no such explicit finding, the majority concludes that this inference is the only way to explain the district court’s focus on probable cause, rather than on Long. In other words, because the court conducted only a probable cause analysis, and since a search for contraband can only be justified by probable cause, the majority believes the district court must have concluded, sub silentio, that the search was for contraband. The majority thus assumes that the district court had the Long argument in mind while conducting its probable cause analysis, and uses an inferred factual finding to reconcile any possible contradiction. While I normally might be inclined to give the district court such a benefit of the doubt, I am unwilling to do so in this case for the following reasons.
First, the inferred factual finding cannot be squared with the record, which suggests that the officers consistently suspected that the car contained weapons. It is uncontested that upon entering the scene, the officers frisked all of the defendants. Cf. Lott, 870 F.2d at 785 (noting that the officers’ failure to frisk suspects after they exited the vehicle demonstrated that they did not fear for their safety). In fact, Officer Cellucci’s first action upon entering the vehicle was to frisk the third occupant of the car and to ask him to exit the vehicle. Significant as well is Officer Ball’s reason — found in the contemporaneous police report as well as the hearing transcript — for not letting Nee back in the car. Ball testified various times that he did not want Nee to enter the car because he “did not know what [Nee] had in there.” Officer Gillis shared this concern, stating, “We didn’t know if there was [sic] any weapons or anything that could be used to hurt us in that vehicle.” Furthermore, after asking Cellucci to retrieve Nee’s license, Ball asked Nee to step to the rear of the vehicle. Ball later testified that “for [reasons of] officer safety, [he] wanted [Nee] at the back of the car so that [he] could watch him and watch the two officers in the car.” All of these measures attest to the officers’ subjective belief that weapons, not contraband, were located in the car. Nevertheless, the majority’s reading of the district court opinion suggests that upon entering the vehicle, this fear simply vanished and the officers were suddenly motivated to “find evidence of a crime,” presumably that the car was stolen.3 Even if the court did conclude, sub silen-tio, that the officers were searching for such evidence, I believe this conclusion requires explicit justification by the district court given the record in this case.
Second, other aspects of the district court’s opinion tend to support a conclusion that the officers had a reasonable fear for their safety. For example, in its conclusions of law, the court states that “where [the officers’] suspicions went be*89yond traffic violations, they had a right to be concerned about weapons.” Earlier in that same paragraph, the court acknowledges that the officers had the right to conduct an investigatory stop for a “possible stolen car.” Similarly, the decision indicates that “[i]f the officers were concerned about weapons, they could have ordered the occupants out of the car.” It is plain from the facts that two of the occupants were already outside of the car when the stop occurred, and the third was frisked and ordered out of the car before the search began. Once again, if the court did in fact conclude that the officers were motivated to look for evidence of a crime, rather than by a concern for their own safety, the district court’s other, contradictory conclusions require further explanation.
Finally, I am concerned by the fact that the Michigan v. Long argument, although raised in the district court, is not so much as cited in the district court’s orders. As the hearing transcript reveals, it was also not addressed orally.4 Even if a finding that the officers intended to search for contraband might invalidate an otherwise legal protective area search under Long,5 it does not absolve the court from addressing the ultimate issue in the analysis, namely, whether the officers actually feared for their safety. See, e.g., Lott, 870 F.2d at 784 (taking into account “all facts gathered up to the time of the search” before deciding that the search was “fatally undercut” due to the officers’ improper motivations). Only when the court considers the legal theories raised by the parties below and has addressed the implications of its own factual findings upon them can this Court properly review the decisions appealed therefrom.
To conclude, I believe that the district court failed to complete its analysis below by neglecting to determine whether the officers’ intentional search was nevertheless a permissible area search under Michigan v. Long. For this reason, I would remand this case for further findings in that vein. I respectfully dissent.
. But see United States v. Menard, 95 F.3d 9, 11 (8th Cir.1996); United States v. Baker, 47 F.3d 691, 694 (5th Cir.1995).
. Given that the officers conceded from the outset that they had no probable cause to search the car, moreover, the Long analysis appears to be the only analysis pertinent to this case.
. Interestingly, the majority, like the district court, suggests that the officers’ suspicion was pretextual, noting that if they really believed the car was stolen, they "could have called in the license plate of the car to determine if it had been reported.” It is paradoxical that a disingenuous suspicion could constitute the main focus of the subsequent search.
. The only reference to Long occurred in the colloquy regarding standing, during which the court stated, "[The officers are] not even suggesting that they had probable cause to search the car, and they're not suggesting that they had probable cause to, quote, frisk the car, whatever that means, to look for weapons.” An "automobile frisk,” of course, is a common way of referring to protective area searches under Long. See Lott, 870 F.2d at 782. The court’s cursory dismissal of this theory, along with its apparent association of this type of search with an incorrect standard reflects, in my opinion, some confusion on the part of the court between a probable cause analysis and the Long analysis urged by the government, and is a further reason I disagree with the inference constructed by the majority.
. This is the paradigm suggested by Lott, though it was not squarely addressed since in that case, this Court found that the officers did not exhibit a fear for their safety in addition to having an improper motivation for the search. Lott, 870 F.2d at 785. At least one case on point has concluded otherwise. See, e.g., United States v. Gonzalez, 954 F.Supp. 48, 50 (D.Conn.1997) ("[The officer’s] statement in the incident report that he believed the car contained narcotics is not inconsistent with his testimony that he feared he could be shot.”).