HOWARD, Circuit Judge.

A jury convicted the appellant, Andres De León-Quiñones ("De León"), of robbing two banks in Puerto Rico, the Euro-Bank and the Doral Bank. See 18 U.S.C. § 2113(a). The jury also convicted De León of carrying a firearm during and in relation to the EuroBank robbery. See 18 U.S.C. § 924(c)(1)(A).

De León appeals his convictions on the firearms count and the Doral Bank robbery count. He makes three arguments, two of which challenge the sufficiency of the evidence presented at trial. De León argues that the evidence presented at trial was insufficient to establish that he carried a real firearm during the EuroBank robbery and that he was the person who robbed the Doral Bank. Part and parcel of his argument that he was not sufficiently identified as the Doral Bank robber is De León's claim that the district court violated his due process rights when it allowed two witnesses, Doral Bank employees Sasha González ("González") and Jaime Massanet ("Massanet"), to identify him during trial as the robber. De León's third argument, also related to his insufficient identification, is that the court erred when it allowed the prosecutor to ask leading questions when examining the identification witnesses. After review, we affirm both of the challenged counts of conviction.

I. Background

We provide the bulk of the facts here, adding more or elaborating when necessary in our later discussion of the issues. We state these facts in a manner consistent with record support.

A. The robberies

On a January morning in 2006, De León arrived at the EuroBank branch bank located in Canovanas, Puerto Rico and waited for it to open. After the bank manager let him in, De León pulled out a gun and informed the manager and the other bank employees that he was holding up the bank. De León then shepherded the manager and employees into the manager's office. Once there, De León asked the employees where the money was located. In response to this query, the manager sent two employees to take De León to the bank's safe. Once the safe was open, De León stuffed approximately $60,000 into bags. He then directed all of the bank employees to lay down near the register area and left the bank.

Another bank in Puerto Rico, the Doral Bank, had been robbed by two men just a few weeks earlier. The two men had entered the bank shortly after it opened and loitered in the lobby area. The bank's senior officer, Massanet, approached one of the men, later identified as De León, and asked the man if he needed assistance. De León responded that he did and asked for the bank's manager. When Massanet informed De León that the manager had yet to arrive, De León told Massanet that he was holding up the bank. Massanet ushered De León and the other man through a door into the vault area. As the other man waited by the door, Massanet, with De León behind him, approached the vault door. Crouching down, Massanet tried to open the door, which had two combination locks. De León put on latex gloves as he watched Massanet work the combinations. Opening the vault door, however, proved to be a two-man job and Massanet called out for another employee to help him. At this point, the other robber complained that Massanet was taking too long and told De León to get money from the tellers instead. After taking money from one teller, De León approached another teller, González, and took money from her drawer. De León and the other man then left the bank. Immediately after the robbery, González told authorities that the man who took money from her drawer wore a red shirt, a red cap, and latex gloves over his hands. She said that the other robber wore a black shirt. Photographs taken from the bank's surveillance video corroborated these descriptions.

Some time after the Doral Bank robbery, authorities asked Massanet and González, individually, whether they could identify one of the robbers from an array of six photographs, including one of De León. González could not identify anyone. Massanet, after narrowing his choices to two photographs, ultimately identified De León.

B. The trial

At trial, the government first presented evidence on the EuroBank robbery, calling the manager and two other employees to testify. In addition to discussing the robbery generally, the three employees testified specifically that De León carried a gun during the robbery, collectively referring to it as a "pistol," "revolver," and "firearm." One of the bank employees who accompanied De León to the bank's safe further described the gun as "nickel plated." Each of the three employees explained that they had the opportunity to view De León and the weapon at close range.

Later, the government presented evidence regarding the Doral Bank robbery. The government first called Massanet. He testified that the man in the lobby with whom he spoke wore a red cap and later put on latex gloves. Massanet also stated that he "stared" at this man when he first approached him in the bank lobby. Nevertheless, when the prosecutor asked Massanet if this man was present in the court-room, Massanet testified that he did not see him. The government then called González. She discussed the robbery, testifying that the man who took the money from her drawer was "very close up" to her and that she looked at him for approximately three seconds before he told her to look away. But, similar to Massanet, González could not identify De León in the court room. When González finished testifying, the court took a brief recess.

At some point during this recess, González approached the prosecutor. She told him that when De León left the courtroom during the recess, she recognized him as the man who had robbed her. Around this same time, Massanet approached a government law enforcement agent and told him the same thing. Shortly after passing this information along, both González and Massanet saw De León being led back into the courtroom in handcuffs.

When proceedings resumed, the government informed the court of these developments. With the court's permission, the government recalled both witnesses, starting with González. The prosecutor asked González if it was true that she recognized De León as he left the courtroom. De León objected to this question as leading but the court permitted it. González answered affirmatively. The prosecutor then asked González whether or not the person who robbed her was in the courtroom. González again said yes and identified De León. When the prosecutor asked her why she had been unable to identify De León previously, she indicated that computer monitors in front of De León had obscured her view of him. A similar exchange occurred between the prosecutor and Massanet. Massanet identified De León as the Doral Bank robber and testified that he had been unable to identify De León previously because De León's head was down, and he thought that De León was just another lawyer. When cross-examined, both witnesses acknowledged that they had seen De León return to the courtroom in handcuffs.

The jury ultimately convicted De León on all three counts of the indictment. This appeal ensued.

II. Discussion A. Sufficiency of the evidence: firearms count

De León argues that the evidence presented at trial was insufficient to allow a reasonable jury to convict him of carrying a firearm during the EuroBank robbery. Because he moved for an acquittal on these grounds, our review is de novo. See United States v. Cruz-Rodriguez, 541 F.3d 19, 26 (1st Cir. 2008). In assessing sufficiency, "we examine the evidence, both direct and circumstantial, in the light most favorable to the prosecution and decide whether that evidence, including all plausible inferences drawn therefrom, would allow a rational factfinder to conclude beyond a reasonable doubt that the defendant committed the charged count or crime." United States v. Cruz-Díaz, 550 F.3d 169, 172 n. 3 (1st Cir. 2008).

A conviction under 18 U.S.C. § 924(c) requires proof that the defendant used a real firearm when committing the predicate offense. See United States v. Taylor, 54 F.3d 967, 975 (1st Cir. 1995) (noting that "a toy or replica will not do"). "Although § 924(c) requires proof that the gun is real, the government's proof need not 'reach a level of scientific certainty.'" United States v. Roberson, 459 F.3d 39, 47 (1st Cir. 2006) (quoting Taylor, 54 F.3d at 976). Indeed, as we have said many times, "[d]escriptive lay testimony can be sufficient to prove that the defendant used a real gun." Cruz-Díaz, 550 F.3d at 173.

Here, there was sufficient evidence, both direct and circumstantial, that De León used a real firearm during the EuroBank robbery. The direct evidence included the testimony of three bank employees. These employees, each of whom observed the object carried by De León at close range, called it either a "revolver," "pistol," or a "firearm." See Taylor, 54 F.3d at 967 ("Three eyewitnesses to the Bay-Bank robbery, each of whom observed the object gripped by appellant at close range, testified that it was a gun. This evidence is enough to allow a rational jury to find that appellant carried a real gun."). One employee further testified that the "pistol or revolver" carried by De León was "nickel plated," a description which is consistent with the jury's finding that De León carried a real gun. See Cruz-Díaz, 550 F.3d at 173 (holding that the evidence was sufficient to establish that the defendant used a real firearm where, among other things, a witness described the gun as being "nickel plated"). Moreover, none of the witnesses called the gun a "toy gun," or "replica gun" or otherwise described it in a way that would indicate that the gun was not real. See id. (finding that the evidence was sufficient to establish that the defendant used a real firearm where, among other things, "none of the witnesses in this case, all of whom had ample time to view the gun, described it as a BB or toy gun").

There was also circumstantial evidence indicating that De León carried a real firearm. At trial, some of the employees stated that they were "afraid" that De León might hurt someone with the gun. And, throughout the robbery, the employees at the bank reacted as if the gun was real, following De León's various orders. See id. From the totality of the evidence, including the reactions of the witnesses, the jury was entitled to infer that De León carried a real firearm. See id.

De León's opening salvo is that the prosecutor was legally required to ask the witnesses whether they thought that De León's firearm was real. This argument is a non-starter. De León does not cite, nor are we are aware of, any precedent that requires the prosecution to specifically ask witnesses whether the firearm carried was real in order to establish a violation of § 924(c). Although asking such a question might inure to the government's benefit, particularly in cases where the government is relying solely on eyewitness testimony to prove that a real firearm was used, the government is entitled, within reason, to present its case as it sees fit.

De León's next argument, tangentially related to the first, is similarly unavailing. He begins by observing that the witnesses never specified whether his gun was real, merely testifying that he carried a "pistol," "revolver," or "firearm." He argues that because the witnesses never used the word real, or some comparable adjective when describing the object he carried, the testimony was not specific enough to enable a rational factfinder to convict him of the firearms offense.

Through this argument, De León asks us to divorce the words "pistol," "revolver," and "firearm" from their natural meanings. We decline the invitation. These words are most naturally understood to refer to real firearms, and the jury was entitled to take the words at face-value when reaching its verdict. See Taylor, 54 F.3d at 967. We note that this conclusion that the government satisfied its burden in no way handicaps a defendant's ability to cast doubt on the government's proof. During cross-examination, De León easily could have tested the witnesses' perceptions of the object he was carrying or further could have emphasized the lack of specificity in witness answers during closing argument. His failure to do either, although lost opportunities to be sure, does not somehow render the evidence the government presented insufficient.

B. Sufficiency of the evidence: identification

De León argues that the district court erred when it did not suppress the identification evidence on due process grounds. Without this identification evidence, his argument continues, the evidence presented was insufficient to convict him of the Doral Bank robbery. Because the success of De León's sufficiency claim hinges on the success of his due process argument, we examine the due process argument first.

Typically, the district court's ultimate decision to admit or suppress identification evidence is subject to a plenary, de novo standard of review, with the underlying findings of fact reviewed for clear error. See United States v. Rivera-Rivera, 555 F.3d 277, 283 (1st Cir. 2009). But because De León never sought to suppress the identification evidence below, our review is for plain error only. See United States v. Henderson, 320 F.3d 92, 102 (1st Cir. 2003); see also United States v. Sanders, 547 F.2d 1037, 1040 (8th Cir. 1976). To establish plain error, a defendant "must show an error that was plain, (i.e., obvious and clear under current law), prejudicial (i.e., affected the outcome of the district court proceedings), and seriously impaired the fairness, integrity, or public reputation of the judicial proceedings." United States v. Griffin, 524 F.3d 71, 76 (1st Cir. 2008).

Identification evidence is for the jury in all but "extraordinary cases." Henderson, 320 F.3d at 100. That said, a trial court should suppress identification evidence on due process grounds where there is a "very substantial likelihood" that there was an "irreparable misidentification." Rivera-Rivera, 555 F.3d at 282; see also Simmons v. United States, 390 U.S. 377, 395-96, 88 S.Ct. 967, 19 L.Ed.2d 1247 (1968). Determining whether suppression is necessary involves a two-step analysis. Rivera-Rivera, 555 F.3d at 283; United States v. Holliday, 457 F.3d 121, 125 (1st Cir. 2006). First, the court should determine whether the identification procedure that preceded the identification was "unnecessarily suggestive." Rivera-Rivera, 555 F.3d at 283. If it was, the court must then decide whether the identification itself is reliable "notwithstanding the suggestive procedure." United States v. Lopez-Lopez, 282 F.3d 1, 10 (1st Cir. 2002). If the court finds the identification to be reliable, it is admissible. See Henderson, 320 F.3d at 101; see also United States v. Alexander, 868 F.2d 492, 495 (1st Cir. 1989) (explaining that even if the procedure at issue "could be termed unnecessarily suggestive, a court need not suppress the identification unless it lacked a sufficient basis for reliability").

As the sequence of this analysis makes clear, reliability is the key. See Manson v. Brathwaite, 432 U.S. 98, 114, 97 S.Ct. 2243, 53 L.Ed.2d 140 (1977) (observing that "reliability is the linchpin in determining the admissibility of identification testimony"). Reliability is assessed by taking into account the "totality of the circumstances." Neil v. Biggers, 409 U.S. 188, 199, 93 S.Ct. 375, 34 L.Ed.2d 401 (1972). Among other things, the reliability assessment entails considering "(1) the opportunity of the witness to view the criminal at the time of the crime; (2) the witness' degree of attention; (3) the accuracy of the witness' prior description of the defendant; (4) the level of certainty demonstrated by the witness at the confrontation; [and] (5) the length of time between the crime and the confrontation." Henderson, 320 F.3d at 100 (citing Neil, 409 U.S. at 199-200, 93 S.Ct. 375).

The two-step inquiry into suggestiveness and reliability applies to both the in-court and out-of-court identifications, provided that the defendant claims that an unnecessarily suggestive, extra-judicial confrontation or procedure "tainted" the identification at issue. Id. at 100. That analysis applies here, as De León claims that the in-court identifications were tainted by an unnecessarily suggestive confrontation that occurred outside the presence of the jury — before identifying him in court, both identification witnesses saw De León returning to the courtroom in handcuffs. De León further argues, as he must, that the identifications are unreliable under the totality of the circumstances.

1. Suggestiveness

The encounter that De León labels as unnecessarily suggestive may be likened to a "one-man show up," a classically suggestive identification procedure. See Stovall v. Denno, 388 U.S. 293, 302, 87 S.Ct. 1967, 18 L.Ed.2d 1199 (1967). Both witnesses saw De León, and only De León, in handcuffs before they identified him in court.

The government asserts that this confrontation cannot be deemed unnecessarily suggestive because it was not orchestrated or staged by the government. This argument is not particularly persuasive, however, and we have rejected a similar one in the past. In United States v. Bouthot, we explained: "Because the due process focus in the identification context is on the fairness of the trial and not exclusively on police deterrence, it follows that federal courts should scrutinize all suggestive identification procedures, not just those orchestrated by the police, to determine if they would sufficiently taint the trial so as to deprive the defendant of due process." 878 F.2d 1506, 1516 (1st Cir. 1989); but see Lopez-Lopez, 282 F.3d at 10-11 (holding that no impermissibly suggestive confrontation occurred where the witnesses identified a handcuffed defendant after inadvertently confronting him at a police station). Accordingly, we will assume that the encounter was unnecessarily suggestive and turn to the second step of the analysis.

2. Reliability of the in-court identifications

The reliability analysis is inherently witness-specific. González's identification of De León presents a closer case than Massanet's does, and we begin with her. As noted above, five factors guide the inquiry.

Application of the first four factors supports a finding of reliability. First, during the commission of the crime, González had the opportunity to view the robber from a close vantage point. The photographs taken from the bank's surveillance video show that the robber stood next to González when he took money from her drawer. And González testified at trial that the robber was "very close up" to her, and that she looked at him for three seconds before he told her to look away.

Second, González's testimony further indicates that she paid a high degree of attention to both the defendant and her surroundings during the robbery. She stated that, after the robbery, she told the authorities that the other robber wore a black shirt and that the person who took money from her drawer wore a red shirt, a red cap, and latex gloves. These detailed descriptions were corroborated by Massanet's testimony and photographs taken from the bank's surveillance video. See Rivera-Rivera, 555 F.3d at 284 (finding the identification reliable where, among other things, the witness' "recollection of detail reflect[ed] attentiveness to his surroundings"). Third, González also provided an accurate description of the defendant, and fourth, she expressed certainty when finally identifying De León in open court.

The fifth and final factor, the length of time between the crime and the in-court identification, points in neither direction. The in-court identification was not particularly fresh, coming seven months after the robbery. But this lapse of time does not severely undermine the reliability of the in-court identification, especially since we have found similar lapses to be "de minimis compared to other cases." See Rivera-Rivera, 555 F.3d at 284-85 (discussing a six-month lapse between the crime and the in-court identification).

For his part, De León attacks the reliability of González's in-court identification in three ways. First, he suggests that the opportunity for observation factor cannot support a finding of reliability since González only looked at the robber for a few seconds. Second, he notes that González initially failed to identify him in open court and had to be recalled to the stand to identify him. Finally, he presses the fact that González was unable to pick him out of a pre-trial photographic lineup arranged by the authorities.

We begin with the last point, the appellant's strongest. If González had an adequate opportunity to observe him during the robbery and paid close attention during this time, it stands to reason that she would have been able to identify De León from the photographic line-up, something she failed to do. Naturally, this casts some doubt upon the reliability of her in-court identification. But the prevailing view is that where there has been a prior "failed" identification, it is typically grist for the jury's mill. See 2 Wayne R. LaFave et al., Criminal Procedure § 7.4 (3d ed. Supp.2008-2009). As one court observed, "a witness's prior inability to identify a defendant goes to the credibility of the in-court identification and not to its admissibility, and thus raises a proper question of fact for the jury to determine." United States v. Briggs, 700 F.2d 408, 413 (7th Cir. 1983); see also United States ex rel. Kosik v. Napoli, 814 F.2d 1151, 1160 (7th Cir. 1987) ("[A] previous failure to make a positive identification from a photo array does not necessarily, or even normally, make the later identification less certain.") (citations omitted); United States v. Douglas, 489 F.3d 1117, 1126 (11th Cir. 2007); State v. King, 156 N.H. 371, 934 A.2d 556, 562 (2007). The view is sound. A variety of reasons might exist for a witness's previous inability to identify the defendant, none of which would cast serious doubt on the reliability of a later identification. For example, the previous opportunity to identify the defendant could have come on the heels of the crime, at a time when the witness was too traumatized to think clearly. Or the witness may have been unable to identify the defendant not out of uncertainty, but because the witness feared retaliation upon a positive identification. To allow a failed identification to always bar a later identification would make little sense.

Of course, in some cases a witness's failure to identify the defendant on a prior occasion, in conjunction with other factors, might create enough doubt about the reliability of a later identification to preclude its admission. The question is whether this is such a case, given González's relatively brief opportunity to view the robber and her initial failure to identify De León in open court.

We have our doubts. González's initial in-court failure to identify was partially explained at trial. She testified that a computer monitor obscured her view of De León. The district court further found — and De León does not contest — that De León avoided looking at González during her initial identification attempt. And although González's original encounter with De León was brief, she observed him from a very close distance and carefully enough to allow her to recall specific details about his clothing. At the least, given these tensions, we cannot say that the district court committed plain or obvious error in allowing González to identify De León, especially in light of the rule that a court should only withhold identification evidence from the jury in "extraordinary cases." Henderson, 320 F.3d at 100.

As to Massanet's identification of De León, we may be brief. In all material respects, Massanet's in-court identification of De León is more reliable than González's. He had an equal, if not better, opportunity to observe De León during the robbery. He spent more time with De León during the robbery and acknowledged that he "stared" at De León when approaching him in the bank lobby. And, unlike González, Massanet successfully identified De León on a previous occasion, picking De León's photograph out of a six-picture photographic lineup.

As with González's identification, we conclude that the district court committed no clear or obvious error in allowing Massanet to identify De León. Because we reject De León's due process argument, his sufficiency argument also is doomed. A rational factfinder could have concluded, based on testimony from Massanet and González, that De León was the individual who robbed the Doral Bank. We therefore uphold De León's conviction on the count charging him with that robbery.

C. Leading questions

De León advances an additional, related, evidentiary argument. He asserts that the district court abused its discretion when it allowed the prosecutor to ask González and Massanet leading questions when they returned to the stand to identify him. The exchange between the prosecutor and González was as follows:

Prosecutor: When I questioned you earlier and asked you if you recognized the person who robbed you on December 30th, 2005 and you said no; is that correct?

González: My answer was that I did not recognize the person.

Prosecutor: Thank you. And is it correct that it was brought to my attention that as a particular person was leaving the courtroom that you did recognize that person?

González: Yes, that's correct.

Although the questions asked were undoubtedly leading, any error in allowing them was harmless. The questions were geared toward explaining why the witnesses were back on the stand, not toward garnering a positive identification of De León. The questions that the witnesses were asked immediately before they identified De León were not leading questions. Moreover, "[t]he evil of leading a friendly witness is that the information may supply a false memory." United States v. Hansen, 434 F.3d 92, 105 (1st Cir. 2006). Here, De León does not claim that the leading questions prompted inaccurate testimony from the witnesses, nor does the record support such a claim.

III. Conclusion

For the reasons provided above, the convictions are affirmed.

AFFIRMED.