concurring in part and dissenting in part:

I join parts I and II of the majority opinion. However, I think that it is premature to reverse outright at this point on this record. Both parties here were well aware that the Batson issue might be key on appellate review, and the possibility of a need for further inquiry would not come as a surprise. I would remand to give the trial judge and the parties an opportunity to determine whether a fair and reliable Batson inquiry can now be undertaken or whether the trial judge must instead order a new trial. The application of Batson depends in great measure on the sound judgment of the trial court. See, e.g., Hernandez v. New York, 500 U.S. 352, 363-70, 111 S.Ct. 1859, 114 L.Ed.2d 395 (1991). I would permit that judgment to be exercised here in light of full knowledge as to the current situation. See Harris v. Kuhlmann, 346 F.3d 330, 347-49 (2d Cir. 2003) (in collateral attack more than fifteen years after trial where state trial court rejected prima facie showing of discrimination, district court abused discretion in fading to hold reconstruction hearing to determine “if possible” whether prosecutor had legitimate reasons for strikes; district court has discretion to order new trial only where it is “demonstrably” true that passage of time has impaired trial court’s ability to make a reasoned determination).

The exercise of peremptory challenges to discriminate against prospective jurors on the basis of race or gender is unconstitutional. Batson v. Kentucky, 476 U.S. 79, 88, 106 S.Ct. 1712, 90 L.Ed.2d 69 (1986); J.E.B. v. Alabama, 511 U.S. 127, 145-46, 114 S.Ct. 1419, 128 L.Ed.2d 89 (1994). Each type of discrimination offends the Equal Protection Clause of the Fourteenth Amendment and “the equal protection component of the Fifth Amendment’s Due Process Clause.” Edmonson v. Leesville Concrete Co., 500 U.S. 614, 616, 111 S.Ct. 2077, 114 L.Ed.2d 660 (1991) (citing Bolling v. Sharpe, 347 U.S. 497, 74 S.Ct. 693, 98 L.Ed. 884 (1954)). This case concerns a claim of discrimination in jury selection based on race and gender, as the prosecutor at appellant Leon Robinson’s trial used a majority of his peremptory strikes to exclude black women from the jury.

Appellant asserts that he made a prima facie showing in the trial court of unconstitutional, purposeful discrimination by the prosecutor against the black women in the jury venire. His primary contention on appeal is that the trial court committed reversible error in overruling his objections without determining whether the prosecutor had acceptable justifications for removing these prospective jurors.1

In ruling as it did, the trial court reasoned that race-and-gender combinations are not “suspect categories” for equal protection purposes, and that Batson and J.E.B. therefore do not prohibit discrimination in jury selection against black women or other groups defined by the intersection of racial and gender identity. In our view, the trial court erred in its application of equal protection principles. We hold that the purposeful exclusion of prospective jurors because they are black and female is discrimination on account of both race and gender in direct violation of Bat-son and J.E.B. We further hold that appellant made a strong prima facie showing of such purposeful discrimination. The trial court therefore erred by not demanding a *1277satisfactory explanation from the prosecutor for his strikes.

We decline to remand for a belated inquiry into the prosecutor’s motives. In this case, an inquiry years after memories of jury selection have faded would be inherently unreliable and unfair to appellant, the party with the burden of proof. Accordingly, we reverse appellant’s convictions. He is entitled to a new trial before a petit jury selected without the taint of unconstitutional discrimination.2

I.

Appellant was a seventeen-year-old black male tried in 2002 as an adult on charges of assault with intent to kill while armed with a shotgun and related offenses. It was undisputed that appellant shot S.M., also a young black male, moments after S.M. struck appellant’s mother in the face with a metal pole and knocked her to the ground. The prosecution charged that appellant shot S.M. in an act of revenge as S.M. was fleeing the scene, ie., without legal justification. Appellant claimed that he shot S.M. to protect his sister, whom S.M. was attacking with the same metal pole he had used against appellant’s mother. Both appellant and his mother testified for the defense.

Before the commencement of voir dire, counsel and the court reviewed jury selection procedures and discussed the questions that would be posed to potential jurors to uncover possible biases or other grounds for exclusion. The questions approved by the court covered the usual areas, including prospective jurors’ knowledge of the case, parties, and witnesses; their past jury and grand jury service; their personal involvement with law enforcement and attitudes toward police testimony; their experience with violent crime; their personal beliefs that might interfere with serving on the jury; and other relevant matters. The prosecutor stated that his “one area of concern” was with jurors who would sympathize unduly with a young male defendant.3 To address this concern, the court agreed to ask jurors whether their feelings about “crimes committed by or against young men in this *1278age group” would make it difficult for them to be fair.

Thereafter, the potential jurors assembled in the courtroom and were handed a written copy of the agreed-upon questions. The court asked each question out loud, instructing the jurors to circle the numbers of any questions to which they would respond in the affirmative. A total of fifty prospective jurors then were brought up to the bench one-by-one for follow-up inquiries directed to the questions they had marked. Of these fifty jurors, twelve had not circled any questions, and their interviews were perfunctory.

Based on the individual questioning at the bench, the court struck nine members of the venire for cause. The prosecutor moved to strike two additional jurors for cause — Juror # 390, a white male, and Juror # 162, a white female — but the court declined to do so.4 The court placed four jurors (all male) who claimed hardship at the end of the list, where they would not be reached even if each party exercised all his peremptory strikes. This decision effectively left a panel of thirty-seven potential jurors — nineteen males and eighteen females — from which appellant’s jury would be selected.

■ The first fourteen jurors on the list then were seated in the jury box. Twelve of these jurors were male, including the two in the seats designated for the alternate jurors. In selecting the twelve regular jurors, each side was entitled to exercise up to ten peremptory strikes, with the first strike in each round belonging to the prosecutor. Following each round, jurors struck from the jury box were replaced by the next jurors in line from the panel. The parties were free to strike jurors from outside the jury box, ie., “from the panel,” on the understanding that if both sides did so in the same round, the twelve jurors who were seated in the box at that time would become the jury. After the jury was selected, the parties had one additional strike apiece for the alternate jurors.

The prosecutor used his first strike against Juror # 390, the white male whom the court had declined to strike for cause. In the second round, the prosecutor struck Juror # 376, a black female, from the panel. This juror had circled no questions on her questionnaire and neither counsel had asked her anything at the bench.5 In the third round, the prosecutor struck Juror # 876, who was also a black female. The prosecutor had not asked this juror any questions either.6 In the fourth round, the *1279prosecutor struck Juror # 898, again a black female. This was another juror who had circled no questions on her questionnaire. In her individual interview, the prosecutor had elicited only that she worked for the Army National Guard.7

At this juncture, appellant’s counsel interrupted the exercise of peremptory challenges to seek an acceptable explanation for the prosecutor’s strikes:

I would just ask the government to proffer a reason [sic8], gender neutral reason for the last three strikes. This is a panel that’s probably two-thirds white, with maybe five or six black females and I believe two of them gave no answers to any questions posed by the court at all. The court acknowledged that the prosecutor’s last three strikes were of black females, and no one disputed that the panel had only five or six black females and was two-thirds white. Nonetheless, the court thought it “too soon” to discern a pattern that would establish “a prima facie case under Batson” and so denied counsel’s request that the prosecutor be required to proffer a “neutral” reason for his strikes.

The exercise of peremptory challenges resumed. In rounds five and six, the prosecutor struck two white females, one from the jury box and one from the'panel.9 In the seventh round, the prosecutor struck Juror # 883, another black female, from the panel. She had circled no questions on her questionnaire and, at the bench, had revealed only that she worked as a cook.10 *1280In the next round, round eight, the prosecutor struck a fifth black female juror (Juror #371). She, too, had circled no questions on her answer sheet. At the bench, she had disclosed only that she worked as a physical therapist.11 Except to ascertain Juror # 833’s employment, the prosecutor had' not questioned either of these two jurors.

At this point, citing the fact that the prosecutor’s seventh and eighth strikes were of black females, appellant’s counsel renewed his request for an acceptable explanation:

Again, neither of them [the two black female jurors] gave any answers. I think of all the black females who have been stricken only one gave any answer to any of the court’s questions and that was only just that they had been on a jury before.12 So I would ask the government to proffer race neutral reasons for their strikes.

The court rejected this request, giving this rationale:

It is my belief legally that there’s not a suspect category of black females or black males or white males or white females. This prohibits the exercise of peremptory challenges based on race, based on gender and maybe certain other suspect factors, but when you consolidate the factors, it completes [sic] the suspect factors in a way to make it impossible to determine if it’s a race neutral reason. •

For example, while [the prosecutor] may have exercised a couple of challenges against black females, he has not exercised peremptory challenges against black males and, therefore, it cannot be said that his strikes are based on race and while he has exercised a number of strikes against black females, he has not exercised a number of strikes against white females, so it cannot be strikes based on gender.

He has stricken about an equal number of blacks and whites and although a disproportionate number of females and males [sic ], the first strike he exercised was against a white male and, therefore, I do not think there’s a basis under Batson to demand a race neutral reason at this stage.

Accordingly, the court did not require the prosecutor to explain why he struck five black female jurors in his first eight strikes.

After the court’s ruling, the parties finished exercising their peremptory challenges. The prosecutor ' used ' his next strike, in the ninth round, to remove the sixth, and presumably the last remaining, black female juror in the venire.13 Like *1281her predecessors, this juror had not circled anything on her questionnaire, and at the bench the prosecutor had asked only about her job as a publication manager.14

With his tenth strike, the prosecutor struck Juror # 162 from the panel. This was the white female juror whom the prosecutor previously had sought to remove for cause. The prosecutor then used his single alternate strike against Juror # 708, a black male who was seated in the panel. He was another juror who had not answered any of the written voir dire inquiries and whom the prosecutor had not questioned dining his interview at the bench.15 By this time it was late in the afternoon, and the court recessed for the day.

The next morning, appellant’s counsel advised the court that he had reviewed his notes and determined that fifteen of the fifty potential jurors questioned at the bench during voir dire were African-American. Neither the court nor the prosecutor objected to that representation, which was consistent with counsel’s earlier estimate that the panel was “two-thirds white.” The prosecutor then asked to make “a very brief record on the Batson issue.” He noted, first, that four of his strikes were from the panel,16 which implied that he was satisfied with the jurors who then were seated in the jury box. Moreover, the prosecutor stated, he made his seventh round strike of (black female) Juror #833 from his notes without even knowing her race or gender, “for reasons I will not give to the court, but I just didn’t like that person.”17 However, the prosecutor solemnly assured the court, he did not strike this juror because of her race or gender. The prosecutor offered to respond to further inquiries, but the court chose instead to bring the discussion to an end. “If I wanted [to pose] further inquiries [to] you,” the court explained, “I would have sustained the prima facie challenge made by counsel.” The court proceeded to swear in the jury.

II.

A.

In Batson v. Kentucky, the Supreme Court held that “the Equal Protection Clause forbids the prosecutor to challenge potential jurors solely on account of their *1282race.” 476 U.S. at 88, 106 S.Ct. 1712. The Court emphasized that “the Constitution prohibits all forms of purposeful racial discrimination in selection of jurors.” Id. In J.E.B. v. Alabama, the Court extended Batson, holding that “the Equal Protection Clause prohibits discrimination in jury selection on the basis of gender” as well as race. 511 U.S. at 129, 146, 114 S.Ct. 1419. As the Court has reiterated over the years since Batson, “[discrimination in jury selection, whether based on race or on gender, causes harm to the litigants, the community, and the individual jurors who are wrongfully excluded from participation in the judicial process.” Id. at 140, 114 S.Ct. 1419. “[T]he very integrity of the courts is jeopardized when a prosecutor’s discrimination ‘invites cynicism respecting the jury’s neutrality,’ ... and undermines public confidence in adjudication.” Miller-El v. Dretke, — U.S.-, -, 125 S.Ct. 2817, 2324, — L.Ed.2d -, 73 USLW 4479, 4481 (2005). “The overt wrong ... casts doubt over the obligation of the parties, the jury, and indeed the court to adhere to the law throughout the trial of the cause.” Powers v. Ohio, 499 U.S. 400, 412, 111 S.Ct. 1364, 113 L.Ed.2d 411 (1991).

In Batson and subsequent decisions, the Supreme Court outlined a three-step process for evaluating claims that a prosecutor has used peremptory challenges in an unconstitutionally discriminatory manner. “First, the defendant must make out a prima facie case ‘by showing that the totality of the relevant facts gives rise to an inference of discriminatory purpose,’ ” ie., an inference that the prosecutor has struck jurors on the basis of their race or gender.18 Johnson v. California, — U.S.-,-, 125 S.Ct. 2410, 2416, — L.Ed.2d -, 73 USLW 4460, 4461 (quoting Batson, 476 U.S. at 93-94, 106 S.Ct. 1712); J.E.B., 511 U.S. at 144, 114 S.Ct. 1419. Second, if the requisite showing is made, the burden shifts to the prosecutor to “give a clear and reasonably specific explanation of his legitimate reasons” for the strikes, Batson, 476 U.S. at 98 n. 20, 106 S.Ct. 1712 (internal quotation marks and citation omitted), ie., to offer race- and gender-neutral “reason[s] that do[] not deny equal protection.” Purkett v. Elem, 514 U.S. 765, 769, 115 S.Ct. 1769, 131 L.Ed.2d 834 (1995).19 Third, if the prosecutor tenders such reasons, the trial court must decide whether the defendant has proved purposeful racial or gender discrimination. Id. The resolution of this factual question “comes down to whether the trial court finds the prosecutor’s race-neutral [and gender-neutral] explanations to be credible” or pretextual in light of all the relevant evidence. Miller-El v. Cockrell, 537 U.S. 322, 339, 123 S.Ct. 1029, 154 *1283L.Ed.2d 931 (2003).20

In the present case, however, the trial court did not follow this process to the second and third steps. The court did not require the prosecutor to justify his strikes of black female jurors or determine whether the prosecutor’s reasons were valid, because it ruled that appellant failed in the first step to make a prima facie showing of racial or gender discrimination.

B.

“Whether a defendant has satisfied the burden of making a prima facie case is a question of law, namely, whether the voir dire record of the government’s peremptory strikes, as shown by the defendant, raised ‘the necessary inference of purposeful discrimination.’ ” Little, supra note 18, 613 A.2d at 885 (quoting Batson, 476 U.S. at 96, 106 S.Ct. 1712). “[T]he burden of establishing a prima facie showing is not onerous,” id., and is satisfied by significantly less than a preponderance of the evidence. In its most recent decision on the subject, the Supreme Court clarified that it Johnson, 125 S.Ct. at 2417, 73 USLW at 4462. As the Court went on to explain, reasonable “suspicions and inferences” that discrimination “may have” occurred are enough to trigger the Batson inquiry:

did not intend the first step to be so onerous that a defendant would have to persuade the judge — on the basis of all the facts, some of which are impossible for the defendant to know with certainty — that the challenge was more likely than not the product of purposeful discrimination. Instead, a defendant satisfies the requirements of Batson’s first step by producing evidence sufficient to permit the trial judge to draw an inference that discrimination has occurred.

The Batson framework is designed to produce actual answers to suspicions and inferences that discrimination may have infected the jury selection process. See [Batson,] 476 U.S., at 97-98, and n. 20, 106 S.Ct. 1712. The inherent uncertainty present in inquiries of discriminatory purpose counsels against engaging in needless and imperfect speculation when a direct answer can be obtained by asking a simple question.

Id. at 2418, 73 USLW at 4463.

Batson “spoke of the methods by which prima facie cases could be proved in permissive terms.” Id. at 2416 n. 5, 73 USLW at 4462 n. 5. “[A] prima facie case of discrimination can be made out by offering a wide variety of evidence, so long as the sum of the proffered facts gives ‘rise to an inference of discriminatory purpose.’ ” Id. at 2412-13, 73 USLW at 4462 (quoting Batson, 476 U.S. at 94, 106 S.Ct. 1712). “In deciding whether the defendant has made the requisite showing, the trial court should consider all relevant circumstances,” including the pattern of the prosecutor’s strikes in light of the composition of the venire and the prosecutor’s questions and statements during jury selection. Batson, 476 U.S. at 96, 106 S.Ct. 1712.

In the present case, the trial court’s ruling that appellant did not make a prima facie showing of discrimination hinged largely on its threshold determina*1284tion that “black females” are not a “suspect category” for equal protection purposes. The critical question under Batson and J.E.B., however, is not whether “black females” constitute a “suspect category” (or, as it is sometimes said, a “cognizable group”). Rather, the critical question is whether the purposeful use of peremptory strikes to exclude black females (or other groups defined in terms of race plus gender) involves racial and/or gender discrimination. If it does, then it offends basic principles of equal protection and is prohibited under Batson and J.E.B.

While the Supreme Court and the lower federal courts have not yet held explicitly that the purposeful exercise of peremptory challenges to exclude race-gender groups from jury service violates equal protection, state courts have condemned the practice. See, e.g., Commonwealth v. Jordan, 439 Mass. 47, 785 N.E.2d 368, 380 (2003); People v. Garcia, 217 A.D.2d 119, 636 N.Y.S.2d 370 (1995). After the Supreme Court’s decision in J.E.B., there can be no doubt that it is unconstitutional. “Allowing parties to remove racial minorities from the jury ... because of their gender, contravenes well-established equal protection principles.” J.E.B., 511 U.S. at 145, 114 S.Ct. 1419. By definition, discrimination against black females in jury selection is both discrimination against certain female jurors solely because they are black and discrimination against certain black jurors solely because they are female. It thus is both racial discrimination of the kind condemned in Batson and gender discrimination of the kind condemned in J.E.B.

When racial bias and gender bias unite to motivate discrimination against black female jurors, it makes no difference for Batson purposes that neither type of bias alone accounts for the strikes. If it is impermissible to exclude jurors because of their race or their gender, it is impermissible to exclude jurors because of their race and their gender. Two bad partial reasons for a peremptory strike do not add up to a good reason; they simply equate to a reason that is doubly bad. To prove a Batson violation, a defendant need not show that a prosecutor’s strikes were motivated solely by racial or gender bias, to the exclusion of all other considerations. Such a requirement would render Batson a virtual nullity and divorce it from the real world of jury selection, for the motivations behind peremptory strikes are seldom so crystallized and singular. Mixed motives are the norm. However, even if the prosecutor acted from mixed motives, some of which were non-discriminatory, his actions deny equal protection and violate Batson if race or gender influenced his decision. A peremptory challenge may not be based even partially on an unlawful discriminatory reason. See Tursio v. United States, 634 A.2d 1205, 1213 n. 7 (D.C. 1993) (“[R]ace is an impermissible factor, even if a minor one, in exercising peremptory strikes. It need not be the sole reason for discrimination ... to be impermissible.”) (internal quotation marks and citations omitted); Howard v. Senkowski, 986 F.2d 24, 30 (2d Cir. 1993) (“Batson challenges may be brought by defendants who can show that racial discrimination was a substantial part of the motivation for a prosecutor’s peremptory challenges.”); Kesser v. Cambra, 392 F.3d 327, 337-38 (9th Cir. 2004) (citing cases); McCormick v. State, 803 N.E.2d 1108, 1112-13 (Ind. 2004) (citing cases).21

*1285Accordingly, we hold that the purposeful exercise of peremptory strikes against black females, or any other group defined by the combination of race and gender, is a violation of equal protection that is prohibited under Batson and its progeny. A prima facie showing of such discrimination is therefore a prima facie showing sufficient to trigger the full Batson inquiry.

We turn, then, to the question whether appellant made a prima facie showing of discrimination against black females. While we shall consider all the relevant circumstances, we begin with the statistical evidence before the trial court. Appellant’s counsel made a more than sufficient record of the composition of the venire, and neither the trial court nor the prosecutor disagreed with it in any respect. We take it as established, therefore, that the relevant venire consisted of thirty-seven jurors — nineteen males and eighteen females. This venire was approximately two-thirds white and one-third black. Six of the eighteen female jurors were black.

The prosecutor used 60% of his peremptory strikes (6 out of 10) to eliminate all six of the black female jurors, who made up only 16% of the venire (6 out of 37). If the prosecutor had distributed his strikes in proportion to the representation of jurors by race and gender in the venire, he would have struck only one or two of the black female jurors. Thus, we see a sizable statistical disparity and a discriminatory impact against black females, resulting in the exclusion of 100% of the black females from the jury. The inference that the prosecutor discriminated on the basis of race and gender is only bolstered by other statistical disparities in the prosecutor’s exercise of his challenges. Sixty percent of the prosecutor’s strikes (or 63.6%, if the alternate strike is considered) were against black jurors, who made up only about a third of the venire; and 90% of the strikes were against female jurors, who were only 49% of the venire. (This latter statistic is the principal basis for appellant’s alternative claim, which we do not reach, that he established a prima facie case of pure gender discrimination.) Unexplained statistical disparities of this magnitude have been held to establish a prima facie case of discrimination. See, e.g., Johnson, 125 S.Ct. at 2410, 2418, 73 USLW at 4460, 4463 (prosecutor used three of his twelve peremptory challenges to remove all three of the forty-three eligible jurors who were black); Miller-El v. Cockrell, 537 U.S. at 342, 123 S.Ct. 1029 (where prosecutors used ten of fourteen peremptory strikes to exclude 91% of eligible African-American venire members, and only one served on the jury, “the statistical evidence alone raises some debate as to whether the prosecution acted with a race-based reason”); Capitol Hill Hosp. v. Baucom, 697 A.2d 760, 765, 770-71 (D.C. 1997) (counsel used all his strikes against white jurors, who were eighteen percent of the venire); Tursio, 634 A.2d at *12861210 (“the prosecutor used ninety percent of his peremptory challenges to strike a group — whites—that constituted only twenty-six percent of the venire”); United States v. Alvarado, 923 F.2d 253, 256 (2d Cir. 1991) (prosecutor used half of his strikes against minority jurors, who composed an estimated 29% of the venire; “a challenge rate nearly twice the likely minority percentage of the venire strongly supports a prima facie case under Bat-son”); Fernandez v. Roe, 286 F.3d 1073, 1078 (9th Cir. 2002) (prosecutor used 21% of his strikes against Hispanics, who constituted 12% of the venire).

In dismissing the significance of the prosecutor’s disproportionate concentration of his strikes against black females, the trial court cited the following factors in addition to its belief that “there’s not a suspect category of black females:” (1) “he has not exercised peremptory challenges against black males and, therefore, it cannot be said that his strikes are based on race;” (2) “while he has exercised a number of strikes against black females, he has not exercised a number of strikes against white females, so it cannot be strikes based on gender;”22 (3) “[h]e has stricken about an equal number of blacks and whites;”23 and (4) “although [the prosecutor struck] a disproportionate number of females ..., the first strike he exercised was against a white male.” None of these factors, either individually or in the aggregate, refuted the strong prima facie showing of discrimination on account of race and gender in the prosecutor’s peremptory elimination of black females from the jury. It is well-settled that “[t]he discrimination condemned by Bat-son need not be as extensive as numerically possible.” Alvarado, 923 F.2d at 256. “A prosecutor may not avoid the Batson obligation to provide race-neutral explanations for what appears to be a statistically significant pattern of racial peremptory challenges simply by foregoing the opportunity to use all of his challenges against minorities.” Id.; cf. Batson, 476 U.S. at 95, 106 S.Ct. 1712 (“A single invidiously discriminatory governmental act is not immunized by the absence of such discrimination in the making of other comparable decisions.”) (internal quotation marks and citation omitted). Further, “the mere fact that the prosecutor ... also struck a white juror (or jurors)” or, as the trial court observed here, one white male juror, is not significant. Little, 613 A.2d at 886. “A single strike of a white juror can be a means to conceal (or attempt to conceal) a pattern of striking black jurors, or it simply may be unrelated to other strikes used discriminatorily.” Id.

We are unmoved by the -government’s observation that, on three occasions when there was a black female juror seated in the jury box, the prosecutor struck a juror from the panel instead. The government posits that if the prosecutor had intended to discriminate against black females, “he would not have left a black female in the box while exercising strikes from the panel, risking that, if the defense also struck from the panel or passed, the black female would be empaneled.” But this argument does not change the fact that the prosecutor focused the strikes he made on black female jurors so as to eliminate them from the jury, and it does not explain why he did so. It is fallacious to assume that a prosecutor’s temporary willingness to “risk” a single black female juror negates *1287all the other indications of discriminatory intent. See Little, supra note 18, 618 A.2d at 887 (“a prosecutor’s decision to pass and not use a peremptory challenge does not, standing alone, signal a mind that is free from discrimination....”). Upon inspection, moreover, the prosecutor’s strikes from the panel are less meaningful than the government supposes. On two of the three occasions to which the government points (in rounds two and seven), the prosecutor struck a black female from the panel; that evidence hardly supports the government’s claim. On the third occasion, which occurred in the fifth round of peremptory strikes (early enough in the process, one might think, for the prosecutor to perceive the “risk” to be minimal), the prosecutor struck Juror #281 from the panel. She was the white female juror who had said that she was ill and had a conflicting dental appointment. The prosecutor’s election to strike this one juror from the panel ahead of the black female juror who was seated in the box at the time proves little, if anything. It certainly does not nullify the prima facie showing of discrimination that is apparent when the prosecutor’s strikes are viewed in their entirety.

Indeed, other factors of which the trial court was apprised strengthened appellant’s prima facie showing. It is especially noteworthy that the black female jurors had furnished no information in voir dire that would have afforded any reason to strike them, and that the prosecutor had asked them virtually no questions at the bench to uncover such reasons. So far as the record indicates, the prosecutor knew nothing about these jurors other than that they were black and female.24 See, e.g., Morse v. Hanks, 172 F.3d 983, 985 (7th Cir. 1999) (holding that “when the voir dire is as perfunctory as this one was — the potential jurors were asked little more than whether they would treat both sides equally — a prima facie case is established” where the prosecutor strikes the only black juror in the venire); cf. Miller-El v. Dretke, 125 S.Ct. at 2328, 73 USLW at 4483-84 (“[T]he State’s failure to engage in any meaningful voir dire examination on a subject the State alleges it is concerned about is evidence suggesting that the explanation is a sham and a pretext for discrimination.”) (citing Ex parte Travis, 776 So.2d 874, 881 (Ala. 2000)).

We also cannot ignore certain salient facts about the case. Appellant was accused of shooting someone who had assaulted and injured his mother and, he claimed, had attacked his sister as well. The theory of appellant’s defense was that he shot the complainant to protect his sister, and appellant and his mother were to be key defense witnesses. We infer from the record that appellant’s mother and sister were black. Appellant himself was a young black man being tried as an adult for acts he committed when he was still a minor. The prosecutor openly had expressed concern about jurors who would sympathize with a young male defendant. These facts certainly do not detract from the inference that the prosecutor was motivated to strike black female jurors who, on the basis of the stereotypes that we *1288deprecate, might be considered most sympathetic to appellant, his witnesses, and his defense.

We hold that appellant made a prima facie showing that the prosecutor unconstitutionally exercised peremptory challenges against six jurors because they were black and female. The trial court therefore erred in not proceeding to the second and third steps of the Batson inquiry.

III.

We next consider the appropriate remedy for the trial court’s failure to conduct the necessary Batson inquiry. The question of remedy is committed to our discretion and has been addressed previously in two of our cases. In each of those cases, unlike this one, the trial court had proceeded to the second and third steps of the Batson inquiry, and a record had been made of the prosecutor’s reasons and the trial court’s assessment. In both cases we found that the court’s evaluation was flawed but reached different conclusions as to the best course of action to remedy the deficiency. In Capitol Hill Hospital, we remanded the case for “a proper redeter-mination, if still possible,” directing “the trial judge [to] accord appropriate weight to the statistical evidence, and [to] apply a more rigorous standard of scrutiny to counsel’s explanations for his strikes.” 697 A.2d at 760. We did this with misgivings, as explained in Judge Ruiz’s concurring opinion:

Although the case is being returned to the trial court to develop further the record and make more specific findings of fact and conclusions of law, it may be impracticable to do so years after the fact. Moreover, the value of an on-the-spot exploration of counsel’s motives, including the trial court’s opportunity to assess counsel’s candor, is likely to be significantly diminished once the appellate court has set out its difficulties with the explanations already given.

Id. at 767 n. 21.

In Tursio, on the other hand, we concluded that an effort on remand to reconstruct the prosecutor’s motives would be futile. We explained:

Theoretically, we could remand for the trial court to conduct the kind of probing inquiry required on this record and to determine, after the inquiry, whether the prosecutor rebutted the prima facie showing of racial discrimination. We do not do so, however, for two related reasons: the explanations the prosecutor proffered, ... as discussed in this opinion, leave little if any room for amplification that would produce a discernibly non-racial motivation-for the peremptory strikes; and, in any event, it is unlikely that the prosecutor could cultivate his memory at this late date to elaborate further justifications that would credibly refine the reasons already proffered. While we do not say that a remand for further inquiry in a case like this could never be the appropriate remedy, we believe that under the circumstances here such an exercise would be futile.

634 A.2d at 1213. Accordingly, in Tursio we simply reversed the judgment of conviction and remanded for a new trial.

In the present case, the trial court did not proceed to the second and third steps of the Batson inquiry. As a result, no record was made of the prosecutor’s justifications for his strikes (other than the prosecutor’s uninformative statement that he “just didn’t like” Juror # 833), or of whether the facts supported or undermined those justifications. The parties disagree as to how we now should proceed. Appellant asks us to reverse his conviction and grant him a new trial, while the government urges us to remand for the trial *1289court to hold a hearing to determine the prosecutor’s reasons for his peremptory strikes of the six black female jurors.

We deem it necessary to reverse appellant’s conviction outright. A Batson inquiry at this late date simply cannot replicate the probing inquiry to which appellant was entitled, and it is appellant who would be most prejudiced as a result. The Batson inquiry is supposed to be a rigorous evaluation of the credibility of the prosecutor’s explanations for his challenged peremptory strikes. See Miller-El v. Cockrell, 537 U.S. at 339, 123 S.Ct. 1029 (“the issue comes down to whether the trial court finds the prosecutor’s race-neutral explanations to be credible”); Tursio, 634 A2d at 1211 (“unless the trial court rigorously scrutinizes the prosecutor’s race-neutral explanations, Batson’s promise of eliminating racial discrimination in jury selection will be an empty one”). In this case, however, a delay of over three years undoubtedly has “impair[ed] the [trial court’s] ability to make an informed assessment of [the prosecutor’s] reasons.” Owens-Corning Fiberglas Corp. v. Henkel, 689 A.2d 1224, 1228 (D.C. 1997) (holding that a Batson challenge must be made before the jury is empaneled, since a prompt hearing is “essential if a Batson issue is to be addressed in an effective and meaningful manner”).

For one thing, while the Batson procedure relies heavily on “the value of an on-the-spot exploration of counsel’s motives,” Capitol Hill Hospital, 697 A.2d at 767 n. 21, it is no longer possible to obtain the prosecutor’s contemporaneous explanations for striking all six black female jurors, or to observe his demeanor in giving such explanations. Whatever explanations the prosecutor may offer years later, for the first time and with the benefits of hindsight and preparation, are a less than satisfactory substitute, if only because “a prosecutor’s explanations will be fresher, and perhaps more candid, if they are given before the prosecutor has time for research and reflection.” Tursio, 634 A.2d at 1211.

Furthermore, since the written record contains nothing that would justify the prosecutor’s concentrated peremptory challenges of black female jurors, any plausible justifications he might offer at a belated Batson hearing would be based on matters not reflected in the record — such as the mannerisms, tone of voice, or other unknown factors that caused the prosecutor to say he “just didn’t like” Juror # 833.25 A meaningful assessment of such justifications would have to rely heavily on demeanor evidence as to the prospective jurors and as to the prosecutor himself. See, e.g., Epps v. United States, 683 A.2d 749, 753 (D.C. 1996) (noting that explanations for peremptory challenges commonly “focus upon a venireperson’s body language or demeanor,” and that such subjective explanations require the closest scrutiny) (citation omitted); Miller-El v. Cockrell, 537 U.S. at 339, 123 S.Ct. 1029 (“[T]he best evidence often will be the demeanor of the attorney who exercises the challenge.”) (quoting Hernandez v. *1290New York, 500 U.S. 352, 365, 111 S.Ct. 1859, 114 L.Ed.2d 395 (1991)). With the passage of years, this important but evanescent evidence has been lost. As memories of the voir dire and jury selection fade, “it becomes difficult (if not impossible) for the court and counsel to recreate in their minds the circumstances of each strike” or even to remember individual jurors. Owens-Corning Fiberglas Corp., 689 A.2d at 1228; see also Tursio, 634 A.2d at 1211; Miller-El v. Cockrell, 537 U.S. at 343, 123 S.Ct. 1029. In short, “[t]he practicalities of the situation ... cry out for judicial insistence on timeliness.” Owens-Corning Fiberglas, 689 A.2d at 1228.

The loss of potentially critical evidence is most problematic for appellant, since it is he who would have the burden of proof in the third stage of a Batson inquiry. See Purkett, 514 U.S. at 768, 115 S.Ct. 1769 (“[T]he ultimate burden of persuasion regarding racial [or gender] motivation rests with, and never shifts from, the opponent of the strike”). For example, “[i]n order for defense counsel to meaningfully perform his' role with regard to rebutting the prosecution’s supposed neutral reasons for striking minority jurors, he needs to be given the opportunity to: (1) point out that the prosecutor’s claims about the particular juror are false; (2) point out that although the prosecutor’s claims about an excluded juror are true, similar claims can be made about non-excluded jurors who are not minorities, which should raise the suspicion of bad faith; and (3) argue that claims about the juror, although true, are so irrational as a reason for striking a juror that they might be pretexts for some undisclosed discriminatory reason.” United States v. Alcantar, 897 F.2d 436, 438 (9th Cir. 1990). That critical opportunity has been impaired. Never having been informed of the prosecutor’s reasons for his strikes, appellant has not been in a position to preserve facts about jurors and other evidence that would be relevant in contesting those reasons.

The government too may be prejudiced by the delay and the ensuing loss of evidence, see, e.g., Owens-Corning Fiberglas, 689 A.2d at 1228, though as a general matter we think it fair to say that the adverse consequences of delaying a Batson inquiry typically will fall most heavily on the party with the burden of proof. Since the government knows why it struck jurors as it did, it has the upper hand, and less reason, therefore, to complain about being disadvantaged by the delay.

We do not hold that a remand for a belated Batson fact finding inquiry never can be appropriate when the trial court has erroneously refused to require a prosecutor to respond to a prima facie showing of race or gender discrimination in jury selection. Such an inquiry might still be feasible in some circumstances, for example where the prosecutor voluntarily disclosed the reasons for his strikes in spite of the trial court’s ruling and the record supports those reasons or otherwise allows them to be tested meaningfully.26 Those circumstances are not before us now, however. In this case.there is amply justified concern that a belated inquiry into the prosecutor’s motivations would be inherently unreliable and unfair to appellant. The burden must be placed on the government to allay that concern, but while it has asked us to remand for a hearing, it has not argued that the explanations for the prosecutor’s seemingly discriminatory *1291strikes could be tested fairly and adequately at this late stage.

When all is said and done, we must recognize that “there are cases where the passage of time may impair a trial court’s ability to make a reasoned determination of the prosecutor’s state of mind when the jury was selected. Where such demonstrably exists, there must be a new trial.” Brown v. Kelly, 973 F.2d 116, 121 (2d Cir. 1992). We deem this to be such a case; the failure to conduct the prompt hearing to which appellant was entitled cannot be rectified by conducting an unavoidably inferior substitute for such a hearing now. In our view, the only fair remedy on the facts before us is to reverse appellant’s conviction and grant him a new trial.

So ordered.