concurring:
I disagree with both of my colleagues on the novel issue presented by this case: whether, without applying Drew analysis, evidence of other crimes of the defendant is admissible to rehabilitate a complaining witness who has recanted her story. However, because the contested evidence in this ease is, I believe, admissible under our usual analysis of other crimes evidence, I concur in affirming the judgment.1
It is established that before evidence of other crimes of the defendant can be admitted, it must first hurdle the prohibition in Drew v. United States, 118 U.S.App.D.C. 11, 331 F.2d 85 (1964), that the evidence not be admitted for the purpose of showing disposition to commit crime. See id. at 15; Samuels v. United States, 605 A.2d 596, 597 (D.C. 1992). Evidence of other crimes may be *543admitted for other, limited purposes, such as to prove motive, intent, absence of mistake or accident, a common scheme or plan, or the identity of the person charged with the crime, if the other crimes evidence is “relevant and important to any one of these five issues.” Drew, supra, 118 U.S.App.D.C. at 16, 331 F.2d at 90. Because of its presumed prejudicial effect, however, the admission of other crimes evidence for even one of these limited purposes is subject to the safeguards set out in Roper v. United States, 564 A.2d 726, 731 (D.C.1989).2 See (William) Johnson v. United States, 683 A.2d 1087, 1092-93 (D.C.1996) (en banc).
It is essential to understand that even when admissible under Drew, evidence of other crimes may never be admitted solely for the purpose of proving that the defendant committed the other crime because that fact is irrelevant as a matter of law to the offense of which the defendant is charged. See Drew, supra, 118 U.S.App.D.C. at 15 & n. 7, 331 F.2d at 89 & n. 7; see also Campbell v. United States, 450 A.2d 428, 429 (D.C.1982) (“It is fundamental to a system of criminal justice based on the presumption of innocence, that the process of adjudication be insulated from evidence of past criminal or wrongful conduct of an accused other than the charged offense.”). That is exactly, however, the purpose for which M.A.’s testimony that she “was beaten up by the same guys with bats” was admitted in this case: to prove that Bolanos and Guzman had beaten M.A. about a week before she claimed they raped her, thereby engendering such fear in her that they would do violence to her again, that she subsequently recanted her complaint of rape to the police.
Judge Reid’s analysis is based on the incorrect assumption that because the use of the other crimes evidence against Bolanos and Guzman was not substantive, but merely an explanation of the complaining witness’ recantation, it is unnecessary to test the evidence against Drew strictures. The underlying assumption is wrong, however, because M.A.’s testimony that Bolanos and Guzman had beaten her was introduced for the truth of its content. In order to effectively rehabilitate M.A. by explaining that she recanted out of fear of being beaten again, the government had to persuade the jury that Bolanos and Guzman had in fact beaten her at an earlier time. Otherwise, M.A.’s fear would not have been credible. Therefore, even though the other crimes evidence was used in the context of rehabilitating the witness in anticipation of her being impeached by the defense with her recantation, the evidence was undisputedly used for substantive purposes. The flaw in Judge Reid’s analysis lies in equating the nonsubstantive use for impeachment purposes of the complaining witness’ recantation (a prior inconsistent statement) with the substantive use of the other crimes evidence to rehabilitate the witness.3
Reliance on Reed v. United States, 452 A.2d 1173 (D.C.1982), cert. denied, 464 U.S. 839, 104 S.Ct. 132, 78 L.Ed.2d 127 (1983), is for naught on the issue before us. All Reed permits is the government’s anticipatory rehabilitation of its own witness by bringing out that witness’ prior inconsistent statements as a matter of strategy, in order to “take the sting out” before the defense does so on cross-examination. See id. at 1179; Kitt v. United States, 379 A.2d 973, 975 n. 2 (D.C.1977) (noting that eliciting a witness’ prior convictions on direct examination is not *544impeachment, but rather is an effort to enhance the witness’ credibility). Reed emphasized that although the government’s attempt to take the sting out is not impermissible impeachment of one’s own witness, but rather an effort to “enhance” the witness’ credibility, the government may not cross the line and try to “bolster” an unimpeaehed witness with prior consistent statements. See Reed, supra, 452 A.2d at 1179 (citing Johnson v. United States, 434 A.2d 415, 420 (D.C.1981)).4 Judge Reid would allow the government not only to “take the sting out” by anticipating the impeachment material, which is what Reed permitted, but to apply the salve of an explanation, which is what Reed expressly disallowed. Reed, moreover, says nothing about whether other crimes evidence of the accused can be used for the purpose of either impeaching or rehabilitating another witness.
Judge Reid improperly relies on an impeachment line of cases which do not support carving oút yet another exception to Drew analysis for the distinct purpose of rehabilitation presented here. See Sherer v. United States, 470 A.2d 732, 738 n. 5 (D.C.1983) (“These impeachment rules [permitting impeachment of witnesses with specific instances of the witnesses’ ‘bad conduct’] should not be confused with the analytically distinct doctrine that restricts the introduction, as substantive evidence, of an accused’s prior crimes or bad acts.”) Our law recognizes a narrow exception, where other crimes evidence may be admissible without implicating Drew concerns, where such evidence is used “strictly for impeachment purposes.” See Samuels, supra, 605 A.2d at 597 (stating that Drew “ ‘restricts the introduction ... of an accused’s prior crimes or bad acts’ ” as substantive evidence and that Drew analysis is inapplicable to evidence admitted “strictly for impeachment purposes”) (quoting Sherer, supra, 470 A.2d at 738 n. 5). That exception, however, is limited to the use of prior bad acts of the witness,5 and then only for non-substantive impeachment purposes.6
The significant issue in this case, therefore, is not whether the government could “take the sting out” of the complaining witness’ recantation by bringing it out on direct examination, but whether the government could bolster its witness by antieipatorily rehabilitating her, not with her prior consistent statements — which Reed found to be impermissible with respect to both its timing and *545content — but with the presumptively prejudicial evidence of the accused’s other crimes. Whether the government can do so bypassing Drew analysis altogether, as Judge Reid decides, is an issue we have never before addressed. To the extent we have considered a somewhat similar situation, we have applied Drew analysis and decided that the evidence was inadmissible. See Campbell v. United States, 450 A.2d 428, 431-82 (D.C. 1982) (evidence of defendant’s past wrongful conduct toward complaining witness was inadmissible under Drew to explain her fear of him because prejudice outweighed probative value).7
Judge Schwelb would conclude that the evidence is admissible because consent is always an issue in a rape case, and evidence that Guzman and Bolanos had earlier beaten M.A. was probative of M.A.’s lack of consent to the sexual activities in the bathroom or, put another way, of the defendants’ intent to use force in their sexual encounter with her.8 See ante at 541-42. The government argued that the prior beating was admissible, under Drew, to the issue of lack of consent. Judge Schwelb nonetheless concludes, citing (William) Johnson, supra, 683 A.2d at 1098, that evidence of the prior beating to show the accused’s intent is not subject to Drew strictures, because it is “necessary to place the charged crime in an understandable context.” 9 Judge Schwelb describes this theory “as a sort of ‘vertical Toliver ’ doctrine,”10 referring to Toliver v. United States, 468 A.2d 958, 961 (D.C.1983), which held that *546evidence of other crimes that is “inextricably intertwined with evidence of the charged offense,” may be admitted without a cautionary instruction that the other crimes evidence may not be used for the purpose of proving criminal propensity.11
In my view, Judge Sehwelb’s “vertical To-liver ” theory, as applied to the facts of this ease, is not supported by the reasons we have identified for excluding some evidence of other crimes from Drew strictures. In (William) Johnson, the en banc court summarized that not all evidence that would support a separate prosecution of an offense other than the one charged is “other crimes” evidence in the sense that it must first be subject to Drew analysis and the applicable safeguards before it may be admitted. Specifically, the court stated that the following classes of evidence of other crimes are not subject to Drew strictures:
where such evidence (1) is direct and substantial proof of the charged crime, (2) is closely intertwined with the evidence of the charged crime, or (3) is necessary to place the charged crime in an understandable context.
683 A.2d at 1098 (emphasis added). The reason for treating these classes of evidence differently from traditional “other crimes” evidence is founded on the very concern for prejudice that engendered the Drew rule. In other words, when evidence is so closely connected to the charged offense that it constitutes direct evidence of that offense or when evidence is necessary to make direct evidence comprehensible to the jury, the risk of prejudice that the jury will use the evidence to convict the defendant for his criminal disposition or bad character, as reflected by commission of another, independent crime, is significantly reduced.12
We have no such close connection here between the prior beating and the rape a week later. Evidence of the beating was not direct evidence of the rape and should not, on that score, have been exempted from Drew analysis. Cf. Wilson, supra, 690 A.2d at 469 (holding that evidence of accused’s repeated threats to kill victim less than three days before murder admissible as direct evidence, without Drew strictures). Nor is evidence of the prior beating closely intertwined with evidence of the charged crime or necessary to “ ‘complete the story’ ” of “ ‘the charged offense.’ ” Id. at 474 (quoting Holmes v. United States, 580 A.2d 1259, 1266 (D.C. 1990)) (citations omitted); Cf. id. at 470 n. 3 (“rejecting] the government’s efforts to squeeze the threats evidence ... into the narrow category of ‘Toliver’ evidence, for which we have required much closer temporal and spatial proximity than shown here.”) (citation omitted). As the government candidly explained, the reason for introducing evidence of the prior beating of M.A. in this case was not to describe the circumstances surrounding the charged offense, rape, but rather to explain the complaining witness’ subsequent recantation of her rape allegation several days later. Thus, here there was not such an intimate connection between the other crime and the charged crime required by our eases as would lead us to conclude that the potential for prejudice would be sufficiently mitigated in the mind of the jury as to make Drew analysis unnecessary.
As we noted in (William) Johnson, the Drew rule grew out of a claim of improper joinder. (William) Johnson, supra, 683 A.2d at 1096. (“[T]he Drew court concluded in relevant part that ‘when two crimes arose out of a continuing transaction or the same set of events’ [ie., when joinder would be proper] the danger of admitting evidence of both in *547one trial is minimized.” (citing Drew, supra, 118 U.S.App.D.C. at 16, 331 F.2d at 90)). Just as examination of mutual admissibility of evidence is instructive in determining the propriety of joinder, so the propriety of joinder can help answer whether evidence of another crime should be excluded. Here, joinder of the prior beating and the rape would have been improper, even though both incidents involved the same participants, because the offenses are not of “the same or similar character,” nor are they “based on the same act or transaction or on two or more acts or transactions connected together or constituting parts of a common scheme or plan.” Super.Ct.Crim.Pro.R. 8(a).
I have previously explained why I believe that, in deciding what safeguards are required for admissibility of evidence of other crimes, we should do so, based not on formulaic notions, but guided by the reason for the Drew rule. See Wilson, supra, 690 A.2d at 470-76 (Ruiz, J. concurring); (William) Johnson, supra, 683 A.2d at 1109-10 (Ruiz, J., concurring). That reason is the acknowledged prejudice inherent in other crimes evidence that is inimical to our fundamental notions of criminal justice. See Old Chief v. United States, 519 U.S. 172, 180-85, 117 S.Ct. 644, 650-52, 136 L.Ed.2d 574 (1997). In this case, I see no sound reason to depart from Drew analysis.
. See note 8 infra.
. The four safeguards are that: 1) there must be clear and convincing evidence that the defendant committed the other crime, 2) the other crimes evidence must be directed to a genuine, material and contested issue in the case, 3) the evidence must be logically relevant to prove a material and contested issue for a reason other than its power to demonstrate propensity, and 4) the evidence’s probative value must not be substantially outweighed by the danger of unfair prejudice, see (William) Johnson, supra, 683 A.2d at 1092-93. Two prophylactic measures are employed to further safeguard against undue prejudice: the jury is to be given a cautionary instruction on the permissible and impermissible uses of the other crimes evidence, see id. at 1097 n. 10, and the trial court is to defer ruling on the admissibility of other crimes evidence until it is clear that the evidence is necessary to the case, see Wilson v. United States, 690 A.2d 468, 472 (D.C.1997) (Ruiz, J., concurring) (citing cases).
. Judge Reid inconsistently states that evidence of the prior beating "was not introduced as substantive evidence” yet "was not used to impeach.” Ante at 538-39.
. Reed expressly disallowed the use of prior consistent statements to rehabilitate a witness on either the government’s direct examination or even on redirect, after the witness had been impeached, because those statements did not come under the two "exceptional situations” where prior consistent statements are permissible, "1) where the witness has been impeached with a portion of a statement and the rest of the statement contains relevant information that could be used to meet the force of the impeachment, and 2) where there is a charge of recent fabrication.” Reed, supra, 452 A.2d at 1180.
. “The first, and probably the most effective and most frequently employed ... [mode of attack upon the credibility of a witness] is an attack by proof that the witness on a previous occasion has made statements inconsistent with his present testimony.” 1 McCormick on Evidence § 33, at 111-12 (John William Strong, ed., 4th ed. 1992) (emphasis added). With respect to impeachment with other crimes, Sherer, in setting out the applicable rules, similarly makes clear that it is limited to the other crimes of the witness:
The general credibility of a witness can be impeached by evidence that the witness has been convicted of a crime punishable by death or an imprisonment in excess of a year, or of a crime involving dishonesty or false statement regardless of the punishment ... a witness may be cross-examined on a prior bad act that has not resulted in a criminal conviction only where: (1) the examiner has a factual predicate for such question, and (2) the bad act bears directly upon the veracity of the witness in respect to the issues involved in the trial.... Moreover, where such impeachment is permitted, evidence of the prior misconduct may be elicited only by cross-examination of the witness; it may not be proved by extrinsic evidence.
470 A.2d at 738 (citations and internal quotations omitted).
.As McCormick states:
The theory of attack by prior inconsistent statements is not based on the assumption that the present testimony is false and the former statement true but rather upon the notion that talking one way on the stand and another way previously is blowing hot and cold, and raises a doubt as to the truthfulness of both statements. More particularly, the prior statement, assuming it is inadmissible as substantive evidence under the hearsay rule, may be used in this context only as an aid in judging the credibility of the testimony with which the previous statement is inconsistent.
McCormick, supra note 5, § 34, at 114.
.Even if Drew were deemed to include the use of the accused’s other crimes to rehabilitate another witness, I would have to conclude that the other crimes evidence presented in this case failed the fourth Roper prong, the probative/prejudicial balancing test, even as modified in (William) Johnson, supra, 683 A.2d at 1092-93, to require that the evidence’s probative value be substantially outweighed by the danger of unfair prejudice. The chronology of events here is that M.A. was beaten by Bolanos and Guzman; that about a week later she was raped by them, at which time she complained to police; and that some time after receiving entreaties and, perhaps, threats, from the family and- friends of the accused, she recanted her claim of rape in a sworn statement to defense counsel. In light of this chronology, it is difficult to see how a beating more than a week before the alleged rape would better explain M.A.’s fear of the accused after the rape than the rape itself. Moreover, evidence that M.A. was approached by several members of the family and friends of the accused after the rape to persuade her to "reject the rape” complaint and even accompanied her to defense counsel’s office, was more recent in time and directly explained why she gave a sworn recantation. Thus, as in Campbell, supra, 450 A.2d at 431, "other means of proof, especially the fact that the reasonableness of [M.A.’s] fear is better explained by ... [other facts], limits the probative value of ... [the other crimes evidence] for this purpose.” For this reason, even if Drew analysis were not employed, the trial court would have been well advised not to admit evidence of the prior beating, because " ‘whenever’ — and not simply in the Drew context — ‘relevant evidence poses a danger of unfair prejudice,’ the trial court ‘must weigh the apparent probative value of the evidence against the unfairly prejudicial effect that it is likely to have.' ” See Wilson, supra, 690 A.2d at 470 (quoting (William) Johnson, supra, 683 A.2d at 1098).
. I agree with Judge Schwelb that consent, or the use of force, is always an issue in a rape case such as this where both the fact of the sexual activity and the identity of the participants is not in dispute. In this regard, evidence of the prior beating, which M.A. testified had been with "bats” or a "stick”, was probative in light of evidence that M.A. had been threatened with a stick, as well as with a gun, at the time of the rape. As Judge Reid points out, the trial court ruled that if consent were an issue in the case (which, of course, it was), the prior beating would be relevant to the complaining witness' lack of consent. As Judge Schwelb explains, the evidence is also probative of the other side of the same coin, the accused’s intent to use force, thereby skirting the unresolved issue noted in Judge Reid's footnote 4. The trial court made a preliminary ruling, applying the Roper safeguards, that if consent became an issue, evidence of the prior beating would be admissible. As I believe that the trial court did not abuse its discretion in coming to this conclusion, I would affirm the conviction on this alternative ground. Although the trial court did not give a cautionary instruction, as we have required, we review its failure to do so sua sponte for plain error. See Gilliam v. United States, 707 A.2d 784, 785-86 (D.C.1998) (en banc). In this case, there was no plain error as the jury is unlikely to have convicted Bolanos and Guzman for rape because of evidence of the prior beating rather than because of the compelling evidence of the rape itself.
. Judge Schwelb also argues that the other crimes evidence could be considered "direct and substantial proof of the charged crime” or "closely intertwined with the evidence of the charged crime.” See (William) Johnson, supra, 683 A.2d at 1098.
. See Ante at 542 note 4.
. The standard cautionary instruction states, in relevant part, that the jury is to consider the evidence of other crimes only for the specific purpose (motive, intent, identity, etc.) for which it was admitted, and that the jury is prohibited from considering the evidence of other crimes as evidence "that the defendant has a bad character, or that the defendant has a criminal personality.” See Criminal Jury Instructions for the District of Columbia, No. 2.51(A) (4th ed. 1996).
. Even in such circumstances, the Johnson court advised that the trial court needs to be alert to the risk of possible prejudice and the need for prophylactic measures, such as the advisability, or indeed necessity in some cases, to give a cautionary instruction. See (William) Johnson, supra, 683 A.2d at 1097 n. 10.