A jury found appellant guilty on two counts each of enticing a minor and sodomy on a minor (D.C.Code §§ 22-3501(b), -3502 (1989)), and six counts of taking indecent liberties with a minor (id., § 22-3501(a)). All of the acts were alleged to have been committed on the same victim, seven-year-old L.B., in September and December of 1992. On appeal, appellant contests the sufficiency of the evidence supporting the sodomy convictions, claims reversible error in the combined effect of two changes which the trial judge made in the reasonable doubt instruction, and assigns other errors in the conduct of the trial. Finding merit in the first two contentions, we reverse.
*736I.
L.B. first met appellant on her way to school on the first day of school in September 1992, when he was giving away candy to children in front of Martin Luther King Elementary School. L.B. went over to appellant and received some bubble gum from him, and also received gum from him the next day. That same month, she was on the way to school one morning when appellant stopped her and asked her to go with him. Thinking he would give her candy again, she followed him to the rear of a budding near the school. After they went down a stairwell, appellant told her to pull down her pants, which she did, and he pulled down his own pants. He then rubbed his penis against her vagina and touched her vagina with his finger. He told her to “put [her] mouth on his penis” and to place her hand on his penis; she complied with both demands. When these acts were done he gave her some bubble gum.
Appellant repeated these acts another morning in the same place, and “the same thing” also happened several times later after school. Confirming L.B.’s testimony that some of her meetings with appellant made her late for school, attendance records showed that she was late for school on September 11 and September 22,1992.
The final acts took place in December 1992 when L.B. saw appellant after school and he told her to come with him. This time they went down some steps and entered a front door of the building behind which the earlier acts had occurred. There, the child testified, appellant “pulled down his pants and I pulled down mine. He told me to put my — put my mouth on his penis and touch it, and he put his penis — rubbed it against my vagina and he put his hands there, too.” He then gave her a stick of bubble gum.
When L.B. was late returning home from school that December day, her mother went looking for her. She located the child walking towards home and asked her where she had been. L.B. told her she had been at school finishing her work. When her mother asked about the gum, the child stated that a teacher had given it to her, then that “a little girl” had done so. The next day, when asked again, she told her mother that a man had given her the gum. She described him to her mother and her stepfather, including the fact that he “walked with a limp,”1 and told them about the assaults.
After hearing the description of the man who assaulted her, the stepfather took L.B. downstairs to appellant’s apartment. Appellant denied knowing the child, and on the way back to their apartment L.B. told her stepfather that appellant was not her attacker. L.B. testified that, though she had recognized appellant that day as the man who had assaulted her repeatedly, she lied to the stepfather because she was afraid, since appellant had warned her not to tell and threatened to take her “far away” if she did.
On December 8, 1992, the child’s mother saw appellant at their apartment braiding and noticed that he fit the description L.B. had given of her attacker. She took the girl downstairs to appellant’s apartment and asked her to tell her if Proctor was the man who had assaulted her. On seeing him L.B. replied yes, and when asked if she was sure, again identified appellant as her attacker.
L.B. later made a showup identification of appellant, and again identified him in court as her assailant. When a police detective took her around the neighborhood, she pointed out the locations where appellant had assaulted her and a candy store they had gone to after some of the assaults.
II.
Appellant was found guilty of two counts of sodomy, one in September and one in December of 1992. As the trial judge instructed the jury, “both of [those] charges relate to the allegation that the defendant placed his penis in the mouth of the complaining witness.” Appellant contends the evidence was insufficient to support these convictions because it failed to establish that he placed his penis in the mouth of L.B., and so did not show penetration as required by the then-existing crime. We agree.
*737The sodomy statute, at the time of appellant’s conduct, stated in relevant part as follows:
(a) Every person who shall be convicted of ... placing his or her sexual organ in the mouth ... of any other person ... shall be fined not more than $1,000 or [if having committed such act with a person under the age of 16] be imprisoned for a period not exceeding 20 years.
(b) Any penetration, however slight, is sufficient to complete the crime specified in this section. Proof of emission shall not be necessary.
D.C.Code § 22-3502 (1989) (emphasis added).2 Therefore, as the trial judge told the jury (elucidating his earlier-quoted statement), the government had to prove “that the defendant placed his penis into the mouth of the complaining witness” (emphasis added). Of course, the judge also explained that “any penetration, however slight,” is enough to meet that requirement and that “[proof] of ejaculation is not required.” But while the “slight penetration” requirement is undemanding, see Barrera v. United States, 599 A.2d 1119, 1125 n. 4 (D.C.1991) (on appellate review, “any evidence tending to show the slightest penetration ... is sufficient to require denial of a motion for judgment of acquittal”), it remains a differentiating feature between sodomy and crimes punished less severely such as taking indecent liberties.3 In ordinary use, the verb “to penetrate” means “to pass into or through.” WEBSTER’S THIRD NEW INTERNATIONAL DICTIONARY 1670 (1986). Decisional law reflects this meaning by providing that, while slight penetration is enough, “a mere touching does not make out the offense.” 70A Am. Jur. 2d Sodomy § 23 (1987). See, e.g., Ashby v. Commonwealth, 208 Va. 443, 158 S.E.2d 657, 658 (1968), cert. denied, 393 U.S. 1111, 89 S.Ct. 884, 21 L.Ed.2d 808 (1969) (where evidence showed only that victim was “told to ‘put my mouth on his [penis],’ ” proof of sodomy was insufficient); People v. Angier, 44 Cal.App.2d 417, 112 P.2d 659, 660 (1941) (kissing not sufficient evidence of penetration). In this case, then, there must have been evidence that the penis “pass[ed] into or through” the lips of the victim, even slightly, or the statutory requirement of a “placing ... in the mouth” was not met.
In describing the September acts, however, L.B. testified only that she “put [her] mouth on his penis.” She described the December incident in the same way through leading questions: she saw appellant’s penis, and agreed that she “put [her] mouth there.” The prosecutor asked no questions further clarifying the point, and did not ask the child to demonstrate at all the contact between her mouth and the penis. Cf. State v. Walker, 252 Kan. 117, 843 P.2d 203 (1992) (complainant made in-court demonstration of how penis touched her lips, saying “[j]ust right there” and pointing; since record was “devoid of any amplification [by defense] of her demonstration,” court could not exclude reasonable inference by jury that penis penetrated victim’s Ups). Though we agree with the government, both anatomically and legally, that “[t]he Ups constitute the entrance to, and are a part of, the mouth,” id., 843 P.2d at 217 (quoting instruction),4 stiU there was no evidence from which the jury could reasonably find that the contact with appeUant’s penis had been made with enough pressure to cause the child’s Ups to part, even sUghtly, amounting to an entry into (a “placing in”) the mouth. The government’s argument thus comes down to the assertion that, given *738a cylindrical object like the penis shaft, any oral contact between it and the mouth infer-ably would have caused the lips to open. That seems to us a matter of surmise and not a rational inference beyond a reasonable doubt from the fact alone that the victim placed her lips “on” the penis.5 The evidence was insufficient to sustain appellant’s convictions for sodomy.6 See Ashby v. Commonwealth, supra.
III.
In two material respects, the trial judge altered the standard jury instruction on reasonable doubt in this jurisdiction. Appellant argues that, under recent decisions of this court, these combined changes were reversible error. We agree.
The judge’s instruction on reasonable doubt was as follows:
Reasonable doubt, as the name implies, is a doubt based on reason. It is a doubt for which you can give a reason. It is such a doubt as would cause a juror after careful and candid and impartial consideration of all of the evidence to be so undecided that he or she cannot say that they have an abiding or deep-seated belief of the defendant’s guilt. Reasonable doubt is not a fanciful doubt, nor a doubt based on fantasy, nor a doubt based on a whimsical doubt, or a whim, nor a doubt based on conjecture or guesswork. It is a doubt which is based on reason. [Emphasis added.]
The standard jury instruction, by contrast, states as follows:
Reasonable doubt, as the name implies, is a doubt based on reason, a doubt for which you can give a reason. It is such a doubt as would cause a juror, after careful and candid and impartial consideration of all the evidence, to be so undecided that he or she cannot say that he or she has an abiding conviction of the defendant’s guilt. It is such a doubt as would cause a reasonable person to hesitate or pause in the graver or more important transactions in life. However, it is not a fanciful doubt, nor a whimsical doubt, nor a doubt based on conjecture. It is a doubt which is based on reason.
Cbiminal Jury Instructions FOR the District of Columbia, No. 2.09 (4th ed.1993) (emphasis added) (hereafter “Redbook”).7 As is apparent, the judge deviated from the Redbook instruction in two material ways. For the phrase “abiding conviction” he substituted “an abiding or deep-seated belief.” And he deleted the sentence defining reasonable doubt as “such a doubt as would cause a reasonable person to hesitate or pause in the graver or more important transactions in life.” Appellant objected to the substitution but not the deletion.
In Foreman v. United States, 633 A.2d 792 (D.C.1993), we “disapproved” of a trial judge’s “use ... of the word ‘belief,’ even intensified by [the adjective] ‘deep rooted,’ ” as a substitute for an “abiding conviction.” We concluded that “[t]he potential ambiguity *739in that change leaves too much room for error, especially if combined with any other shortcoming in the instruction [given],” because a “belief,” however intensified, “may convey less confidence or certainty of the existence of a fact than [the word] ‘conviction.’” Id. at 794. At issue in Foreman, however, was whether that substitution constituted “plain error” since the defendant had failed to object to the change. We acknowledged that, under Sullivan v. Louisiana, 508 U.S. 275, 113 S.Ct. 2078, 124 L.Ed.2d 182 (1993), “some defective reasonable doubt instructions would constitute plain error, since a ‘structural error’ of that kind undermines the ‘reliability]’ and hence integrity of the criminal trial.” Foreman, 633 A.2d at 796 (emphasis and alteration in Foreman) (quoting Sullivan, 508 U.S. at 281, 113 S.Ct. at 2082). But we held that “the judge’s single modification of the instruction did not work a ‘structural defeet[ ] in the constitution of the trial mechanism’ and so deprive appellant of ‘[t]he right to trial by jury...Id. at 797 (quoting Sullivan, 508 U.S. at 281, 113 S.Ct. at 2082) (citations omitted in Foreman). In the present case, as mentioned, appellant objected to the “abiding or deep-seated belief’ instruction.
In Butler v. United States, supra note 7, decided a year later, the defendant objected to the judge’s deletion, as in this case, of the sentence explaining a reasonable doubt as one that “would cause a reasonable person to hesitate or pause in the graver or more important transactions in life.” 646 A.2d at 334. Relying on Sullivan and the Supreme Court’s intervening decision in Victor v. Nebraska, 511 U.S. 1, 114 S.Ct. 1239, 127 L.Ed.2d 583 (1994), we expressed the controlling principle as “whether there has been ‘a misdescription of the burden of proof,’ producing a ‘reasonable likelihood’ that the jury understood the instructions to allow conviction based on a lesser standard than proof beyond a reasonable doubt.” Butler, 646 A.2d at 334 (citations omitted). If so, “then the instruction is constitutionally deficient and, at least where objection has been made to the instruction, we must reverse appellant’s conviction.” Id.8 In holding that the judge’s omission of the standard sentence “was improper but did not create a constitutionally deficient instruction under the particular circumstances of this case,” id. at 337 (footnote omitted), we relied on three factors. First, os an omission, the defect in the instruction was “less likely to be prejudicial than [an affirmative] misstatement of the law.” Id. at 338 (citation omitted). Second,
it was not necessary in this case, as it had been in Foreman, for the trial court to “save” the reasonable doubt definition by using the definitional clause omitted here. The instruction in this case contained no improper substitutions of phrases or other infirmities that the definitional clause clarifying reasonable doubt would have helped to cure.
Id. Third, relatedly, “[t]he jury was not left without a workable definition of the term [reasonable doubt]” because the instruction given preserved two other features of the standard instruction, including explication of reasonable doubt as such indecision that a juror “cannot say that he [or she] is firmly convinced of the Defendant’s guilt.” Id. (emphasis added; alterations in Butler ).9
The government points out that Foreman did not hold that the substitution of “deep rooted belief’ for “abiding conviction” was reversible error. That is true, since the context in which the instruction was challenged — unobjected-to error — made that decision unnecessary. But our discussion left little room for asserting that a judge can override objection to substituting “belief’ for “conviction” and avoid reversal. We expressed agreement with United States v. Merlos, 299 U.S.App.D.C. 401, 984 F.2d 1239 (1993) (Merlos I), in which the Circuit Court found constitutional error in an instruction that replaced “abiding conviction” with *740“strong belief’; and we noted the subsequent history of that ease in which the court, after holding the error originally to be harmless, reversed itself in light of Sullivan v. Louisiana in the ease of a co-defendant who had objected to the change, and ordered a new trial. Foreman, 633 A.2d at 794 & n. 4 (citing United States v. Loriano, 302 U.S.App. D.C. 158, 996 F.2d 424 (1993) (per curiam)). Moreover, even reviewing for plain error in Foreman, we held that “the judge’s single modification of the [reasonable doubt] instruction” did not warrant reversal. Id. at 797 (emphasis added). With that we conveyed our understanding that any additional weakening of the 'definition would require reversal at least in the less demanding context of objected-to error. Butler confirmed this understanding when faced with a deletion of the same “prescribed definitionf ]” that the judge omitted here. After stressing the value of the “hesitate or pause” definition as “ ‘a eommonsense benchmark’ for understanding reasonable doubt,” 646 A.2d at 336 (quoting Victor, 511 U.S. at 20, 114 S.Ct. at 1250), we found no constitutional error only because, in critical part, that definition was not needed to “save” the overall instruction from the very “misstatement of the law” to which appellant objected in the present case.
Foreman and Butler together dictate that the combined changes to the standard definition here created “a ‘reasonable likelihood’ that the jury understood the instructions to allow conviction based on a lesser standard than proof beyond a reasonable doubt.” Butler, 646 A.2d at 334. Judge Schwelb dismisses much of the language in those decisions as dictum, but the foregoing analysis demonstrates that we would drain them of their substance if we affirmed these convictions. Butler, in particular, applied the Supreme Court’s analysis in Victor which Judge Schwelb says controls “a fortiori ” here. In reversing, we adhere to the holdings of our past decisions.
The government only half suggests that appellant has not preserved the error in this case: it concedes that he objected to replacement of “conviction” with “belief’ (the judge responding that he thought “a deep-seated belief is synonymous with abiding conviction”). Since appellant identified the principal defect in the instruction, we consider the objection preserved. Moreover, since we review for objected-to error (hence mooting whether the error was “plain” or “obvious”),10 it does not matter that this case was tried before Foreman and Butler. Even if those cases had announced a new constitutional rule, appellant would receive the benefit of it. Griffith v. Kentucky, 479 U.S. 314, 107 S.Ct. 708, 93 L.Ed.2d 649 (1987). The remaining charges against appellant must therefore be retried.11
* Hi * * * 4s
The remaining remarks are those only of the author of this opinion. This is the fourth time in the past few years where this court has been confronted with changes by individual trial judges to the standard Redbook definition of reasonable doubt. Other such cases may be in the pipeline. That fact raises a critical question about the supervisory role of this court. In the course of his dissent, Judge Schwelb states that “[appellate judges ought not [to] substitute their prejudices regarding jury instructions or their notions of apt phraseology for the experience of trial judges in such matters.” Post at 744 (quoting United States v. Yunis, 288 U.S.App. D.C. 129, 139, 924 F.2d 1086, 1096 (1991)). As a general observation, that is indisputable. It is also true, as Judge Schwelb states, that “the Constitution requires no ... conformity to any prescribed recitation” of the reasonable doubt standard. *741Post at 744. But neither observation ends the matter. We do not deal here with an ordinary jury instruction, but with what is “perhaps the most important aspect of the closing instruction to the jury in a criminal trial.” Dunn v. Perrin, 570 F.2d 21, 25 (1st Cir.), cert. denied, 437 U.S. 910, 98 S.Ct. 3102, 57 L.Ed.2d 1141 (1978). Moreover, as explained, this court faces a recurrent problem of dissatisfaction by trial judges with the traditional version of that instruction which has prompted them to modify or even discard entirely the text of the instruction in their courtrooms.12 In these circumstances, I am of the view that supervisory action by the en banc court consistent with our authority under D.C.Code § 11-102 (1995) is called for. If we do not act, and especially if the dissent’s view were to prevail as to the relative insignificance of deviations from the current instruction like those made in this case, I have no doubt we will continue to be met with appeals challenging modifications by individual judges who are convinced that they can do it better.
The dissatisfaction with the Redbook instruction stems from language there which judges consider awkward, archaic, hard-to-comprehend, or misleading. Their criticism is justified. In Foreman we recognized the unclarity of the phrase “abiding conviction,” 13 as well as its placement in a double negative sentence which the jury may have difficulty grasping. 633 A.2d at 794-95, 796 n. 8. In Victor, supra, Justice Ginsburg’s concurrence pointed out that “distinguished federal judges” have criticized the traditional “hesitate to act” formula that corresponds to our “hesitate or pause” language. 511 U.S. at 24, 114 S.Ct. at 1252. And, in my view, a jury receives little additional help by being told that a reasonable doubt is a doubt “based on reason” or “for which you can give a reason.”
This court should respond to the well-founded dislike of the Redbook instruction while ending the sort of improvisation by trial judges that will force us to review many such variations for conformity to an uncertain constitutional standard. We should adopt and mandate the use of a new reasonable doubt instruction in the courts of this jurisdiction. We have the supervisoxy power to take that step. See, e.g., Winters v. United States, 317 A.2d 530, 532 (D.C.1974) (en banc) (adopting, “in the exercise of our superintendent responsibility,” a standard anti-deadlock instruction for future cases as “serv[ing] the administration of justice”); Kinard v. United States, 416 A.2d 1232, 1235 & n. 3 (D.C.1980) (holding, after poll of the en banc court, that “no longer shall ... use [of the falsus in uno instruction] be deemed appropriate in the District of Columbia court system”). And we should use it. To paraphrase the United States Court of Appeals, this court “should no longer be burdened with the necessities and niceties — and the concomitant uncertainties — of gauging various [Redbook]-type renditions [of the reasonable doubt instruction] in terms of’ their risk of misdescribing the government’s burden of proof. United States v. Thomas, 146 U.S.App.D.C. 101, 110, 449 F.2d 1177, 1186 (1971) (en banc) (adopting standard anti-deadlock charge in exercise of its supervisory authority). Subject to being further educated on the point, I believe we should adopt the “clear, straightforward, and accurate” instruction proposed by the Federal Judicial Center. Victor, 511 U.S. at 26, 114 S.Ct. at 1253 (Ginsburg, J., concurring; quoting instruction). While citing this instruction, both the majority and Justice Ginsburg in Victor explained that the Supreme Court has “no supervisory power over the state courts”— implying that if it faced a recurrent problem of experimentation in the federal trial courts, it would consider using that power. Id. at 17, 27,114 S.Ct. at 1248, 1253. The issue for this court too is whether to use authority we unquestionably have; and while no one would disagree that we should do so very sparingly, *742this is the extraordinary situation that calls for its use.
Reversed.
. Police officers testified that appellant walked with “a very obvious limp.”
. D.C.Code § 22-3502 was repealed on May 23, 1995, by the "Anti-Sexual Abuse Act of 1994,” D.C.Code §§ 22-4101 et seq. (1996), D.C. Law 10-257. Under the new act, "sodomy” is no longer an offense; rather, "sexual abuse" is a prohibited crime involving participation in a "sexual act” involving force, threats, drugs or lack of consent. Id. "Sexual act” is defined as including "[c]ontact between the mouth and the penis." D.C.Code § 22-4101(8)(B).
. More precisely, this pertains to "forms of sodomy involving a penis.” Roundtree v. United States, 581 A.2d 315, 330 (D.C.1990). The penetration requirement is inapplicable to cunnilingus or oral sodomy on a woman's genitalia. Id.
.In the related context of rape or carnal knowledge, it is well settled that "[e]ntry of the anterior of the female genital organ, known as the vulva or labia, is sufficient penetration ... it is not necessary that the vagina itself be penetrat-ed_” Charles E. Torcía, 3 Wharton’s Criminal Law § 278, at 17-18 (15th ed.1995).
.In State v. Pettijohn, 541 S.W.2d 74 (Mo.App.1976) cited by the government, a police officer testified that he saw the defendant asleep in a car with his “penis out of his pants and a little girl with her mouth on his penis.” Id. at 75. "[V]ig-orous" questioning by the attorneys "could not clarify the officer’s testimony as to what ‘on’ ” meant in those circumstances. Id. While therefore there was "a total lack of direct evidence” of penetration, State v. Jenkins, 733 S.W.2d 790, 791 (Mo.App.1987) (discussing Pettijohn), there nonetheless was evidence that the child had engaged in oral sodomy with the defendant while he was asleep on another occasion, and whiskey arid petroleum jelly were found in the car on the present occasion. In these circumstances the Pettijohn court found that the evidence gave "rise to an inference that penetration had occurred.” 541 S.W.2d at 76. No such circumstantial evidence was offered in this case.
. Thus, we need not consider the proposition— which would render the proof even more insufficient here — that only proof of insertion of "the glans or virile end of the organ" into the opening is enough to satisfy the penetration requirement, so that proof even of “nibbling" on the defendant's penis by the victim would have been insufficient. Rozar v. State, 93 Ga.App. 207, 91 S.E.2d 131, 133 (1956).
. The fourth edition of the standard jury instructions does not differ significantly from the third edition in existence at the time of trial. See Butler v. United States, 646 A.2d 331, 334 n. 2 (D.C.1994), cert. denied, - U.S. -, 115 S.Ct. 1326, 131 L.Ed.2d 206 (1995).
. Reversal is necessary in that event, we said, because "absent such misdescription” there is no constitutional error at all, but "with such misde-scription, there is per se reversible error.” 646 A.2d at 337 n. 7 & accompanying text (analyzing Victor and Sullivan).
. In Foreman, we had approved in dicta, as the legitimate equivalent of "abiding conviction,” the phrase "firmly convinced" employed in the Federal Pattern Criminal Jury Instructions. 646 A.2d at 794-95.
. See United States v. Merlos, 303 U.S.App. D.C. 395, 398-99, 8 F.3d 48, 51-52 (1993) (Merlos II) (holding that the error found in Merlos I was not plain error as to the defendant who had not objected), cert, denied, - U.S. -, 114 S.Ct. 1635, 128 L.Ed.2d 358 (1994).
. Our disposition of the case makes it unnecessary to consider appellant’s remaining claims of error, with one exception. The government concedes, and we agree, that the fruits of the police-initiated showup identification are not available to it on retrial, since the police entered appellant’s apartment without a warrant, consent, or exigent circumstances. See Bryant v. United States, 599 A.2d 1107 (D.C.1991) (post-seizure identification evidence inadmissible where it was the product of a warrantless entry into defendant’s rooming house).
. A fourth case pending before a division of this court involves complete rejection of the Redbook instruction by a judge in favor of the model federal instruction on reasonable doubt, discussed infra.
. "The adjective ‘abiding’ borders on the archaic, hence may cany little precise meaning to modem ears; and 'conviction' has its own potential for confusion with, say, conviction for a crime.” Foreman, 633 A.2d at 794-95.