dissenting;
In this distinctly dubious “drop-see”1 drug case, the trial judge excluded relevant testimony on the legally insupportable ground that it did not necessarily establish Collins’ innocence. The apparent misapprehension that evidence has to be decisive in order to be relevant characterized the judge’s ruling both with regard to the proposed testimony of Maurice Lewis and that of Monsen McConnell. The judge also sustained Ms. McConnell’s invocation of the privilege against self-incrimination with respect to matters that were not potentially incriminating.
*496In my opinion, both witnesses should have been permitted to testify. Because the prosecution evidence in this case, which rested on a single officer’s assumption that the drugs fell from Collins’ pocket, was marginal at best, I am satisfied that the errors were not harmless. Accordingly, Collins’ conviction should be reversed and a new trial should be ordered.
I
RELEVANCY
The right to call witnesses in one’s own behalf is basic to our jurisprudence. Chambers v. Mississippi, 410 U.S. 284, 294, 93 S.Ct. 1038, 1045, 35 L.Ed.2d 297 (1973); King v. United States, 550 A.2d 348, 353 (D.C.1988). It is a “fundamental element of due process of law.” Taylor v. Illinois, 484 U.S. 400, 409, 108 S.Ct. 646, 652, 98 L.Ed.2d 798 (1988) (quoting Washington v. Texas, 388 U.S. 14, 19, 87 S.Ct. 1920, 1923, 18 L.Ed.2d 1019 (1967)). The Sixth Amendment speaks in terms of compulsory process, but Collins was constitutionally entitled not only to bring his witnesses to the courtroom but also to put them on the witness stand. King, supra, 550 A.2d at 353.2
The Constitution does not, of course, guarantee a criminal defendant the right to adduce testimony which is irrelevant or otherwise inadmissible under applicable rules of evidence. Taylor, supra, 484 U.S. at 410, 108 S.Ct. at 653; see also Round-tree v. United States, 581 A.2d 315, 329 n. 34 (D.C.1990). When a defendant is on trial and his liberty is at stake, however, courts should not be guided by what Judge Posner has characterized as “crabbed notions of relevance or excessive mistrust of juries.” Riordan v. Kempiners, 831 F.2d 690, 698 (7th Cir.1987). The right to present witnesses on one’s own behalf has a constitutional foundation, and courts should approach it generously rather than grudgingly.
Ordinarily, any evidence which is logically probative of some fact in issue is admissible, Ahrens v. Broyhill, 117 A.2d 452, 455-56 (D.C.1955), unless it conflicts with some settled exclusionary rule. Fowel v. Wood, 62 A.2d 636, 637 (D.C.1948). “[I]f the evidence offered conduces in any reasonable degree to establish the probability or improbability of the fact in controversy, it should go to the jury.” Home Ins. Co. v. Weide, 78 U.S. (11 Wall.) 438, 440, 20 L.Ed. 197 (1870). As Professor McCormick has written,
[i]t is enough if the item could reasonably show that a fact is slightly more probable than it would appear without that evidence. Even after the probative force of the evidence is spent, the proposition for which it is offered still can seem quite improbable. Thus, the common objection that the inference for which the fact is offered “does not necessarily follow” is untenable. It poses a standard of conclusiveness that very few single items of circumstantial evidence ever could meet. A brick is not a wall.
E. Cleary, McCormick on Evidence § 185, at 542-43 (3d ed. 1984) (footnotes omitted).
If the trial judge had adhered to the principles articulated in these authorities, the evidence proffered by Collins would have been admitted, and the result of the trial might well have been different.
II
MAURICE LEWIS’ PROPOSED TESTIMONY
Maurice Lewis was a friend both of Ms. McConnell and of appellant Collins. He visited frequently in the home which Collins had shared with Ms. McConnell until a few weeks before the incident which precipitated this prosecution. Collins’ counsel proffered that Lewis would testify that Ms. *497McConnell had a crack cocaine habit. Counsel argued that “[t]he fact that she has a crack cocaine habit would show ... the regularity with which she would use drugs and the regularity [with] which she would keep them in her home.”
The trial judge viewed this testimony as irrelevant because Lewis had not seen drugs in the apartment “on the day in question.” She stated that the fact that Ms. McConnell had a drug habit “by itself doesn’t mean necessarily [that] there were drugs on the table.... That’s a quantum leap to make.”
The judge’s use of the word “necessarily” was not an aberration. After counsel had presented further argument, the judge remarked again that “the fact that she has a drug problem doesn’t necessarily mean that the drugs were hers." Several pages later in the transcript, the judge put the same thought in a slightly different way:
it would be speculating that because she lives there and in the past has had some drugs there ... that the drugs that were found were definitely hers.
Necessity and certainty were evidently deemed to be preconditions for admissibility.
The judge repeated the same general theme in connection with the defense proffer of Ms. McConnell’s testimony. She declared that whether or not Ms. McConnell displayed drugs in her apartment for friends in May or June “really doesn’t make any difference.” Indeed, in the judge’s view,
[i]t makes no difference whether there were drugs the day before or whatever else. The only question is whether there were any drugs on the table that particular day. Okay. That’s the only thing that’s pertinent. What happened on pri- or occasions doesn’t make one bit of difference.
Courts ought not to be uncharitable when assessing what has been said by a participant in a fast-moving trial. We have generally declined in that context to attribute to remarks made by a prosecutor the most unfavorable interpretation which could possibly be given to them. Irick v. United States, 565 A.2d 26, 33 (D.C.1989). I think the same principle should apply with at least equal force to the comments of a trial judge. Nevertheless, I find the conclusion inescapable that the judge’s evi-dentiary rulings in this case were predicated on a misconception of the proper standard in determining relevance. Repeatedly, the judge stated or implied that if proffered evidence did not “necessarily” prove the point Collins was trying to establish, then it was irrelevant.3
The government claims that the correct analytical rubric here is the line of cases dealing with attempts by a defendant to show that a different individual committed the crime. See, e.g., Brown v. United States, 409 A.2d 1093, 1097 (D.C.1979) (“before evidence of the guilt of another can be deemed relevant and thereby admissible, the evidence must clearly link that other person to the commission of the crime”). More recently, however, this court has explained that
[w]hat we mean by “clearly link,” as used first by this court in Brown, supra, 409 A.2d at 1097, is proof of facts or circumstances which tend to indicate some reasonable possibility that a person other than the defendant committed the *498charged offense. This proof permits the admission of evidence which otherwise is generally excluded because it is too remote in time and place, completely unrelated or irrelevant to the offense charged, or too speculative with respect to the third party’s guilt.
* * * * * *
There is no requirement that the proffered evidence must prove or even raise a strong probability that someone other than the defendant committed the offense. Rather, the evidence need only tend to create a reasonable doubt that the defendant committed the offense. In this regard, our focus is on the effect the evidence has upon the defendant’s culpability, and not the third party’s culpability.
Johnson v. United States, 552 A.2d 513, 516-17 (D.C.1989) (citations omitted) (emphasis in original). See also IA J. Wig-MORE, Evidence § 139, at 1724 (Tillers rev. ed. 1983):
[I]t is not a question of absolute proof, nor even of strong probability, but only of raising a reasonable doubt about A’s commission [of the offense], and for this purpose the slightest likelihood of B’s commission may suffice or at least assist. ... [I]f the evidence is in truth calculated to cause the jury to doubt, the court should not attempt to decide for the jury that this doubt is purely speculative and fantastic but should afford the accused every opportunity to create that doubt. A contrary rule is unfair to a really innocent accused.
Even if the sole purpose of Lewis’ proposed testimony had been to show that the drugs had probably been in the possession of Ms. McConnell rather than of Collins, I am satisfied that it would have been receivable under the Johnson test.
In any event, I do not think we have to reach the Brovm-Johnson line of analysis at all. The law ought not to be an arcane and remote intellectual discipline in which the lessons of real life are sedulously ignored. Everybody knows that drug addicts tend to associate with other drug users and are more likely than, say, Mother Theresa, to have cocaine in their homes and, specifically, on tables in their homes. If Ms. McConnell was a cocaine addict a few months before the incident in question, then this made it more probable than it would otherwise have been that the officer was mistaken in his assumption that Collins was the source of the packet of cocaine.
My colleagues correctly point out that defense counsel failed to proffer precisely when it was that Lewis claimed to have seen Ms. McConnell use cocaine.4 It cannot be gainsaid that the defense posture would have been stronger if counsel had proffered an approximate date. Under the particular circumstances of this case, however, it would be unreasonable to penalize Collins on that account. The judge repeatedly stated that only the events on the day of the arrest were relevant. These remarks would have made any inquiry into the precise dates of Lewis’ visits to the McConnell residence appear futile. Moreover, Collins’ attorney specifically told the court that she would have to speak further with Lewis to obtain more detailed information about these visits, but the judge made her ruling that the proposed testimony was inadmissible before counsel was able to consult with her witness.
Even if Lewis last visited the apartment several months before the incident,51 think evidence to that effect would have been admissible. It is a matter of common knowledge that drug addicts have great difficulty in “kicking the habit.” As I show below, an individual who was a cocaine addict in January is far more likely than a life-long non-user to have the dread substance in her home the following June. *499Ms. McConnell’s use of cocaine in January would thus represent a brick in the “wall” of reasonable doubt, McCORMiCK, supra, § 185, at 542-43, which Collins was attempting to construct, one section at a time. Moreover, any information from Lewis that Ms. McConnell was involved with cocaine a few months before Collins’ arrest would have tended to corroborate Collins’ own testimony that he last saw drugs when he moved out of Ms. McConnell’s apartment at the end of April, 1988.6
Whether Ms. McConnell’s past drug use was relevant is surely one of those issues which might best be approached with what Justice Frankfurter has called the “saving grace of common sense,” Bell v. United States, 349 U.S. 81, 83-84, 75 S.Ct. 620, 622-23, 99 L.Ed. 905 (1955); see also Riggs Nat’l Bank v. District of Columbia, 581 A.2d 1229, 1262 (D.C.1990), rather than by resort to ruminations regarding temporal remoteness or by doctrinal adherence to purported rules of evidence which do not bear the slightest relation to reality. No matter how forbidding our ebony-colored robes may appear, we ought not to be blind as judges to what we know as men and women. Edwards v. Habib, 130 U.S.App. D.C. 126, 140, 397 F.2d 687, 701 (1968), cert. denied, 393 U.S. 1016, 89 S.Ct. 618, 21 L.Ed.2d 560 (1969). We adjudicate in the real world, and our decisions have real consequences for real people.
Although invocation of scholarly works on the nature of drugs and addiction is hardly necessary to resolve the question confronting us, I pause to belabor the obvious. According to the World Health Organization, the characteristics of drug addiction include:
1. An overpowering desire or need (compulsion) to continue taking the drug and to obtain it by any means;
2. A tendency to increase the dose;
3. A psychic (psychological) and, sometimes, a physical dependence on the effects of the drug.
G.F. Uelmen & Y.G. Haddox, Drug Abuse and the Law § 2.3(a), at 2-4 (1990). The Council of the District of Columbia has told us that an addict is one who, among other things, has “lost the power of self-control with reference to his [or her] addiction.” D.C.Code § 33-501(24) (1988).
Drug addiction is an illness. Robinson v. California, 370 U.S. 660, 667, 82 S.Ct. 1417, 1420-21, 8 L.Ed.2d 758 (1962). It results, among other things, in significant changes in the chemistry of the brain, which cannot easily be reversed. Uelmen & Haddox, supra, § 2.3(b) at 2-13 — 2-14. It is surely no secret in 1991 that persistent abuse of drugs is not a malady for which the cure, even for those who seek one, comes simply, reliably or swiftly. Indeed, although short-term detoxification is not difficult,
the problem is that after detoxification — even years after — the ex-addict remains at high risk of relapse. This is the fact that is recognized so correctly and effectively with respect to alcohol addiction by AA, when they regard alcohol addicts as always at risk, no matter how many years of abstinence are behind them. There seems to be something quite fundamental other than tolerance or physical dependence, that drives addicts to begin self-administering an addicting drug and that puts them at risk for relapse long after any physical dependence has worn off.
Dr. Avram Goldstein, The Sidney Cohen Lecture (1986), quoted in Uelmen & Haddox, supra, § 2.3(b) at 2-14 (emphasis added).
Any judge, lawyer, probation officer or treatment specialist who has worked with drug offenders for any appreciable period can attest to the accuracy of Dr. Gold-stein’s analysis. If my colleagues in the majority were hiring a law clerk or a babysitter or even a domestic worker and were advised that he or she abused cocaine some time last year, I suspect that they would discern an appreciable possibility that such *500abuse might recur and that other bad things might happen too. They would surely deem the report significant enough to warrant a thorough investigation of the situation, and would doubtless be grateful to the person who brought the facts to their attention.
The jurors in the present case were making a decision affecting a man’s liberty. They ought not to have been denied access to such important evidence when they were exercising this sobering responsibility.
We have held, as the majority says, that a trial judge’s ruling excluding evidence as insufficiently probative is reviewable only for abuse of discretion. Shepard v. United States, 538 A.2d 1115, 1116 (D.C.1988). But exercise of judicial discretion must be founded on correct legal principles, and a trial court abuses its discretion when it rests its conclusion on incorrect legal standards. In re J.D.C., 594 A.2d 70, 75, (D.C.1991); Conrad v. Medina, 47 A.2d 562, 565 (D.C.1946); Jett v. Sunderman, 840 F.2d 1487, 1496 (9th Cir.1988). This is what happened here. In my opinion, Lewis’ testimony should have been admitted.7
Ill
MS. McCONNELL’S PROPOSED TESTIMONY
As I have noted in the preceding discussion, see page 496, supra, the trial judge invoked, in connection with the proposed testimony of Ms. McConnell, the same constricted notions of relevance which had led her not to permit Maurice Lewis to testify. Relevance rulings became to some extent academic in regard to Ms. McConnell, however, because her testimony was ultimately excluded for a different reason. After counsel was appointed for her, Ms. McConnell invoked her privilege against self-incrimination and declined to testify. The trial judge sustained her claim of privilege with respect to all of the questions which counsel for Collins proposed to propound to Ms. McConnell.
I have no quarrel with that decision insofar as any inquiry regarding Ms. McConnell’s drug use was concerned. Indeed, Collins claims no error in this regard, except that he says, contrary to controlling authority, that Ms. McConnell should have been required to “take the Fifth” in the presence of the jury. Judge Kotelly also declined, however, to permit Ms. McConnell to testify on subjects which had no potential for incriminating her. The judge so ruled on the ground that if Ms. McConnell were permitted to testify on direct examination with respect to matters which would not incriminate her, cross-examination would spill over into areas with respect to which the privilege applied, and could not be fairly limited without impairing the rights of the prosecution.
My colleagues acknowledge, maj. op. at 491, that this decision was incorrect. I agree with them. “Although a criminal defendant has the absolute right not to testify, a witness may invoke the privilege only as to those specific questions to which [the] answers would incriminate [her].” Wilson v. United States, 558 A.2d 1135, 1141 (D.C.1989). The trial judge must make a “searching inquiry” to determine whether the testimony sought to be elicited would in fact be incriminating and, if so, whether the risk of prosecution is “real and substantial.” Id. A witness may not assert a blanket privilege when a narrower assertion will suffice to protect her rights. Id. “When a Fifth Amendment claim is asserted by someone other than a defendant, the court must ordinarily permit examination of the witness (out of the presence of the jury in a jury trial) and rule on the claim of privilege one question at a time.” Id. at 1142 (citation omitted).
*501A blanket ruling that a witness cannot be asked non-incriminating questions because a cross-examiner may later wish to cross into protected territory runs afoul of these principles. If the judge was concerned that it might be unfair to the prosecution to restrict both direct examination and cross-examination of Ms. McConnell to non-incriminating subjects, a voir dire examination outside the presence of the jury would have helped to flesh out the issue. Some reasonable accommodation could surely have been made between Collins’ interest in calling his witness and the government’s right to cross-examine her. There was no reason to believe, short of such a “dry run,” that reconciliation of these interests would not be possible. I am therefore compelled to conclude that Collins’ right to present non-privileged testimony was not accorded the weight to which it was entitled.
IV
HARMLESS ERROR ANALYSIS
A. The Excluded Evidence.
My colleagues say, however, that any error in this regard was harmless. I disagree.
Collins sought to present testimony by Ms. McConnell which would have corroborated his account of his encounter with the police in two important respects. First, he proposed to show that, after the police arrived at the apartment, he declined to come out immediately and, instead, remained in the bedroom for several minutes. Second, he wanted Ms. McConnell to confirm that he had Ms. McConnell’s children with him in the bedroom for most of that time.
If Ms. McConnell had confirmed that Collins waited several minutes before he emerged and that he did not come out immediately, as the officers testified he did, then this would have provided him with a “common sense” argument against the notion that it was he who possessed the packet of cocaine. With a substantial period of time to think about it, no person of reasonable intelligence would keep cocaine in his pocket until he was in the presence of the police. Ms. McConnell having called the police about a man with a gun, Collins, being that man, had ample reason to expect to be searched upon the arrival of the officers. It would thus have made sense for him to dispose in the bedroom of any cocaine then in his possession, rather than to bring it with him to the room where the officers were waiting for him, and thus virtually assure his own apprehension with the contraband.8
The second assertion for which Collins sought confirmation of his own testimony from Ms. McConnell was that her two children were with him for some appreciable time in the bedroom before he complied with the officers’ directive to come out. With due respect to my colleagues, this was not by any stretch of the imagination “a fact too remote to have had any effect on the verdict.” Maj. op. at 492. On the contrary, it was of critical importance.
Collins’ point here is that there is a tremendous difference between a routine call to the police from a woman about a domestic argument and a far more harrowing (and thus less forgettable) situation in which children are in apparent danger from a man who reportedly has a gun and who refuses for some period of time either to free the children or to surrender. It would surely be reasonable for impartial jurors to *502conclude that while busy officers might not recollect much about the details of an incident which fell into the first, more routine category, they would surely remember that, for several minutes, two young lives appeared to be hanging in the balance.9
If Collins’ account was true, then for a significant period of time, the officers must have had serious concerns about the safety of the two youngsters. An impartial jury could reasonably conclude that if the officers did not remember so critical a fact about their visit to Ms. McConnell’s apartment, they probably did not remember much about the occasion, especially about such relatively minor details as precisely what was on Ms. McConnell’s table immediately before the packet of cocaine fell to the ground. If Ms. McConnell — the woman who reported Collins to the police — had verified his account of that aspect of the incident, and if the jury had believed her, the credibility of the officers, especially with respect to their ability to recollect details of events that happened half a year earlier, would have been potentially shattered.
B. The Weakness of the Prosecution Case.
To warrant a holding that trial court error is harmless, the court must be satisfied “with fair assurance, after pondering all that happened without stripping the erroneous action from the whole, that the judgment was not substantially swayed by the error.” Kotteakos v. United States, 328 U.S. 750, 765, 66 S.Ct. 1239, 1248, 90 L.Ed. 1557 (1946); see also Giles v. United States, 432 A.2d 739, 746 (D.C.1981). As we recently explained in Clark v. United States, 593 A.2d 186, 192-193 (D.C.1991),
[t]he Supreme Court’s choice of the words “fair assurance” precludes us from affirming on a mere hunch that the case would have ended with the same verdict if the erroneous ruling had not been made. To conclude that an error is harmless, we must find it highly probable that [that] error did not contribute to the verdict.... Harmless error analysis should not be limited to superficial inquiry as to whether the same verdict would have been possible absent the tainted evidence.
(Citations and internal quotation marks omitted).
As the irrepressible Judge Irving Goldberg had occasion to remark for the court in Chapman v. United States, 547 F.2d 1240, 1250 (5th Cir.), cert. denied, 431 U.S. 908, 97 S.Ct. 1705, 52 L.Ed.2d 393 (1977) (perhaps with his tongue not too far from his cheek), infusion of “harmlessness” into error must be the exception rather than the rule, and “[a] minuscule error must coalesce with gargantuan guilt, even where the accused displays an imagination of Pan-tagruelian dimensions.” See also Clark, supra, at 192 n. 8 (discussing Chapman). Error may be prejudicial rather than harmless even where the prosecution case was strong. United States v. Tussa, 816 F.2d 58, 67 (2d Cir.1987). Obviously, however, courts are less ready to reverse a conviction where the government has presented a strong case rather than a weak one. See Kotteakos, supra, 328 U.S. at 764, 66 S.Ct. at 1247-48.
As I have attempted to demonstrate above, the significance of the evidence which the judge excluded in this case was hardly “minuscule”. Chapman, supra, 547 F.2d at 1250. At the same time, the evidence of guilt was anything but “gargantuan.” Id. Even if we treat these adjectives as rhetorical, we cannot say with “fair assurance” that the error did not substantially sway the jury. In fact, this may *503be the most troubling “drop-see” case which has come to my attention in more than eleven years as a judge.
In this jurisdiction, it is at least the norm in such cases, if not the invariable practice, that at least one prosecution witness is able to testify that he or she observed the drugs in the defendant’s possession prior to the ' alleged drop. Assuming the correctness of my spelling, that is what “drop-see” means. Indeed, it is my impression that as a practical matter, cases in which no such evidence is available are seldom, if ever, approved for prosecution by the United States Attorney. There is ample reason for hesitation and caution where the officers did not see the drop. Most “drop-sees” occur in locations where drug users and traffickers congregate. There are often many people who might have discarded the contraband upon the approach of the police. When drugs are found near an individual in a “high narcotics area,” this does not necessarily mean that this individual possessed them.
In the present case, Officer Taylor did not observe the drop at all. Officer Savage testified that he saw the packet “fall from the direction of the defendant.” He did not actually see the cocaine come out of Collins’ jacket pocket, although it came from that direction. Indeed, he first became aware of the packet just before it hit the ground. Officer Savage “assume[d] that [the cocaine] fell from [Collins’] pocket because there was no other place [from which] it could fall.”
In spite of the foregoing, Officer Savage testified that he told Officer Taylor that he had seen the drop.10 Officer Taylor subsequently completed an arrest report in which he recited that “a small plastic bag fell from the defendant’s right jacket pocket.” There was nothing in the report which disclosed that no one had seen the cocaine fall from Collins’ pocket, but that one officer had assumed that this was what had occurred.11
In any event, the evidence that Collins ever possessed the cocaine was altogether circumstantial. Circumstantial evidence may often be as persuasive as direct evidence, or more so. Janifer v. Jandebeur, 551 A.2d 1351, 1352 (D.C.1989). It is imperative, however, that all of the relevant circumstances be disclosed to the jury, and that necessarily includes exculpatory circumstances.
In the present case, one of the government’s principal contentions is that the cocaine could not have fallen from the table, as Collins suggests, because neither officer had observed it there. The accounts of the two officers were not consistent with each other, however, regarding what was actually on the table. Officer Savage recalled seeing “maybe two large books and nothing else.” Officer Taylor observed some books and “maybe a satchel or briefcase on the left side of the table.” If Collins could have established, through the testimony of Ms. McConnell, that at the time of these observations the officers had to be concerned with the safety of two children who were in a bedroom with an uncooperative gunman, the jurors might reasonably have wondered whether the officers were in a position to devote their attention to the *504table and to the items located on it.12
But that is not all. Where the jurors are asked to deduce that the cocaine “must have” fallen from Collins’ pocket even though nobody saw this happen, simple fairness requires that they be apprised of circumstances which may point to an entirely different conclusion. Evidence from witnesses other than the defendant that Ms. McConnell was or had been a drug user, and that Collins had plenty of time to discard the drugs in the bedroom after the officers arrived, could potentially have damaged the government’s case quite severely and allowed the jury to draw inferences favorable to Collins. Unfortunately, it was never admitted.
V
CONCLUSION
It may be that the jury’s verdict would have been the same if the excluded testimony had been received as evidence. In my opinion, however, it is just as possible that the decision would have been different if correct evidentiary rulings had been made. The prosecution case appears to me to have been marginal or worse. For the reasons stated, I believe that the exclusion of probative evidence tending to show innocence, or at least to create a reasonable doubt of guilt, was prejudicial rather than harmless. I am therefore compelled to conclude that Collins did not receive the fair trial to which he was entitled. Accordingly, I respectfully dissent.13
. "Drop-see” (or dropsie?) is the nickname given by veterans and aficionados of the drag wars in the Superior Court to cases in which it is claimed that the defendant dropped the drags in order to forestall the disadvantageous prospect of being apprehended with contraband on his or her person. Sometimes the stratagem works and sometimes it doesn't.
. It is not the defendant’s interest alone which is implicated by the exclusion of relevant testimony. "For more than three centuries it has now been recognized as a fundamental maxim that the public (in the words sanctioned by Lord Hardwicke) has a right to every man’s [and woman’s] evidence." VIII J. Wigmore, Evidence § 2192, at 70 (McNaughton rev. ed. 1961). This rule is required by, and basic to, “the public interest in the search for truth.” United States v. Bryan, 339 U.S. 323, 331, 70 S.Ct. 724, 730, 94 L.Ed. 884 (1950).
. In fairness to the trial judge, I must agree with the majority that Collins’ counsel’s persistent casting of a simple issue of relevance in terms of complex and inapplicable doctrines, e.g. habit, did not make the judge’s task any easier. As this court stated in Johnson v. United States, 398 A.2d 354, 365-66 n. 8 (D.C.1979) (quoting with deletions MAGRUDER, The Trials and Tribulations of an Intermediate Appellate Court, 44 CORNELL L.Q. 1, 3 (1958)):
we must always bear in mind that [trial judges] may be as good lawyers as we are or better. They are under the disadvantage of often having to make rulings off the cuff ... in the press and urgency of a trial....
A more precise identification and articulation of the issue by counsel would have reduced the prospect of an incorrect ruling by the court. Nevertheless, counsel did make an argument based on relevance, and in my opinion the remarks by the trial judge quoted in the text unambiguously reject the proffered testimony because it was inconclusive and thus, in the judge’s view, irrelevant, and not on the basis of arcane considerations applicable to the law of habit.
. The judge asked counsel how frequently Lewis had visited the apartment. Counsel responded that he had done so on "many, many occasions.” The judge responded that "many, many occasions doesn’t make it. Tell me something more specific.” Counsel responded that she would have to question her witness but, as noted in the text, page 498, infra, never had the opportunity to do so.
. That is what Collins’ counsel suggested at oral argument in this court.
. On cross-examination, Collins testified that he had knowledge, but not personal knowledge, that Ms. McConnell continued to have cocaine in the apartment through the month of May. There was no objection to this hearsay evidence.
. The majority’s remand on this issue is, in my view, a more reasonable disposition than the affirmance suggested by the government. Nevertheless, the majority’s view of relevance (and trust of juries), see Riordan, supra, 831 F.2d at 698, is far narrower than mine. At least in my view, my colleagues leave the trial judge far too much leeway on remand to exclude the evidence. If Maurice Lewis saw Ms. McConnell use drugs at the apartment a few months before Collins’ arrest, elementary fairness requires that the jury must be apprised of this information. Our mandate ought to leave no doubt on that score.
. Concededly, it is not impossible that Collins emerged with cocaine on his person even if he had several minutes after the arrival of the officers to think of the best way to dispose of it. As the majority points out, Collins first had the opportunity to discard the cocaine during the period after Ms. McConnell called the police but before the officers arrived at the apartment. He would also have had time to get rid of the incriminating substance even if he had walked out of the bedroom seconds after being directed to do so by the police. The government’s evidence, if credited, would thus have been sufficient to prove guilt even if Ms. McConnell’s proposed testimony had been admitted. This does not, however, render the erroneous exclusion of that testimony harmless. A jury might reasonably find that, no matter what Collins had done before the police arrived, several minutes in another room when officers were already in the apartment would be a sufficient time for him to remember to dispose of the cocaine. See the authorities discussed in the text at 496-497, supra.
. My colleagues take me to task, maj. op. at 492 n. 5, for suggesting without "record basis” that if Collins’ version was accurate, the children were in apparent danger. I do not agree that bland articulation is de rigueur when a more vivid expression accurately captures the scenario. Surely my colleagues do not believe that if a man with a gun refuses for several minutes, in defiance of a police directive, either to come out or to release two young children who are with him, a reasonable police officer would be unconcerned about the children’s safety. Such a suggestion would necessarily have its source, to use Judge Wisdom's captivating phrase, in the “eerie atmosphere of never-never land.” See Meredith v. Fair, 298 F.2d 696, 701 (5th Cir. 1962).
. The cross-examination proceeded as follows:
Q Officer Savage, did you ever at any time tell Officer Taylor that you saw the white rock substance fall from Mr. Collins’ right jacket pocket?
A Yes, I did.
Q You did tell him that?
A Yes.
Q And under what circumstances did you tell him and when?
A I told him on the scene, I believe right after it hit the ground, I told him.
Q Now, you testified earlier, though, didn’t you, Officer, that you didn’t actually see it fall from the pocket; isn’t that right?
A I didn’t actually see it fall out of the pocket, that’s correct.
Q But then you turned around and told Officer Taylor that you did see it fall from the right pocket?
A I don’t know if I told him the pocket, I may have said it fell from the defendant’s — I can’t recall whether I said pocket ...
. It may well be that the prosecution in this case would never have been authorized if the officer’s assumption had been disclosed in the report.
. The prosecution also presented, as circumstantial evidence of Collins’ possession of the cocaine, the state of his right jacket pocket. Officer Savage testified that “the white lining on the inside was partially sticking out, just about half an inch, as if when you pull your hands out of your pocket in a hurry and you pull your pocket halfway out, it was like that." Officer Taylor made essentially the same point. Officer Savage’s testimony established, however, that this contention proved too much. Only one bag of cocaine was found, and Officer Savage testified that "they were kind of puffed out at the time, both pockets.” Something other than the cocaine must therefore have “puffed out” the left pocket.
. The government served enhancement papers in this case, and Collins was sentenced to imprisonment for twenty-four months, of which all but eight months were suspended. He thus may have completed serving his sentence long ago. He was both on parole and on probation at the time of this alleged offense; the record does not reveal whether revocation proceedings were instituted.
Following his conviction, Collins applied for bond pending appeal. He noted that possession of one bag of cocaine was not a dangerous crime and that he was presenting significant issues on appeal. Citing Barry v. Washington Post Co., 529 A.2d 319 (D.C.1987), a civil action brought pursuant to the Freedom of Information Act, as authority on the question whether a stay of the sentence should be granted, the judge denied bond on the ground that Collins had not shown a likelihood of success on the merits, or irreparable injury, or that "the public interest favors the granting of the stay.” Aside from the question whether cases involving stays of orders in civil proceedings have any bearing at all on this kind of situation — and I do not think they have, for the governing statute is D.C.Code § 23-1325(c) (1989) — loss of liberty pending appeal represents, at least to me, the quintessential irreparable injury. If the conviction is reversed, Collins can never get back the time he spent under lock and key.
I therefore suggest that this is the kind of case in which bond pending appeal might have been favorably considered, assuming that a sufficient showing could be made pursuant to § 23-1325(c). If a conviction is ultimately reversed, the defendant’s victory is more illusory than real if he has already served his sentence. To a wrongfully convicted and incarcerated individual, the right to appeal under such circumstances may understandably seem little more than a hollow mockery.