(concurring in part and dissenting in part):
The majority concludes that specification 1 of Charge I must be amended by striking the words “at divers times, from on or about November 1987 to”; and that the words “at or near Houston, Texas, and” as well as the words “at divers times” must be stricken from specification 2. I agree with this conclusion, but I would go further because, unlike the majority, I am unpersuaded that the other evidence corroborated appellant’s confession that he had used marijuana. On the other hand, I agree with the majority that appellant’s confession to using cocaine on New Year’s Eve in Houston was adequately corroborated by independent evidence.
I
Mil.R.Evid. 304(g), Manual for Courts-Martial, United States, 1984, prescribes:
An admission or a confession of the accused may be considered as evidence against the accused on the question of guilt or innocence only if independent evidence, either direct or circumstantial, has been introduced that corroborates the essential facts admitted to justify sufficiently an inference of their truth.
(Emphasis added.) The majority of this Court has recognized that, to satisfy this rule, “ ‘the quantum of evidence’ needed to raise such an inference is ‘slight.’ ” United States v. Yeoman, 25 MJ 1, 4 (CMA 1987), quoting S. Saltzburg, L. Schinasi, & D. Schlueter, Military Rules of Evidence Manual 139 (2d ed.1986). Accord United States v. Melvin, 26 MJ 145, 146 (CMA 1988) . It suffices that “there is substantial independent evidence that the offense has been committed, and the evidence as a whole proves beyond a reasonable doubt that defendant is guilty.” Id., quoting Smith v. United States, 348 U.S. 147, 156, 75 S.Ct. 194, 199, 99 L.Ed. 192 (1954). The independent evidence need not support every element of the offense or itself prove the defendant’s guilt beyond a reasonable doubt.
It is not the quantity of corroborating evidence that concerns me here; rather, it is the quality of that evidence. The whole purpose served by the requirement for corroborative evidence is to ensure the reliability of the confession as a basis for conviction. As the Supreme Court observed in Smith, not only the involuntary, coerced confession is the. focus of concern, for there are “separate doctrines [that] exclude involuntary confessions from consideration by the jury____” As the Court put it:
[T]hough a statement may not be “involuntary” within the meaning of this exclusionary rule, still its reliability may be suspect if it is extracted from one who is under the pressure of a police investigation — whose words may reflect the strain and confusion attending his predicament rather than a clear reflection of his past. Finally, the experience of the courts, the police and the medical profession recounts a number of false confessions voluntarily made____ These are the considerations which justify a restriction on the power of the jury to convict, for this experience with confessions is not shared by the average juror.
348 U.S. at 153, 75 S.Ct. at 197.
To satisfy these reservations about reliability of confessions — without, at the same time, unduly restricting appropriate use of confessions by factfinders — the Supreme Court has established the rule that a confession may support a conviction “as long as there is substantial independent evidence that the offense has been committed, and the evidence as a whole proves beyond a reasonable doubt that defendant is guilty.” Id. at 156, 75 S.Ct. at 199 (emphasis added.) When I consider the nature of appellant’s confession to using marijuana, I am convinced that the quality of the evidence cited by the majority to “corroborate” that confession does not satisfy this rule.
Part of the difficulty arises from the extremely general nature of appellant’s confession. Where a confession is replete with detail, the opportunity for corrobora*84tion is rather generous. For example, in their Analysis of Mil.R.Evid. 304(g), the drafters of that provision offer this hypothetical:
[P]resume that an accused charged with premeditated murder has voluntarily confessed that, intending to kill the alleged victim, she concealed herself so that she might surprise the victim at a certain place and that when the victim passed by, she plunged a knife in his back. At trial, the prosecution introduces independent evidence that the victim was found dead as a result of a knife wound in his back at the place where, according to the confession, the incident occurred. This fact would corroborate the confession because it would support an inference of the truth of the essential facts admitted in the confession.
Drafters’ Analysis, Manual, supra at A2213 (Change 3). In this example, the prosecution’s opportunity for corroborating the confession is ample due to the detail of the confession: identity of the victim; instrument used to kill him (a knife); locus of the fatal wound (the back); and place where the murder occurred. Independent evidence of all the detail of the confession adequately corroborates the trustworthiness of the confession and its only uncorroborated factor: identity of the culprit.
The prosecution in this case, though, was not so fortunate. Here the beginning and the end of appellant’s confession to marijuana use of any sort was this passage: “I have smoked marijuana since being here. Why? I don’t know.” I interpret this language to mean that Rounds had used marijuana after being transferred to Randolph Air Force Base. However, it provides no basis for inferring where he smoked the marijuana — whether at Randolph Air Force Base, elsewhere in San Antonio, or somewhere else in Texas. In my view, when a confession is this general, the quality of the independent evidence offered to corroborate the fact that the offense was committed must be rather high if the corroboration is to satisfy the concerns discussed earlier.
To illustrate — and by analogy to the Drafter’s Analysis of Mil.R.Evid. 304(g) quoted above — assume that an accused confessed that he had used marijuana in his barracks with Private Doe and PFC Roe on a number of occasions since arriving at his current installation. The prosecution then offers evidence that someone had seen appellant in his barracks room on that installation, with Private Doe and PFC Roe also in the room, and that a burning marijuana cigarette lay in an ashtray within reach of all three. Even though there was no independent evidence directly showing that appellant used the marijuana on the occasion mentioned, the “dovetailing” (in the majority’s words) of the detail of the confession and the detail of the independent evidence adequately meets the requirement that the latter corroborate that the offense confessed to occurred.
What is the nature of the “corroborating” evidence offered here to support appellant’s detail-deficient confession? It consists of testimony of two fellow airmen — Longworth and Sax.1
Longworth testified that, during December 1987, he saw appellant and Sax standing near a burning marijuana cigarette in an ashtray in the bathroom/passageway between Longworth’s and Sax’ room; he testified that he never saw appellant have anything to do with the cigarette.2 Sax *85testified that, sometime in late November or early December 1987, he had smoked marijuana in this passageway. He and three friends (including appellant) had returned from drinking; appellant left at some point to go turn on some music and, while he was gone, Sax and the others lit and smoked a marijuana cigarette; Sax then left to turn on some music and returned to find his friends (including appellant) in the passageway with a marijuana cigarette burning in an ashtray. He testified that he never saw appellant have anything to do with the cigarette.
It is not apparent to me that Longworth and Sax are addressing the same incident; the dates do not coincide; and the accounts differ in other respects, such as the number of persons present. The two versions are not mutually supportive in a way that gives either one greater weight as corroboration than it would have on its own.
Furthermore, the details of these incidents are not necessarily such as to constitute great high-quality evidence that Rounds used marijuana. The locus in each case was a shared bathroom passageway between the rooms of Longworth and Sax. Although appellant shared Longworth’s room as his roommate, this passageway was not uniquely under appellant’s control but, instead, was equally under the control of and accessible to Longworth and Sax, as well. In each instance, the marijuana cigarette was in an ashtray that was not uniquely accessible to Rounds. On each occasion, at least one other person was near the ashtray; and so, there is no necessary inference that the burning cigarette had been lit or used by Rounds. Neither Longworth nor Sax could testify that Rounds either smoked or touched a marijuana cigarette.
If appellant’s confession had been more detailed and if those details had “dovetailed” with the details in Longworth’s and Sax’ testimony, my conclusion would be different. Or, perhaps, even with a confession as general as this one, if the independent evidence were more indicative of use by appellant (e.g., if he had held the cigarette or had demonstrated physical manifestations of having used marijuana), I could find the necessary corroboration. However, here the prosecution evidence does not constitute the “substantial independent evidence that the offense has been committed” (see Smith); and for this reason, I must dissent from affirming any finding of guilty of marijuana use.
II
Applying a similar approach to the alleged use of cocaine in Houston at a New Year's Eve party, I agree that the independent evidence is substantial enough — both in quantity and quality — to corroborate appellant’s confession. As summarized by the majority, the detail of that evidence (such as the unique time and the uncommon place) satisfactorily “dovetails” with the detail of that incident as reflected in appellant’s confession.3
III
Accordingly, I would reverse the decision below except to the extent it concerns appellant’s use of cocaine in Houston on New Year’s Eve; and I would dismiss specification 2 of Charge I. Since appellant then would remain convicted only of a single drug offense — not the “multiple drug of*86fenses” upheld by the majority, see 30 MJ 76, 82 — I would return the record for the Court of Military Review to take appropriate action on the sentence.
. The military judge instructed the court members on accomplice testimony and advised them that "[w]hether ... Airmen Longworth and Sax, who testified as witnesses in this case, were accomplices is a question for you to decide.”
. On direct examination of Longworth, trial counsel brought out, over defense objection, that Longworth had made a statement to the OSI that Rounds and Sax had been smoking the marijuana cigarette when he arrived and that then they smoked it together. Apparently this statement was not made under oath, so it would be considered hearsay. See Mil.R.Evid. 801(d)(1). Thus, it was not admissible as substantive evidence (Mil.R.Evid. 802), although it could be used to impeach Longworth. The defense objection was not made specifically on hearsay grounds; and it could be argued that any hearsay objection was waived. See Mil.R. Evid. 103. In that event, Longworth’s testimony *85would be legally sufficient to sustain the findings. However, since the Government did not advance this theory at trial or on appeal, it would be inappropriate to rely on it at this stage in the proceedings.
. Airman Sax described how, at the party in Houston, Rounds had provided him a dollar bill that he could use in snorting some cocaine that was present. Under the circumstances, this assistance would probably make Rounds criminally liable as a principal for Sax' use of cocaine. Art. 77, Uniform Code of Military Justice, 10 USC § 877. The Government, however, did not advance this theory, and the military judge did not instruct thereon. Therefore, I need not decide whether independent evidence of being a principal to a crime constitutes corroboration of a confession to committing the crime as the active participant.