with whom GOD-FREY, Justice, joins, concurring.
I concur in part I of the majority opinion and in the result reached in part II. I find it necessary, however, to state separately my views concerning the missing witness instruction discussed in part II. This is the first case since the adoption of the Maine Rules of Evidence where we are squarely faced with the question of the propriety of a jury instruction on a defendant’s failure to call a witness. Because of the change effected by adoption of the evidence rules, we should now reexamine the limitations upon the missing witness instruction in a criminal case.
The significance of the adoption of the Maine Rules of Evidence lies in rule 607, which states: “The credibility of a witness may be attacked by any party, including the party calling him.” This rule changed traditional Maine practice, under which a party “vouched” for the credibility of a witness called by him, see State v. Fournier, Me., 267 A.2d 638, 640 (1970); R. Field and P. Murray, Maine Evidence § 607.1 (1970), and thus was unlikely to call a witness whose testimony might be adverse.
The relationship between the vouching concept and creating an inference from a missing witness is illustrated by the approach taken by the First Circuit Court of Appeals in United States v. Wright, 573 *209F.2d 681 (1st Cir. 1978). There, the court found no error in the prosecutor’s comments on the defendant’s failure to call two witnesses, even though, as the defendant argued, the witnesses were present in court and thus were equally available to the prosecution. The court adopted the position that “a witness’s practical and legal availability is to be determined on the basis of his disposition and relationship towards the parties.” Id. at 684. Pointing to the close association of the witnesses to the defendant, extending to their sitting with him during trial, the court found that their relationship with the defendant and the expected nature of their testimony made them not equally available to the prosecutor, so his comment was permissible. Id. See also United States v. Johnson, 467 F.2d 804, 809 (1st Cir. 1972); McClanahan v. United States, 230 F.2d 919 (5th Cir. 1956) (defendant’s attorney); Dent v. United States, 404 A.2d 165, 170 (D.C.App.1979) (family members).
Even prior to the change in our rules of evidence there was a logical inconsistency in saying (1) the prosecution can rely on the adverse inference to be drawn from a defendant’s failure to call a witness supposed to be friendly to the defendant and (2) the prosecutor can excuse his failure to call that witness, who was equally available to him, on the basis of practical unavailability because the witness is predisposed to testify favorably to the defendant. If the one inference that the witness is predisposed to testify favorably to the defendant justifies the prosecutor’s failure to call the witness, then how can the prosecution ask the jury to draw the opposite inference, i. e., that the witness’s testimony would have been unfavorable to the defendant? I do not believe that these circumstances support any inference at all.
Even where the witness is not equally available to the prosecutor, allowing an inference to be drawn from the defendant’s failure to introduce evidence distorts the allocation of the burden of proving his guilt. Ordinarily, the defendant has no obligation to present evidence on his behalf. Yet, according to the missing witness instruction, the state may be able to rely on an inference arising from his failure to produce evidence as evidence against him, thus effectively “creatpng] evidence from none-vidence.” Dent v. United States, 404 A.2d 165, 171 (D.C.App.1979). Absent action by the defendant that by itself demonstrates a consciousness of his guilt, as, for example, by actively procuring the witness's absence, the state should not be able to gain an advantage from the exercise of his right not to be compelled to produce evidence.
My concern is not with impeaching the defendant’s credibility or attacking the weight of his testimony by commenting on its lack of corroboration. Therefore, the prosecutor’s comment on the lack of support for the defendant’s alibi might have been allowable here, except that he knew the witness was unavailable to either party. See State v. Gilcott, Me., 420 A.2d 1238, 1240 (1980). The trial court, however, went further and specifically instructed the jury that it could infer that the absent witness’s testimony would be adverse to the defendant. Instructing that the defendant’s failure to call a witness can rise to the level of affirmative evidence assisting in the prosecution of the case has the effect of instructing that the defendant has assumed a burden of proof or, at the least, that the state’s burden has decreased.1 I find that proposition unacceptable. See Jackson v. Virginia, 443 U.S. 307, 99 S.Ct. 2781, 61 L.Ed.2d 560 (1979); In re Winship, 397 U.S. 358, 90 S.Ct. 1068, 25 L.Ed.2d 368 (1970).
. As the majority opinion points out at page 207, the use of the inference can be subject to abuse. Some courts have promulgated prophylactic measures to prevent abuse. See Dent v. United States, 404 A.2d 165 (D.C.App.1979).
Such measures frequently involve an exploration of collateral issues to an extent unwarranted by any probative value inherent in the inference and frequently fail to eliminate the abuse.