dissenting:
*422It was improper for the District Court to make the factual determination during sentencing that the defendant did not tell the truth when he testified at another trial. Accordingly, I would reverse the order of the District Court and order defendant be resentenced without regard to whether a trial judge determined that defendant committed perjury.
The majority entirely misconceives the nature of the problem presented and the standards of review to be applied to the District Court proceedings. The issue is not a factual one of determining whether there is substantial evidence in the record to ultimately support the District Court’s finding that another District Court had sufficient evidence before it to justify sentencing defendant to the maximum of 40 years in prison for robbery. Due to the nature of the offense and defendant’s previous criminal background, there is no doubt defendant was a well qualified candidate for a lengthy stay in prison. However, that is not the issue. Rather, the issue is strictly a legal one —whether a District Court can add 15 years to a defendant’s sentence because it believes the defendant committed perjury, or in the softer terminology some courts prefer to use, the defendant was not candid in his testimony. Neither the reviewing District Court nor this Court came to grips with this issue.
I believe we are bound by the unequivocal testimony of the sentencing judge in determining whether the sentence was improper. At the post-conviction hearing he testified:
“Q. But you did then in this particular case put heavy emphasis in the sentencing on the fact that you felt that the defendant was not candid in the testimony at the Melody Boykin trial? A. About 15 years worth, I think.” (Emphasis added.)
The sentencing judge also explained what he meant by defendant’s lack of candor:
“He wasn’t candid to the Court in that he didn’t tell the truth. I don’t know if you would define it as perjury or I would have done it as perjury. There were things about what transpired with this event that just were not explained by him that were incredible to me.” (Emphasis added.)
*423Clearly, this is an unequivocal admission from the sentencing judge that he added 15 years to defendant’s sentence solely because “he didn’t tell the truth” at the Melody Boykin trial.
The majority gave no factual background of the circumstances surrounding the Melody Boykin trial. Defendant and four others, including Melody Boykin, were arrested for robbery and aggravated assault. The victim was lured into an alley, then was rolled and severely beaten. Melody Boykin was apparently defendant’s girlfriend. Three of these people plead guilty to robbery and the aggravated assault charges were dismissed without prejudice. It was understood that they would testify against Melody Boykin who plead innocent, and they did. Defendant also plead guilty and the aggravated assault charge was dismissed without prejudice. Defendant also testified at the trial, but Melody Boykin was acquitted, much to the disappointment of the District Court judge who presided over the trial and who was later to sentence defendant. The trial court obviously thought defendant did not tell the truth during the Melody Boykin trial and thereby helped gain her acquittal. For this testimony the trial court added 15 years onto defendant’s sentence. This is an intolerable abuse of sentencing discretion.
The majority erroneously relied on the findings made by the trial court at the conclusion of the post-conviction proceeding, quoting in full from finding number nine. However, these findings were not made on the basis of the sentencing court’s testimony at the post-conviction hearing. Rather, they were based entirely on the record of the sentencing hearing. Surely this Court is in a better position to review the sentencing record. I say better position because it is extremely unrealistic to assume that one District Court will fairly and impartially review the decisions of another District Court judge. Human nature simply does not work that way.
In Worden v. Alexander (1939), 108 Mont. 208, 90 P.2d 160, a trial judge who did not preside over the trial itself ruled on a motion for a new trial. As to reviewing the record of the trial on appeal, this Court stated:
“ * * * where a motion for a new trial is heard by a judge who did *424not try the case, such judge * * * was compelled to gain his knowledge of the case from the record alone, and was therefore in no better position to determine the motion than is this court, and hence his order does not carry with it the presumption usually indulged in favor of such order. * * *” 108 Mont. 211-212, 90 P.2d 161-162.
This Court extended this principle of review to findings and conclusions where one trial judge enters findings and conclusions in a case over which he did not preside. Phalen v. Rilley (1970), 154 Mont. 399, 403-404, 465 P.2d 102. Under these cases this Court could just as easily review the record of the sentencing hearing and come to its own conclusions as to the propriety of the sentence.
On the other hand, if the District Court made its findings and conclusions in part in reliance on the testimony introduced at the post-conviction hearing, he was clearly in error that the entire sentence was made for a variety of justifiable reasons. It is obvious that the District Court totally ignored the testimony of the sentencing judge who admitted he added 15 years to the sentence because the defendant did not tell the truth at the Melody Boykin trial. This fact is not even mentioned in his findings. I do not believe this Court can also in good conscience ignore that testimony.
But the majority does ignore the record when it states:
“The record shows the sentencing judge considered many factors in sentencing petitioner. The comments of the sentencing judge relied upon by petitioner indicate only that the judge believed petitioner was not candid with the court and took this into consideration while imposing sentence. * * *” (Emphasis added.)
It is simply not a matter of taking defendant’s candor into consideration; it is a matter of sentencing him to an additional 15 years because the defendant displayed a lack of candor. Certainly, it is logical to assume by the sentencing judge’s testimony that, if he had not considered defendant’s testimony at the Melody Boykin trial, the sentence would have been 25 years in prison rather than 40.
The apparent reasoning of the majority is that a defendant’s can*425dor may be considered and apparently used to augment a sentence up to the maximum because candor reflects on one’s character and capacity for reformation. This reasoning ignores the more fundamental constitutional protections to which one is entitled, but which have now been stripped away.
Presumably the sentencing judge determined that the defendant either committed perjury during the Boykin trial (although he admitted in effect it could not be proved), or that the testimony involved was such as to adversely reflect on defendant’s capacity for reformation. In either situation, the District Court had no right to make this factual determination and punish defendant without benefit of trial.
The constitutional hazards of increasing a sentence for alleged perjury committed during the course of a trial were discussed at 66 Yale Law Journal 204, 212-213:
“The perjury rationale for increasing sentence may be viewed from two different perspectives. It may be said that the judge is awarding the defendant a given punishment for the crime of which he has been convicted, and then, within the limits of his discretion to fix punishment for this offense, is imposing an additional sentence because the defendant has committed the second crime of perjury. Or the court may be said to consider the occurrence of perjury as a culpable act bearing upon the character of the accused; accordingly, the defendant is given a longer sentence for the crime of which he stands convicted because his perjurious conduct increases the difficulty of reformation. Both of these justifications seem unsound.
“Penalizing the defendant for the substantive crime of perjury by increasing the sentence for another offense contradicts basis tenets of criminal law. Since perjury is properly punishable in a separate criminal proceeding, a summary adjudication by the court of the defendant’s guilt is an inadequate substitute for the constitutional safeguards inherent in a new indictment and jury trial. Moreover, even though the defendant’s conviction must be taken as a repudiation of his testimony, the judgment in the initial case would un*426doubtedly be inadmissible evidence in a subsequent perjury proceeding. In light of this doctrine, the practice of conclusively presuming the commission of perjury from the fact of conviction is particularly suspect.”
In addition to a charge of perjury, the District Court, of course, would have the choice of proceeding against a defendant for criminal contempt. But even in such situations the defendant would have his procedure protections. As stated in 66 Yale Law Journal 204, 213-214:
“A defendant is not always entitled to a separate trial on the issue of perjury; lying under oath in a judicial proceeding may at times be punishable as criminal contempt. Under rule 42(a) of the Federal Rules of Criminal Procedure, criminal contempt may be punished summarily if the judge certifies that the conduct constituting the contempt was committed in the presence of the court. However, the Supreme Court has held that perjury standing alone does not ‘obstruct the administration of justice’ and hence does not by itself constitute criminal contempt under the applicable federal statute. To be punishable, the perjury must be such that it blocks the performance of judicial duty. Although the application of this doctrine has evoked considerable confusion, it is questionable whether the requisite obstruction of justice could be said to occur when the defendant has been convicted in spite of his perjured testimony. In addition, cases have indicated that the bona fide belief of a presiding judge that a witness has committed perjury does not by itself justify summary punishment of the contemnor without due notice and hearing.”
However, when a sentencing court punishes the defendant by an increased sentence as though he had committed perjury, he deprives the defendant of the procedural protections that he would have had in a criminal contempt proceeding. In 66 Yale Law Journal 204, 214, the legal consequences are described:
“When a judge increases the sentence awarded for the charged crime as a penalty for perjury not summarily punishable as contempt, he utilizes his discretion to accomplish a result he could not *427have effected directly. The rigid requirements which must be met for perjury to qualify as criminal contempt reflect the policy that the contempt power should be strictly applied, since the contemnor is sentenced without the normal safeguards of procedural due process. Such a policy is jeopardized when a judge increases sentence for another crime to discipline a defendant whose perjured testimony, though not qualifying as contempt, has been felt to offend the dignity of the court.
“Assuming that the perjury of the accused would be subject to punishment as criminal contempt, the court that awards a more severe sentence on the basis of the defendant’s commission of perjury may be said to substitute one form of summary punishment for another concededly within its power. At first glance, such a practice does not seem objectionable. However, it should be noted that a sentence within the legal limits provided for the proven crime is not usually subject to appellant review. Accordingly, a defendant whose sentence has been increased as a penalty for perjury felt to constitute contempt is denied the right of appeal which he would have enjoyed if the court has followed the standard procedure for summary punishment of contempt. Such a deprivation is particularly significant in view of the frequent misapplication by trial courts of the standards for punishing perjury as criminal contempt.”
See also, Scott v. United States, 136 U.S.App.D.C. 377, 419 F.2d 264, where the problems involved in considering the candor of a defendant, either as perjury or as criminal contempt, are discussed in detail. Scott unequivocally condemns this practice.
Assuming moreover that a court sentences a defendant to additional time in prison either because it believes the defendant has committed perjury at his own trial, or, as in this case, at another’s trial, is it really an ironclad indicia of the defendant’s capacity for reformation or repentance? This point has also been discussed in 66 Yale Law Journal 204, 216-217, where the author states:
“The prior criminal conduct of a defendant awaiting sentence is acknowledged to be a gauge of his antisocial propensities, and *428hence useful in predicting the sentence necessary to effect reformation. Accordingly judges consider, in addition to convictions, other purported illegal activities of the accused as revealed by presentencing reports or other sources of information. Viewed from this perspective, perjury apparently committed at trial may be classified an illegal activity properly incorporated into the defendant’s criminal record. And, since past criminal behavior is a factor influencing length of sentence, a pro tanto increase in the punishment accorded a defendant suspected of perjury may at first glance seem defensible.
“However, when a judge awards a defendant believed to be a perjurer a more severe sentence than he would have given a defendant pleading guilty to the same offense but otherwise possessing an identical criminal record, such a discrimination is justified only if the perjury indicates a distinction in the character of the two defendants. An individual willing to commit a crime would quite likely have no moral scruples against subsequently denying under oath its commission in order to escape punishment. In the usual case, a defendant is motivated to admit guilt not by an aversion to perjury but by the realization that his plea may be effective in mitigating sentence. For an accused confronted with incriminating evidence, a guilty plea may appear to be a far more profitable choice than an effective false denial at trial. When perjury is avoided for reasons of expediency, not principle, it is debatable whether the defendant pleading guilty is a better prospect for reformation than one who perjures himself at trial in an unsuccessful effort to obtain acquittal. On the other hand, the defendant whose sentence has been increased on the basis of suspected perjury has, in the opinion of the court, actually perjured himself; his counterpart who pleaded guilty may have entertained the propensity to commit perjury but did not in fact do so. Thus in evaluating the character of the two defendants, it may sometimes be reasonable to give greater weight to the commission of perjury than to the mere propensity to commit perjury.”
It appears that all the cases which have discussed perjury and its *429effect on a sentence have been situations where the trial judge, in his infinite wisdom, has determined that the defendant perjured himself during his own trial. No cases have involved situations where the judge has made the same determination at someone else’s trial. Accordingly, the cases cited by the majority, and in this dissent, have involved alleged perjury committed during the defendant’s trial. However, I feel the reasons why the sentencing court should not consider perjury to augment a sentence are even more compelling in situations where the judge believes a defendant has committed perjury at another’s trial. Where a defendant has testified at his own trial and is convicted the rationale is that the jury’s verdict is an implicit finding that he was not telling the truth. But that same reasoning does not hold forth in a situation where the defendant has testified at another’s trial.
I suggest another reason why a judge should not consider perjury as a reason to augment a sentence. From practical experience, judges know that rare is the day if testimony is being heard all day, the perjury has not been committed, or at least that a witness has been more than a little careless with the truth. Like it or not, it is a fact of life. And yet judges, when they believe someone is not telling the truth, do not order a prosecuting attorney to file perjury charges. It is equally as rare that the trial judge holds a witness in contempt for not telling the truth. The fact is that the trial judge does nothing about this perjury. I suggest the only reason it is done in criminal cases, after a defendant is convicted, is the sentencing judge then has the hammer over the defendant’s head— and, at sentencing, the hammer falls. Even though he knows perjury could not be proven if a criminal charge were filed, and perhaps even a criminal contempt charge would not hold up, he then decides to consider perjury as a factor in the sentence. Surely, this is an uneven application of the law. But to so apply the law denies to defendant his constitutional protections as well as other defenses he may have either to a perjury charge or to a criminal contempt charge.
Using this kind of leverage against a defendant whom the sentencing judge believes has committed perjury can only be attribut*430ed to the arrogance of the court system where the trial court is permitted to play god with the life of a defendant. Nowhere should this kind of arbitrary conduct be tolerated by an appellate court.
The cases cited by the majority indicate in one way or another that it is permissible for a sentencing court to conside defendant’s candor. The apparent reason is that it reflects on a defendant’s capacity for reformation. However, almost all the cases cited are readily distinguishable from the situation here. Some do not state how many years, if any, were added because the trial judge believed the defendant committed perjury, or, more delicately, was not candid. In most of the situations the defendant was sentenced to far less than the maximum. Additionally, in most cases the sentence imposed does not indicate that the trial court placed undue emphasis on the defendant’s lack of candor. None of them came close to adding 15 years to a sentence because of a defendant’s lack of candor.
The first case cited by the majority, United States v. Lustig (9th Cir. 1977), 555 F.2d 737, was a situation where the defendant was convicted of distributing and conspiracy to distribute cocaine. The maximum penalty was not stated. After the verdict the prosecuting attorney commented that defendant may have perjured himself testifying in his defense. Defendant was given nine years in prison, but there is no indication of how much time was added, if any, for the defendant’s lack of candor. The appellate court simply ruled it was permissible for the sentencing court to consider the defendant’s candor on the witness stand at his trial.
In United States v. Cluchette (9th Cir. 1972), 465 F.2d 749, the defendant was convicted of passing counterfeit bills. The maximum sentence was 15 years on each count. The sentencing record showed that the sentencing judge believed defendant had been “less than candid” in his trial testimony. Defendant was sentenced to 18 months on each of two counts, to run concurrently. The appellate court stated that there was “no evidence that the sentencing judge was in any way biased or prejudiced”, and that the sentence imposed “forecloses any thought that the judge was unfair * * *465 F.2d 754.
*431In United States v. Sneath (8th Cir. 1977), 557 F.2d 149, the defendant was convicted of interstate transportation of stolen motor vehicles. The maximum sentence was' not stated. The presentencing investigation alleged defendant had lied to the FBI. The court sentenced defendant to 18 months in prison. Defendant contended the trial court relied on this presentence report stating he had lied to the FBI. The appellate court stated that he did not deny the allegations in the report and that they were apparently not a major factor in his sentence. The court concluded that “* * *. A sentencing judge is entitled to consider a wide range of information regarding defendant’s character and background in fixing sentence. * * *” 557 F.2d 151.
In United States v. Levine (7th Cir. 1967), 372 F.2d 70, the defendant was convicted of armed robbery and sentenced to the maximum of 25 years. Defendant contended on appeal that he received the maximum sentence because he took the witness stand. However, the appellate court stated that although the sentencing judge may have believed defendant perjured himself, and may have considered that (it did not knowAvhether he had), the record shows that the trial court considered many factors, several of which independently supported the sentence'.
In United States v. Wallace (6th Cir. 1969), 418 F.2d 876, the defendant was convicted of removing $3.00 in currency from a letter entrusted to him. Defendant was convicted in a nonjury trial, and at sentencing the judge commented that defendant’s guilt was proven to “a mathematical certainty”. 418 F.2d 877. The court also commented it would almost certainly have granted probation if defendant had come into court and admitted his guilt. Instead, the court imprisoned the defendant, the term not being mentioned in the opinion. It is not clear what issue the defendant raised on appeal, but the appellate court did hold it was proper to consider defendant’s testimony at trial in imposing sentence.
In United States v. Nunn (5th Cir. 1976), 525 F.2d 958, the defendant was convicted of transporting illegal aliens. Neither the maximum sentence nor the sentence actually given was disclosed. *432The appellate court stated, without explanation, that consideration of perjury in sentencing is permissible.
In United States v. Moore (4th Cir. 1973), 484 F.2d 1284, the defendant was convicted of receiving a stolen vehicle. He was sentenced to four years in prison. The maximum sentence was greater. In sentencing the judge considered that the defendant had “testified falsely” at this trial. The appellate court allowed the four year sentence to stand on the broad ground “* * * that a sentencing judge may consider evidence of crimes for which the defendant has not been convicted. * * *” 484 F.2d 1287. The court actually sidestepped the issue, but clearly stated its preference for the future when it stated:
“* * * It is better in the usual case for the trial judge who suspects perjury to request an investigation. Then, if the facts warrant it, the United States Attorney may institute prosecution for this separate and distinct crime.” 484 F.2d 1288.
In his dissent, Judge Craven was emphatic in stating that evidence of perjury should neve be considered as a factor in sentencing.
In United States v. Hendrix (2nd Cir. 1974), 505 F.2d 1233, the defendant was convicted of possession with intent to distribute cocaine and marijuana. The maximum penalty was not disclosed in the opinion. He was sentenced to ten years in prison, two years of which were attributable to the judge’s belief that defendant perjured himself. In affirming the sentence, the appellate court stated “* * * Impressions * * * are * * * central factors to be appraised; under our theory of ‘individualized’ sentencing. * * *” 505 F.2d 1236.
Clearly, these cases cannot withstand the scrutiny of closer analysis, and neither should they be the foundation for law in this state which gives carte blanche to sentencing judges to sentence for perjury, or the more delicate term of “lack of candor”.
There are a multitude of policy factors which militate against allowing a sentencing judge to impose additional punishment because of his belief that the defendant has committed perjury, either *433at his own trial, or at the trial of another. However, when it comes down to the core, I believe that Justice Craven, in his dissent in United States v. Moore, supra, best stated it:
“I suggest one more reason why a trial judge should never impose additional punishment because of his belief that a defendant lied in his own defense: he may be wrong. * * *” 484 F.2d 1288. (Emphasis added.)