This is an appeal from an order of the Court of Common Pleas of Chester County dismissing a Petition under the Post Conviction Hearing Act after an evidentiary hearing. The appellant alleged inadequacy of counsel because of improper dual representation.

The proceedings arose out of an incident which occurred at an apartment complex on April 19, 1970, involving the appellant, Irvin Knight, Herbert Handy and John Mitchell. The appellant was represented in all proceedings, except in the instant appeal, by Attorney William McLaughlin who also represented Handy. Mitchell had separate counsel throughout the trial.

At the time of his arrest, Handy gave a written statement to the officers which incriminated himself and the other co-defendants. All were tried together and all were convicted. The appellant and Mitchell were convicted of robbery with an accomplice, assault and battery, larceny, rape and conspiracy; Handy was convicted of all these charges plus aggravated assault and battery.

*340At the trial, the Handy statement was not introduced. Handy did not take the stand but the appellant did. The appellant in his testimony incriminated Handy as to the aggravated assault and battery, of which Handy alone was subsequently convicted. The appellant’s testimony was to the effect that Handy was seen beating one of the victims with a belt buckle. The defense in both Handy’s and the appellant’s cases was consent to the entry on the premises and consent to the intercourse.

After the entry of the verdicts and the denial of post-trial motions, both appellant and Handy were represented by Attorney McLaughlin at sentencing. At the sentencing, Attorney McLaughlin said of Handy:

“. . . And I think because of his age and he seemed to be under the instructions of that older man that he should not be dealt with on the same footing with them.”

At the time of the offense appellant was 19 years old; Handy was 21 and Mitchell was 31. Both the appellant and Handy received the same sentence.

Both convictions, on appeal, were sustained by this Court per curiam. Commonwealth v. Handy, 225 Pa.Super. 721, 306 A.2d 340, (1973) and Commonwealth v. Knight, 225 Pa.Super. 729, 306 A.2d 340, (1973).

The sole question presented in this appeal is whether the dual representation by Attorney McLaughlin denied the appellant his constitutional right to effective counsel. In the light of Commonwealth v. Breaker, 456 Pa. 341, 318 A.2d 354 (1974), a new trial is indicated with counsel unhampered by a conflict of interest, potential or actual.

In Commonwealth v. Breaker, supra, the Supreme Court cited the various cases outlining the four (4) principles that control dual representation. The court stated at page 356, 318 A.2d at page 356:

“Our dual representation cases make several principles clear. First, ‘[i]f, in the representation of more *341than one defendant, a conflict of interest arises, the mere existence of such conflict vitiates the proceedings, even though no actual harm results. The potentiality that some harm may result, rather than that such harm did result, furnishes the appropriate criterion’ .. . [citations omitted] . . . Second, a defendant must demonstrate that a conflict of interest actually existed at trial, because ‘dual representation alone does not amount to a conflict of interest’ . [citations omitted] . . . Third, [t]o make the dual representation rise to a true conflict, appellant need not show that actual harm resulted . . . but he must at least show the possibility of harm’ . . . [citations omitted] . . . Fourth, appellant will satisfy the requirement of demonstrating possible harm, if he can show, inter alia, ‘that he had a defense inconsistent with that advanced by the other client, or that counsel neglected his case in order to give the other client a more spirited defense’ . . . [citations omitted] . . .”

The instant case is not as clearly defined factually as Breaker, supra, which involved the appellant being urged by his counsel to plead guilty while same counsel also represented the informant-co-defendant. However, the instant case falls within the four principles enumerated in Breaker, supra.

The co-defendants were represented by the same counsel and both relied on the defense of consent as to entry and consent to intercourse with the victim. Because the defenses are nominally identical as to both defendants does not mean they were similar in effectiveness for both defendants. The appellant, at trial, took the stand in his own defense and testified as to Handy’s assault with a belt buckle. Handy, for whatever reason, although urged to do so by his counsel, did not take the stand and his prior written statement incriminating the appellant was not introduced. However, the possibility of its introduction for purposes of cross-examination or *342impeachment, if he were to have testified, is a reality. Once the co-defendant to charges such as are involved here testified to violence to a person the effectiveness of the defense of consent is weakened. The impact this had on the jury may have been mitigated had Handy taken the stand and testified. Appellant’s defense was at least potentially eroded by Handy’s failure to testify. The defenses of both defendants rose and fell together. It was for such situations where the harm is incalculable that the prophylactic rule of Whitling v. Russell, 406 Pa. 45, 176 A.2d 641 (1962) and Breaker, supra, was devised. See also, Commonwealth v. Johnson, 223 Pa.Super. 307, 299 A.2d 367, 369 (1973). The potentiality that such harm may result rather than that such harm did result, furnishes the appropriate criterion. The defense of both co-defendants, although nominally consistent, were practically antagonistic.

Reference should also be made to the comment of counsel at sentencing as to Handy’s being led astray by the older man. In Breaker, supra, 456 Pa. at page 357, 318 A.2d at page 357, the Supreme Court pointed out:

“. . . Indeed, at sentencing the assistant district attorney informed the court that the police had communicated to him that Mangold ‘was instrumental in the breaking of the case, that he was extremely cooperative with the police and [that the police] wanted the court advised of that fact.’ ”

In footnote 5 on the same page the Court noted that stressing one co-defendant’s role as the informer was in fact the trial strategy decided upon by dual representing counsel. Although we cannot speculate as to counsel’s motives, the attitude of counsel at sentencing in favoring one client over the other poses a serious question about the intention of counsel during the conduct of the trial. Commonwealth v. Johnson, supra. Clearly if the Supreme Court was concerned as to comments of the Commonwealth in Commonwealth v. Breaker, supra, we *343should be concerned with the appellants’ own counsel commenting upon his co-defendant’s being led astray absent any redeeming factors or mitigating comments about the appellant.

As we said in Commonwealth v. Cullen, 216 Pa.Super. 23, 260 A.2d 818, 820 (1970):

“At sentencing, appellant’s trial counsel was under a duty to put each of his clients in the best possible light before the court. In the instant case, trial counsel felt compelled to picture the co-defendant as the less guilty party, the party that had been led. He could do no less for the co-defendant. But, by doing so, his duty to present appellant in the best possible light was neglected. He pictured the appellant as the party worthy of strict treatment and the co-defendant as the party worthy of lenient treatment.”

The order is reversed and a new trial granted.

concurring:

I think it clear that appellant was denied effective counsel at sentencing, for counsel did not fulfill his duty to put appellant before the court in the best possible light. Commonwealth v. Cullen, 216 Pa.Super. 23, 260 A.2d 818 (1970). Whether appellant was denied effective counsel at trial, however, seems to me a more difficult issue. While I agree with the majority that he was, I arrive at that conclusion by reasoning different from the majority’s.

The majority gives two reasons for its conclusion that trial counsel’s representation of both appellant and Handy involved the possibility of harm to appellant.

First, the majority notes that appellant as part of his testimony said Handy had assaulted the victim with a *344belt buckle. I agree that because of this testimony, “the effectiveness of [appellant’s] defense of consent [was] weakened.” Majority opinion at 342. That, however, does not show any harm to appellant arising from the fact that he and Handy were represented by the same attorney. Handy might argue that it did: Why should his attorney, by questioning appellant, bring out that he had assaulted the victim? Or, to put it differently: If Handy had had his own, separate, attorney, would not that attorney have cross-examined appellant on his testimony that Handy had assaulted the victim ? Handy, however, is not before us. So far as appellant is concerned, it only appears that his (and Handy’s) attorney asked appellant about a fact (Handy assaulting the victim) that hurt both appellant and Handy. This may show that the attorney was ineffective, but it does not show that his ineffectiveness arose from his representation of both appellant and Handy.

Second, the majority notes that Handy’s decision not to testify “at least potentially eroded” appellant’s defense. Majority opinion at p. 342. Again, I agree; the jury might have thought, despite the judge’s instruction to the contrary, that Handy’s decision not to testify implied his guilt; and from this the jury might have gone on to think that because Handy was guilty, so was appellant. However, this danger is present in every case involving two or more defendants where one defendant decides not to testify. The question here is whether Handy’s decision not to testify was motivated by advice given him by his attorney. If it was, reversal would be required, for since Handy’s attorney was also appellant’s, a conflict of interest would arise: in order to help his client Handy, the attorney might have to hurt his client appellant. Here, however, at the PCHA hearing the attorney testified that he had advised Handy that he should testify, and the hearing judge accepted this testimony.

*345Further in regard to Handy’s decision not to testify, the majority notes that Handy had made an incriminating statement, which, had he testified, might have been used to impeach him. I do not see how this fact helps appellant. The statement incriminated not only Handy but also appellant. Therefore, if by not testifying Handy was able to preclude introduction of the statement, he helped not only himself but also appellant.

Despite these difficulties, I have concluded that a new trial should be awarded. The fact that appellant testified that Handy had assaulted the victim shows — or at least strongly suggests — that appellant was hostile to Handy: he was trying to persuade the jury that it was not he but Handy who had acted criminally. This shows —or at least strongly suggests — that Handy was in turn hostile to appellant. (It is conceivable, I admit, that Handy said to himself that appellant was justified in being hostile to him, but it seems unlikely.) Thus it appears that one attorney represented two clients, each hostile to the other, which is a situation “ ‘at least show[ing] the possibility of harm.’ ” Commonwealth v. Breaker, 456 Pa. 341, 345, 318 A.2d 354, 356 (1974).

This appearance, moreover, gains substance from the fact that shortly before trial, the attorney wished to withdraw. He did not withdraw because, to quote the lower court, “the writer of this opinion indicated to him that he should remain in the case so as to avoid a further continuance of the trial, which would have been required if he had been permitted to withdraw.” Opinion of lower court at 2. It is regrettable that the lower court exerted such pressure on the attorney; the attorney was in a better position than was the court to appraise the possibility of harm arising from dual representation. It is regrettable for another reason. This matter comes before us as a Post Conviction Hearing Act proceeding. The hearing judge in that proceeding was the same judge as had been the trial judge. Thus, in ruling upon *346appellant’s post-conviction claim that there had been a conflict of interest, the judge was in the position of having already decided that there had not been a conflict. He had, in short, pre-judged the case. Ordinarily, unless abuse of discretion appears, we are bound by findings made in a PCHA proceeding. Commonwealth ex rel. Johnson v. Rundle, 440 Pa. 485, 270 A.2d 183 (1970). Given the circumstances here, I do not think we are bound. Thus we are left, as I see the case, with two facts: the attorney’s judgment that he ought to withdraw; and the apparent hostility between his clients. I think these are enough to require a new trial.