Appellant was convicted of narcotics offenses in 1975 and thereafter filed a motion to vacate his sentence (28 U.S.C. § 2255) on the ground of ineffective assistance of counsel. He asserted that unknown to him his trial attorney had simultaneously been representing a DEA agent who had testified against him at the trial. The district judge denied relief and this court, relying on Cooper v. Fitzharris, 586 F.2d 1325 (9th Cir. 1978), affirmed judgment, finding an insufficient showing of prejudice. Brown v. United States, 605 F.2d 561 (9th Cir. 1979). The Supreme Court, 446 U.S. 2936, 100 S. Ct. 2936, 64 L. Ed. 2d 821, vacated the judgment and remanded for further consideration in the light of Cuyler v. Sullivan, 466 U.S. 335, 100 S. Ct. 1708, 64 L. Ed. 2d 333, 625 F.2d 210 (1980). We thereafter remanded the case to the district court to permit it to rule on the question.

On remand the district judge found explicitly that a conflict of interest did exist in the dual representation. But the district judge continued:

"However, the court cannot find, based on a thorough re-examination of the whole record in this case, that such conflict adversely affected defense counsel's performance. The government's incriminating evidence in this case was massive to the hundreds of exhibits. The defendant was awash in a sea of adverse evidence, which, in short overwhelmingly established this defendant's guilt. The court, then, specifically finds that the existing conflict of interest, under all the circumstances of this case, did not adversely affect counsel's performance, and the court so rules in response to the remand of the Court of Appeals."

Appellant has again appealed and says that the district judge did not use the correct test under Cuyler. We must agree. The weight of evidence against Brown (even such overwhelming evidence as the district judge found here) cannot, of itself, justify the denial of relief under § 2255 once there has been a finding of actual conflict of interest from cases (e. g. Cooper v. Fitzharris, supra ) where conflict of interest is not in issue:

"We read Sullivan (Cuyler v. Sullivan ) to define an actual, as opposed to a potential, conflict as one which in fact adversely affects the lawyer's performance. But the requirement that the petitioner show this adverse effect is not the same as the requirement of Cooper v. Fitzharris, 586 F.2d 1325 (9th Cir. 1978), cert. denied, 440 U.S. 974 (99 S. Ct. 1542, 59 L. Ed. 2d 793) (1979), that the petitioner show that counsel's incompetent assistance resulted in actual prejudice. But such evidence would be completely irrelevant to an inquiry whether the same error, if caused by an actual conflict of interest, showed an adverse effect on counsel's performance." (Emphasis supplied.)

The test, as stated in Cuyler, is whether the conflict of interest adversely affected the performance of Brown's attorney. This is a question that a trial judge, who has seen the attorney in action, is uniquely equipped to answer. We, therefore, remand the case again to permit the trial judge this opportunity, now with the added guidance that Hearst (U. S. v. Hearst, 638 F.2d 1190 (9th Cir. 1980)) provides.

Remand for proceedings not inconsistent herewith.

I fully concur in the above opinion. I write separately only to highlight the standards that should govern the district court's determination on remand. The district court concluded that the appellant's representation was not adversely affected because further cross-examination of the D.E.A. agent was highly unlikely to change the verdict against the appellant. Although this is the proper inquiry under the "actual prejudice" test of Cooper v. Fitzharris, 586 F.2d 1325 (9th Cir. 1978) (en banc), cert. denied, 440 U.S. 974, 99 S. Ct. 1542, 59 L. Ed. 2d 793 (1979), it is not the inquiry required by Cuyler v. Sullivan, 446 U.S. 335, 100 S. Ct. 1708, 64 L. Ed. 2d 333 (1980). To satisfy Cuyler, the district court must determine whether the appellant's attorney refrained from a more vigorous cross-examination of the D.E.A. agent because of his divided loyalties, and, if so, whether the appellant's representation would have benefited even marginally from a more aggressive cross-examination.