Cleophus Davis, Jr. appeals from the district court's1 denial of his 28 U.S.C. § 2255 motion, in which he sought to vacate his robbery and firearm convictions based on alleged violations of his Fifth and Sixth Amendment rights. We affirm.

The facts underlying Davis's convictions are detailed in his direct appeal, in which we affirmed his convictions. See United States v. Davis, 103 F.3d 660 (8th Cir. 1996), cert. denied, 520 U.S. 1258, 117 S. Ct. 2424, 138 L. Ed. 2d 187 (1997). We repeat the facts here to the extent they are relevant to Davis's § 2255 claims. Davis was convicted by a jury of three counts of armed bank robbery, 18 U.S.C. § 2113(a) and (d), and three counts of using a firearm during and in relation to a crime of violence, 18 U.S.C. § 924(c)(1). Some eyewitness testimony was offered at trial, and the prosecutor used a ballistics expert to link Davis to two of the three robberies.

Investigators recovered bullets from two of the robbery locations, and the prosecutor used that evidence, in part, to identify Davis as the robber by tying the two bullets to bullets later found in Davis's car and in a parking lot near his home. The prosecution's expert, John Riley, worked for the FBI Laboratory Division in its Elemental and Metals Analysis Unit, and had extensive experience analyzing the elemental composition of various materials, including the lead used in manufacturing bullets. Prior to trial, Riley performed a forensic analysis on the two bullets found at the scenes of the robberies as well as one spent bullet found in a parking lot near Davis's apartment and ten bullets recovered from a partial box of ammunition found in Davis's car. Riley testified that he measured the trace elements in each of the bullets using Inductively Coupled Plasma-Atomic Emission Spectrometry ("ICP") in an attempt to match particular bullets to each other. Based on the similarity among six trace elements found in the bullets tested, Riley opined that the bullets were "analytically indistinguishable," and that "the three expended bullets either came from this partial box of cartridges, made by Remington, or from another box of cartridges that was manufactured by Remington on the same day that this box was."

Prior to trial, Davis's attorney moved to exclude Riley's testimony on foundational grounds. The district court rejected Davis's motion to exclude the testimony, and during trial, Davis's attorney extensively cross-examined Riley about his methodology and his findings. Toward the end of trial, Davis's attorney hired an analytical chemist, Dr. David Dobberpuhl, as an expert to refute Riley's testimony. During his testimony, Dr. Dobberpuhl criticized Riley's use of the ICP analysis to support his opinion and challenged Riley's methodology and conclusions.

As noted, the government also offered some eyewitness testimony at trial, which the magistrate judge characterized as weak. Because the robber wore a mask, no one at the scene of any of the robberies got a clear view of the robber. One teller was unable to pick Davis out of a lineup, though she narrowed it down to two people in the lineup, one of whom was Davis. She then made a positive voice identification of Davis after hearing the men in the lineup repeat statements similar to what the robber had said to her during the robbery. A motorist, who noticed a man jogging away from the vicinity of one of the robberies shortly after it occurred, later identified Davis in a lineup and testified at trial that there was an 80-90% probability that Davis was the man he saw leaving the vicinity of the robbery.

After the district court denied Davis's motion for acquittal, a jury convicted Davis on all counts, and we affirmed the conviction on direct appeal. Davis then filed a § 2255 motion raising four issues: (1) ineffective assistance of trial counsel for failing to investigate or hire an appropriate expert witness to rebut the government's expert; (2) ineffective assistance of trial counsel for failing to interview any witnesses prior to trial; (3) ineffective assistance of appellate counsel for failing to raise the ineffective assistance of trial counsel issue on direct appeal; and (4) violation of the Fifth Amendment's Indictment Clause. The district court rejected the claims alleging ineffective assistance of appellate counsel and Indictment Clause violations, but held an evidentiary hearing on the two claims related to the alleged ineffective assistance of trial counsel.

Davis presented testimony at the § 2255 hearing challenging the testimony and conclusions of the prosecution's metals expert. He also presented evidence that his trial attorney had not interviewed the two eyewitnesses prior to trial, nor had he contacted the officers who conducted the lineups concerning the eye witnesses' level of certainty at the time of the lineups. Having considered the additional testimony, the magistrate judge recounted trial counsel's cross-examination of Agent Riley at trial, as well as Dr. Dobberpuhl's testimony that refuted some of Agent Riley's opinions. The magistrate judge concluded that "defense counsel forcefully challenged Agent Riley's conclusions ... [and did not] blunder by retaining an analytical chemist rather than a metallurgist." (Report and Recommendation at 19-20.) The magistrate judge also concluded that the results of the trial would not have been different if Davis's trial attorney had hired a different expert. (Id. at 20.) The magistrate judge further noted that the concerns raised about the FBI's bullet analysis by the § 2255 expert, who discredited the methodology used by the FBI, were not widely disseminated within the scientific community in 1995 at the time of Davis's trial. The magistrate judge recommended to the district court that Davis's § 2255 motion be denied.

The district court agreed with the magistrate judge's thorough analysis, adopted the Report and Recommendation, and denied Davis's § 2255 motion. The district court granted Davis a certificate of appealability on all four issues raised in the § 2255 motion.

We review the legal issues raised by the district court's denial of Davis's § 2255 motion de novo, but any underlying fact-findings are reviewed for clear error. See Covey v. United States, 377 F.3d 903, 906 (8th Cir.2004). Ineffective assistance of counsel claims raise mixed questions of law and fact, and we accordingly review those claims de novo. Id.

To prevail on an ineffective assistance of counsel claim, Davis must satisfy the familiar Strickland standard by establishing both that his counsel's performance "fell below an objective standard of reasonableness," and that the deficient performance prejudiced his defense. See Strickland v. Washington, 466 U.S. 668, 688, 691, 104 S. Ct. 2052, 80 L. Ed. 2d 674 (1984); United States v. Robinson, 301 F.3d 923, 925 (8th Cir.2002), cert. denied, 537 U.S. 1238, 123 S. Ct. 1367, 155 L. Ed. 2d 207 (2003). "In determining whether counsel's conduct was objectively reasonable, there is a `strong presumption that counsel's conduct falls within the wide range of reasonable professional assistance.'" Nguyen v. United States, 114 F.3d 699, 704 (8th Cir.1997) (quoting Strickland, 466 U.S. at 689, 104 S. Ct. 2052). To satisfy the prejudice prong of the Strickland analysis, Davis must establish a "reasonable probability that, but for counsel's unprofessional errors, the result of the proceeding would have been different." Strickland, 466 U.S. at 694, 104 S. Ct. 2052.

Davis raises two instances of allegedly ineffective assistance of trial counsel: his counsel's handling of the government's expert witness including his failure to obtain a competent rebuttal expert, and his counsel's failure to interview witnesses prior to trial. Following a hearing on both of these issues, the district court adopted the magistrate judge's Report and Recommendation, finding that Davis's counsel was not deficient on the expert witness issue, and that even if his counsel's failure to interview witnesses equated to deficient performance, such deficiency did not prejudice Davis's defense.

At the § 2255 hearing, Davis offered the testimony of Dr. Erik Randich, a metallurgist, who challenged the FBI's alleged ability to conclusively identify bullets as coming from a specific box of bullets based on the bullets' composition. The magistrate judge determined that the information relied upon by Dr. Randich to challenge the FBI's use of ICP to support its bullet-identification theory was not readily available in 1995, at the time of Davis's trial. Davis has failed to establish that this factual finding was clearly erroneous. Dr. Randich did not begin his own study challenging the FBI's ballistics analysis until 1998, three years after Davis's trial. Dr. Randich admitted that he was unaware of any study or report in 1995 that was critical of the FBI's methodology. Davis's trial counsel cannot be said to be ineffective for failing to challenge the FBI's methodology on a basis that was not advanced by the scientific community at the time of trial. Cf. Dees v. Caspiri, 904 F.2d 452, 455 (8th Cir.) ("At most counsel might be deficient for failing to discover and make use of a widely held scientific view." (emphasis added)), cert. denied, 498 U.S. 970, 111 S. Ct. 436, 112 L. Ed. 2d 419 (1990).

Davis alleges that his trial counsel's investigation of the case was constitutionally deficient because he did not hire Dr. Dobberpuhl until after the trial had started, giving the witness little time to prepare, and that his counsel should have hired a metallurgist instead of an analytical chemist. Again, Davis offers Dr. Randich's testimony as an example of the type of expert his trial counsel should have hired. Dr. Randich took issue with Riley's conclusion that if the composition of two bullets are found to be "analytically indistinguishable" they must have come from the same source or pour of lead. He also took issue with the underinclusiveness of Riley's supporting data.

Contrary to Davis's assertions, Dr. Dobberpuhl provided similar testimony. He challenged Riley's conclusion that a finding that bullets were analytically indistinguishable necessarily means that the bullets came from the same pour of lead. Dr. Dobberpuhl also challenged the underinclusiveness of Riley's sample, pointing out that Riley did not test the composition of other bullets available in the community to see if he could find other analytically indistinguishable bullets. Dr. Dobberpuhl further disagreed with Riley's conclusion that the bullets were in fact analytically indistinguishable, pointing out that six of the ten bullets from the partial box did not have the same levels of copper. Nor has Davis established that the district court clearly erred in finding that Davis's trial counsel effectively and vigorously cross-examined Riley during trial. As explained in the magistrate judge's detailed Report and Recommendation, Davis's trial counsel researched ballistics analysis and became familiar with the scientific evidence, such that he was able to raise many of the same issues addressed by Dr. Randich with Riley on cross-examination. We agree with the district court that Dr. Dobberpuhl was an effective expert witness, and that Davis's counsel acted reasonably in employing him.

Davis also alleges that his trial counsel's performance was substandard when he failed to interview either the eyewitnesses or the officers involved in the pretrial identification procedures prior to trial. However, Davis's counsel reviewed all notes of interviews with the eyewitnesses, the officers' reports of the lineups, and the sketch artist's notes and sketches, and used the information in an attempt to suppress the witnesses' in-court identification prior to trial. Davis's counsel also used inconsistencies contained in the police interview notes during cross-examination in an attempt to impeach the witnesses, with some success. Both eyewitnesses testified on cross-examination that their pretrial identifications were less certain than their in-court identifications. Davis's counsel got one witness to concede that her description of the robber's clothes differed from the video tape of the robbery. In short, Davis's counsel elicited basically the same type of testimony that Davis says he would have been able to use had he interviewed the eyewitnesses prior to trial. Davis has failed to establish prejudice, that is, a reasonable probability that had his trial counsel interviewed the eyewitnesses prior to trial, the outcome of his trial would have been different. See Payne v. United States, 78 F.3d 343, 348 (8th Cir. 1996) (rejecting ineffective assistance claim where, even if counsel's performance was deficient for failing to interview or investigate witnesses, defendant failed to establish that interviewing witnesses would have changed the outcome).

Having carefully reviewed the entire record, we agree with the magistrate judge's careful and detailed analysis regarding the sufficiency of Davis's trial counsel's performance. Hindsight can often reveal a better tactical line of questioning, and time can often uncover a more effective expert witness, especially as the body of scientific knowledge grows. The Constitution does not guarantee the best representation, especially when measured through the lens of hindsight, but rather effective representation. We cannot say on this record that Davis's trial attorney provided him with constitutionally ineffective representation.

B. Ineffective Assistance of Appellate Counsel

Davis argues that his direct appeal counsel (who was a different lawyer than his trial counsel and a different lawyer than his current postconviction counsel) was constitutionally ineffective for failing to raise his trial counsel's ineffective assistance as an argument on direct appeal. As Davis recognizes, ineffective assistance claims are generally not raised on direct appeal. In fact, we often refuse to address such claims if they are raised, directing appellants to raise the issue in a postconviction proceeding. See United States v. Pherigo, 327 F.3d 690, 696 (8th Cir.) ("We decline Piercefield's invitation and state our oft repeated refrain-claims of ineffective assistance of counsel are best evaluated on facts developed outside of the record on direct appeal and are properly raised in a post-conviction motion under 28 U.S.C. § 2255 and not on direct appeal."), cert. denied, 539 U.S. 969, 123 S. Ct. 2663, 156 L. Ed. 2d 674 (2003). Davis suggests that the record was sufficiently developed at the time of the direct appeal to have allowed such a claim to be reviewed. Davis's reliance in this § 2255 proceeding on the expert testimony elicited at the postconviction hearing belies his position. We would not have had the benefit of Dr. Randich's postconviction testimony refuting the bullet analysis used by the FBI had we considered an ineffective assistance of counsel claim as part of the direct appeal. Because we have concluded that trial counsel was not constitutionally ineffective (using the same test and a better developed record than would have been used on direct appeal), there would have been no viable claim for appellate counsel to raise on direct appeal concerning trial counsel's performance. Davis cannot meet either Strickland prong, and his claim is rejected.

Bank robbery is a federal crime only if the institution robbed is a federally chartered financial institution or is insured by a federal deposit insurer. See 18 U.S.C. § 2113(f)-(h) (defining "bank," "credit union," and "savings and loan"). The indictment returned by the grand jury charged that one of the banks robbed by Davis was insured by the National Credit Union Administration (NCUA), when in actuality it was insured by the Federal Deposit Insurance Corporation (FDIC). At trial, the government established that the bank was insured by FDIC, and the jury instructions and verdict form conformed to that evidence. On direct appeal, Davis challenged his conviction on the basis that the indictment charged a different jurisdictional element (NCUA-insured status) than that proved at trial (FDIC-insured status). We rejected Davis's argument, holding that "what occurred in the present case was not a jurisdictional defect but a mere variance between the indictment and the proof, which was not prejudicial to the defendant and therefore is not fatal to his conviction." Davis, 103 F.3d at 675.

Davis attempts to resurrect the issue by arguing that the jury instructions constituted a constructive amendment of the indictment in violation of the Fifth Amendment. (See Appellant's Br. at 52-55, quoting Stirone v. United States, 361 U.S. 212, 215-16, 80 S. Ct. 270, 4 L. Ed. 2d 252 (1960), for the proposition that "after an indictment has been returned its charges may not be broadened through amendment except by the grand jury itself.") As noted, we rejected Davis's argument on direct appeal that the divergence between the indictment and the jury instructions was a constructive amendment as opposed to a mere variance in the proof. We find the present claim to be the same claim in different clothes, and we will not reconsider the issue on collateral review. See Dall v. United States, 957 F.2d 571, 572 (8th Cir.1992) ("Claims which were raised and decided on direct appeal cannot be relitigated on a motion to vacate pursuant to 28 U.S.C. § 2255." (quoted source and internal marks omitted)).

The district court's judgment denying Davis's § 2255 motion is affirmed.