This case involves a procedurally barred habeas petition. The issue is whether the petitioner has presented evidence of actual innocence sufficiently strong to avoid the procedural barriers to raising an issue that would otherwise be procedurally barred.

In 1986, Sistrunk was convicted of forcibly raping a little girl. His jury trial took place in an Oregon state court, and he was sentenced to thirty years of imprisonment with a fifteen year minimum. He appealed to the Oregon Court of Appeals and petitioned for review to the Oregon Supreme Court without success.1 He then petitioned for post-conviction relief in state court. Both that petition and his appeal to the Oregon Court of Appeals were denied,2 but he did not seek review by the Oregon Supreme Court of his failed petition. He then filed a second state post-conviction relief proceeding with unfavorable results, lost on his appeal to the Oregon Court of Appeals, and unsuccessfully petitioned for review to the Oregon Supreme Court.3 Finally, he filed this federal petition for a writ of habeas corpus, claiming ineffective assistance of counsel, insufficiency of evidence, and prosecutorial misconduct. He conceded procedural default in the district court, but argued that the default should be excused because he is actually innocent, and failure to consider his claims would result in a fundamental miscarriage of justice. This district court carefully considered his evidence, and concluded that it was not so strong as to get petitioner through the " Schlup gateway," as is necessary for the court to consider his claims. We agree and affirm the district court's denial of the petition.

The eleven year old victim knew the petitioner, enough so that when she saw him the day of the rape, shortly after getting out of school, she ran away from him. He had assaulted her once before, and she testified that she was scared of him, because of the assault, and because he had threatened to kill her family if she told. She ran back to her school, but it was locked, and he caught her. He grabbed her by the arm and walked her over a freeway and into an open garage. He laid his coat on a large, square pan on the floor, pulled down his pants and hers, and forced her on top of him, penetrating her painfully, and ejaculated. Then he slapped her in the face with a five dollar bill, took her back to school, and warned her that he would kill her and her family if she told. She got an after-school activity bus that took her home. She bought flowers with the five dollars and took them home to her mother. Her mother realized that something was not right, got the girl to tell her what was wrong, and called the police and took the girl to the hospital. The girl had a fresh abrasion in her vaginal area. Subsequently, she began to have a burning sensation when she urinated and some discharge.

Sistrunk claims ineffective assistance of counsel because his trial attorney did not object to impermissible expert witness testimony, and his appellate attorney did not raise issues of failure to give a lesser included offense instruction on first degree sex abuse. He also claims that the trial court erred in denying a defense motion to photograph Sistrunk's penis and show that photograph to the jury. Sistrunk argues that his newly presented evidence demonstrates that the state's expert witness was biased in favor of the prosecution and improperly vouched for the victim's testimony by testifying falsely that scientific evidence proved that young children never lie about sexual molestation. Moreover, he argues that his penis did not have the distinctive appearance described by the victim and that introduction of the photograph would impeach her testimony.

The state's main expert witness, Dr. Jan Bays, testified falsely. She testified that a scientific study proved that "it is very, very rare that a child lies about sex abuse " and that the chance of such a lie is only with teenagers, "never with the younger children." She testified that the study established that "if the child comes forward with the story, themselves [sic], then it is the truth. If the child is younger than a teenager, then it is the truth."

The "scientific study" to which Dr. Bays referred had not been published at the time of trial. Because the study was not published at the time of trial and because it contradicts Dr. Bays's testimony, petitioner claims that it is newly available evidence. The article in question is in the Journal of Interpersonal Violence,4 and does not say what Dr. Bays testified that it said, nor is it a scientific study establishing anything at all. The article describes reports of suspected child abuse to the Denver Department of Social Services in 1983, and says that eight fictitious reports were made by children, but as for the accuracy of the rest, says that "[w]e do not have an absolutely reliable test"5 and "the definition of fictitious used in this study was that [social work] professionals did not consider that the abuse had occurred. This is subject to error."6 The fictitious cases cited in the report established the falsehood of Dr. Bays's testimony that "children never lie. " The study expressly said that "children of all ages made false allegations,"7 and the report itself "suggest[s] that the results be used as a basis for further study and not as a definitive basis for proving that a case is or is not `true.' "8 There is no evidence in the record suggesting that the prosecutor knew that Dr. Bays was testifying falsely.

After the trial, Dr. Bays also co-authored an article in a law review, suggesting that "clinical experience and systematic studies confirm that deliberately false allegations of sexual abuse are infrequent."9 Sistrunk says that this amounts to newly discovered evidence that Dr. Bays was biased.

The issue of whether a photograph of Sistrunk's penis should have been taken and submitted to the jury arises from a condition he has called neurofibromatosis. It causes pigmented spots and bumps on various areas of the patient's skin.10 The victim said that Sistrunk's penis "had bumps on it." This may have been inaccurate, because although Sistrunk had neurofibromas on his chest, he apparently had none on his penis, or it may have been an accurate description of how the girl perceived the anatomy of a penis, i.e., "bumpy," having not been previously familiar with the organ. Sistrunk's attorney wanted to have a photograph taken of the penis at the jail, in an erect state, with Sistrunk lying on his back in the position the girl described him, to show the jury that it had no bumps, but the judge would not permit it. However, the judge did allow Sistrunk's girlfriend to testify that he had no bumps on his penis. The claimed newly discovered evidence is a subsequent medical examination reporting that no neurofibromas were found on Sistrunk's genitals.

The district court denied the petition, on the grounds that Sistrunk procedurally defaulted on his claims and has not made a showing of actual innocence sufficient to raise them despite the procedural default. The magistrate judge, despite recommending denial of the petition, expressed her concern about the case, particularly because of Dr. Bays's false testimony vouching for the victim's credibility.

The case did not turn entirely on the victim's testimony. There was corroboration. A medical examination found abrasion of her vaginal area. Her subsequent discharge and pain on urination were consistent with rape. The girl's description of the garage where Sistrunk took her and raped her was consistent with the garage of a person in the location. The garage door did not lock, and the homeowner often left it open. There was a dry, square, oil pan with leaves on the floor, in the place where the girl had described a large pan onto which Sistrunk tossed his coat. The garage had a coiled utility cord hanging where the girl described a "rope" on the wall. The unusual event of an eleven year old girl going to a flower shop and buying her mother five dollars worth of flowers on no special occasion makes the flowers chilling corroboration of the girl's testimony that Sistrunk slapped her in the face with a five dollar bill he left her after raping her.

We review denial of the writ de novo and may affirm on any ground supported by the record.11 This petition was filed after the amendment of the habeas statute by the enactment of the Antiterrorism and Effective Death Penalty Act (AEDPA).12

Petitioner concedes that his claim would be procedurally barred but for the "Schlup gateway. " The Supreme Court held in Schlup v. Delo13 that if a petitioner otherwise procedurally barred "presents evidence of innocence so strong that a court cannot have confidence in the outcome of the trial unless the court is also satisfied that the trial was free of non-harmless constitutional error, the petitioner should be allowed to pass through the gateway and argue the merits of his underlying claims."14 The standard for actual innocence for purposes of the "Schlup gateway" is lower than the Herrera v. Collins15 standard applicable where the petitioner had a fair trial but is nevertheless actually innocent. To get through the Schlup gateway so that he can argue his otherwise procedurally barred constitutional claims, the petitioner must show that "it is more likely than not that no reasonable juror would have convicted him in light of the new evidence."16 Under this standard, "the newly presented evidence may indeed call into question the credibility of the witnesses presented at trial."17 We held in Carriger v. Stewart18 that where the post-conviction evidence "casts a vast shadow of doubt over the reliability" of the conviction but only by undercutting the reliability of the proof of guilt, not by affirmatively proving innocence, that can be enough to get through the Schlup gateway and allow consideration of otherwise barred constitutional claims,19 despite being insufficient for a "freestanding" Herrera claim of innocence for a trial free of prejudicial constitutional error.20 The AEDPA now provides that a claim that is procedurally barred because the petitioner "failed to develop the factual basis of [the] claim in State court proceedings must be dismissed unless "the facts underlying the claim would be sufficient to establish by clear and convincing evidence that but for constitutional error, no reasonable fact finder would have found the applicant guilty of the underlying offense."21

The briefs do not discuss whether the elaborate Schlup analysis applies to cases where the death penalty is not at stake, or whether the AEDPA standard supplants the Schlup standard. Both sides argue on the assumption that the Schlup standard applies. Because we ultimately decide that the requirements for the Schlup gateway are not met, we assume without deciding that Schlup applies,22 and analyze the facts of this case in terms of the Schlup gateway. We would reach the same result under the AEDPA language.

Sistrunk's first argument is that the district court erred by requiring that "new evidence" for purposes of the Schlup gateway be newly available rather than newly presented. We need not decide whether he is correct,23 because, as we explain below, he is not entitled to relief even with consideration of the new evidence.

We also need not decide on the strength of Sistrunk's legal claims, including his claim that the jury should have been instructed on a lesser included offense. The issue in this case is whether Sistrunk's evidence of actual innocence is strong enough to get him through the Schlup gateway so that his claims can be considered, which is a question independent of the legal force of his arguments about the claimed errors. We assume without deciding for purposes of analysis that Sistrunk could show that nonharmless constitutional error tainted the trial, with respect to the expert's vouching and falsification, and with respect to the denial of the opportunity to show the jury a photograph demonstrating that his penis was not as the victim described it.

Sistrunk's evidence does not purport to show that he did not rape the little girl, just that Dr. Bays should not have been allowed to vouch for the girl's testimony, and that the girl should have been impeached by a photograph that would have shown that his penis was not bumpy, as the girl had testified. Thus the evidence that Sistrunk seeks to introduce as proof of his innocence is merely impeachment evidence. Where the new evidence "casts a vast shadow of doubt over the reliability"24 of the conviction such that "it is more likely than not that no reasonable juror would have found petitioner guilty beyond a reasonable doubt,"25 it gets a petitioner through the Schlup gateway even if it merely impeaches credibility of prosecution evidence rather than affirmatively establishing actual innocence.26 We therefore consider whether the subsequently published journal articles showing that Dr. Bays testified falsely and was biased in favor of the prosecution, and the medical evidence that Sistrunk's penis had no neurofibromas on it, make it "more likely than not that no reasonable juror would have found petitioner guilty beyond a reasonable doubt."27

In this case there is not "a reasonable probability that the outcome of [petitioner's] trial would have been different" had the jury been furnished with the material undermining the expert's credibility and showing what Sistrunk's penis looked like.28 The facts do not "establish by clear and convincing evidence that but for [presumed] constitutional error, no reasonable fact finder would have found the applicant guilty of the underlying offense."29

The victim's account, not the expert's testimony, was most central to establishing guilt. The victim's account was plausible, internally consistent, and well corroborated. Her accurate description of the garage, the unlikely expenditure of $5 by an eleven year old for flowers for her mother when no occasion called for flowers, her prompt disclosure of the rape, the abrasion found at the hospital on the victim's vaginal area, and petitioner's implausible responses all contributed to the credibility of her account.

Likewise, evidence that Sistrunk's penis was free of neurofibromas does not necessarily impeach the victim's credibility regarding the rape, despite her description of Sistrunk as having a "bumpy" penis. First,"bumpy" does not necessarily refer to neurofibromas but, rather, could reflect an initial reaction to the appearance of male genitalia. Second, an honest and reasonably observant little girl, being raped by a man with neurofibromas conspicuously spotting his body might not notice that they are absent from his penis.

It is one thing to hold, as we did in Carriger, that new evidence that undermines the credibility of the prosecution's case may alone suffice to get an otherwise barred petitioner through the Schlup gateway. It would be quite another to hold, as we did not in Carriger, that such evidence necessarily must get a petitioner through the Schlup gateway. Whether it does or does not depends on whether the evidence is such that it is "more likely than not that no reasonable juror would have found petitioner guilty beyond a reasonable doubt."30 In this case, it does not. Unlike the new evidence in Carriger, which tended to show that the critical witness against Carriger had himself committed the murder and was an utterly unreliable witness, the new evidence of Dr. Bays's false testimony and bias, and of the non-bumpy condition of Sistrunk's penis, does not cast the "vast shadow of doubt over the reliability of his conviction."31 Nor does the new evidence show that "the facts underlying the claim would be sufficient to establish by clear and convincing evidence that but for constitutional error, no reasonable fact finder would have found the applicant guilty of the underlying offense."32 Dr. Bays's testimony was disgraceful, but it was icing on the cake, not critical to the persuasive force of the remaining evidence. Similarly, a photograph of Sistrunk's penis would not have had much persuasive force with respect to whether the girl was testifying falsely about the rape.

Sistrunk argues that he should pass through the Schlup gateway under our decision in Mach v. Stewart .33 The case is not in point, because it involves a permissible, not a procedurally barred, habeas claim, so we had no occasion to consider the Schlup gateway. Also, the question in Mach was whether a jury was tainted because the judge's questioning of the first venireperson, a social worker who emphatically insisted that children never lie about sexual assault against them, which is a different question from whether the impeachment of a witness saying that so undermines the reliability of the trial result as to get Sistrunk through the Schlup gateway. The case at bar is more like Wood v. Hall.34 Wood holds that a medical report improperly hidden from the defense, though it "could have supported reasonable doubt," did not so undermine reliability of the trial result as to get the petitioner through the Schlup gateway. As in Wood, the petitioner has presented insufficient evidence of actual innocence to allow consideration of his procedurally barred claims.

Sistrunk has failed to demonstrate that he can pass through the Schlup gateway to allow consideration of his procedurally barred claims.

AFFIRMED.

Emanuel Sistrunk has served almost fifteen years for a crime he probably did not commit. Sistrunk was convicted of raping Jane Roe,* an eleven-year-old girl. Roe testified that the crime occurred in a garage owned by a third party unconnected to either Sistrunk or herself. No one else saw Sistrunk with the victim, or anywhere near the garage on the day of the crime. The prosecution presented no semen, blood, hair, DNA or fibers connecting Sistrunk to the crime. A trench coat allegedly worn by the perpetrator, and used by him during the crime, was never recovered. The prosecution's entire case hinged on Roe's testimony and her identification of Sistrunk as the rapist.

The testimony of a victim--particularly a very young one--is a highly tenuous ground on which to rest a conviction. A jury might develop a reasonable doubt from the total absence of corroborating evidence. If the jury nevertheless convicts, we are bound by that determination. However, when the state's case is so heavily dependent on a single witness, errors affecting the witness's credibility take on enormous significance. Here, there is strong reason to believe that the jury's decision to believe Roe beyond a reasonable doubt was heavily influenced by the false testimony of a prosecution expert. Moreover, the trial court improperly denied defendant the opportunity to present evidence that would have undermined the victim's credibility. I don't share the majority's confidence that the jury would have convicted anyway; no reasonable jury would have convicted defendant but for these serious errors.

The expert witness issue is clear-cut and dispositive. Admission of the expert's testimony was highly questionable to begin with. The expert had nothing to say about the crime, nor about any connection Sistrunk might have had to it. Her only function was to lend credence to Roe's testimony. To that end, Dr. Jan Bays** testified about a supposedly scientific study she had conducted which--she claimed--showed that "it is very, very rare that a child lies about sex abuse" and "never with the younger children." Supplanting the jury's fact-finding responsibilities by anointing the prosecution's star witness with the aura of scientific infallibility is highly suspect. See State v. Middleton, 657 P.2d 1215, 1219 (Or. 1983). However, Sistrunk was aware of this issue at the time of trial, and it therefore can't form the basis for passage through the Schlup gateway. See Schlup v. Delo, 513 U.S. 298 (1995).

What is new is the discovery that the expert lied. As the majority bluntly puts it, the "study" in question "does not say what Dr. Bays testified that it said, nor is it a scientific study establishing anything at all." Maj. Op. at 1177. In other words, the expert fabricated the supposedly scientific proof on which she relied in persuading the jury that the prosecution's witness was being truthful. We have here not merely improper vouching, not merely supplanting of the jury by an expert, but doing so by means of perjury. Because the prosecution's case against Sistrunk depended on having the jury believe that the complaining witness was both accurate and truthful, I cannot agree that Dr. Bays's testimony was merely "icing on the cake." Id. at 1181. Jurors seldom have experience with sexually abused children. If an expert testifies, based on an allegedly scientific study, that such children never lie, jurors would be hard put to reach the opposite conclusion.

The problem of the expert's false vouching was compounded when the trial judge improperly prevented petitioner from gathering and introducing evidence that would have undermined the complaining witness's credibility. Roe testified that her assailant had a bumpy penis. As the jury could easily observe, petitioner suffers from neurofibromatosis, which creates bumps or lumps in his skin. The most plausible interpretation of Roe's statement is that she saw her assailant's penis, and that it had the types of bumps the jury could plainly see on visible portions of defendant's body. The clear inference the jury could draw is that petitioner was the assailant because he must have a bumpy penis.

Petitioner sought to introduce a photograph of his erect penis showing that it has no bumps. This would have been strong graphic evidence that the complaining witness was confused or untruthful. The trial court refused to allow defense counsel to take such a photograph, describing the necessary arrangements as "cumbersome" and the evidence as "misleading and confusing." Because the evidence would have been highly relevant and easily procured, I can imagine nothing except the trial judge's personal prissiness that could have motivated his decision to prohibit it. But when a man's life is at stake--and to an adult man, a thirty-year sentence pretty much is life--the judge's personal distaste for the evidence cannot be a sufficient reason to prevent the jury from seeing it.

New evidence, consisting of a medical examination performed after trial, discloses "no fibromas of [Sistrunk's] external genitalia"; in other words, his penis isn't bumpy, just as the photograph would have shown. The substitute of allowing petitioner's girlfriend to testify as to the appearance of his penis was plainly inadequate. The jury could have inferred that the girlfriend was biased in his favor and would tailor her testimony accordingly; they may have wondered why petitioner did not present a photograph or the result of a medical exam. Allowing petitioner to present the photograph would have forced the jury to consider whether the young witness had guessed that petitioner must have bumps on his penis and embellished her story to make herself more credible.

The supposedly corroborating evidence that my colleagues find so persuasive does nothing to cure these problems affecting the credibility of the prosecution's principal witness. Evidence that the witness had suffered some trauma to her vagina merely proves that someone may have raped her; in no way does it confirm her identification of petitioner as the assailant. The same is true of her description of the garage and the bizarre story of the five dollars petitioner supposedly gave her after the rape. It may be that whatever sexual assault Roe suffered happened in the garage in question, but nothing of petitioner's was found in or near the garage, so her knowledge of the garage does nothing to tie petitioner to her story. The five dollar payment may or may not support the claim of rape, but it does nothing to point the finger at petitioner any more than at anyone else who might have had five dollars in his pocket on the day in question.

As the majority recognizes, we held in Carriger v. Stewart, 132 F.3d 463, 477-78 (9th Cir. 1997), that a petitioner can pass through the Schlup gateway with evidence that undermines a key witness's testimony. The prosecution's case against Carriger was much stronger than that against Sistrunk. For one thing, there was no doubt that a crime had been committed: The victim was found dead, tied up, bludgeoned and strangled. Here, by contrast, it's unclear that a rape occurred, as no semen or pubic hair were found on the victim's body; Roe may have injured herself some other way. In Carriger, there was strong circumstantial proof that petitioner had participated in the crime, because his fingerprint was found on the adhesive tape that bound the victim's body. Here, there is nothing at all, other than the word of the victim, to connect petitioner to whatever crime did occur. In Carriger, the en banc court held that impeachment of the principal witness's testimony was sufficient to satisfy Schlup. Carriger, 132 F.3d at 478. This must be doubly true here, where the witness's testimony was the whole of the government's case against petitioner, and the so-called expert who vouched for that testimony lied through her teeth. If petitioner cannot pass through the Schlup gateway, there is no gateway.

We have here a miscarriage of justice. No reasonable jury would have convicted petitioner in a trial free of the serious errors affecting the complaining witness's credibility. Because I cannot join my colleagues in their contrary conclusion, I respectfully dissent.