Richard Hall, a member of the Three Affiliated Tribes of North Dakota, was convicted of the involuntary manslaughter of eleven-month-old Tory Big Crow in Indian country, in violation of 18 U.S.C. § 1112(b). The trial court1 set aside the verdict and entered a judgment of acquittal on the ground that the evidence, being equally consistent with the guilt of defendant's unindicted wife, was insufficient to support the verdict. The government appeals and we affirm.

The government produced testimony from no fewer than seven physicians, and a number of other health-care providers, that tended to prove that Tory Big Crow was the victim of a vicious assault that resulted in his death. The witnesses provided impressive proof that Tory had suffered an extensive injury due to a forcible disturbance of his brain. The doctors described Tory's brain as exhibiting cerebral edema or severe swelling; several of them stated that his brain was scrambled or liquified. There was further testimony that these injuries resulted from a trauma produced by an externally-applied massive force or, as one witness put it, by some type of acceleration event. One physician testified that the injury was so extensive that a force equivalent to that generated by a car crashing at a speed of 30 to 40 miles per hour, or a fall from a three-story building, would have been necessary to cause it. There was very general agreement by the experts that it was impossible for Tory's injuries to have been produced by a fall from a bed, which is what the defendant, who did not testify, attributed them to in conversations with a number of people, including an F.B.I. agent who testified at the trial. We offer the observation that common experience more than confirms this last conclusion.

This evidence, of course, amply supports a finding beyond a reasonable doubt that Tory Big Crow was battered to death. Indeed, as the trial court noted, such a conclusion is virtually irresistible. The difficulty that the trial court saw, however, was that all the available evidence was to the effect that Jamie Hall, the defendant's wife and Tory's mother, was in the house with defendant when the trauma was inflicted. Both defendant and Jamie Hall told absurd (but different) stories about how the injury had occurred. Jamie did not testify nor was she present in the courtroom during the trial. The trial court, in its opinion accompanying the order setting aside the verdict, stated that "[b]oth were physically capable of inflicting the injury," basing this observation, no doubt, on Dr. Wilbur Smith's testimony that any adult of normal stature and strength could have committed the crime. There was thus no evidence, the trial judge concluded, from which a reasonable mind could find that either the defendant or his wife was guilty beyond a reasonable doubt.

We are constrained to agree with this conclusion. If the state of the evidence were such that it would support a finding that the two possible perpetrators were jointly responsible, then this verdict could stand. The evidence is certainly consistent with that conclusion, but that, of course, is not enough. We think that a conscientious mind would have to have entertained a reasonable doubt about whether the defendant and his wife acted in concert, if for no other reason than that physical proximity to the commission of a crime is simply insufficient to establish responsibility for it. The government argued in its brief that the proof was sufficient to establish that both defendant and his wife were guilty; but at oral argument the government conceded the point, as we believe it had to.

The government calls our attention to the case of United States v. Boise, 916 F.2d 497 (9th Cir.1990), cert. denied, --- U.S. ----, 111 S. Ct. 2057, 114 L. Ed. 2d 462 (1991), a case which upheld a conviction of second-degree murder of a six-week-old baby based on circumstantial evidence. We think Boise is inapposite because, as the court in that case repeatedly pointed out, the defendant there had sole custody of the murdered boy when the injuries from which he died were evidently inflicted. Boise is therefore of no particular relevance to the circumstances of our case.

Finally, the government maintains that the jury could have reasonably believed the defendant's statement to the F.B.I. agent that Tory was making decerebrate (reflex, instinctive) movements when defendant found him on the floor, movements that the doctors established could be expected immediately after the infliction of an injury causing severe and irreversible cerebral damage. The government further says that the jury might have concluded from this that the defendant was the one who inflicted the injury, because that person would have indeed observed exactly what the defendant testified that he observed. The jury, the government correctly points out, may choose to believe some parts of a witness's testimony and reject other parts. So, the argument runs, the jury could properly have believed the defendant's statement as to his observation of Tory's condition and disbelieved his explanation of its cause. We understand the logic of the government's position and admit, for the sake of argument, that the jury may have well done all of these things, and we admit, too, that it was privileged to do so. But the government's conclusion, namely, that the verdict is therefore sustainable, misses the mark. We believe that while a reasonable mind might have followed the path outlined by the government, it would nevertheless still have had to entertain a reasonable doubt with respect to the identity of the perpetrator or perpetrators of the crime.

Because we conceive it to be our duty, when the sufficiency of the evidence is challenged, to peruse the entire record for evidence to support the verdict, we choose to deal briefly here with a matter not discussed by the parties on appeal but which we find suggested by the record. In his closing statement, defendant's counsel several times entreated the jury not to hold the defendant's sex against him as a means of overcoming the evident difficulty that there was no way to decide which of the two suspects was guilty. At one point, for instance, counsel ventured, "You're being asked, ladies and gentlemen of the jury, to believe and have faith ... that a mother could never kill its [sic] child. That's what the government's asking you to do." At another point, counsel urged the jury, "The only thing that the government could possibly be asking you to do is make some kind of sexual distinction between men and women, that women just wouldn't do things like that versus men...." The government, which had not broached the subject at all before, addressed the matter in its rebuttal closing. The prosecutor stated:

Well, ladies and gentlemen, we're not inexperienced adults as we approach this courtroom. We don't park our common sense and our common experiences at the door. One doesn't have to make any kind of sexist analogy to point the obvious out. Women are better mothers than men are. It's in the genes. We all know that. That isn't any startling thing. Women are better mothers than men are. Women do it much better than we do. The instantaneous reaction of most women to a crying child is to pick it up and comfort it. We've all seen that thousands of times in our lives. The instantaneous reaction of men is generally to find the child's mother. It's as simple as that. I've done it. All of you have done it. You've seen it. You've seen it in your families. That's simply the way it is.

We construe the prosecutor's remarks as an invitation to the jury to consider the defendant's sex in deciding his guilt. The jury, of course, is entitled to notice such facts as it has acquired knowledge of in its daily experience, but we are more than a little disturbed by the prospect that a jury might be authorized to put a person's sex in the balance in deciding whether to convict him. We wonder, too, about the propriety, in some future similar case, of admitting into evidence statistics and testimony about sex-based behavioral differences, including the propensity, if any, to commit certain kinds of crime. We view with special alarm the government's appeal to some alleged common knowledge about sex differences that are genetically determined.

It would be wrong, on the basis of so thin a record as this, to conclude that sex could never be relied on as proof of some relevant characteristic, and we have no need to do so. We do note, however, that evidence of a defendant's character, because of its potential to have unfairly prejudicial effects, is very strictly regulated by the Federal Rules of Evidence. We also observe that large equal protection difficulties are patent in the argument that "men are like that and women aren't." In any event, we are confident that, at a minimum, jury notice of alleged sex differences is insufficient, as a matter of law, to overcome the reasonable doubt that a rational juror must have entertained on the evidence presented in this case.

Finally, we note the possibility that the jury relied on the biological connection between victim and mother to conclude that she was less likely to have committed the act than the defendant, who was not Tory's father. We understand how such an assumption might have motivated the jury, but conclude, without commenting on the validity of that assumption, that it could not have furnished sufficient weight to overcome the presumption of innocence.

For the reasons given, we affirm the district court.