dissenting:
Because defendant has the requisite amount of Indian blood, the only question is whether he has “tribal or government recognition as an Indian.” United States v. Bruce, 394 F.3d 1215, 1223 (9th Cir.2005) (quoting United States v. Broncheau, 597 F.2d 1260, 1263 (9th Cir.1979) for the *852“generally accepted test,” derived from United States v. Rogers, 45 U.S. (4 How.) 567, 573, 11 L.Ed. 1105 (1846)). He plainly does. The record discloses that the Blackfeet tribal authorities have accorded Cruz “descendant” status, which entitles him to many of the benefits of tribal membership, including medical treatment at any Indian Health Service facility in the United States, certain educational grants, housing assistance and hunting and fishing privileges on the reservation.
That Cruz may not have taken advantage of these benefits doesn’t matter because the test is whether the tribal authorities recognize him as an Indian, not whether he considers himself one. That they do is confirmed by the fact that, when he was charged with an earlier crime on the reservation, the tribal police took him before the tribal court rather than turning him over to state or federal authorities. How that case was finally resolved is irrelevant; what matters is that the tribal authorities protected him from a state or federal prosecution by treating him as one of their own. Finally, Cruz was living on the reservation when he was arrested, another piece of evidence supporting the jury’s verdict.
The majority manages to work its way around all of this evidence by taking a stray comment in Bruce to the effect that certain factors have been considered in “declining order of importance” and turning it into a four-part balancing test. But Bruce was not announcing a rule of law; it was merely reporting what it thought other courts had done: “[Cjourts have considered, in declining order of importance, evidence of [four factors].” 394 F.3d at 1224. Bruce did not adopt this as any sort of standard, nor did it have any cause to do so, as nothing in Bruce turned on the relative weight of the factors. The majority strains hard to make this part of Bruce’s holding, but a fair reading of the opinion discloses that it’s not even dicta because it’s descriptive rather than prescriptive. We recognized this the last time we applied the test by omitting any reference to the declining order of importance. See United States v. Ramirez, 537 F.3d 1075, 1082 (9th Cir.2008).
Bruce borrowed the “declining order of importance” language from United States v. Lawrence, 51 F.3d 150, 152 (8th Cir.1995), and Lawrence itself was quoting the observation of a district judge in an earlier case, St. Cloud v. United States, 702 F.Supp. 1456, 1461-62 (D.S.D.1988). The district judge in St. Cloud did not cite most of the cases he relied on, so it’s hard to tell whether his observation is correct, but he did offer a note of caution that my colleagues overlook: “These factors do not establish a precise formula for determining who is an Indian. Rather, they merely guide the analysis of whether a person is recognized as an Indian.” Id. at 1461.
This is the opposite of what my colleagues do today: They turn the four factors into a rigid multi-part balancing test, with the various prongs reinforcing or offsetting each other, depending on how they are analyzed. This is not what the judge in St. Cloud had in mind, and certainly nothing like what Bruce adopted as the law of our circuit. It is an invention of the majority in our case, designed to take power away from juries and district judges and give it to appellate judges. Nothing in the law, dating back to the Supreme Court’s opinion in Rogers, justifies this fine mincing of the evidence. The question we must answer is whether there is enough evidence from which a rational jury could have concluded beyond a reasonable doubt that Cruz was recognized as an Indian. Clearly there was, and that’s the end of our task.
*853The majority misreads Bruce and misrepresents my position: “Given Bruce’s clear admonition that ‘tribal enrollment,’ and therefore a fortiori descendant status, ‘is not dispositive of Indian status,’ we reject the dissent’s argument that mere descendant status with the concomitant eligibility to receive benefits is effectively sufficient to demonstrate ‘tribal recognition.’” Maj. op. at 847 (quoting Bruce, 394 F.3d at 1224-25). Bruce certainly doesn’t hold that tribal enrollment is insufficient to support a finding of Indian status. Bruce holds the converse: that the absence of tribal enrollment does not preclude finding that defendant is an Indian — which was the question presented here. To suggest, as does the majority, that an individual who is enrolled as a member of a tribe might not be an Indian after all is not only preposterous, it’s unnecessary, as no one claims that Cruz was enrolled.
Nor do I maintain, as the majority makes believe, that Cruz’s descendant status is enough to make him an Indian. Whether or not it is, there are additional facts here: Cruz’s residence on the reservation and the fact that he was previously arrested and brought before the tribal court. The latter is a fact that the Bruce majority held to be highly significant. Bruce did not consider the disposition of prior tribal court cases relevant and we are not free to disregard the arrest and prosecution by tribal authorities on this spurious basis.
Worse still, after huffing and puffing for 11 hefty paragraphs and 12 chubby footnotes trying to explain why the district court erred at all, the majority concludes in a single opaque sentence that the error is “plain.” Just how plain can this error be when the majority has to struggle so long and hard to find any error at all? After complaining bitterly about pointy-headed judges who “slic[e] ever finer and finer distinctions whose practical consequences are seemingly minuscule, if not microscopic,” maj. op. at 845, my colleagues pull out a scalpel of their own and proceed to engage in the same exercise, so that “our standards of review continue to multiply, the relationships between them growing more obscure with each iteration.” Id. at 845. Before reading today’s opinion, no one could have guessed its outcome and methodology. Saying that the error is plain eviscerates the “plain” part of the plain error standard. If this is plain error, no error isn’t.
Not satisfied with merely reversing the verdict, the majority goes a bridge too far by converting its novel four-part test into a jury instruction. This is wholly unnecessary, as Cruz cannot be tried again for violating 18 U.S.C. § 1153 because of double jeopardy. It is also wrong. We don’t instruct juries as to how to weigh the evidence; that is their function, not ours. Yet the majority now requires jurors to assign relative weight to various pieces of evidence presented to them. I am aware of no such instruction anywhere else in our jurisprudence and the majority points to none. It is a bold step into uncharted territory and, in my judgment, an unwise one.
The majority engages in vigorous verbal callisthenics to reach a wholly counter-intuitive-and wrong-result. Along the way, it mucks up several already complex areas of the law and does grave injury to our plain error standard of review. I hasten to run in the other direction.