with whom Mr. Justice Stevens joins, dissenting.
Although I agree with Parts I and II of the Court’s opinion, I believe that a prima facie case of purposeful discrimination was made out and was not rebutted by the State. I therefore dissent from Parts III and IY and from the judgment. On the basis of the evidence presented at the evidentiary hearing in state court, the District Court concluded that respondents “appear [ed]” to have made out a prima facie case of discrimination in the selection of the foreman of the grand *589jury that indicted them. App. 99. However, upon the affidavits submitted by the State in response, the court concluded that in fact the foreman had been chosen for other than racial reasons, that he had not voted on the indictment, and thus that there had not been a violation of the Equal Protection Clause. Id., at 122. The Court of Appeals agreed that a prima facie case was shown, interpreting the record testimony to the effect that the recollections of those testifying were that there had never been a black chosen as foreman of a grand jury in Tipton County, and pointing out the potential for discrimination in a system which leaves the selection of the foreman to the discretion of a single judge who has not “really given any thought to appointing” a black, id., at 113. See 570 F. 2d 129, 134-135 (1978). The Court of Appeals disagreed, however, that this prima facie case had been rebutted by the testimony of the selecting judge that he had “no feeling against” appointing a black to be foreman, and found irrelevant that the foreman did not vote on respondents’ indictment. Id., at 131. Because we do not sit to redetermine the factfindings of lower courts, and because the Court of Appeals correctly enunciated and applied the law governing proof of discrimination in the context of grand jury selection, I dissent.
The only difference between this case and our previous cases voiding a conviction due to discriminatory selection of members of the grand jury is that in this case it has been shown only that the grand jury foreman, who did not vote on the indictment, was chosen in a manner prohibited by the Equal Protection Clause. I agree with the Court of Appeals that given the vital importance of the foreman in the functioning of grand juries in Tennessee,1 a conviction based on an *590indictment where the foreman was chosen in a discriminatory fashion is void just as would be a conviction where the entire grand jury' is discriminatorily selected, whether or not there is a showing of actual prejudice, see Castaneda v. Partida, 430 U. S. 482 (1977); Alexander v. Louisiana, 405 U. S. 625 (1972); Arnold v. North Carolina, 376 U. S. 773 (1964); Eubanks v. Louisiana, 356 U. S. 584 (1958); Cassell v. Texas, 339 U. S. 282 (1950); Patton v. Mississippi, 332 U. S. 463 (1947); Hill v. Texas, 316 U. S. 400 (1942); Pierre v. Louisiana, 306 U. S. 354 (1939); Bush v. Kentucky, 107 U. S. 110 (1883).
That this case involves only the foreman, rather than the entire grand jury, does have implications for the manner in which respondents may meet their burden of proving discrimination. In the context of racial discrimination in the selection of juries, “the systematic exclusion of Negroes is itself such an 'unequal application of the law ... as to show intentional discrimination/ ” a necessary component of any equal protection violation. Washington v. Davis, 426 U. S. 229, 241 (1976). Generally, in those cases in which we have found unconstitutional discrimination in jury selection, those alleging discrimination have relied upon a significant statistical discrepancy between the percentage of the underrepresented group in the population and the percentage of this group called to serve as jurors, combined with a selection procedure “that is susceptible of abuse or is not racially neutral.” Castaneda v. Partida, supra, at 494. See, e. g., Alexander v. Louisiana, supra; Turner v. Fouche, 396 U. S. 346 (1970); Carter v. Jury Comm’n, 396 U. S. 320 (1970). Once this *591showing is made, the burden shifts to the State to rebut the inference of discriminatory purpose. Castaneda v. Par-tida, supra, at 495. This method of proof, sometimes called the “rule of exclusion,” 430 U. S., at 494, may not be well suited when the focus of inquiry is a single officeholder whose term lasts two full years, as is true of the Tipton County grand jury foreman. For instance, in Castaneda v. Partida, we considered statistics relating to an 11-year period showing that 39% of the 870 persons selected for grand jury duty were Hispanic, from a general population that was over 79% Hispanic. The likelihood that this statistical discrepancy could be explained on the basis of chance alone was less than 1 in 10140. See id., at 495-496, and n. 17. The sample size necessarily considered in a case of discrimination in the selection of a foreman simply does not permit a statistical inference as overwhelming as that in Castaneda. During any 11-year period, there would be only five or six opportunities for selecting jury foremen in Tipton County, assuming that every foreman selected serves at least the full 2-year term.2
Despite the inherent difficulty of any statistical presentation with respect to discrimination in filling a particular grand jury spot, respondents nonetheless have made a strong showing of underrepresentation supporting an inference of purposeful discrimination. This Court is not in a position to reject the finding, explicitly made by the Court of Appeals and implicitly made by the District Court,3 that those who testified believed *592there had never been a black foreman during the period 1951— 1973. See Berenyi v. Immigration Director, 385 U. S. 630, 635 (1967); Graver Mfg. Co. v. Linde Air Products Co., 336 U. S. 271, 275 (1949). Assuming that 11 foreman selections were made during this period,4 the expected number of black foremen would be more than 3 — and the likelihood of no blacks being chosen would be less than 1 in 50 — if blacks, who constituted nearly a third of the county’s population, and whites had an equal.chance of being selected. I do not see how respondents could be expected to make a stronger statistical showing.5
In any event, any possible weakness in respondents’ statistical presentation was more than overcome by the additional evidence before the District Court. First, the selection of a foreman is left to the complete discretion of a single person— the circuit judge. The potentialitiés for abuse in such a system are obvious, cf. Castaneda v. Partida, supra, at 497; Carter v. Jury Comm’n, supra; Hernandez v. Texas, 347 U. S. 475, 479 (1954) (“key man” system). Moreover, the particular judge who chose the foreman of respondents’ grand jury had *593never chosen a black in any of the five counties for which he appointed foremen over a 6-year period, App. 113. Finally, the judge himself admitted that he had never even considered appointing a black foreman. Ibid.6 Although these facts are not necessarily inconsistent with an ultimate conclusion that respondents’ foreman was not chosen on racial grounds, they raise, in conjunction with the previously described statistical presentation, a strong inference of intentional racial discrimination, shifting the burden to the State. Clearly the Court of Appeals is correct that the Circuit Judge’s further self-serving statement that he had “nothing against” appointing blacks is not sufficient rebuttal, see Alexander v. Louisiana, 405 U. S., at 632; Turner v. Fouche, 396 U. S., at 361; Hernandez v. Texas, supra, at 481-482. It can hardly be said that the judge, as the official authorized by the State to appoint grand jury foremen, performed his “constitutional duty . . . not to pursue a course of conduct in the administration of [his] office which would operate to discriminate in the selection of jurors on racial grounds.” Hill v. Texas, 316 U. S., at 404.
See 570 F. 2d 129, 136 (1978):
“The foreman or forewoman is vitally important to the functioning of grand juries in Tennessee, being 'the thirteenth member of each grand jury organized during his term of office, having equal power and author*590ity in all matters coming before the grand jury with the other members thereof.’ Tenn. Code Ann. §40-1506. He or she is expected to assist the district attorney in investigating crime, may administer oaths to all witnesses, conduct the questioning of witnesses, must indorse and sign all indictments, and like every other chairperson is in a position to guide, whether properly or improperly, the decision-making process of the body. . . .” (Footnote omitted.)
The key numbers to compare are the number of blacks selected to be foremen and the total number of opportunities to select a foreman. The latter number may be greater than the number of different individuals who serve if the appointing judge has an inclination to reappoint those who have previously served.
The District Court did not. make written findings of fact explaining the basis of its conclusion that a prima facie case appeared to have been established. However, the Court of Appeals was in a position to dispose of the appeal, without the necessity of a remand to the District Court, because the record and the District Court’s conclusions of law clearly *592reveal the basis for its conclusion, see Finney v. Arkansas Board of Correction, 505 F. 2d 194 (CA8 1974). This was the failure of any of the foremen who testified at the state-court hearing to recollect there having been a black foreman, and the inference therefrom — not clearly erroneous, see Fed. Rule Civ. Proc. 52 (a) — that these witnesses believed there had nevér been a black foreman.
See n. 2, supra.
If there were any doubt that the evidence adduced in the state-court hearing on respondents’ plea in abatement was insufficient — perhaps because it did not unequivocally establish the race of every foreman chosen since 1950 — the appropriate course would be for the District Court to hold an evidentiary hearing. See Townsend v. Sain, 372 U. S. 293, 313 (1963) (evidentiary hearing must be held “unless the state-court trier of fact has after a full hearing reliably found the relevant facts”); 28 U. S. C. § 2254 (d) (3) (determination of merits of factual issue by state court shall be presumed to be correct unless it appears “that the material facts were not adequately developed at the State court hearing”).
Clearly, it is irrelevant that the admissions on the part of the selecting judge that he had never given thought to appointing, and indeed had never appointed, a black foreman came as part of the petitioner’s written response to respondents’ petitions for writs of federal habeas corpus. In ascertaining whether a plaintiff has carried his burden of proof, all the evidence must be considered. It is not unusual that an affidavit or other evidence submitted by one party to a lawsuit turns out to be of primary, and perhaps even determinative, aid to the other party.