with whom Mr. Justice Powell and Mr. Justice Rehnquist join, dissenting.
In addition to the views expressed in Mr. Justice Powell’s dissent, I identify one other flaw in the Court’s opinion. What the majority characterizes as a prima facie case of discrimination simply will not “wash.” The decisions of this Court suggest, and common sense demands, that eligible population statistics, not gross population figures, provide the relevant starting point. In Alexander v. Louisiana, 405 U. S. 625, 630 (1972), for example, the Court in an opinion by Mr. Justice White looked to the “proportion of blacks in the eligible population . . . .” (Emphasis supplied.)
The failure to produce evidence relating to the eligible population in Hidalgo County undermines respondent’s claim that any statistical “disparity” existed in the first instance. Particularly where, as here, substantial numbers of members *505of the identifiable class actually served on grand jury panels, the burden rightly rests upon the challenger to show a meaningful statistical disparity. After all, the presumption of constitutionality attaching to all state procedures has even greater force under the circumstances presented here, where exactly one-half the members of the grand jury list now challenged by respondent were members of the allegedly excluded class of Mexican-Americans.
The Court has not previously been called upon to deal at length with the sort of statistics required of persons challenging a grand jury selection system. The reason is that in our prior cases there was little doubt that members of identifiable minority groups had been excluded in large numbers. In Alexander v. Louisiana, supra, the challenger’s venire included only one member of the identifiable class and the grand jury that indicted him had none. In Turner v. Fouche, 396 U. S. 346 (1970); Jones v. Georgia, 389 U. S. 24 (1967); Sims v. Georgia, 389 U. S. 404 (1967); and Whitus v. Georgia, 385 U. S. 545 (1967), there was at best only token inclusion of Negroes on grand jury lists. The case before us, in contrast, involves neither tokenism nor absolute exclusion; rather, the State has used a selection system resulting in the inclusion of large numbers of Spanish-surnamed citizens on grand jury lists. In this situation, it is particularly incumbent on respondent to adduce precise statistics demonstrating a significant disparity. To do that, respondent was obligated to demonstrate that disproportionately large numbers of eligible individuals were excluded systematically from grand jury service.
Respondent offered no evidence whatever in this respect. He therefore could not have established any meaningful case of discrimination, prima facie or otherwise. In contrast to respondent’s approach, which the Court’s opinion accepts without analysis, the Census Bureau’s statistics for 1970 demonstrate that of the adults in Hidalgo County, 72%, not *50679.1% as respondent implies, are Spanish surnamed. At the outset, therefore, respondent’s gross population figures are manifestly overinclusive.
But that is only the beginning. Respondent offered no evidence whatever with respect to other basic qualifications for grand jury service.1 The statistics relied on in the Court’s opinion suggest that 22.9% of Spanish-surnamed persons over age 25 in Hidalgo County have had no schooling at all. Ante, at 488-489, n. 8. Since one requirement of grand jurors in Texas is literacy in the English language, approximately 20% of adult-age Mexican-Americans are very likely disqualified on that ground alone.
The Court’s reliance on respondent’s overbroad statistics is not the sole defect. As previously noted, one-half of the members of respondent’s grand jury list bore Mexican-American surnames. Other grand jury lists at about the same time as respondent’s indictment in March 1972 were predominantly Mexican-American. Thus, with respect to the September 1971 grand jury list, 70% of the prospective grand jurors were Mexican-American. In the January 1972 Term, 55% were Mexican-American. Since respondent was indicted in 1972, by what appears to have been a truly representative grand jury, the mechanical use of Hidalgo County’s practices some 10 years earlier seems to me entirely indefensible. We do not know, and on this record we cannot know, whether respondent’s 1970 gross population figures, which served as the basis for establishing the “disparity” complained of in this case, had any applicability at all to the period prior to 1970. Accordingly, for all we know, the 1970 figures may be totally *507inaccurate as to prior years;2 if so, the apparent disparity alleged by respondent would be increased improperly.
Therefore, I disagree both with the Court’s assumption that respondent established a prima facie case and with the Court’s implicit approval of respondent’s method for showing an allegedly disproportionate impact of Hidalgo County’s selection system upon Mexican-Americans.
The burden of establishing a prima facie case obviously rested on respondent. It will not do to produce patently overinclusive figures and thereby seek to shift the burden to the State. Cf. ante, at 486-487, n. 6, 488-489, n. 8. Rather, a prima facie case is established only when the challenger shows a disparity between the percentage of minority persons in the eligible population and the percentage of minority individuals on the grand jury.
Indeed, Judge Reynaldo Garza in this case referred to Hidalgo County as “rapidly changing” and as experiencing “rapid growth.”