Drawn from hamlet and metropolis alike, jurors daily give life to the ancient right, enshrined in the Magna Carta and the Constitution, to a trial by jury of one's peers. Although now summoned by statute instead of the medieval sheriff's writ of venire facias, a representative jury array remains the expression of the community's role in securing this fundamental right. Accordingly, both the sixth and fourteenth amendments mandate that prospective jurors be chosen free of the taint of racial discrimination.

It is the interplay of these two provisions that is now before us. The State of Connecticut appeals from judgments of the United States District Court for the District of Connecticut, granting Alston's and Haskins's petitions for writs of habeas corpus. Using modern statistical data, the district court ruled that the jury selection system employed in 1975 in Connecticut state court, where the petitioners were tried, violated the equal protection clause. Agreeing with Chief Judge Daly's analysis, we affirm.

A brief factual statement is necessary to understand the law in this complex area. In Connecticut, each county compiles a jury array. Conn.Gen.Stat. Secs. 51-217 et seq. Each town in the county, in turn, furnishes a number of prospective jurors. Unfortunately, section 51-220 established a strict quota system which favored representation of the smaller towns. As an illustration, the smallest town in New Haven County, Beacon Falls, contributed 4.2% of its adult population to the array. In contrast, New Haven, the largest town, was limited by statute to assigning a mere 1.1% of its adults. It is undisputed that a larger concentration of the black population in Connecticut lives in the more populated urban settings. Accordingly, Sec. 51-220 was repealed in 1982. Conn. Public Act 82-307 Sec. 2.

The jury that convicted Michael Alston and James Haskins, however, was chosen from an array skewed in favor of the white majority. Alston, Haskins, and a co-defendant were convicted in state court of assaulting two police officers while fleeing the scene of a New Haven bank robbery. Each had previously been convicted in federal court of the underlying robbery. Both Alston and Haskins appealed their convictions to the Connecticut Supreme Court, presenting, inter alia, the same constitutional grounds supporting their subsequent habeas corpus petitions. The supreme court ruled that the jury selection system passed constitutional muster. State v. Haskins, 188 Conn. 432, 436-41, 450 A.2d 828 (1982). Currently, Alston is on parole from both his state and federal convictions. Haskins is still incarcerated in federal prison and has not yet begun to serve his state sentence of ten years.

Subsequently, Alston and Haskins filed habeas corpus petitions in United States district court. The petitions contended, inter alia, that the statutory scheme for selecting the array evidenced an intent to discriminate against blacks in contravention of the equal protection clause. They also argued that blacks were substantially under-represented among the potential jurors, a violation of the sixth amendment. At a later hearing, a professor of statistics from Yale University testified to the adverse effect upon black representation produced by the existing town quotas. The statistician estimated the number of blacks summoned was 368. In the absence of Sec. 51-220, however, the number would have been closer to 501. The total number in the array was 8,405.

Relying on these figures, the district judge decided the Connecticut plan violated the fourteenth amendment. Employing the three-prong test set forth in Castaneda v. Partida, 430 U.S. 482, 97 S. Ct. 1272, 51 L. Ed. 2d 498 (1977), Chief Judge Daly found the petitioners had established a prima facie case of racial discrimination, and that the State had failed to rebut this showing. Alston v. Lopes, 621 F. Supp. 992 (D.Conn.1985).

The district court easily found the first and third branches of the Castaneda test had been satisfied. The first requirement is that the excluded group be cognizable. Because blacks comprised the group underrepresented, this standard was clearly met. Similarly, the third part of the standard, establishing that the selection procedure was not racially neutral, was also on firm footing. All parties agree a high proportion of blacks live in the City of New Haven, and that the quota system could not help but partially exclude them.

The district court, therefore, devoted most of its analysis to the second Castaneda standard: whether the underrepresentation of blacks was substantial. Through the use of an established and accepted data device, Statistical Decision Theory, the district judge determined that the odds against chance alone accounting for the low number of blacks were simply astronomical. Indeed, Chief Judge Daly noted the likelihood of only 368 blacks appearing as potential jurors was three chances in one billion.

Having found that Alston and Haskins had established a prima facie case of intentional discrimination, the district court then inquired whether the State had rebutted that showing. Connecticut, however, persisted in asserting that the discrepancy was due to the voter registration lists instead of the town quota system. But, Chief Judge Daly observed that blacks would be underrepresented even if the use of voter rolls was abandoned, and accordingly deemed the State's argument inapposite. As a result, the district court found that Connecticut presented no evidence of a neutral intent motivating the system and granted the petitions for habeas corpus. These appeals followed.

Before us now is the seemingly narrow question whether the underrepresentation of blacks in the 1975 array was "substantial." Specifically, we must determine whether the presence of an estimated 368 blacks instead of 501 gave rise to a presumption of discriminatory intent. Our answer, however, requires a careful delineation of the constitutional standards governing selection of potential jurors.

The equal protection clause of the fourteenth amendment condemns underrepresentation of minorities only if it is the product of intentional discrimination. The Supreme Court, however, has squarely held that substantial underrepresentation of a cognizable group can establish a prima facie case of discriminatory purpose. Castaneda v. Partida, supra, 430 U.S. 482, 97 S. Ct. 1272. In Castaneda, a habeas corpus petitioner alleged that his indictment by a grand jury on which Mexican-Americans were underrepresented violated the fourteenth amendment. The High Court established a tripartite analysis for analyzing this claim. A prima facie case is made, stated the Court, if a cognizable group is substantially underrepresented and if the selection procedure is not racially neutral. Id. at 494-95, 97 S.Ct. at 1280. In Castaneda, as here, the Court found that the State was unable to rebut the presumption of discrimination raised by statistical evidence. Id. at 501, 97 S.Ct. at 1283.

We agree with the district court that the three branches of the Castaneda test were satisfied. Blacks, of course, form a cognizable group. Carter v. Jury Commission, 396 U.S. 320, 90 S. Ct. 518, 24 L. Ed. 2d 549 (1970). Moreover, the third requirement of Castaneda--that the jury selection system is not racially neutral--is also fulfilled. The ineluctable effect of the statutory scheme was to reduce the number of potential jurors drawn from large towns, especially New Haven, where it is undisputed that a larger proportion of blacks resided.

Turning to the central issue of this appeal, we hold that blacks were substantially underrepresented as a result of Connecticut's earlier jury selection system. The reliable tool referred to as Statistical Decision Theory is an appropriate instrument for analyzing whether minority underrepresentation in the jury array is substantial. The goal of such an inquiry is to determine if chance alone could account for a meager representation of minorities. See, e.g., Castaneda, 430 U.S. at 494 n. 13, 97 S. Ct. at 1280 n. 13; Smith v. Texas, 311 U.S. 128, 131, 61 S. Ct. 164, 165, 85 L. Ed. 84 (1940). The Statistical Decision Theory formulation resulted from an effort to determine what one effect, chance, could have on the flux of events, and is therefore an appropriate and acceptable augmentation to the legal arsenal. It has been adopted in the context of employment discrimination. See Board of Education v. Califano, 584 F.2d 576, 584 n. 29 (2d Cir.1978). Moreover, modern statistical approaches have been employed by other courts in jury selection cases. Castaneda, 430 U.S. at 496 n. 17, 97 S. Ct. at 1281 n. 17; Moultrie v. Martin, 690 F.2d 1078 (4th Cir.1982); Villafane v. Manson, 504 F. Supp. 78 (D.Conn.1980); see generally Finkelstein, The Application of Statistical Decision Theory to the Jury Discrimination Cases, 80 Harv.L.Rev. 338 (1966) (mathematical description of theory).

The Statistical Decision Theory formulation may be summarized as a set of mathematical calculations that indicate whether an observed outcome could reasonably be ascribed to chance. The district court found, for example, that the expected number of blacks in the array was 501. Surely, however, a violation would not be found if the actual number turned out to be 500 or 490. Such a result could easily have resulted from random factors. In contrast, the estimated number of blacks who would have actually served on the array in light of the quota system was 368. The odds of such a result occurring through happenstance are an incredible three in one billion. Accordingly, in light of the scant likelihood of the Connecticut result, the district court correctly ruled that the State had the burden of producing a plausible justification for Sec. 51-220.1

Although not disputing the figures relied upon below, the State objects that the proper standard was the absolute disparity of blacks excluded. Under this view, one would first note that the array should have been 5.96% black. Instead, 4.38% of the prospective jurors were black. Connecticut argues the resulting discrepancy of 1.58% is not significant. United States v. Jenkins, 496 F.2d 57 (2d Cir.1974), cert. denied, 420 U.S. 925, 95 S. Ct. 1119, 43 L. Ed. 2d 394 (1975).

The State, however, has confused equal protection and sixth amendment standards. In an analysis pursuant to the sixth amendment, a disparity can constitute a per se violation of an accused's right to trial by jury. Duren v. Missouri, 439 U.S. 357, 99 S. Ct. 664, 58 L. Ed. 2d 579 (1979). In Jenkins, supra, we ruled that a difference of one black in a pool of 60 jurors was not substantial. 496 F.2d at 66. In the instant case, however, the use of statistics was not to establish a violation itself, but to raise the presumption of discriminatory intent. The Statistical Decision Theory analysis is ideally suited for shedding light on this issue because it reveals the possible role of chance and works well where a small sample is involved, as here. Villafane, supra, 504 F.Supp. at 86.

Nonetheless, in light of Jenkins, our result might appear anomalous. The State argues that the town quota system would not violate the sixth amendment, although it is generally thought to set forth a more stringent standard than the equal protection clause. The sixth amendment is stricter because it forbids any substantial underrepresentation of minorities, regardless of whether the State's motive is discriminatory. The fourteenth amendment, however, imposes the additional requirement of discriminatory purpose. In this case, it appears that a constitutional transgression was more easily established pursuant to the equal protection clause, despite the added burden of proving intent.

The answer to this seeming quandray is easily available. First, as the appellees strenuously urge, the absolute disparity approach employed in Jenkins may be outmoded and should be discarded. In light of our fourteenth amendment analysis, however, we need not reach this question, and therefore decline to do so. Nonetheless, if this Court should subsequently apply modern statistical theory to the sixth amendment as well as the fourteenth, then the two constitutional analyses would be more congruent. In that case, it would be unlikely that a jury selection system would be found to violate the equal protection clause, but not the sixth amendment.

More importantly, the statistics herein employed raised only a presumption of discrimination. The State was free to rebut this presumption, but failed to do so. Indeed, Chief Judge Daly flatly stated that "[n]o attempt has been made to demonstrate that the town quota system serves a legitimate state purpose." 621 F. Supp. at 998.

Accordingly, the judgments of the district court granting Alston's and Haskins's petitions for habeas corpus are affirmed.

Because I believe that the statistical analysis employed by the petitioners was seriously flawed and therefore insufficient to raise the presumption of discriminatory intent, I respectfully dissent.

Since the Connecticut quota system for summoning jurors favored representation of the smaller towns, the district court concluded that the statute discriminates against blacks, who reside predominately in the more populated regions of the county. The conclusion was based entirely on the expert testimony of a statistician, who estimated that 368 blacks would be summoned under the quota system instead of the 501 who would appear if they were summoned in proportion to their representation in the entire county population.

Apparently lacking information as to the number of blacks registered to vote and therefore eligible for jury duty, the statistician presumed black voter registration (as well as white voter registration) to be 100% of the adult population for the purposes of his model. This approach was believed to be more than fair, because the effect of the voting registration requirement was to decrease the number of blacks actually summoned. Indeed, the State conceded that the use of voter lists "has traditionally resulted in an under-representation of blacks for blacks as a group tend to register to vote less frequently than do whites." Brief for the Appellant at 5.

In order to find intentional discrimination on the part of the State of Connecticut, however, it is not enough simply to demonstrate an underrepresentation of blacks in the array of jurors. What must be shown is that the Connecticut statute brought about that underrepresentation. In the absence of actual information as to the number and distribution of blacks registered to vote, i.e., eligible for jury service, in the county population, that showing cannot be made. The reason it cannot is that the black proportion of those registered to vote in the towns may be greater than the black proportion of those registered to vote in the city, and the statute, since it favors the towns, would therefore have no pernicious effect insofar as jury service by blacks is concerned. Underrepresentation of blacks, if found under those circumstances, would be due to some factor other than the statute challenged by these petitioners.

Two other considerations favor the denial of the writ sought by the petitioners. First, the number of blacks in the array summoned under the challenged statute is estimated. In previous cases involving fourteenth amendment violation claims, the actual number of the minority members summoned was known. Vasquez v. Hillery, *--- U.S. ---, 106 S. Ct. 617, 88 L. Ed. 2d 598 (1986) (total absence of blacks from grand juries was known fact); Castaneda v. Partida, 430 U.S. 482, 97 S. Ct. 1272, 51 L. Ed. 2d 498 (1977) (actual number of Mexican-Americans summoned over an extended period was "observed"); Villafane v. Manson, 504 F. Supp. 78 (D.Conn.) (actual percentage of Puerto Ricans in the electorate, as well as actual number summoned, was known), aff'd mem., 639 F.2d 770 (2d Cir.1980). Here, not only is the actual number of blacks summoned under the quota system unknown, but the estimate is not even based on the universe of blacks eligible for jury duty. Second, the use of statistical models to prove claims of discrimination in jury selection procedures has been limited heretofore to cases involving a subjective selection process. J. Nowak, R. Rotunda & J. Young, Constitutional Law 602 (2d ed. 1983); see Brown v. Allen, 344 U.S. 443, 73 S. Ct. 397, 97 L. Ed. 469 (1953). When, as here, jurors are selected randomly, in accordance with objective statutory criteria, there is no basis for the use of statistical analysis to raise a presumption of discriminatory intent. Under these circumstances, we are required to presume that the statute meets constitutional standards, absent a showing of discriminatory intent on the part of the legislature. Arlington Heights v. Metropolitan Housing Development Corp., 429 U.S. 252, 265, 97 S. Ct. 555, 563, 50 L. Ed. 2d 450 (1977).

We deal here with two people who stand convicted of serious crimes involving the wounding of police officers by gunfire in the course of a robbery. Their convictions were affirmed after appeal to the Connecticut Supreme Court, where the same argument presented here was advanced and rejected. Before granting the writ to these petitioners, thereby making possible the release of an unknown number of offenders convicted by juries drawn under the challenged statute, I would at least remand to the district court for further findings as to the actual operation of Connecticut's statutory jury quota system.