BOWNES, Circuit Judge.
Petitioner, James Barber, seeks a writ of habeas corpus pursuant to 28 U.S.C. § 2254 on the ground that an alleged systematic exclusion of young people from his Massachusetts state court jury venire violated his constitutional right to an impartial jury drawn from a cross-section of the community.
In September of 1980 Barber, a prisoner at Massachusetts Correctional Institution at Norfolk, was charged with unlawful possession of a hypodermic syringe, unlawful possession of two hypodermic needles, unlawful possession of heroin and unlawful possession of marijuana. Prior to his jury trial, Barber moved to dismiss the venire on the grounds that an alleged systematic exclusion of young people violated Mass. GemLaws Ann. ch. 234, § 4 (1969) and the Massachusetts and United States Constitutions. The court held a hearing on February 1, 1981. The only evidence presented at the hearing on the pretrial motion was a copy of a report on the representation of persons age eighteen to thirty-four on Norfolk County juries that had been generated in the case of Commonwealth v. Flaherty, Criminal No. 76318 (Norfolk Sup.Ct.). The report indicated that for the twenty-five-month period between October 1, 1978, and October 31, 1980, young adults (age 18-34) were underrepresented by 50% in the jury-selection process.
The case went to trial on April 21, 1981, before a six-person jury (plus one alternate) that included two persons under the age of thirty-five. The jury found defendant not guilty on the count of possession of marijuana and guilty on all other counts. Petitioner appealed his conviction which was affirmed by the Massachusetts Appeals Court in Commonwealth v. Barber, 14 Mass.App. 1008, 441 N.E.2d 763 (1982), and unsuccessfully sought further appellate review from the Supreme Judicial Court. See Commonwealth v. Barber, 388 Mass. 1101 (1983). Petitioner thereafter brought a habeas petition in federal district court. The district court denied the petition and a motion for certificate of probable cause. Petitioner sought a certificate of probable cause from this court, which we granted.
The Supreme Court has long recognized that the vitality of a defendant’s fundamental right to a trial by jury is dependent upon the composition of the jury.1 In Glas-ser v. United States, the Supreme Court declared:
[984]*984Lest the right of trial by jury be nullified by the improper constitution of juries, the notion of what a proper jury is has become inextricably intertwined with the idea of jury trial____ Our notions of what a proper jury is have developed in harmony with our basic concepts of a democratic society and a representative government. For it is part of the established tradition in the use of juries as instruments of public justice that the jury be a body truly representative of the community.
315 U.S. 60, 85, 62 S.Ct. 457, 471, 86 L.Ed. 680 (1942) (citations omitted). See also Thiel v. Southern Pacific Co., 328 U.S. 217, 220, 66 S.Ct. 984, 985, 90 L.Ed. 1181 (1946) (“The American tradition of trial by jury, considered in connection with either criminal or civil proceedings, necessarily contemplates an impartial jury drawn from a cross-section of the community.”); Smith v. Texas, 311 U.S. 128, 130, 61 S.Ct. 164, 165, 85 L.Ed. 84 (1940).
The Supreme Court has grounded the requirement of a jury drawn from a cross-section of the community on the equal protection clause, Hernandez v. Texas, 347 U.S. 475, 476-82, 74 S.Ct. 667, 669-72, 98 L.Ed. 866 (1954); Strauder v. West Virginia, 10 Otto 303, 100 U.S. 303, 307-10, 25 L.Ed. 664 (1879), the supervisory power of the court over the right to jury trials in federal courts, Ballard v. United States, 329 U.S. 187, 193, 67 S.Ct. 261, 264, 91 L.Ed. 181 (1946); Thiel v. Southern Pacific Co., 328 U.S. 217, 66 S.Ct. 984, 90 L.Ed. 1181 (1946), and the sixth amendment which binds the states through the fourteenth amendment,2 Taylor v. Louisiana, 419 U.S. 522, 95 S.Ct. 692, 42 L.Ed.2d 690 (1975).3
In order to establish a prima facie violation of the cross-section requirement, a petitioner must show: (1) that the group alleged to be excluded is a “cognizable” or “distinctive” group in the community; (2) that the representation of this group in venires from which juries are selected is not fair and reasonable in relation to the number of such persons in the community; and (3) that the underrepresentation in petitioner’s venire is due to the systematic exclusion of the group in the jury-selection process. Duren v. Missouri, 439 U.S. 357, 364, 99 S.Ct. 664, 668, 58 L.Ed.2d 579 (1979); Hernandez v. Texas, 347 U.S. at 480, 74 S.Ct. at 671.
I. COGNIZABILITY
The Supreme Court first considered the issue of cognizability in Strauder v. West Virginia, 10 Otto 303, 100 U.S. 303, 25 L.Ed. 664 (1879), an equal protection case. In sustaining defendant’s equal protection argument the Court assumed that blacks were cognizable for jury selection purposes and took judicial notice of the effects of prejudice. The Court identified two constitutional values impaired by the exclusion of a cognizable group: (1) potential prejudice against the defendant, and (2) stigmatization of the group excluded from service.
In Glasser v. United States, 315 U.S. 60, 62 S.Ct. 457, 86 L.Ed. 680 (1942), a case involving discretion in the selection of women jurors, the Court spelled out the prerogatives of jury commissioners in the jury selection process:
Th[e] duty of selection ... must always accord with the fact that the proper functioning of the jury system, and, indeed, our democracy itself, requires that the jury be a “body truly representative of the community” and not the organ of any special group or class. If that requirement is observed [the jury commissioners] ... may exercise some discretion to the end that competent jurors may be called. But they must not allow the desire for competent jurors to lead them into selections which do not comport with the concept of the jury as a cross-section of the community.
[985]*985315 U.S. at 86, 62 S.Ct. at 472. Four years later, in Thiel v. Southern Pacific Co., 328 U.S. 217, 220, 66 S.Ct. 984, 985, 90 L.Ed. 1181 (1946), the Court identified additional cognizable groups including economic, social, religious, racial, political, and geographic groups within a community, and held that economic distinctions in jury selection resulted in the exclusion of a cognizable group. The plaintiff in Thiel, suing for damages in á federal diversity tort case, moved to strike the jury venire on the ground that poor people were systematically excluded from the jury rolls. The Court held that exclusion of people who were paid daily wages rather than paid by salary failed to satisfy the general principles underlying proper jury selection because the exemption or exclusion was completely irrelevant to a person’s eligibility and capacity to serve as a juror.4 Id. at 223, 66 S.Ct. at 987. The Court did not declare that daily wage earners had a particular community of interest or values distinct from the rest of the population. Rather, it grounded its decision on the possibility that those responsible for the selection of jury panels could discriminate against persons of low economic and social status. Id. at 224, 66 S.Ct. at 987.
In Ballard v. United States, 329 U.S. 187, 67 S.Ct. 261, 91 L.Ed. 181 (1946), the Court sustained a challenge based on systematic exclusion of women. The Court assumed that women are a cognizable group and responded to the argument that women jurors behave similarly to male jurors.
It is said ... that an all male panel drawn from the various groups within a community will be as truly representa-five as if women were included. The thought is that the factors which tend to influence the action of women are the same as those which influence the action of men — personality, background, economic status — and not sex. Yet it is not enough to say that women when sitting as jurors neither act nor tend to act as a class. Men, likewise, do not act as a class. But, if the shoe were on the other foot, who would claim that a jury was truly representative of the community if all men were intentionally and systematically excluded from the panel? The truth is that the two sexes are not fungible; a community made up exclusively of one is different from a community composed of both; the subtle interplay of influence one on the other is among the imponderables. To insulate the courtroom from either may not in a given case make an iota of difference. Yet a flavor, a distinct quality is lost if either sex is excluded.
Id. at 193-94, 67 S.Ct. at 264.
Eight years later in Hernandez v. Texas, 347 U.S. 475, 74 S.Ct. 667, 98 L.Ed. 866 (1954), an equal protection challenge to the jury venire, the Court began to suggest the means for factually proving that a group of excluded people constitute a cognizable class so that litigants would not have to depend upon courts taking judicial notice of the cognizability of a particular group. The Court rejected the contention that systematic exclusion challenges brought under the fourteenth amendment were limited to race and color and found that Mexiean-Americans in Jackson County constituted a cognizable group stating that “community [986]*986prejudices are not static, and from time to time other differences from the community norm may define other groups which need the same protection. Whether such a group exists within a community is a question of fact.” Id. at 478, 74 S.Ct. at 670. The Court instructed, “[0]ne method by which [cognizability] ... may be demonstrated is by showing the attitude of the community.” Id. at 479, 74 S.Ct. at 671.
II. THE COGNIZABILITY OF YOUNG ADULTS
The Supreme Court has yet to rule on whether young adults are a cognizable group for jury selection purposes.5 This court has twice judicially noticed that young adults constitute a cognizable group. See LaRoche v. Perrin, 718 F.2d 500, 504 (1st Cir.1983); United States v. Butera, 420 F.2d 564, 570 (1st Cir.1970). A number of state courts have also held that jurors cannot be systematically excluded on the basis of age. See, e.g., Julian v. State, 134 Ga.App. 592, 215 S.E.2d 496, 499 (1975) (pool in which average age of jurors was sixty-nine cannot be considered a fairly representative cross-section of the community); People v. Fujita, 117 Cal.Rptr. 757, 770, 43 Cal.App.3d 454 (1974), cert. denied, 421 U.S. 964, 95 S.Ct. 1952, 44 L.Ed.2d 451 (1975) (citing age as among groups constituting cognizable classes); Paciona v. Marshall, 45 A.D.2d 462, 359 N.Y.S.2d 360, 362-63, aff'd, 35 N.Y.2d 289, 360 N.Y.S.2d 882, 319 N.E.2d 199 (1974) (systematic exclusion of students on grand jury array illegal); People v. Marr, 67 Misc.2d 113, 324 N.Y.S.2d 608 (1971) (exemption of students resulted in systematic exclusion of young adults 21 to 29); State v. Holstrom, 43 Wis.2d 465, 168 N.W.2d 574, 578 (Wise.1969) (court stated that systematic exclusion based on age would render jury defective but defendant failed to produce data indicating underrepresentation in pool). Because, however, the state appeals court, relying on Massachusetts Supreme Judicial Court’s decision in Commonwealth v. Bastarache, 382 Mass. 86, 414 N.E.2d 984, 993 (1980), found that underrepresentation of young people on the venire did not constitute ground for relief and a number of other courts have rejected jury selection challenges based on the exclusion of young people, we again focus our attention on the considerations relevant to a determination of cognizability.6 Such considerations include: (1) whether the group can be adequately defined; (2) whether the group is sufficiently cohesive to be cognizable; and (3) whether the exclusion of the group might cause juries to be biased against a defendant from that group or could encourage public prejudice against the group. [987]*987See Willis v. Zant, 720 F.2d 1212, 1216 (11th Cir.1983), cert. denied, — U.S.-, 104 S.Ct. 3546, 82 L.Ed.2d 849 (1984); United States v. Test, 550 F.2d 577, 584 (10th Cir.1976). See also Zeigler, Young Adults as a Cognizable Group in Jury Selection, 76 U.Mich.L.Rev. 1045, 1071-72 (1978).
A. Definition of Group
The difficulty in delimiting a group of “young adults” has posed a problem to some courts. See, e.g., United States v. Ross, 468 F.2d 1213, 1217 (9th Cir.1972), cert. denied, 410 U.S. 989, 93 S.Ct. 1500, 36 L.Ed.2d 188 (1973); United States v. Guzman, 337 F.Supp. 140, 143 (S.D.N.Y.), aff'd, 468 F.2d 1245 (2d Cir.1972), cert. denied, 410 U.S. 937, 93 S.Ct. 1397, 35 L.Ed.2d 602 (1973). While we recognize that the boundaries of any age group are necessarily somewhat arbitrary, it does not follow that no lines should be drawn. Society regularly makes decisions based on age distinctions despite the fact that the lines drawn are arbitrary to some degree. See, e.g., U.S. Const, art. I, § 2, cl. 2; § 3, cl. 4; art. II, § 1, cl. 5 (requiring that a member of the House of Representatives be twenty-five years of age, a Senator be thirty years old and the President be thirty-five years old); Massachusetts Board of Retirement v. Murgia, 427 U.S. 307, 96 S.Ct. 2562, 49 L.Ed.2d 520 (1976) (upholding statute requiring that state police retire at age fifty despite the arbitrariness of particular age limit)7; Johnson v. Mayor and City Council of Baltimore, 731 F.2d 209 (4th Cir.1984) (upholding Baltimore pension provision requiring firefighters to retire at fifty-five), cert. granted, — U.S.-, 105 S.Ct. 901, 83 L.Ed.2d 917 (1985).
While we recognize that legislative line drawing of precise boundaries is generally more acceptable than similar judicial efforts, cf. Roe v. Wade, 410 U.S. 113, 93 S.Ct. 705, 35 L.Ed.2d 147 (1973), defining the parameters of a group excluded from jury selection is not the type of line drawing that a legislature is likely to undertake. We, therefore, conclude that if young adults can generally be conceived of as an identifiable group, “[w]e cannot allow the requirements of a ‘distinct’ group to be applied so stringently with regard to age grouping that possible discrimination against a large class of persons ... will be insulated from attack.” United States v. Butera, 420 F.2d at 570. See also LaRoche v. Perrin, 718 F.2d at 504. Here, petitioner alleged systematic exclusion of young adults eighteen to thirty-four. We think that such a comprehensive grouping which encompasses virtually all whose chronological age entitles them to be called young adults, and which we accepted in LaRoche, is certainly one, if not the only, means of delimiting the group.8 Accordingly, we take judicial notice of the fact that eighteen to thirty-four year olds represent all persons who may be characterized as “young adults.”
B. Cohesiveness
The state alleges that recognition of any young adult group is not possible because young adults do not have sufficient cohesiveness to constitute an identifiable group [988]*988for jury selection purposes. They cite a number of cases from other circuits holding that young adults do not have a sufficient community of interest to qualify as a cognizable group. See, e.g., United States v. Potter, 552 F.2d 901, 905 (9th Cir.1977); Brown v. Harris, 666 F.2d 782, 784 (2d Cir.1981), cert. denied, 456 U.S. 948, 102 S.Ct. 2017, 72 L.Ed.2d 472 (1982). We decline to follow Massachusetts’ argument for several reasons.
Although conclusive evidence of the experimental or attitudinal distinctiveness of any group is difficult, if not impossible, to obtain many sociologists contend that young people do comprise a class with distinctive values, outlooks, manners, roles, and behavior patterns.9 A study of juror attitudes undertaken by Professor Zeigler of Pace University and submitted at trial in Johnson v. Durante, 387 F.Supp. 149 (E.D.N.Y.1975), suggests that jury deliberations are affected by the exclusion of young people. In fact, Zeigler’s study showed greater differences of opinion by age than by race or gender in such matters as: the inference of guilt from silence; the presumption of innocence; the predisposition of the young to break the law; and the performance and prerogatives of the police. See Zeigler, Young Adults as a Cognizable Group in Jury Selection, 76 Mich.L.Rev. 1045, 1090 (1978).
Differences in opinion absent from a jury because of gross underrepresentation of young people could be expected to affect a large number of cases, particularly those involving young adult defendants. More young adults are involved in the criminal process than any other age group. In 1983, approximately 6.4 million adults between eighteen and thirty-four were arrested compared to 2.2 million adults over thirty-four. Uniform Crime Reports for the United States — 1983 179-180 (Released Sept. 9, 1984). Similarly, in 1982, 6.2 million persons between the ages of eighteen and thirty-four were reported arrested by federal and state officers while 2 million persons arrested were thirty-five years of age or older. And in 1981, 6.1 million persons between eighteen and thirty-four were arrested while 2.1 million persons over thirty-five were reported arrested. Statistical Abstract of the United States 1984, 183-84 (104th ed. 1984). Such a high proportion of young arrestees causes us to be particularly concerned about the possibility of prejudice.
As we noted in Butera, 420 F.2d at 570, and LaRoche, 718 F.2d at 504, society at large clearly distinguishes between young adults and middle aged and older adults. The draft, the right to vote, insurance rates, and restrictions on the ages of elected federal officials are just a few of the many facets of our lives geared to age. A causal glance at service and product advertising, hiring and seniority systems, organizational hierarchies, and the media reveals that almost no area of society is age blind or age neutral. 10
To demand that a litigant show more precisely that the members of a cognizable [989]*989group would be influenced by the same factors or would cast their vote as a lot would make a mockery of the values the jury system is designed to protect. In essence it would require litigants to stereotype and overgeneralize the attributes and experiences of a group in order to “save” it from prejudice and discrimination. The Supreme Court has steadfastly refused to do that with racial and economic groups and women, see, e.g., Peters v. Kiff, 407 U.S. 493, 503-04, 92 S.Ct. 2163, 2168-69, 33 L.Ed.2d 83 (1972), and we see no reason to follow a different path for age-groups.
C. Underrepresentation Is Statistical: ly Significant
In order to establish the second element in his prima facie case petitioner must show that young adults were significantly underrepresented in the challenged venires. Duren v. Missouri, 439 U.S. at 364, 99 S.Ct. at 668. To that end, petitioner presented statistical evidence compiled by a court appointed special master in Commonwealth v. Flaherty, Norfolk County Criminal No. 76318. The parties stipulated that the report be made part of the record. The report indicated that the venires for the period October 1, 1978, to October 31, 1980, contained approximately 20% fewer young adults than an age-blind selection process should have produced,11 and that the venires contained less than half the young adults one could expect in a venire proportional to either the general population or the general population under the age of seventy.12 The master reported that the discrepancy is clearly statistically significant, that the probability of a 20% discrepancy occurring in an age-neutral selection process is less than one in 100 quadrillion.13 The Supreme Court has found 20% underrepresentation to be constitutionally cognizable. See Smith v. Texas, 311 U.S. 128, 129, 61 S.Ct. 164, 164, 85 L.Ed. 84 (1940) (20% underrepresentation of blacks); Hernandez v. Texas, 347 U.S. 475, 74 S.Ct. 667, 98 L.Ed. 866 (1954) (exclusion of Mexican-Americans who comprised 14% of the community). We, therefore, find that petitioner has presented sufficient evidence of the under-representation of young adults to meet the second requirement of his prima facie case.
D. Systematic Exclusion
Finally, to establish his prima facie case, petitioner must show that the underrepresentation in his venire was due to the systematic exclusion of young adults from the jury pools. Duren v. Missouri, 439 U.S. at 366, 99 S.Ct. at 669. In Duren, the Supreme Court held that proof of intentional or purposeful discrimination is not required to show systematic exclusion. A large discrepancy occurring over a sustained period of time where there is an opportunity for arbitrary selection is sufficient to demonstrate that the exclusion of the underrepresentation is systematic — that is, inherent in the particular jury selection process utilized. 439 U.S. at 366, 99 S.Ct. at 669. See also United States v. Benmuhar, 658 F.2d 14, 19 (1st Cir.1981), cert. denied, 457 U.S. [990]*9901117, 102 S.Ct. 2927, 73 L.Ed.2d 1328 (1982); Murrah v. State of Arkansas, 532 F.2d 105 (8th Cir.1976); Smith v. Yeager, 465 F.2d 272, 280 (3d Cir.), cert. denied, 409 U.S. 1076, 93 S.Ct. 685, 34 L.Ed.2d 665 (1972) ; Stephens v. Cox, 449 F.2d 657, 659 (4th Cir.1971).
Norfolk County employs the keyman system to select jury venires. Mass.Gen.Laws Ann. ch. 234, § 4 (1970). Under the key-man system, the board of selectmen prepares a list of inhabitants of good moral character, sound judgment, free from legal exceptions, not exempt, and whom the selectmen think qualified to serve as jurors. The list may contain only the names of persons determined to be qualified upon the knowledge of one of its members or after personal appearances or in response to examination by questionnaire under oath. The discretion required on the part of the selectmen in compiling the jury pool lends itself to nonrandom selection. See Note, Federal Court — Juror Selection— Underrepresentation of Young Adults on Juror Source Lists, 19 Wayne L.Rev. 1287 (1973) . While there is no evidence that the exclusion of young people on the jury ve-nire was purposeful or motivated by animus, it is natural, if not inevitable, that a system that is structured, in large part, around the community connections of the members of the board of selectmen would yield venires that underrepresent groups of people that are less likely to be known to the jury selectors than other groups. See Murrah v. State of Arkansas, 532 F.2d at 109. Certainly, the opportunity for intentional or unintentional discrimination exists in Norfolk County where the venire is heavily weighed in favor of adults thirty-five to seventy. Petitioner has presented evidence that the underrepresentation between 1975 and 1978 was even more severe. We find that petitioner has made a prima facie case of systematic exclusion of young adult jurors in Norfolk County.
III. BURDEN UPON THE STATE
Once a petitioner establishes a prima fa-cie violation the burden shifts to the state to justify the underrepresentation. Duren v. Missouri, 439 U.S. at 367, 99 S.Ct. at 670. While the states remain free to prescribe relevant qualifications and reasonable exemptions for jury selection, the right to a proper jury cannot be overcome on merely rational grounds. Rather, it requires that a significant state interest be manifestly and primarily advanced by those aspects of the selection process that result in the disproportionate exclusion of a distinctive group. Duren v. Missouri, 439 U.S. at 367-78, 99 S.Ct. at 670-75; LaRoche v. Perrin, 718 F.2d at 503; United States v. Benmuhar, 658 F.2d at 19. Here, the state has advanced no explanation for the underrepresentation.
IV. RELEVANCE OF THE COMPOSITION OF THE PETIT JURY
The state claims, and the district court below held, that even if petitioner proves a constitutional violation of the cross-section requirement in jury selection, petitioner is precluded from mounting a successful challenge because his ultimate jury contained two persons out of seven under thirty-five, a near perfect representation of young adults based on the population of Norfolk County. Petitioner claims he was nonetheless affected by the underrepresentation on the jury rolls because he did not have the same opportunity to exercise a challenge against either of the young jurors that a proper venire would have offered.
The Supreme Court has clearly stated that it is unnecessary for a petitioner to show that he was in any.way prejudiced by a wrongful exclusion or that he was a member of the excluded class. Taylor v. Louisiana, 419 U.S. at 527, 95 S.Ct. at 696; Peters v. Kiff 407 U.S. at 503, 92 S.Ct. at 2169; Thiel v. Southern Pacific Co., 328 U.S. at 225, 66 S.Ct. at 988. This is due, at least in part, to the multiple purposes of the cross-section requirement.
Trial by jury presupposes a jury drawn from a pool broadly representative of the community as well as impartial in a specific case____ The broad representative character of the jury should be main-[991]*991tamed, partly as assurance of a diffused impartiality and partly because sharing in the administration of justice is a phase of civic responsibility.
Taylor v. Louisiana, 419 U.S. at 530, 95 S.Ct. at 697 (citing Thiel v. Southern Pacific Co., 328 U.S. at 227, 66 S.Ct. at 989 (Frankfurter, J., dissenting)). In Thiel v. Southern Pacific, the Court reversed a verdict because working people paid on a daily basis had been excluded, finding that “it ... is immaterial that the jury which actually decided the factual issue in this case was found to contain at least five members of the laboring class.” Thiel, 328 U.S. at 225, 66 S.Ct. at 988. The Court stated:
The evil lies in the admitted wholesale exclusion of a large class of wage earners in disregard of the high standards of jury selection. To reassert those standards, to guard against the subtle undermining of the jury system, requires a new trial by a jury drawn from a panel properly and fairly chosen.
Id.
In the eases after Thiel, the Court has continued to focus on the composition of the jury pool and not the specific jury drawn by the challenger. For instance, in Peters v. Kiff, 407 U.S. at 502-03, 92 S.Ct. at 2168-69, the Court found “[ijllegal and unconstitutional jury selection procedures cast doubt on the integrity of the whole judicial process. They create the appearance of bias in the decision of individual cases, and they increase the risk of actual bias as well.” See also Taylor v. Louisiana, 419 U.S. at 530, 95 S.Ct. at 697. Recently, in Duren v. Missouri, 439 U.S. at 366, 99 S.Ct. at 669, the Court reaffirmed its focus: “[I]n order to establish a prima facie case, it was necessary for petitioner to show that the underrepresentation of women, generally and on his venire, was due to their systematic exclusion in the jury-selection process.” The Court did not require that the petitioner state or be able to show that the jury drawn was also underrepresented, nor did the Court suggest that the state' could defeat a challenge by showing that the actual jury, by happenstance, contained a representative number of the underrepresented group. Indeed, in his dissent in Duren, Justice Rehnquist stated: “[UJnder the majority’s fair cross section analysis the underrepresentation of women on venires in Jackson County, Missouri, would entitle Duren to reversal of his conviction even if the jury chosen in his case had been composed of all women.” Id. at 373, n. 1, 99 S.Ct. at 673, n. 1 (Rehnquist, J., dissenting).
Although the Court has not had occasion to directly consider a case where chance has, against great odds, dealt the defendant a representative jury, the Fifth Circuit has addressed the issue. In Rabinowitz v. United States, 366 F.2d 34 (5th Cir.1966) (in banc), the court found that departures from federal jury selection procedures that resulted in a systematic exclusion of blacks from the jury pool could not be cured by proof from the government that the defendant’s jury contained a representative number of blacks. There, the court stated:
The focus of the law is on the list from which the grand jury is drawn and not on the composition of a particular jury or grand jury. The efforts by Congress to broaden the base of the jury system in federal courts was an attempt to do more than improve the administration of justice at the point where it most directly touches members of the legal profession and litigants. It constituted an effort to improve the judicial system where it most directly touches the lives of the average citizen. For many citizens their only contact with the court arises from their service on grand or petit juries.
Even more important is the fact that many citizens will have no direct contact with the administration of justice, but will judge its efficacy on how the judicial process functions. Congress and the courts are aware of this fact. The injury is not limited to the defendant — there is injury to the jury system, to the law as an institution, to the community at large, and to the democratic ideal reflected in the processes of our courts. When the basic jury list was poisoned, the fruits of [992]*992that list were also infected. To cure the infection, it is necessary to start the process anew.
Id. at 59-60 (citations omitted).14 Lower courts have directly held, or in dicta stated, that the proper focus in jury discrimination cases is on the source list from which the jury is chosen and not the composition of any particular jury. Mobley v. United States, 379 F.2d 768, 770 (5th Cir.1967); McGinnis v. M.I. Harris, 486 F.Supp. 750, 756 (N.D.Tex.1980); See also United States v. Zirpolo, 450 F.2d 424, 428-31 (3d Cir.1971) (dicta); Ford v. Hollowell, 385 F.Supp. 1392, 1398 (N.D.Miss.1974) (reaffirming focus on jury list, not composition of particular jury).
In addition to requiring a showing of underrepresentation in the pool or general list, Duren v. Missouri appears to require petitioner to show that the systematic exclusion also infected petitioner’s venire. But there is no suggestion that the composition of a particular jury is relevant. See 439 U.S. at 366, 99 S.Ct. at 669. The Court’s directive in Duren indicates that when the public’s interest in broad participation in the administration of justice is weighed against the public’s interest in enforcing judgments against defendants who in fact receive a representative venire, the scale tips in favor of some specificity. The state claims that the considerations militating in favor of requiring the petitioner to show that his venire was tainted by exclusion also require petitioner to show that the cognizable group was underrepresented on the jury drawn notwithstanding the accepted rule that courts do not look to the particular jury to determine whether systematic exclusion has occurred. The state cites Harris v. Wyrick, 644 F.2d 710 (8th Cir.1981), in support. There, the court did state that Wyrick’s jury was not underrepresented, but the holding is not as the state suggests. The decision in that case rested on a lack of proof of exclusion in defendant’s venire. Id. at 713.
The venire is important because that is the group from which the jurors for the case are chosen. When that venire is tainted, the ability to challenge is affected. The right to challenge jurors both for cause and peremptorily has been perceived to be a critical feature of the right to a jury trial for over two hundred years and the Supreme Court has declared that denial or impairment of a defendant’s right to exercise a peremptory challenge is cause for reversal without a showing of prejudice. Swain v. Alabama, 380 U.S. 202, 212-21, 85 S.Ct. 824, 831-36, 13 L.Ed.2d 759 (1965); Pointer v. United States, 151 U.S. 396, 406, 14 S.Ct. 410, 413, 38 L.Ed. 208 (1894); Lewis v. United States, 146 U.S. 370, 376, 13 S.Ct. 136, 138, 36 L.Ed. 1011 (1892). See also Commonwealth v. Soares, 377 Mass. 461, 387 N.E.2d 499 (1967) (affirming right of peremptory challenge consonant with cross-section requirement). Even when chance has dealt a litigant a relatively representative jury, he is put in the position of having to choose between challenging a member of a grossly underrepresented group that he would like to eliminate from the jury or being tried by a jury that reflects the systematic exclusion of the jury selection process.15 Forcing litigants to elect biased or unrepresentative juries in order to preserve their jury exclusion [993]*993claims does not seem to be a sensible means of implementing the cross-section requirement. Nor is it harmonious with the federal jury selection statute and many state statutes which require a challenge to the composition of the jury venire prior to trial. See, e.g„ 28 U.S.C. § 1867 (1982); Mass.R.Crim.P. 13 (1980).
Moreover, the Supreme Court has repeatedly stated that the defendant’s interest in an impartial trial is not the only consideration involved in the cross-section requirement. See, e.g., Taylor v. Louisiana, 419 U.S. at 530-31, 95 S.Ct. at 697-98; Peters v. Kiff, 407 U.S. at 498-500, 92 S.Ct. at 2166-67. The appearance of fairness and the rights and duties of all groups to take part in the civic responsibility of jury duties have caused the Court to stress the importance of fair and representative selection procedures. Finally, to hold that a litigant is not entitled to a representative jury when the jury venires are drawn from a fair cross-section of the community but that the cross-section requirement can be dispensed with when the dice fall a particular way in an individual case undermines the analytical foundation upon which the right to a jury drawn from a cross-section of the community is based.16 We, therefore, continue to focus on the general venire (jury pool) and the petitioner’s venire in determining whether a jury selection process systematically excludes a cognizable group.
V. PETITIONER’S VENIRE
The state also challenges the sufficiency of the petition because petitioner did not present evidence that young adults were underrepresented on venires at the time of petitioner’s trial. Petitioner claims that the statistics he produced did include the venire from which his jury was selected because Mass.Gen.Laws Ann. ch. 234, § 5 (1970) requires town jury lists to be annually drawn in August. Petitioner’s study included venires drawn in October of 1980 from the August 1980 source list that was used to produce his jury panel in April of 1981. Petitioner also points out that he moved to expand his record in district court to include evidence of the representation of young adults on venires between November 1980 and April 1981. Nonetheless, petitioner’s evidence raises two issues.
In Commonwealth v. Bastarache, decided on December 12, 1980, the Supreme Judicial Court considered a challenge to the composition of jury lists drawn under the keyman system in Massachusetts where young adults were substantially underrepresented on the lists. Although the court did not find a constitutional violation in Bastarache, it expressed concern with the composition of the lists and asked the attorney general to prescribe new jury selection procedures for the compilation of jury lists in those cities and towns not using a random selection process. The court suggested that in some instances jury lists be reconstituted as soon as practicable. Although the juror selection statute which reflects the keyman system has not been amended to secure randomness in the jury selection process, it is possible that the attorney general prescribed procedures which resulted in a recompilation of the jury list in Norfolk County prior to the drawing of petitioner’s venire.17 If such a recompilation occurred, then petitioner’s case must be dismissed; it is grounded on statistics that are not relevant to the selec[994]*994tion procedures in use at the time of his trial.
If the jury list in use at the time of petitioner’s trial in April 1981 was the list drawn in August of 1980, then petitioner must be afforded the opportunity to produce evidence of underrepresentation of young adults on the venire from which his jury was chosen. See Duren v. Missouri, 439 U.S. at 366, 99 S.Ct. at 669; Harris v. Wyrick, 644 F.2d at 713; United States v. Berry, 627 F.2d 193, 196 (9th Cir.1980), cert. denied, 449 U.S. 1113, 101 S.Ct. 925, 66 L.Ed.2d 843 (1981). Although the parties do not indicate whether data on the petitioner’s venire was available at the time of the state court challenge, the petitioner sought to introduce such information in the district court and we assume it is readily available. If the data on petitioner’s venire indicates that young adults age eighteen to thirty-four were significantly underrepresented, then petitioner is entitled to a new trial drawn from a fair cross section of the community. While the Supreme Court has never identified the precise degree of underrepresentation required to meet the prima facie test for systematic exclusion, in Swain v. Alabama, 380 U.S. at 208-09, 85 S.Ct. at 829-30, it stated that a showing of up to 10% underrepresentation of a cognizable group was not enough, standing alone, to show systematic exclusion. While courts have occasionally found systematic exclusion where a cognizable group was underrepresented by less than 10%, the majority of decisions involve underrepresentation in excess of 10%. The percentages of underrepresentation that have been presented to courts have varied widely,18 and courts have generally preferred to examine the underrepresentation statistics on a case-by-ease basis rather than deciding upon a threshold percentage. We also believe that a determination of whether a petitioner’s pool and venire indicate systematic exclusion may turn upon particular factors such as the size of the community, the size of the pool and venire, history of underrepresentation, and opportunity for discrimination in selection. See, e.g., Blackwell v. Thomas, 476 F.2d 443, 446-47 (4th Cir.1973) (considering opportunity for discrimination and degree of exclusion over time); Smith v. Yeager, 465 F.2d 272, 279 (3d Cir.1972) (discussing underrepresentation in connection with opportunity to discriminate).
Therefore, we remand to the district court for a determination of whether (1) petitioner proffered statistical evidence relevant to the composition of his particular jury venire; and, if so, (2) whether young adults were significantly underrepresented on petitioner’s venire.
Remanded.