Randy Gometz was convicted in the United States District Court for the Southern District of Illinois of assaulting a guard in the federal penitentiary at Marion, Illinois and was sentenced to three years in prison. His appeal presents an important question under the Jury Selection and Service Act of 1968, 28 U.S.C. Secs. 1861 et seq.: whether, whenever a large fraction of the persons to whom the clerk of a district court sends juror qualification forms fail to complete and return them, the Act requires the clerk to take steps to assure himself that a jury list composed of those who do respond will be a representative cross-section of the eligible population. The appeal raises a number of other questions as well, but this opinion will discuss only those that relate to jury composition; the others have no possible merit or precedential significance and are therefore being decided in an unpublished opinion pursuant to Circuit Rule 35. The appeal was originally argued before a three-judge panel but the full court granted rehearing en banc on its own initiative before publication of the panel opinion, in accordance with Circuit Rule 16(e).

Gometz's challenge to the selection of the jury that tried him (his challenge to the selection of the grand jury is identical and does not require separate consideration), duly made under 28 U.S.C. Sec. 1867(a), is twofold: (1) the Sixth Amendment was violated because blacks may have been underrepresented on the master jury wheel; (2) the Jury Selection and Service Act entitles him to a hearing to determine whether the qualified jury wheel, composed of names of people who voluntarily filled in and returned their juror qualification forms, was representative. The "master jury wheel," based usually on voter lists, is the source from which names are picked at random to be mailed juror qualification forms (questionnaires). See 28 U.S.C. Secs. 1863(b)(4), 1864(a). The "qualified jury wheel" contains the names of those persons whose responses on the forms indicate that they are qualified for jury service. It is from the qualified wheel that the panels (venires) are selected (again at random) from which grand and petit juries are picked. See 28 U.S.C. Sec. 1866(a). The focus of Gometz's statutory argument is that the people who receive juror qualification forms are not forced to complete and return them, and apparently many do not.

The Sixth Amendment has been interpreted to forbid racial discrimination in the selection of jurors; and although Gometz is white, and the jury that tried him was white (the venire had one black person on it, but the prosecutor used one of his peremptory challenges to exclude him from the jury), he has standing to challenge discrimination against blacks. Peters v. Kiff, 407 U.S. 493, 498-500, 92 S. Ct. 2163, 2166-2167, 33 L. Ed. 2d 83 (1972); see Taylor v. Louisiana, 419 U.S. 522, 526, 95 S. Ct. 692, 695, 42 L. Ed. 2d 690 (1975). But his only evidence of racial discrimination is some minor discrepancies between the population of particular counties in the Southern District of Illinois and the representation of those counties on the master jury wheel. For example, Alexander County, which has a large nonwhite population (31.63 percent), has 2.38 percent of the population of the Southern District but only 2.31 percent of the names on the master jury wheel. But Pulaski County, which has the largest percentage of nonwhites in its population of any county in the district (33.52 percent), is overrepresented on the master wheel: it has 1.71 percent of the district's population but 1.84 percent of the names on the wheel. Some counties with large nonwhite populations are underrepresented on the master wheel, some overrepresented; no clear pattern emerges. Systematic discrimination against counties with large black populations has not been shown.

Gometz makes a different argument with respect to the racial composition of the qualified jury wheel. He conjectures that blacks are overrepresented among those who do not return their juror qualification forms and therefore are underrepresented on the qualified jury wheel, and he wants a hearing in the district court to explore the issue. No evidence supports his conjecture. True, one of the 30 veniremen from among whom Gometz's petit jury was picked was black, which is 3.33 percent, while the population of the Benton Division of the Southern District, from which the venire was drawn, is 3.90 percent black. But the discrepancy is too small to be significant. If there had been two blacks on the venire, blacks would have been overrepresented (6.67 percent), and you cannot get perfect representativeness by cutting an individual juror in two. Cf. United States v. Armstrong, 621 F.2d 951, 956 (9th Cir.1980).

Gometz also argues that a certain type of personality is bound to be underrepresented on the qualified jury wheel when some people do not return their juror qualification forms--namely the "anti-authoritarian" personality, who thumbs his nose at the law and is therefore likely to ignore the requirement (see 28 U.S.C. Sec. 1864(a)) of completing and returning the form. But nothing in the Jury Selection and Service Act suggests that Congress thought that a cross-section of the community, to be fairly representative, must include the anti-social elements in the community. Although for obvious practical reasons people who have "anti-authoritarian" personalities are not forbidden to serve on juries (as convicted felons whose civil rights have not been restored are, see 28 U.S.C. Sec. 1865(b)(5)), neither the letter nor the spirit of the Act requires that they be represented on the qualified jury wheel, let alone in the same proportion that they bear to the population.

Still, a low rate of response to juror questionnaires could lead to the underrepresentation of a group that is entitled to be represented on the qualified jury wheel. Gometz alleges that in the period when the qualified jury wheel was created from which the juries that indicted and convicted him were drawn, 70 percent of those who received jury questionnaires did not respond and the qualified wheel was made up entirely of the names of the 30 percent who did. We do not know whether this allegation is true, but for purposes of this appeal must assume it is. And although, as we have said, it does not appear that blacks were underrepresented on the venire from which Gometz's petit jury was drawn, we do not know the percentage of blacks on the grand jury venire (Gometz says there were six blacks on the venire but does not say what the size of the venire was), the representation of other groups, such as women, in either venire, or the composition of the qualified jury wheel itself.

Gometz does not argue that the Sixth Amendment entitles a criminal defendant to demand that people be coerced to respond to a jury questionnaire if otherwise some group might be underrepresented on the jury that indicts or convicts him. He rests his case entirely on the Jury Selection and Service Act. And the issue under the Act is of course not whether it would be a good thing to follow up on persons who do not respond to a jury questionnaire, or to conduct an investigation of the effect of nonresponse on the representativeness of the jury wheel, or to place jury service on a basis of true conscription, but only whether the Act imposes any duty on the clerk or judges of the district court to follow up on the nonresponders. If it does not, no purpose would be served by the hearing that Gometz requests.

Before the passage of the Jury Selection and Service Act most federal districts used the "key man" system, whereby people believed to have extensive contacts in the community would suggest names of prospective jurors and the qualified jury wheel would be made up from those names. S.Rep. No. 891, 90th Cong., 1st Sess. 10, U.S.Code Cong. & Admin.News 1968, p. 1792 (1967). This system was believed to foster discrimination, and the Act substituted for it the principle of "random selection of juror names from the voter lists of the district or division in which court is held," id. at 15, subject to qualifications not pertinent here, see 28 U.S.C. Sec. 1863(b)(2). Randomness in a statistical sense was not sought, however. See S.Rep. No. 891, supra, at 16 n. 9; United States v. Bearden, 659 F.2d 590, 602 (5th Cir.1981). Congress not only recognized, see S.Rep. No. 891, supra, at 17, but desired, that venires drawn from lists of registered voters would not be random: "Voter lists contain an important built-in screening element in that they eliminate those individuals who are either unqualified to vote or insufficiently interested in the world about them to do so." Id. at 22.

Therefore, although section 1861 makes it "the policy of the United States that all litigants in Federal courts entitled to trial by jury shall have the right to grand and petit juries selected at random from a fair cross section of the community in the district or division wherein the court convenes," and a self-selected sample is not a random sample, it is apparent from the legislative history that the words "selected at random" were not intended literally and that self-screening was not anathematized. (The word "fair" before "cross section" may well have been intended to refer to the pragmatic compromise embodied in the use of voter lists as usually the only source of prospective jurors' names.) Moreover, to the extent that section 1861 is more than merely declaratory of a federal criminal defendant's Sixth Amendment rights, it is a preamble to the Act's specific provisions on qualification, selection, and compensation of jurors; and both committee reports state: "If the voter lists are used and supplemented where necessary, and if the procedures outlined in the bill are otherwise rigorously followed, it is no departure from the standards of the legislation that the qualified jury wheel, the venire or array, or the jury itself, may not reflect a community cross section. The act ... does not require that at any stage beyond the initial source list the selection process shall produce groups that accurately mirror community makeup. Thus, no challenge lies on that basis." S.Rep. No. 891, supra, at 17; H.R.Rep. No. 1076, 90th Cong., 2d Sess. 5, U.S.Code Cong. & Admin.News 1968, p. 1794 (1968). It seems, therefore, that the general requirements of section 1861 are not intended to provide a basis, independent of the specific statutory procedures, for setting aside a conviction.

Those procedures include no requirement that the district court clerk take measures to correct a low response rate, so long as it is high enough to generate enough names for the qualified jury wheel to enable staffing the required number of juries. The Act empowers--not requires--the clerk to pursue those who fail to return their juror qualification forms: "Any person who fails to return a completed juror qualification form as instructed may be summoned by the clerk or jury commission forthwith to appear before the clerk or jury commission to fill out a juror qualification form." 28 U.S.C. Sec. 1864(a) (emphasis added). This power is necessary to keep the jury system from collapsing in any district where widespread noncompliance with the requirement of completing and returning the juror qualification form prevents the clerk from maintaining an adequate qualified jury wheel. But it does not follow that Congress was worried about "no shows" in cases where there was no problem with filling the wheel.

Congress wanted to make it possible for all qualified persons to serve on juries, which is different from forcing all qualified persons to be available for jury service. To the criticism that using voter lists to create the pool from which jurors are selected discriminates against persons who are otherwise qualified to serve on a jury but who do not register and vote, the committees replied: "This is not unfair, however, because anyone with minimal qualifications--qualifications that are relevant to jury service--can cause his name to be placed on the lists simply by registering or voting." H.R.Rep. No. 1076, supra, at 5; S.Rep. No. 891, supra, at 17, U.S.Code Cong. & Admin.News 1968, p. 1795. "Can"--not "must"; and filling out and returning a juror qualification form is no more difficult or burdensome than registering to vote. The statement we quoted earlier from the Senate Report, approving the "built-in screening element" of voter lists that consists in "eliminat[ing] those individuals who are ... insufficiently interested in the world about them," suggests that Congress, rather than disapproving the element of nonrandomness implicit in any form of self-screening, wanted people who lacked a sense of civic obligation not to serve on federal juries, unless the number of "no shows" was so great that the qualified jury wheel could not be filled up. In that event, but only in that event, the clerks could be expected to use the coercive powers that the Act gave them. It is therefore not surprising that the criticisms that have been made of Congress's decision to base jury selection on lists of registered voters, see, e.g., Statutory Note, 4 St. Mary's L.J. 470, 475 (1972), sound a good deal like Gometz's criticisms of limiting jury selection to those persons who complete and return the juror qualification form. Self-screening may indeed reduce the representativeness of federal juries, but Congress crossed that Rubicon when it approved (with narrow exceptions not applicable here) the use of voter lists as the basic source for names of prospective jurors.

Even if we did not have these clues from the structure and legislative history of the Jury Selection and Service Act, we could not assume that it was through oversight that Congress did not command the district court clerks to operate federal jury selection as a true system of conscription. Most district court clerks lack the resources to issue thousands of summonses every year to persons who do not return juror qualification forms, and then to follow up on all the summonses that are ignored; nor should our overworked district judges be required to cite nonresponders for contempt. And anyone with experience as a trial judge knows that a person forced against his will to serve on a jury is apt to be an angry juror and that an angry juror is a bad juror. These considerations increase our reluctance to infer from the provision empowering court clerks to follow up on nonresponders a legislative intent that the power must be used to eliminate possible nonresponse bias. The provision would seem to have, as we have suggested, the more modest function of enabling the jury system to be preserved in the unusual situation where the response rate is so low that there are not enough responders to stock the juries required for trials in the district.

The principal cases that Gometz cites, Broadway v. Culpepper, 439 F.2d 1253 (5th Cir.1971), and Berry v. Cooper, 577 F.2d 322 (5th Cir.1978), did not arise under the Jury Selection and Service Act, and involve the distinct question whether requiring that nonresponders be followed up is a suitable remedy for discrimination established by other evidence. See 439 F.2d at 1254-58; 577 F.2d at 324, 327-28. The controlling principle in these cases is, as stated in Berry, that "Once discrimination is recognized to exist, the District Court has 'not merely the power but the duty to render a decree which will so far as possible eliminate the discriminatory effects of the past as well as bar like discrimination in the future.' " 577 F.2d at 324 (emphasis in original). If the clerk of the Southern District of Illinois had excluded black people from service on federal grand and petit juries in the past, a decree merely enjoining the exclusion and requiring him to send juror qualification forms to all registered voters, black or white, might not "eliminate the discriminatory effects of the past ...." Since by assumption black people would have had little or no history of jury service in the district and would assume they were unwelcome as jurors, they could be expected to have a disproportionately low response rate. And since this would be an effect of the very discrimination that the lawsuit was intended to eliminate, requiring follow-up would be an appropriate part of the decree. The principle of Broadway and Berry has no application to this case: there is no evidence that blacks (or any other group) are underrepresented among juries in the Southern District of Illinois as a result of discrimination.

United States v. Santos, 588 F.2d 1300 (9th Cir.1979), supports our position; it held that the failure to pursue nonresponders did not violate the Act, because "Sec. 1864(a) clearly provides for discretionary follow-up," id. at 1303 (emphasis added). The court did add, however, that "the clerk apparently believed that an 87% response to the questionnaires was sufficient. We do not disagree." Id. And this could mean that the court thought that the clerk's failure to follow up was judicially reviewable under an abuse of discretion standard. (United States v. Carmichael, 685 F.2d 903, 912 (4th Cir.1982), which rejected a challenge to a 70 percent response rate, can be interpreted similarly, although the court's discussion is cryptic.) Whether the court in Santos did mean this is unclear--we have quoted its entire discussion of the question. But the language of section 1864(a) and the legislative history quoted earlier persuade us that Congress did not want to put court clerks under a duty to require that jury questionnaires be completed and returned. A power is not a duty; and courts rarely interfere with the exercise of prosecutorial discretion, as they would be doing, indirectly but effectively, if they exerted pressure on clerks to issue summonses to nonreturners.

The best case for Gometz, though not cited by him, is United States v. Kennedy, 548 F.2d 608, 610-12 (5th Cir.1977), which held that the clerk of the district court had substantially failed to comply with the statutory procedures by his "practice of filling gaps in the month's jury pool with volunteers from last month's jurors ...." Id. at 610. See also United States v. Branscome, 682 F.2d 484, 485 (4th Cir.1982) (per curiam); cf. Government of the Virgin Islands v. Rosado, 699 F.2d 121, 124 (3d Cir.1983). There is a sense in which those who return completed juror qualification forms to the clerk of the Southern District of Illinois are "volunteers," since there is no follow-up of those who fail to respond. But whereas section 1866(f) provides a specific procedure for dealing with an unanticipated shortage of jurors, which was violated by the practice used in Kennedy, there was no shortage here and we do not read the Act to establish a requirement of randomness so rigorous as to forbid any trace of voluntarism. Some voluntarism is, as we have said (and more important as the Senate Report said), implicit in the use of voter lists, since there is no legal duty in this country to vote; and the statute contains a strong although not absolute presumption in favor of voter lists as the basis for juror lists.

Although we do not think the clerk's action is reviewable even for abuse of discretion, we also are not persuaded that the clerk did abuse his discretion in deciding that a 30 percent return rate (if indeed that was the return rate) was good enough in the circumstances. The mailing generated more than 4,000 names for the qualified jury wheel. This is a large sample of the eligible population, bearing in mind that it is the absolute size of a sample rather than its ratio to the population from which it is drawn that determines the sample's reliability. See Freedman, Pisani & Purves, Statistics 332-33 (1980). It is true that a sample, however large, that is limited to those who respond to a mail inquiry is not a random sample of the population surveyed, and, more important, that such a sample may be biased (statistically biased, not necessarily racially or ethnically or sexually biased)--may contain nonresponse bias. See id. at 304, 313, 346. But Congress was not concerned with anything so esoteric as nonresponse bias when it enacted the Jury Selection and Service Act. Against a background of widespread exclusion of the members of particular groups from jury service, see, e.g., S.Rep. No. 891, supra, at 11, Congress decided to broaden eligibility for jury service and to prevent discriminatory exclusions from it but did not attempt to solve the distinct problem of people who refuse to perform their civic duties, though it gave the clerks and judges power to coerce such people.

But it is argued that while some nonresponse may be tolerable, a 70 percent nonresponse rate is not; that we should strain to find somewhere in the Act a prohibition against the district court clerk's inaction in so extreme a case. As we have noted, however, it is not the relative but the absolute size of a sample that determines its reliability, and here the absolute size was great. This shows, moreover, that the principle for which Gometz contends cannot be limited to very low response rates. Suppose the rate had been not 30 but 95 percent; and suppose further, as is the case, that the population from which the sample was drawn--the population of the Benton Division of the Southern District--was less than 5 percent nonwhite. Then if nonresponders were heavily concentrated among nonwhites, as Gometz conjectures may be the case (though the racial composition of his petit jury venire does not support his conjecture), even a 95 percent response rate might produce substantial underrepresentation, or even complete nonrepresentation, of nonwhites. Yet extreme measures would be required to obtain a higher rate. Gometz's counsel said at argument that he would be content with our holding that any response rate above 90 percent is per se adequate under the Jury Selection and Service Act. But the logic of his position belies his concession. A 90 percent rule would be arbitrary. Legislatures are permitted to be arbitrary; courts are not. It is therefore apparent that the hearing Gometz has asked for would be just the first step toward far-reaching changes in the method of federal jury selection that Congress did not contemplate when it passed the Jury Selection and Service Act.

AFFIRMED.

CUDAHY, Circuit Judge, with whom CUMMINGS, Chief Judge, joins, concurring in part and dissenting in part.1

If the discrepancy alleged here were not striking, I would not propose to place limitations on the discretion of the court clerks to ignore nonresponse to jury questionnaires. But, based on the level of nonresponse which has been alleged, the courts have a clear duty, imposed by Congress, to determine the facts and enforce limitations if the facts warrant. The theory adopted by Judge Rosenn's proposed panel opinion was simple: a 30% return rate of questionnaires, coupled with substantial underrepresentation of non-whites or other legally cognizable groups among those who returned questionnaires, would constitute a substantial failure to comply with the Jury Selection and Service Act. Because Gometz alleged these facts in accordance with 28 U.S.C. Sec. 1867, he is entitled to a chance to prove them.

The basic principle of the Jury Selection and Service Act is that, beginning with a list which represents a fair cross section of the community in the district or division where the court convenes, jury panels and grand and petit juries are to be developed by a process of random selection. So long as the basic source list is adequate and the prescribed procedure is thereafter faithfully pursued, there can be no complaint merely because the ultimate panel produced by random selection is somehow unrepresentative in result. The relevant legislative history states:

In short, the act attempts to achieve its cross sectional aim by insuring that the basic source list is adequate [as a community cross section] and that no procedure is employed that would impermissibly diminish the likelihood that a cross section will be attained.

H.R.Rep. No. 1076, 90th Cong.2d Sess. 5 ["House Report"] reprinted in 1968 U.S.CODE CONG. & AD.NEWS 1792, 1794 (emphasis supplied). The majority cites no authority for its conclusion that sending a single questionnaire which results in jury selection from a small minority of the community can under no circumstances be an impermissible deviation from the cross-sectional goal.

With respect to the procedure for selecting the jury panel from a properly constituted (cross-sectional) source list, 28 U.S.C. Sec. 1861 provides,

that all litigants in Federal courts entitled to trial by jury shall have the right to grand and petit juries selected at random from a fair cross section of the community in the district or division wherein the court convenes.

(Emphasis supplied). The majority admits and it is quite indisputable that selection on a basis of who returns a questionnaire and who does not is not random selection. The issue then becomes how egregious a condition of nonresponse must exist before "substantial noncompliance" with the Jury Selection and Service Act has been shown. The majority reaches the rather extreme position that nonresponse can never become a problem and that the lower limit on response is reached only if there are not enough potential jurors to fill the required places on juries. There is no authority whatever for this conclusion.

The majority's discussion of whether the clerks of court must resort to compulsory process to secure responses to questionnaires (and the suggestion of protesting citizens being dragged to the courthouse) is irrelevant. It is unlikely that raising a return rate from 30% to some more acceptable level would require compulsion. A simple reminder would probably be adequate to secure a much better return. And, in any event, the powers and procedures prescribed in 28 U.S.C. Sec. 1864 to secure return of questionnaires suggest that Congress thought they might need to be resorted to under some circumstances. I agree with the majority that these compulsory methods were not provided to secure a 100% return of jury qualification forms but there is certainly nothing to suggest (particularly in light of the statutory powers provided him) that the clerk's discretion to ignore the problem of non-return is unlimited.

The government has not denied that only 30% of the questionnaires sent out have been returned.2 This is far below the level suggested in any reported case which has been brought to my attention. See, e.g., United States v. Santos, 588 F.2d 1300, 1303 (9th Cir.) (87% response), cert. denied, 441 U.S. 906, 99 S. Ct. 1994, 60 L. Ed. 2d 374 (1979); United States v. Carmichael, 685 F.2d 903, 912 (4th Cir.1982) (70% response), cert. denied, --- U.S. ----, 103 S. Ct. 1187, 75 L. Ed. 2d 434 (1983); Broadway v. Culpepper, 439 F.2d 1253, 1257 (5th Cir.1971) (60% response). In the case before us almost three-quarters of the potential jurors have selected themselves out of the process. Gometz merely seeks an opportunity under these conditions to prove at a hearing that substantial nonrepresentativeness had resulted from this extreme degree of nonresponse. If nonrepresentativeness of the sample in some legally cognizable category is found to correlate positively with nonresponse, a nonresponse rate of 70% would almost certainly create a potential for substantial nonrepresentativeness of the persons selected for jury service. That is the simple proposition which Gometz has put, and I believe that he should have an opportunity to prove it if he can. His statistics may be incorrect and he may not be able to carry his burden of proof, but I do not believe that we can prejudge that matter here.

The majority seems to suggest that the choices before this court are either to require 100% response with recourse to every sanction available in order to achieve it or, as noted, to accept a minimal response sufficient only to fill up the requisite juries. The black and white approach of the majority opinion is quite misleading. Under 28 U.S.C. Sec. 1867 the touchstone of a defendant's rights are a "substantial failure" to comply with the provisions of the Jury Selection and Service Act. Congress has delegated enforcement of the Act to the judicial branch through the procedures provided in 28 U.S.C. Sec. 1867 (of which Gometz here has taken advantage) for challenge to jury selection methods. The applicable House Committee Report specifically "leave[s] the definition of 'substantial' to the process of judicial decision." House Report at 5, reprinted in 1968 U.S.CODE CONG. & AD.NEWS at 1794. Therefore, we judges are empowered to formulate standards to guide the clerks in their efforts to comply with the Jury Selection and Service Act. Common sense suggests that we need not require 100% return of jury questionnaires but, when the return rate drops as low as 30% so as to conspicuously impair the principle of randomness, it seems to me that we may reasonably find at least one prong of a "substantial failure to comply." The defendant seeks a hearing to establish the second prong, substantial unrepresentativeness of persons in the jury pool. The majority thus ignores the statutory requirement of "substantial compliance," as defined by the courts. "Substantial compliance" suggests some reasonable standard, less than 100% but appreciably greater than zero.

In United States v. Santos, supra, the court said that "the clerk apparently believed that an 87% response to the questionnaires was sufficient. We do not disagree." Id., 588 F.2d at 1303. This statement does not suggest to me that the Ninth Circuit would regard a response rate of 30% or less as judicially unreviewablewhich seems to be the position of the majority. Similarly, in United States v. Carmichael, supra, the court stated at 685 F.2d at 912:

Finally, appellants argue that the Plan requires a 95% return rate of grand jury questionnaires and because the return rate for this grand jury was less--approximately 70% average--the fair cross section requirement of the Act was not met. A low return rate of questionnaires without some evidence that such rate caused underrepresentation of some group, does not amount to a substantial violation of the Act. Without other evidence, unfair cross section simply cannot be inferred.

In the case before us, the return rate is not 70%; it is 30%. And, in any event, the Carmichael court does not suggest that even a 70% return is adequate; it merely notes the need for other evidence to establish the existence of an unfair cross section. Here, Gometz does not ask us to find that a 30% return is a per se violation of the Act. He simply seeks a hearing to demonstrate that this extremely low return rate has, in fact, resulted in unrepresentativeness.

In Broadway v. Culpepper, supra, the Fifth Circuit said:

In the posture of this record the showing that this voter list would produce a fair cross-section is itself inadequate. Out of 2,500 names only 550 were effectively available. And assuming, without deciding, that [all] of the 485 living and resident and the 500 returning but not signing the questionnaire were available, the fact remains that 40% could not even be used in the process of selective or random choosing because 1,000 questionnaires were returned as undeliverable. A list--this constitutes the "universe"--which is only 60% useable is hardly a source reflecting the community from which a fair cross-section may be obtained unless there is proof--lacking here--that the composition of the 40% remnant is comparable to the 60% available.

439 F.2d at 1257. As Judge Rosenn noted in the proposed panel opinion,

statistically speaking, the situation here is even worse than the one in Broadway, supra. Plainly, we cannot simply assume that the responding 30% of the population is characteristically and attitudinally a fair cross-section of the population.

There are, as the majority points out, distinctions between Broadway and the instant case, but Broadway 's statistical analysis seems quite unexceptionable and in point.

In principle, there is no difference between the case before us and United States v. Kennedy, 548 F.2d 608 (5th Cir.), cert. denied, 434 U.S. 865, 98 S. Ct. 199, 54 L. Ed. 2d 140 (1977). In that case there was a deficiency in the number of jury prospects for a particular month and the jury clerk sought volunteers from the list of persons serving during an earlier month. Volunteering to serve as occurred in Kennedy, insofar as it affects randomness, is in principle no different than responding to a questionnaire. The Kennedy court said:

We need not speculate as to what sort of biases will be reflected in the jury chosen on the basis of its members' willingness to depart from their daily business and serve as jurors. It is enough to recognize that a substantial variable, not contemplated by the Act's few, narrow categories of qualifications, exemptions, and excuses, has confounded the selection process.

548 F.2d at 612 (footnote omitted). In the case at bar, the 30% of the jury list who responded to questionnaires are, in effect, volunteers. The 70% who failed to respond are, in effect, non-volunteers. Gometz merely seeks an opportunity to show that this highly non-random process has resulted in substantial nonrepresentativeness of the jury wheel.

Even if the language of the Jury Selection and Service Act were susceptible of two interpretations, mine and the interpretation the majority gives it (that Congress was indifferent to randomness in the operation of the questionnaire system), we should opt for the interpretation that effectuates the plain Congressional intent. And there are clear expressions of Congressional commitment to random selection. We should thus decide what constitutes a substantial failure to comply with the Act in light of the overall Congressional purpose of randomness. The majority ignores that purpose, sweeping it under the rug by characterizing any attention to questionnaire return as "conscription." Congress may not have intended a generally applicable system of "conscription." But given the repeated references in the legislative history to the goal of "random selection" there is no doubt that Congress intended some reaction by the clerk to a response rate as extremely low as that alleged in this case.

The majority's references to Gometz' sixth amendment rights are mystifying and lead away from the central problem of Congressional intent. This court has acknowledged that the Act may require a more perfect cross section than the constitution. See United States v. Dellinger, 472 F.2d 340, 365 (7th Cir.1972), cert. denied, 410 U.S. 970, 93 S. Ct. 1443, 35 L. Ed. 2d 706 (1974). Whatever the relationship between the sixth amendment and the Act, Gometz has not made a sixth amendment claim and the majority's discussion of the amendment serves merely to confuse.3 Certainly the majority does not believe that Congress lacks power to prescribe jury selection standards more rigorous than the minimum intended by the framers.

In summary, I do not think that any level of nonresponse constitutes a per se violation of the Act, but a 70% nonresponse coupled with a showing of substantial unrepresentativeness of the jury panel would so far depart from the principles of the Jury Selection and Service Act that a violation would be shown. The only thing that is before us now is whether a hearing should be provided in which the defendant would have the burden of showing the second prong of the requirements for a violation--substantial unrepresentativeness of the resulting panel in a cognizable category. This seems to me to be a minimalist view of our obligations to enforce the Jury Selection and Service Act.

For these reasons, I respectfully dissent with respect to the claim based on nonresponse to questionnaires.