This case, and the case of Porter v. Freeman,1 also decided today, are class actions alleging discrimination in the selection of persons to serve on grand and petit juries. While the targets of the discrimination differ — the plaintiff class in Dr. Berry alleges racial discrimination while the plaintiff class in Porter claims discrimination on the basis of sex — and the Georgia statutory selection method involved in Dr. Berry is not identical to the Alabama procedure used in Porter, the two cases raise essentially the same question: once it is recognized that the jury lists have been compiled in a discriminatory manner, what remedial measures are necessary to meet the constitutional command that a jury be drawn from a fairly representative cross-section of the community?2 In both suits, the District Court Judges found that the defendant jury commissioners’ revisions of the jury lists were sufficient to meet constitutional and statutory requirements. For the reasons that follow, we hold that in both cases the steps taken to remedy the discrimination were inadequate. We therefore reverse and remand.
We approach both cases from the premise that there is a distinction between proving discrimination by race or sex in the first instance and measuring the adequacy of a remedy for such discrimination once it has been found to exist.3 In Dr. Berry, the *324District Court Judge recognized at the initial hearing that the jury lists were unconstitutionally composed; in Porter, the defendants revised the jury lists shortly after the complaint was filed to increase the number of women and blacks. 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.” Carter v. Jury Comm’n of Greene County, 1970, 396 U.S. 320, 340, 90 S.Ct. 518, 529, 24 L.Ed.2d 549, 563 (emphasis added), quoting Louisiana v. United States, 1965, 380 U.S. 145, 85 S.Ct. 817, 13 L.Ed.2d 709. The Judge in Dr. Berry acted under this duty when he ordered the preparation of new and better lists. However, this duty also sets the standard by which the adequacy of the recompiled lists is to be measured — the elimination of the discriminatory disparities as far as possible. We hold that in approving the recompiled lists, the District Court Judges confused the initial proof of discrimination with the adequacy of the remedy for such discrimination. Accordingly, we reverse and remand for the defendants to compile jury lists that will adequately remedy the discrimination that has tainted the lists in the past.
I.
On May 27, 1976, black and female residents of Peach County, Georgia filed suit against the county jury commissioners, claiming that they had been prevented from serving as jurors because of their race or sex.4 The plaintiffs sought to have the District Court declare that the lists of grand and traverse (petit) jurors were unconstitutionally composed, to enjoin their continued use, and to order new lists prepared.5 On June 1, 1976, the District Court Judge held a hearing to determine whether a preliminary injunction should issue. At this hearing, the defendants presented an order signed by Judges Bell and Morgan of the Superior Court for Peach County, Georgia. Stating that “the [Superior] Court finds as a matter of fact that both [the grand and traverse] Jury boxes are unconstitutionally constituted in that there is under representation of black . . . and female citizens in the Grand Jury box and black citizens in the Traverse Jury box,” the state judges ordered that use of the boxes to select jurors be discontinued and that new boxes be prepared. The Superior Court specified the procedures to be followed in preparing the new boxes:
in making up these jury boxes the Jury Commissioners shall use the method of questionnaires to prospective jurors and shall use the responses to these questionnaires and the said Commissioners shall further be guided by the Statutes of the State of Georgia governing their duties as well as guidance set out in State and Federal cases on the subject of composition of Jury Boxes.
[R., vol. I, at 15-16].6 On June 1, the District Court ordered the defendants to comply with the state judges’ order and compile a new jury list.
The jury commissioners prepared new lists and presented them to the District Court with reports detailing the procedures followed and the resulting racial composi*325tion of the jury boxes. On August 27,1976 the Court approved the revised jury lists and dissolved the injunctions. The Court found no merit to the plaintiffs’ objections that blacks continued to be underrepresented and that the new lists were inadequate to eliminate the present effects of past racial discrimination. The plaintiffs press these objections on appeal.
II.
The Jury Boxes: Before And After
The statutory scheme of county government in Georgia provides that the judge of the state superior court for the circuit in which a county is located shall appoint six “discreet persons” to serve as jury commissioners. Ga.Code Ann. § 59-101.7 Using the voters’ list as the major source of names, the commissioners compile a traverse jury list of “intelligent and upright citizens of the county,” supplementing the source only if the resulting jury list is “not a fairly representative cross-section.” From the traverse list, the commissioners then select “a sufficient number of the most experienced, intelligent and upright citizens, not exceeding two-fifths of the whole number” who are 21 years of age and older for the grand jury list. Ga.Code Ann. §§ 59-106, 59 — 201.8 Grand and traverse jurors are then selected at random from these lists.
The plaintiffs bottomed their complaint on the disparity between the number of blacks summoned to jury duty and the number of blacks eligible to serve as jurors in the county population. According to the 1970 census, there were 10,042 citizens in Peach County 18 years of age or older, of whom 5,567 (55.4%) were black, and 8,439 citizens of 21 years of age or older (i. e., eligible to serve on a grand jury), of whom 4,220 (50%) were black. From August 1974 to March 1976,187 persons were summoned for grand jury service; of these, 29, or 15.5%, were black. During this same period, 1,247 persons were summoned for traverse jury service; of these, approximately 195, or 15.6%, were black. On the basis of these figures, the Superior Court and the District Court found that blacks were so underrepresented as to make the jury lists unconstitutional, and ordered that new lists be compiled.
In revising the jury lists, the commissioners sent questionnaires, a letter urging prompt return of the form, and a pread-dressed envelope to all citizens registered to vote in Peach County in the May 5, 1976 presidential preference primary. Of the 9,218 questionnaires sent, 4,378 were ad*326dressed to black voters (48.17%). Only 4,827 questionnaires were completed and returned to the commissioners. The defendants traced the fate of the questionnaires in their report to the District Court.9 1,879 were returned by the Post Office marked “undeliverable” or “addressee unknown.” Of this number, 1,230, or 65%, had been sent to black voters. 2,545 were delivered by the Post Office but not returned to the commissioners; of these, 1,378 or 54% were originally addressed to black voters. From the 4,827 questionnaires completed and returned, the defendants eliminated 1,701 as statutorily unfit or exempt from jury service,10 of whom 592, or 35%, were black. Of the 3,142 on the “reduced” list, 1,180, or 37.6% were black. From this list, the commissioners selected every third numbered juror questionnaire and then every fourth numbered questionnaire until 2,344 names were chosen; these named comprised the traverse jury list. From this list, the 223 persons viewed as the most intelligent and upright were culled for the grand jury list. The resulting traverse list contained the names of 896 black registered voters, or 38.1% of the list; the grand jury list included 92 black names, or 41.2% of the list.
The plaintiffs compare the percentage of black residents in the county of an age to serve as jurors (55.4% aged 18 and over, 50% of an age of 21 and over) with the percentage of blacks on the jury lists (38.1% of the traverse jury list and 41.2% of the grand jury list). Under the method of comparison used in most of the cases in this Circuit, there is in this county a 17.34% differential between blacks in the voting age population and blacks on the traverse list, and an 8.8% differential on the grand jury list. Under a different measure, blacks are underrepresented by 31.28% on the traverse list and 17.60% on the grand jury list.11
The defendants remind us that the Constitution does not require a precise conformity between a jury list and the propor-
*327tionate strength of each identifiable group in the community, see Swain v. Alabama, 1964, 380 U.S. 202, 208, 85 S.Ct. 824, 13 L.Ed.2d 759, and argue that these disparities are within ranges of deviation found acceptable by the Supreme Court and this Court.12 But the plaintiffs do not ask us to hold that the remaining disparities are sufficiently substantial to create a prima facie case of discrimination in the first instance. Rather, they attack the sufficiency of the means used and the results achieved as a remedy for an established and admitted constitutional defect in the jury lists. On this basis, we agree with the plaintiffs and hold that additional corrective steps must be taken.
Ill
Remedying The Remedy
It is clear that a significant disparity between the demographic patterns in a county and the relative percentages of each cognizable, distinct demographic group on the jury lists may warrant corrective action by a federal court, absent a countervailing explanation by the jury commission. Castaneda v. Partida, 1977, 430 U.S. 482, 494, 97 S.Ct. 1272, 51 L.Ed.2d 498, 510-11; Turner v. Fouche, 1970, 396 U.S. 346, 90 S.Ct. 532, 24 L.Ed.2d 567; Foster v. Sparks, 1975, 506 F.2d 805, 810; Broadway v. Cul-pepper, 5 Cir., 1971, 439 F.2d 1253; Preston v. Mandeville, 5 Cir., 1970, 428 F.2d 1392. The issue here is the degree of corrective action required. We have held that “in injunctive cases the Court may exact from officials responsible for the construction of jury lists a high standard of comparability between demographic percentages and those of the jury list.” Broadway v. Cul-pepper, supra, 439 F.2d at 1259; Foster v. Sparks, supra, 506 F.2d at 810. This high standard was not met in the present case.
The District Court Judge apparently based his approval of the revised lists on the premise that even if they would be inadequate to remedy intentional discrimination in jury selection, the plaintiffs had not proved that such intentional discrimination ever occurred. Rather, they had only presented a prima facie case of statistical underrepresentation.13 We find that this is a distinction without a constitutional difference. . The Supreme Court has recently clarified that evidence based solely on statistics is sufficient to establish a prima facie case of discriminatory purpose, without additional positive indicia of discrimination or evidence that the selection procedures followed provided an opportunity to discriminate. Castaneda v. Partida, supra, 430 U.S. at 494-95, 97 S.Ct. 1272.14 In this case, the *328District Court found that the plaintiffs had succeeded in showing significant underrep-resentation of blacks, and the defendants made no effort to rebut this prima facie case of discrimination.15 Because the Court measured the adequacy of the revised lists to remedy the discrimination under an erroneous legal standard, we must reverse and remand. The defendants must take steps to produce a jury list in which the percentage of black jurors more nearly approaches the percentage of eligible blacks residing in the community.
The defendant jury commissioners urge that this result is inconsistent with Thompson v. Sheppard, 5 Cir., 1974, 490 F.2d 830, in which this Court approved a revised jury list as a constitutionally sufficient remedy for unconstitutionally composed jury lists in Dougherty County, Georgia. The jury commissioners in Thompson followed a similar procedure in revising the lists to that employed in this suit. However, several factors distinguish Thompson from the present case. The disparity between age eligible blacks in Dougherty County and those on the revised lists was 11%, as compared to a 17.34% disparity in the case before us. More important, the Court in Thompson stated that “[t]he record is silent as to how much of this disparity was due to exemptions claimed, inability to deliver the questionnaires, and failure to return them.” 490 F.2d at 832. In the present case, the record speaks directly to these points. Of 1,879 questionnaires returned as undeliverable, 1,230 were sent to blacks; of 2,545 unanswered questionnaires, 1,378 were sent to blacks. As a result, 2,608 blacks, or 46.8% of the age eligible black population in the county and 59.6% of all black registered voters were eliminated from possible inclusion on the jury roll. This compares to the 36.7% of the white registered voters similarly lost to the jury lists. We have held that a list only partly usable — here, only 4,794 names out of 9,218 were effectively available — is not a “source reflecting the community from which a fair cross-section may be obtained” unless it can be shown that the available population is comparable to the unusable remnant. Broadway v. Culpepper, supra, 439 F.2d at 1257. No such showing can be made here.
The jury commissioners made no attempt to follow up undelivered or unreturned questionnaires to minimize or reduce this disproportionate result. Until such procedures are implemented, the defendants will be unable to justify disparities remaining on a revised jury list. The Georgia Code provides that if the procedures used result in too wide a disparity, the defendants must find ways to supplement their sources of potential jurors’ names. Ga.Code, § 59-106. We believe that a statutorily and constitutionally sufficient list for the future is unlikely to be produced unless an entirely new list is compiled, subject to the showing that it meets the stringent commands of the Constitution. Broadway v. Culpepper, supra, 493 F.2d 1253. At a minimum, this showing must demonstrate that the defendants added to the procedures previously employed steps to follow up on those whose questionnaires are not delivered and those who fail to return their questionnaires.16
REVERSED and REMANDED.
. Porter v. Freeman, 5 Cir., 1978, 577 F.2d 329.
. E. g., Castaneda v. Partida, 1977, 430 U.S. 482, 97 S.Ct. 1272, 51 L.Ed.2d 498; Taylor v. Louisiana, 1975, 419 U.S. 522, 95 S.Ct. 692, 42 L.Ed.2d 690; Alexander v. Louisiana, 1972, 405 U.S. 625, 92 S.Ct. 1221, 31 L.Ed.2d 536; Carter v. Jury Comm’n of Greene County, 1970, 396 U.S. 320, 90 S.Ct. 518, 24 L.Ed.2d 549; Swain v. Alabama, 1965, 380 U.S. 202, 85 S.Ct. 824, 13 L.Ed.2d 759.
. This distinction between proving discrimination and remedying such discrimination is a familiar one in constitutional litigation. E. g., Milliken v. Bradley, 1977, 433 U.S. 267, 97 S.Ct. 2749, 53 L.Ed.2d 745; Swann v. Charlotte-Mecklenburg Bd. of Educ., 1971, 402 U.S. 1, 15, 91 S.Ct. 1267, 28 L.Ed.2d 554. The equitable principles that guide courts in the formulation of remedies are also familiar: an injunctive decree must “be related to ‘the condition alleged to offend the Constitution’ . . . must indeed be remedial in nature, that is, it *324must be designed as nearly as possible ‘to restore the victims of discriminatory conduct to the position they would have occupied in the absence of such conduct’. . . [and] must take into account the interests of state and local authorities in managing their own affairs, consistent with the Constitution.” Id., 433 U.S. at 280, 97 S.Ct. at 2757, 53 L.Ed.2d at 756.
. No issue of discrimination on the basis of sex is presented on appeal.
. The complaint was based on 42 U.S.C. §§ 1981, 1983, 1985(2), and 1988, and on the Sixth, Thirteenth, and Fourteenth Amendments.
. The Superior Court issued this order after the District Attorney of the Macon Judicial Circuit “brought to the Court’s attention that certain figures have been compiled which indicates that certain classes of citizens, to-wit: blacks and females, are underrepresented in the . . . Jury Box and requested the Court to hear evidence regarding the composition of the Jury boxes.” [R., vol. 1, at 15-16.]
. In Peach County, three of the six jury commissioners were black.
. Ga.Code Ann. § 59-106 provides:
At least biennially, or, if the senior judge of the superior court shall direct, at least annually, the board of jury commissioners shall compile and maintain and revise a jury list of intelligent and upright citizens of the county to serve as jurors. In composing such list the commissioners shall select a fairly representative cross-section of the intelligent and upright citizens of the county from the official registered voters’ list of the county as most recently revised by the county board of registrars or other county election officials. If at any time it appears to the jury commissioners that the jury list ... is not a fairly representative cross-section of the intelligent and upright citizens of the county, they shall supplement such list by going out into the county and personally acquainting themselves with other citizens of the county, including intelligent and upright citizens of any significantly identifiable group in the county which may not be fairly representative thereon.
After selecting the citizens to serve as jurors, the jury commissioners shall select from the jury list a sufficient number of the most experienced, intelligent and upright citizens, not exceeding two-fifths of the whole number, to serve as grand jurors. . Ga.Code Ann. § 59-201 provides:
All citizens of this State, above the age of 21 years, being neither idiots, lunatics, nor insane, who have resided in the county for six months preceding the time of serving, and who are the most experienced, intelligent, and upright persons, are qualified, and liable to serve as grand jurors.
The Supreme Court, in sustaining the statute against a challenge to its facial constitutionality, held that these provisions are not “inherently unfair, or necessarily incapable of administration without regard to race.” Turner v. Fouche, 1970, 396 U.S. 346, 355, 90 S.Ct. 532, 537, 24 L.Ed.2d 567, 576.
. Defendants’ Exhibit F, R., vol. I, at 144-47.
. The commissioners rejected 1,070 for poor health or old age, of whom 326 were black; eliminated 305 teachers and law enforcement officials as statutorily exempt and electing not to serve, of whom 128 were black; struck 29 as illiterate, including 22 blacks; eliminated 301 as residing outside the county, of whom 54 were black, struck 10 who were deceased, including 4 blacks; eliminated 21 with criminal records, of whom 18 were black; struck 35 for incomplete questionnaires, including 23 blacks; and, finally, struck 30 as “not upright, intelligent or jury material,” of whom 12 were black. Defendants’ Exhibit F.
. The first method adopts the so-called absolute disparity standard, which measures representativeness by the difference between the proportion of the cognizable groups in the population from which jurors are drawn and the proportion of the groups on the jury list. See, e. g., Turner v. Fouche, 1970, 396 U.S. 346, 90 S.Ct. 532, 24 L.Ed.2d 567; Swain v. Alabama, 1965, 380 U.S. 202, 85 S.Ct. 824, 13 L.Ed.2d 759; Black v. Curb, 5 Cir., 1972, 464 F.2d 165; Gibson v. Blair, 5 Cir., 1972, 467 F.2d 842; Preston v. Mandeville, 5 Cir., 1976, 428 F.2d 1392. The second counting method adopted by some courts is the comparative disparity standard, which looks to the percentage by which the probability of serving as a juror is reduced for people in the cognizable group. The Supreme Court has referred to this approach without comment, Alexander v. Louisiana, 1972, 405 U.S. 625, 629-30, 92 S.Ct. 1221, 31 L.Ed.2d 536; see also, Stephens v. Cox, 4 Cir., 1971, 449 F.2d 657.
Critics of these approaches appear to favor the comparative standard as the more informative. See, e. g., Gewin, An Analysis of Jury Selection Decision [appendix to the opinion] Foster v. Sparks, 5 Cir., 1975, 506 F.2d 805, 811-37, at 834-35; Kairys, Kadane, and Le-hoczky, Jury Representativeness: A Mandate for Multiple Source Lists, 1977, 65 Calif.L.Rev. 776, 793-99. The lack of critical and judicial concensus has recently been exacerbated by the emergence of a new measure, advocated primarily by the Ninth Circuit. This test considers the effect of any deviation on the absolute numerical composition of grand juries: “to determine substantially we look to people not percentages.” United States v. Kleifgen, 9 Cir., 1977, 557 F.2d 1293, 1297 (where blacks constituted 7% of the population, and 5.1% of the jury list, and the absolute deviation was 2.9%, the comparative deviation 27%, but blacks would be underrepresented by fewer than one juror on 23-member grand jury panel, no relief was warranted); United States v. Potter, 5 Cir., 1977, 552 F.2d 901, 906. This Circuit has followed this standard in one opinion. In United States v. Goff, 5 Cir., 1975, 509 F.2d 825, we approved the District Court’s conclusion that an underrepresentation of 1.4 persons on a 23-person grand jury is not sufficiently substantial to required revised jury lists.
. E. g., Swain v. Alabama, 1964, 380 U.S. 202, 85 S.Ct. 824, 13 L.Ed.2d 759 (10% absolute disparity between age-eligible blacks and blacks on jury list not sufficient to warrant federal court decree); Thompson v. Sheppard, 5 Cir., 1974, 490 F.2d 830 (11% absolute disparity between jury list and proportionate population representation not unconstitutional).
. Transcript of hearing of August 27, 1976, R., Vol. Ill, App. at 59-60.
. In Castaneda, the Supreme Court reiterated its understanding of the impact of Washington v. Davis, 1976, 426 U.S. 229, 96 S.Ct. 2040, 48 L.Ed.2d 597, holding that official action is not deemed unconstitutional solely because it results in a racially disproportionate impact, on jury selection cases;
“It is also clear from the cases dealing with racial discrimination in the selection of juries that the systematic exclusion of Negroes is itself such an ‘unequal application of the law . as to show intentional discrimination . . A prima facie case of discriminatory purpose may be proved as well by the absence of Negroes on a particular jury combined with the failure of the jury commissioners to be informed of eligible Negro jurors in a community, ... or with racially non-neutral selection procedures . . With a prima facie case made out, ‘the burden of proof shifts to the State to rebut the presumption of unconstitutional action by showing that permissible racially neutral selection criteria and procedures have produced the monochromatic result.’ Alexander v. Louisiana, 405 U.S. at 632, 92 S.Ct. 1221, 31 L.Ed.2d 536.” 426 U.S. at 241, 96 S.Ct. [2040] at 2048, 48 L.Ed.2d 597.
430 U.S. at 493-94, 97 S.Ct. at 1279, 51 L.Ed.2d at 510. The nature of the statistical showing sufficient to give rise to an inference of discriminatory purpose had previously been considered in Arlington Heights v. Metropolitan Housing Corp., 1977, 429 U.S. 252, 266 at n. 13, 97 S.Ct. 555, 564 at n.13, 50 L.Ed.2d 450:
“Because of the nature of the jury-selection task, however, we have permitted a finding *328of constitutional violation even when the statistical pattern does not approach the extremes of Yick Wo v. Hopkins, 1886, 118 U.S. 356, 6 S.Ct. 1064, 30 L.Ed. 220; or Gomillion v. Lightfoot, 1960, 364 U.S. 339, 81 S.Ct. 125, 5 L.Ed.2d 110.”
. “[T]here is no question about the fact that defendants admitted your prima facie showing as to the jury box . . . [before it was] . . . completely revised as a result of an Order of this Court.” Id., at 68-69.
. The plaintiffs also contend that, because of the long periods during which blacks were effectively disenfranchised in Georgia, they are still underrepresented on voters lists, and that the defendants therefore had a duty to use sources other than the voter list in recompiling the jury boxes. The plaintiffs recognize that voters lists are almost unanimously favored as the principal, if not the sole, source of names for jury duty under the federal Jury Selection and Service Act of 1968, 28 U.S.A. §§ 1861 et seq., and most state laws. However, they urge us to hold that the presumption of representativeness accorded a jury list drawn randomly from a voter registration list, see Thompson v. Sheppard, supra, 490 F.2d at 833, does not *329apply — or is rebutted — when the jurisdiction is covered by the Voting Rights Act of 1965, 42 U.S.C.A. § 1973b.
Because of our disposition of this case, it is not appropriate for us to pass on the adequacy of voter lists as the sole source for jury lists. This question must await a situation in which no other factor but the use of the voter list accounts for a statistically underrepresentative jury list. However, we would point out that the general judicial approval of voter lists as the sole source of prospective juror names was recently affirmed by the United States Judicial Conference which, in April 1976, recommended to Congress that § 1863(b)(2) of the jury selection and service act be amended to establish a presumption that names of prospective jurors contained in voter lists represent a fair cross-section of the community and to require court finding that the voter lists do not represent such a cross-section before they may be supplemented by other sources of juror names. This recommendation was based on the frequency of challenges to jury lists and on the fact that no federal court of appeals has held that exclusive reliance on voter lists to select jurors is either constitutionally or statutorily deficient.